Targeted Immune Modulators for Ulcerative Colitis: Effectiveness and Value Final Evidence Report and Meeting Summary October 16, 2020 Prepared for ©Institute for Clinical and Economic Review, 2020 ICER Staff and Consultants University of Washington Modeling Group Daniel A. Ollendorf, PhD Lisa Bloudek, PharmD, MS Director, Value Measurement & Global Health Senior Research Scientist Initiatives University of Washington Center for the Evaluation of Value and Risk in Health Tufts University Medical Center Josh J. Carlson, PhD, MPH Associate Professor, Department of Pharmacy Rajshree Pandey, PhD, MPH University of Washington Research Lead, Evidence Synthesis Institute for Clinical and Economic Review The role of the University of Washington is limited to the development of the cost-effectiveness model, and Katherine Fazioli the resulting ICER report does not necessarily (Former) Research Lead, Evidence Synthesis represent the views of the University of Washington. Institute for Clinical and Economic Review Rick Chapman, PhD, MS Director of Health Economics Institute for Clinical and Economic Review Pamela Bradt, MD, MPH Chief Scientific Officer Institute for Clinical and Economic Review Steven D. Pearson, MD, MSc President Institute for Clinical and Economic Review None of the above authors disclosed any conflicts of interest. HOW TO CITE THIS DOCUMENT: Ollendorf DA, Bloudek L, Carlson JJ, Pandey R, Fazioli K, Chapman R, Bradt P, Pearson SD. Targeted Immune Modulators for Ulcerative Colitis: Effectiveness and Value; Evidence Report. Institute for Clinical and Economic Review, October 16, 2020. https://icer- review.org/topic/ulcerative-colitis/. DATE OF PUBLICATION: October 16, 2020 Daniel A. Ollendorf served as the lead author for the report, led the systematic review and authorship of the comparative clinical effectiveness section, and wrote the background, other benefits, and contextual considerations sections of the report. Lisa Bloudek and Josh J. Carlson developed the cost-effectiveness model and authored the corresponding section of the report. Rick Chapman developed the potential budget impact analysis and authored Section 8. Laura Cianciolo authored the section on coverage policies and clinical guidelines, managed the timeline and public process, and performed quality controls. Pamela Bradt and Steven D. Pearson provided methodologic guidance on the clinical and economic evaluations. We would also like to thank Serina Herron-Smith, Avery McKenna, David M. Rind, Monica Frederick, Foluso Agboola, and Tom Chen for their contributions to this report. ©Institute for Clinical and Economic Review, 2020 Page i Final Report – Targeted Immune Modulators for UC About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at http://www.icer-review.org. The funding for this report comes from government grants and non-profit foundations, with the largest single funder being the Laura and John Arnold Foundation. No funding for this work comes from health insurers, pharmacy benefit managers, or life science companies. ICER receives approximately 19% of its overall revenue from these health industry organizations to run a separate Policy Summit program, with funding approximately equally split between insurers/PBMs and life science companies. Life science companies relevant to this review who participate in this program include Genentech, Johnson & Johnson, Merck, and Pfizer. For a complete list of funders and for more information on ICER's support, please visit http://www.icer-review.org/about/support/. About CTAF The California Technology Assessment Forum (CTAF) – a core program of ICER – provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. CTAF seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. The CTAF Panel is an independent committee of medical evidence experts from across California, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy. All Panel members meet strict conflict of interest guidelines and are convened to discuss the evidence summarized in ICER reports and vote on the comparative clinical effectiveness and value of medical interventions. More information about CTAF is available at https://icer- review.org/programs/ctaf/. The findings contained within this report are current as of the date of publication. Readers should be aware that new evidence may emerge following the publication of this report that could potentially influence the results. ICER may revisit its analyses in a formal update to this report in the future. The economic models used in ICER reports are intended to compare the clinical outcomes, expected costs, and cost effectiveness of different care pathways for broad groups of patients. Model results therefore represent average findings across patients and should not be presumed to represent the clinical or cost outcomes for any specific patient. In addition, data inputs to ICER models often come from clinical trials; patients in these trials and provider prescribing patterns may differ in real-world practice settings. ©Institute for Clinical and Economic Review, 2020 Page ii Final Report – Targeted Immune Modulators for UC In the development of this report, ICER's researchers consulted with several clinical experts, patients, manufacturers, and other stakeholders. The following clinical expert provided input that helped guide the ICER team as we shaped our scope and report. Dr. Siddharth Singh is not responsible for the final contents of this report or should be assumed to support any part of this report, which is solely the work of the ICER team and its affiliated researchers. For a complete list of stakeholders from whom we requested input, please visit: https://icer- review.org/wp-content/uploads/2019/09/ICER_UC_Key_Stakeholder_List_052620.pdf Expert Reviewers Siddharth Singh, MD Assistant Professor of Medicine University of California, San Diego Health Dr. Siddharth Singh has received research grants from AbbVie, and consulting fees from AbbVie, Takeda, Pfizer, AMAG Pharmaceuticals. ©Institute for Clinical and Economic Review, 2020 Page iii Final Report – Targeted Immune Modulators for UC Table of Contents Executive Summary ............................................................................................................................ ES1 Background .................................................................................................................................... ES1 Comparative Clinical Effectiveness ................................................................................................ ES4 Long-Term Cost Effectiveness ...................................................................................................... ES12 Potential Other Benefits and Contextual Considerations............................................................ ES21 Health-Benefit Price Benchmarks ................................................................................................ ES22 Potential Budget Impact .............................................................................................................. ES24 1. Introduction ....................................................................................................................................... 1 1.1 Background .................................................................................................................................. 1 1.2 Scope of the Assessment ............................................................................................................. 3 1.3 Definitions .................................................................................................................................... 7 1.4 Potential Cost-Saving Measures in Ulcerative Colitis .................................................................. 9 2. Patient Perspectives......................................................................................................................... 10 2.1 Methods ..................................................................................................................................... 10 2.2 Impact on Patients ..................................................................................................................... 11 3. Summary of Coverage Policies and Clinical Guidelines ................................................................... 16 3.1 Coverage Policies ....................................................................................................................... 16 3.2 Clinical Guidelines ...................................................................................................................... 19 4. Comparative Clinical Effectiveness .................................................................................................. 23 4.1 Overview .................................................................................................................................... 23 4.2 Methods ..................................................................................................................................... 23 4.3 Results ........................................................................................................................................ 26 4.4 Summary and Comment ............................................................................................................ 64 5. Long-Term Cost Effectiveness .......................................................................................................... 68 5.1 Overview .................................................................................................................................... 68 5.2 Methods ..................................................................................................................................... 68 5.3 Results ........................................................................................................................................ 84 5.4 Summary and Comment .......................................................................................................... 102 6. Potential Other Benefits and Contextual Considerations .............................................................. 104 ©Institute for Clinical and Economic Review, 2020 Page iv Final Report – Targeted Immune Modulators for UC 6.1 Potential Other Benefits .......................................................................................................... 105 6.2 Contextual Considerations ....................................................................................................... 106 7. Health-Benefit Price Benchmarks .................................................................................................. 107 8. Potential Budget Impact ................................................................................................................ 109 8.1 Overview .................................................................................................................................. 109 8.2 Methods ................................................................................................................................... 109 8.3 Results ...................................................................................................................................... 110 8.4 Summary .................................................................................................................................. 112 References ......................................................................................................................................... 124 Appendix A. Search Strategies and Results........................................................................................ 136 Appendix B. Previous Systematic Reviews and Technology Assessments ........................................ 141 Appendix C. Ongoing Studies ............................................................................................................. 147 Appendix D. Comparative Clinical Effectiveness Supplemental Information .................................... 152 Appendix E. Comparative Value Supplemental Information ............................................................. 190 Appendix F. Network Meta-Analysis Supplemental Information ...................................................... 205 ©Institute for Clinical and Economic Review, 2020 Page v Final Report – Targeted Immune Modulators for UC List of Acronyms and Abbreviations Used in this Report ACG American College of Gastroenterology AE Adverse event AHRQ Agency for Healthcare Research and Quality BCBS Blue Cross Blue Shield CI Confidence interval CMS Centers for Medicare and Medicaid Services EIM Extraintestinal manifestation evLYG Equal value of life-years gained FDA United States Food and Drug Administration GDP Gross domestic product HBPB Health-benefit price benchmark HCSC Health Care Service Corporation HCUP Healthcare Cost and Utilization Project HR Hazard ratio IBD Inflammatory bowel disease IBDQ Inflammatory Bowel Disease Questionnaire ICER Institute for Clinical and Economic Review IV Intravenous LCD Local Coverage Determination LY Life year NCD National Coverage Determination NICE National Institute for Health and Care Excellence NIS National Inpatient Sample NMA Network meta-analysis OPUS Observational Post-Marketing Ulcerative Colitis Study PCE Personal consumption expenditure PICOTS Population, Intervention, Comparator, Outcome, Timing, Setting PHC Personal health care PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses PUCAI Pediatric Ulcerative Colitis Activity Index QALY Quality-adjusted life year RCT Randomized controlled trial SF-36 Short Form Health Survey SPEC Specialty Drug Evidence and Coverage TIM Targeted immune modulator TNF Tumor necrosis factor UC Ulcerative colitis UHC UnitedHealthcare US United States USPSTF United States Preventive Services Task Force WAC Wholesale acquisition cost WPAI Work Productivity and Activity Impairment Questionnaire WTP Willingness to pay ©Institute for Clinical and Economic Review, 2020 Page vi Final Report – Targeted Immune Modulators for UC Executive Summary Background Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) that affects the mucosa, the innermost lining of the intestinal wall in the large bowel (i.e., the colon and rectum).1 The disease causes long-lasting inflammation and ulcers in the digestive tract and is typically marked by periods of remission and recurrence of symptoms. Symptoms may include frequent diarrhea, sometimes with blood or pus, abdominal and/or rectal pain, weight loss, and fatigue.2 When the disease affects children, it can have a detrimental impact on growth, nutritional status, and psychosocial development.3 It is estimated that approximately 900,000 individuals in the United States (US) have UC.4 The economic burden of UC is significant, ranging between an estimated $15-32 billion per year.5 The management of UC in adults is dependent on the severity of symptoms. The goal of treatment is to induce a clinical response to treatment (as evidenced by a reduction of the disease's key symptoms) or effect a complete remission of the symptoms during a short-term (six to 14 weeks) "induction" phase of treatment, and maintain response or remission via long-term "maintenance" therapy, often at a lower dose. Colectomy (surgical removal of the colon) may be considered in patients whose disease does not respond to maximal medical management. In patients with mild disease, local or topical use of aminosalicylates may induce and maintain remission. Once symptoms become moderate-to-severe, however, the use of oral or ileal/colonic preparations of budesonide as well as systemic corticosteroids is typically warranted.6 Those whose disease does not respond to or recurs despite systemic therapy are candidates for a number of targeted immune modulators (TIMs) to induce and/or maintain remission. These agents affect a number of different targets on the inflammatory cascade associated with UC and are summarized in Table ES1 on the following page. ©Institute for Clinical and Economic Review, 2020 Page ES1 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES1. TIMs for the Treatment of UC Estimated Treatment Brand Name Route Dose Annual Cost* TNF Inhibitors 160 mg on day 1, then 80 mg 2 weeks later, Adalimumab Humira® Injection $46,933 then 40 mg EOW 200 mg at week 0, 100 mg at week 2, then Golimumab Simponi® Injection $41,332 100 mg EOW Infliximab Remicade® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $14,614 Infliximab-abda Renflexis® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $13,883 Infliximab-dyyb Inflectra® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $13,451 JAK Inhibitor Tofacitinib Xeljanz® Oral 10 mg BID for 8 weeks, then 5 mg BID $35,506 IL-12/23 Inhibitor Weight-based IV dose (≤55 kg: 260 mg, >55 Infusion and kg to 85 kg: 390 mg; >85 kg: 520 mg) before Ustekinumab Stelara® $91,609 injection administering 90 mg SC at week 8, then q8w α4β7 Integrin Inhibitor Vedolizumab Entyvio® Infusion 300 mg at 0, 2, and 6 weeks, then q8w $44,224 BID: twice daily, EOW: every other week, IL: interleukin, IV: intravenous, JAK: Janus kinase, kg: kilogram, mg: milligram, q8w: every 8 weeks, TNF: tumor necrosis factor *Estimated based on average net price to payers per maintenance year (obtained from SSR Health, LLC) for oral and subcutaneously administered products and Medicare ASP pricing for IV-administered products. Potential Cost-Saving Measures in Ulcerative Colitis Stakeholder feedback on potential cost-saving measures in UC was limited. The Crohn's and Colitis Foundation described repeated use of steroid therapy to be of low value given its potential for serious adverse effects, and also mentioned insurance-mandated step therapy, in which patients are required to try agents other than the one they or their clinician prefer, even if the use of that medication is contraindicated. Clinical expert input indicated that continued use of aminosalicylates in patients who have been failed by such therapy, and escalated to use of TIMs, is a pervasive and low-value intervention. Patient Perspectives From the beginning of this assessment, we sought input from patients, caregivers, and representatives from patient advocacy organizations on the research design of this review (i.e., the PICOTS framework; Population, Intervention, Comparators, Outcomes, Timing, and Setting). We also sought insight on the patient experience of UC and its treatment, including benefits of treatment that may not be described in the clinical literature, any broader potential other benefits or disadvantages associated with treatments, and contextual considerations related to UC, which ©Institute for Clinical and Economic Review, 2020 Page ES2 Final Report – Targeted Immune Modulators for UC Return to Table of Contents are summarized here and reported in detail in Sections 2 and 6. Information was provided primarily by the Crohn's and Colitis Foundation, which described impacts and challenges associated with living with UC in four distinct categories: heterogeneity of disease presentation, benefits and risks of UC treatment, access to care challenges, and gaps in the current management of UC. As with many chronic diseases, the presentation of symptoms and disease course can vary substantially among patients. In some, the disease course may reflect periods of active disease and remission, while in others, the symptoms are persistent despite escalating medical therapy. A minority of patients may present with a rapidly progressive form of the disease known as fulminant colitis. As noted previously, children as young as five years of age or less may develop UC, with additional complications such as growth failure and delayed puberty. In addition, there are noted but currently poorly understood differences in how racial and ethnic minorities experience UC. Clinical differences in disease presentation by severity as well as patient characteristics have profound impacts on those suffering with UC. Patients with long enough periods of remission may be able to resume normal work, school, or leisure activity, while others with more progressive disease require increasing levels of caregiver support. Response to treatment is also heterogeneous. Treatments that are successful in some patients may not work for others, and patients frequently report loss of response, development of intolerable side effects, and the need to cycle off a medication. For example, patients may respond differently to the tumor necrosis factor (TNF) inhibitors based on drug composition (e.g., mouse- vs. human- derived, possibly as a result of greater potential immunogenicity with the former) as well as clinical considerations (body mass index, disease severity). Some patients do not clinically respond to TNF inhibitor therapy while others develop immunogenicity to a specific agent; a switch outside of class may be indicated for the former, while a switch in-class may be the best course of action for the latter. The overall lack of information and evidence in this area is frustrating to patients as it potentially delays effective treatment, which may negatively impact various aspects of a patient's quality of life. In addition, caution was urged with regard to surgical intervention. Though colectomy may be curative in some patients, there are long-term complications to consider, which may have substantial and distressing implications for a patient's quality of life and activities of daily living. Several challenges with accessing appropriate care were noted. For one, requirements for complex treatment regimens may pose a challenge for physicians and, consequently, patients. Primary care providers and gastroenterologists without specific UC experience may not, for example, appreciate the broader benefits of psychosocial and dietary support for long-term care management. Estimates of average annual patient out-of-pocket costs for UC services may total $2,000 to $4,000 annually, which may pose a substantial burden for some.7 Costs are especially onerous for patients without insurance or those who are unable to work due to their disease. The Crohn's and Colitis Foundation reiterated patient concerns regarding step therapy and its various negative ©Institute for Clinical and Economic Review, 2020 Page ES3 Final Report – Targeted Immune Modulators for UC Return to Table of Contents consequences. Patients are often undertreated if they are failed by multiple biologics, and by the time they gain access to a treatment subject to numerous step therapy protocols, their disease may have progressed considerably, limiting the efficacy of the drug. Other insurance mechanisms may produce additional barriers and sources of stress for UC patients. For example, some UC treatments are covered as a medical benefit and others as a pharmacy benefit, with different criteria for treatment authorization and different expectations for patient financial contribution. Step therapy algorithms typically found in UC are not concordant with authoritative clinical guidelines (see Section 3) and are also not portable; individuals who switch health plans (or whose employer does) often must repeat earlier steps in the treatment algorithm even if those treatments were previously unsuccessful. Finally, stakeholders noted major gaps in the current management of the disease, manifested primarily in a lack of high-quality evidence. There is a lack of head-to-head clinical trials of TIMs, for example, which limits the ability of patients and their families to make fully informed decisions about their treatment goals and desires. It is a source of frustration for many that the one large head-to-head trial that is available-one that demonstrated vedolizumab's superiority to adalimumab-has not resulted in any appreciable change in insurer coverage policy. Moreover, there is a general need for more comparative effectiveness research given the plethora of drug classes, mechanisms of action, routes of administration, and safety profiles now available. It was also noted that patient participants in Crohn's and Colitis Foundation-sponsored focus groups voiced concern with UC symptoms not routinely collected in clinical trials, which tend to focus on stool frequency and rectal bleeding. Beyond these, patients also reported concerns with pain and fatigue, ability to concentrate, and fecal urgency. Further, patients also experience challenges with interpersonal relationships and educational and career goals due to the impact of the disease on activities of daily living. Relatedly, patients often face difficulty when sharing the experience of UC symptoms with friends, family, and colleagues, which may lead to feelings of isolation. Comparative Clinical Effectiveness We systematically reviewed and synthesized existing evidence from clinical studies to assess the comparative clinical effectiveness of TIMs for moderate-to-severe UC. Full PICOTS criteria are described in Section 1.2. We compared the efficacy, safety, and effectiveness of TIMs (adalimumab, golimumab, infliximab, tofacitinib, ustekinumab, and vedolizumab [intravenous]) to ongoing background conventional therapy (i.e., placebo arms of clinical trials) and to each other. Effectiveness was defined primarily in terms of clinical response and remission (summarized here), and secondarily in terms of endoscopic improvement, health-related quality of life, rates of UC hospitalization, and other outcomes (described in detail in Section 4). Based in part on stakeholder feedback, we assessed benefits and harms separately for patients without evidence of prior TIM use (i.e., "biologic-naïve") and those with such evidence ("biologic- ©Institute for Clinical and Economic Review, 2020 Page ES4 Final Report – Targeted Immune Modulators for UC Return to Table of Contents experienced"). Within each of these populations, we also examined findings separately for the induction and maintenance phases of available studies. We identified a total of 19 randomized controlled trials (RCTs) of the TIMs of interest, including one head-to-head trial (VARSITY) of vedolizumab and adalimumab.8 Where feasible, we conducted network meta-analysis (NMA) to combine available direct evidence comparing the TIMs with indirect evidence. Importantly, maintenance phases came from two major categories of study design: so-called "treat-through" trials (in which patients remained on the therapy they were randomized to at the beginning of induction) and "re-randomized" trials (where responders to induction therapy were re-randomized for maintenance therapy). We adjusted data from treat-through trials to more closely resemble re- randomized designs, the more common design employed in our analytic set. Further details on the adjustments made and the conduct of these analyses may be found in Section 4 and Appendix F. Clinical Benefits Biologic-Naïve: Induction A total of 15 RCTs informed the induction phase in the biologic-naïve population. In the VARSITY head-to-head trial comparing the efficacy of vedolizumab and adalimumab, the rate of response was higher with vedolizumab 300 mg compared to adalimumab 160/80 mg at the end of induction (70.1% vs. 49.5%, difference: 20.6; 95% CI for difference: 12.9 to 28.2); however, the rates of achieving remission did not statistically differ.8 In placebo-controlled trials, TIMs generally demonstrated superior rates of response, remission, or both at the end of induction periods ranging from six to 10 weeks in duration. We note that, while tofacitinib has been studied in biologic-naïve patients, it is indicated for use only after TNF inhibitor therapy and was not, therefore, included in our biologic-naïve NMA. Results from the NMA showed all TIMs were more likely to achieve response and remission compared to placebo. Specifically, TIMs were 1.4 to 1.9 times more likely to achieve clinical response (Table ES2) and 1.8 to 3.2 times more likely to achieve remission (Table ES3). Additionally, infliximab and vedolizumab were statistically more likely to achieve response and remission compared to adalimumab. No other statistical differences among TIMs were observed. ©Institute for Clinical and Economic Review, 2020 Page ES5 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES2. Risk Ratios for Response at the End of the Induction Phase in Biologic-Naïve Patients IFX Pooled* 1.07 (0.93, 1.24) VEDO 1.10 (0.93, 1.34) 1.02 (0.84, 1.27) UST 1.15 (1.00, 1.35) 1.08 (0.92, 1.27) 1.05 (0.85, 1.28) GOL Pooled* 1.37 (1.18, 1.62) 1.28 (1.14, 1.45) 1.25 (1.00, 1.53) 1.18 (1.00, 1.42) ADA 1.88 (1.67, 2.12) 1.76 (1.54, 2.02) 1.71 (1.41, 2.06) 1.63 (1.43, 1.86) 1.38 (1.21, 1.56) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg/kg and 10 mg/kg pooled; golimumab 200/100 mg and 400/200 mg pooled. Table ES3. Risk Ratios for Remission at the End of the Induction Phase in Biologic-Naïve Patients IFX Pooled* 1.15 (0.87, 1.54) VEDO 1.21 (0.85, 1.79) 1.05 (0.71, 1.59) UST 1.34 (1.00, 1.80) 1.16 (0.84, 1.62) 1.10 (0.73, 1.63) GOL Pooled* 1.84 (1.39, 2.50) 1.60 (1.29, 1.98) 1.52 (1.00, 2.26) 1.38 (1.00, 1.92) ADA 3.22 (2.60, 3.96) 2.79 (2.18, 3.58) 2.66 (1.86, 3.73) 2.41 (1.89, 3.08) 1.76 (1.38, 2.19) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg/kg and 10 mg/kg pooled; golimumab 200/100 mg and 400/200 mg pooled. Biologic-Naïve: Maintenance Data from 11 RCTs were available with data on maintenance therapy through follow-up timepoints ranging from 52 to 60 weeks. In the VARSITY head-to-head trial, vedolizumab 300 mg every eight weeks had higher rates of remission compared to adalimumab 40 mg at the end of maintenance (34.2% vs. 24.3%, p=0.006).8 Findings from placebo-controlled trials indicated that all TIMs were superior to placebo in achieving clinical response and remission. The biologic-naïve NMA was adjusted for placebo response rates given observed variability across RCTs. All TIMs were more likely to achieve clinical response (Table ES4) and remission (Table ES5) at the end of maintenance compared to placebo, although no statistical differences from placebo were observed for adalimumab or golimumab. Vedolizumab was shown to be more likely to achieve clinical response and remission compared to adalimumab and golimumab. No other statistical differences between TIMs were observed. ©Institute for Clinical and Economic Review, 2020 Page ES6 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES4. Risk Ratios for Response at the End of the Maintenance Phase in Biologic-Naïve Patients UST 1.12 (0.62, 1.77) VEDO Pooled* 1.16 (0.77, 1.54) 1.04 (0.69, 1.58) IFX Pooled* 1.38 (0.72, 2.18) 1.21 (1.05, 1.43) 1.18 (0.74, 1.79) ADA 1.46 (0.75, 2.54) 1.30 (1.01, 1.77) 1.26 (0.77, 2.03) 1.07 (0.81, 1.50) GOL 1.80 (1.13, 2.50) 1.64 (1.16, 2.02) 1.56 (1.10, 2.14) 1.34 (0.94, 1.78) 1.25 (0.79, 1.70) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Table ES5. Risk Ratios for Remission at the End of the Maintenance Phase in Biologic-Naïve Patients UST 1.18 (0.52, 2.26) VEDO Pooled* 1.24 (0.71, 1.83) 1.05 (0.60, 1.86) IFX Pooled* 1.56 (0.64, 3.00) 1.30 (1.06, 1.62) 1.25 (0.67, 2.22) ADA 1.70 (0.68, 3.55) 1.43 (1.01, 2.15) 1.37 (0.71, 2.61) 1.10 (0.75, 1.71) GOL 2.22 (1.17, 3.57) 1.93 (1.22, 2.58) 1.80 (1.13, 2.86) 1.47 (0.92, 2.14) 1.35 (0.74, 2.04) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Biologic-Experienced: Induction We found no clinical trial evidence for golimumab or infliximab in the biologic-experienced population. Seven of the included RCTs measured the efficacy of TIMs in achieving clinical response and remission at the end of the induction phase (six to 14 weeks). In the VARSITY head-to-head trial comparing the efficacy of vedolizumab and adalimumab,8 the rate of response was higher with vedolizumab compared to adalimumab at the end of induction (55.7% vs. 32.1%, difference: 23.6; 95% CI for difference: 8.5 to 38.7); however, the rates of achieving remission did not statistically differ. Placebo-controlled data were somewhat more mixed in the biologic-experienced population. In one available RCT of adalimumab, no statistical differences were observed for either response or remission. Two RCTs of vedolizumab showed contrasting results, with findings favorable for response (but not remission) for one and no statistical differences observed in the other. Treatment group imbalances were reported in this latter trial, however, as a possible explanation ©Institute for Clinical and Economic Review, 2020 Page ES7 Final Report – Targeted Immune Modulators for UC Return to Table of Contents for findings.9 Trials of tofacitinib and ustekinumab reported statistically significantly better rates of response and remission versus placebo. All seven RCTs were included in the NMA. Tofacitinib, ustekinumab, and vedolizumab were more likely to achieve clinical response and remission compared to placebo; there were no statistical differences between adalimumab and placebo (Table ES6 and ES7). Additionally, tofacitinib, ustekinumab, and vedolizumab were more likely to achieve clinical response and remission compared to adalimumab. Table ES6. Risk Ratios for Response at the End of the Induction Phase in Biologic-Experienced Patients UST 1.01 (0.77, 1.33) TOF 1.32 (0.97, 1.87) 1.31 (0.96, 1.84) VEDO 2.01 (1.39, 3.13) 2.00 (1.36, 3.08) 1.53 (1.11, 2.15) ADA 2.11 (1.71, 2.62) 2.11 (1.68, 2.60) 1.61 (1.20, 2.05) 1.05 (0.71, 1.46) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. Table ES7. Risk Ratios for Remission at the End of the Induction Phase in Biologic-Experienced Patients UST 1.02 (0.58, 1.82) TOF 1.75 (0.94, 3.45) 1.74 (0.92, 3.38) VEDO 3.82 (1.88, 8.49) 3.75 (1.80, 8.08) 2.18 (1.22, 3.95) ADA 4.13 (2.71, 6.26) 4.10 (2.62, 6.18) 2.38 (1.38, 3.80) 1.09 (0.57, 1.97) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. Biologic-Experienced: Maintenance Seven RCTs were available with maintenance data in the biologic-experienced population. No statistical differences were noted between vedolizumab and adalimumab in the VARSITY head-to- head trial.8 Statistically significant differences in favor of TIMs were noted in available placebo- controlled RCTs of adalimumab, tofacitinib, ustekinumab, and vedolizumab. All seven RCTs were included in the NMA; however, placebo adjustment was not performed because the statistical model proved unstable regardless of the number of iterations attempted. All TIMs were more likely to achieve clinical response and remission compared to placebo. Specifically, TIMs were 1.9 to 2.4 times more likely to achieve clinical response (Table ES8), and 2.5 to 3.5 times ©Institute for Clinical and Economic Review, 2020 Page ES8 Final Report – Targeted Immune Modulators for UC Return to Table of Contents more likely to achieve remission (Table ES9). No other statistical differences among TIMs were observed. Table ES8. Risk Ratios for Clinical Response at the End of the Maintenance Phase in Biologic- Experienced Patients VEDO Pooled* 1.12 (0.87, 1.55) ADA 1.25 (0.88, 1.86) 1.11 (0.72, 1.72) UST 1.26 (0.88, 1.90) 1.12 (0.70, 1.75) 1.00 (0.63, 1.61) TOF Pooled* 2.40 (1.87, 3.00) 2.14 (1.48, 2.89) 1.92 (1.34, 2.55) 1.92 (1.32, 2.55) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Tofacitinib 5 and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Table ES9. Risk Ratios for Clinical Remission at the End of the Maintenance Phase in Biologic- Experienced Patients VEDO Pooled* 1.19 (0.80, 1.94) ADA 1.41 (0.83, 2.5) 1.17 (0.62, 2.27) UST 1.41 (0.82, 2.59) 1.18 (0.60, 2.31) 1.01 (0.51, 1.98) TOF Pooled* 3.48 (2.43, 5.03) 2.91 (1.72, 4.65) 2.49 (1.49, 3.82) 2.49 (1.46, 3.79) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Tofacitinib 5 and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Harms Severe and serious adverse events were rare during the induction and maintenance phases of available RCTs. Upper respiratory tract infections and headaches were the most reported adverse events across the TIMs. There was no indication of increased rates of serious infections, tuberculosis, and mortality for any of the agents in available RCTs. Data on serious adverse events in available long-term RCT extensions generally showed similar rates to those observed during randomized periods. Section 4 also includes a summary of large comparative observational studies. While some showed an increased risk of serious infection in selected comparisons between TIMs and to conventional therapy, differences in study design, data source, and study quality make it difficult to draw firm conclusions regarding long-term safety. Of note, however, the label for tofacitinib was modified in July 2019 given the results from a long- term clinical trial for safety in rheumatoid arthritis patients.10 Modifications included warnings ©Institute for Clinical and Economic Review, 2020 Page ES9 Final Report – Targeted Immune Modulators for UC Return to Table of Contents related to thrombotic events (deep vein thrombosis, arterial thrombosis, and pulmonary embolism) as well as cardiovascular mortality in patients receiving the 10 mg twice daily dose. We also note that the three TNF inhibitors (adalimumab, golimumab, and infliximab) as well as tofacitinib carry a black box warning in their FDA labels for an increased risk of lymphomas and other malignancies observed in children, adolescents, and/or young adults based on clinical trials and real-world evidence for these TIMs when studied for other indications (e.g., rheumatoid arthritis).11-14 In our evidence base for UC, overall rates of new malignancy were very low (<3%) and no study in our sample indicated an increased risk for any TIM. Controversies and Uncertainties While the evidence for each TIM of focus in this review suggests a potential net health benefit in comparison to conventional therapy among patients with moderate-to-severe UC, there are several concerns with trial design and results as well as gaps in evidence that bear mention. First and foremost, a single trial of the 19 RCTs identified included direct evidence comparing the TIMs. Our comparisons were therefore driven almost exclusively by the conduct of NMAs. Our NMAs were limited by several factors, including differences in the definitions of biologic-naïve and biologic-experienced populations. Likewise, the network of evidence was sparse for some of our populations of interest (e.g., maintenance data in the biologic-experienced population), adding uncertainty to our relative estimates of response and remission. There is currently very limited information with which to ascertain the optimal sequence of treatment, as the definition of "experienced" varied across trials. In some cases, the focus was on failure to achieve response alone, while in others, those achieving and losing response were also included. Some agents have no evidence in the populations of interest, posing challenges in the interpretation and application of data. We note other differences in population characteristics and measurement in available trials. While populations were broadly similar with respect to age and baseline disease severity, variation was noted in disease duration as well as use of conventional therapy at baseline. In addition, substantial variation has been noted in definitions of endoscopic factors, histologic features, and serum or fecal biomarkers across UC trials.15 As is often the case with therapies for chronic inflammatory diseases, evidence of long-term safety comes from the conduct of observational studies using large datasets. While such data exist for the more established therapies in our set, there are limited to no long-term data on newer UC therapies, such as tofacitinib, ustekinumab, and vedolizumab. There are uncertainties around not only what has been measured in available UC studies but what has not. Concerns were raised around the extraintestinal manifestations (EIMs) of the disease, but we found very little empiric data on the frequency of these manifestations and, more importantly, the impact of treatment on them. Other concerns raised by patients, such as chronic pain and fatigue as well as fecal urgency, have gone largely unaddressed by available studies. Finally, even though a substantial proportion ©Institute for Clinical and Economic Review, 2020 Page ES10 Final Report – Targeted Immune Modulators for UC Return to Table of Contents of cases of UC are diagnosed in childhood and adolescence, there is essentially a complete absence of robust comparative evidence to inform treatment strategies in this population. Summary and Comment Our evidence ratings were based on a combined evaluation of the clinical benefits and potential harms of TIMs across all four of our populations of interest (i.e., induction and maintenance within the biologic-naïve and biologic-experienced populations). We have opted to rate infliximab-dyyb and infliximab-abda, the two biosimilars to infliximab, as comparable ("C") to the originator product, and so the evidence ratings in Table ES10 involve comparisons to infliximab as a single entity. This rating is based on the Food and Drug Administration's (FDA) determination that the biosimilars are therapeutically equivalent in UC. We also note some context around ratings of the TIMs versus placebo. These were generally "A" (superior) ratings, although they apply only to the biologic-naïve population for golimumab and infliximab. We opted to rate tofacitinib at "B+" (small or superior net benefit) given the recent safety warnings; this also applied only to the biologic- experienced population for which its use is indicated. Table ES10. Summary of ICER Evidence Ratings TIM Comparator Rating Infliximab Infliximab biosimilars C Infliximab Placebo A* Golimumab Placebo A* Tofacitinib Placebo B+† All other TIMs Placebo A Vedolizumab Adalimumab B+ Ustekinumab Adalimumab C+ Infliximab Adalimumab C+* Tofacitinib Adalimumab P/I† Vedolizumab Golimumab C+* All other TIM -- I Comparisons ICER: Institute for Clinical and Economic Review, TIM: targeted immune modulator *Biologic-naïve only. †Biologic-experienced only. The head-to-head VARSITY study showed substantial and statistically significant differences in remission, response, and other measures of health benefit in favor of vedolizumab versus adalimumab. These findings were generally bolstered by the addition of indirect evidence in our NMAs. However, with just one head-to-head trial available, we concluded that there was only moderate certainty of a small or substantial net health benefit for vedolizumab ("B+"). ©Institute for Clinical and Economic Review, 2020 Page ES11 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Other comparisons to adalimumab are based on indirect evidence only. We found the evidence directionally consistent across the four populations for both infliximab (in biologic-naïve patients only) and ustekinumab to indicate a net health benefit that is at least comparable, and likely incremental, relative to adalimumab ("C+"). While we generally concluded the same for tofacitinib (in biologic-experienced patients only), the safety concerns associated with this agent resulted in a promising but inconclusive ("P/I") rating versus adalimumab. In the comparison of vedolizumab and golimumab (among biologic-naïve patients only), NMA findings were directionally in favor of vedolizumab for both induction and maintenance, with results reasonably robust for maintenance. We judged this evidence to suggest a net health benefit for vedolizumab that was a least comparable, and likely incremental, relative to golimumab ("C+"). All other comparisons between TIMs reflected insufficient ("I") evidence of a net health benefit. Long-Term Cost Effectiveness A decision analytic model was developed to estimate the cost effectiveness of TIMs for moderate- to-severe UC in two populations: a biologic-naïve population and a biologic-experienced population. The model compared adalimumab, golimumab, infliximab, infliximab-dyyb, infliximab-abda, tofacitinib, ustekinumab, and vedolizumab to conventional treatment and to each other. The model was structured as a Markov model with eight-week cycles consisting of the health states of active UC, clinical response without remission, clinical remission, post-colectomy (with and without complications), and death. The base-case analysis took a health care sector perspective (i.e., focused on direct medical care costs only), over a lifetime time horizon. Costs and outcomes were discounted at 3% per year. All patients enter the model in an active moderate-to-severe state. At the end of induction, patients with response (both with and without remission) continue to receive the TIM or conventional treatment. Those without response or those who discontinue after initial response begin induction with a subsequent treatment, represented by a market basket of all TIMs with data in a biologic-experienced population. At the end of induction with subsequent treatment, responders continue to receive the subsequent treatment. Patients who do not respond during the induction phase of subsequent treatment discontinue treatment with TIMs and follow the transition probabilities of the conventional treatment arm for the remainder of the model time horizon. A proportion of patients in the active UC state are assumed to opt for colectomy during each cycle. ©Institute for Clinical and Economic Review, 2020 Page ES12 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure ES1. Model Framework UC: ulcerative colitis Efficacy data were derived from the results of the ICER NMAs of RCTs of TIMs in moderate-to- severe UC (one each for biologic-naïve and biologic-experienced), with data from the placebo arms used as a proxy for conventional treatment. We used consistent health state utility values across treatments evaluated in the model and across induction and maintenance. Health state utility values are taken from a published multi-center cross-sectional study of moderate-to-severe UC patients in Australia with mean EQ-5D utility values for patients in remission, active mild UC, and active moderate-to-severe UC.16 Utility for the post-colectomy health state is based on EQ-5D scores from a cross-sectional survey of UC patients in Canada, Australia, and the United Kingdom with a history of colectomy within the prior 10 years.17 The model also includes an increased risk of mortality associated with UC, which is not impacted by treatment, discontinuation for reasons other than lack of efficacy, disutility associated with serious infections, early complications of colectomy, and late complications (chronic pouchitis). The model considered cost of drugs, administration, colectomy procedures, direct costs attributable to health states, and adverse events. In a modified societal perspective scenario, we also considered the cost of lost productivity due to intravenous (IV) administration, colectomy procedure, and indirect costs attributable to health states. For TIMs with oral or subcutaneous modes of administration, we obtained net pricing estimates from SSR Health, LLC.18 For IV- ©Institute for Clinical and Economic Review, 2020 Page ES13 Final Report – Targeted Immune Modulators for UC Return to Table of Contents administered TIMs, we used Centers for Medicare and Medicaid Services Average Sales Prices (ASP) plus 9.5%.19 Model outcomes included total costs, life years (LYs), quality-adjusted life years (QALYs), and equal value of life years (evLYG). Base-Case Results Biologic-Naïve Population Total discounted costs ranged from $434,000 to $545,000 over the lifetime time horizon compared to a total cost of $421,000 for conventional treatment. Discounted life expectancy was 22.08 LYs across all treatments. Projected discounted QALYs for TIMs ranged from 15.60 to 15.68 compared with 15.57 QALYs for conventional treatment. The incremental cost per QALY for TIMs compared to conventional treatment in the biologic-naïve population ranged from $186,000 (infliximab-dyyb) to $1,870,000 (adalimumab). All TIMs had lower or equal LYs and QALYs compared to infliximab at a higher cost with the exception of ustekinumab, which had slightly higher QALYs at a higher cost with a resulting cost per QALY exceeding $2.9 million. Table ES11. Base-Case Results for TIMs and Conventional Treatment: Biologic-Naïve Parameter Total Cost QALYs LYs* Adalimumab $461,000 15.596 22.077 Golimumab $458,000 15.600 22.078 Infliximab $435,000 15.644 22.080 Infliximab-dyyb $434,000 15.644 22.080 Infliximab-abda $434,000 15.644 22.080 Ustekinumab $545,000 15.681 22.081 Vedolizumab $480,000 15.641 22.080 Conventional Treatment $421,000 15.574 22.075 LY: life year, QALY: quality-adjusted life year, TIM: targeted immune modulator Costs rounded to nearest $1,000. *Small differences in LYs across comparators are not displayed due to rounding. ©Institute for Clinical and Economic Review, 2020 Page ES14 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES12. Pairwise Results for TIMs Compared to Conventional Treatment: Biologic-Naïve ICER vs. Incremental Treatment Incremental Cost Incremental QALYs Conventional LYs Treatment Adalimumab $40,710 0.02 <0.01 $1,870,000 Golimumab $37,734 0.03 <0.01 $1,455,000 Infliximab $14,741 0.07 <0.01 $212,000 Infliximab-dyyb $12,936 0.07 <0.01 $186,000 Infliximab-abda $13,606 0.07 <0.01 $195,000 Ustekinumab $123,883 0.11 <0.01 $1,163,000 Vedolizumab $59,597 0.07 <0.01 $887,000 Conventional Treatment Reference Reference Reference Reference ICER: incremental cost-effectiveness ratio, LY: life year, QALY: quality-adjusted life year, TIM: targeted immune modulator Cost per QALY gained rounded to nearest $1,000. Biologic-Experienced Population Total discounted costs ranged from $460,000 to $504,000 over the lifetime time horizon compared to a total cost of $434,000 for conventional treatment. Discounted life expectancy was 22.07 LYs across all treatments. Projected discounted QALYs for TIMs ranged from 15.41 to 15.45 for TIMs compared with 15.39 QALYs for conventional treatment. The incremental cost per QALY for TIMs compared to conventional treatment in the biologic- experienced population ranged from $495,000 (tofacitinib) to $1,885,000 (adalimumab). Treatment with tofacitinib resulted in greater QALYs at a lower cost compared with adalimumab. The cost per additional QALY for ustekinumab and vedolizumab compared with adalimumab was $996,000 and $464,000, respectively. Table ES13. Base-Case Results for TIMs and Conventional Treatment: Biologic-Experienced Parameter Total Cost QALYs LYs* Adalimumab $465,000 15.410 22.070 Tofacitinib $460,000 15.445 22.074 Ustekinumab $504,000 15.449 22.074 Vedolizumab $482,000 15.446 22.073 Conventional Treatment $434,000 15.393 22.070 LY: life year, QALY: quality-adjusted life year, TIM: targeted immune modulator Costs rounded to nearest $1,000. *Small differences in LYs across comparators are not displayed due to rounding. ©Institute for Clinical and Economic Review, 2020 Page ES15 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES14. Pairwise Results for TIMs Compared to Conventional Treatment: Biologic-Experienced ICER vs. Treatment Incremental Cost Incremental QALYs Incremental LYs Conventional Treatment Adalimumab $30,462 0.02 <0.01 $1,885,000 Tofacitinib $25,672 0.05 <0.01 $495,000 Ustekinumab $70,216 0.06 <0.01 $1,252,000 Vedolizumab $47,333 0.05 <0.01 $902,000 Conventional Treatment Reference Reference Reference Reference ICER: incremental cost-effectiveness ratio, LY: life year, QALY: quality-adjusted life year, TIM: targeted immune modulator Cost per QALY gained rounded to nearest $1,000. Sensitivity Analyses To demonstrate effects of uncertainty on both costs and health outcomes, we varied input parameters using available measures of parameter uncertainty (i.e., 95% confidence interval [CI]) or reasonable ranges (±10%) to evaluate changes in cost per additional QALY for each TIM compared to conventional treatment. The tornado charts for vedolizumab in a biologic-naïve population and biologic-experienced population are presented in Figure ES2 on the following page as an example. Across all TIMs, utility values and risk ratios for induction and maintenance were the most influential drivers of model results. The results for other comparators were similar and are presented in the Appendix. ©Institute for Clinical and Economic Review, 2020 Page ES16 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure ES2. One-way Sensitivity Analysis of Vedolizumab versus Conventional Treatment (Cost per QALY) Biologic-Naive Biologic-Experienced bio: biologic, exp: experienced, IV: intravenous, resp: response without remission, RR: risk ratio, UC: ulcerative colitis, vedo: vedolizumab The proportion of probabilistic sensitivity analysis iterations with an incremental cost-effectiveness ratio below $250,000 per QALY gained were <1% for adalimumab, golimumab, vedolizumab, and ustekinumab for the biologic naïve population and <1% for all TIMs in the biologic-experienced population. The cost-effectiveness acceptability curve for infliximab and infliximab biosimilars compared to conventional treatment is presented in Figure ES3; acceptability estimates for the four TIMs with evidence in the biologic-experienced population are presented in Figure ES4. ©Institute for Clinical and Economic Review, 2020 Page ES17 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure ES3. Cost-Effectiveness Acceptability Curve for TIMs Compared to Conventional Treatment: Biologic-Naïve Population Figure ES4. Cost-Effectiveness Acceptability Curve for TIMs Compared to Conventional Treatment: Biologic-Experienced Population ©Institute for Clinical and Economic Review, 2020 Page ES18 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Threshold Analyses The dosing, mode of administration, and frequency of administration of some TIMs differs in the induction period compared with the maintenance period. In order to generate a single annual threshold price across TIMs, threshold pricing was conducted for price per maintenance year. The annual drug cost per maintenance year for each TIM to reach an incremental cost per QALY gained compared to conventional treatment at thresholds of $50,000 per QALY, $100,000 per QALY, and $150,000 per QALY are presented in Tables ES15 and ES16. Annual TIM prices at the $150,000 per QALY threshold ranged from $6,824 (adalimumab, biologic-experienced population) to $16,624 (ustekinumab, biologic-naïve population). We note, however, that these results are highly sensitive to what amount to very minor differences in estimated QALYs. For example, the somewhat higher threshold prices for ustekinumab in the biologic-naïve population are driven by a QALY difference of 0.037 in comparison to the next most-effective TIM, which is the equivalent of less than two additional weeks of life. This difference is itself based on parameter inputs, which are based on point estimate differences from our NMA that were not statistically different across ustekinumab and other TIMs. It is also important to note that the threshold prices for ustekinumab are for the subcutaneous injection used in the maintenance period and are independent of the IV product price used in the induction period. This is why, for example, we see the threshold price in the biologic-experienced population rising faster with increasing cost per QALY thresholds versus other TIMs (e.g., vedolizumab). With these details in mind, the overall results demonstrate that for all TIMs, except infliximab and the infliximab biosimilars, annual net price estimates were far higher than threshold prices all the way up to prices at $150,000 per QALY. Table ES15. Resulting Prices per Maintenance Year for TIMs to Reach Cost per QALY Thresholds for the Biologic-Naïve Population Drug Base-Case Cost $50,000/QALY $100,000/QALY $150,000/QALY Adalimumab $46,933 $4,616 $5,778 $6,941 Golimumab $42,332 $4,991 $6,320 $7,649 Infliximab $14,614 $6,754 $8,813 $10,872 Infliximab-dyyb $13,451 $6,754 $8,813 $10,872 Infliximab-abda $13,883 $6,754 $8,813 $10,872 Ustekinumab $91,609 $9,220 $12,922 $16,624 Vedolizumab $44,224 $7,247 $9,454 $11,662 QALY: quality-adjusted life year, WAC: wholesale acquisition cost Prices rounded to nearest $1,000. ©Institute for Clinical and Economic Review, 2020 Page ES19 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES16. Resulting Prices per Maintenance Year for TIMs to Reach Cost per QALY Thresholds for the Biologic-Experienced Population Drug Base-Case Cost $50,000/QALY $100,000/QALY $150,000/QALY Adalimumab $46,933 $4,512 $5,668 $6,824 Tofacitinib $35,506 $9,429 $12,360 $15,292 Ustekinumab $91,609 $4,523 $8,144 $11,766 Vedolizumab $44,224 $6,738 $8,939 $11,140 QALY: quality-adjusted life year, WAC: wholesale acquisition cost Prices rounded to nearest $1,000. Model Validation Model validation followed standard practices in the field. We tested all mathematical functions in the model to ensure that they were consistent with the report. We also conducted sensitivity analyses with extreme input values to ensure the model was producing findings consistent with expectations. Further, independent modelers tested the mathematical functions in the model as well as the specific inputs and corresponding outputs. Summary and Comment We estimated the cost effectiveness of TIMs in biologic-naïve and biologic-experienced adult patients with moderate-to-severe UC. Patient time spent in health states of active UC, response without remission, and response with remission was summed to provide estimates of life expectancy, quality-adjusted life expectancy, and evLY gained. Annual net health care costs, including net price drug acquisition, administration, adverse events, and colectomy were summed to estimate lifetime costs for TIMs and conventional treatment. Based on these calculations, the cost effectiveness of TIMs was estimated to range from $186,000 to $1,870,000 per QALY in the biologic-naïve population and $495,000 to $1,885,000 per QALY in the biologic-experienced population. In general, the incremental cost of TIMs versus conventional treatment was modest given the annual price of TIMs, the use of subsequent treatments, and a lifetime time horizon. In our model, initial non-responders and those who lose response to a TIM are assumed to discontinue treatment and receive additional treatments. With this assumption, few patients remain on the initial TIM longer than a year. Resulting QALYs were quite similar between TIMs, so even small differences in cost drove the variability in cost per QALY ratios. Adalimumab, golimumab, ustekinumab, and vedolizumab were unlikely to be cost effective in the biologic-naïve population, with incremental cost-effectiveness ratios exceeding $800,000 in the base case. Infliximab and infliximab biosimilars had the lowest cost-effectiveness ratios in the biologic-naïve population but remained above the $150,000 cost per QALY threshold in the base case. ©Institute for Clinical and Economic Review, 2020 Page ES20 Final Report – Targeted Immune Modulators for UC Return to Table of Contents No TIMs were found to be cost effective compared with conventional treatment in the biologic- experienced population. However, tofacitinib resulted in greater QALYs at a lower cost compared with adalimumab. Results for both populations were tested under a variety of assumptions and alternative sources of model inputs, none of which drove the incremental cost per QALY below the threshold of $150,000 per QALY gained for TIMs other than infliximab and infliximab biosimilars in the biologic-naïve population. Potential Other Benefits and Contextual Considerations Our reviews seek to provide information on potential other benefits offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. These elements are listed in the table below and on the following page. Potential Other Benefits Table ES17. Potential Other Benefits Other Benefits Description This intervention offers reduced complexity that will N/A significantly improve patient outcomes. This intervention will reduce important health disparities across racial, ethnic, gender, socio- N/A economic, or regional categories. The benefits of TIMs relative to conventional therapy This intervention will significantly reduce caregiver or may translate into significant and durable periods of broader family burden. clinical remission, allowing patients to resume normal activities and reducing caregiver impact. Novel mechanisms of action provide additional This intervention offers a novel mechanism of action options for patients with UC whose disease has or approach that will allow successful treatment of stopped responding to other TIM classes. In addition, many patients for whom other available treatments available UC therapies include oral, self-injectable, and have failed. infused products; patients tend to have clear preferences for method of delivery. The benefits of TIMs relative to conventional therapy This intervention will have a significant impact on may translate into significant and durable periods of improving return to work and/or overall productivity. clinical remission, allowing patients to resume normal activities and reducing caregiver impact. Other important benefits or disadvantages that should have an important role in judgments of the value of N/A this intervention. ©Institute for Clinical and Economic Review, 2020 Page ES21 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Contextual Considerations Table ES18. Potential Contextual Considerations Considerations Description This intervention is intended for the care of individuals UC poses a significant lifetime burden on quality of with a condition of particularly high severity in terms life, and many patients fear the prospect of surgical of impact on length of life and/or quality of life. intervention and its attendant complications. This intervention is intended for the care of individuals UC poses a significant lifetime burden on quality of with a condition that represents a particularly high life, and many patients fear the prospect of surgical lifetime burden of illness. intervention and its attendant complications. This intervention is the first to offer any improvement N/A for patients with this condition. While the safety profile of the TNF inhibitors has been Compared to "the comparator," there is significant relatively well established in UC and other chronic uncertainty about the long-term risk of serious side inflammatory conditions, these data are sparse for effects of this intervention. newer TIMs, such as tofacitinib, ustekinumab, and vedolizumab. While RCTs are comparatively long in duration in UC Compared to "the comparator," there is significant generally, infliximab remains the only agent with an uncertainty about the magnitude or durability of the FDA indication in children and adolescents, and as long-term benefits of this intervention. such, there is substantial uncertainty about the long- term benefits and risks of other TIMs in these patients. There are additional contextual considerations that should have an important role in judgments of the N/A value of this intervention. Health-Benefit Price Benchmarks The health benefit price benchmark (HBPB) is a price range suggesting the highest price that a manufacturer should charge for a treatment based on the amount of improvement in overall health that patients receive from that treatment when a higher price would cause disproportionately greater losses in health among other patients due to rising overall costs of health care and health insurance. Annual prices of each TIM that would achieve incremental cost-effectiveness ratios of $100,000 and $150,000 per QALY for the biologic-naïve and biologic-experienced populations are presented in Tables ES19 and ES20, respectively. The cost per evLYG price ranges are almost identical to the cost per QALY ranges due to the very similar evLYG and QALYs gained in the base-case analysis, and as such, are not presented separately here. ©Institute for Clinical and Economic Review, 2020 Page ES22 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES19. Annual Cost-Effectiveness Threshold Prices per Maintenance Year for TIMs for the Treatment of UC in the Biologic-Naïve Population Discount from Annual Price at Annual Price at WAC Annual Estimated Annual WAC $100,000 $150,000 to Reach Net Price Threshold Threshold Threshold Prices Adalimumab $72,400 $46,900 $5,800 $6,900 90%-92% Golimumab $75,300 $42,300 $6,300 $7,600 90%-92% Infliximab $27,900 $14,600 $8,800 $10,900 61%-68% Infliximab-dyyb $22,600 $13,500 $8,800 $10,900 52%-61% Infliximab-abda $18,000 $13,900 $8,800 $10,900 40%-51% Ustekinumab $150,400 $91,600 $12,900 $16,600 89%-91% Vedolizumab $43,800 $44,200 $9,500 $11,700 73%-78% WAC: wholesale acquisition cost Prices rounded to nearest $100. Table ES20. Annual Cost-Effectiveness Threshold Prices per Maintenance Year for TIMs for the Treatment of UC in the Biologic-Experienced Population Discount from Annual Price at Annual Price at WAC Annual Estimated Annual WAC $100,000 $150,000 to Reach Net Price Threshold Threshold Threshold Prices Adalimumab $72,400 $46,900 $5,700 $6,800 91%-92% Tofacitinib $57,200 $35,500 $12,600 $15,300 73%-78% Ustekinumab $150,400 $91,600 $8,100 $11,800 92%-95% Vedolizumab $43,800 $44,200 $8,900 $11,100 75%-80% WAC: wholesale acquisition cost Prices rounded to nearest $100. Across both populations, the HBPB ranges for adalimumab (approximately $6,000 to $7,000 per year) would require discounts of approximately 90% to 92% from the wholesale acquisition cost (WAC). Golimumab would require discounts of approximately 90% from WAC, for a HBPB range from $6,300 to $7,600 per year. The HBPB range for infliximab ($8,800 to $10,900) would require discounts of 61% to 68% from WAC, with smaller discounts required for infliximab-dyyb (52% to 61%) and infliximab-abda (40% to 51%). The HBPB range for ustekinumab across both populations ranged from $8,100 to $16,600, representing discounts of 89% to 95%. As noted previously, however, the higher threshold prices for ustekinumab must be viewed in the context of the minor differences among TIMs in summary measures of effectiveness, the use of point estimates from NMAs that are not statistically significant, and the sensitivity of the model to variability in costs. For vedolizumab, the HBPB range across both populations is $8,900 to $11,700, requiring 73% to 80% discounts from WAC. ©Institute for Clinical and Economic Review, 2020 Page ES23 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Potential Budget Impact We used the cost-effectiveness model to estimate the potential total budgetary impact of the recently expanded indication of ustekinumab for prevalent individuals in the US with moderate-to- severe UC. In our estimates of potential budget impact, we used the WAC, estimated net price, and $50,000, $100,000, and $150,000 cost-effectiveness threshold prices that were weighted averages of the threshold prices for the biologic-naïve and biologic-experienced populations eligible for ustekinumab. We did not include the other therapies modeled above in this potential budget impact analysis given their established presence on the market for UC. This potential budget impact analysis includes the estimated number of individuals with UC in the US who would be eligible for treatment with ustekinumab. To estimate the size of the potential candidate populations for treatment, we used an estimate by Turner et al. of the prevalence of individuals with UC in the US of approximately 900,000 patients.4 The Crohn's and Colitis Foundation has reported that approximately 22% of UC patients have moderate-to-severe disease activity in a given year, which would equate to approximately 198,000 patients with moderate-to- severe UC in the US.20 To estimate the proportions of these patients who would be biologic-naïve versus biologic-experienced, we used the weighted average of the baseline distribution of patients in the relevant trials that enrolled a "mixed" population (i.e., both biologic-naïve and biologic- experienced), resulting in an estimate of approximately 55% of patients who were not using biologics and 45% who had previously used biologics. For the purposes of this analysis, we assumed that 20% of these patients would initiate ustekinumab in each of the five years, or 21,780 biologic- naïve patients per year and 17,820 biologic-experienced patients per year. Among patients eligible for ustekinumab, we assumed that patients could be drawn from all other available treatment options for biologic-naïve and biologic-experienced patients, respectively. The average annualized potential budgetary impact when using the list price of ustekinumab was an additional per-patient cost of approximately $36,100 and approximately $15,600 using the net price (Table ES21 on the following page). The weighted-average threshold prices for $50,000 to $150,000 per QALY were estimated to produce cost savings relative to the treatment market basket because of the relatively higher cost offset from the comparator mix ($37,500 per patient), which includes several biologics at their net prices. ©Institute for Clinical and Economic Review, 2020 Page ES24 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table ES21. Annualized Per-Patient Potential Budget Impact over a Five-Year Time Horizon for Ustekinumab in a UC Population Assuming 55% Biologic-Naïve and 45% Biologic-Experienced Average Annual Per Patient Budget Impact At $150,000/ At $100,000/ At $50,000/ At List Price At Net Price QALY Price QALY Price QALY Price Ustekinumab $73,600 $53,100 $28,800 $27,600 $26,400 55% Naïve/45% Experienced $37,500 Market Basket Net Impact $36,100 $15,600 -$8,700 -$9,900 -$11,100 QALY: quality-adjusted life year All annualized costs include drug and non-drug health care costs. Numbers may not sum due to rounding. As shown in Figure ES5, approximately 21% of eligible patients could be treated in a given year without crossing the ICER budget impact threshold of $819 million at the WAC of ustekinumab. Approximately 51% of eligible patients could be treated without crossing the budget impact threshold at its estimated net price. All eligible patients could be treated at the $150,000, $100,000, and $50,000 threshold prices, with potential budget impact estimated to be cost saving across the $150,000 to $50,000 threshold prices. Figure ES5. Potential Budget Impact Scenarios of Ustekinumab versus Market Basket Treatment Mix at Different Acquisition Prices $160,000 WAC $140,000 $120,000 $100,000 Annual Price Net Price $80,000 $60,000 $40,000 $20,000 $0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percentage of Patients Treated Without Crossing BI Threshold Each Year BI: budget impact, WAC: wholesale acquisition cost ©Institute for Clinical and Economic Review, 2020 Page ES25 Final Report – Targeted Immune Modulators for UC Return to Table of Contents CTAF Votes The CTAF Panel deliberated on key questions raised by ICER's report at a virtual public meeting on September 24, 2020. The results of these votes are presented below, and additional information on the deliberation surrounding the votes can be found in the full report. 1) Is the evidence adequate to demonstrate that the net health benefit of vedolizumab is superior to that provided by adalimumab? 1 Yes: 12 votes No: 2 votes 2) Is the evidence adequate to demonstrate that the net health benefit of ustekinumab is superior to that provided by adalimumab? Yes: 0 votes No: 15 votes 3) Is the evidence adequate to distinguish the net health benefit among tofacitinib, ustekinumab, and vedolizumab? Yes: 1 vote No: 14 votes 4) When compared to conventional therapy, does treating patients with TIMs offer one or more of the following potential "other benefits"? These interventions offer reduced complexity that will significantly improve patient outcomes. 2/15 These interventions will significantly reduce caregiver or broader family burden. 9/15 These interventions offer a novel mechanism of action or approach that will allow successful 13/15 treatment of many patients for whom other available treatments have failed. These interventions will have a significant impact on improving patients' ability to return to work 12/15 and/or their overall productivity. There are other important benefits or disadvantages that should have an important role in 2/15 judgments of the value of these interventions. 1 Due to technological issues, one CTAF Panelist was unable to cast a vote for the first question. ©Institute for Clinical and Economic Review, 2020 Page ES26 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 5) Are any of the following contextual considerations important in assessing the long-term value for money of TIMs? These interventions are intended for the care of individuals with a condition of particularly high 12/15 severity in terms of impact on length of life and/or quality of life. These interventions are intended for the care of individuals with a condition that represents a 13/15 particularly high lifetime burden of illness. These interventions are the first to offer any improvement for patients with this condition. 0/15 Compared to conventional therapy, there is significant uncertainty about the long-term risk of 13/15 serious side effects of these interventions. Compared to conventional therapy, there is significant uncertainty about the magnitude or 12/15 durability of the long-term benefits of these interventions. There are additional contextual considerations that should have an important role in judgments 0/15 of the value of these interventions. 6) Given the available evidence on comparative effectiveness and incremental cost effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long-term value for money of treatment at current pricing with infliximab, infliximab-abda, and infliximab- dyyb versus conventional treatment? Low: 3 votes Intermediate: 10 votes High: 2 votes ©Institute for Clinical and Economic Review, 2020 Page ES27 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Key Policy Implications Following its deliberation on the evidence, the CTAF Panel engaged in a moderated discussion with a Policy Roundtable about how best to apply the evidence on TIMs for UC to policy and practice. The Policy Roundtable members included two patients/patient advocates, two clinical experts, two payers, and two representatives from manufacturers. The discussion reflected multiple perspectives and opinions, and therefore, none of the statements below should be taken as a consensus view held by all participants. The top-line policy implications are presented on the following pages, and additional information can be found in the full report. Clinicians, Payers, Manufacturers, and Patient Organizations • The significantly lower prices seen for infliximab and its biosimilars speaks to the important potential for improved value with broader availability and uptake of biosimilar treatment options. All stakeholders should collaborate to ensure that TIM biosimilars have an increasing and comprehensive role in the UC treatment landscape. Manufacturers and Payers • The "bundled rebate" approach, in which rebates are provided at the drug level across all of its possible indications, should be abolished and replaced with an indication- and value- based pricing approach. Payers • Insurance coverage should be structured to prevent situations in which patients are forced to choose a treatment approach on the basis of cost. • Specialty society guidelines and drug labels should be monitored for changes, with coverage policy adjusted accordingly. • Because there are no clear biomarkers or predictors of the success for any given treatment in UC, it is not unreasonable to consider prior authorization criteria in order to manage the costs of expensive medications and negotiate prices for TIMs priced beyond a fair range. However, prior authorization criteria should be based on clinical evidence, specialty society guidelines, and input from clinical experts and patient groups. The process for authorization should be clear and efficient for providers. Patient Organizations • Patient advocacy organizations should be an active voice in noting the potentially negative effects of TIM pricing on patient access. ©Institute for Clinical and Economic Review, 2020 Page ES28 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Specialty Societies • Consensus guidelines should be developed across the major gastroenterology societies, in collaboration with patient groups, to ensure a common voice for UC treatment guidance. Regulators • Given the maturity and longstanding use of several of the TIMs of focus in this review, the FDA should require the inclusion of active control arms in Phase III clinical trials of UC treatments. Researchers • The research community should make a strong commitment to generate real-world evidence that can fill in the gaps from available RCTs and allow for comprehensive comparisons of TIMs. • Further clinical study should be conducted to ascertain the optimal sequencing of TIM therapy in UC. ©Institute for Clinical and Economic Review, 2020 Page ES29 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 1. Introduction 1.1 Background Background Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) that affects the mucosa, the innermost lining of the intestinal wall in the large bowel (i.e., the colon and rectum).1 The disease causes long-lasting inflammation and ulcers in the digestive tract and is typically marked by periods of remission and recurrence of symptoms. Symptoms may include frequent diarrhea, sometimes with blood or pus, abdominal and/or rectal pain, weight loss, and fatigue.2 A number of extraintestinal manifestations (EIMs) have also been associated with UC, including musculoskeletal, ocular, dermatologic, hepatobiliary, and psychological effects.21 When the disease affects children, it can have a detrimental impact on growth, nutritional status, and psychosocial development.3 It is estimated that approximately 900,000 individuals in the United States (US) have UC.4 Most individuals are diagnosed between the ages of 15 and 35.5 The economic burden of UC is significant, ranging between an estimated $15-32 billion per year.5 UC is diagnosed based on the presence of symptoms with confirmation of disease via colonoscopy and biopsy. Other disease processes that may cause similar symptoms, such as infection and cancer, should be excluded.6 UC is typically distinguished from Crohn's disease, another form of IBD, based on diffuse inflammation rather than the focal or patchy patterns typical of Crohn's, and confining of the disease to the colon (i.e., the large intestine), while Crohn's can also affect the small intestine, and particularly the terminal ileum, and often spares the rectum.22 In a number of cases, differential diagnosis is difficult, and 5% to 15% of patients may be diagnosed as "IBD unclassified."22 Management The management of UC in adults is dependent on the severity of symptoms. The goal of treatment is to induce a clinical response to treatment (as evidenced by reduction of the disease's key symptoms) or effect a complete remission of the symptoms during a short-term (six to 14 weeks) "induction" phase of treatment, and maintain response or remission via long-term "maintenance" therapy, often at a lower dose. Colectomy (surgical removal of the colon) may be considered in patients whose disease does not respond to maximal medical management. Symptoms of interest and overall disease status are typically defined using the Mayo Score, a combined clinical and endoscopic tool, and include stool frequency, rectal bleeding, mucosal status, and the physician's global impression. The tool documents clinical response to treatment (significant improvement from baseline), remission (low total score and individual sub-scores), and endoscopic improvement ©Institute for Clinical and Economic Review, 2020 Page 1 Final Report – Targeted Immune Modulators for UC Return to Table of Contents (low mucosa sub-score).23 In patients with mild disease, local or topical use of aminosalicylates may induce and maintain remission. Once symptoms become moderate-to-severe, however, the use of oral or ileal/colonic preparations of budesonide as well as systemic corticosteroids is typically warranted.6 Those whose disease does not respond to or recurs despite systemic therapy are candidates for a number of targeted immune modulators (TIMs) to induce and/or maintain remission. These agents affect a number of different targets on the inflammatory cascade associated with UC and are summarized in Table 1.1 below. TIMs may be used alone or in combination with other systemic immunomodulators, such as azathioprine, to prevent relapse.6 Table 1.1. TIMs for the Treatment of UC Estimated Annual Treatment Brand Name Route Dose Cost† TNF Inhibitors 160 mg on day 1, then 80 mg 2 weeks Adalimumab Humira® Injection $46,933 later, then 40 mg EOW 200 mg at week 0, 100 mg at week 2, Golimumab Simponi® Injection $41,332 then 100 mg EOW Infliximab Remicade® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $14,614 Infliximab-abda Renflexis® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $13,883 Infliximab-dyyb Inflectra® Infusion 5 mg/kg at 0, 2, and 6 weeks, then q8w $13,451 JAK Inhibitor Tofacitinib Xeljanz® Oral 10 mg BID for 8 weeks, then 5 mg BID $35,506 IL-12/23 Inhibitor Weight-based IV dose (≤55 kg: 260 mg, Infusion and >55 kg to 85 kg: 390 mg; >85 kg: 520 Ustekinumab Stelara® $91,609 injection mg) before administering 90 mg at week 8, then q8w α4β7 Integrin Inhibitor Vedolizumab Entyvio® Infusion 300 mg at 0, 2, and 6 weeks, then q8w $44,224 BID: twice daily, EOW: every other week, IL: interleukin, IV: intravenous, JAK: Janus kinase, kg: kilogram, mg: milligram, q8w: every 8 weeks, TNF: tumor necrosis factor *Average including biosimilars infliximab-dyyb (Inflectra®, Pfizer) and infliximab-abda (Renflexis®, Merck). †Estimated based on average net price to payers per maintenance year (obtained from SSR Health, LLC) for oral and subcutaneously administered products and Medicare ASP pricing for IV-administered products. As shown in the above table, available TIMs differ substantially in terms of how treatment is delivered, dosing levels and their frequency, and their estimated annual cost. A subcutaneous injectable form of vedolizumab was under review by the US Food and Drug Administration (FDA) when this review began; however, on December 20, 2019, the manufacturer announced receipt of a complete response letter from the FDA denying the application, with no details given on whether and how the FDA's concerns will be addressed.24 ©Institute for Clinical and Economic Review, 2020 Page 2 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 1.2 Scope of the Assessment The scope for this assessment is described on the following pages using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Settings) framework. Evidence was abstracted from randomized controlled trials (RCTs) (see "Data Extraction and Quality Assessment" in Section 4); high-quality comparative cohort studies were also included, particularly for long-term outcomes and uncommon adverse events. Our evidence review included input from patient advocacy organizations, data from regulatory documents, information submitted by manufacturers, and other grey literature when the evidence met ICER standards (for more information, see https://icer- review.org/methodology/icers-methods/icer-value-assessment-framework-2/grey-literature- policy/). All relevant evidence was synthesized qualitatively or quantitatively. Wherever possible, we sought head-to-head studies of the interventions and comparators of interest. We also combined direct and indirect evidence in network meta-analyses (NMAs) of selected outcomes where data permitted. Based on counsel from patient advocacy groups, clinical experts, and manufacturers, most comparisons are stratified by experience with biologic therapy (i.e., naïve vs. experienced) as well as phase of treatment (i.e., induction of response or remission vs. maintenance following response/remission). Full details regarding the literature search, screening strategy, data extraction, and evidence synthesis are provided in a research protocol published on the Open Science Framework website (https://osf.io/cwyn5/). Analytic Framework The analytic framework for this assessment is depicted in Figure 1.1 on the following page. ©Institute for Clinical and Economic Review, 2020 Page 3 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure 1.1. Analytic Framework AE: adverse event, SAE: serious adverse event, TEAE: treatment-emergent adverse event, UC: ulcerative colitis The diagram above begins with the population of interest on the left. Actions, such as treatment, are depicted with solid arrows, which link the population to outcomes. For example, a treatment may be associated with specific clinical or health outcomes. Outcomes are listed in the shaded boxes: those within the rounded boxes are intermediate outcomes (e.g., clinical remission), and those within the squared-off boxes are key measures of clinical benefit (e.g., health-related quality of life). The key measures of clinical benefit are linked to intermediate outcomes via a dashed line, as the relationship between these two types of outcomes may not always be validated. Curved arrows lead to the adverse events of an action (typically treatment), which are listed within the blue ellipse. Populations The population of focus for the review was adults with moderate-to-severe UC, whose disease has either inadequate response or intolerance to conventional therapy, such as systemic corticosteroids, azathioprine, or mercaptopurine. While controlled comparative data were extremely limited in children (ages six to 17), we nevertheless summarized the available information. Additionally, as noted above, given that outcomes may differ according to prior use of biologic therapy as well as whether TIM use is intended for induction or maintenance, we stratified our comparisons according to these factors in both the synthesis of available evidence as well as the economic evaluation. ©Institute for Clinical and Economic Review, 2020 Page 4 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Other subgroups of interest included age (e.g., ≥65), presence of EIMs (e.g., arthritic symptoms, psychological effects), and other key comorbidities. Interventions The interventions of interest developed with input from clinicians and patient organizations included: • Adalimumab (Humira, AbbVie) • Golimumab (Simponi, Janssen) • Infliximab (Remicade, Janssen) • Infliximab-dyyb (Inflectra, Pfizer) • Infliximab-abda (Renflexis, Merck) • Tofacitinib (Xeljanz, Pfizer) • Ustekinumab (Stelara, Janssen) • Vedolizumab (Entyvio, Takeda), intravenous (IV) formulation We included all FDA-approved biosimilars of originator products that are currently available on the US market. Importantly, our focus was on patient-centric data for UC only in comparisons of biosimilars to originator products; information limited to pharmacodynamics, pharmacokinetics, or other laboratory parameters was not considered. We did not include other FDA-approved biosimilars (e.g., biosimilars for adalimumab) as their entry to the US marketplace has been substantially delayed due to patent litigation. Comparators Based on data availability, we compared the interventions of interest to ongoing background conventional therapy (i.e., placebo arms of clinical trials) and to each other. Outcomes The following outcomes of interest were evaluated: Efficacy • Clinical remission • Clinical response • Endoscopic improvement (often referred to as "mucosal healing") • Steroid-free remission • Health-related quality of life • Work/school productivity ©Institute for Clinical and Economic Review, 2020 Page 5 Final Report – Targeted Immune Modulators for UC Return to Table of Contents • Other patient-reported outcomes • Use of rescue medication • UC-related hospitalization • Surgery • Mortality Safety • Serious adverse events • Adverse events leading to discontinuation • Treatment-emergent adverse events o Infections o Headache o Nausea o Fatigue o Pain o Pharyngitis o Respiratory o Autoimmune o Demyelinating disease o Malignancy o Injection reactions o Development of neutralizing antibodies Timing Evidence on intervention efficacy, safety, and effectiveness was collected from studies testing treatments with at least six weeks' exposure duration. Settings Evidence from all relevant settings was considered, with a focus on outpatient settings as well as ambulatory and hospital-based settings. ©Institute for Clinical and Economic Review, 2020 Page 6 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 1.3 Definitions Clinical Outcome Measures Outcomes of clinical trials of UC commonly include clinical response, clinical remission, and endoscopic improvement. Clinical response is defined as a reduction of greater than or equal to 3 points and greater than or equal to 30% from the baseline in total Mayo Score (see definition below) along with a decrease in the rectal bleeding sub-score of greater than or equal to 1 point or an absolute rectal bleeding sub- score of less than or equal to 1 point. Clinical remission is defined as a Mayo Score of less than or equal to 2 with no individual sub-score greater than 1. Endoscopic improvement is defined as a Mayo endoscopic sub-score of 0 or 1. Corticosteroid-free remission is defined as clinical remission in patients using oral corticosteroids at baseline who have discontinued corticosteroids and are in clinical remission at the end of the study. Delayed response in clinical trials has been defined as clinical response and remission (via partial Mayo Score) achieved by the non-responders to induction therapy. Mayo Score The Mayo Score is a disease activity index developed for assessing the severity of UC. It comprises four sub-scores of 3 points each (stool frequency, rectal bleeding score, mucosal appearance at endoscopy, and physical global assessment). The higher the score (maximum 12 points), the more severe the UC. A Mayo Score between 6 and 12 classifies the disease as moderate-to-severe. Pediatric Ulcerative Colitis Activity Index The Pediatric Ulcerative Colitis Activity Index (PUCAI) is a disease activity index to assess the severity of UC in children. It is comprised of six sub-scores: abdominal pain, rectal bleeding, stool consistency of most stools, number of stools per 24 hours, nocturnal stools, and activity level. A PUCAI score of 35 to 64 is classified as moderate and 65 and above is classified as severe. Inflammatory Bowel Disease Questionnaire The Inflammatory Bowel Disease Questionnaire (IBDQ) is a 32-item questionnaire the measures the overall health-related quality of life in patients with IBD. Scores range from 32 to 224, with higher scores indicating better health-related quality of life.25 Although there is no minimal clinically ©Institute for Clinical and Economic Review, 2020 Page 7 Final Report – Targeted Immune Modulators for UC Return to Table of Contents important difference (MCID) established for patients with UC, the MCID for patients living with Crohn's disease is an improvement of at least 16 points. In addition, data have shown that patients with Crohn's in remission generally have an IBDQ score of at least 170 points.26 Trials in patients with UC have used the thresholds established in patients with Crohn's disease to measure the rates of meaningful improvements in IBDQ score. Short Form Health Survey The 36-item short form survey (also known as SF-36)27 is a generic self-reporting tool to assess functional health and wellbeing. SF-36 consists of 36 questions aggregated across eight domains, namely, physical problems and physical functioning, social functioning, bodily pain, mental health, role limitations due to emotional problems, vitality, and general health perceptions. It captures health-related quality of life with two components, a physical component summary (PCS) and mental component summary (MCS), providing global metrics for physical and mental health, respectively. SF-36 is scored from 0 to 100 and then using the mean scores (mean ± standard deviation [SD]) for the general US population (50±10), the norm-based summary component scores are constructed. Higher summary scores indicate better quality of life. The SF-36 MCID thresholds for clinically meaningful improvement for UC have not been established. However, based on the sample of the general population and Crohn's disease, the range of MCID estimates for the PCS and MCS are between 1.6 and 8.7.28 Work Productivity and Activity Impairment Questionnaire The Work Productivity and Activity Impairment Questionnaire (WPAI) is a six-item questionnaire that collects information on both paid and unpaid work. It measures the impact of health problems on absenteeism, presenteeism, and unpaid activity over the past seven days.29,30 The WPAI has been validated in several conditions, including IBD.29,31 The scoring on the WPAI ranges from 0% (no impairment) to 100% (total loss of productivity), with a decrease in scores indicating improvement, and lower scores signifying little impact of disease on work and activity. EuroQol-5 Dimensions EuroQol-5 Dimensions (EQ-5D) is a generic three- or five-level tool to assess health-related quality of life.32 EQ-5D is used to generate a health utility score that varies from 0 (death) to 1 (perfect health). The tool also includes a visual analogue scale (EQ-5D VAS) that varies from 0 to 100 (worst to best). The MCID for UC has not been established; however, EQ-5D VAS MCID estimates ranging from 4.2 to 14.8 have been reported for Crohn's disease.28 ©Institute for Clinical and Economic Review, 2020 Page 8 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Biosimilar The FDA defines a biosimilar as "a biological product that is highly similar to and has no clinically meaningful differences from an existing FDA-approved reference product."33 1.4 Potential Cost-Saving Measures in Ulcerative Colitis ICER includes in its reports information on wasteful or lower-value services in the same clinical area that could be reduced or eliminated to create headroom in health care budgets for higher-value innovative services (for more information, see https://icer-review.org/final-vaf-2017-2019/). These services are ones that would not be directly affected by therapies for UC (e.g., reduced need for surgical colectomy), as these services will be captured in the economic model. Rather, we seek services used in the current management of UC beyond the potential offsets that arise from a new intervention. During stakeholder engagement and public comment periods, ICER encouraged all stakeholders to suggest services (including treatments and mechanisms of care) currently used for patients with UC that could be reduced, eliminated, or made more efficient. The Crohn's and Colitis Foundation described repeated use of steroid therapy to be of low value given its potential for serious adverse effects. The Crohn's and Colitis Foundation also mentioned insurance-mandated step therapy, in which patients are required to try agents other than the one they or their clinician prefer, even if the use of that medication is contraindicated. Clinical expert input indicated that continued use of aminosalicylates in patients who have been failed by such therapy and escalated to use of TIMs is a pervasive and low-value intervention. Finally, patients may prefer a particular route of administration for therapy, and not taking this into account may reduce adherence to therapy. We also note that a Canadian version of the Choosing Wisely campaign for IBD has published a list of potential low-value services, including the use of steroids for maintenance of remission, prolonged use of IV steroids in the absence of clinical response, and the use of abdominal CT in the acute setting without suspicion of an IBD complication or non-IBD etiology.34 ©Institute for Clinical and Economic Review, 2020 Page 9 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 2. Patient Perspectives 2.1 Methods From the beginning of this assessment, we sought input from patients, caregivers, and representatives from patient advocacy organizations on the research design of this review (i.e., the PICOTS framework; Population, Intervention, Comparators, Outcomes, Timing, and Setting). We also sought insight on the patient experience of UC and its treatment, including benefits of treatment that may not be described in the clinical literature, any broader potential other benefits or disadvantages associated with treatments, and contextual considerations related to UC, the details of which are reported in this section as well as in Section 6. We received input on the patient and caregiver perspective from the national Crohn's and Colitis Foundation, which is both a patient advocacy organization and a major sponsor of IBD research. The Crohn's and Colitis Foundation is not responsible for the final content of ICER's report, and it should not be assumed that they support any part of the report, which is solely the work of the ICER team and affiliated researchers. The Crohn's and Colitis Foundation provided input during the following contacts with ICER: • Submission of a letter with both research considerations and patient/caregiver concerns during the open input period • Conference call discussion with ICER research team during the scoping period • Submission of a letter reacting to the posted draft scoping document • Participation in ICER's preliminary model presentation, with both written and verbal feedback provided • Submission of a letter reacting to the posted draft evidence report. Initial input received informed the selection of the PICOTS elements of the review. For example, our focus on stratified results (e.g., biologic-naïve vs. experienced) rather than on overall intent-to- treat findings was informed in large part by a discussion of the heterogeneity of the patient experience and outcomes in the Crohn's and Colitis Foundation's open input letter. The Crohn's and Colitis Foundation was also instrumental in highlighting specific impacts of UC and its treatment on children and adolescents. During a conference call conducted prior to the posting of the draft scope, participants noted that there is much interest in understanding the most appropriate sequence of therapy for the moderate-to-severe patient, as unnecessary delays in effective treatment could have catastrophic effects in terms of hospitalization and requirement for surgical intervention. Many of these delays are felt to be caused by insurance-mandated step therapy, in which patients are required to step through one or two biologic treatments before a therapy of choice, even if they have previously been failed by treatments in the same class or have ©Institute for Clinical and Economic Review, 2020 Page 10 Final Report – Targeted Immune Modulators for UC Return to Table of Contents contraindications to them. At the moment, however, there is little to no empiric data on appropriate sequencing to counteract these policies. The Crohn's and Colitis Foundation also highlighted the importance of EIMs in UC, which could include joint/arthritic symptoms, uveitis (painful inflammation of the middle layer of eye tissue), and lung complications. The Crohn's and Colitis Foundation also responded to our draft scope with comments, and we revised the scope in response to this feedback. Specifically, we added subgroups of particular interest, including by age (in addition to pediatric patients, those ages 65 and older were felt to be of interest), presence of EIMs, and other comorbidities. We also added steroid-free remission as an outcome of interest and were advised to remove pain relief, given its inconsistent and variable collection. Finally, we added demyelinating disease as an adverse event of interest. We received feedback after the preliminary model presentation suggesting three modifications to our planned approach: • More than two switches between TIMs to better reflect clinical practice • A scenario analysis in which infliximab is used in the second line after vedolizumab • Allow the per-cycle rate of discontinuation to vary over time. We considered these modifications, but we did not find empiric data with which to estimate them for the base-case analysis. However, we did conduct scenario analyses to address them, including a) evaluating the effect of shorter time horizons on model results, where two TIM switches would represent a more realistic expectation; b) allowing a lower discontinuation rate after one year for those remaining on therapy; and c) use of infliximab in the second line, assuming comparable efficacy for infliximab to that observed in biologic-naïve trials. We also reviewed a publication on the direct health care costs of IBD that was sponsored by the Crohn's and Colitis Foundation7 for possible inclusion in the model, but opted to use an alternative source that stratified costs the presence of response and remission. 2.2 Impact on Patients Several themes emerged from conversations and documents received. We have organized them into multiple sections, including heterogeneity in disease presentation; the benefits and risks of UC treatment; challenges of access to care; and gaps in current UC management. Heterogeneity in Disease Presentation As with many chronic diseases, the presentation of symptoms and disease course can vary substantially among patients. In some, the disease course may reflect periods of active disease and remission, while in others, the symptoms are persistent despite escalating medical therapy. A ©Institute for Clinical and Economic Review, 2020 Page 11 Final Report – Targeted Immune Modulators for UC Return to Table of Contents minority of patients may present with a rapidly progressive form of the disease known as fulminant colitis. UC may also manifest in different locations. In some patients, inflammation may be limited to the rectum ("ulcerative proctitis"), while in others, it may extend from the rectum to the splenic fixture ("left-sided colitis"), and still others may experience inflammation throughout the colon ("pancolitis"). As noted previously, children as young as five years of age or less may develop UC, with additional complications, such as growth failure and delayed puberty. In addition, there are noted but currently poorly understood differences in how racial and ethnic minorities experience UC. Those of Hispanic and Asian descent appear to present more commonly with the pancolonic form of disease. However, despite an estimate of approximately 700,000 minority individuals suffering from UC in the US, very little is known about the epidemiology and progression of disease in these groups. The Crohn's and Colitis Foundation is currently working with the US Centers for Disease Control on a longitudinal study to address these populations. The clinical differences in disease presentation by severity as well as patient characteristics have profound impacts on those suffering with UC. Patients with long enough periods of remission may be able to resume normal work, school, or leisure activity, while others with more progressive disease require increasing levels of caregiver support. The presence of more severe disease among certain racial and ethnic minorities can exacerbate disparities in access to appropriate care that they may already experience generally. Benefits and Risks of Ulcerative Colitis Treatment Input from the Crohn's and Colitis Foundation reiterated that, regardless of presentation or symptoms, the goal of treatment for any UC patient should be sustained and durable steroid-free remission, along with appropriate psychosocial support, normal health-related quality of life, and prevention of morbidity, including hospitalization, surgery, and colorectal cancer. The Crohn's and Colitis Foundation recommends that patient-clinician conversations include preferences for certain medications and/or routes of administration, ability to adhere to a medication regimen, the impact of treatment on daily life, and the patient's expected out-of-pocket costs associated with medical and surgical treatment. As with the disease itself, the Crohn's and Colitis Foundation mentioned that response to treatment is also heterogeneous. Treatments that are successful in some patients may not work for others, and patients frequently report loss of response, development of intolerable side effects, and the need to cycle off a medication. Patients may respond differently to the TNF inhibitors based on drug composition (e.g., mouse- vs. human-derived, possibly as a result of greater potential immunogenicity with the former) as well as clinical considerations (body mass index, disease severity). Some patients do not clinically respond to TNF inhibitor therapy while others develop immunogenicity to a specific agent; a switch outside of class may be indicated for the former, while ©Institute for Clinical and Economic Review, 2020 Page 12 Final Report – Targeted Immune Modulators for UC Return to Table of Contents a switch in-class may be the best course of action for the latter. It is also possible that some patients may do well on combination biologic therapy, although this has not been extensively studied. The overall lack of information and evidence in this area is frustrating to patients as it potentially delays effective treatment, which may negatively impact various aspects of a patient's quality of life. While all TIMs have side effects that patients must weigh when selecting treatment, the Crohn's and Colitis Foundation urged particular caution with regard to surgical intervention. We heard from multiple stakeholders that patients view surgery as a last-resort option. Though colectomy may be curative in some patients, there are long-term complications to consider, which may have substantial and distressing implications for a patient's quality of life and activities of daily living. Further, because UC is an immune-mediated condition, even after surgery, patients may still contend with devastating EIMs. Up to 80% of patients report symptoms of pouchitis (inflammation of the lining of the pouch created during a colectomy), and up to 20% of these individuals will develop refractory or rapidly relapsing disease. Some patients receiving "J-pouch" surgery (creation of an internal pouch that eliminates the need for an external ostomy) have been later diagnosed with Crohn's disease in other parts of their intestines. The invasiveness of the procedure in conjunction with the potential for long-term complications is a significant source of fear and anxiety among patients living with UC. At the same time, there is also apprehension toward biologic use and its potential effects, which may further complicate how patients weigh these treatment options. Access to Care Challenges Several challenges with accessing appropriate care were noted by the Crohn's and Colitis Foundation. For one, requirements for complex treatment regimens may pose a challenge for physicians and, consequently, patients. Primary care providers and gastroenterologists without specific UC experience may not, for example, pursue inpatient treatment with cyclosporine for the patient with acute severe UC, or may not appreciate the broader benefits of psychosocial and dietary support for long-term care management. The direct medical costs associated with managing UC are substantial. A recent analysis of the Optum Research Database commissioned by the Crohn's and Colitis Foundation found that annual management costs associated with IBD were nearly $23,000, or threefold higher than among a matched set of non-IBD controls.7 Importantly, estimates of average annual patient out-of-pocket costs for these services and lost wages associated with UC were approximately $2,000 and $3,000, which may pose a substantial burden for some.7 Costs are especially onerous for patients without insurance or those who are unable to work due to their disease. Reports suggest that out-of-pocket costs for UC diagnostic work-up alone exceed $4,000 in uninsured patients; even after receiving insurance, medication costs may exceed $350 per month.35 36 ©Institute for Clinical and Economic Review, 2020 Page 13 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Finally, the Crohn's and Colitis Foundation reiterated patient concerns regarding step therapy and its various negative consequences. Patients are often undertreated if they are failed by multiple biologics, and by the time they gain access to a treatment subject to numerous step therapy protocols, their disease may have progressed considerably, limiting the efficacy of the drug. The Crohn's and Colitis Foundation referenced vedolizumab to further illustrate this issue. Vedolizumab is a gut-selective biologic agent that is recommended as a potential first-line treatment for both induction and maintenance in UC but is often not covered by insurance until failure by other treatment options. In some cases, patients must be failed by treatment with at least two TNF inhibitors before vedolizumab is considered, despite clinical evidence that they do not respond to this biologic class. Other insurance mechanisms may produce additional barriers and sources of stress for UC patients. For example, some UC treatments are covered as a medical benefit and others as a pharmacy benefit, with different criteria for treatment authorization and different expectations for patient financial contribution. Step therapy algorithms typically found in UC are not concordant with authoritative clinical guidelines (see Section 3) and are also not portable; individuals who switch health plans (or whose employer does) often must repeat earlier steps in the treatment algorithm even if those treatments were previously unsuccessful. Step therapy requirements are exacerbated in the pediatric population, where many TIMs do not carry an FDA-approved indication for use in young patients. There is some evidence that pediatric patients present more frequently with aggressive forms of UC, so time is of the essence in providing effective treatment. Yet, many pediatric patients do not have access to the full range of UC treatments due to such requirements. These delays may cause irreversible impairments in growth and early requirements for surgical intervention, with its own set of consequences. Gaps in Current Ulcerative Colitis Management Finally, the Crohn's and Colitis Foundation noted major gaps in the current management of the disease, manifested primarily in a lack of high-quality evidence. There is a lack of head-to-head clinical trials of TIMs, for example, which limits the ability of patients and their families to make fully informed decisions about their treatment goals and desires. It is a source of frustration for the Crohn's and Colitis Foundation, patients, and clinicians that the one large head-to-head trial that is available-one that demonstrated vedolizumab's superiority to adalimumab-has not resulted in any appreciable change in insurer coverage policy. Moreover, there is a general need for more comparative effectiveness research given the plethora of drug classes, mechanisms of action, routes of administration, and safety profiles now available. As mentioned previously, UC management is increasingly involving shared decision-making between the patient and clinician. Evidence generated from Crohn's and Colitis Foundation- sponsored focus groups suggested room for improvement, as many patients reported that initial therapy was prescribed without discussion or consultation.37 These same patients, however, ©Institute for Clinical and Economic Review, 2020 Page 14 Final Report – Targeted Immune Modulators for UC Return to Table of Contents reported that they were able to bring their own research and preferences into discussions at subsequent visits, suggesting that treatment should be highly individualized and integrated into a shared decision-making process. That said, there are some clear factors that influence patient- provider decisions, such as the presence of heart failure or a history of melanoma (contraindications to TNF inhibitors)37 as well as expected time to clinical response. It was also noted that patient participants in Crohn's and Colitis Foundation-sponsored focus groups voiced concern with UC symptoms not routinely collected in clinical trials, which tend to focus on stool frequency and rectal bleeding. Beyond these, patients also reported concerns with pain and fatigue, ability to concentrate, and fecal urgency. Further, patients also experience challenges with interpersonal relationships and educational and career goals due to the impact of the disease on activities of daily living. Relatedly, patients often face difficulty when sharing the experience of UC symptoms with friends, family, and colleagues, which may lead to feelings of isolation. Finally, both the Crohn's and Colitis Foundation and others have noted that management of pain and its sequelae is a major challenge for patients with UC and Crohn's disease. A recent cross- sectional study of nearly 300 patients with IBD (~40% of whom had UC) found that 40% of respondents met criteria for chronic pain and nearly 20% reported opioid use.38 A number of psychosocial factors were associated with greater pain severity, including depression, anxiety, and reduced self-efficacy. ©Institute for Clinical and Economic Review, 2020 Page 15 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 3. Summary of Coverage Policies and Clinical Guidelines 3.1 Coverage Policies We reviewed the Tufts Medical Center Specialty Drug Evidence and Coverage (SPEC) Database for US commercial health plan coverage policies for adalimumab, golimumab, infliximab, infliximab- dyyb, infliximab-abda, tofacitinib, ustekinumab, and vedolizumab. Developed by the Center for Evaluation of Value and Risk in Health, the SPEC database features data from more than 290 specialty drugs, more than 175 disease areas, and more than 25,000 decisions from 17 of the largest US national and regional commercial payers: Aetna, Anthem, Blue Cross Blue Shield (BCBS) of Florida (FL), Massachusetts (MA), Michigan (MI), North Carolina (NC), New Jersey (NJ), and Tennessee (TN), CareFirst, Centene, Cigna, Emblem, Health Care Service Corporation (HCSC), Highmark, Humana, Independence Blue Cross (IndepBC), and UnitedHealthcare (UHC). We also searched for National or Local Coverage Determinations (NCDs or LCDs) from the Centers for Medicare and Medicaid Services (CMS) and from the California Department of Health Care Services. We located two NCDs that describe indications and limitations of coverage for fecal occult blood tests (Record 190.34) and colorectal cancer screening (Record 210.3). In addition, we located five LCDs pertaining to diagnostic and therapeutic colonoscopy (Record L34213), diagnostic colonoscopy and sigmoidoscopy (Record L34614), and Prometheus IBD sgi Diagnostic® test (L37299/L37539).39-44 Table 3.1 on the following page summarizes the benefit designs for representative commercial payers. We were unable to locate publicly available coverage policies for adalimumab from BCBSNJ, BCBSTN, Emblem, or HCSB; for golimumab from BCBSNJ, BCBSTN, and Emblem; for tofacitinib from BCBSNJ, BCBSTN, CareFirst, Emblem, HCSC, and Highmark; and for ustekinumab from Aetna, BCBSFL, BCBSTN, CareFirst, Cigna, HCSC, IndepBC, and UHC due to its recent approval. As a note, all therapies except ustekinumab and vedolizumab are approved by the FDA as second- line treatments. The TNF inhibitors are indicated as per the FDA label after inadequate response to conventional therapy and tofacitinib is indicated as per the FDA label following an inadequate response to a TNF inhibitor. Lastly, we do not report several percentages for tofacitinib and vedolizumab because the FDA recently changed the labeled indications of these drugs, and most coverage policies have not yet been updated. ©Institute for Clinical and Economic Review, 2020 Page 16 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 3.1. Benefit Design for Treating Moderate-to-Severe UC across Representative Commercial Payers Step Edits After First-Line Therapy % of Plans with % of Plans % of Plans with Coverage Criteria Excluding Drug 0 1 2 3+ Prescriber More Restrictive from Coverage Restrictions than FDA Label Adalimumab 0% 25% 83% 8% 8% 0% 25% Golimumab 0% 43% 43% 43% 14% 0% 21% Infliximab 0% 24% 71% 18% 12% 0% 18% Infliximab-dyyb 0% 59% 29% 47% 24% 0% 18% Infliximab-abda 0% 65% 24% 53% 18% 6% 18% Tofacitinib 0% -- -- -- -- -- 27% Step Edits Imposed by Payer 1 2 3 Ustekinumab 0% 100% 75% 13% 13% 38% Vedolizumab 0% -- -- -- -- -- FDA: Food and Drug Administration, IL: interleukin, JAK: Janus kinase, TNF: tumor necrosis factor Adalimumab Twelve out of the 17 payers surveyed have publicly available coverage policies for adalimumab. Only a quarter of payers list policies more restrictive than the FDA label. All plans surveyed require a documented diagnosis of moderate-to-severe UC, and all payers list an age restriction consistent with the FDA label (18 years of age or older). Out of all the drugs surveyed, plans overall have the least number of step edits required for access to adalimumab after failure by conventional therapy (83%). The most common step therapy requirement is a trial of multiple conventional therapies, such as corticosteroids, aminosalicylates, and thiopurines.45 Golimumab Of the 17 commercial payers surveyed, 14 companies have publicly available coverage policies for golimumab. All plans surveyed require a documented diagnosis of moderate-to-severe UC, and all payers list an age restriction consistent with the FDA label (18 years of age or older). Compared to the other TNF inhibitors (adalimumab and infliximab), more plans have coverage criteria that is narrower than the FDA label (24-25% vs. 43%, respectively), and a larger number of plans necessitate additional step edits following failure by conventional therapy. Several plans require a trial with adalimumab or tofacitinib before accessing golimumab, while others require trials with multiple conventional therapies.45 ©Institute for Clinical and Economic Review, 2020 Page 17 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Infliximab All 17 payers surveyed have issued coverage policies for infliximab. As noted in the table above, less than a quarter of payers have coverage policies that are more restrictive than the FDA label. All plans surveyed require a documented diagnosis of moderate-to-severe UC. Because infliximab is the only TIM approved in pediatric patients in addition to adults, most plans specify that the patient must be six years of age or older to receive treatment. Similar to adalimumab, most patients are able to access infliximab following failure by conventional treatment with no additional step edits, but some payers require that patients try at least two conventional agents, such as corticosteroids, aminosalicylates, and thiopurines.45 Infliximab-dyyb and Infliximab-abda Most commercial payers surveyed maintain differing coverage criteria for accessing infliximab versus its biosimilars. Compared to infliximab, plans generally have much more restrictive coverage conditions for the biosimilars. A larger set of plans require additional step edits (in contrast to the FDA label) in order to access either infliximab-dyyb or infliximab-abda. Most commonly, plans necessitate that patients undergo a trial with infliximab first. Only five of the 17 payers surveyed have coverage policies equivalent to the labeled indication of infliximab-dyyb, and even fewer for infliximab-abda. The most restrictive plan requires that patients seeking infliximab-abda are 1) refractory to or require continuous immunosuppression with high-dose corticosteroids, 2) inadequate responders to aminosalicylates, 3) inadequate responders to thiopurines, 4) inadequate responders to infliximab and infliximab-dyyb, and 5) inadequate responders to additional preferred alternatives, including vedolizumab, golimumab, or tofacitinib.45,46 Tofacitinib Eleven out of the 17 plans surveyed have issued coverage policies for tofacitinib, the only small- molecule drug under review. As noted above, we do not report several percentages because the labeled indication recently changed, and most plans have not yet updated their policies. Currently, according to the FDA label, in order to access tofacitinib, patients must demonstrate an inadequate response or intolerance to a TNF inhibitor. Previously, as indicated in the 2018 FDA label, patients did not have to step through a TNF inhibitor to access tofacitinib.45,47 Ustekinumab Ustekinumab was approved for the treatment of patients with moderate-to-severe UC in October 2019. Currently, ustekinumab has an FDA indication for first-line treatment (i.e., prior to conventional therapy). However, all plans surveyed list coverage criteria narrower than the FDA label; 75% of plans surveyed, require that patients try and be failed by one treatment (either a conventional agent or a TNF inhibitor) before utilizing ustekinumab. A small percentage of plans ©Institute for Clinical and Economic Review, 2020 Page 18 Final Report – Targeted Immune Modulators for UC Return to Table of Contents (13%) require three step edits; the most restrictive plan requires that patients demonstrate an inadequate response or intolerance to at least two conventional agents and one preferred drug (adalimumab) before accessing ustekinumab.45 Vedolizumab The labeled indication of vedolizumab was changed on March 31, 2020. Previously, in order to access vedolizumab, patients had to first step through a TNF inhibitor. As of the publication of this report, vedolizumab is indicated as a first-line treatment. We surveyed the five largest commercial payers and currently only Aetna and UHC have updated their polices since March 31, 2020; Cigna, Anthem, and Humana have not. Although Aetna and UHC have issued new policies since the label change, neither explicitly allow for first-line use. Aetna considers vedolizumab medically necessary for patients who have either 1) previously received the treatment or any other biologic or tofacitinib or, 2) were failed by at least one conventional therapy option.48 UHC requires a history of failure, contraindication, or intolerance to TNF inhibitors, azathioprine or mercaptopurine, or corticosteroids.49 3.2 Clinical Guidelines Below, we review clinical guidelines pertaining to UC from the American Gastrological Association (AGA), the American College of Gastroenterology (ACG), the Toronto Consensus, and National Institute for Health and Care Excellence (NICE). Though all four guidelines contain recommendations for patients with mildly active and acute UC, the summaries below include recommendations targeted only toward non-hospitalized, chronic patients with moderate-to- severe UC. American Gastrological Association Clinical Practice Guidelines on the Management of Moderate-to-Severe Ulcerative Colitis (2020)50 The AGA released its clinical practice guidelines for the treatment of moderate-to-severe UC in January 2020. According to the AGA, "moderate-to-severe disease" refers to patients who are dependent on or refractory to systemic corticosteroids, have severe endoscopic disease activity (including the presence of ulcers), or at high risk of colectomy. For such patients, the AGA recommends the use of adalimumab, golimumab, infliximab, tofacitinib, ustekinumab, or vedolizumab over no treatment due to clinical trial results that demonstrate the superiority of TIMs over placebo. Due to the increased risk of pulmonary embolism and all-cause mortality, the AGA notes that the recommended induction dose of tofacitinib is 10 mg twice daily for eight weeks, and for maintenance, 5 mg twice daily. ©Institute for Clinical and Economic Review, 2020 Page 19 Final Report – Targeted Immune Modulators for UC Return to Table of Contents The AGA's guidance is further stratified by a patient's previous experience with biologics. In patients who are naïve to biologics, infliximab or vedolizumab is recommended over adalimumab for induction of remission. However, adalimumab may be preferred in patients with less severe disease or those who favor self-administration. For patients with previous experience with infliximab, the AGA recommends ustekinumab or tofacitinib (rather than vedolizumab or adalimumab) for induction, especially for patients with primary nonresponse to infliximab. The AGA recommends that all patients with moderate-to-severe UC combine TNF inhibitors, vedolizumab, or ustekinumab with thiopurines or methotrexate, as opposed to biologic monotherapy or thiopurine monotherapy. Lastly, the AGA recommends the early use of biologics rather than step therapy with aminosalicylates, which may delay effective treatment in patients at high risk of complications, hospitalization, and colectomy. American College of Gastroenterology Clinical Guideline: Ulcerative Colitis in Adults (2019)51 The ACG published its clinical guidelines for the treatment of UC in adults in 2019. After a diagnosis of UC is determined, the ACG recommends treating patients with a treat-to-target approach to achieve endoscopic improvement, which is most likely to produce steroid-free remission and prevent hospitalizations and colectomy. Treatment begins with induction, and the ACG outlines several recommendations for treatment split by induction and maintenance. Induction of Remission The ACG offers numerous recommendations for the management of moderately-to-severely active UC. Major recommendations are summarized below: • For induction of remission in moderately-to-severely active UC, the ACG recommends oral systemic corticosteroids, TNF inhibitors (adalimumab, golimumab, or infliximab), vedolizumab, or tofacitinib • For patients who achieved induction of remission using infliximab, the ACG recommends the addition of a thiopurine for added clinical efficacy • The ACG recommends the use of vedolizumab or tofacitinib for induction of remission • For patients who have been previously failed by a TNF inhibitor, vedolizumab is recommended for induction of remission. The ACG does not offer specific recommendations concerning sequencing of treatment as head-to- head data is limited. However, it is noted that higher rates of remission and lower rates of corticosteroid usage have been observed with infliximab compared to adalimumab. In addition, adalimumab has been associated with higher rates of all-cause and UC-related hospitalizations compared to other TNF inhibitors. ©Institute for Clinical and Economic Review, 2020 Page 20 Final Report – Targeted Immune Modulators for UC Return to Table of Contents In addition, the ACG acknowledges that the use of treatments with fewer systemic effects (i.e., vedolizumab) in UC is an emerging clinical practice. Vedolizumab may be preferable for some patient populations, such as older patients who may have a higher risk of infection. In contrast, for patients with several EIMs, systemic therapies may be preferred. Overall, decisions regarding which therapy to use to achieve initial remission should be made by the patient and clinician and should be based on prognostic factors, disease extent and severity, and EIMs among other factors important to the patient. Maintenance of Remission Once induction of remission is achieved, patients and providers must work to maintain it. The ACG offers guidance based on how initial remission was achieved. Please see below for major recommendations: • Systemic corticosteroids are not recommended for maintenance of remission • Patients who achieved induction with a TNF inhibitor (adalimumab, golimumab, infliximab) should continue to use a TNF inhibitor as maintenance therapy • Patients who achieved induction of remission with vedolizumab should continue to use vedolizumab as maintenance therapy • Patients who achieved induction of remission with tofacitinib should continue to use tofacitinib as maintenance therapy. Clinical Practice Guidelines for the Medical Management of Non-Hospitalized Ulcerative Colitis: The Toronto Consensus (2015)52 The Toronto Consensus was released in 2015 and provides guidance for the medical management of patients with moderate-to-severe UC. Of note, the Toronto Consensus was published before Health Canada's approval of tofacitinib for UC,53 and as such, the guidelines focus primarily on the TNF inhibitors and vedolizumab. Following a diagnosis of moderate or severe UC, the Toronto Consensus recommends oral corticosteroids as first-line therapy to induce remission. The guidelines advise against thiopurines or methotrexate for induction. Once a patient has achieved remission, the guidelines recommend against the continued use of oral corticosteroids to maintain remission. If a patient presents with a contraindication to corticosteroids, TNF inhibitors or vedolizumab may be considered as first-line to induce remission and should be combined with thiopurines or methotrexate. Because there are no head-to-head trials among the TNF inhibitors, the guidelines do not offer guidance on which TNF inhibitor to choose. After TNF inhibitor induction therapy, patients should be evaluated at eight to 12 weeks to determine response. If a patient is responding to a TNF inhibitor, the patient should begin maintenance therapy using the same TNF inhibitor used ©Institute for Clinical and Economic Review, 2020 Page 21 Final Report – Targeted Immune Modulators for UC Return to Table of Contents for induction. In patients with a suboptimal response to TNF inhibitors, the guidelines recommend dose escalation, and if a patient loses response to TNF inhibitor maintenance, the guidelines recommend dose optimization informed by drug monitoring. If patients are ultimately failed by a TNF inhibitor, the guidelines recommend a switch to vedolizumab over a switch to a different TNF inhibitor. Patients on vedolizumab should be assessed between eight and 14 weeks and should continue vedolizumab for maintenance therapy if they demonstrate response. National Institute for Health and Care Excellence – Ulcerative Colitis: Management54 NICE recommends the use of adalimumab, golimumab (only if the manufacturer provides the 100 mg dose at the same cost as the 50 mg dose), and infliximab in patients with moderate-to-severe UC whose disease has not responded adequately to conventional therapy (i.e., corticosteroids, mercaptopurine, azathioprine) or who present with contraindications to conventional therapy. Because there are no head-to-head trials among the TNF inhibitors, NICE recommends that the choice of treatment be made on an individual basis following discussions between the clinician and patient. The choice of treatment should account for therapeutic need and issues of adherence, and if more than one treatment is appropriate, the least costly option should be selected. All three TNF inhibitors should be given as a planned course of therapy until treatment fails or until 12 months after beginning treatment. Clinical symptoms, biological markers, and results from an endoscopy are key to assessment at 12 months to determine whether a patient should continue treatment. If a patient is in remission, the clinician may discuss halting treatment, but should be easily able to resume if the patient relapses.55 NICE recommends vedolizumab in patients with moderate-to-severe UC only if the manufacturer provides the treatment with the agreed discount. If the discount is honored, NICE recommends that vedolizumab be used until it stops working or surgery becomes indicated. Similar to the TNF inhibitors, NICE recommends assessment at 12 months following the start of treatment to determine whether treatment should continue. If patients are in complete remission, the clinician may consider stopping treatment, but should resume if the patient relapses.56 Lastly, NICE recommends tofacitinib in patients with moderate-to-severe UC if conventional therapy or a biologic agent cannot be tolerated or if the disease has not responded adequately to prior treatment. Similar to vedolizumab, tofacitinib is only recommended if the manufacturer honors the agreed discount.57 ©Institute for Clinical and Economic Review, 2020 Page 22 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 4. Comparative Clinical Effectiveness 4.1 Overview To inform the analysis of the comparative clinical effectiveness of TIMs for moderate-to-severe UC, we systematically reviewed and synthesized existing evidence from available clinical studies. Full PICOTS criteria are described in Section 1.2. The drugs and regimens of interest for this review are included in Table 1.1 in Section 1. In this review, we compared the efficacy, safety, and effectiveness of TIMs (adalimumab, golimumab, infliximab, tofacitinib, ustekinumab, and vedolizumab [IV]) to ongoing background conventional therapy (i.e., placebo arms of clinical trials) and to each other. Our review focused on the clinical benefits important to patients living with UC as well as potential harms. We sought evidence on all outcomes listed in Section 1. The methods and findings of our review of the clinical evidence are described in the sections that follow. 4.2 Methods Data Sources and Searches Procedures for the systematic literature review assessing the evidence on TIMs for moderate-to- severe UC followed established best research methods.59,60 The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.61 These guidelines include a checklist of 27 items, which are described further in Appendix Table A1. We searched MEDLINE and EMBASE for relevant studies through July 2020. Each search was limited to English-language studies of human subjects and excluded articles indexed as guidelines, letters, editorials, narrative reviews, case reports, or news items. We included abstracts from conference proceedings identified from the systematic literature search. All search strategies were generated utilizing the Population, Intervention, Comparator, and Study Design elements described previously. The proposed search strategies included a combination of indexing terms (MeSH terms in MEDLINE and EMTREE terms in EMBASE) as well as free-text terms. To supplement the database searches, we performed manual checks of the reference lists of included trials and systematic reviews and invited key stakeholders to share references germane to the scope of this project. We also supplemented our review of published studies with data from conference proceedings, regulatory documents, information submitted by manufacturers, and other grey literature when the evidence met ICER standards (for more information, see https://icer- review.org/methodology/icers-methods/icer-value-assessment-framework-2/grey-literature- ©Institute for Clinical and Economic Review, 2020 Page 23 Final Report – Targeted Immune Modulators for UC Return to Table of Contents policy/). Where feasible and deemed necessary, we also accepted data submitted by manufacturers "in-confidence," in accordance with ICER's published guidelines on acceptance and use of such data (https://icer-review.org/use-of-in-confidence-data/). Study Selection After removal of duplicate citations, references went through two levels of screening at both the abstract and full-text levels. Three reviewers independently screened the titles and abstracts of all publications identified using DistillerSR (Evidence Partners, Ottawa, Canada) and disagreements were resolved through consensus. Studies that did not meet PICOTS criteria were excluded. We included evidence from RCTs and high-quality comparative observational studies, where available (see below for details on quality assessment). Single-arm studies and early clinical phase development studies (i.e., Phase I), were excluded. Further, abstracts that report duplicate data available in the published articles or results from observational studies presented in conference abstracts with insufficient information to evaluate methodological quality were excluded. Only studies that evaluated an FDA-approved dose were included; however, we also included treatment arms with higher dosing levels given the potential for dose escalation in UC. Finally, while concomitant use of conventional systemic agents (e.g., aminosalicylates, thiopurines) was permitted in available trials, we excluded any trial that randomized patients to treatment with TIMs in combination with other agents, given our focus on the incremental benefits of TIM therapy.62 Data Extraction and Quality Assessment Two reviewers extracted data into evidence tables. Extracted data were verified by another researcher. Elements include study name, study year, study design, phase of the trial, study inclusion and exclusion criteria, description of patient populations, sample size, duration of follow- up, funding source, interventions (agent, dosage, frequency, schedules), concomitant therapy allowed and used (agent, dosage, frequency, schedules), outcome assessments, results, and quality assessment for each study. The report utilized the criteria published by the US Preventive Services Task Force (USPSTF) to assess the quality of clinical trials and cohort studies, using the categories "good," "fair," or "poor."63 For more information on data extraction and quality assessment, refer to Appendix D. Assessment of Level of Certainty in Evidence We used the ICER Evidence Rating Matrix to evaluate the level of certainty in the available evidence of a net health benefit among each of the interventions of focus (see Appendix D).64 ©Institute for Clinical and Economic Review, 2020 Page 24 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Assessment of Bias As a part of quality assessment, we examined the evidence base for the presence of potential publication bias. Given the emerging nature of the evidence base for newer treatments, ClinicalTrials.gov was scanned to identify studies completed more than two years ago. Studies that met the inclusion criteria, and for which no findings have been published or presented publicly, were selected. We found no evidence of publication bias. However, we identified a Phase III RCT (NCT01551290) for infliximab in Chinese patients sponsored by Xian-Janssen that was completed in November 2014; results are only available from a clinical trial report linked to the ClinicalTrials.gov page and have not been published in a peer-reviewed journal. Data Synthesis and Statistical Analyses Data for the available comparisons of TIMs for the FDA-approved dose or higher were abstracted in evidence tables (see Appendix D) and synthesized in the text on the following pages. In addition, comparative efficacy of TIMs for patients living with moderate-to-severe UC was assessed by means of NMA, where feasible. Trials that were deemed sufficiently similar in terms of population, intervention type, duration, and outcome definitions were included in the NMAs. Below, we briefly summarize the characteristics of our NMAs. Appendix F contains a more detailed description of the NMA methods. NMAs focused on clinical efficacy outcomes, including clinical response, clinical remission, and endoscopic improvement, were conducted. Given the expected differences in the clinical efficacy of treatment in patients with and without prior biologic exposure, separate networks were developed for biologic-naïve and biologic-experienced populations. In addition, outcomes were analyzed separately for the induction phase (six to 14 weeks) and maintenance phase (52-60 weeks). Clinical response and remission were analyzed in both induction and maintenance phases. Endoscopic improvement was analyzed only during the induction phase due to limited data availability and trial design differences. The evidence base of the included trials in the maintenance phase is a combination of "treat- through" designs, where patients were randomized at baseline and followed through until the end of maintenance, and "re-randomized" designs, where responders to treatment from one or two induction trials were re-randomized in the maintenance phase. The re-randomized trials report clinical response and remission at the end of maintenance among induction responders. In order to analyze all trials in a comparable fashion in a single network, results from treat-through trials were adjusted to more closely resemble results from re-randomized trials. ©Institute for Clinical and Economic Review, 2020 Page 25 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Data were available for all TIMs (adalimumab, golimumab, infliximab, tofacitinib, ustekinumab, and vedolizumab) in the biologic-naïve population. However, it was noted that the use of tofacitinib is no longer feasible in a biologic-naïve population based on an FDA-enforced label change (July 2019) that now requires that tofacitinib use be reserved for "…patients who have failed or do not tolerate tumor necrosis factor (TNF) blockers."65 Based on this information, tofacitinib was not included in the NMAs for induction or maintenance outcomes within the biologic-naïve population. Data were available for adalimumab, tofacitinib, ustekinumab, and vedolizumab in the biologic-experienced population; data were not available for golimumab or infliximab, so we were unable to generate comparative efficacy estimates for these drugs. The population of focus for this review included patients with moderate-to-severe UC who had inadequate response to conventional treatment. Despite this relatively narrow definition, trial populations may differ in terms of prior conventional therapies used, other demographic or clinical risk factors, timing of trial assessments, and other concerns. Adjusting for placebo response in an NMA design is frequently performed as means of controlling for differences in population characteristics and baseline risk; we considered placebo adjustment in situations where a) model fit and convergence was not compromised; and b) inclusion of such an adjustment materially changed model findings. While different doses of some of TIMs were studied in available trials, we found no statistically significant differences in rates of key response or remission outcomes between doses in any relevant trial, so these data were pooled at the drug level for our primary NMAs. Additionally, we included trials conducted exclusively in Asian populations in our primary NMAs. To explore the impact of these characteristics on our results, we conducted two sensitivity analyses: 1) using unpooled doses for relevant TIMs; and 2) restricting included trials to those that enrolled subjects from multiple countries. Results from these analyses were generally consistent with the primary results (see Appendix F for more details). All NMAs were conducted in a Bayesian framework using either the gemtc or R2jags package in R.66,67 4.3 Results Study Selection Our literature review search identified a total of 6,836 potentially relevant references. We included 53 references, of which 42 related to 19 unique RCTs in adults,8,9,68-106 two references related to one RCT and one observational study in children,107,108 and nine references related to nine high-quality comparative observational studies in adults.109-117 The primary reasons for study exclusion were that the intervention or comparators used were outside the scope of this review, another study population was of focus (e.g., patients with mild-to-moderate UC), the study design was non- ©Institute for Clinical and Economic Review, 2020 Page 26 Final Report – Targeted Immune Modulators for UC Return to Table of Contents comparative, or conference abstracts reported duplicative data to the full publications. In the results that follow, we focus on the comparative efficacy and safety of TIMs in the adult population; the RCT conducted in children is described later under Special Populations. Of the 19 included trials in the adult population, one trial was a head-to-head trial comparing vedolizumab and adalimumab (VARSITY), and the other trials were placebo controlled. The trials enrolled patients with moderate-to-severe UC (Mayo Score ≥6 with an endoscopic sub-score ≥2) whose disease had not responded to conventional systemic agents. The trials assessed disease severity using the Mayo Score at the end of induction (week six to 14) or at the end of maintenance (week 52-60), or both. Although we preferred the use of the total Mayo Score to measure disease severity, we use data based on the partial Mayo Score (i.e., all components except the endoscopic sub-score) if data based on the total Mayo Score were not available. Use of conventional systemic agents (e.g., azathioprine, aminosalicylates) was permitted alongside active and placebo therapy in all trials, although we note substantial variation in levels of utilization of conventional treatments across trials. For example, baseline corticosteroid use ranged from 40-65% across trials, and immunomodulator use from 12-44%; in certain trials, detailed use of conventional treatments at baseline was not reported. Across the trials, the demographic and clinical characteristics of trial populations were broadly similar. Overall, the included trials were comparable with respect to age (range: 34-43 years) and disease severity as measured by the Mayo Score (range: 8.0-9.1). However, there was some variation in the disease duration across trials, ranging from 3.7 to 8.3 years (see Table 4.1). Trials excluded patients who had been diagnosed with Crohn's disease, severe or extensive colitis requiring colectomy, and with a history of malignancy. Additional details of included references and their study characteristics have been summarized in Appendix D, and the trials included in the review are summarized in Table 4.1 on the following page. ©Institute for Clinical and Economic Review, 2020 Page 27 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.1. Study Design and Baseline Characteristics of Included RCTs8,9,68,77,79-81,86,90,93-96,99,101 Trial Mean Baseline Inclusion in Randomized Treatment Arms† (n) (IND/MAINT Naïve (%) Characteristics Primary Response/ Timepoints) Exp (%) Disease Mayo Endpoint** Remission Induction Maintenance Age TT or RR* Duration Score NMA§ Head-to-Head VARSITY Naïve 1) VEDO 300 mg (n=383) 1) VEDO 300 mg q8w (n=383) Remission at IND (14/52 Weeks), (79%) 40.6 6.8 8.7 2) ADA 160/80 mg (n=386) 2) ADA 40 mg (n=386) week 52 MAINT TT Exp (21%) Adalimumab ULTRA 1 Naïve 1) ADA 160/80 mg (n=130) Remission at -- 37.8‡ 6.1‡ 8.8 IND (8 Weeks) (100%) 2) Placebo (n=130) week 8 Naïve Remission at ULTRA 2 1) ADA 160/80 mg (n=248) 1) ADA 40 mg (n=248) IND (60%) 40.4 8.3 8.9 week 8 and (8/52 Weeks), TT 2) Placebo (n=246) 2) Placebo (n=246) MAINT Exp (40%) at week 52 Suzuki 2014 Naïve 1) ADA 160/80 mg (n=87) 1) ADA 40 mg (n=177) Not 42.7 7.9 8.5 IND§ (8/52 Weeks), TT (100%) 2) Placebo (n=96) 2) Placebo (n=96) specified Golimumab 1) GOL 200/100 mg (n=294) PURSUIT-SC Naïve Response at 2) GOL 400/200 mg (n=298) -- 40.0 6.3 8.5 IND (6 Weeks) (100%) week 6 3) Placebo (n=292) Response PURSUIT-M Naïve 1) GOL 100 mg (n=154) -- 40.2 7.0 8.3 through MAINT (54 Weeks), RR (100%) 2) Placebo (n=156) week 54 Response PURSUIT-J Naïve 1) GOL 100 mg (n=32) -- 41.1 5.5‡ 8.0‡ through --§ (54 Weeks), RR (100%) 2) Placebo (n=31) week 54 Infliximab 1) IFX 5 mg/kg (n=121) 1) IFX 5 mg/kg (n=121) ACT 1 Naïve Response at IND 2) IFX 10 mg/kg (n=122) 2) IFX 10 mg/kg (n=122) 41.8 6.8 8.4 (8/54 Weeks), TT (100%) week 8 MAINT 3) Placebo (n=121) 3) Placebo (n=121) ©Institute for Clinical and Economic Review, 2020 Page 28 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial Mean Baseline Inclusion in Randomized Treatment Arms† (n) (IND/MAINT Naïve (%) Characteristics Primary Response/ Timepoints) Exp (%) Disease Mayo Endpoint** Remission Induction Maintenance Age TT or RR* Duration Score NMA§ 1) IFX 5 mg/kg (n=121) 1) IFX 5 mg/kg (n=121) ACT 2 Naïve Response at 2) IFX 10 mg/kg (n=120) 2) IFX 10 mg/kg (n=120) 40.0 6.6 8.4 IND§ (8/30 Weeks), TT (100%) week 8 3) Placebo (n=123) 3) Placebo (n=123) Kobayashi 2016 Naïve 1) IFX 5 mg/kg (n=104) 1) IFX 5 mg/kg (n=73) Response at 38.9 7.6 8.6 IND§ (8/30 Weeks), TT (100%) 2) Placebo (n=104) 2) Placebo (n=72) week 8 Jiang 2015 Naïve 1) IFX 5 mg/kg (n=41) 1) IFX 5 mg/kg (n=41) Response at 34.4 4.4 NR IND§ (8/30 Weeks), TT (100%) 2) Placebo (n=41) 2) Placebo (n=41) week 8 NCT01551290 Naïve 1) IFX 5 mg/kg (n=49) 1) IFX 5 mg/kg (n=50) Response at 37‡ 3.7‡ 8‡ --§ (8/26 Weeks), TT (100%) 2) Placebo (n=50) 2) Placebo (n=49) week 8 Tofacitinib Naïve OCTAVE 1 1) TOF 10 mg (n=476) Remission at (47%) -- 41.6 6.3‡ 9.1 IND (8 Weeks) 2) Placebo (n=122) week 8 Exp (53%) Naïve OCTAVE 2 1) TOF 10 mg (n=476) Remission at (45%) -- 40.8 6.1‡ 9.0 IND (8 Weeks) 2) Placebo (n=112) week 8 Exp (55%) Naïve 1) TOF 5 mg (n=198) OCTAVE SUSTAIN Remission at (52%) -- 2) TOF 10 mg (n=197) 42.7 6.8‡ NR¤ MAINT (52 Weeks), RR week 52 Exp (48%) 3) Placebo (n=198) Ustekinumab Naïve Remission at UNIFI 1) UST 6 mg/kg (n=322) 1) UST 90 mg q8w (n=176) IND (49%) 41.7 8.1 8.9 week 8 and (8/52 Weeks), RR 2) Placebo (n=319) 2) Placebo (n=175) MAINT Exp (51%) at week 52 Vedolizumab Naïve 1) VEDO 300 mg q8w (n=122) Response at GEMINI 1 1) VEDO 300 mg (n=225) IND (52%) 2) VEDO 300 mg q4w (n=125) 40.3 6.9 8.6 week 6 and (6/52 Weeks), RR 2) Placebo (n=149) MAINT Exp (48%) 3) Placebo (n=126) at week 52 ©Institute for Clinical and Economic Review, 2020 Page 29 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial Mean Baseline Inclusion in Randomized Treatment Arms† (n) (IND/MAINT Naïve (%) Characteristics Primary Response/ Timepoints) Exp (%) Disease Mayo Endpoint** Remission Induction Maintenance Age TT or RR* Duration Score NMA§ Response at Motoya 2019 Naïve 1) VEDO 300 mg (n=164) 1) VEDO 300 mg q8w (n=41) week 10 and IND (10/60 Weeks), (49%) 42.9 8.3 8.2 2) Placebo (n=82) 2) Placebo (n=42) remission at MAINT RR Exp (51%) week 60 Naïve VISIBLE 1†† 1) VEDO 300 mg q8w (n=54) Remission at (61%) -- 39.3 7.9 9.0‡ MAINT (6/52 Weeks), RR 2) Placebo (n=56) week 52 Exp (39%) ADA: adalimumab, GOL: golimumab, Exp: experienced, IFX: infliximab, IND: induction, kg: kilogram, MAINT: maintenance, M week: maintenance week, mg: milligram, N: total number, n: number, q4w: every 4 weeks, q8w: every 8 weeks, PBO: placebo, RR: re-randomized, TOF: tofacitinib, TT: treat-through, UST: ustekinumab, VEDO: vedolizumab *Treat-through or re-randomized design for trials with maintenance phases. †Only including randomized treatment arms of FDA-approved dosing or higher. ‡Median reported. §Refer to Appendix F for more details on reasons for exclusion from induction or maintenance NMAs. ¤The mean Mayo Score at the start of treatment was not reported; however, the mean Mayo Score at beginning of the maintenance phase was reported (3.3). **Response and remission based on the Mayo Score. ††In our review, we only include evidence from the IV vedolizumab arm and not the subcutaneous vedolizumab arm. ©Institute for Clinical and Economic Review, 2020 Page 30 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Quality of Individual Studies We used the USPSTF criteria to rate the quality of the included RCTs (see Appendix D). The RCTs were rated "good" or "fair." Of note, we did not rate the one trial of infliximab that was only available in grey literature (NCT01551290) and this trial was not included in our NMA. Generally, good, and fair quality trials had comparable groups at baseline, clear definitions of outcomes and interventions, and valid outcome measurements. For efficacy endpoints, most RCTs performed intent-to-treat analysis while some performed modified intent-to-treat analysis; additionally, most RCTs employed appropriate methods to handle missing data (e.g., non-responder imputation for response and remission). Randomized patients who received at least one dose of the study drug were evaluated for safety endpoints. Furthermore, we noted that some trials had imbalances across arms in certain baseline characteristics, which may have affected findings. Specifically, Motoya and colleagues noted imbalances in key disease characteristics that could have led to unusually high placebo response rates, specifically in the biologic-experienced group.9 Additionally, we noted that across the trials, the rates of trial and treatment discontinuation were generally high; in some trials, the rates were uneven, which may have affected results. Also, in the re-randomized trials, patients randomized to placebo in the maintenance phase had already received active treatment. This could have potentially resulted in unblinding to maintenance treatment among patients, although we note that the primary measures of effectiveness in this review were objective in nature and unlikely to be materially affected by patient unblinding. Quality of the Network Meta-Analysis Evidence Base Generally, the trials were comparable in terms of populations studied and outcomes measured. However, we did note some differences across trials. First, the trials conducted in a "mixed" population used different criteria regarding prior exposure to biologics to define their strata when reporting subgroup results. Some trials defined subgroups by prior use of a biologic (e.g., "biologic- naïve" and "biologic-experienced"), while the other trials defined subgroups by prior failure by a biologic (e.g., "biologic-failure" and "biologic non-failure"). Throughout our review, we refer to the populations across the trials as "biologic-naïve" and "biologic-experienced"; however, it should be noted that differences in how the populations were defined could potentially affect findings. Further, the clinical efficacy outcomes of response and remission were generally defined comparably across trials (see Section 1) with a few exceptions. Importantly, for the head-to-head VARSITY trial, the rate of response at the end of maintenance used in our NMA was based on the partial Mayo Score (including all components except the endoscopic sub-score) and not the total Mayo Score. This may lead to higher effect sizes relative to total Mayo results and possibly biased estimates of treatment effect. Secondly, the OCTAVE trials used a stricter definition of clinical remission compared to other trials by requiring the rectal bleeding score to be 0 (rather than 0 or 1) ©Institute for Clinical and Economic Review, 2020 Page 31 Final Report – Targeted Immune Modulators for UC Return to Table of Contents in addition to the other criteria. Moreover, while most trials used a local endoscopy reading, the OCTAVE trials used a centralized system.96 Lastly, the adalimumab trials used the worst rank method to measure Mayo Scores (i.e., taking the highest score in a three-day period) as opposed to using an average of the scores for their primary and key secondary endpoints, which may have underestimated effect sizes relative to average scores.101 We attempted placebo adjustment for all four populations of interest in our NMAs. Material changes in findings were observed in both maintenance populations (i.e., biologic-naïve and biologic-experienced populations). However, results were highly unstable in the biologic- experienced maintenance NMA, regardless of the number of iterations attempted, in all likelihood due to the sparsity of the available network. We therefore included a placebo adjustment only in the maintenance NMA for the biologic-naïve population. We conducted inconsistency tests for each of the NMA populations using the node-splitting approach. We found no statistical differences between direct and indirect estimates in any of our populations of interest. Further details are available in Appendix F. Clinical Benefits Adults Biologic-Naïve Induction Outcomes from Randomized Controlled Trials • In the head-to-head VARSITY trial, vedolizumab had a higher rate of clinical response compared to adalimumab, although rates of remission were similar. • In placebo-controlled trials, TIMs generally had higher rates of clinical response and remission compared to placebo, a finding that was also reflected in NMAs. • The NMA showed infliximab and vedolizumab had higher rates of response and remission compared to adalimumab. • The NMA showed all TIMs had higher rates of endoscopic improvement compared to placebo; additionally, infliximab had higher rates of endoscopic improvement compared to adalimumab. Response and Remission Fifteen of the included RCTs8,9,68,77,80,81,86,90,93,95,96,99,101 measured the efficacy of TIMs in achieving clinical response and remission at the end of the induction phase (six to 14 weeks) in the biologic- naïve population. One trial was head-to-head (vedolizumab vs. adalimumab) and 14 were placebo ©Institute for Clinical and Economic Review, 2020 Page 32 Final Report – Targeted Immune Modulators for UC Return to Table of Contents controlled. Data were available for all TIMs included in our review. The rates in each of the individual 15 RCTs are described below and reported in Table 4.2. Head-to-Head Trial Vedolizumab versus Adalimumab In the VARSITY head-to-head trial comparing the efficacy of vedolizumab and adalimumab, the rate of response was higher with vedolizumab 300 mg compared to adalimumab 160/80 mg at the end of induction (70.1% vs. 49.5%, difference: 20.6; 95% CI for difference: 12.9 to 28.2); however, the rates of achieving remission did not statistically differ (Table 4.2).8 Placebo-Controlled Trials Adalimumab Three RCTs measured the efficacy of adalimumab compared to placebo (ULTRA 1, ULTRA 2, and Suzuki 2014).86,93,101 In one RCT (ULTRA 2), adalimumab 160/80 mg had higher rates compared to placebo of clinical response (59.3% vs. 36.8%, p<0.0001) and remission (21.3% vs. 11.0%, p=0.017) at the end of induction.93 Evidence was mixed in the other RCTs; in one RCT (ULTRA 1), adalimumab 160/80 mg had higher rates compared to placebo of remission (18.5% vs. 9.2%, p=0.031) but not clinical response,86 and in the other RCT (Suzuki 2014), adalimumab 160/80 mg had higher rates compared to placebo of clinical response (50.0% vs. 35.4%, p=0.044) but not remission (Table 4.2).101 As mentioned earlier, all adalimumab trials used the worst rank method to assign Mayo Scores, which likely led to lower effect sizes. Golimumab In the one RCT of golimumab versus placebo (PURSUIT-SC),95 both golimumab 200/100 mg and golimumab 400/200 mg had higher rates compared to placebo of clinical response (51.0% and 54.9% vs. 30.3%, respectively, p<0.0001 for both) and remission (17.8% and 17.9% vs. 6.4%, respectively, p<0.0001 for both) in the Phase III population at the end of induction (Table 4.2). In our NMA, we used rates from the pooled Phase II and III population, which were similar to the rates observed in the Phase III population. Both golimumab 200/100 mg and golimumab 400/200 mg had numerically higher rates compared to placebo of clinical response (50.0% and 54.7% vs. 30.5%, respectively) and remission (17.7% and 18.8% vs. 6.8%, respectively); the significance of the difference was not reported for the pooled analysis. ©Institute for Clinical and Economic Review, 2020 Page 33 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Infliximab Five RCTs measured the efficacy of infliximab compared to placebo (ACT 1, ACT 2, Jiang 2015, Kobayashi 2016, and NCT01551290);68,80,81,90 all five of the trials assessed infliximab 5 mg/kg and two trials also assessed infliximab 10 mg/kg.90 In all five RCTs, infliximab had higher rates of clinical response at the end of induction compared to placebo. In three RCTs, infliximab had higher rates of clinical remission at the end of induction compared to placebo,80,90 while one RCT found marginal significance (p=0.05) and another RCT found no difference (Kobayashi 2016 and NCT01551290) (Table 4.2).68,81 As noted earlier, NCT01551290 was not included in the NMA as it was only available in grey literature. Tofacitinib Two RCTs measured the efficacy of tofacitinib compared to placebo (OCTAVE 1 and OCTAVE 2).96 In a pooled analysis of the two trials, tofacitinib 10 mg had higher rates compared to placebo of clinical response (64.5% vs. 39.1%, p<0.0001) and remission (24.1% vs. 11.8%, p<0.01) at the end of induction (Table 4.2).72 As noted previously, we have not included these data in our NMA due to the FDA label change. Ustekinumab In the one RCT of ustekinumab versus placebo (UNIFI), ustekinumab 6 mg/kg had higher rates compared to placebo of clinical response (66.7% vs. 35.4%, p<0.001) and remission (18.6% vs. 9.5%, p=0.022) at the end of induction (Table 4.2).98 Vedolizumab Two RCTs measured the efficacy of vedolizumab (IV) compared to placebo (GEMINI 1 and Motoya 2019).9,77 In one RCT, vedolizumab 300 mg had higher rates compared to placebo of clinical response (53.1% vs. 26.3%, difference: 26.4, 95% CI for difference: 12.4 to 40.4) and remission (23.1% vs. 6.6%, difference: 15.5, 95% CI for difference: 5.1 to 25.9) at the end of induction (GEMINI 1),77 but these measures did not differ between vedolizumab and placebo in the other RCT (Motoya 2019) (Table 4.2).9 As previously noted, Motoya and colleagues describe several treatment group imbalances that may have contributed to unusually high placebo response and remission rates. ©Institute for Clinical and Economic Review, 2020 Page 34 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.2. Response and Remission Rates at the End of Induction in the Biologic-Naïve Population8,9,68,77,80,81,86,90,93,95,96,99,101 Response Remission Trial Week N Arm % Significance % Significance Head-to-Head 304 VEDO 300 mg 70.1 Diff. (95% CI): 27.6 Diff. (95% CI): VARSITY 14 305 ADA 160/80 mg 49.5 20.6 (12.9, 28.2) 23.6 4.0 (-2.9, 10.9) Adalimumab 130 ADA 160/80 mg 54.6 NS 18.5 p=0.031 ULTRA 1 8 130 PBO 44.6 -- 9.2 -- 150 ADA 160/80 mg 59.3 p<0.001 21.3 p=0.017 ULTRA 2 8 145 PBO 38.6 -- 11.0 -- 90 ADA 160/80 mg 50.0 p=0.044 10.0 NS Suzuki 2014 8 96 PBO 35.4 -- 11.5 -- Golimumab 294 GOL 200/100 mg 51.0 p<0.0001 17.8 p<0.0001 PURSUIT-SC 6 298 GOL 400/200 mg 54.9 p<0.0001 17.9 p<0.0001 (Phase III)* 292 PBO 30.3 -- 6.4 -- Infliximab 121 IFX 5 mg/kg 69.4 p<0.001 38.8 p<0.001 ACT 1 8 122 IFX 10 mg/kg 61.5 p<0.001 32.0 p=0.002 121 PBO 37.2 -- 14.9 -- 121 IFX 5 mg/kg 64.5 p<0.001 33.9 p<0.001 ACT 2 8 120 IFX 10 mg/kg 69.2 p<0.001 27.5 p<0.001 123 PBO 29.3 -- 5.7 -- 41 IFX 5 mg/kg 78.1 p=0.00 53.7 p=0.003 Jiang 2015 8 41 PBO 36.6 -- 21.9 -- 104 IFX 5 mg/kg 54.8 p=0.005 20.2 NS Kobayashi 2016 8 104 PBO 35.6 -- 10.6 -- 50 IFX 5 mg/kg 64.0 p=0.0021 22.0 NS NCT0155129† 8 49 PBO 32.7 -- 10.2 -- Tofacitinib OCTAVE 1 & 2 440 TOF 10 mg 64.5 p<0.0001 24.1 p<0.01 8 (Pooled)‡ 110 PBO 39.1 -- 11.8 -- Ustekinumab 8 156 UST 6 mg/kg 66.7 p<0.001 18.6 p=0.022 UNIFI 158 PBO 35.4 -- 9.5 -- Vedolizumab 130 VEDO 300 mg 53.1 Diff. (95% CI): 26.4 23.1 Diff. (95% CI): 15.5 GEMINI 1 6 76 PBO 26.3 (12.4 to 40.4) 6.6 (5.1 to 25.9) 79 VEDO 300 mg 53.2 Diff. (95% CI): 16.6 27.8 Diff. (95% CI): 13.2 Motoya 2019 10 41 PBO 36.6 (-1.8 to 35.0) 14.6 (-1.4 to 27.9) ADA: adalimumab, CI: confidence interval, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, NR: not reported, NS: not significant, PBO: placebo, UST: ustekinumab, VEDO: vedolizumab *Pooled Phase II and III results were used in NMA. †Data for trial only available in grey literature and not included in NMA. ‡Data not included in NMA given tofacitinib's label change. ©Institute for Clinical and Economic Review, 2020 Page 35 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Network Meta-Analyses Twelve of the 15 available RCTs were included in the induction NMA for the biologic-naïve population (all but the tofacitinib trials and the infliximab trial available in grey literature [NCT01551290]); refer to Appendix F for more details on reasons for exclusion from the NMA. Results from the NMA showed all TIMs were more likely to achieve response and remission compared to placebo. Specifically, TIMs were 1.4 to 1.9 times more likely to achieve clinical response (Table 4.3) and 1.8 to 3.2 times more likely to achieve remission (Table 4.4) compared to placebo. Additionally, infliximab and vedolizumab were more likely to achieve response and remission compared to adalimumab (Table 4.3 and 4.4). No other statistical differences among TIMs were observed. Table 4.3. Risk Ratios for Response at the End of the Induction Phase in Biologic-Naïve Patients IFX Pooled* 1.07 (0.93, 1.24) VEDO 1.10 (0.93, 1.34) 1.02 (0.84, 1.27) UST 1.15 (1.00, 1.35) 1.08 (0.92, 1.27) 1.05 (0.85, 1.28) GOL Pooled* 1.37 (1.18, 1.62) 1.28 (1.14, 1.45) 1.25 (1.00, 1.53) 1.18 (1.00, 1.42) ADA 1.88 (1.67, 2.12) 1.76 (1.54, 2.02) 1.71 (1.41, 2.06) 1.63 (1.43, 1.86) 1.38 (1.21, 1.56) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg/kg and 10 mg/kg pooled; golimumab 200/100 mg and 400/200 mg pooled. Table 4.4. Risk Ratios for Remission at the End of the Induction Phase in Biologic-Naïve Patients IFX Pooled* 1.15 (0.87, 1.54) VEDO 1.21 (0.85, 1.79) 1.05 (0.71, 1.59) UST 1.34 (1.00, 1.80) 1.16 (0.84, 1.62) 1.10 (0.73, 1.63) GOL Pooled* 1.84 (1.39, 2.50) 1.60 (1.29, 1.98) 1.52 (1.00, 2.26) 1.38 (1.00, 1.92) ADA 3.22 (2.60, 3.96) 2.79 (2.18, 3.58) 2.66 (1.86, 3.73) 2.41 (1.89, 3.08) 1.76 (1.38, 2.19) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg/kg and 10 mg/kg pooled; golimumab 200/100 mg and 400/200 mg pooled. ©Institute for Clinical and Economic Review, 2020 Page 36 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Endoscopic Improvement Fourteen of the included placebo-controlled RCTs9,68,77,80,81,90,93,95,96,99,101 measured the efficacy of TIMs in achieving endoscopic improvement at the end of the induction phase (six to 14 weeks) in the biologic-naïve population. Results for endoscopic improvement stratified by prior biologic exposure at week 14 were not available for the VARSITY head-to-head trial. Data were available for all TIMs included in our review. The rates in each of the individual 14 RCTs are reported in Appendix Table D5. Generally, all TIMs showed higher rates of endoscopic improvement compared to placebo at the end of induction. Network Meta-Analysis Eleven trials were included in our NMA (all but the tofacitinib trials and the infliximab trial available in grey literature [NCT01551290]); refer to Appendix F for more details on reasons for exclusion from the NMA. Results from our NMA showed all TIMs were significantly more likely to achieve endoscopic improvement compared to placebo (Table 4.5). Specifically, TIMs were 1.3 to 1.8 times more likely to achieve endoscopic improvement compared to placebo. Infliximab was more likely to induce endoscopic improvement than adalimumab. No other statistical differences were observed among TIMs. Table 4.5. Risk Ratios for Endoscopic Improvement at the End of the Induction Phase in Biologic- Naïve Patients IFX Pooled* 1.07 (0.74, 1.51) VEDO 1.17 (0.77, 1.75) 1.1 (0.67, 1.80) UST 1.19 (0.92, 1.55) 1.12 (0.77, 1.65) 1.02 (0.67, 1.57) GOL Pooled* 1.43 (1.13, 1.81) 1.34 (0.94, 1.95) 1.22 (0.81, 1.85) 1.2 (0.92, 1.57) ADA 1.83 (1.56, 2.17) 1.71 (1.27, 2.4) 1.56 (1.08, 2.3) 1.53 (1.26, 1.9) 1.28 (1.08, 1.53) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg/kg and 10 mg/kg pooled; golimumab 200/100 mg and 400/200 mg pooled. Maintenance Outcomes from Randomized Controlled Trials • The head-to-head VARSITY trial showed the rate of clinical remission was higher with vedolizumab compared to adalimumab. • In placebo-controlled trials, TIMs generally had higher rates of clinical response and remission compared to placebo, a finding that was also reflected in the NMAs. • The NMA also showed that vedolizumab had higher rates of clinical response and remission compared to adalimumab and golimumab. ©Institute for Clinical and Economic Review, 2020 Page 37 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Response and Remission Eleven RCTs measured the efficacy of TIMs in achieving clinical response and remission at the end of the maintenance phase (52-60 weeks) in the biologic-naïve population. One RCT was head-to- head (vedolizumab vs. adalimumab), and 10 trials were placebo controlled. Data were available for all TIMs included in our review. Four of the trials had a treat-through design8,90,93,101 and seven utilized a re-randomized approach.9,77,79,92,94,96,99 In the following section, we describe results for the key outcome measures in the biologic-naïve population as reported in the published trials (i.e., measures designated as primary and key secondary outcomes for the overall population). As discussed earlier, we adjusted the rates from the treat-through trials to more closely resemble results from re-randomized trials to enable comparisons in our NMA. Additionally, for some re-randomized trials, we preferred manufacturer- submitted inputs or other published secondary outcomes over the results described below, as they provided us with more comparable outcomes for our NMA (e.g., response at end of maintenance vs. response through the end of maintenance). For more details on the inputs used in the NMAs, refer to Appendix F. Published Results from Treat-Through Trials Head-to-Head Trial Vedolizumab versus Adalimumab In the VARSITY head-to-head trial, vedolizumab 300 mg every eight weeks had higher rates of remission compared to adalimumab 40 mg at the end of maintenance (Table 4.6).8 Placebo-Controlled Trials Both adalimumab 40 mg and infliximab (5 mg/kg and 10 mg/kg doses) had higher rates of response and remission compared to placebo at the end of maintenance in available trials (Table 4.6).90,93,101 ©Institute for Clinical and Economic Review, 2020 Page 38 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.6. Response and Remission Rates Among all Patients in Treat-Through Trials in Biologic- Naïve Patients8,90,93,101 Response Remission Trial Week N Arm % Significance % Significance 304 VEDO 300 mg q8w 34.2 Diff. (95% CI): VARSITY 52 NR 305 ADA 40 mg 24.3 9.9 (2.8 to 17.1) 150 ADA 40 mg 36.7 p=0.019 22.0 p=0.029 ULTRA 2 52 145 PBO 24.1 -- 12.4 -- 177 ADA 40 mg 31.0 p=0.021 23.3 p=0.011 Suzuki 2014 52 96 PBO 17.7 -- 7.3 -- 121 IFX 5 mg/kg 45.5 p<0.001 34.7 p=0.001 ACT 1 54 122 IFX 10 mg/kg 44.3 p<0.001 34.4 p=0.001 121 PBO 19.8 -- 16.5 -- ADA: adalimumab, CI: confidence interval, IFX: infliximab, kg: kilogram, mg: milligram, NR: not reported, PBO: placebo, q8w: every 8 weeks, VEDO: vedolizumab Published Results from Re-Randomized Trials Placebo-Controlled Trials Golimumab Golimumab 100 mg had higher rates of maintaining response through week 54 and of achieving remission at both week 30 and 54 compared to placebo in PURSUIT-J and PURSUIT-M (Table 4.7).79,94 Tofacitinib Tofacitinib 5 mg and 10 mg had higher rates of achieving response and remission at week 52 compared to placebo in OCTAVE SUSTAIN (Table 4.7).72 As noted previously, we have not included tofacitinib in our biologic-naïve NMA due to the FDA label change. Ustekinumab Ustekinumab 90 mg had higher rates of maintaining response through week 52 and remission at maintenance week 52 compared to placebo in UNIFI (Table 4.7).98 Vedolizumab Vedolizumab 300 mg every eight weeks had higher rates of response compared to placebo in GEMINI 1 and Motoya 2019, and higher rates of remission compared to placebo in GEMINI 1 but not in Motoya 2019 at week 52.9,77 Additionally, vedolizumab 300 mg every four weeks showed higher rates of response and remission compared to placebo in GEMINI 1. In VISIBLE 1, rates of ©Institute for Clinical and Economic Review, 2020 Page 39 Final Report – Targeted Immune Modulators for UC Return to Table of Contents remission were numerically higher with vedolizumab 300 mg every eight weeks compared to placebo, although the significance of the difference was not reported (Table 4.7).92 Table 4.7. Response and Remission Rates Among Induction Responders in Re-Randomized Trials in Biologic-Naïve Patients8,9,77,79,90,92-94,99 Response Remission Trial Week N Arm % Significance % Significance 151 GOL 100 mg 49.7 p<0.001 27.8 p=0.004 PURSUIT-M 54 154 PBO 31.2 -- 15.6 -- 32 GOL 100 mg 56.3 Diff. (95% CI): 36.9 50.0 Diff. (95% CI): 43.6 PURSUIT-J 54 31 PBO 19.4 (14.8, 59.0) 6.5 (24.2, 62.9) 115 TOF 5 mg 56.6 p<0.0001 41.7 p<0.0001 OCTAVE 52 104 TOF 10 mg 64.4 p<0.0001 44.2 p<0.0001 SUSTAIN 109 PBO 24.8 -- 11.0 -- 85 UST 90 mg q8w 77.6 48.2 UNIFI 52 p<0.001 p=0.024 87 PBO 50.6 31.0 Diff. (95% CI): 9.2 Diff. (95% CI): 28.4 73 VEDO 300 mg q4w 56.2 47.9 (13.7, 44.7) (13.7, 43.1) GEMINI 1 52 Diff. (95% CI): 38.2 Diff. (95% CI): 26.6 72 VEDO 300 mg q8w 65.3 45.8 (22.6, 53.8) (11.8, 41.4) 79 PBO 26.6 -- 19.0 -- Motoya 24 VEDO 300 mg q8w 66.7 Diff. (95% CI): 31.0 54.2 Diff. (95% CI): 18.5 60 2019 28 PBO 35.7 (5.1, 56.9) 35.7 (-8.2, 45.1) 32 VEDO 300 mg q8w NR 53.1 VISIBLE 1 52 NR NR 37 PBO NR 18.9 CI: confidence interval, GOL: golimumab, IFX: infliximab, mg: milligram, NR: not reported, PBO: placebo, q4w: every 4 weeks, q8w: every 8 weeks, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Network Meta-Analysis Eight RCTs were included in our NMA (all trials except for Suzuki 2014, PURSUIT-J, and OCTAVE SUSTAIN); please refer to Appendix F for more details on reasons for exclusion from the NMA. As mentioned previously, placebo adjustment was performed for the maintenance NMA in this population. Our NMA showed all TIMs were more likely to achieve clinical response and remission at the end of maintenance compared to placebo, although no statistical differences from placebo were observed for adalimumab or golimumab. Vedolizumab was shown to be more likely to achieve clinical response and remission compared to adalimumab and golimumab. No other statistical differences between TIMs were observed. ©Institute for Clinical and Economic Review, 2020 Page 40 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.8. Risk Ratios for Response at the End of the Maintenance Phase in Biologic-Naïve Patients UST 1.12 (0.62, 1.77) VEDO Pooled* 1.16 (0.77, 1.54) 1.04 (0.69, 1.58) IFX Pooled* 1.38 (0.72, 2.18) 1.21 (1.05, 1.43) 1.18 (0.74, 1.79) ADA 1.46 (0.75, 2.54) 1.30 (1.01, 1.77) 1.26 (0.77, 2.03) 1.07 (0.81, 1.50) GOL 1.80 (1.13, 2.50) 1.64 (1.16, 2.02) 1.56 (1.10, 2.14) 1.34 (0.94, 1.78) 1.25 (0.79, 1.70) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Table 4.9. Risk Ratios for Remission at the End of the Maintenance Phase in Biologic-Naïve Patients UST 1.18 (0.52, 2.26) VEDO Pooled* 1.24 (0.71, 1.83) 1.05 (0.60, 1.86) IFX Pooled* 1.56 (0.64, 3.00) 1.30 (1.06, 1.62) 1.25 (0.67, 2.22) ADA 1.70 (0.68, 3.55) 1.43 (1.01, 2.15) 1.37 (0.71, 2.61) 1.10 (0.75, 1.71) GOL 2.22 (1.17, 3.57) 1.93 (1.22, 2.58) 1.80 (1.13, 2.86) 1.47 (0.92, 2.14) 1.35 (0.74, 2.04) PBO ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Infliximab 5 mg and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Endoscopic Improvement One head-to-head and eight placebo-controlled RCTs measured the efficacy of TIMs in achieving endoscopic improvement at the end of the maintenance phase in the biologic-naive population. Compared to placebo, patients treated with adalimumab, golimumab, infliximab, tofacitinib, and ustekinumab reported achieving higher rates of endoscopic improvement.8,9,77,79,90,93,94,96,99,101 All TIMs reported the rates of achieving endoscopic improvement at the end of maintenance, while the PURSUIT-M and PURSUIT-J trials for golimumab used a stricter reporting criterion for maintaining endoscopic improvement at both week 30 and week 54. As mentioned previously, an NMA was not conducted for endoscopic improvement at the end of the maintenance phase. ©Institute for Clinical and Economic Review, 2020 Page 41 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Biologic-Experienced Induction Outcomes from Randomized Controlled Trials • In the head-to-head VARSITY trial, vedolizumab had a higher rate of response compared to adalimumab, although rates of remission were similar. • In placebo-controlled trials, rates of response and remission were higher with tofacitinib and ustekinumab compared to placebo but similar with adalimumab compared to placebo. Evidence was mixed for vedolizumab, with one trial reporting higher rates of response compared to placebo. • The NMA showed tofacitinib, ustekinumab, and vedolizumab had higher rates of clinical response and remission compared to placebo and adalimumab; adalimumab did not statistically differ from placebo. • The NMA showed tofacitinib and ustekinumab had higher rates of endoscopic improvement; vedolizumab or adalimumab did not statistically differ from placebo. Response and Remission Seven of the included RCTs measured the efficacy of TIMs in achieving clinical response and remission at the end of the induction phase (six to 14 weeks) in the biologic-experienced population. One RCT was a head-to-head trial (vedolizumab vs. adalimumab) and six were placebo controlled. Data were available for adalimumab, tofacitinib, ustekinumab, and vedolizumab, but not for golimumab or infliximab. Head-to-Head Trial Vedolizumab versus Adalimumab In the VARSITY head-to-head trial comparing the efficacy of vedolizumab and adalimumab,8 the rate of response was higher with vedolizumab compared to adalimumab at the end of induction (55.7% vs. 32.1%, difference: 23.6; 95% CI for difference: 8.5 to 38.7); however, the rates of achieving remission did not statistically differ (Table 4.10). Placebo-Controlled Trials Adalimumab In a single trial of adalimumab 160/80 mg versus placebo (ULTRA 2),93 there were no significant differences in achieving clinical response or remission at the end of induction (Table 4.10). ©Institute for Clinical and Economic Review, 2020 Page 42 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Tofacitinib Two RCTs measured the efficacy of tofacitinib compared to placebo (OCTAVE 1 and OCTAVE 2).96 In a pooled analysis of the two trials,72 tofacitinib 5 mg and 10 mg had higher rates compared to placebo of clinical response (51% vs. 23.4%, p <0.0001) and remission (11.4% vs. 0.8%, p<0.01) at the end of induction (Table 4.10). Ustekinumab In the one RCT of ustekinumab versus placebo, ustekinumab 6 mg/kg had higher rates compared to placebo of clinical response (57.2% vs. 27.3%, p<0.001) and remission (12.7% vs. 1.2%, p<0.001) at the end of induction (Table 4.10).99 Vedolizumab Two RCTs measured the efficacy of vedolizumab compared to placebo.9,77 In one RCT, vedolizumab 300 mg had higher rates compared to placebo of clinical response (39.0% vs. 20.6%, difference: 18.1, 95% CI for difference: 2.8 to 33.5) but not remission at the end of induction,77 and the other RCT found no difference in clinical response and remission rates (Motoya 2019) (Table 4.10).9 As previously noted, Motoya and colleagues describe several treatment group imbalances that may have contributed to unusually high placebo response and remission rates. Table 4.10. Response and Remission Rates at the End of Induction in the Biologic-Experienced Population8,9,77,93,96,99 Response Remission Trial Week N Arm % Significance % Significance 79 VEDO 300 mg 55.7 Diff. (95% CI): 22.8 Diff. (95% CI): VARSITY 14 81 ADA 160/80 mg 32.1 23.6 (8.5, 38.7) 12.3 10.3 (-1.5, 22.2) 98 ADA 160/80 mg 36.7 NS 9.2 NS ULTRA 2 8 101 PBO 28.7 -- 6.9 -- OCTAVE 1 & 2 465 TOF 10 mg 51.0 p<0.0001 11.4 p<0.01 8 (Pooled) 124 PBO 23.4 -- 0.8 -- 166 UST 6 mg/kg 57.2 p<0.001 12.7 p<0.001 UNIFI 8 161 PBO 27.3 -- 1.2 -- 82 VEDO 300 mg 39.0 Diff. (95% CI): 9.8 Diff. (95% CI): GEMINI 1 6 63 PBO 20.6 18.1 (2.8, 33.5) 3.2 7.0 (-1.3, 15.2) 85 VEDO 300 mg 27.1 Diff. (95% CI): 9.4 Diff. (95% CI): Motoya 2019 10 41 PBO 29.3 -2.2 (-19.0, 41.6) 9.8 -0.3 (-11.3, 10.7) ADA: adalimumab, CI: confidence interval, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, NS: not specified, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 43 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Network Meta-Analysis All seven RCTs were included in the NMA. Results from our NMA showed tofacitinib, ustekinumab, and vedolizumab were more likely to achieve clinical response and remission compared to placebo; there were no statistical differences between adalimumab and placebo (Table 4.11 and Table 4.12). Additionally, tofacitinib, ustekinumab, and vedolizumab were more likely to achieve clinical response and remission compared to adalimumab. Table 4.11. Risk Ratios for Response at the End of the Induction Phase in Biologic-Experienced Patients UST 1.01 (0.77, 1.33) TOF 1.32 (0.97, 1.87) 1.31 (0.96, 1.84) VEDO 2.01 (1.39, 3.13) 2.00 (1.36, 3.08) 1.53 (1.11, 2.15) ADA 2.11 (1.71, 2.62) 2.11 (1.68, 2.60) 1.61 (1.20, 2.05) 1.05 (0.71, 1.46) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. Table 4.12. Risk Ratios for Remission at the End of the Induction Phase in Biologic-Experienced Patients UST 1.02 (0.58, 1.82) TOF 1.75 (0.94, 3.45) 1.74 (0.92, 3.38) VEDO 3.82 (1.88, 8.49) 3.75 (1.80, 8.08) 2.18 (1.22, 3.95) ADA 4.13 (2.71, 6.26) 4.10 (2.62, 6.18) 2.38 (1.38, 3.80) 1.09 (0.57, 1.97) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. Endoscopic Improvement Six of the included placebo-controlled RCTs measured the efficacy of TIMs in achieving endoscopic improvement at the end of the induction phase in the biologic-experienced population.8,9,77,93,96 Data were available for adalimumab, tofacitinib, ustekinumab, and vedolizumab. Higher rates of endoscopic improvement compared to placebo were achieved with tofacitinib (22.0% vs. 6.1%, p<0.001) and ustekinumab (21.1% vs. 6.8%, p<0.01). Network Meta-Analysis All six RCTs were included in the NMA. Results from our NMA showed that patients treated with tofacitinib and ustekinumab were more likely to achieve endoscopic improvement compared to placebo. Similarly, tofacitinib and ustekinumab were also more likely to achieve endoscopic ©Institute for Clinical and Economic Review, 2020 Page 44 Final Report – Targeted Immune Modulators for UC Return to Table of Contents improvement than vedolizumab and adalimumab. We note, however, that credible intervals are particularly wide for this analysis, in large part due to the sparseness of the network and relatively smaller sample size of the biologic-experienced population included in these trials. Table 4.13. Risk Ratios for Endoscopic Improvement at the End of the Induction Phase in Biologic- Experienced Patients TOF 1.22 (0.48, 3.30) UST 3.24 (1.50, 7.84) 2.65 (1.25, 5.93) VEDO 3.60 (1.64, 8.85) 2.94 (1.37, 6.70) 1.11 (0.60, 2.07) ADA 3.92 (2.09, 8.65) 3.21 (1.75, 6.46) 1.21 (0.81, 1.89) 1.1 (0.70, 1.73) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. Maintenance Outcomes in Randomized Controlled Trials • In the head-to-head VARSITY trial, the rate of clinical remission was similar with vedolizumab and adalimumab. • In placebo-controlled trials, rates of response and remission were higher with adalimumab, tofacitinib, and ustekinumab. Evidence was mixed for vedolizumab, with two trials reporting higher rates of remission compared to placebo. • The NMA showed all TIMs had higher rates of response and remission compared to placebo. No differences among TIMs were observed. Response and Remission Seven RCTs measured the efficacy of TIMs in maintaining clinical response and remission in the biologic-experienced population. Six of the RCTs were placebo-controlled, and one RCT was head- to-head (VARSITY).8 Data was available for adalimumab, tofacitinib, ustekinumab, and vedolizumab, but not for golimumab or infliximab. Two of the trials had a treat-through design (ULTRA 2 and VARSITY)8,93 and five had re-randomized designs (GEMINI 1, Motoya 2019, VISIBLE 1, OCTAVE SUSTAIN, and UNIFI).9,77,92,96,99 In the section that follows, we describe results for the key outcomes in the biologic-experienced population as reported in the published trials. As discussed earlier, we adjusted the rates from the treat-through trials to resemble results more closely from re-randomized trials to enable comparisons in our NMA. Additionally, for some re-randomized trials, we preferred manufacturer- submitted inputs or other published secondary outcomes over the results described below, as they provided us with more comparable outcomes for our NMA (e.g., response at end of maintenance ©Institute for Clinical and Economic Review, 2020 Page 45 Final Report – Targeted Immune Modulators for UC Return to Table of Contents vs. response through the end of maintenance). For more details on the inputs used in the NMAs, refer to Appendix F. Published Results from Treat-Through Trials Head-to-Head Trial Vedolizumab versus Adalimumab In the VARSITY head-to-head trial, the rates of remission were similar with vedolizumab 300 mg every eight weeks and adalimumab 40 mg at week 52 (Table 4.14).8 Placebo-Controlled Trials Adalimumab Adalimumab 40 mg had higher rates of response and remission at the end of maintenance compared to placebo at week 52 in ULTRA-2 (Table 4.14).93 Table 4.14. Response and Remission Rates Among all Patients in the Treat-Through Trials in Biologic-Experienced Patients8,93 Response Remission Trial N Arm % Significance % Significance 79 VEDO 300 mg q8w 20.3 Diff. (95% CI): VARSITY 52 NR 81 ADA 40 mg 16.0 4.2 (-7.8, 16.5) 98 ADA 40 mg 20.4 0.038 10.2 p=0.039 ULTRA 2 52 101 PBO 9.9 -- 3.0 -- ADA: adalimumab, CI: confidence interval, mg: milligram, NR: not reported, PBO: placebo, q8w: every 8 weeks, VEDO: vedolizumab Published Results from Re-Randomized Trials Placebo-Controlled Trials Tofacitinib Tofacitinib 5 mg and 10 mg had higher rates of response and remission at week 52 compared to placebo in OCTAVE SUSTAIN (Table 4.15).72 Ustekinumab Ustekinumab 90 mg had higher rates of response through week 52 and remission at week 52 compared to placebo in UNIFI (Table 4.15).98 ©Institute for Clinical and Economic Review, 2020 Page 46 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Vedolizumab Vedolizumab 300 mg every eight weeks had higher rates of remission compared to placebo in GEMINI 1 and Motoya 2019 and had higher rates of response compared to placebo in GEMINI 1 but not in Motoya 2019.9,77 Additionally, vedolizumab 300 mg every four weeks had significantly higher rates of response and remission compared to placebo in GEMINI 1. In VISIBLE 1, rates of remission were numerically higher with vedolizumab compared to placebo, although the significance of the difference was not reported (Table 4.15).92 Table 4.15. Response and Remission Rates in Re-Randomized Trials in Biologic-Experienced Patients9,72,77,92,98 Response Remission Trial Week Arm N % Significance % Significance TOF 5 mg 83 44.6 p<0.0001 24.1 p<0.05 OCTAVE 52 TOF 10 mg 93 59.1 p<0.0001 36.6 p<0.0001 SUSTAIN PBO 89 14.6 -- 11.2 -- UST 90 mg q8w 91 64.8 39.6 UNIFI 52 p<0.001 p<0.001 PBO 88 38.6 17.0 Diff. (95% CI): Diff. (95% CI): VEDO 300 mg q4w 43 42.5 35.0 25.9 (5.8, 45.9) 31.3 (13.2, 49.3) GEMINI 1 52 Diff. (95% CI): Diff. (95% CI): VEDO 300 mg q8w 40 46.5 37.2 26.8 (7.4, 46.2) 27.8 (10.6, 45.0) PBO 38 15.8 -- 5.3 -- Motoya VEDO 300 mg q8w 17 64.7 Diff. (95% CI): 58.8 Diff. (95% CI): 60 2019 PBO 14 35.7 29.0 (-4.9, 62.8) 21.4 37.4 (5.6, 69.2) VEDO 300 mg q8w 22 27.3 VISIBLE 1 52 NR NR PBO 19 5.3 CI: confidence interval, mg: milligram, N: number, NR: not reported, PBO: placebo, q4w: every 4 weeks, q8w: every 8 weeks, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Network Meta-Analysis All seven RCTs were included in the NMA. All TIMs were more likely to achieve clinical response and remission compared to placebo. Specifically, TIMs were 1.9 to 2.4 times more likely to achieve clinical response compared to placebo (Table 4.16), and 2.5 to 3.5 times more likely to achieve remission compared to placebo (Table 4.17). No other statistical differences among TIMs were observed. ©Institute for Clinical and Economic Review, 2020 Page 47 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.16. Risk Ratios for Clinical Response at the End of the Maintenance Phase in Biologic- Experienced Patients VEDO Pooled* 1.12 (0.87, 1.55) ADA 1.25 (0.88, 1.86) 1.11 (0.72, 1.72) UST 1.26 (0.88, 1.90) 1.12 (0.70, 1.75) 1.00 (0.63, 1.61) TOF Pooled* 2.40 (1.87, 3.00) 2.14 (1.48, 2.89) 1.92 (1.34, 2.55) 1.92 (1.32, 2.55) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Tofacitinib 5 and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Table 4.17. Risk Ratios for Clinical Remission at the End of the Maintenance Phase in Biologic- Experienced Patients VEDO Pooled* 1.19 (0.80, 1.94) ADA 1.41 (0.83, 2.5) 1.17 (0.62, 2.27) UST 1.41 (0.82, 2.59) 1.18 (0.60, 2.31) 1.01 (0.51, 1.98) TOF Pooled* 3.48 (2.43, 5.03) 2.91 (1.72, 4.65) 2.49 (1.49, 3.82) 2.49 (1.46, 3.79) PBO ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Each box represents the estimated risk ratios and 95% credible interval for the combined direct and indirect comparisons between two drugs. Estimates in bold signify that the 95% credible interval does not contain 1. *Tofacitinib 5 and 10 mg pooled; vedolizumab 300 mg every 8 weeks and every 4 weeks pooled. Endoscopic Improvement Five of the included placebo-controlled RCTs and one head-to-head trial measured the efficacy of TIMs in achieving endoscopic improvement at the end of the maintenance phase in the biologic- experienced population. Data stratified by prior biologic exposure were available for adalimumab, tofacitinib, ustekinumab, and vedolizumab. Higher rates of endoscopic improvement were reported with tofacitinib 5 mg and 10 mg (30.1% and 39.8% vs. 12.4% for placebo, p<0.01 and <0.001, for both comparisons), ustekinumab (45.1% vs. 22.7%, p<0.001) and vedolizumab (44.6% vs. 7.9%, difference: 34.9; 95% CI for difference: 17.1 to 52.8). Differences were not statistically significant in the ULTRA 2 trial of adalimumab. ©Institute for Clinical and Economic Review, 2020 Page 48 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Other Outcomes Delayed Responders As noted previously, some patients may respond to induction therapy after FDA-mandated timepoints for clinical trials. We sought evidence on delayed response from all available RCTs and observational studies, and identified data for adalimumab, tofacitinib, and vedolizumab on an overall basis (i.e., not stratified by prior biologic use). Delayed response stratified by prior biologic use was available for golimumab, tofacitinib, and ustekinumab. Data are summarized in Appendix Table D7. Overall, with vedolizumab, 39% of induction non-responders achieved delayed response by week 14, based on data from GEMINI 1.56 With tofacitinib, 50.1% of induction non-responders achieved delayed response, of which 13.9% achieved remission by week 16, based on the combined data for OCTAVE 1 and 2.118 Among adalimumab recipients in ULTRA 2, 4.1% achieved delayed remission (data were not available for delayed response).93 When stratified by prior biologic exposure, among biologic-experienced ustekinumab induction non-responders in the UNIFI trial, 43.1% and 1.7% achieved delayed response and remission respectively by week 16. Additionally, in biologic-naïve populations, delayed response (79.1%) and remission (18.6%) was achieved with ustekinumab at week 16.71 Further, among biologic- experienced tofacitinib induction non-responders, 42.6% and 39.7% achieved delayed response and remission. Among biologic-naïve induction non-responders with tofacitinib, 56.2% and 18.8% achieved response and remission.118 A total of 28.1% of golimumab induction non-responders (all biologic-naïve) achieved response at the end of week 16 in PURSUIT-SC.70 Corticosteroid-Free Remission Concomitant corticosteroid use at baseline was permitted across the trials and the dose remained unaltered through the induction phase. The use of corticosteroids was then tapered, and corticosteroid-free remission was assessed at the end of the maintenance phase. Overall, one head-to-head and nine placebo-controlled trials reported corticosteroid-free remission, with six trials reporting corticosteroid-free remission stratified by prior biologic exposure (see Appendix Table D10 for details). In the VARSITY trial, a numerically higher proportion of patients treated with adalimumab achieved corticosteroid-free remission relative to vedolizumab, but the results were not statistically significant.8 Further, corticosteroid-free remission was higher with tofacitinib (both 5 mg and 10 mg) in the overall population as compared to placebo (27.6% vs. 10.9, p=0.003).96 Findings were similar when stratified by prior biologic use. In the placebo-controlled trials, a higher proportion of biologic-naïve patients achieved corticosteroid-free remission at the end of maintenance when treated with infliximab 5 mg (25.7% vs. 8.9%; p=0.006), vedolizumab (44.6% vs. 18.6%; difference: 26.3; 95% CI: 8.7 to 43.9), and ustekinumab (37.4 vs. 15.9, p<0.001). A higher proportion of biologic-experienced patients achieved corticosteroid-free remission at the end of ©Institute for Clinical and Economic Review, 2020 Page 49 Final Report – Targeted Immune Modulators for UC Return to Table of Contents maintenance when treated with tofacitinib (27.6% vs. 10.9%, p=0.002), vedolizumab (26.7% vs. 4.3%; difference: 21.3; 95% CI: 1.7-40.8), and ustekinumab (49.4 vs. 32.1, p=0.03).77,90,99 Ulcerative Colitis-Related Hospitalization and Surgeries The rates of UC-related hospitalizations and surgeries were not commonly reported in trials included in this review. Where available, results were reported in the overall population and not stratified by biologic-naïve and biologic-experienced. Data were available from the head-to-head VARSITY trial and placebo-controlled trials of adalimumab, infliximab, and ustekinumab. In the VARSITY trial, the rates of UC-related hospitalizations and procedures were slightly numerically higher for adalimumab compared to vedolizumab at week 52 (5.2% vs. 3.9% for hospitalizations and 2.1% vs. 1.8% for procedures); these rates were not compared statistically.8 The rates of UC-related hospitalizations were lower with adalimumab compared to placebo (incidence ratio [IR]: 0.12 vs. 0.22, p=0.002)73 and with infliximab compared to placebo (20 vs. 40 events per 100 patient years [PY], p=0.003).91 In UNIFI, the rates of UC-related hospitalizations were numerically lower for ustekinumab 90 mg every eight weeks compared to placebo (1.7% vs. 5.7%); these rates were not statistically compared.99 The rates of UC-related procedures were lower with infliximab compared to placebo (21 vs. 34 events per 100 PY, p=0.03, respectively).91 However, there was no difference in colectomy rates between adalimumab and placebo (IR: 0.04 vs. 0.05, p-value: NR).73 In UNIFI, the rates of surgeries were numerically higher with placebo compared to ustekinumab 90 mg every eight weeks at week 52 (1.7% vs. 0.6%); these rates were not statistically compared.99 Quality of Life Trials measured health-related quality of life (HRQoL) using various measures including the IBDQ, the SF-36, and the EQ-5D. In general, results were presented on an overall basis alone; stratified findings for the biologic-naïve and biologic-experienced populations are reported where available. Inflammatory Bowel Disease Questionnaire As described previously, an MCID on the IBDQ has not been established in UC. However, a 16-point or better improvement from baseline is used in Crohn's disease, and a 170-point or better score is also indicative of Crohn's remission. These thresholds have been used in available UC trials. Rates of achieving a change ≥16 points ("IBDQ response") or a score ≥170 points ("IBDQ remission") at the end of maintenance were reported for VARSITY and placebo-controlled trials of adalimumab, golimumab, tofacitinib, ustekinumab, and vedolizumab.8,74,79,93,97,101,119 In the VARSITY head-to-head trial, the rates of achieving "IBDQ response" and "IBDQ remission" were higher with vedolizumab 300 mg compared to adalimumab 40 mg at the end of maintenance. In the placebo-controlled trials, adalimumab, golimumab, tofacitinib, and ustekinumab had higher rates of IBDQ response compared to placebo. Additionally, patients treated with tofacitinib, ©Institute for Clinical and Economic Review, 2020 Page 50 Final Report – Targeted Immune Modulators for UC Return to Table of Contents ustekinumab, and vedolizumab achieved higher rates of IBDQ remission compared to placebo at the end of the maintenance. Finally, stratification by prior biologic exposure was available for adalimumab and higher rates of IBDQ response were reported among biologic-naïve patients but not biologic-experienced patients. A summary of relevant results can be found in Appendix Tables D8 and D11. 36-Item Short Form Survey SF-36 was assessed in the placebo-controlled trials of infliximab, tofacitinib, and vedolizumab stratified by prior biologic exposure. The trials reported improvement in the quality of life by MCID thresholds and mean component summary scores. As described previously, MCID for the SF-36 PCS and MCS summary scores ranged from 1.6 to 7.0 and 2.3 to 8.7, respectively, depending on the approach.28 However, the MCID threshold on the SF-36 has not been established in UC; the MCID thresholds ranged from 3 to 5 points in the included trials. A higher proportion of patients achieved ≥3-point MCID for both physical and mental component scores with infliximab compared to placebo (PCS: 59.1% vs. 40.6%; MCS: 50% vs. 34%; p<0.05 for both comparisons) and ≥5-point (PCS: 49% vs. 32.4%; MCS: 43% vs. 29.9%; p<0.05 for both comparisons) improvement.75 Similarly, higher proportions of patients were reported to have achieved ≥5-point improvements on the physical component score with vedolizumab every eight weeks compared to placebo (65% vs. 48%).74 Further, among the biologic-naïve population, significant improvements in mean PCS and MCS patients were reported with infliximab, tofacitinib, and vedolizumab, compared to placebo at the end of induction and maintenance phases.75,74,83 Additionally, in the biologic-experienced population, statistically significant improvement in PCS and MCS scores was reported with tofacitinib compared to placebo.83 A summary of relevant results can be found in Appendix Tables D9 and D12. EuroQol-5D EQ-5D results were reported in vedolizumab trials (GEMINI 1 and VISIBLE 1).77,92 The EQ-5D results were reported as the differences in mean scores between vedolizumab and placebo for the utility index and MCID for the VAS. While an MCID threshold has not been established in UC, a change of ≥10 in the MCID score was used as the threshold in the GEMINI 1 trial as per the MCID estimates ranging from 4.2-14.8 established for Crohn's disease.28 In biologic-naïve patients, significant differences in the EQ-5D VAS (10.6, 95% CI: 4.9-16.3) and utility scores (0.062, 95% CI: 0.003-0.120) were reported in the vedolizumab treated group compared to placebo at the end of the maintenance phase.74 Similar results were reported for EQ-5D VAS in the vedolizumab IV group in the VISIBLE 1 trial (LS mean difference: 13.1, 95% CI: 5.5-20.8), compared to placebo.92 ©Institute for Clinical and Economic Review, 2020 Page 51 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Work Productivity and Activity Impairment Three trials assessed work productivity among patients living with UC using the WPAI-UC questionnaire at the end of induction.96,99,106 Work productivity in the trials was reported as the mean percent change from baseline, at the end of induction and maintenance, for the overall population. MCID thresholds for UC were not reported. At the end of induction, patients treated with tofacitinib and ustekinumab were associated with statistically significant reductions in work- related impairments. At the end of the maintenance phase, statistically significant improvements with tofacitinib 5 mg and 10 mg, ustekinumab, and vedolizumab were reported for some or all components of the WPAI relative to placebo (see Appendix, Tables D13 and D14 for details). Improvements achieved at the end of induction were consistent through the end of maintenance. However, the WPAI-UC score of patients in the placebo group for all four domains worsened (increased) at the end of the maintenance phase. Special Populations Pediatric Population The efficacy and safety of TIMs for UC have not been studied widely in children and adolescents. Among the TIMs of interest for our review, only infliximab has been approved by the FDA for use among patients ages <18 years. Only one Phase III, open-label RCT evaluating the efficacy and safety of infliximab in both induction (open-label) and maintenance phases was identified; this was, in fact, a comparison of multiple doses of infliximab with no placebo or active comparator.107 No other RCTs or comparative observational studies were identified based on the inclusion criteria for our review, including the intervention of interest, outcomes of interest, and minimum sample size (n≥20) requirement for the UC cohort. The patient population was six to 17 years old living with moderate-to-severe UC, with a median PUCAI of 55 and a median Mayo Score of 8.0. The patients received infliximab open-label dosing of 5 mg/kg, before being randomized equally based on the response at week eight to one of the two infliximab regimens (i.e., infliximab 5 mg/kg every eight weeks or every 12 weeks). Across treatment groups, the mean age was 15 years, was primarily female, and had a mean disease duration of one to two years. Patients were also required to have prior experience with at least one conventional agent (including aminosalicylates, azathioprine, mercaptopurine, or oral or IV corticosteroids). All the enrolled patients received an open-label induction regimen of infliximab 5 mg/kg. At week eight, the clinical response defined by Mayo Score was achieved by 73.3% (44 of 60 patients). Clinical remission at week eight was achieved by 40% (24 of 60) and 33.3% (17 of 51) of patients ©Institute for Clinical and Economic Review, 2020 Page 52 Final Report – Targeted Immune Modulators for UC Return to Table of Contents based on Mayo Score and PUCAI, respectively. Remission was maintained at comparable levels at weeks 30 and 54 among patients in the every-eight-week group at rates that were numerically higher than those in the every-12-week group. Of all the treated patients in this RCT, 95% experienced more than one adverse event. The number of patients experiencing at least one serious adverse event was similar in both infliximab every- eight-week and every-12-week arms: 18.2% (4/22) and 21.7% (5/23), respectively. More than half of patients treated with infliximab reported the occurrence at least one infection (51.7%). In addition, we identified one observational study conducted in Australia that compared colectomy rates among 204 children diagnosed with UC in two time periods: 2005 to 2010 (group 1, n=71) and 2011 to 2016 (group 2, n=133). The 2011 cutoff was chosen based on the timing of approval of infliximab for children in Australia. The two-year cumulative probability of colectomy significantly decreased between these two time periods from 14% to 3% (p=0.02). The use of infliximab was independently associated with the decreased rates of colectomy (OR: 3.7; 95% CI: 1.1 to 11.7, p=0.02).108 Geriatric Population Geriatric patients with IBD (UC and Crohn's disease) can be divided into two groups: patients with long-standing disease (i.e., first diagnosis at a younger age) and patients with a late onset of disease.120,121 The symptoms of disease activity including abdominal pain, fecal urgency, and rectal bleeding are largely similar to those of younger patients.20,122 While the prevalence of UC in elderly populations has been increasing steadily, the availability of evidence is very limited in elderly patients as compared to younger patient populations. Most published trials do not report outcomes stratified by age (>65 years). Only two trials included in this review reported subgroup analyses stratified by age group, one head-to-head trial8 and one placebo-controlled trial (GEMINI 1).123 In GEMINI 1, of the total randomized population (n=895), only 15 patients ages 65 years or older were included. At the end of maintenance, one patient in each of the vedolizumab every-eight- week and placebo arms and two patients in vedolizumab every-four-week arm achieved clinical remission. Further, in VARSITY,8 of the total randomized patients (n=769), a total of 19 patients over the age of 65 years were randomized to the vedolizumab group only. At the end of maintenance, 26.3% of these patients achieved clinical remission, with 42.1% achieving endoscopic improvement. Among the five elderly patients receiving corticosteroids at baseline, only one patient achieved corticosteroid-free remission. ©Institute for Clinical and Economic Review, 2020 Page 53 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Extraintestinal Manifestations Systemic EIMs can affect other organ systems (including dermatological, hepatopancreatobiliary, ocular, oral, and musculoskeletal) and occur independently of colon-related disease activity. Despite the increased interest in assessing the role of TIMs in the effective management of EIMs, the evidence base for the effects of treatment on these outcomes is very limited. In the updated search, we identified a subgroup analysis of the three double-blind RCTs (OCTAVE 1, 2, and SUSTAIN) for tofacitinib. In OCTAVE Induction 1&2 and OCTAVE SUSTAIN, 27.0% and 9.0% of patients had experienced EIMs at baseline, respectively. The most common active EIM was peripheral arthritis, for which the majority of patients in the induction and maintenance periods reported either no change or improvement from baseline.124 Comparative Observational Studies Nine comparative observational studies met our inclusion criteria (sample size ≥500 in adults) and were deemed to be of high quality. All studies included at least 500 UC patients (before propensity- score matching), presented results from analyses adjusted for disease characteristics, and generally had similar follow-up across groups. The key study characteristics are listed in Table 4.18. Eight of the studies had a retrospective study design,109-111,113,116,117 and one was prospective.112 Seven studies compared originator TIMs to each other,109,110,113,116,117 one study compared the infliximab originator to an infliximab biosimilar (infliximab-dyyb; CT-P13),111 and one study compared infliximab to conventional therapy.112 ©Institute for Clinical and Economic Review, 2020 Page 54 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.18. Key Characteristics and Design of Included Comparative Observational Studies109-117 Study Design Location Population Comparison Outcomes • N=804 Sandborn Retrospective • ≥18 years with moderate-to- • Partial Mayo Score US ADA vs. IFX 2016 chart review severe UC • Remission • TNF-naïve A propensity- • N=1,719 (before matching); • Hospitalizations Singh matched 275 (after matching) • Colectomy Denmark ADA vs. IFX 2017 retrospective • ≥15 years with UC • Steroid use registry study • Biologic-naïve • Serious infections • N=1,400 (before matching); A propensity- • Hospitalizations 816 after matching Singh matched • Colectomy US • ≥18 years with UC ADA vs. IFX 2016 database study • Steroid use • Had not received biologic (Optum) • Serious infections therapy in prior 12 months Retrospective • Colectomy • N=862 Han database study South • ER visits • UC: age not specified ADA vs. IFX 2020* (administrative Korea • Hospitalization • Biologic-naïve claims) • Steroid use • N=3,562 ADA vs. IFX vs. Retrospective • Remaining steroid • ≥18 years with UC GOL vs. VEDO Long 2019 database study US free • Had not received biologic vs. IMM (IBM) • Hospitalizations therapy in prior 12 months therapy • N=2,239 Prospective • ≥18 years with moderate-to- IFX vs. Panes Multi- • AEs including serious registry study severe UC conventional 2019 national infection (OPUS) • IFX-naïve or IFX-free for ≥90 therapy days • N= 527 (VEDO pts= 325; anti- • Effectiveness and AEs Yarur Retrospective Multi- TNF-202) VEDO vs. TNFs including serious AEs 2019* chart review national • ≥18 years with UC and serious infections • Biologic-naïve • N=26,505 Retrospective Dubinsky • ≥18 years with IBD (CD, • Rate of newly database study US VEDO vs. TNFs 2018 n=18,055; UC, n=8,450) diagnosed EIMs (Truven) • TNF-naïve • Composite endpoint (death, UC-related Retrospective • N=3,112 surgery, all-cause Meyer database study France • ≥15 years with UC IFX vs. IFX-dyyb hospitalization, or 2019 (administrative • IFX naïve reimbursement of claims) another TIM) • Serious infections ADA: adalimumab, AE: adverse event, EIM: extraintestinal manifestation, ER: emergency room, GOL: golimumab, IBD: inflammatory bowel disease, IFX: infliximab, IMM: immunomodulator, N: number, OPUS: Observational Post-Marketing Ulcerative Colitis Study, TIM: targeted immune modulator, TNF: tumor necrosis factor, UC: ulcerative colitis, US: United States, VEDO: vedolizumab *Identified in updated search. ©Institute for Clinical and Economic Review, 2020 Page 55 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Four retrospective studies compared infliximab and adalimumab (Singh 2016, Singh 2017, Sandborn 2016, and Han 2020). Singh 2017 and Singh 2016 were similarly designed retrospective, propensity score-matched studies comparing adalimumab and infliximab on outcomes including hospitalizations, colectomy, steroid use, and serious infections.114,115 Singh 2017 included 1,719 biologic-naïve patients with UC in Denmark. After propensity score matching based on variables including demographic characteristics and disease severity, 275 patients were included in the analysis, 171 in the infliximab group and 104 in the adalimumab group. Adalimumab was shown to have a higher risk of all-cause hospitalizations compared to infliximab (HR: 1.84; 95% CI: 1.18 to 2.85) and a trend towards a higher risk of UC-related hospitalizations (HR: 1.71; 95% CI: 0.95 to 3.07).114 Rates of colectomy did not differ between the groups. Adalimumab had a higher risk of serious infections requiring hospitalization compared to infliximab (HR: 5.11 95% CI: 1.20 to 21.80).114 Singh 2016 included 1,400 patients with UC that had not been treated with a biologic in the previous 12 months. After propensity score matching, 816 patients were analyzed, 544 in the infliximab group, and 272 in the adalimumab group. In contrast to the evaluation above, there were no differences in the risk of all-cause hospitalizations, UC-related hospitalizations, steroid use, or serious infections.115 One study compared the safety of infliximab and conventional therapy (Panes 2019).112 This study was a prospective study, called the Observational Post-Marketing Ulcerative Colitis Study (OPUS), in 2,239 patients treated with infliximab (n=1,059) or conventional therapy (n=1,180). Patients were followed for up to five years. In the time-to-event analysis adjusted for multiple patient characteristics, such as disease severity, infliximab was shown to have a significantly higher risk of serious infections compared to conventional therapy (hazard ratio [HR]: 1.98; 95% CI: 1.34 to 2.91). Sandborn 2016113 was a chart review of 804 TNF-naïve adults with UC that compared the effectiveness of adalimumab (n=380) and infliximab (n=424) in achieving remission, as measured by partial Mayo Score. At six months, the rates of remission with adalimumab and infliximab were similar (76.8% vs. 71.0%, respectively, no statistical differences). Additionally, in a time-to-event analysis adjusted for multiple patient characteristics, such as disease severity, the likelihood of remission was similar for adalimumab and infliximab (HR: 1.07; 95% CI: 0.87 to 1.31). Of note, the rates of clinical remission observed in this real-world study are higher than those observed in RCTs. The authors suggest that higher rates of remission observed in this study could be attributed to a number of factors; these factors include using the partial Mayo Score as opposed to full the Mayo Score to measure remission, the inclusion of only biologic-naïve patients, and shorter disease duration at baseline (mean disease duration in this study was 4.4 to 4.8 years). Han 2020 was a database study of 862 biologic-naïve patients with UC treated with infliximab (n=630) or adalimumab (n=232) in South Korea. After a median follow-up of 1.8 years, there were no significant differences between infliximab and adalimumab in the risk of colectomy (HR: 1.87; 95% CI: 0.30 to 11.63), emergency room visits (HR: 1.58; 95% CI: 0.79 to 3.16), hospitalizations (HR: 0.83; ©Institute for Clinical and Economic Review, 2020 Page 56 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 95% CI: 0.59 to 1.17), or corticosteroid use (HR: 1.16; 95% CI: 0.76 to 1.78) in analyses adjusted for multiple patient characteristics. Three retrospective studies compared vedolizumab and the TNF inhibitors; one study compared all agents to each other (Long 2019) and two compared vedolizumab to the TNF inhibitors as a class (Dubinksy 2018 and Yarur 2019). Long 2019110 was a database study in 3,562 adults with UC that compared the effectiveness of adalimumab (n=1,291), golimumab (n=127), infliximab (n=810), and vedolizumab (n=103); additionally, the TIMs were also compared to immunosuppressant therapy alone (N=1,231) in unadjusted analyses. The proportions of patients remaining steroid-free and the proportion of patients experiencing a UC-related hospitalization at 12 months are reported in Table 4.19. Results are shown for unadjusted analyses as well as analyses adjusted for multiple patient characteristics, such as disease severity. The significance of the differences of the rates is only reported in adjusted analyses; adalimumab was chosen as the reference cohort given its larger sample size. Additionally, the rates with immunosuppressant therapy are only reported in the unadjusted analysis for remaining steroid-free. In the adjusted analysis, the proportion of patients remaining steroid-free at 12 months was moderately higher for infliximab compared to adalimumab (43.9% vs. 39.4%, p<0.05). In contrast, the adjusted proportion of patients with a UC-related hospitalization was higher with infliximab compared to adalimumab (20.4% vs. 16.3%, p<0.05). Table 4.19. Effectiveness Outcomes from Long 2019 at 12 Months110 Remaining Steroid-Free UC-Related Hospitalizations Unadjusted Unadjusted Treatment N Adjusted Analysis* Adjusted Analysis* Analysis Analysis % % p-Value % % p-Value ADA 1,291 43.6 39.4 Reference 12.9 16.3 Reference GOL 127 41.7 38.2 NS 8.7 10.5 NS IFX 810 48.3 43.9 <0.05 18.3 20.4 <0.05 VEDO 103 46.6 41.4 NS 9.7 15.6 NS IMM 1,231 52.7 NR NR NR NR ADA: adalimumab, GOL: golimumab, IFX: infliximab, IMM: immunomodulator, N: number, NR: not reported, NS: not specified, UC: ulcerative colitis, VEDO: vedolizumab *Adjusted for patient characteristics, such as disease severity. Dubinsky 2018109 was a retrospective database study of 26,505 TNF-naïve adults with IBD (Crohn's disease, n=18,055; UC, n=8,450) conducted in the US. The study compared vedolizumab and TNF inhibitors for incidence of EIMs. In patients with UC, there were no significant differences in the rate of newly diagnosed EIMs between the vedolizumab group (n=554) and the TNF inhibitor group (n=7,896) in an analysis adjusted for disease characteristics (incidence rate ratio [IRR]: 1.20, 95% CI: 0.91 to 1.59). However, UC patients were reported to be 3.67 times more likely to develop aphthous stomatitis with vedolizumab compared to TNF inhibitors (IRR, 3.67; 95% CI, 1.30-10.34). ©Institute for Clinical and Economic Review, 2020 Page 57 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Yarur 2019117 was a comparative chart review (the EVOLVE study) of 527 biologic-naïve UC patients treated with vedolizumab (n=325) or TNF inhibitors (n=202 [adalimumab: 58, infliximab: 120, and golimumab: 24]). At 24 months, the cumulative rates of clinical response (91% vs. 86%), remission (79% vs. 66%), and mucosal healing (92% vs. 84%) did not differ significantly between the vedolizumab and TNF inhibitor groups. However, the risk of experiencing UC exacerbation (HR: 0.7, 95% CI: 0.5 to 0.9, p=0.02) and serious adverse events (HR: 0.5, 95% CI: 0.3 to 0.8, p<0.01) was statistically significantly lower for vedolizumab in comparison to the TNF inhibitors. Lastly, Meyer 2019111 compared the infliximab originator product to an infliximab biosimilar (infliximab-dyyb/CT-P13). This study was a retrospective database study in 3,112 infliximab-naïve patients treated with infliximab (n=1,434) or infliximab-dyyb (n=1,678) in France. The main outcome was a composite endpoint (death, UC-related surgery, all-cause hospitalization, or reimbursement of another TIM). The cumulative incidence (95% CI) of the composite endpoint was 43.0% (40.5-45.6) and 57.5% (54.9‐60.0) for infliximab at 12 and 24 months, respectively, compared to 45.1% (42.7-47.5) and 59.8% (57.5‐62.1) for infliximab-dyyb; these rates did not statistically differ between groups. Regarding safety, however, fewer serious infections were observed with infliximab-dyyb compared to infliximab (HR: 0.65; 95% CI: 0.48 to 0.88). Harms • Severe and serious adverse events were rare during the induction and maintenance phases. Upper-respiratory tract infections and headaches were the most reported adverse events across the TIMs. There was no indication of increased rates of serious infections, tuberculosis, and mortality for any of the agents in available RCTs. Randomized Controlled Trials Data on adverse events and discontinuations due to adverse events as well as administration reactions and infections observed in induction (see Appendix Table D15) and maintenance (see Table 4.20) phases of the clinical trials have been summarized. Since safety data were not consistently reported according to stratification by prior biologic exposure, we summarized these data on the overall population for the relevant doses used in each trial phase. Mortality rates across trials were low during both induction as well as maintenance phases. Overall, one event was reported in PURSUIT-SC (0.3%), GEMINI (2.2%), OCTAVE 1 (0.2%), and UNIFI (0.3%) during the induction phase. In the maintenance phase, the PURSUIT-M trial reported three (1.9%) deaths. Of note, the label for tofacitinib was modified in July 2019 given the results from a long- term clinical trial for safety in rheumatoid arthritis patients.10 Modifications included warnings related to thrombotic events (deep vein thrombosis, arterial thrombosis, and pulmonary embolism) as well as cardiovascular mortality in patients receiving the 10 mg twice daily dose. We note that the three TNF inhibitors (adalimumab, golimumab, and infliximab) as well as tofacitinib carry a black ©Institute for Clinical and Economic Review, 2020 Page 58 Final Report – Targeted Immune Modulators for UC Return to Table of Contents box warning in their FDA labels for an increased risk of lymphomas and other malignancies observed in children, adolescents, and/or young adults based on clinical trials and real-world evidence for these TIMs when studied for other indications (e.g., rheumatoid arthritis).11-14 In our evidence base for UC, overall rates of new malignancy were very low (<3%) and no study in our sample indicated an increased risk for any TIM. The most frequently reported adverse events across the placebo-controlled trials were mild infections (e.g., upper respiratory tract infections), administration reactions, including injection site reaction, headache, and nausea. In the induction trials, the rates of adverse events were low across the trials. Analyses of the maintenance phase of ACT 1 indicated a higher proportion of patients experiencing upper respiratory tract infections in the infliximab 10 mg/kg arm (23.8%) than in the infliximab 5 mg /kg arm (16.5%), but rates were comparable to the placebo arm (23.1%). In the VISIBLE 1 trial for vedolizumab, a higher proportion of patients experienced adverse events in the placebo arm than the vedolizumab IV arm (32.1% vs. 11.1%). The rates of serious infection, serious adverse events, and discontinuations due to adverse events were generally low and comparable in the head-to-head trial comparing vedolizumab with adalimumab. In placebo-controlled trials, the risk of severe or serious adverse events and serious infections was low and generally comparable between the treatment and placebo groups. There was one report of tuberculosis in Suzuki 2014 (1.1%). There was no evidence on the increased risk of serious infections in the placebo-controlled trials. In the infliximab trials, no serious infections were reported for the placebo arms of the Asian trials (Jiang 2015, Kobayashi 2016, and NCT01551290).68,80,81 In the PURSUIT-M trial, the rate of serious infections was marginally higher in the golimumab treatment arms than the placebo arm (3.2% vs. 1.9%), but this was not statistically tested.94 Further, the rates of new-onset autoimmune disease and demyelinating disease were low across the trials. Immunogenicity to TIM therapy (i.e., development of neutralizing antibodies) is a relatively commonly identified effect that may negatively impact rates of response and/or remission.125,126 The development of neutralizing antibodies was reported in nine trials. Rates ranged tightly from 3- 6% at some point after the first infusion, with no one TIM showing significantly higher rates.80,90 Long-Term Extensions of Randomized Controlled Trials Most adverse events in the open-label extension studies were mild-to-moderate in severity. The long-term safety data from the RCTs and open-label extension studies report low rates of serious infections (see Appendix Table D16). Higher rates for discontinuations (249 per 110 PY), serious adverse events (414 per 100 PY), and serious infections (79 per 100 PY) were reported in the ULTRA-2 adalimumab long-term trial, however. Furthermore, the FDA label for tofacitinib was modified in July 2019 given the results from the long-term clinical trial for safety in rheumatoid arthritis patients.10 In the OCTAVE open-label extension study, rates of serious infection (3.4%) and ©Institute for Clinical and Economic Review, 2020 Page 59 Final Report – Targeted Immune Modulators for UC Return to Table of Contents serious adverse events (11.8%) were slightly higher than those reported for tofacitinib 10 mg during the RCT. The incidence rate of malignancy for excluding nonmelanoma skin cancer was reported to be 0.69 and 0.75 per 100 patient years for tofacitinib 5 mg and 10 mg, respectively.103 The risk of thromboembolic events during tofacitinib exposure of up to 6.1 years was estimated for deep vein thrombosis (IR: 0.04, 95% CI: 0.00‐0.23) and pulmonary embolism (IR:0.16; 95% CI: 0.04‐0.41).104 In the PURSUIT maintenance extension study, seven of nine (overall IR: 0.44, 95% CI: 0.18, 0.90) deaths occurred in patients treated with golimumab 100 mg through week 212.84 The safety profile for ustekinumab was largely similar to that observed in the RCT. No malignancy was reported, and the rates of serious infections remained low (1.4%) in the long-term extension (through week 96) among responders to ustekinumab IV induction.105 ©Institute for Clinical and Economic Review, 2020 Page 60 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.20. Adverse Events During the Maintenance Phase AE Infusion or Overall Serious Antibodies Trial Arm N Any AE SAE Leading Injection Site URTI Headache Infections Infections n/N (%) to D/C Reaction Head-to-Head VEDO 300 mg 383 240 (62.7) 42 (11.0) 10 (2.6) 103 (23.4) 7 (1.6) NR 55 (12.5) NR NR VARSITY ADA 40 mg 386 267 (69.2) 53 (13.7) 13 (3.4) 124 (34.6) 8 (2.2) NR 65 (18.1) NR NR Infliximab IFX 5 mg/kg 121 106 (87.6) 26 (21.5) 10 (8.3) 53 (43.8) 3 (2.5) 12 (9.9) 20 (16.5) 22 (18.2) 14/243 ACT 1 IFX 10 mg/kg 122 111 (91.0) 29 (23.8) 11 (9.0) 60 (59.2) 8 (6.6) 15 (12.3) 29 (23.8) 18 (14.8) (5.8) PBO 121 103 (85.1) 31 (25.6) 11 (9.1) 47 (38.8) 5 (4.1) 13 (10.7) 28 (23.1) 27 (22.3) -- IFX 5 mg/kg 121 99 (81.8) 13 (10.7) 2 (1.7) 33 (27.3) 2 (2.5) 14 (11.6) 16 (13.2) 19 (21.7) 12/241 ACT 2 IFX 10 mg/kg 120 96 (80.0) 11 (9.2) 5 (4.2) 34 (28.3) 3 (1.7) 14 (11.7) 14 (11.7) 26 (15.7) (5.0) PBO 123 90 (73.2) 24 (19.5) 12 (9.8) 29 (23.6) 1 (0.8) 10 (8.1) 14 (11.4) 18 (14.6) -- IFX 5 mg/kg 41 17 (41.5) 3 (7.3) 1 (2.4) 6 (14.6) 1 (2.4) 3 (7.3) NR NR 2/40 (5.0) Jiang 2015 PBO 41 16 (39.0) 4 (9.8) 2 (4.9) 5 (12.2) 0 (0) 2 (4.9) NR NR -- Kobayashi IFX 5 mg/kg 73 100 (96.2) 18 (17.3) 7 (6.7) 62 (59.6) 1 (1) 16 (15.4) NR NR NR 2016 PBO 72 94 (90.4) 19 (18.3) 8 (7.7) 51 (49.0) 2 (1.9) 11 (10.6) NR NR -- IFX 5 mg/kg 50 33 (66.0) 7 (14.0) 4 (8.0) 13 (26.0) 0 (0) NR NR NR NCT01551290 PBO 49 31 (63.3) 4 (8.2) 2 (4.1) 7 (14.3) 0 (0) NR NR NR -- Adalimumab ADA 40 mg 248 213 (82.8) 31 (12.1) 23 (8.9) 116 (45.1) 4 (1.6) 31 (12.1) NR NR 7/245 (2.9) ULTRA 2 PBO 246 218 (83.8) 32 (12.3) 34 (13.1) 103 (39.6) 5 (1.9) 10 (3.8) NR NR -- Suzuki 2014 ADA 40 mg 177 538 (547.9) 33 (33.6) 22 (13.4) 134 (136.5) 8 (8.1) 20 (20.4) NR NR 5/177 (2.8) (Per 100 PY) PBO 96 273 (609.4) 14 (31.3) 6 (22.4) 70 (156.3) 2 (4.5) 4 (8.9) NR NR -- Golimumab 32/1103 GOL 100 mg 154 113 (73.4) 22 (14.3) 14 (9.1) 60 (39.0) 5 (3.2) 11 (7.1) 9 (5.8) 12 (7.8) PURSUIT-M (2.9) PBO 156 103 (66.0) 12 (7.7) 10 (6.4) 44 (28.2) 3 (1.9) 3 (1.9) 4 (2.6) 14 (9.0) -- GOL 100 mg 32 31 (96.9) 1 (3.1) 0 (0) 21 (65.6) NR 6 (18.8) NR NR 4/63 (2.5) PURSUIT-J PBO 31 22 (71.0) 4 (12.9) 1 (3.2) 11 (35.5) NR 0 (0) NR NR -- ©Institute for Clinical and Economic Review, 2020 Page 61 Final Report – Targeted Immune Modulators for UC Return to Table of Contents AE Infusion or Overall Serious Antibodies Trial Arm N Any AE SAE Leading Injection Site URTI Headache Infections Infections n/N (%) to D/C Reaction Vedolizumab VEDO 300 mg 125 101 (81) 11 (9) NR 90 (72) 2 (2) 10 (11) 12 (10) NR NR q4w GEMINI 1 VEDO 300 mg 122 100 (82.0) 10 (8.0) NR 87 (71.0) 3 (2.0) 7 (5.7) 12 (9.8) NR NR q8w PBO 126 106 (84.0) 20 (16.0) NR 89 (71.0) 4 (3.0) 2 (1.6) 13 (10.3) NR -- VEDO 300 mg 41 36 (87.8) 4 (9.8) 2 (4.9) NR 9 (2.4) NR 3 (7.3) 2 (4.9) NR Motoya 2019 PBO 42 33 (78.6) 3 (7.1) 6 (14.3) NR 10 (2.4) NR 1 (2.4) 1 (2.4) -- VEDO 300mg 54 41 (75.6) 7 (13.0) 2 (3.7) 15 (27.8) NR 1 (1.8) 2 (3.7) 0 3/54 (6.0) VISIBLE 1 q8w PBO 56 43 (76.8) 6 (10.7) 5 (8.9) 14 (25.0) NR 0 1 (1.8) 6 (10.7) -- Tofacitinib TOF 5 mg 198 143 (72.2) 10 (5.1) 18 (9.1) 71 (35.9) 2 (1.0) NR NR 17 (8.6) NR OCTAVE TOF 10 mg 196 156 (79.6) 11 (5.6) 19 (9.7) 78 (39.8) 1 (0.5) NR NR 6 (3.1) NR SUSTAIN PBO 198 149 (75.3) 13 (6.6) 37 (18.7) 48 (24.2) 2 (1.0) NR NR 12 (6.1) -- Ustekinumab 23/505 UST 90mg q8w 176 136 (77.3) 15 (8.5) 5 (2.8) 86 (48.9) 3 (1.7) 5 (2.8) 16 (9.1) 18 (10.2) UNIFI (4.6) PBO 175 138 (78.9) 17 (9.7) 20 (11.4) 81 (46.3) 4 (2.3) 4 (2.3) 8 (4.6) 7 (4.0) -- ADA: adalimumab, AE: adverse event, D/C: discontinuation, GOL: golimumab, I: induction, IFX: infliximab, kg: kilogram, M: maintenance, mg: milligram, NA: not applicable, NR: not reported, PBO: placebo, PY: patient year, q4w: every 4 weeks, q8w: every 8 weeks, SAE: serious adverse event, TOF: tofacitinib, URTI: upper respiratory tract infection, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 62 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Controversies and Uncertainties While the evidence for each TIM of focus in this review suggests a potential net health benefit in comparison to conventional therapy among patients with moderate-to-severe UC, there are several concerns with trial design and results as well as gaps in evidence that bear mention. First and foremost, a single trial of the 19 RCTs identified included direct evidence comparing the TIMs. Our comparisons were therefore driven almost exclusively by the conduct of NMAs. Our NMAs were limited by several factors, including differences in the definitions of biologic-naïve and biologic-experienced populations. Likewise, the network of evidence was sparse for some of our populations of interest (e.g., maintenance data in the biologic-experienced population), adding uncertainty to our relative estimates of response and remission. As noted by the Crohn's and Colitis Foundation, there is currently very limited information with which to ascertain the optimal sequence of treatment. Some insight can be gleaned from assessing results for the biologic-naïve and biologic-experienced populations, but the definition of "experienced" varied across trials. In some cases, the focus was on failure to achieve response alone, while in others, those achieving and losing response were also included. Still, other trials focused on experience only with certain drug classes (e.g., TNF inhibitors). Finally, some agents have no evidence in the populations of interest, posing challenges in the interpretation and application of data. For example, clinicians mentioned great interest in the use of infliximab following failure by vedolizumab, but there are no randomized data for this treatment sequence, likely due to infliximab's status as the first biologic agent approved for UC. We note other differences in population characteristics and measurement in available trials. While populations were broadly similar with respect to age and baseline disease severity, variation was noted in disease duration as well as use of conventional therapy at baseline. Regarding measurement, some studies used central endoscopy reading to inform Mayo Scores, while others allowed these readings to occur at study sites. While variability in local readings is unlikely to have biased within-trial comparisons, the impact on indirect comparisons in our NMA is less clear. We also had only partial Mayo Scores to inform maintenance rates used in our NMA in the one head-to- head trial available. In addition, substantial variation has been noted in definitions of endoscopic factors, histologic features, and serum or fecal biomarkers across UC trials.15 As is often the case with therapies for chronic inflammatory diseases, evidence of long-term safety comes from the conduct of observational studies using large datasets. While such data exist for the more established therapies in our set, there are limited to no long-term data on newer UC therapies, such as tofacitinib, ustekinumab, and vedolizumab. We note, for example, that the FDA recently changed the label for tofacitinib to allow use in UC only after initial TNF inhibitor therapy and added boxed warnings for elevated risks of pulmonary embolism and all-cause mortality. But these changes were based on observations in populations with rheumatoid arthritis, not UC. ©Institute for Clinical and Economic Review, 2020 Page 63 Final Report – Targeted Immune Modulators for UC Return to Table of Contents There are uncertainties around not only what has been measured in available UC studies but what has not. Concerns were raised around the EIMs of the disease, but we found very little empiric data on the frequency of these manifestations and, more importantly, the impact of treatment on them. Other concerns raised by patients, such as chronic pain and fatigue as well as fecal urgency, have gone largely unaddressed by available studies. Finally, even though a substantial proportion of cases of UC are diagnosed in childhood and adolescence, there is essentially a complete absence of robust comparative evidence to inform treatment strategies in this population. Only one of the TIMs, infliximab, carries an FDA indication for treatment in patients age <18 years, and this comes only on the strength of a dose-finding study without comparison to alternative TIMs or even to conventional therapy. While there is no obvious reason to believe that TIMs would operate differently in children and adolescents, the absence of comparative data among those most likely to benefit from early intervention is nevertheless troubling. 4.4 Summary and Comment As mentioned previously, we organized our review and synthesis of the evidence by a) prior biologic use and b) use of therapy during the induction versus maintenance phases of treatment. However, while clinical guideline statements also employ these strata in their reviews of the evidence, recommendations are not specific to patient population or phase of treatment. For example, if a therapy is recommended to induce remission, continued maintenance treatment with that therapy is also recommended. In addition, FDA labeling and payer coverage policies (see Section 3) do not make these distinctions. Therefore, we produced a single set of evidence ratings for all therapy comparisons using the ICER Evidence Rating Matrix, considering therapy performance across all strata. We note the few instances in which our evidence ratings apply to a specific population due to FDA labeling or evidentiary limitations. As mentioned earlier, there was only one head-to-head trial available in the evidence base, so all remaining comparisons between TIMs were based on our assessment of their benefits in placebo-controlled trials and observational studies as well as our NMAs. Further, safety data were not consistently reported across studies and populations and were therefore not quantitatively synthesized. Nevertheless, safety concerns identified in RCTs, open-label extension studies, and observational or real-world evidence played a role in determining the evidence ratings for each agent. Note that we have opted to rate infliximab-dyyb and infliximab-abda, the two biosimilars to infliximab, as comparable ("C") to the originator product, and so the evidence ratings that follow involve comparisons to infliximab as a single entity. This rating is based on the FDA's determination that the biosimilars are therapeutically equivalent in UC. Finally, while there is no obvious reason to suggest that TIM performance would markedly differ in children and adolescents versus adults, we note the complete lack of robust comparative evidence ©Institute for Clinical and Economic Review, 2020 Page 64 Final Report – Targeted Immune Modulators for UC Return to Table of Contents in pediatric populations for both between-TIM and placebo-controlled comparisons. The ratings summarized below should therefore be considered relevant to adult populations only. Figure 4.1. ICER Evidence Rating Matrix ©Institute for Clinical and Economic Review, 2020 Page 65 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 4.21. Summary of ICER Evidence Ratings TIM Comparator Rating Infliximab Infliximab biosimilars C Infliximab Placebo A* Golimumab Placebo A* Tofacitinib Placebo B+† All other TIMs Placebo A Vedolizumab Adalimumab B+ Ustekinumab Adalimumab C+ Infliximab Adalimumab C+* Tofacitinib Adalimumab P/I† Vedolizumab Golimumab C+* All other TIM -- I Comparisons TIM: targeted immune modulator *Biologic-naïve only. †Biologic-experienced only. Targeted Immune Modulators versus Placebo Across populations stratified by prior biologic use and phase of treatment, the addition of TIMs to conventional therapy consistently performed better than conventional therapy alone in RCTs. For all placebo comparisons other than tofacitinib, we felt there was high certainty of a substantial net benefit (an "A" rating), as no serious safety signals were evident from RCTs (note that this rating applies to the biologic-naïve population for golimumab and infliximab, given a lack of evidence in biologic-experienced patients). Given the recent black-box warnings regarding thrombosis and cardiovascular death for tofacitinib, however, our certainty was only moderate. We judged the evidence to indicate a small or substantial benefit ("B+"), given the uncertainty in how the benefits seen with tofacitinib would trade-off against its risks. We further note that this rating is applicable to the biologic-experienced population only given the recent FDA labeling changes. Between Targeted Immune Modulator Comparisons The only comparison between TIMs that involved direct head-to-head evidence is that of vedolizumab versus adalimumab. The VARSITY study showed substantial and statistically significant differences in remission, response, and other measures of health benefit in favor of vedolizumab. These findings were generally bolstered by the addition of indirect evidence in our NMAs. However, with only one head-to-head trial available, and variation in apparent effect size across the four NMAs, we concluded that there was only moderate certainty of a small or substantial net health benefit for vedolizumab in relation to adalimumab ("B+"). ©Institute for Clinical and Economic Review, 2020 Page 66 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Other comparisons to adalimumab are based on indirect evidence only. We found the evidence directionally consistent across the four populations for both infliximab (in biologic-naïve patients only) and ustekinumab to indicate a net health benefit that is at least comparable, and likely incremental, relative to adalimumab ("C+"). While we generally concluded the same for tofacitinib (in biologic-experienced patients only), the safety concerns associated with this agent resulted in a promising but inconclusive ("P/I") rating versus adalimumab. In comparison of vedolizumab and golimumab (among biologic-naïve patients only), NMA findings were directionally in favor of vedolizumab for both induction and maintenance, with results reasonably robust for maintenance. We judged this evidence to suggest a net health benefit for vedolizumab that was a least comparable, and likely incremental, relative to golimumab ("C+"). For all other comparisons between TIMs, there was variability in the rankings between them across our populations of interest, wide confidence intervals around the estimates of effect, or both. Because of this variability, we judged all these comparisons to reflect insufficient ("I") evidence of a net health benefit. ©Institute for Clinical and Economic Review, 2020 Page 67 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 5. Long-Term Cost Effectiveness 5.1 Overview A decision analytic model was developed for this evaluation, informed by key clinical trials and prior relevant economic models, to estimate the cost effectiveness of TIMs for moderate-to-severe UC in biologic-naïve and biologic-experienced populations. The model compared all eight treatments to each other and to conventional treatment. The base-case analysis took a health care sector perspective (i.e., focused on direct medical care costs only), over a lifetime time horizon. Due to uncertainty of treatment patterns over a lifetime time horizon, shorter time horizons of two, five, and 10 years were explored as additional scenario analyses. The model was structured as a Markov model with eight-week cycles, based on a common point of assessment in clinical trials to mark the end of induction and beginning of maintenance treatment. Costs and outcomes were discounted at 3% per year. The TIMs evaluated include: • Adalimumab • Golimumab • Infliximab • Infliximab-dyyb • Infliximab-abda • Tofacitinib • Ustekinumab • Vedolizumab (IV) Model outcomes included total costs, life years (LYs), quality-adjusted life years (QALYs), and equal value of life years gained (evLYG). Methods used to calculate evLYG can be found in Appendix E.127 5.2 Methods Model Structure A Markov model was developed in Microsoft Excel 365 consisting of the health states of active UC, clinical response without remission, clinical remission, post-colectomy (with and without complications), and death (Figure 5.1). The model structure and health states were chosen based on the disease course, the impact of treatment, and prior economic models in UC. Moderate-to- severe UC patients enter the model at the beginning of induction of a TIM or conventional treatment. At the end of induction, patients with response (both with and without remission) ©Institute for Clinical and Economic Review, 2020 Page 68 Final Report – Targeted Immune Modulators for UC Return to Table of Contents continue to receive the TIM or conventional treatment. Those without response or those who discontinue after initial response begin induction with a subsequent treatment, represented by a market basket of all TIMs with data in a biologic-experienced population. At the end of induction with subsequent treatment, responders continue to receive the subsequent treatment. Patients who do not respond during the induction phase of subsequent treatment discontinue treatment with TIMs and follow the transition probabilities of the conventional treatment arm (e.g., corticosteroids, other systemic immunomodulators). A proportion of patients in the active UC state are assumed to opt for colectomy during each cycle.128 Patients remain in the model until death. All patients can transition to death from all causes from any of the alive health states. In addition, patients can die from surgical complications of colectomy. Figure 5.1. Model Framework UC: ulcerative colitis Target Population The economic evaluation includes two target populations: 1) biologic-naïve and 2) biologic- experienced. These two groups are generally similar in age, gender, and weight but have been shown to differ in clinical response to TIMs. The cost-effectiveness and threshold prices were evaluated for both patient populations, with an eventual health benefit price benchmark (HBPB) ©Institute for Clinical and Economic Review, 2020 Page 69 Final Report – Targeted Immune Modulators for UC Return to Table of Contents weighted by the estimated proportions of patients with and without prior biologic use. When TIM efficacy data were not available for one of the populations or if a TIM was not labeled for use in a specific population, the TIM was not assessed in that population. Table 5.1. Baseline Population Characteristics Mean Age Percent Male Weight Source Biologic-Naïve 40 years 59% 74.3 kg GEMINI 177 Biologic-Experienced 40 years 59% 74.3 kg GEMINI 177 kg: kilogram Treatment Strategies The list of interventions was developed with input from patient organizations, clinicians, manufacturers, and payers. The full list of interventions is as follows: • Adalimumab • Golimumab • Infliximab • Infliximab-dyyb • Infliximab-abda • Tofacitinib • Ustekinumab • Vedolizumab (IV) • Conventional treatment (corticosteroids for induction followed by azathioprine or mercaptopurine) Key Model Characteristics and Assumptions Below is a list of key model choices in the base-case analysis. • Cycle length of eight weeks • Lifetime time horizon • Three percent annual discount rate for costs and outcomes • Probability of achieving response without remission or response with remission at the end of induction informed by the results of the ICER NMA • Given response at the end of induction, probability of being in clinical response without remission or clinical remission at the end of maintenance is informed by the results of the ICER NMA • Patients remaining in active UC at the end of induction will discontinue the treatment and initiate the next line of therapy ©Institute for Clinical and Economic Review, 2020 Page 70 Final Report – Targeted Immune Modulators for UC Return to Table of Contents • Constant per-cycle rate of discontinuation unrelated to clinical response (e.g., adverse events, other reasons) based on discontinuation rates for reasons other than lack of efficacy from RCTs • The model considers initial treatment and subsequent treatment with a TIM; after failure of subsequent treatment, patients are assumed to discontinue to conventional treatment • Constant per-cycle probability of elective colectomy only from the active UC health state • Direct health state costs, including hospitalization, emergency department visits, and outpatient visits, based on published claims analysis • Mortality based on US Life Tables, adjusted for elevated risk of mortality due to UC • Model includes one-time risk of mortality associated with colectomy • Costs and disutility of serious infection with TIMs and conventional treatment and complications of colectomy are included Our model includes several key assumptions stated in Table 5.2 below. Table 5.2. Key Model Assumptions Assumption Rationale Conventional treatment is represented by the control In the real world and in controlled trials, background arm of RCTs. conventional treatment is taken in addition to TIMs. Multiple TIMs may be tried over a lifetime time All patients will cycle through initial treatment and horizon; however, the focus of this evaluation is to subsequent treatment before discontinuing to estimate the cost effectiveness of the initial choice of conventional treatment. TIM in a biologic-naïve or biologic-experienced population rather than entire treatment pathways. Choice of the subsequent treatment is represented by a market basket of equal distribution of all TIMs with data in a biologic-experienced population. Incomplete data exists on the efficacy of specific Efficacy, safety, and cost are informed by an average treatment sequences; thus, the market basket of treatments in the biologic-experienced population. approach is taken to focus the analysis on the Probabilities for any given patient to receive any intervention of interest. given therapy in the market basket are assumed to be equal in the absence of contemporary real-world data on utilization patterns in the US. The transition probabilities for clinical response Data is only available for transitions from the end of without remission and clinical remission to induction to the end of maintenance for response subsequent health states in the maintenance phase (with and without remission) combined. are equal. RCT: randomized controlled trial, TIM: targeted immune modulator, US: United States ©Institute for Clinical and Economic Review, 2020 Page 71 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Model Inputs Clinical Inputs Efficacy data were derived from the results of the ICER NMAs of RCTs of TIMs in moderate-to- severe UC (one each for biologic-naïve and biologic-experienced), with data from the placebo arms used as a proxy for conventional treatment. Clinical Probabilities/Response to Treatment For each of the biologic-naïve and biologic-experienced populations, a set of transition probabilities was created based on the placebo group of the NMA for the induction phase (Cycle 1) and maintenance phases (Cycles 2+). In the induction phase, we calculated the probability of achieving response without remission and clinical remission at the end of induction. Conditional upon achieving response entering the maintenance phase, we calculated the probability of achieving clinical response without remission, clinical remission, or losing response at the end of maintenance. Table 5.3. Conventional Treatment Induction Outcomes Population Active UC Response Remission Source Biologic-Naive 65% 26% 9% ICER NMA Biologic-Experienced 74% 22% 4% ICER NMA ICER: Institute for Clinical and Economic Review, NMA: network meta-analysis, UC: ulcerative colitis Table 5.4. Conventional Treatment Maintenance Outcomes Biologic-Naïve Active UC Response Remission Source Clinical Remission 60% 14% 26% ICER NMA Clinical Response without Remission 60% 14% 26% ICER NMA No Response (Active UC) 88% 9% 3% ULTRA 2*93 Biologic-Experienced Active UC Response Remission Source Clinical Remission 73% 14% 13% ICER NMA Clinical Response without Remission 73% 14% 13% ICER NMA No Response (Active UC) 95% 4% 1% ULTRA 2*93 ICER: Institute for Clinical and Economic Review, NMA: network meta-analysis, UC: ulcerative colitis *Calculated based on the probability of clinical remission and clinical response without remission among placebo non-responders at the end of maintenance. Based on the results of the NMAs for the biologic-naïve population and biologic-experienced population, risk ratios for each TIM relative to placebo were calculated and applied to the probability of achieving clinical response (with and without remission) with conventional treatment (Table 5.5 through Table 5.8). The risk ratio for clinical response (with and without remission) at the end of the maintenance for each TIM relative to placebo was applied to the probability of achieving ©Institute for Clinical and Economic Review, 2020 Page 72 Final Report – Targeted Immune Modulators for UC Return to Table of Contents clinical response (with and without remission) with conventional treatment. The probability of maintaining response (with and without remission) at the end of the maintenance phase was then converted to eight-week cycles. Due to data limitations, we were unable to calculate the risk ratio for maintenance of response without remission specifically among those with response without remission at the start of induction for all treatments. Therefore, the same risk ratio for clinical remission and clinical response without remission at the end of maintenance was applied to patients entering the maintenance phase, regardless of entering maintenance in the clinical response without remission or clinical remission health states. We also note that, consistent with findings from Section 4, we assumed that clinical performance for infliximab biosimilars was identical to that of the originator product. Table 5.5. Induction Risk Ratios for the Biologic-Naïve Population Clinical Response without Remission Clinical Remission Placebo Reference Reference Adalimumab 1.24 (1.13, 1.36) 1.76 (1.38, 2.19) Golimumab 1.34 (1.22, 1.50) 2.41 (1.89, 3.08) Infliximab 1.39 (1.22, 1.58) 3.22 (2.60, 3.96) Infliximab-dyyb 1.39 (1.22, 1.58) 3.22 (2.60, 3.96) Infliximab-abda 1.39 (1.22, 1.58) 3.22 (2.60, 3.96) Ustekinumab 1.36 (1.21, 1.54) 2.66 (1.86, 3.73) Vedolizumab 1.37 (1.23, 1.55) 2.79 (2.18, 3.58) Table 5.6. Maintenance Risk Ratios for the Biologic-Naïve Population Response without Remission Response with Remission Placebo Reference Reference Adalimumab 1.09 (0.93, 1.20) 1.47 (0.92, 2.14) Golimumab 1.08 (0.85, 1.19) 1.35 (0.74, 2.04) Infliximab 1.09 (0.86, 1.20) 1.80 (1.13, 2.86) Infliximab-dyyb 1.09 (0.86, 1.20) 1.80 (1.13, 2.86) Infliximab-abda 1.09 (0.86, 1.20) 1.80 (1.13, 2.86) Ustekinumab 1.01 (0.49, 1.17) 2.22 (1.17, 3.57) Vedolizumab 1.08 (0.88, 1.20) 1.93 (1.22, 2.58) Table 5.7. Induction Risk Ratios for the Biologic-Experienced Population Response without Remission Response with Remission Placebo Reference Reference Adalimumab 1.04 (0.74, 1.37) 1.09 (0.57, 1.97) Tofacitinib 1.75 (1.49, 2.07) 4.10 (2.62, 6.18) Ustekinumab 1.76 (1.50, 2.09) 4.13 (2.71, 6.26) Vedolizumab 1.47 (1.17, 1.78) 2.38 (1.38, 3.80) ©Institute for Clinical and Economic Review, 2020 Page 73 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.8. Maintenance Risk Ratios for the Biologic-Experienced Population Response without Remission Response with Remission Placebo Reference Reference Adalimumab 1.42 (1.20, 1.69) 2.91 (1.72, 4.65) Tofacitinib 1.40 (1.18, 1.64) 2.49 (1.46, 3.79) Ustekinumab 1.40 (1.19, 1.65) 2.49 (1.49, 3.82) Vedolizumab 1.47 (1.17, 1.78) 3.48 (2.43, 5.03) Patients have a constant per-cycle risk of elective colectomy from the active UC health state. The per-cycle probability of colectomy is based on a cumulative rate of 25.4% at 20 years (95% CI, 19.8% to 30.8%) from UC diagnosis observed from a cohort of UC patients in Olmstead County, Minnesota from 1997 to 2004, converted to an annual probability of 1.45% (or 0.23% per eight-week cycle).128 Delayed Response/Extended Induction Evidence exists for a few TIMs that some additional patients achieve response if the induction period is extended beyond eight weeks. Furthermore, clinicians engaged during the scoping period of this review also commented that it is not uncommon in clinical practice to wait longer than eight weeks to determine response and non-response to TIMs. Therefore, delayed response was explored in a scenario analysis for TIMs where there is evidence of benefit with extended induction. In this scenario, a 16-week induction period was allowed for a proportion of patients to mirror real- world practice patterns. In the extended induction scenario, patients discontinue at week eight only for reasons other than lack of efficacy. The gain in response with the extended induction scenario is presented in Table 5.9. Detailed calculations to derive these estimates may be found in the Appendix. Table 5.9. Gain in Response or Remission at 16 Weeks for the Extended Induction Efficacy Scenario Clinical Response Clinical Remission Source without Remission Adalimumab No data 5.7% Sandborn 201293 Golimumab No data 28.1% Colombel 201670 Infliximab No evidence of additional benefit with extended induction Infliximab-dyyb* No evidence of additional benefit with extended induction Infliximab-abda* No evidence of additional benefit with extended induction Tofacitinib 36.3% 13.9% Rubin 201989 Biologic-naïve: 56.6% Biologic-naïve: 18.9% Ustekinumab Danese 201971 Biologic-experienced: 41.4% Biologic-experienced: 1.7% Vedolizumab 39.0% No data NICE FAD56 *Assumed to be equivalent to infliximab. ©Institute for Clinical and Economic Review, 2020 Page 74 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Discontinuation At the end of induction, patients who remain in active UC discontinue treatment with the TIM or conventional treatment and initiate the next line of therapy. In the maintenance phase, patients who revert to active UC will discontinue treatment with the TIM or conventional treatment and initiate the next line of therapy. The model also considers discontinuation for reasons other than loss of efficacy. For the health states of clinical response without remission and clinical remission, a per-cycle probability of discontinuation was applied based on rates observed in RCTs (Table 5.10). Discontinuation is based on rates of discontinuation for reasons other than loss of efficacy and converted to a per-cycle probability. Based on clinical expert opinion, discontinuation rates are lower among patients who remain on TIMs for a year or more. Elimination of certain discontinuation rates after one year for the initial TIM is explored as a scenario analysis. Note that in this scenario, patients still can discontinue the initial TIM due to loss of efficacy. Table 5.10. Discontinuation Per-Cycle Parameter Discontinuation Source Discontinuation Adalimumab 31/254 (12.2%) had discontinued at 52 weeks 1.98% ULTRA 293 Golimumab 23/154 (14.9%) had discontinued at 54 weeks 2.37% PURSUIT94 Infliximab 59/229 (25.7%) had discontinued at 152 weeks 1.56% ACT-1 and ACT-288 Infliximab-dyyb Assume same as infliximab -- -- Infliximab-abda Assume same as infliximab -- -- Tofacitinib 34/394 (8.6%) had discontinued at 52 weeks 1.38% OCTAVE Sustain96 Ustekinumab Manufacturer Vedolizumab 14/122 (11.5%) had discontinued at 52 weeks 1.86% GEMINI 177 Mortality Gender- and age-specific mortality are sourced from the Human Mortality Database's US-specific tables.129 In addition, an elevated risk of mortality is assumed among patients with UC based on a published meta-analysis, which showed a slightly elevated risk of mortality among patients with UC.130 UC has been associated with an increased risk of colorectal cancer and colorectal cancer death in large epidemiologic studies.130,131 The exact pathology of excess colorectal cancer risk is unknown but may be linked to mucosal inflammation.132,133 The impact of TIMs on the risk of colorectal cancer incidence and mortality is unknown and early research does not suggest a clear beneficial effect.134 However, TIMs have a hypothetical potential to reduce incidence of colorectal cancer in UC patients through reduced mucosal inflammation. Colorectal cancer incidence and deaths among UC patients have declined over time, which may be attributed partly to improved treatments for UC, among other factors.135 ©Institute for Clinical and Economic Review, 2020 Page 75 Final Report – Targeted Immune Modulators for UC Return to Table of Contents A mortality multiplier for colorectal cancer-related mortality was applied for all patients with a colon (i.e., without colectomy) to SEER colorectal cancer-related death rates by age to generate the proportion of elevated mortality risk in UC that is attributable to colorectal cancer.136 We removed this elevated risk of mortality due to colorectal cancer in the remission health state to capture a potential long-term beneficial effect of endoscopic improvement. No impact of TIMs on non- colorectal cancer-related mortality will be assumed, as data to date does not provide conclusive evidence of impact.137 Table 5.11. Mortality Inputs Parameter Value Source All-Cause Mortality Gender- and age-specific US Life Tables138 Standardized Mortality Ratio for UC 1.19 (95% CI 1.06-1.35) Bewtra 2013130 Standardized Mortality Ratio for Colorectal 2.82 (95% CI 1.30-1.63) Bewtra 2013130 Cancer-Related Death in UC One-Time Probability of Mortality Among 0.2% Causey 2013139 Patients Undergoing Laparoscopic Colectomy One-Time Probability of Mortality Among 1.7% Causey 2013139 Patients Undergoing Open Colectomy Procedure CI: confidence interval, UC: ulcerative colitis, US: United States Health State Utilities We used consistent health state utility values across treatments evaluated in the model and across induction and maintenance. Health state utility values are taken from a published multi-center cross-sectional study of moderate-to-severe UC patients in Australia with mean EQ-5D utility values for patients in remission, active mild UC, and active moderate-to-severe UC.16(Table 5.12). The impact of alternative sources of utility estimates was evaluated by extending the parameter range of a one-way sensitivity analysis to include lower estimates for active UC and post-colectomy health states from an alternative source of utility values, which have been used in previous cost- effectiveness evaluations in UC.140,141 An alternative scenario was explored whereby health state utility for active moderate-to-severe UC was informed by an average of EQ-5D scores at baseline across the InspireADA, GEMINI 1, and GO-COLITIS trials, 0.6, 0.675, and 0.7, respectively (average of 0.658).74,142,143 Considerable uncertainty exists regarding utility for the post-colectomy health state. In the base case, utility is based on EQ-5D scores from a cross-sectional survey of UC patients in Canada, Australia, and the United Kingdom with a history of colectomy within the prior 10 years.17 Disutility associated with short-term complications of colectomy and chronic pouchitis is also captured (see Table 5.12). Other long-term complications and the long-term impact of colectomy are assumed to be captured within the utility score applied to the post-colectomy health state. We assume in the model that colectomy induces remission of UC, but at a lower utility compared to drug-induced ©Institute for Clinical and Economic Review, 2020 Page 76 Final Report – Targeted Immune Modulators for UC Return to Table of Contents remission because of the health consequences of removing the colon and the potential for long- term complications and adverse events not otherwise captured in the disutility of perioperative complications and chronic pouchitis (e.g., fertility, fatigue). The utility value used in the base case (0.79) is similar to the average EQ-5D utility index score reported for another overall post- colectomy UC cohort in the United Kingdom (0.74). This same study also reported a lower utility score for the subset of patients who experienced long-term complications (0.71), which was explored as the lower bound of the point estimate in one-way sensitivity analysis.144 Table 5.12. Utility Values for Health States Parameter Value (95% CI) Source Range for OWSA Source Active UC 0.68 (0.63, 0.73) 16 0.41 to 0.73 140 Clinical Response without 0.78 (0.71, 0.85) 16 0.71 to 0.85 95% CI (calculated) Remission* Clinical Remission 0.81 (0.77, 0.85) 16 0.77 to 0.85 95% CI (calculated) Post-Colectomy 0.79 (0.77, 0.81) 17 0.71 to 0.81 140,144,17 CI: confidence interval, OWSA: one-way sensitivity analysis, UC: ulcerative colitis *Response without remission utility value represented by mild UC cohort in Malinowski et al. 2016. Adverse Events Disutility associated with serious infections with TIMs and conventional treatment (e.g., respiratory infections, gastrointestinal infections, sepsis), early complications of colectomy, and late complications (chronic pouchitis) are included in the model (Table 5.13). The rate of serious infection is informed by the rates for the maintenance phase of clinical trials and adjusted to an eight-week cycle length. A reduction in utility of 52% is applied to the patients' health state and is assumed for the duration of the cycle with a serious infection event. A small but statistically significant increased risk of malignancy with the use of TIMs has been documented in the literature, with potential for long-term impacts on quality of life and cost.145 However, long-term studies have not consistently demonstrated an elevated risk, and the comparative risk across agents is unknown.107,112,146 Due to data limitations, risk of malignancy is not included in the model. Multiple different surgical techniques are used to perform colectomy procedures, with differing safety outcomes and rates of complications. While subtle differences may exist among various procedures, the most notable division exists between open procedures and laparoscopic (closed) procedures. Table 5.13 presents the incidence of early complications in open and laparoscopic colectomy procedures as well as the probability of the long-term complication of chronic pouchitis that is applicable to both procedures. ©Institute for Clinical and Economic Review, 2020 Page 77 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.13. Adverse Events Parameter Value Source Utility Source Serious Infection (Per Year) Adalimumab 1.6% ULTRA 293 Golimumab 3.2% PURSUIT147 Infliximab, Infliximab-dyyb, and 2.1% ACT 1 and 290 Infliximab-abda 5 mg/kg -52% disutility NICE STA34256 Tofacitinib 5 mg 1.0% OCTAVE SUSTAIN96 Ustekinumab 90 mg q8w 1.7% UNIFI99 Vedolizumab 1.9% GEMINI 177 Conventional Treatment 2% * Colectomy Complications Early Complications of 25.3% Arseneau Zogg 2016148 0.49 Colectomy (Open Procedure) incidence 2006149 Early Complications of 17.3% Arseneau Zogg 2016148 0.49 Colectomy (Laparoscopic) incidence 2006149 15.5% Arseneau Chronic Pouchitis Zogg 2016148 0.40 prevalence 2006149 kg: kilogram, mg: milligram, q8w: every 8 weeks *Based on pooled placebo arms of treat through trials. Economic Inputs Drug Utilization Table 5.14 on the following page outlines inputs for dose and frequency of administration of TIMs in UC. Conventional treatment is modeled as an induction period of prednisone 40 mg orally once daily, followed by mercaptopurine 1-1.5 mg/kg per day or azathioprine 2-3 mg/kg per day (assuming a 50:50 split between mercaptopurine and azathioprine). ©Institute for Clinical and Economic Review, 2020 Page 78 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.14. Treatment Regimen Recommended Dosage Generic ADA GOL IFX IFX-dyyb IFX-abda TOF UST VEDO Name Brand Name Humira Simponi Remicade Inflectra Renflexis Xeljanz Stelara Entyvio Manufacturer AbbVie Janssen Janssen Pfizer Merck Pfizer Janssen Takeda Route of SC SC IV IV IV Oral SC IV Admin. 160 mg on 200 mg at 10 mg One 390 5 mg/kg 5 mg/kg day 1, then week 0, 5 mg/kg twice daily mg IV 300 mg at at 0, 2, at 0, 2, 80 mg 100 mg at at 0, 2, for 8 dose, 0, 2, Dosing 6 weeks, 6 weeks, 2 weeks week 2, 6 weeks, weeks, followed 6 weeks, then then later, then then 100 then q8w then 5 mg by 90 mg then q8w q8w q8w 40 mg EOW mg q4w twice daily SC q8w ADA: adalimumab, admin.: administration, EOW: every other week, GOL: golimumab, IFX: infliximab, IV: intravenous, kg: kilogram, mg: milligram, q4w: every 4 weeks, q8w: every 8 weeks, SC: subcutaneous, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Dose Escalation Dose escalation is considered in a scenario as an additional cost through a higher dose and/or frequency of administration for a proportion of patients in the maintenance phase. The frequency of dose escalation for each TIM was taken from a recent claims analysis of greater than expected dosing of TIMs in patients with IBD.150 As no data were available for tofacitinib, the prevalence of dose escalation was assumed as the average of all other TIMs. Table 5.15. Assumptions for Dose Escalation Escalated Maintenance Parameter Proportion of Patients Source Regimen in the Model Adalimumab 40 mg every week 27.5% MacDougall 2019150 Golimumab 200 mg q4w 14.3% MacDougall 2019150 Infliximab, Infliximab-dyyb, 5 mg/kg q4w 39.4% MacDougall 2019150 Infliximab-dyyb Tofacitinib 10 mg twice daily 25% Assumption Ustekinumab 90 mg q4w 21.6% MacDougall 2019150 Vedolizumab 300 mg q4w 22.7% MacDougall 2019150 kg: kilogram, mg: milligram, q4w: every 4 weeks Drug Acquisition Costs For TIMs with oral or subcutaneous modes of administration we obtained net pricing estimates from SSR Health, LLC, which combines data on unit sales with publicly-disclosed US sales figures that are net of discounts, rebates, concessions to wholesalers and distributors, and patient assistance programs, to derive a net price.18 We estimated net prices by comparing the four- quarter averages of both net prices and wholesale acquisition cost (WAC) per unit to arrive at a ©Institute for Clinical and Economic Review, 2020 Page 79 Final Report – Targeted Immune Modulators for UC Return to Table of Contents mean discount from WAC for the drug. Finally, we applied this average discount to the most recent available WAC (May 6, 2020) to arrive at an estimated net price per dose. For IV-administered TIMs, we used Centers for Medicare and Medicaid Services Average Sales Prices (ASP) plus 9.5%.19 Infliximab and infliximab biosimilar doses, which are weight based, were rounded up to the nearest vial. Table 5.16. Drug Unit Costs Source of Discount from Drug Price per Unit Units Cost Data* WAC18 Adalimumab (40 mg) $5,556.97 WAC Two 40 mg doses 35.2% Golimumab (100 mg) $5,779.26 WAC One 100 mg dose 43.8% Infliximab (5 mg/kg) $51.20 ASP 10 mg N/A Infliximab-dyyb (5 mg/kg) $47.13 ASP 10 mg N/A Infliximab-abda (5 mg/kg) $48.64 ASP 10 mg N/A Tofacitinib (10 mg) $4,700.18 WAC Sixty 10 mg tabs 37.9% Ustekinumab IV (130 mg) $12.17 ASP 1 mg N/A Ustekinumab SC (90 mg) $23,082.84 WAC One 90 mg dose 39.1% Vedolizumab (300 mg) $ 20.66 ASP 1 mg 18.3% One hundred 20 Prednisone (20 mg) $20.72 WAC -- mg tabs Mercaptopurine $79.25 WAC 25 tabs N/A Azathioprine $45.00 WAC 100 tabs N/A ASP: average sale price, IV: intravenous, kg: kilogram, mg: milligram, N/A: not applicable, SC: subcutaneous, WAC: wholesale acquisition cost *WAC as of May 6, 2020151; ASP pricing effective July 1, 2020 through September 30, 2020.19 Please refer to the ICER Reference Case for more details on drug pricing. ©Institute for Clinical and Economic Review, 2020 Page 80 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.17. Cost of Induction and Maintenance Net Price Net Price Units Per Eight Units per Drug per Eight Week per Maintenance Week Induction Maintenance Year Induction Year Adalimumab (40 mg) 8 $14,403.67 26.1 $46,933 Golimumab (100 mg) 4 $12,991.78 13.0 $42,332 3 weight-based 6.5 weight-based Infliximab (5 mg/kg) $6,727.68 $14,614 doses doses 3 weight-based 6.5 weight-based Infliximab-dyyb (5 mg/kg) $6,192.36 $13,451 doses doses 3 weight-based 6.5 weight-based Infliximab-abda (5 mg/kg) $6,390.90 $13,883 doses doses Tofacitinib (5 or 10 mg) 112 $5,448.45 729.9 $35,506 Ustekinumab IV (130 mg) 3 $5,197.20 N/A N/A Ustekinumab SC (90 mg) N/A N/A 6.5 $91,609 Vedolizumab (300 mg) 3 $20,358.46 6.5 $44,224 Prednisone (20 mg) 3 $23.21 N/A N/A Mercaptopurine N/A N/A 729.9 $2,314 Azathioprine N/A N/A 1094.8 $493 IV: intravenous, kg: kilogram, mg: milligram, N/A: not applicable, SC: subcutaneous Direct Cost of Administration and Monitoring Administration costs are included for IV formulations at a cost of $72.80 per infusion based on the Centers for Medicare and Medicaid Use Physician Fee Schedule average non-facility price for CPT 96365 (IV infusion, for therapy, prophylaxis, or diagnosis [specify substance or drug]; initial, up to one hour).152 The average non-facility price for CPT 96366 ($21.98) was added for infliximab, infliximab-dyyb, and infliximab-abda to account for a second hour of infusion time per product labeling.153 Direct Cost of Colectomy Procedure A cost of $33,871 per admission for colectomy is assumed, calculated by the weighted average cost of emergent and non-emergent cost per admission from the Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) database, after inflation to 2019 US dollars using the personal health care (PHC) expenditure deflator up to 2017 then the personal consumption expenditure (PCE) price index to update from 2017 to 2019.154-156 Other Costs of UC Management Non-drug costs of UC management include the average cost of hospitalization, emergency department visits, and outpatient visits by health state (Table 5.18). Direct health care utilization costs were calculated based on previously published estimates of mean unadjusted annual costs ©Institute for Clinical and Economic Review, 2020 Page 81 Final Report – Targeted Immune Modulators for UC Return to Table of Contents among privately-insured employed people with UC and a matched cohort without UC.157 The active UC health state is informed by the moderate-to-severe cohort of the study, defined by investigators as having a hospitalization with a primary diagnosis of UC or treatment with biologics, immunosuppressants, or systemic corticosteroids. The clinical response without remission health state is informed by the overall UC study population, which includes all patients with at least two diagnoses of UC. Finally, the clinical remission health state is informed by the non-UC control cohort. Table 5.18. Annual Direct Health Care Costs by Health State* Emergency Hospitalization Outpatient/Other Parameter Department Source Value (95% CI) Value (95% CI) Value (95% CI) Active UC $8,048 ($33,476) $581 ($1,722) $10,114 ($16,787) Cohen 2015157 Clinical Response without $4,461 ($23,436) $410 ($1,436) $7,506 ($14,561) Cohen 2015157 Remission† Clinical Remission‡ $969 ($8,137) $223 ($992) $2,722 ($8,266) Cohen 2015157 CI: confidence interval, UC: ulcerative colitis Health states apply to both induction and maintenance phases. *Inflated 2019 US dollars using the PHC expenditure deflator up to 2017 then the PCE price index to update from 2017 to 2019. †Informed by overall UC population. ‡Informed by non-UC control group. Adverse Event Costs Costs associated with serious infections with TIMs, early complications of colectomy, and late complications (chronic pouchitis) are included (Table 5.19). Serious infection events are assigned a cost of $10,238 based on HCUP net mean hospital costs for an inpatient stay with ICD-10 diagnosis of pneumonia (J12-J18), as pneumonia is a commonly reported serious infection, an approach that has been taken in another recent model of UC.158 Table 5.19 also presents the cost per short-term complication in open and laparoscopic colectomy procedures as well as chronic pouchitis. Table 5.19. Cost of Adverse Events Parameter Cost Source Serious Infection $10,238 AHRQ159 Early Complications of Colectomy (Open Procedure) $11,435* Zogg 2016148 Early Complications of Colectomy (Laparoscopic) $8,293* Zogg 2016148 Chronic Pouchitis $1,581 per month* Park 2012160 AHRQ: Agency for Healthcare Research and Quality *Inflated 2019 US dollars using the PHC expenditure deflator up to 2017 then the PCE price index to update from 2017 to 2019. ©Institute for Clinical and Economic Review, 2020 Page 82 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Productivity Costs A modified societal perspective including indirect costs of disability, medical-related absenteeism, and presenteeism by time spent in clinical response without remission, remission, or active UC health states is included as a scenario analysis. Days of disability and medical-related absenteeism are sourced from the claims analysis as described for direct health state costs, assuming eight hours per day lost due to disability or absenteeism and average hourly wage in the US ($28.18).157,161 For presenteeism, WPAI presenteeism estimates from a US patient survey are combined with US Bureau of Labor Statistics average working hours per week (38.6) and average hourly wage ($28.18).161-163 Table 5.20. Annual Indirect Health Care Costs by Health State Medical- Total Indirect Parameter Disability Related Presenteeism Source Costs per Year Absenteeism Active UC $2,773 $2,593 $13,694 $19,059 Cohen 2015,157 Ding Clinical Response $1,781† $1,916† $7,843† $11,540* 2019,163 US Bureau of without Remission Labor Statistics145-147 Clinical Remission $1,014‡ $1,195‡ $3,056 $5,266 UC: ulcerative colitis, US: United States Health states apply to both induction and maintenance phases. *Response without remission utility value represented by mild UC cohort. †Informed by overall UC population. ‡Informed by non-UC control group. The modified societal perspective also includes lost productivity related to UC treatments, including patient time spent to and from a medical visit to receive an IV infusion (vs. self-administration of subcutaneous or oral TIMs) (121 minutes per infusion)164 and absenteeism due to colectomy procedure (conservative approach of 100 days of absenteeism related to open procedures).165 For both inputs, a human capital approach was taken whereby the average time spent was multiplied by average hourly wage ($28.18), assuming eight hours of absenteeism per day, to estimate indirect cost. Sensitivity Analyses We ran one-way sensitivity analyses to identify the key drivers of model outcomes, using available measures of parameter uncertainty (i.e., standard errors) or reasonable ranges for each input described in the model inputs section above. Probabilistic sensitivity analyses were also performed by jointly varying all model parameters over 1,000 simulations, then calculating 95% credible range estimates for each model outcome based on the results. We used normal distributions for costs, rates, multipliers, and ages; gamma distributions for disutilities; and beta distributions for probabilities and utilities. For the TIM risk ratios, we sampled from the NMA codas directly. ©Institute for Clinical and Economic Review, 2020 Page 83 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Additionally, we performed a threshold analysis by systematically altering the price of each TIM to estimate the maximum prices that would correspond to given willingness-to-pay (WTP) thresholds in the biologic-naïve and biologic-experienced populations. Scenario Analyses We conducted the following scenario analyses: • Time horizons of two, five, and 10 years • Dose escalation • Extended induction/delayed response • Modified societal perspective that includes productivity losses • Elimination of reduced risk of colorectal cancer death for patients in remission state • Alternative sources of health state utility estimates • Lower rates of discontinuation after one year, with discontinuation due to loss of response only • Use of infliximab following initial vedolizumab therapy, assuming comparable efficacy to that in a biologic-naïve population. Model Validation We used several approaches to validate the model and followed standard practices in the field. First, we provided preliminary methods and results to manufacturers, patient groups, and clinical experts. Based on feedback from these groups, we refined data inputs used in the model. We also shared the draft model with participating manufacturers during the public commenting period. We tested all mathematical functions in the model to ensure that they were consistent with the report. We also conducted sensitivity analyses with extreme input values to ensure the model was producing findings consistent with expectations. Finally, we compared results to other cost- effectiveness models in this therapy area. 5.3 Results Base-Case Results Biologic-Naïve Population Total discounted initial TIM drug costs, total costs, LYs, and QALYs for the biologic-naïve population over the lifetime time horizon are shown in Table 5.21. Undiscounted results are presented in Appendix Tables E2-E4. Discounted drug costs for initial TIMs ranged from $22,000 to $138,000 and total costs ranged from $434,000 to $545,000 over the lifetime time horizon compared to a total cost of $421,000 for conventional treatment. Discounted life expectancy from age of initiation (age ©Institute for Clinical and Economic Review, 2020 Page 84 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 40 years) was 22.08 LY across all treatments, although small differences exist across TIMs that are not presented in Table 5.21 due to rounding. Projected discounted QALYs for TIMs ranged from 15.60 to 15.68 for TIMs compared with 15.57 QALYs for conventional treatment. As the TIMs resulted in minimal gains in LY, QALYs closely approximated evLY. The incremental cost per QALY for TIMs compared to conventional treatment in the biologic-naïve population ranged from $186,000 (infliximab-dyyb) to $1,870,000 (adalimumab) (Table 5.21). The incremental cost per QALY for TIMs compared to infliximab, a market share leader in the biologic- naïve population, is presented in Table 5.22. All TIMs had lower or equal LYs and QALYs compared to infliximab at a higher cost with the exception of ustekinumab, which had slightly higher QALYs at higher cost with a resulting cost per QALY exceeding $2.9 million. Full pairwise comparisons are presented in Appendix Table E8. Infliximab and its biosimilars had lower costs and greater QALYs than adalimumab, golimumab and vedolizumab. Ustekinumab resulted in the greatest number of QALYs but was not cost effective compared with infliximab-dyyb. Table 5.21. Results for the Base Case for TIMs and Conventional Treatment: Biologic-Naïve Parameter Initial Drug Cost Total Cost LYs* QALYs evLY Adalimumab $44,000 $461,000 22.077 15.596 15.598 Golimumab $41,000 $458,000 22.078 15.600 15.602 Infliximab $24,000 $435,000 22.080 15.644 15.646 Infliximab-dyyb $22,000 $434,000 22.080 15.644 15.646 Infliximab-abda $23,000 $434,000 22.080 15.644 15.646 Ustekinumab $138,000 $545,000 22.081 15.681 15.683 Vedolizumab $68,000 $480,000 22.080 15.641 15.644 Conventional Treatment $600 $421,000 22.075 15.574 15.576 evLY: equal value of life year, LY: life year, N/A: not applicable, QALY: quality-adjusted life year, TIM: targeted immune modulator Costs rounded to nearest $1,000. *Small differences in LYs across comparators are not displayed due to rounding. ©Institute for Clinical and Economic Review, 2020 Page 85 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.22. Incremental Cost-Effectiveness Ratios for the Base Case Compared to Conventional Treatment: Biologic-Naïve Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab $21,624,000 $1,870,000 $1,847,000 Golimumab $12,981,000 $1,455,000 $1,432,000 Infliximab $3,021,000 $212,000 $209,000 Infliximab-dyyb $2,651,000 $186,000 $184,000 Infliximab-abda $2,788,000 $195,000 $193,000 Ustekinumab $23,615,000 $1,163,000 $1,155,000 Vedolizumab $13,767,000 $887,000 $880,000 Conventional Treatment Reference Reference Reference evLY: equal value of life year, LY: life year, QALY: quality-adjusted life year Cost per LY gained rounded to the nearest $10,000; cost per QALY gained rounded to nearest $1,000. Table 5.23. Incremental Cost-Effectiveness Ratios for the Base Case Compared to Infliximab: Biologic-Naïve Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab Higher cost, fewer LYs Higher cost, fewer QALYs Higher cost, fewer evLYs Golimumab Higher cost, fewer LYs Higher cost, fewer QALYs Higher cost, fewer evLYs Infliximab Reference Reference Reference Infliximab-dyyb Lower cost, equal LYs Lower cost, equal QALYs Lower cost, equal evLYs Infliximab-abda Lower cost, equal LYs Lower cost, equal QALYs Lower cost, equal evLYs Ustekinumab $297,831,000 $2,960,000 $2,956,000 Vedolizumab Higher cost, fewer LYs Higher cost, fewer QALYs Higher cost, fewer evLYs evLY: equal value of life year, LY: life year, QALY: quality-adjusted life year Cost per LY gained rounded to the nearest $10,000; cost per QALY gained rounded to nearest $1,000. ©Institute for Clinical and Economic Review, 2020 Page 86 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure 5.2. Cost-Effectiveness Frontier for TIMs in the Base Case: Biologic-Naïve IV: intravenous, QALY: quality-adjusted life year Drugs that are farther to the right in Figure 5.2 above provide greater clinical benefit and drugs higher on the y-axis are more expensive. The line on the graph depicts the cost-effectiveness efficiency frontier. Those therapies that lie to the left of the frontier are dominated by therapies that lie on the frontier. Thus, therapies to the left of the frontier, using only the deterministic findings, are considered to not be as cost effective as those therapies on the frontier. Biologic-Experienced Population Total discounted initial TIM drug costs, total costs, LYs, and QALYs for the biologic-experienced population over the lifetime time horizon are shown in Table 5.24. Undiscounted results are presented in Appendix Tables E5-E7. Discounted drug costs for initial TIMs ranged from $32,000 to $76,000 and total costs ranged from $460,000 to $504,000 over the lifetime time horizon compared to a total cost of $434,000 for conventional treatment. Discounted life expectancy from age of initiation (age 40 years) was 22.07 LYs across all treatments, although small differences exist across TIMs that are not presented due to rounding. Projected discounted QALYs for TIMs ranged from 15.41 to 15.45 for TIMs compared with 15.39 QALYs for conventional treatment. As the TIMs resulted in minimal gains in LYs, QALYs closely approximated evLY. The incremental cost per QALY for TIMs compared to conventional treatment in the biologic- experienced population ranged from $495,000 (tofacitinib) to $1,885,000 (adalimumab). The incremental cost per QALY for TIMs compared to each other is located in Appendix E9 and Table 5.25 for TIMs compared to adalimumab, a market share leader in the biologic-experienced population. Treatment with tofacitinib resulted in greater QALYs at a lower cost compared with adalimumab. Ustekinumab had an incremental cost-effectiveness ratio of $996,000 per QALY ©Institute for Clinical and Economic Review, 2020 Page 87 Final Report – Targeted Immune Modulators for UC Return to Table of Contents gained compared to adalimumab. The additional cost per QALY was $464,000 for vedolizumab compared with adalimumab. Although QALYs were very similar between ustekinumab and vedolizumab, ustekinumab had an incremental cost-effectiveness ratio of $6,422,000 per QALY gained compared to vedolizumab. Please note that due to the incremental cost-effectiveness calculation being a ratio, it increases rapidly with small differences in QALYs as is the case here. Table 5.24. Results for the Base Case for TIMs and Conventional Treatment: Biologic-Experienced Initial Parameter Total Cost LYs* QALYs evLY Drug Cost Adalimumab $33,000 $465,000 22.070 15.410 15.412 Tofacitinib $32,000 $460,000 22.074 15.445 15.448 Ustekinumab $76,000 $504,000 22.074 15.449 15.452 Vedolizumab $54,000 $482,000 22.073 15.446 15.448 Conventional Treatment $323 $434,000 22.070 15.393 15.395 evLY: equal value of life year, LY: life year, N/A: not applicable, QALY: quality-adjusted life year, TIM: targeted immune modulator Costs rounded to nearest $1,000. *Small differences in LYs across comparators are not displayed due to rounding. Table 5.25. Incremental Cost-Effectiveness Ratios for the Base Case Compared to Conventional Treatment: Biologic-Experienced Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab $79,822,000 $1,885,000 $1,878,000 Tofacitinib $6,281,000 $495,000 $489,000 Ustekinumab $16,640,000 $1,252,000 $1,239,000 Vedolizumab $17,707,000 $902,000 $895,000 Conventional Treatment Reference Reference Reference evLY: equal value of life year, LY: life year, QALY: quality-adjusted life year Cost per LY gained rounded to the nearest $10,000; cost per QALY gained rounded to nearest $1,000. Table 5.26. Incremental Cost-Effectiveness Ratios for the Base Case Compared to Adalimumab: Biologic-Experienced Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab Reference Reference Reference Tofacitinib Lower cost, greater LYs Lower cost, greater QALYs Lower cost, greater evLYs Ustekinumab $10,358,000 $996,000 $982,000 Vedolizumab $7,362,000 $464,000 $460,000 evLY: equal value of life year, LY: life year, QALY: quality-adjusted life year Cost per LY gained rounded to the nearest $10,000; cost per QALY gained rounded to nearest $1,000. ©Institute for Clinical and Economic Review, 2020 Page 88 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure 5.3. Cost-Effectiveness Frontier for TIMs in the Base Case: Biologic-Experienced IV: intravenous, QALY: quality-adjusted life year Drugs that are farther to the right in Figure 5.3 provide greater clinical benefit and drugs higher on the y-axis are more expensive. The line on the graph depicts the cost-effectiveness efficiency frontier. Those therapies that lie to the left of the frontier are dominated by therapies that lie on the frontier. Thus, therapies to the left of the frontier, using only the deterministic findings, are considered to not be as cost effective as those therapies on the frontier. Sensitivity Analysis Results To demonstrate effects of uncertainty on both costs and health outcomes, we varied input parameters using available measures of parameter uncertainty (i.e., 95% CI) or reasonable ranges (±10%) to evaluate changes in cost per additional QALY for each TIM compared to conventional treatment. The tornado charts for vedolizumab in a biologic-naïve population and biologic- experienced population are presented in Figure 5.4 as an example. The results for other comparators were similar and are presented in the Appendix. For all comparisons, the health state utility of active UC was the most important driver of model results, with the cost per QALY for infliximab falling below the $150,000 per QALY threshold when the low estimate for utility was applied. Other key drivers across most comparisons were utility of the response and remission health states, efficacy of TIMs, and cost of TIMs. ©Institute for Clinical and Economic Review, 2020 Page 89 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure 5.4. Tornado Diagram for One-Way Sensitivity Analyses of Vedolizumab versus Conventional Treatment (Cost per QALY) Biologic-Naive Biologic-experienced bio: biologic, exp: experienced, IV: intravenous, resp: response without remission, QALY: quality-adjusted life year, RR: risk ratio, UC: ulcerative colitis, vedo: vedolizumab The proportion of probabilistic sensitivity analysis iterations with an incremental cost-effectiveness ratio below thresholds ranging from $50,000 to $250,000 per QALY gained are presented in Table 5.27 for the biologic-naïve population and Table 5.28 for the biologic-experienced population. Full results of probabilistic sensitivity analysis are presented in Appendix Tables E10 and E11. The mean probabilistic estimate of cost per QALY was above $150,000 for all TIMs in both populations. For the biologic-naïve population, infliximab-dyyb had the highest probability of being cost effective at $150,000 per QALY at 29%. The lower bound of the 95% credible interval for adalimumab, golimumab, ustekinumab, and vedolizumab did not fall below $150,000 in the biologic-naïve population. For the biologic-experienced population, no TIMs were likely (>50%) to be cost effective at any of the thresholds and the lower bound of the 95% credible interval for the incremental cost-effectiveness ratios did not fall below $150,000 for any TIMs. ©Institute for Clinical and Economic Review, 2020 Page 90 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.27. Probabilistic Sensitivity Analysis Results: TIMs versus Conventional Treatment: Biologic-Naïve Cost Effective Cost Effective Cost Effective Cost Effective Cost Effective at $50,000 at $100,000 at $150,000 at $200,000 at $250,000 per QALY per QALY per QALY per QALY per QALY Adalimumab 0% 0% 0% 0% 0% Golimumab 0% 0% 0% 0% 0% Infliximab 0% 3% 17% 44% 69% Infliximab-dyyb 1% 6% 29% 59% 81% Infliximab-abda 1% 5% 24% 54% 77% Ustekinumab 0% 0% 0% 0% 0% Vedolizumab 0% 0% 0% 0% 0% QALY: quality-adjusted life year Table 5.28. Probabilistic Sensitivity Analysis Results: TIMs versus Conventional Treatment: Biologic-Experienced Cost Effective Cost Effective Cost Effective Cost Effective Cost Effective at $50,000 at $100,000 at $150,000 at $200,000 at $250,000 per QALY per QALY per QALY per QALY per QALY Adalimumab 0% 0% 0% 0% 0% Tofacitinib 0% 0% 0% 0% 1% Ustekinumab 0% 0% 0% 0% 0% Vedolizumab 0% 0% 0% 0% 0% QALY: quality-adjusted life year Scenario Analyses Results Shorter Time Horizons The model assumes that patients try a second TIM after lack of response, loss of response, or discontinuation of the first TIM. In clinical practice, patients may try many interventions for UC over a lifetime time horizon. Our model is a simplification of a complex and highly individualized treatment pathway in order to focus on the cost effectiveness of the TIM of interest. In order to increase focus on the TIM of interest, analyses with time horizons of two, five, and 10 years were conducted. Shorter time horizons had a variable impact on cost per QALY estimates, but not alter base-case conclusions. ©Institute for Clinical and Economic Review, 2020 Page 91 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.29. Incremental Cost-Effectiveness Ratios for the Shorter Time Horizon Scenario: Biologic- Naïve Base-Case 10 Year Time Five Year Time Two Year Time Treatment Cost per QALY Horizon Cost per Horizon Cost per Horizon Cost per Gained QALY Gained QALY Gained QALY Gained Adalimumab $1,870,000 $1,743,564 $1,870,602 $2,781,164 Golimumab $1,455,000 $1,390,522 $1,443,954 $1,921,569 Infliximab $212,000 $187,276 $176,265 $198,221 Infliximab-dyyb $186,000 $161,834 $147,553 $154,055 Infliximab-abda $195,000 $171,270 $158,202 $170,436 Ustekinumab $1,163,000 $1,125,847 $1,269,998 $1,784,835 Vedolizumab $887,000 $848,232 $934,151 $1,418,486 Conventional Reference Reference Reference Reference Treatment QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.30. Incremental Cost-Effectiveness Ratios for the Shorter Time Horizon Scenario: Biologic- Experienced 10 Year Time Five Year Time Two Year Time Cost per QALY Treatment Horizon Cost per Horizon Cost per Horizon Cost per Gained QALY Gained QALY Gained QALY Gained Adalimumab $1,885,000 $1,623,246 $1,798,937 $3,516,694 Tofacitinib $495,000 $453,742 $454,364 $553,737 Ustekinumab $1,252,000 $1,180,697 $1,235,494 $1,629,824 Vedolizumab $902,000 $831,989 $910,179 $1,509,651 Conventional Reference Reference Reference Reference Treatment QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Dose Escalation A higher dose and/or frequency of administration may be required for a proportion of patients in the maintenance phase to maintain response. Higher cost of TIMs led to an increase in cost per QALY across all TIMs compared to conventional treatment. ©Institute for Clinical and Economic Review, 2020 Page 92 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.31. Incremental Cost-Effectiveness Ratios for the Dose Escalation Scenario: Biologic-Naïve Cost per QALY Gained Treatment Base-Case Cost per QALY Gained Considering Dose Escalation Adalimumab $1,870,000 $2,248,519 Golimumab $1,455,000 $1,605,081 Infliximab $212,000 $302,295 Infliximab-dyyb $186,000 $269,279 Infliximab-abda $195,000 $281,524 Ustekinumab $1,163,000 $1,433,705 Vedolizumab $887,000 $1,048,395 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.32. Incremental Cost-Effectiveness Ratios for the Dose Escalation Scenario: Biologic- Experienced Cost per QALY Gained Treatment Base-Case Cost per QALY Gained Considering Dose Escalation Adalimumab $1,885,000 $2,202,210 Tofacitinib $495,000 $492,377 Ustekinumab $1,252,000 $1,529,360 Vedolizumab $902,000 $1,043,476 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Extended Induction An additional efficacy benefit from extended induction was calculated for adalimumab, golimumab, tofacitinib, ustekinumab, and vedolizumab. Allowing for an additional eight weeks of induction did not substantially alter conclusions based on incremental cost-effectiveness ratios and all remained above the $150,000 per QALY threshold. It is important to note that our cost-effectiveness analysis was not constructed with the intent of evaluating the most appropriate duration of induction for TIMs in UC. The results of this scenario analysis are limited by assumptions surrounding the efficacy of the second TIM and the market basket and are not intended to inform quantity limits or treatment decisions. ©Institute for Clinical and Economic Review, 2020 Page 93 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.33. Incremental Cost-Effectiveness Ratios for the Extended Induction Scenario: Biologic- Naïve Cost per QALY Gained Treatment Base-Case Cost per QALY Gained Considering Extended Induction Adalimumab $1,870,000 Higher cost, fewer QALYs Golimumab $1,455,000 Higher cost, fewer QALYs Ustekinumab $1,163,000 $1,519,412 Vedolizumab $887,000 $1,714,386 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.34. Incremental Cost-Effectiveness Ratios for the Extended Induction Scenario: Biologic- Experienced Cost per QALY Gained Treatment Base-Case Cost per QALY Gained Considering Extended Induction Adalimumab $1,885,000 Higher cost, fewer QALYs Tofacitinib $495,000 $1,244,000 Ustekinumab $1,252,000 $3,926,000 Vedolizumab $902,000 $1,687,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Modified Societal Perspective A modified societal perspective scenario was undertaken, which considered absenteeism, presenteeism, and disability due to UC. In this scenario, the cost per QALY was reduced substantially for all TIMs, with infliximab and infliximab biosimilars falling to or below the $100,000 per QALY threshold for the biologic-naïve population. ©Institute for Clinical and Economic Review, 2020 Page 94 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.35. Incremental Cost-Effectiveness Ratios for the Modified Societal Perspective Scenario: Biologic-Naïve Including Indirect Costs per Treatment Base-Case Cost per QALY Gained QALY Gained Adalimumab $1,870,000 $1,745,283 Golimumab $1,455,000 $1,334,762 Infliximab $212,000 $100,932 Infliximab-dyyb $186,000 $75,023 Infliximab-abda $195,000 $84,633 Ustekinumab $1,163,000 $1,041,891 Vedolizumab $887,000 $775,740 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.36. Incremental Cost-Effectiveness Ratios for the Modified Societal Perspective Scenario: Biologic-Experienced Including Indirect Costs per Treatment Base-Case Cost per QALY Gained QALY Gained Adalimumab $1,885,000 $1,753,157 Tofacitinib $495,000 $384,312 Ustekinumab $1,252,000 $1,142,920 Vedolizumab $902,000 $793,415 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Exclusion of Reduced Risk of Colorectal Cancer Death for Patients in Remission A risk of mortality in UC, which excludes colorectal cancer-related mortality, was applied to patients in the remission health state to capture a potential long-term benefit effect of endoscopic improvement. Excluding this impact had a negligible impact on the incremental cost-effectiveness ratios due to the relatively low incidence of colorectal cancer. Alternative Health State Utility Estimates Health state utility estimates for active UC, response without remission, and response with remission are a major driver of model results. As a scenario analysis, we applied an alternative estimate for active UC based on an average of baseline EQ-5D scores across the InspireADA, GEMINI 1, and GO-COLITIS trials (0.658).74,142,143 The lower estimate of utility for active UC led to an increase in QALYs gained with TIMs compared to conventional treatment, corresponding to lower incremental cost-effectiveness ratios for all TIMs. No TIMs fell below the $150,000 per QALY threshold. ©Institute for Clinical and Economic Review, 2020 Page 95 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.37. Incremental Cost-Effectiveness Ratios for the Alternative Utility for Active UC Scenario: Biologic-Naïve Using Alternative Utility Cost per Treatment Base-Case Cost per QALY Gained QALY Gained Adalimumab $1,870,000 $1,612,000 Golimumab $1,455,000 $1,259,000 Infliximab $212,000 $185,000 Infliximab-dyyb $186,000 $162,000 Infliximab-abda $195,000 $170,000 Ustekinumab $1,163,000 $1,017,000 Vedolizumab $887,000 $775,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.38. Incremental Cost-Effectiveness Ratios for the Alternative Utility for Active UC Scenario: Biologic-Experienced Alternative Utility Cost per Treatment Base-Case Cost per QALY Gained QALY Gained Adalimumab $1,885,000 $1,600,000 Tofacitinib $495,000 $428,000 Ustekinumab $1,252,000 $1,082,000 Vedolizumab $902,000 $780,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Lower Discontinuation Rates after One Year The scenario of assuming lower discontinuation rates after one year for the initial TIM had reduced the cost per QALY for all TIMs. No TIMs fell below the $150,000 per QALY threshold. ©Institute for Clinical and Economic Review, 2020 Page 96 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.39. Incremental Cost-Effectiveness Ratios for Lower Discontinuation Rates Scenario: Biologic-Naïve Excluding Discontinuation for Reasons Treatment Base-Case Cost per QALY Gained Other Than Lack of Efficacy After One Year Adalimumab $1,870,000 $1,535,000 Golimumab $1,455,000 $1,224,000 Infliximab $212,000 $183,000 Infliximab-dyyb $186,000 $160,000 Infliximab-abda $195,000 $168,000 Ustekinumab $1,163,000 $1,112,000 Vedolizumab $887,000 $771,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.40. Incremental Cost-Effectiveness Ratios for Lower Discontinuation Rates Scenario: Biologic-Experienced Excluding Discontinuation for Reasons Treatment Base-Case Cost per QALY Gained Other Than Lack of Efficacy After One Year Adalimumab $1,885,000 $1,526,000 Tofacitinib $495,000 $472,000 Ustekinumab $1,252,000 $1,227,000 Vedolizumab $902,000 $788,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Use of Infliximab Following Initial Vedolizumab Therapy or Conventional Treatment No RCT data was identified for infliximab or infliximab biosimilars in a biologic-experienced population, so these therapies were excluded from the market basket of second TIMs, which by definition would be a biologic-experienced population. However, clinicians consulted during the course of this review indicated that using infliximab after failure by vedolizumab was a common treatment sequence used in real-world clinical practice. Therefore, we conducted a scenario whereby all patients who are initiated on vedolizumab or conventional treatment and switch to subsequent treatment are switched to infliximab as the subsequent treatment. As no data are available in a biologic-experienced population for infliximab, we assumed comparable efficacy to that in a biologic-naïve population. Although absolute QALYs are higher for both vedolizumab and conventional treatment in this scenario, the incremental cost per QALY gained for vedolizumab compared with conventional treatment is equal to the base case in both the biologic-naïve and biologic-experienced populations. ©Institute for Clinical and Economic Review, 2020 Page 97 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 5.41. Incremental Cost-Effectiveness Ratios for Infliximab Following Initial Vedolizumab Scenario: Biologic-Naïve Infliximab as Subsequent Treatment Treatment Base-Case Cost per QALY Gained Instead of Market Basket Vedolizumab $887,000 $886,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table 5.42. Incremental Cost-Effectiveness Ratios for Infliximab Following Initial Vedolizumab Scenario: Biologic-Experienced Infliximab as Subsequent Treatment Treatment Cost per QALY Gained Instead of Market Basket Vedolizumab $902,000 $899,000 Conventional Treatment Reference Reference QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Threshold Analyses Results The dosing, mode of administration, and frequency of administration of some TIMs differs in the induction period compared with the maintenance period. In order to generate a single annual threshold price across TIMs, threshold pricing was conducted for price per maintenance year. The annual drug cost per maintenance year for each TIM to reach an incremental cost per QALY gained compared to conventional treatment at thresholds of $50,000 per QALY, $100,000 per QALY, and $150,000 per QALY are presented in Tables ES15 and ES16. Annual TIM prices at the $150,000 per QALY threshold ranged from $6,824 (adalimumab, biologic-experienced population) to $16,624 (ustekinumab, biologic-naïve population). We note, however, that these results are highly sensitive to what amount to very minor differences in estimated QALYs. For example, the somewhat higher threshold prices for ustekinumab in the biologic-naïve population are driven by a QALY difference of 0.037 in comparison to the next most-effective TIM, which is the equivalent of less than two additional weeks of life. This difference is itself based on parameter inputs based on point estimate differences from our NMA that were not statistically different across ustekinumab and other TIMs. It is also important to note that the threshold prices for ustekinumab are for the subcutaneous injection used in the maintenance period and are independent of the IV product price used in the induction period. This is why, for example, we see the threshold price in the biologic-experienced population rising faster with increasing cost per QALY thresholds versus other TIMs (e.g., vedolizumab). ©Institute for Clinical and Economic Review, 2020 Page 98 Final Report – Targeted Immune Modulators for UC Return to Table of Contents With these details in mind, the overall results demonstrate that for all TIMs, except infliximab and the infliximab biosimilars, annual net price estimates were far higher than threshold prices all the way up to prices at $150,000 per QALY. Table 5.43. Threshold Analysis Results – Per Maintenance Year for the Biologic-Naïve Population Price to Price to Price to Achieve WAC Net Price Achieve Achieve $50,000/QALY $100,000/QALY $150,000/QALY Adalimumab $72,427 $46,933 $4,616 $5,778 $6,941 Golimumab $75,324 $42,332 $4,991 $6,320 $7,649 Infliximab $27,930 $14,614* $6,754 $8,813 $10,872 Infliximab- $22,632 $13,451* $6,754 $8,813 $10,872 dyyb Infliximab- $18,018 $13,883* $6,754 $8,813 $10,872 abda Ustekinumab $150,425 $91,609 $9,220 $12,922 $16,624 Vedolizumab $43,842 $44,224* $7,247 $9,454 $11,662 ASP: average sales price, QALY: quality-adjusted life year, WAC: wholesale acquisition cost Prices rounded to nearest $1,000. *Net prices represented by ASP plus 9.5%. Table 5.44. Threshold Analysis Results – Per Maintenance Year for the Biologic-Experienced Population Price to Price to Achieve Price to Achieve WAC Net Price Achieve $100,000/QALY $150,000/QALY $50,000/QALY Adalimumab $72,427 $46,933 $4,512 $5,668 $6,824 Tofacitinib $57,176 $35,506 $9,429 $12,360 $15,292 Ustekinumab $150,425 $91,609 $4,523 $8,144 $11,766 Vedolizumab $43,842 $44,224* $6,738 $8,939 $11,140 QALY: quality-adjusted life year, WAC: wholesale acquisition cost Prices rounded to nearest $1,000. *Net prices represented by ASP plus 9.5%. Prior Economic Models We reviewed the literature for recent cost-effectiveness analyses of treatments for UC, for comparison to the results from our model. A 2018 analysis by Scott et al.166 used a model to simulate outcomes for vedolizumab compared to colectomy. While this analysis included QALY estimates, it did not include costs and used a time horizon of only one to seven years, and so was not directly comparable to our analysis. A cost-effectiveness analysis in the same year by Beilman et al.167 examined the cost effectiveness of vedolizumab compared to infliximab in Canada. They found that vedolizumab and infliximab had similar effectiveness, but vedolizumab was considered ©Institute for Clinical and Economic Review, 2020 Page 99 Final Report – Targeted Immune Modulators for UC Return to Table of Contents more cost effective due to its lower cost over a five-year time horizon. Our model found similar or slightly higher efficacy for vedolizumab compared with infliximab. However, our conclusions would be the opposite, with infliximab considered more cost effective, as vedolizumab has the higher price in the US. This analysis did not include a lifetime horizon scenario. Milev et al.158 have conducted a cost analysis of tofacitinib treatment from the perspective of a US payer over a two- year time horizon, but their analysis did not include QALY estimates. Tappenden et al.168 evaluated adalimumab, golimumab, and infliximab in moderate-to-severe UC patients for whom conventional therapy had failed, using a United Kingdom National Health Services perspective and a lifetime horizon. Their analysis estimated that colectomy would be more effective and less costly than these medical treatments. In patients for whom colectomy was not an option, the cost effectiveness of adalimumab compared to conventional therapy was estimated at just over £50,000 per QALY, with infliximab and golimumab dominated by adalimumab. Their analysis used a similar cohort as ours in terms of starting age, proportion male, and mean body mass index, but did not stratify by prior biologic experience. Their model produced fewer QALYs in general, perhaps because they used a lower utility value for active UC. Their model estimated approximately 0.35 incremental QALYs for adalimumab compared to conventional therapy, which was much higher than our estimates of 0.02, primary due to the availability of subsequent treatment in our conventional therapy arm. Costs were not comparable across the different health care systems of the US and United Kingdom. Wu et al.169 used a Markov model to examine the cost effectiveness of 14 possible treatment sequences (consisting of up to two lines of therapy followed by conventional treatment) for patients with moderate-to-severe UC, using Chinese and United Kingdom perspectives and costs. They reported that treatment sequences including tofacitinib and vedolizumab were most cost effective in the United Kingdom, while treatment with tofacitinib was most cost effective in China. Their cost estimates from China and the United Kingdom were not comparable to those in our US- based analysis, but this analysis also reported lower QALYs than our present analysis, again likely due to the use of lower utility values for active UC (0.42) compared to the value used in the current analysis (0.680) as well as higher discount rates (3.5% for the United Kingdom and 5% for China). This, along with the use of generally higher response rate estimates, led to greater incremental QALY gains (and consequently lower estimated cost-effectiveness ratios) than were found in our analysis. More recently, Lohan et al.170 conducted an NMA and Markov model of tofacitinib compared to biologics and conventional therapy for the treatment of moderate-to-severe UC using a United Kingdom National Health Service perspective and lifetime horizon. Non-responders in their model were assumed to move directly to conventional therapy rather than to another biologic. As with the analyses above, this model used a lower utility value for active UC (0.41) and a higher discount ©Institute for Clinical and Economic Review, 2020 Page 100 Final Report – Targeted Immune Modulators for UC Return to Table of Contents rate (3.5%) than in our analysis. This resulted in lower estimated lifetime QALYs and larger incremental QALY gains from treatment in general. Limitations We have attempted to model TIMs for the treatment of UC to both reflect clinical practice and accommodate the limits of available data. The latter has placed some restrictions on how accurately we can model UC treatment with TIMs. Outcomes for conventional treatment are based on the placebo arm of clinical trials and may not fully reflect the clinical course of disease in the real world. In addition, patients may try several treatment options over a lifetime time horizon whereas our model was limited to only a trial of two TIMs before moving to conventional treatment. For this reason, our model does not reflect a comprehensive disease model of UC but is instead constructed with the intent to evaluate the cost effectiveness of the initial TIM of interest while keeping other factors, such as later line treatment options, relatively constant. The available options for a second TIM are limited to those with efficacy data in a biologic- experienced population. Market shares for the second TIM meeting all inclusion criteria are assumed to be equally distributed. These assumptions may not be reflective of real-world treatment patterns. Some TIMs without RCT data in a biologic-experienced population may be prescribed by some clinicians as a second TIM (e.g., infliximab as a second TIM after non-response or loss of response to vedolizumab as the initial TIM). Without efficacy data in a pre-treated population, we are unable to generate reliable estimates for how well these would perform in a biologic-experienced population and have excluded them from the market basket. This analysis is based on efficacy inputs from the NMA. All limitations of the NMA (e.g., differences in clinical trial design) are also limitations of the model, which heavily relies on these inputs to derive treatment benefit. Finally, we assumed that patients who enter the maintenance phase have the same risk ratio for response without remission and response with remission regardless of whether the patient achieved response with remission or response without remission during induction. We acknowledge that these may differ in reality (e.g., a patient entering in remission may have a higher likelihood of maintaining remission). However, due to data limitations, we are unable to create more granular sets of risk ratios for all comparators. Conclusions In summary, our analyses indicate that TIMs improved health outcomes compared to conventional treatment in both the biologic-naïve and biologic-experienced populations. Using net prices, incremental cost-effectiveness ratio results were above commonly cited thresholds for cost effectiveness for all TIMs in the base-case analysis. ©Institute for Clinical and Economic Review, 2020 Page 101 Final Report – Targeted Immune Modulators for UC Return to Table of Contents In the biologic-naive population, cost-effectiveness ratios for all TIMs were above $150,000 per QALY compared to conventional treatment for nearly all scenarios, with the exception of infliximab and its biosimilars in the modified societal perspective and infliximab-dyyb using a five-year time horizon. None of the TIMs were cost effective compared to infliximab. In the biologic-experienced population, no TIMs had a cost per QALY gained compared to conventional treatment below the $150,000 per QALY threshold. Compared to adalimumab, tofacitinib resulted in lower cost and greater QALYs. Ustekinumab and vedolizumab resulted in greater QALYs but were not cost effective. 5.4 Summary and Comment We estimated the cost effectiveness of TIMs, considering a lifetime time horizon for adult patients with moderate-to-severe UC in biologic-naïve and biologic-experienced populations. Patient time spent in health states of active UC, response without remission, and response with remission was summed to provide estimates of life expectancy, quality-adjusted life expectancy, and evLY gained. Annual net health care costs, including net price drug acquisition, administration, adverse events, and colectomy were summed to estimate lifetime costs for TIMs and conventional treatment. We used a transition matrix for conventional treatment based on the placebo arm of RCTs and applied a risk ratio for TIMs for response without remission and response with remission to derive TIM- specific transition probabilities between health states for induction and maintenance. Based on these assumptions, the cost effectiveness of TIMs was estimated to range from $186,000 to $1,870,000 per QALY in the biologic-naïve population and $495,000 to $1,885,000 per QALY in the biologic-experienced population. It is important to note that the magnitude of difference in the total QALY estimates across interventions was not large (largest difference was 0.11), suggesting an increased importance for the cost differences when interpreting the results. Results for cost per evLY gained were similar but slightly lower than cost per QALY, as a result of a small decrease in mortality with the use of TIMs due to avoidance of colectomy. In general, the incremental cost of TIMs versus conventional treatment was modest given the annual price of TIMs, the use of subsequent treatments, and a lifetime time horizon. In our model, initial non-responders and those who lose response to a TIM are assumed to discontinue treatment. With this assumption, few patients remain on the initial TIM longer than a year. Appendix Table E12 shows the distribution of patients across health states at the end of one, five, and 10 years. Approximately 20% of biologic- naïve patients and 10% of biologic-experienced patients remain in clinical remission or response without remission health states at five years. Scenarios in which TIMs resulted in cost per QALY estimates below the $150,000 per QALY threshold were infliximab and infliximab biosimilars compared with conventional treatment in the biologic-naïve population in the modified societal perspective, infliximab-dyyb using a five-year ©Institute for Clinical and Economic Review, 2020 Page 102 Final Report – Targeted Immune Modulators for UC Return to Table of Contents time horizon, and for tofacitinib compared with adalimumab in the biologic-experienced population. Considerable uncertainty exists in our model, primarily driven by estimates of health utility values, wide confidence intervals of risk ratio estimates produced in the NMA, and subsequent treatment pathways. However, results of the probabilistic sensitivity analysis showed that TIMs are unlikely to be cost effective at the $150,000 per QALY threshold. Based on our analysis, the cost per additional QALY for TIMs is expected to exceed usual thresholds for cost effectiveness. These results were tested under a variety of assumptions and alternative sources of model inputs, few of which drove the incremental cost per QALY below the threshold of $150,000 per QALY gained. ©Institute for Clinical and Economic Review, 2020 Page 103 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 6. Potential Other Benefits and Contextual Considerations Our reviews seek to provide information on potential other benefits offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. We also recognize that there may be broader contextual issues related to the severity of the condition, whether other treatments are available, and ethical, legal, or other societal priorities that influence the relative value of illnesses and interventions. These general elements are listed in Table 6.1, and the subsequent text provides detail about the elements that are applicable to the comparison of TIMs to conventional therapy in UC. We sought input from stakeholders, including individual patients, patient advocacy organizations, clinicians, and manufacturers, to inform the contents of this section. Each ICER review culminates in a public meeting of an independent voting panel of clinicians, patients, and health services researchers. As part of their deliberations, Panel members will judge whether a treatment may substantially impact the considerations listed in Table 6.1. The presence of substantial other benefits or contextual considerations may shift a Panel member's vote on an intervention's long-term value for money to a different category than would be indicated by the clinical evidence and cost-effectiveness analyses alone. For example, a Panel member may initially consider a therapy with an incremental cost-effectiveness ratio of $150,000 per QALY to represent low long-term value for money. However, the Panel member may vote for a higher value category if they consider the treatment to bring substantial other benefits or contextual considerations. Conversely, disadvantages associated with a treatment may lead a Panel member to vote for a lower value category. A Panel member may also determine that there are no other benefits or contextual considerations substantial enough to shift their vote. All factors that are considered in the voting process are outlined in ICER's Value Assessment Framework. The content of these deliberations is described in the last chapter of ICER's Final Evidence Report and Meeting Summary, which is released after the public meeting. This section, as well as the Panel's deliberation, provides stakeholders with information to inform their decisions on a range of issues, including shared decision-making between patients and clinicians, coverage policy development, and pricing negotiations. ©Institute for Clinical and Economic Review, 2020 Page 104 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 6.1. Potential Other Benefits or Contextual Considerations (Not Specific to Any Disease or Therapy) Potential Other Benefits This intervention offers reduced complexity that will significantly improve patient outcomes. This intervention will reduce important health disparities across racial, ethnic, gender, socio-economic, or regional categories. This intervention will significantly reduce caregiver or broader family burden. This intervention offers a novel mechanism of action or approach that will allow successful treatment of many patients for whom other available treatments have failed. This intervention will have a significant impact on improving return to work and/or overall productivity. Other important benefits or disadvantages that should have an important role in judgments of the value of this intervention. Potential Other Contextual Considerations This intervention is intended for the care of individuals with a condition of particularly high severity in terms of impact on length of life and/or quality of life. This intervention is intended for the care of individuals with a condition that represents a particularly high lifetime burden of illness. This intervention is the first to offer any improvement for patients with this condition. Compared to conventional therapy, there is significant uncertainty about the long-term risk of serious side effects of this intervention. Compared to conventional therapy, there is significant uncertainty about the magnitude or durability of the long-term benefits of this intervention. There are additional contextual considerations that should have an important role in judgments of the value of this intervention. 6.1 Potential Other Benefits As noted in Sections 1 and 2, UC confers a significant burden to patients and their caregivers. As a disease that is diagnosed primarily before age 30, UC has an important impact on return to work and/or school as well as overall productivity (i.e., both absenteeism and presenteeism). The benefits of TIMs relative to conventional therapy may translate into significant and durable periods of clinical remission, allowing patients to resume normal activities and reducing caregiver impact. We also heard from both patient advocacy organizations and clinicians that, similar to other chronic inflammatory conditions, UC is a disease with treatment patterns that involve relatively frequent switching due to lack of or loss of response, both within and across classes of TIMs. Novel mechanisms of action, such as that offered by ustekinumab, the newest addition to the armamentarium, provide additional options for patients with UC whose disease has stopped responding to other TIM classes. In addition, available UC therapies include oral, self-injectable, and infused products; patients tend to have clear preferences for method of delivery. For example, some may value the freedom and independence provided by oral or self-injectable treatments, ©Institute for Clinical and Economic Review, 2020 Page 105 Final Report – Targeted Immune Modulators for UC Return to Table of Contents while others may place more weight on the regular clinician interactions that come with scheduled infusions. Finally, we heard from the Crohn's and Colitis Foundation that there are important aspects of the condition (e.g., pain, fatigue, fecal urgency) not adequately captured by trial-based clinical endpoints and therefore not fully reflected in the economic model. If TIM therapy addresses these concerns wholly or in part, there may be benefits to patients and caregivers that have not been captured by this review. 6.2 Contextual Considerations As noted above, UC poses a significant lifetime burden on quality of life, and many patients fear the prospect of surgical intervention and its attendant complications. In comparison to other chronic inflammatory diseases, RCT evidence for UC is actually relatively lengthy, with comparative information available out to one year or more in most circumstances. However, comparative long-term observational data, particularly on safety concerns that may have been raised during clinical development, vary in availability. For example, while the safety profile of the TNF inhibitors has been relatively well established in UC and other chronic inflammatory conditions, these data are sparse for newer TIMs, such as tofacitinib, ustekinumab, and vedolizumab. It should be noted that uncertainty surrounding both clinical effectiveness and safety is most pronounced in pediatric populations where we identified only a single RCT that was not actually a comparison of alternative TIM therapies (i.e., a comparison of two dose regimens of infliximab). Infliximab remains the only agent with an FDA indication in children and adolescents, and as such, there is substantial uncertainty about the long-term benefits and risks of other TIMs in these patients-uncertainty that will hopefully be addressed by future clinical trials and/or high-quality observational studies. ©Institute for Clinical and Economic Review, 2020 Page 106 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 7. Health-Benefit Price Benchmarks The health benefit price benchmark (HBPB) is a price range suggesting the highest price a manufacturer should charge for a treatment, based on the amount of improvement in overall health patients receive from that treatment, when a higher price would cause disproportionately greater losses in health among other patients due to rising overall costs of health care and health insurance. In short, it is the top price range at which a health system can reward innovation and better health for patients without doing more harm than good. Annual prices of each TIM that would achieve incremental cost-effectiveness ratios of $100,000 and $150,000 per QALY for the biologic-naïve and biologic-experienced populations are presented in Tables 7.1 and 7.2, respectively. The cost per evLYG price ranges are almost identical to the cost per QALY ranges due to the very similar evLYG and QALYs gained in the base-case analysis, and as such, are not presented separately here. Table 7.1. Annual Cost-Effectiveness Threshold Prices per Maintenance Year for TIMs for the Treatment of UC in the Biologic-Naïve Population Annual Annual Price at Annual Price at Discount from WAC Annual WAC Estimated $100,000 $150,000 to Reach Threshold Net Price Threshold Threshold Prices Adalimumab $72,400 $46,900 $5,800 $6,900 90%-92% Golimumab $75,300 $42,300 $6,300 $7,600 90%-92% Infliximab $27,900 $14,600 $8,800 $10,900 61%-68% Infliximab-dyyb $22,600 $13,500 $8,800 $10,900 52%-61% Infliximab-abda $18,000 $13,900 $8,800 $10,900 40%-51% Ustekinumab $150,400 $91,600 $12,900 $16,600 89%-91% Vedolizumab $43,800 $44,200 $9,500 $11,700 73%-78% WAC: wholesale acquisition cost Prices rounded to nearest $100. Table 7.2. Annual Cost-Effectiveness Threshold Prices per Maintenance Year for TIMs for the Treatment of UC in the Biologic-Experienced Population Annual Annual Price Annual Price Discount from WAC Annual WAC Estimated at $100,000 at $150,000 to Reach Threshold Prices Net Price Threshold Threshold Adalimumab $72,400 $46,900 $5,700 $6,800 91%-92% Tofacitinib $57,200 $35,500 $12,600 $15,300 73%-78% Ustekinumab $150,400 $91,600 $8,100 $11,800 92%-95% Vedolizumab $43,800 $44,200 $8,900 $11,100 75%-80% WAC: wholesale acquisition cost Prices rounded to nearest $100. ©Institute for Clinical and Economic Review, 2020 Page 107 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Across both populations, the HBPB ranges for adalimumab (approximately $6,000 to $7,000 per year) would require discounts of approximately 90% to 92% from WAC. Golimumab would require discounts of approximately 90% from WAC, for a HBPB range from $6,300 to $7,600 per year. The HBPB range for infliximab ($8,800-$10,900) would require discounts of 61% to 68% from WAC, with smaller discounts required for infliximab-dyyb (52% to 61%) and infliximab-abda (40% to 51%). The HBPB range for ustekinumab across both populations ranged from $8,100 to $16,600, representing discounts of 89% to 95%. For vedolizumab, the HBPB range across both populations is $8,900 to $11,700, requiring 73% to 80% discounts from WAC. ©Institute for Clinical and Economic Review, 2020 Page 108 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 8. Potential Budget Impact 8.1 Overview We used the cost-effectiveness model to estimate the potential total budgetary impact of the recently expanded indication of ustekinumab for prevalent individuals in the US with moderate-to- severe UC. In our estimates of potential budget impact, we used the WAC, estimated net price, and $50,000, $100,000, and $150,000 cost-effectiveness threshold prices that were weighted averages of the threshold prices for the biologic-naïve and biologic-experienced populations eligible for ustekinumab. We did not include the other therapies modeled above in this potential budget impact analysis given their established presence on the market for UC. 8.2 Methods We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact. Potential budget impact was defined as the total differential cost of using this new therapy rather than relevant existing therapies for the treated population, calculated as differential health care costs (including drug costs) minus any offsets in these costs from averted health care events. All costs were undiscounted and estimated over a five-year time horizon, given the potential for cost offsets to accrue over time and to allow a more realistic impact on the number of patients treated with the new therapy. This potential budget impact analysis includes the estimated number of individuals with UC in the US who would be eligible for treatment with ustekinumab. To estimate the size of the potential candidate populations for treatment, we used an estimate by Turner et al. of the prevalence of individuals with UC in the US of approximately 900,000 patients.4 The Crohn's and Colitis Foundation has reported that approximately 22% of UC patients have moderate-to-severe disease activity in a given year, which would equate to approximately 198,000 patients with moderate-to- severe UC in the US.20 To estimate the proportions of these patients who would be biologic-naïve versus biologic-experienced, we used the weighted average of the baseline distribution of patients in the relevant trials that enrolled a "mixed" population (i.e., both biologic-naïve and biologic- experienced), resulting in an estimate of approximately 55% of patients who were not using biologics and 45% who had previously used biologics. Applying these proportions resulted in estimates of 108,900 eligible patients who were biologic-naïve and 89,100 who were biologic- experienced. For the purposes of this analysis, we assumed that 20% of these patients would initiate ustekinumab in each of the five years, or 21,780 biologic-naïve patients per year and 17,820 biologic-experienced patients per year. ©Institute for Clinical and Economic Review, 2020 Page 109 Final Report – Targeted Immune Modulators for UC Return to Table of Contents For patients eligible for ustekinumab, we assumed that patients could be drawn from all other available treatment options for biologic-naïve patients (i.e., adalimumab, golimumab, infliximab, infliximab-dyyb, infliximab-abda, vedolizumab, and conventional treatment), and from all other available treatment options for biologic-experienced patients (i.e., adalimumab, tofacitinib, vedolizumab, and conventional treatment). As in the base-case cost-effectiveness analysis, proportions of treatments in each population were assumed to be equal in the absence of contemporary real-world data on utilization patterns in the US. ICER's methods for estimating potential budget impact are described in detail elsewhere171 and have been recently updated. The intent of our revised approach to potential budgetary impact is to document the percentage of patients who could be treated at selected prices without crossing a potential budget impact threshold that is aligned with overall growth in the US economy. For 2019- 2020, the five-year annualized potential budget impact threshold that should trigger policy actions to manage access and affordability is calculated to be approximately $819 million per year for new drugs. 8.3 Results Table 8.1 illustrates the five-year annualized per-patient potential budget impact of ustekinumab compared to the blended market basket of treatments in the biologic-naive and biologic- experienced populations, assuming a fixed ratio of 55% biologic-naïve and 45% biologic- experienced patients (based on the baseline distribution of patients in the trials). The results are based on the list price ($150,425 per year), the net price ($91,609 per year), and the annual weighted-average threshold prices (i.e., in a mixed population of biologic-naïve and biologic- experienced patients) for cost-effectiveness thresholds of $150,000, $100,000, and $50,000 per QALY versus conventional treatment (approximately $14,400, $10,800, and $7,100, respectively). Note that this analysis uses results from the cost-effectiveness model, which accounts for treatment discontinuation and impact of treatments on total net costs. The average annualized potential budgetary impact when using the list price of ustekinumab was an additional per-patient cost of approximately $36,100 and approximately $15,600 using the net price. The weighted-average threshold prices for $50,000 to $150,000 per QALY were estimated to produce cost savings relative to the treatment market basket, because of the relatively higher cost offset from the comparator mix ($37,500 per patient), which includes several biologics at their net prices. ©Institute for Clinical and Economic Review, 2020 Page 110 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table 8.1. Annualized Per-Patient Potential Budget Impact Over a Five-Year Time Horizon for Ustekinumab in a UC Population Assuming 55% Biologic-Naïve and 45% Biologic-Experienced Average Annual Per Patient Budget Impact At $150,000/ At $100,000/ At $50,000/ At List Price At Net Price QALY Price QALY Price QALY Price Ustekinumab $73,600 $53,100 $28,800 $27,600 $26,400 55% Naïve/45% Experienced $37,500 Market Basket Net Impact $36,100 $15,600 -$8,700 -$9,900 -$11,100 QALY: quality-adjusted life year All annualized costs include drug and non-drug health care costs. Numbers may not sum due to rounding. To estimate potential budget impact in the overall population eligible for ustekinumab, we assumed a fixed ratio of 55% biologic-naïve and 45% biologic-experienced patients, as above. In the overall population eligible for ustekinumab, as shown in Figure 8.1, approximately 21% of eligible patients could be treated in a given year without crossing the ICER budget impact threshold of $819 million at the WAC of ustekinumab. Approximately 51% of eligible patients could be treated without crossing the budget impact threshold at its estimated net price. All eligible patients could be treated at the $150,000, $100,000, and $50,000 threshold prices, with potential budget impact estimated to be cost saving across the $150,000 to $50,000 threshold prices. Figure 8.1. Potential Budget Impact Scenarios of Ustekinumab versus Market Basket Treatment Mix at Different Acquisition Prices $160,000 WAC $140,000 $120,000 $100,000 Annual Price Net Price $80,000 $60,000 $40,000 $20,000 $0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percentage of Patients Treated Without Crossing BI Threshold Each Year BI: budget impact, WAC: wholesale acquisition cost ©Institute for Clinical and Economic Review, 2020 Page 111 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 8.4 Summary Potential budget impact in the overall population eligible for ustekinumab could be relatively high, with treatment of approximately 51% of eligible patients surpassing the budget impact threshold at its estimated net price. In contrast, all eligible patients could be treated at the $150,000, $100,000, and $50,000 threshold prices, with the potential for cost savings at these prices. ©Institute for Clinical and Economic Review, 2020 Page 112 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 9. Summary of the Votes and Considerations for Policy 9.1 About the CTAF Process During CTAF public meetings, the CTAF Panel deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of treatments under examination, and the supplementary information presented. Panel members are not pre- selected based on the topic being addressed and are intentionally selected to represent a range of expertise and diverse perspectives. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting topic and provide input to CTAF Panel members before the meeting to help clarify their understanding of the different interventions being analyzed in the evidence review. The same clinical experts serve as a resource to the CTAF Panel during their deliberation and help to shape recommendations on ways the evidence can apply to policy and practice. After the CTAF Panel votes, a policy roundtable discussion is held with the CTAF Panel, clinical experts, patient advocates, payers, and when feasible, manufacturers. The goal of this discussion is to bring stakeholders together to apply the evidence to guide patient education, clinical practice, and coverage and public policies. Participants on policy roundtables are selected for their expertise on the specific meeting topic, are different for each meeting, and do not vote on any questions. At the September 24 meeting, the CTAF Panel discussed issues regarding the application of the available evidence to help patients, clinicians, and payers address important questions related to the use of TIMs for UC. Following the evidence presentation and public comments (public comments from the meeting can be accessed here, the CTAF Panel voted on key questions concerning the comparative clinical effectiveness, comparative value, and potential other benefits and contextual considerations related to TIMs. These questions are developed by the ICER research team for each assessment to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice, medical policy decisions, and patient decision-making. The voting results are presented below, along with specific considerations mentioned by CTAF Panel members during the voting process. In its deliberations and votes related to value, the CTAF Panel considered the individual patient benefits, and incremental costs to achieve such benefits, from a given intervention over the long term. ©Institute for Clinical and Economic Review, 2020 Page 113 Final Report – Targeted Immune Modulators for UC Return to Table of Contents There are four elements to consider when deliberating on long-term value for money (see Figure 9.1 below): 1. Comparative clinical effectiveness is a judgment of the overall difference in clinical outcomes between two interventions (or between an intervention and placebo), tempered by the level of certainty possible given the strengths and weaknesses of the body of evidence. CTAF uses the ICER Evidence Rating Matrix as its conceptual framework for considering comparative clinical effectiveness. 2. Estimated incremental cost-effectiveness is the average incremental cost per patient of one intervention compared to another to achieve a desired "health gain," such as an additional stroke prevented, case of cancer diagnosed, or gain of a year of life. Alternative interventions are compared in terms of cost per unit of effectiveness, and the resulting comparison is presented as a cost-effectiveness ratio. Relative certainty in the cost and outcome estimates continues to be a consideration. As a measure of cost-effectiveness, the CTAF Panel follows common academic and health technology assessment standards by using cost per quality-adjusted life year (QALY), with formal voting on "long-term value for money" when the base case incremental cost-effectiveness ratio is between $50,000 per QALY and $175,000 per QALY. 3. Potential other benefits refer to any significant benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Examples of potential other benefits include better access to treatment centers, mechanisms of treatment delivery that require fewer visits to the clinician's office, treatments that reduce disparities across various patient groups, and new potential mechanisms of action for treating clinical conditions that have demonstrated low rates of response to currently available therapies. Other disadvantages could include increased burden of treatment on patients or their caregivers. For each intervention evaluated, it will be open to discussion whether potential other benefits or disadvantages such as these are important enough to factor into the overall judgment of long-term value for money. There is no quantitative measure for potential other benefits or disadvantages. 4. Contextual considerations include ethical, legal, or other issues (but not cost) that influence the relative priority of illnesses and interventions. Examples of contextual considerations include whether there are currently any existing treatments for the condition, whether the condition severely affects quality of life or not, and whether there is significant uncertainty about the magnitude of benefit or risk of an intervention over the long term. There is no quantitative measure for contextual considerations. ©Institute for Clinical and Economic Review, 2020 Page 114 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure 9.1. Conceptual Structure of Long-Term Value for Money 9.2 Voting Results 1) Is the evidence adequate to demonstrate that the net health benefit of vedolizumab is superior to that provided by adalimumab? 2 Yes: 12 votes No: 2 votes A majority of the Panel determined that the evidence was adequate to demonstrate a superior net health benefit of vedolizumab versus adalimumab. The Panel was persuaded by the results of the head-to-head VARSITY trial, in which vedolizumab demonstrated higher rates of clinical response compared to adalimumab. 2) Is the evidence adequate to demonstrate that the net health benefit of ustekinumab is superior to that provided by adalimumab? Yes: 0 votes No: 15 votes The Panel unanimously judged that the evidence was inadequate to demonstrate a superior net health benefit of ustekinumab versus adalimumab. The Panel cited the results of the ICER NMA and 2 Due to technological issues, one CTAF Panelist was unable to cast a vote for the first question. ©Institute for Clinical and Economic Review, 2020 Page 115 Final Report – Targeted Immune Modulators for UC Return to Table of Contents indirect evidence that concluded that ustekinumab demonstrates a health benefit that is at least comparable, but not superior, relative to adalimumab 3) Is the evidence adequate to distinguish the net health benefit among tofacitinib, ustekinumab, and vedolizumab? Yes: 1 vote No: 14 votes A majority of the Panel concluded that the evidence was inadequate to distinguish the net health benefit among tofacitinib, ustekinumab, and vedolizumab. The Panel vote was driven by the results of the NMA and the lack of head-to-head trials comparing the TIMs. 4) When compared to conventional therapy, does treating patients with TIMs offer one or more of the following potential "other benefits"? (Select all that apply.) These interventions offer reduced complexity that will significantly improve patient outcomes. 2/15 These interventions will significantly reduce caregiver or broader family burden. 9/15 These interventions offer a novel mechanism of action or approach that will allow successful 13/15 treatment of many patients for whom other available treatments have failed. These interventions will have a significant impact on improving patients' ability to return to work 12/15 and/or their overall productivity. There are other important benefits or disadvantages that should have an important role in 2/15 judgments of the value of these interventions. A majority of the Panel recognized the impact that TIMs may offer on caregiver or broader family burden. Numerous Panel members also stressed that TIMs may have a significant impact on a patient's ability to return to work or school. Panelists cited patient testimony: patient experts on the Policy Roundtable emphasized the potential benefits offered by TIMs, such as long-lasting remission, which supports increased independence, ability to commute to work and/or school and remain focused throughout the day, and attend and graduate high school and college. Panelists also noted that the existence of multiple classes of TIMs each with different mechanisms of action helps ensure that patients are more likely to find an optimal treatment. Though patients may cycle through multiple drugs before they find the best treatment, the ability to switch to a different medication within a different class is essential to achieving and maintaining remission. ©Institute for Clinical and Economic Review, 2020 Page 116 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 5) Are any of the following contextual considerations important in assessing the long-term value for money of TIMs? (Select all that apply.) These interventions are intended for the care of individuals with a condition of particularly high 12/15 severity in terms of impact on length of life and/or quality of life. These interventions are intended for the care of individuals with a condition that represents a 13/15 particularly high lifetime burden of illness. These interventions are the first to offer any improvement for patients with this condition. 0/15 Compared to conventional therapy, there is significant uncertainty about the long-term risk of 13/15 serious side effects of these interventions. Compared to conventional therapy, there is significant uncertainty about the magnitude or 12/15 durability of the long-term benefits of these interventions. There are additional contextual considerations that should have an important role in judgments 0/15 of the value of these interventions. A majority of the Panel recognized that UC is a condition of high severity with substantial impacts on quality of life. The Panel also noted that significant uncertainty remains regarding the magnitude or durability of long-term benefits, as well as the long-term risks associated with TIMs. Though some TIMs, such as the TNF inhibitors, have been on the market for a long period of time, there is more limited evidence regarding the long-term durability and safety around newer treatments like tofacitinib, ustekinumab, and vedolizumab. 6) Given the available evidence on comparative effectiveness and incremental cost effectiveness, and considering other benefits, disadvantages, and contextual considerations, what is the long- term value for money of treatment at current pricing with infliximab, infliximab-abda, and infliximab-dyyb versus conventional treatment? Low: 3 votes Intermediate: 10 votes High: 2 votes A majority of the Panel judged the long-term value for money of treatment with infliximab, infliximab-abda, and infliximab-dyyb as "intermediate." The intermediate vote was driven by the clinical and economic evidence and the various potential other benefits and contextual considerations offered by infliximab and its biosimilars. In a majority of RCTs, infliximab had higher rates of both clinical response and clinical remission compared to placebo. In addition, rates of endoscopic improvement were higher for infliximab compared to placebo as well. In conjunction, Panelists emphasized that the safety and side effect profile is relatively well established as infliximab has been on the market for a while. Though infliximab and its biosimilars carry black box warnings for increased risk of lymphomas and malignancies, the Panel noted that in the evidence base, overall rates of new malignancy were very low (<3%) for TNF inhibitors. The "intermediate" vote was also driven by the results of the base-case economic analysis as well as other scenario analyses in which infliximab and its biosimilars fell near, at, or below commonly cited ©Institute for Clinical and Economic Review, 2020 Page 117 Final Report – Targeted Immune Modulators for UC Return to Table of Contents cost-effectiveness thresholds. Although Panelists acknowledged that the cost of infliximab and its biosimilars is still too high, taking the clinical and economic evidence and potential other benefits and contextual considerations into account, the Panel determined an overall intermediate long- term value for money. 9.3 Roundtable Discussion and Key Policy Implications Following its deliberation on the evidence, the CTAF Panel engaged in a moderated discussion with a policy roundtable about how best to apply the evidence on TIMs for UC to policy and practice. The policy roundtable members included two patients/patient experts, two clinical experts, two payers, and two representatives from pharmaceutical manufacturers. The discussion reflected multiple perspectives and opinions, and therefore, none of the statements below should be taken as a consensus view held by all participants. The names of the Policy Roundtable participants are shown below, and conflict of interest disclosures for all meeting participants can be found in the Appendix. Table 9.1 Policy Roundtable Participants Name Title and Affiliation Thomas Brownlie, PhD, MS Senior Director, US Payer Policy, Pfizer Deborah Gan, MBA Leader, US Payer Marketing, Primary Care, Merck Patrick Gleason, PharmD Assistant Vice President, Health Outcomes, Prime Therapeutics Bruce Sands, MD, MS Professor of Medicine and Chief, Division of Gastroenterology, Mount Sinai Assistant Professor of Medicine, UC San Diego School of Medicine; Siddharth Singh, MD Gastroenterologist, UC San Diego Health Megan Starshak Patient Advocate Director, Inflammatory Bowel Disease Program, Kaiser Permanente Northern Fernando Velayos, MD California Senior Vice President, Education, Support, and Advocacy, Crohn's & Colitis Laura Wingate Foundation Clinicians, Payers, Manufacturers, and Patient Groups The significantly lower prices seen for infliximab and its biosimilars speaks to the important potential for improved value with broader availability and uptake of biosimilar treatment options. All stakeholders should collaborate to ensure that TIM biosimilars have an increasing and comprehensive role in the UC treatment landscape. The biosimilar market is a $2.8 billion annualized business with health care savings now estimated at $5.6 billion per year. A recent analysis by Bernstein analyst Ronny Gal estimated that biosimilar versions of infliximab accounted for 42% of the savings as the price for the brand-name biologic dropped about 48%.172 Although the resulting net price for infliximab and its biosimilars did not reach traditional cost-effectiveness thresholds in the ICER analysis, their cost effectiveness was far superior to that of other treatment options. Roundtable participants emphasized the benefit of ©Institute for Clinical and Economic Review, 2020 Page 118 Final Report – Targeted Immune Modulators for UC Return to Table of Contents having biosimilars to infliximab available in the UC treatment armamentarium and expressed frustration that other biosimilars are FDA-approved (e.g., adalimumab) but not yet available on the US market. In addition, it was noted that the Affordable Care Act (ACA), which provided the regulatory pathway for biosimilar agents, is under threat. Clinicians, payers, manufacturers, and patient groups should collaborate to develop an approach for broader biosimilar access, including a contingency plan in the event the ACA is overturned. Manufacturers and Payers The "bundled rebate" approach, in which rebates are provided at the drug level across all of its possible indications, should be abolished and replaced with an indication- and value-based pricing approach. Several TIMs in this review carry indications for multiple inflammatory conditions. Manufacturers have historically been able to negotiate on a "bundled" basis, offering a single price and rebate across all indications. Our cost-effectiveness analyses indicated that the pricing of all TIMs (with the possible exception of infliximab biosimilars) was far out of alignment with the benefits delivered. Abolishing the bundled rebate approach and replacing it with pricing that is tied to the value brought by a given TIM for each indication, would allow payers to relax certain step therapy requirements and increase patient access to all TIMs. Payers Insurance coverage should be structured to prevent situations in which patients are forced to choose a treatment approach on the basis of cost. Patient input on the Roundtable indicated a variety of views on and experiences with surgical colectomy. Some also suggested that, due to copayment and coinsurance structures, some reluctant patients might nevertheless opt for colectomy as a "cheaper" option if they are having difficulty paying for their medications. Payers should take particular pains to ensure that benefit structures are sufficiently flexible so that patients wishing to avoid a colectomy have other options at their disposal. Specialty society guidelines and drug labels should be monitored for changes, with coverage policy adjusted accordingly. Findings from our review and discussion at the Roundtable noted several instances in which payer coverage policy has not matched changes in clinical guidelines or in the drug label. For example, vedolizumab's label changed in 2019 to remove the requirement for an initial trial of TNF inhibitors. Despite this change, payer coverage policies have generally continued to require use of a TNF ©Institute for Clinical and Economic Review, 2020 Page 119 Final Report – Targeted Immune Modulators for UC Return to Table of Contents inhibitor before vedolizumab can be given. Payers should routinely monitor guidelines and label changes just as they would for new clinical evidence and modify coverage policy accordingly. Because there are no clear biomarkers or predictors of the success for any given treatment in UC, it is not unreasonable to consider prior authorization criteria in order to manage the costs of expensive medications and negotiate prices for TIMs priced beyond a fair range. However, prior authorization criteria should be based on clinical evidence, specialty society guidelines, and input from clinical experts and patient groups. The process for authorization should be clear and efficient for providers. Options for specific elements of coverage criteria within insurance coverage policy are discussed below. Patient Eligibility Criteria • Diagnosis: Because a diagnosis of UC is made based on clinical symptoms and endoscopic investigation, physician attestation is sufficient for diagnosis. • Patient population: Patients eligible for TIMs include those with moderate-to-severe UC whose disease has had an inadequate response to conventional systemic therapy. Patient eligibility criteria should be flexible given that clinical trials used tools (e.g., Mayo Score for disease severity) that are not routinely used in clinical practice. Relying on physician attestation of the level of disease severity is the most common approach taken by insurers. Inadequate response to conventional systemic therapy is the facet of clinical criteria that insurers may choose to define by specifying particular types of systemic therapies, number of attempts, or duration. This approach is reasonable as long as there is a valid citation or reference for the specifications given. Measurement of therapy "failure" in clinical trials is based on the Mayo Score, but as noted, this should not be used as a criterion within insurance coverage. • Exclusions: UC patients with mild disease and those without a prior trial of conventional systemic agents are not eligible for TIM therapy. Step Therapy Given the lack of biomarkers and other predictors of TIM treatment success in UC, it is not unreasonable to use step therapy in this case to manage the costs of treatment. Step therapy among agents for UC appears to meet criteria for reasonable step therapy: • Use of the first-step therapy reduces overall health care spending, not just drug spending. • The first-step therapy is clinically appropriate for all or nearly all patients and does not pose a greater risk of any significant side effect or harm. • Patients will have a reasonable chance to meet their clinical goals with first-step therapy. ©Institute for Clinical and Economic Review, 2020 Page 120 Final Report – Targeted Immune Modulators for UC Return to Table of Contents • Failure of the first-step drug and the resulting delay in beginning the second-step agent will not lead to long-term harm for patients. For step therapy to be reasonable, it must ensure that patients are not required to retry a first-line drug with which they have previously had adverse side effects or an inadequate response at a reasonable dose and duration. In addition, the exception process must be rapid, transparent, and administratively competent; electronic systems for exceptions should be employed whenever possible to minimize time and paperwork burdens for patients and providers. Similarly, any cost savings realized from step therapy protocols (e.g., originator to biosimilar switch) should be returned to the patient as efficiently as possible, using electronic systems as feasible. Required Switching • Required switching of TIM therapy for patients who are stable on current treatment should be limited to switches to another medication with the same mechanism of action or from an originator to a biosimilar agent. Given the availability of multiple classes of TIMs for UC and the benefits that sustained remission provide to patients, requiring a switch to another class for a patient who is currently responding to treatment raises a risk that the patient could have new significant side effects or insufficient response, either of which renders this kind of switch unreasonable. Required switches should be within-class only (e.g., between TNF inhibitors) or from an originator to a biosimilar product (e.g., infliximab). Even for required switches within the same class, if the switch requires the patient to adopt a different route of administration, e.g., IV infusion instead of subcutaneous injection, there should be provisions to allow for exceptions if a patient's living or caregiver situation makes the switch infeasible. In addition, as with step therapy, any switching policy must ensure that patients are not required to switch to a drug that they have used before at a reasonable dose and duration with inadequate response and/or significant side effects, including earlier use under a different payer. We note that switching policies can be deeply resented by patients and clinicians and should only be contemplated if coordinated efforts are also made to educate providers and patients; the success of Kaiser Permanente's initiative to switch patients from the originator infliximab to a biosimilar is an example of such a comprehensive approach. Provider Qualification Restrictions • TIM therapy should be prescribed and managed by gastroenterologists with specific training and expertise in UC. Several stakeholders have noted gaps in clinical practice when the care of UC patients is managed by those without specific training and expertise, including overuse of steroids to manage recurrence of symptoms. ©Institute for Clinical and Economic Review, 2020 Page 121 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Patient Advocacy Organizations Patient advocacy organizations should be an active voice in noting the potentially negative effects of TIM pricing on patient access. Patient Roundtable participants recognized the challenges posed by TIM pricing (and payer response to this) on patients' ability to access the drug they need at the time they need it. Patient groups can represent a strong voice for pricing moderation to align more closely with clinical value, increased use of and access to biosimilars, and other efforts to modulate the pricing-access tension. Specialty Societies Consensus guidelines should be developed across the major gastroenterology societies, in collaboration with patient groups, to ensure a common voice for UC treatment guidance. Several Roundtable participants noted that, unlike other specialties such as cardiology, the major gastroenterology societies (the AGA and the ACG) maintain separate guidelines. Development of common consensus guidelines, with direct input from patient groups, would allow payers to more closely align coverage policy. Such guidelines should involve clinical experts who are free from significant financial or other conflicts of interest to ensure their patient-centricity. Regulators Given the maturity and longstanding use of several of the TIMs of focus in this review, the FDA should require the inclusion of active control arms in Phase III clinical trials of UC treatments. As noted in this review, only one of the 19 RCTs in the available evidence base featured a head-to- head comparison between TIMs. This is despite the longstanding availability of several of these TIMs on the US market. Clinical trials of new agents for other chronic inflammatory conditions such as rheumatoid arthritis and psoriasis now routinely feature an active comparator, including some of the same TIMs featured in this review (e.g., adalimumab). There is no obvious reason why the same approach cannot be taken in UC. ©Institute for Clinical and Economic Review, 2020 Page 122 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Researchers The research community should make a strong commitment to generate real-world evidence that can fill in the gaps from available RCTs and allow for comprehensive comparisons of TIMs. Our review noted several gaps in evidence that RWE may be positioned to fill, such as ongoing monitoring of the safety of newer TIMs as well as information on clinical benefits observed in populations without RCT evidence but with clear real-world experience (e.g., infliximab in biologic- experienced patients). In addition, there is limited understanding of the trajectory of disease and relative effectiveness of TIMs in African American, Asian, and LatinX population. The Crohn's and Colitis Foundation's IBD Plexus initiative is a nationwide registry of over 16,000 patients with UC and Crohn's disease, and research is already underway to gain a better understanding of how IBD affects minority populations, the quality of IBD care, and better understanding of disease severity in pediatric populations. This should be supplemented with other efforts to fill in data gaps both in IBD Plexus and other large data networks such as PCORNet and the Sentinel Research Network. Further clinical study should be conducted to ascertain the optimal sequencing of TIM therapy in UC. Clinical experts on the roundtable noted that there are currently no available tools with which to predict response to specific UC treatments, and there is no current evidence that involves robust comparisons of different sequences of TIM therapy. Future clinical study could compare clinical benefit and safety for sequences involving within-class switching versus switching outside of class, for example, or long-term outcomes when the same agent is used for both induction and maintenance in comparison to a switch after induction. **** This is the first CTAF review of TIMs for UC. ©Institute for Clinical and Economic Review, 2020 Page 123 Final Report – Targeted Immune Modulators for UC Return to Table of Contents References 1. American Society of Colon and Rectal Surgeons. Ulcerative Colitis. 2019; https://www.fascrs.org/patients/disease-condition/ulcerative-colitis. Accessed September 23, 2019. 2. Mayo Clinic. Ulcerative Colitis. 2019; https://www.mayoclinic.org/diseases- conditions/ulcerative-colitis/symptoms-causes/syc-20353326. Accessed September 23, 2019. 3. Bradley GM, Oliva-Hemker M. Pediatric ulcerative colitis: current treatment approaches including role of infliximab. Biologics : targets & therapy. 2012;6:125-134. 4. Turner D, Levine A, Escher JC, et al. Management of pediatric ulcerative colitis: joint ECCO and ESPGHAN evidence-based consensus guidelines. Journal of pediatric gastroenterology and nutrition. 2012;55(3):340-361. 5. Foundation CsaC. Facts about inflammatory bowel diseases. 2020. 6. Rubin DT, Ananthakrishnan AN, Siegel CA, Sauer BG, Long MD. ACG Clinical Guideline: Ulcerative Colitis in Adults. Am J Gastroenterol. 2019;114(3):384-413. 7. Park KT, Ehrlich OG, Allen JI, et al. The Cost of Inflammatory Bowel Disease: An Initiative From the Crohn's & Colitis Foundation. Inflamm Bowel Dis. 2020;26(1):1-10. 8. Sands BE, Peyrin-Biroulet L, Loftus EV, et al. Vedolizumab versus Adalimumab for Moderate-to- Severe Ulcerative Colitis. New England Journal of Medicine. 2019;381(13):1215-1226. 9. Motoya S, Watanabe K, Ogata H, et al. Vedolizumab in Japanese patients with ulcerative colitis: A Phase 3, randomized, double-blind, placebo-controlled study. PloS one. 2019;14(2):e0212989- e0212989. 10. Wollenhaupt J, Lee E-B, Curtis JR, et al. Safety and efficacy of tofacitinib for up to 9.5 years in the treatment of rheumatoid arthritis: final results of a global, open-label, long-term extension study. Arthritis Research & Therapy. 2019;21(1):89. 11. Administration FaD. REMICADE (infliximab) lyophilized concentrate for injection, for intravenous use [package insert]. 1998. 12. Food and Drug Administration. HUMIRA (adalimumab) injection, for subcutaneous use [package insert]. 2002. 13. Food and Drug Administration. SIMPONI (golimumab) injection, for subcutaneous use [package insert]. 2009. 14. Food and Drug Administration. XELJANZ (tofacitinib) tablets, for oral use [package insert]. 2012. 15. Ma C, Panaccione R, Fedorak RN, et al. Heterogeneity in Definitions of Endpoints for Clinical Trials of Ulcerative Colitis: A Systematic Review for Development of a Core Outcome Set. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association. 2018;16(5):637-647.e613. 16. Gibson PR, Vaizey C, Black CM, et al. Relationship between disease severity and quality of life and assessment of health care utilization and cost for ulcerative colitis in Australia: a cross- sectional, observational study. J Crohns Colitis. 2014;8(7):598-606. 17. Brown C, Gibson PR, Hart A, et al. Long-term outcomes of colectomy surgery among patients with ulcerative colitis. Springerplus. 2015;4:573. 18. SSR Health L. Data on File. In:2020. 19. CMS. Payment Allowance Limits for Medicare Part B Drugs Effective July 1, 2020 through September 30, 2020. 2020; https://www.cms.gov/apps/ama/license.asp?file=https%3A//www.cms.gov/files/zip/july-2020- asp-pricing-file.zip. Accessed August 6, 2020. ©Institute for Clinical and Economic Review, 2020 Page 124 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 20. Crohn's & Colitis Foundation of America. The Facts about Inflammatory Bowel Diseases. 2014. 21. Levine JS, Burakoff R. Extraintestinal manifestations of inflammatory bowel disease. Gastroenterology & hepatology. 2011;7(4):235-241. 22. Lamb CA, Kennedy NA. British Society of Gastroenterology consensus guidelines on the management of inflammatory bowel disease in adults. 2019;68(Suppl 3):s1-s106. 23. Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med. 1987;317(26):1625-1629. 24. Takeda. Complete response letter. 2019. 25. Irvine EJ. Development and subsequent refinement of the inflammatory bowel disease questionnaire: a quality-of-life instrument for adult patients with inflammatory bowel disease. Journal of pediatric gastroenterology and nutrition. 1999;28(4):S23-27. 26. Irvine EJ. Quality of life of patients with ulcerative colitis: past, present, and future. Inflamm Bowel Dis. 2008;14(4):554-565. 27. Maruish ME, Maruish M, Kosinski M, et al. User's manual for the SF36v2 Health Survey. 2011. 28. COTEUR G, FEAGAN B, KEININGER DL, KOSINSKI M. Evaluation of the meaningfulness of health- related quality of life improvements as assessed by the SF-36 and the EQ-5D VAS in patients with active Crohn's disease. Aliment Pharmacol Ther. 2009;29(9):1032-1041. 29. Yarlas A, Maher SM, Bayliss MS, Lovley A, Cappelleri JC, DiBonaventura MD. Psychometric validation of the work productivity and activity impairment questionnaire in ulcerative colitis: results from a systematic literature review. Journal of Patient-Reported Outcomes. 2018;2(1):62. 30. Reilly MC, Zbrozek AS, Dukes EM. The Validity and Reproducibility of a Work Productivity and Activity Impairment Instrument. PharmacoEconomics. 1993;4(5):353-365. 31. Reilly MC, Gerlier L, Brabant Y, Brown M. Validity, reliability, and responsiveness of the work productivity and activity impairment questionnaire in Crohn's disease. Clinical Therapeutics. 2008;30(2):393-404. 32. Rabin R, de Charro F. EQ-5D: a measure of health status from the EuroQol Group. Annals of medicine. 2001;33(5):337-343. 33. Administration FAD. Biosimilar and Interchangeable Products. 2017; https://www.fda.gov/drugs/biosimilars/biosimilar-and-interchangeable-products#biosimilar. 34. Nguyen GC, Boland K, Afif W, et al. Modified Delphi Process for the Development of Choosing Wisely for Inflammatory Bowel Disease. Inflamm Bowel Dis. 2017;23(6):858-865. 35. Heather Cruickshank. The Cost of Living with Ulcerative Colitis: Meg's Story. Break Free from UC. 2019. https://www.healthline.com/health/ulcerative-colitis/cost-of-care-megan-wells#1. Accessed 09/01/20. 36. Heather Cruickshank. The Cost of Living with Ulcerative Colitis: Nyannah's Story. Break Free from UC. 2019. https://www.healthline.com/health/ulcerative-colitis/cost-of-care-nyannah#1. Accessed 09/01/20. 37. Chung AE, Vu MB, Myers K, Burris J, Kappelman MD. Crohn's and Colitis Foundation of America Partners Patient-Powered Research Network: Patient Perspectives on Facilitators and Barriers to Building an Impactful Patient-Powered Research Network. Med Care. 2018;56 Suppl 10 Suppl 1(10 Suppl 1):S33-s40. 38. Sweeney L, Moss-Morris R, Czuber-Dochan W, Murrells T, Norton C. Developing a better biopsychosocial understanding of pain in inflammatory bowel disease: a cross-sectional study. European journal of gastroenterology & hepatology. 2020;32(3):335-344. 39. Centers for Medicare & Medicaid Services. National Coverage Determination (NCD) for Fecal Occult Blood Test (190.34). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2003. ©Institute for Clinical and Economic Review, 2020 Page 125 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 40. Centers for Medicare & Medicaid Services. National Coverage Determination (NCD) for Colorectal Cancer Screening Tests (210.3). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2014. 41. Centers for Medicare & Medicaid Services. Local Coverage Determination (LCD): Colonoscopy and Sigmoidoscopy-Diagnostic (L34614). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2015. 42. Centers for Medicare & Medicaid Services. Local Coverage Determination (LCD): Diagnostic and Therapeutic Colonoscopy (L34213). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2015. 43. Centers for Medicare & Medicaid Services. Local Coverage Determination (LCD): MolDX: Prometheus IBD sgi Diagnostic Policy (L37299). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2018. 44. Centers for Medicare & Medicaid Services. Local Coverage Determination (LCD): MolDX: Prometheus IBD sgi Diagnostic Policy (L37539). In: Services USDoHaH, ed. www.cms.gov: Centers for Medicare & Medicaid Services; 2018. 45. Tufts Medical Center. Specialty Drug Evidence and Coverage (SPEC) Database. In: Center TM, ed2020. 46. Aetna. Renflexis® (infliximab-abda) Injectable Medication Precertification Request. 2019; http://www.aetna.com/pharmacy-insurance/healthcare-professional/documents/renflexis- precert-form.pdf. Accessed 03/05/20, 2020. 47. United States Food and Drug Administration. FDA approves Boxed Warning about increased risk of blood clots and death with higher dose of arthritis and ulcerative colitis medicine tofacitinib (Xeljanz, Xeljanz XR): FDA Drug Safety Communication. 2019; https://www.fda.gov/drugs/drug- safety-and-availability/fda-approves-boxed-warning-about-increased-risk-blood-clots-and- death-higher-dose-arthritis-and. Accessed 03/05/20, 2020. 48. Aetna. Vedolizumab (Entyvio). Clinical Policy Bulletins 2020; http://www.aetna.com/cpb/medical/data/800_899/0885.html#:~:text=Aetna%20considers%20 continuation%20of%20vedolizumab,and%20symptoms%20of%20the%20condition. Accessed 09/01/20, 2020. 49. UnitedHealthcare. UnitedHealthcare Commerical Medical Benefit Drug Policy: Entyvio (Vedolizumab). 2020; https://www.uhcprovider.com/content/dam/provider/docs/public/policies/comm-medical- drug/entyvio-vedolizumab.pdf. Accessed 09/01/20, 2020. 50. Feuerstein JD, Isaacs KL, Schneider Y, et al. AGA Clinical Practice Guidelines on the Management of Moderate to Severe Ulcerative Colitis. Gastroenterology. 2020. 51. Rubin DT, Ananthakrishnan AN, Siegel CA, Sauer BG, Long MD. ACG Clinical Guideline: Ulcerative Colitis in Adults. American Journal of Gastroenterology. 2019;114(3):384-413. 52. Bressler B, Marshall JK, Bernstein CN, et al. Clinical practice guidelines for the medical management of nonhospitalized ulcerative colitis: the Toronto consensus. Gastroenterology. 2015;148(5):1035-1058.e1033. 53. Health Canada. Regulatory Decision Summary - Xeljanz - Health Canada. 2018; https://hpr- rps.hres.ca/reg-content/regulatory-decision-summary-detail.php?lang=en&linkID=RDS00449. Accessed 02/24/20, 2020. 54. NICE. Ulcerative colitis: management NICE guideline [NG130]. NICE Guidance [Guideline]. 2019; https://www.nice.org.uk/guidance/ng130. Accessed 02/24/20, 2020. ©Institute for Clinical and Economic Review, 2020 Page 126 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 55. NICE. Infliximab, adalimumab and golimumab for treating moderately to severely active ulcerative colitis after the failure of conventional therapy Technology appraisal guidance [TA329]. 2015; https://www.nice.org.uk/guidance/ta329. Accessed 02/24/20, 2020. 56. NICE. Vedolizumab (Entyvio®) for the Treatment of Adults with Moderate to Severe Active Ulcerative Colitis [STA342]. 2014; https://www.nice.org.uk/guidance/ta342/resources/ulcerative-colitis-moderate-to-severely- active-vedolizumab-id691-evaluation-report-part-22. Accessed December 19, 2019. 57. NICE. Tofacitinib for moderately to severely active ulcerative colitis Technology appraisal guidance [TA547]. 2018; https://www.nice.org.uk/guidance/ta547. Accessed 02/24/20, 2020. 58. NICE. Appraisal consultation document: Ustekinumab for treating moderately to severely active ulcerative colitis. In: Department of Health and Social Care, ed. nice.org.uk: NICE; 2020:1-18. 59. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Ann Intern Med. 1997;126(5):376-380. 60. Higgins JPT GSe. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. Available from http://handbook.cochrane.org.; 2011. 61. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8(5):336-341. 62. Panaccione R, Ghosh S, Middleton S, et al. Combination therapy with infliximab and azathioprine is superior to monotherapy with either agent in ulcerative colitis. Gastroenterology. 2014;146(2):392-400.e393. 63. Agency for Healthcare Research and Quality. U.S. Preventive Services Task Force Procedure Manual. 2008. 64. Ollendorf DA, Pearson SD. An integrated evidence rating to frame comparative effectiveness assessments for decision makers. Med Care. 2010;48(6 Suppl):S145-152. 65. Food and Drug Administration. FDA approves Boxed Warning about increased risk of blood clots and death with higher dose of arthritis and ulcerative colitis medicine tofacitinib (Xeljanz, Xeljanz XR). 2019. 66. van Valkenhoef G, Kuiper J. gemtc: Network Meta-Analysis Using Bayesian Methods. R package version 0.8-2. 2016. 67. Yu-sung S, Masanao Y. R2jags: Using R to Run 'JAGS'. R package version 0.5-7. 2015; https://CRAN.R-project.org/package=R2jags. 68. ClinicalTrials.gov. A Study to Evaluate the Effectiveness and Safety of Infliximab in Chinese Patients With Active Ulcerative Colitis [NCT01551290]. 2014. 69. Colombel J-F, Sandborn WJ, Ghosh S, et al. Four-year maintenance treatment with adalimumab in patients with moderately to severely active ulcerative colitis: Data from ULTRA 1, 2, and 3. The American journal of gastroenterology. 2014;109(11):1771-1780. 70. Colombel J, Reinisch W, Gibson P, et al. Delayed Response to Golimumab Therapy: UC Patient Characteristics and Long-term Clinical Outcome-Post-Hoc Analyses from the PURSUIT Programme. Paper presented at: 11th Congress of the European Crohn's and Colitis Organisation2016; Amsterdam, Netherlands. 71. Danese SS, BE.; O'Brien, CD. Efficacy and Safety of Ustekinumab Through Week 16 in Patients with Moderate-To-Severe Ulcerative Colitis Randomised to Ustekinumab: Results from the UNIFI Induction Trial. J Crohns Colitis. 2019;13:S061–S062. ©Institute for Clinical and Economic Review, 2020 Page 127 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 72. Dubinsky MC, Peyrin-Biroulet L, Melmed GY, et al. Efficacy of Tofacitinib in Patients With Ulcerative Colitis by Prior Tumor Necrosis Factor Inhibitor Treatment Status: Results From OCTAVE Induction and Maintenance Studies: 640. American Journal of Gastroenterology. 2017;112:S354. 73. Feagan B, Sandborn W, Lazar A. Adalimumab Therapy Is Associated With Reduced Risk of Hospitalization in Patients With Ulcerative Colitis. Gastroenterology. 2014;146(1). 74. Feagan BG, Patel H, Colombel JF, et al. Effects of vedolizumab on health-related quality of life in patients with ulcerative colitis: results from the randomised GEMINI 1 trial. Aliment Pharmacol Ther. 2017;45(2):264-275. 75. Feagan BG, Reinisch W, Rutgeerts P, et al. The effects of infliximab therapy on health-related quality of life in ulcerative colitis patients. The American journal of gastroenterology. 2007;102(4):794-802. 76. Feagan BG, Rubin DT, Danese S, et al. Efficacy of Vedolizumab Induction and Maintenance Therapy in Patients With Ulcerative Colitis, Regardless of Prior Exposure to Tumor Necrosis Factor Antagonists. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association. 2017;15(2):229-239.e225. 77. Feagan BG, Rutgeerts P, Sands BE, et al. Vedolizumab as induction and maintenance therapy for ulcerative colitis. N Engl J Med. 2013;369(8):699-710. 78. Han C, Naessens D, Zhang H, Johanns J, Marano C, Sands BE. Ustekinumab Improved Work Productivity in Patients With Moderate to Severe Ulcerative Colitis: Results From the Phase 3 UNIFI Induction and Maintenance Studies: 748. American Journal of Gastroenterology. 2019;114:S438. 79. Hibi T, Imai Y, Senoo A, Ohta K, Ukyo Y. Efficacy and safety of golimumab 52-week maintenance therapy in Japanese patients with moderate to severely active ulcerative colitis: a phase 3, double-blind, randomized, placebo-controlled study-(PURSUIT-J study). Journal of Gastroenterology. 2017;52(10):1101-1111. 80. Jiang X-L, Cui H-F, Gao J, Fan H. Low-dose Infliximab for Induction and Maintenance Treatment in Chinese Patients With Moderate to Severe Active Ulcerative Colitis. J Clin Gastroenterol. 2015;49(7):582-588. 81. Kobayashi T, Suzuki Y, Motoya S, et al. First trough level of infliximab at week 2 predicts future outcomes of induction therapy in ulcerative colitis-results from a multicenter prospective randomized controlled trial and its post hoc analysis. Journal of gastroenterology. 2016;51(3):241-251. 82. Loftus EV, Jr., Colombel J-F, Feagan BG, et al. Long-term Efficacy of Vedolizumab for Ulcerative Colitis. Journal of Crohn's & colitis. 2017;11(4):400-411. 83. Panés J, Vermeire S, Lindsay JO, et al. Tofacitinib in Patients with Ulcerative Colitis: Health- Related Quality of Life in Phase 3 Randomised Controlled Induction and Maintenance Studies. Journal of Crohn's & colitis. 2018;12(2):145-156. 84. Reinisch W, Gibson PR, Sandborn WJ, et al. Long-Term Benefit of Golimumab for Patients with Moderately to Severely Active Ulcerative Colitis: Results from the PURSUIT-Maintenance Extension. Journal of Crohn's & colitis. 2018;12(9):1053-1066. 85. Reinisch W, Osterman MT, Doherty G, et al. Tu1720 – Efficacy of Tofacitinib Maintenance Therapy for Ulcerative Colitis in Remitting Patients Vs Patients with Clinical Response After 8 Weeks of Induction Treatment. Gastroenterology. 2019;156(6, Supplement 1):S-1098. 86. Reinisch W, Sandborn WJ, Hommes DW, et al. Adalimumab for induction of clinical remission in moderately to severely active ulcerative colitis: results of a randomised controlled trial. Gut. 2011;60(6):780-787. ©Institute for Clinical and Economic Review, 2020 Page 128 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 87. Reinisch W, Sandborn WJ, Panaccione R, et al. 52-week efficacy of adalimumab in patients with moderately to severely active ulcerative colitis who failed corticosteroids and/or immunosuppressants. Inflammatory bowel diseases. 2013;19(8):1700-1709. 88. Reinisch W, Sandborn WJ, Rutgeerts P, et al. Long-term infliximab maintenance therapy for ulcerative colitis: the ACT-1 and -2 extension studies. Inflamm Bowel Dis. 2012;18(2):201-211. 89. Rubin DTD, M; Lukas M. Long-term efficacy of tofacitinib in patients who received extended induction therapy: results of the OCTAVE open study for tofacitinib delayed responders. J Crohns Colitis. 2019;13:S050–S052. 90. Rutgeerts P, Sandborn WJ, Feagan BG, et al. Infliximab for induction and maintenance therapy for ulcerative colitis. The New England journal of medicine. 2005;353(23):2462-2476. 91. Sandborn W, Rutgeerts P, Feagan B. Colectomy rate comparison after treatment of ulcerative colitis with placebo or infliximab. Gastroenterology. 2009;137(4). 92. Sandborn WJ, Baert F, Danese S, et al. Efficacy and Safety of Vedolizumab Subcutaneous Formulation in a Randomized Trial of Patients With Ulcerative Colitis. Gastroenterology. 2020;158(3):562-572.e512. 93. Sandborn WJ, van Assche G, Reinisch W, et al. Adalimumab induces and maintains clinical remission in patients with moderate-to-severe ulcerative colitis. Gastroenterology. 2012;142(2):257-265 e251-253. 94. Sandborn WJ, Feagan BG, Marano C, et al. Subcutaneous golimumab maintains clinical response in patients with moderate-to-severe ulcerative colitis. Gastroenterology. 2014;146(1):96-109 e101. 95. Sandborn WJ, Feagan BG, Marano C, et al. Subcutaneous golimumab induces clinical response and remission in patients with moderate-to-severe ulcerative colitis. Gastroenterology. 2014;146(1):85-e15. 96. Sandborn WJ, Su C, Sands BE, et al. Tofacitinib as Induction and Maintenance Therapy for Ulcerative Colitis. N Engl J Med. 2017;376(18):1723-1736. 97. Sands BE, Han C, Zhang H, et al. P681 Ustekinumab therapy induced clinically meaningful improvement and remission as measured by the Inflammatory Bowel Disease Questionnaire: Results from the phase 3 UNIFI induction and maintenance studies. Journal of Crohn's and Colitis. 2019;13(Supplement_1):S460-S460. 98. Sands BE, Peyrin-Biroulet L, Marano C, et al. P312 Efficacy in biologic failure and non-biologic- failure populations in a Phase 3 study of ustekinumab in moderate–severe ulcerative colitis: UNIFI. Journal of Crohn's and Colitis. 2019;13(Supplement_1):S256-S257. 99. Sands BE, Sandborn WJ, Panaccione R, et al. Ustekinumab as Induction and Maintenance Therapy for Ulcerative Colitis. New England Journal of Medicine. 2019;381(13):1201-1214. 100. Suzuki Y, Motoya S, Hanai H, et al. Four-year maintenance treatment with adalimumab in Japanese patients with moderately to severely active ulcerative colitis. Journal of Gastroenterology. 2017;52(9):1031-1040. 101. Suzuki Y, Motoya S, Hanai H, et al. Efficacy and safety of adalimumab in Japanese patients with moderately to severely active ulcerative colitis. Journal of Gastroenterology. 2014;49(2):283- 294. 102. Sandborn WJ, Wolf D, van Assche G, et al. Rapid onset of adalimumab and long-term efficacy among week-8 responders in adults with moderate to severe active Ulcerative Colitis. Inflammatory Bowel Disease. 2011;17(Suppl 2):S4. 103. Lichtenstein G. R. C, M. A.,Rogler, G.,Quirk, D.,Thorpe, A. J.,Pedersen, R. D.,Lawendy, N.,Chan, G.,Panés, J. Tofacitinib for the treatment of ulcerative colitis: An update on the analysis of ©Institute for Clinical and Economic Review, 2020 Page 129 Final Report – Targeted Immune Modulators for UC Return to Table of Contents malignancy rates from the UC clinical program. American Journal of Gastroenterology. 2019;114:S414-S415. 104. Sandborn WJ PJ, Sands BE, et al. . Venous thromboembolic events in the tofacitinib ulcerative colitis clinical development programme. Aliment Pharmacol Ther. 2019;50(10):1068-1076. 105. Abreu MT, Sands, B. E.,Leong, R. W.,Marano, C. W.,O'Brien, C. D.,Zhang, H.,Zhou, Y.,Johanns, J.,Rowbotham, D.,Hisamatsu, T.,Arasaradnam, R. P.,Scherl, E. J.,Danese, S.,Peyrin-Biroulet, L.. . Efficacy and safety of long-term treatment with ustekinumab in moderate-severe ulcerative colitis patients with delayed response to ustekinumab induction: results from the UNIFI 2-year long-term extension. Gastroenterology; 2020. 106. Han CN, Dominik; Zhang, Hongyan; Johanns, Jewel ; Marano, Colleen; Sands, Bruce E. . Ustekinumab Improved Work Productivity in Patients With Moderate to Severe Ulcerative Colitis: Results From the Phase 3 UNIFI Induction and Maintenance Studies. The American Journal of Gastroenterology. 2019;114(P):S438. 107. Hyams JS, Dubinsky MC, Baldassano RN, et al. Infliximab Is Not Associated With Increased Risk of Malignancy or Hemophagocytic Lymphohistiocytosis in Pediatric Patients With Inflammatory Bowel Disease. Gastroenterology. 2017;152(8):1901-1914 e1903. 108. Bolia R, Rajanayagam J, Hardikar W, Alex G. Impact of Changing Treatment Strategies on Outcomes in Pediatric Ulcerative Colitis. Inflammatory Bowel Diseases. 2019;25(11):1838-1844. 109. Dubinsky MC, Cross RK, Sandborn WJ, et al. Extraintestinal Manifestations in Vedolizumab and Anti-TNF-Treated Patients With Inflammatory Bowel Disease. Inflammatory bowel diseases. 2018;24(9):1876-1882. 110. Long MD, Smith TW, Dibonaventura M, et al. Real-world Effectiveness of Advanced Therapies Among Patients With Moderate to Severe Ulcerative Colitis in the United States. Inflammatory bowel diseases. 2019:izz204. 111. Meyer A, Rudant J, Drouin J, Coste J, Carbonnel F, Weill A. The effectiveness and safety of infliximab compared with biosimilar CT-P13, in 3112 patients with ulcerative colitis. Aliment Pharmacol Ther. 2019;50(3):269-277. 112. Panes J, Lindsay JO, Teich N, et al. Five-year Safety Data From OPUS, a European Observational Safety Registry for Adults With Ulcerative Colitis Treated With Originator Infliximab [Remicade(R)] or Conventional Therapy. J Crohns Colitis. 2019;13(9):1148-1157. 113. Sandborn WJ, Sakuraba A, Wang A, et al. Comparison of real-world outcomes of adalimumab and infliximab for patients with ulcerative colitis in the United States. Current medical research and opinion. 2016;32(7):1233-1241. 114. Singh S, Andersen NN, Andersson M, Loftus EV, Jr., Jess T. Comparison of Infliximab and Adalimumab in Biologic-Naive Patients With Ulcerative Colitis: A Nationwide Danish Cohort Study. Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association. 2017;15(8):1218-1225.e1217. 115. Singh S, Heien HC, Sangaralingham LR, et al. Comparative effectiveness and safety of infliximab and adalimumab in patients with ulcerative colitis. Aliment Pharmacol Ther. 2016;43(9):994- 1003. 116. Han M, Jung YS, Cheon JH, Park S. Comparison of Real-World Outcomes of Infliximab versus Adalimumab in Biologic-Naïve Korean Patients with Ulcerative Colitis: A Population-Based Study. Yonsei medical journal. 2020;61(1):48-55. 117. Yarur A MG, Silverberg M, et al. Real-world effectiveness and safety of vedolizumab and anti- TNF in biologic-naive ulcerative colitis patients: Results from the EVOLVE study. Journal of Crohn's and Colitis. 2019;13(Supplement 1):P573. ©Institute for Clinical and Economic Review, 2020 Page 130 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 118. Rubin DT, Dubinsky MC, Lukas M, et al. DOP43 Long-term efficacy of tofacitinib in patients who received extended induction therapy: results of the OCTAVE open study for tofacitinib delayed responders. Journal of Crohn's and Colitis. 2019;13(Supplement_1):S050-S052. 119. Sandborn WJ, Ghosh S, Panes J, et al. Tofacitinib, an Oral Janus Kinase Inhibitor, in Active Ulcerative Colitis. New England Journal of Medicine. 2012;367(7):616-624. 120. Ananthakrishnan AN, Donaldson T, Lasch K, Yajnik V. Management of Inflammatory Bowel Disease in the Elderly Patient: Challenges and Opportunities. Inflamm Bowel Dis. 2017;23(6):882-893. 121. Charpentier C, Salleron J, Savoye G, et al. Natural history of elderly-onset inflammatory bowel disease: a population-based cohort study. Gut. 2014;63(3):423-432. 122. Sturm A, Maaser C, Mendall M, et al. European Crohn's and Colitis Organisation Topical Review on IBD in the Elderly. Journal of Crohn's and Colitis. 2016;11(3):263-273. 123. Yajnik V, Khan N, Dubinsky M, et al. Efficacy and Safety of Vedolizumab in Ulcerative Colitis and Crohn's Disease Patients Stratified by Age. Advances in therapy. 2017;34(2):542-559. 124. David R, Walter, R.,Thomas, G.,Stephen, V.,Gustavo, K. P.,Marcia, P.,Haiyun, F.,Eric, M.,Marc, F.,Nervin, L.,Irene, M.,Gary, L.. . Extraintestinal manifestations at baseline, and effect of tofacitinib in patients with moderate to severe ulcerative colitis in the octave program. American Journal of Gastroenterology. 2019;114:S21. 125. Vermeire S, Gils A, Accossato P, Lula S, Marren A. Immunogenicity of biologics in inflammatory bowel disease. Therapeutic advances in gastroenterology. 2018;11:1756283x17750355. 126. Clark M, Colombel JF, Feagan BC, et al. American gastroenterological association consensus development conference on the use of biologics in the treatment of inflammatory bowel disease, June 21-23, 2006. Gastroenterology. 2007;133(1):312-339. 127. Pickard AS, Law EH, Jiang R, et al. United States Valuation of EQ-5D-5L Health States Using an International Protocol. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research. 2019;22(8):931-941. 128. Samuel S, Ingle SB, Dhillon S, et al. Cumulative incidence and risk factors for hospitalization and surgery in a population-based cohort of ulcerative colitis. Inflamm Bowel Dis. 2013;19(9):1858- 1866. 129. HMD. The Human Mortality Database (HMD). 2020; https://www.mortality.org/. Accessed January 13, 2020. 130. Bewtra M, Kaiser LM, TenHave T, Lewis JD. Crohn's disease and ulcerative colitis are associated with elevated standardized mortality ratios: a meta-analysis. Inflamm Bowel Dis. 2013;19(3):599-613. 131. Eaden JA, Abrams KR, Mayberry JF. The risk of colorectal cancer in ulcerative colitis: a meta- analysis. Gut. 2001;48(4):526-535. 132. Flores BM, O'Connor A, Moss AC. Impact of mucosal inflammation on risk of colorectal neoplasia in patients with ulcerative colitis: a systematic review and meta-analysis. Gastrointest Endosc. 2017;86(6):1006-1011 e1008. 133. Rutter M, Saunders B, Wilkinson K, et al. Severity of inflammation is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology. 2004;126(2):451-459. 134. Biancone L, Armuzzi A, Scribano ML, et al. Cancer Risk in Inflammatory Bowel Disease: A 6-Year Prospective Multicenter Nested Case-Control IG-IBD Study. Inflamm Bowel Dis. 2019. 135. Olen OE, R.; Sachs, M.; Pedersen, L.; Halfvarson, J.; Askling, J. Ekbom, A.; Sorensen, T.; Ludvigsson, JF. Colorectal cancer in ulcerative colitis: a Scandinavian population-based cohort study. Lancet. 2020;395(10218):123-131. ©Institute for Clinical and Economic Review, 2020 Page 131 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 136. SEER. SEER Cancer Statistics Review 1975-2016. Table 6.10 Cancer of the Colon and Rectum (Invasive). SEER Incidence(a) and U.S. Death(b) Rates, Age-Adjusted and Age-Specific Rates, by Race and Sex. 2019; https://seer.cancer.gov/csr/1975_2016/results_merged/sect_06_colon_rectum.pdf. Accessed November 20, 2019. 137. Lewis JD, Scott FI, Brensinger CM, et al. Increased Mortality Rates With Prolonged Corticosteroid Therapy When Compared With Antitumor Necrosis Factor-alpha-Directed Therapy for Inflammatory Bowel Disease. Am J Gastroenterol. 2018;113(3):405-417. 138. Barbieri M, Wilmoth JR, Shkolnikov VM, et al. Data Resource Profile: The Human Mortality Database (HMD). International journal of epidemiology. 2015;44(5):1549-1556. 139. Causey MW, Stoddard D, Johnson EK, et al. Laparoscopy impacts outcomes favorably following colectomy for ulcerative colitis: a critical analysis of the ACS-NSQIP database. Surg Endosc. 2013;27(2):603-609. 140. Woehl AH, A.; McEwan, P. The relation between disease activity, quality of life and health utility in patients with ulcerative colitis. Gut. 2008;57:A153. 141. Archer R, Tappenden P, Ren S, et al. Infliximab, adalimumab and golimumab for treating moderately to severely active ulcerative colitis after the failure of conventional therapy (including a review of TA140 and TA262): clinical effectiveness systematic review and economic model. Health Technol Assess. 2016;20(39):1-326. 142. Travis S, Feagan BG, Peyrin-Biroulet L, et al. Effect of Adalimumab on Clinical Outcomes and Health-related Quality of Life Among Patients With Ulcerative Colitis in a Clinical Practice Setting: Results From InspirADA. J Crohns Colitis. 2017;11(11):1317-1325. 143. Probert CS, Sebastian S, Gaya DR, et al. Golimumab induction and maintenance for moderate to severe ulcerative colitis: results from GO-COLITIS (Golimumab: a Phase 4, UK, open label, single arm study on its utilization and impact in ulcerative Colitis). BMJ Open Gastroenterol. 2018;5(1):e000212. 144. Kosmas CK, C; Marsh, SE; Gentile, B; Doll, HA; Finney-Hayward, T. Health Related Quality of Life (Hrqol) In Patients Who Have Undergone Colectomy for Ulcerative Colitis: Impacts of Complications Post-Surgery. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research. 2015;18(7):A335–A766. 145. Lemaitre M, Kirchgesner J, Rudnichi A, et al. Association Between Use of Thiopurines or Tumor Necrosis Factor Antagonists Alone or in Combination and Risk of Lymphoma in Patients With Inflammatory Bowel Disease. Jama. 2017;318(17):1679-1686. 146. Bonovas S, Fiorino G, Allocca M, et al. Biologic Therapies and Risk of Infection and Malignancy in Patients With Inflammatory Bowel Disease: A Systematic Review and Network Meta-analysis. Clin Gastroenterol Hepatol. 2016;14(10):1385-1397 e1310. 147. Gibson PR, Feagan BG, Sandborn WJ, et al. Maintenance of Efficacy and Continuing Safety of Golimumab for Active Ulcerative Colitis: PURSUIT-SC Maintenance Study Extension Through 1 Year. Clin Transl Gastroenterol. 2016;7:e168. 148. Zogg CK, Najjar P, Diaz AJ, et al. Rethinking Priorities: Cost of Complications After Elective Colectomy. Ann Surg. 2016;264(2):312-322. 149. Arseneau KO, Sultan S, Provenzale DT, et al. Do patient preferences influence decisions on treatment for patients with steroid-refractory ulcerative colitis? Clin Gastroenterol Hepatol. 2006;4(9):1135-1142. 150. MacDougall DG, J.; Ehrenberg R.; et al. Greater than Expected Dosing (GTED) Assessment among Targeted Immunomodulators in Management of Inflammatory Bowel Disease (IBD). AMCP ©Institute for Clinical and Economic Review, 2020 Page 132 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Managed Care & Specialty Pharmacy Annual Meeting; March 25 – 28, 2019, 2019; San Diego, California. 151. RedBook. 2020. Accessed May 6, 2020. 152. CMS. Physician Fee Schedule. 2020; https://www.cms.gov/apps/physician-fee- schedule/overview.aspx. Accessed January 10, 2020. 153. REMICADE. REMICADE (infliximab) Prescribing Information. 2018; http://www.janssenlabels.com/package-insert/product-monograph/prescribing- information/REMICADE-pi.pdf. Accessed January 13, 2020. 154. Xu F, Liu Y, Wheaton AG, Rabarison KM, Croft JB. Trends and Factors Associated with Hospitalization Costs for Inflammatory Bowel Disease in the United States. Appl Health Econ Health Policy. 2019;17(1):77-91. 155. PCE. U.S. Bureau of Economic Analysis, Personal Consumption Expenditures: Chain-type Price Index [PCEPI], retrieved from FRED, Federal Reserve Bank of St. Louis. 2019; https://fred.stlouisfed.org/series/PCEPI. 156. PHC. Table 23: National Health Expenditures; Nominal Dollars, Real Dollars, Price Indexes, and Annual Percent Change: Selected Calendar Years 1980-2017. 2020; https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and- Reports/NationalHealthExpendData/NationalHealthAccountsHistorical.html. Accessed January 13, 2020. 157. Cohen R, Skup M, Ozbay AB, et al. Direct and indirect healthcare resource utilization and costs associated with ulcerative colitis in a privately-insured employed population in the US. J Med Econ. 2015;18(6):447-456. 158. Milev S, DiBonaventura MD, Quon P, et al. An economic evaluation of tofacitinib for the treatment of moderately-to-severely active ulcerative colitis: modeling the cost of treatment strategies in the United States. J Med Econ. 2019;22(9):859-868. 159. AHRQ. Healthcare Cost and Utilization Project (HCUPnet). 2019; https://hcupnet.ahrq.gov. Accessed December 19, 2019. 160. Park KT, Tsai R, Perez F, Cipriano LE, Bass D, Garber AM. Cost-effectiveness of early colectomy with ileal pouch-anal anastamosis versus standard medical therapy in severe ulcerative colitis. Ann Surg. 2012;256(1):117-124. 161. BLS. Table B-3. Average hourly and weekly earnings of all employees on private nonfarm payrolls by industry sector, seasonally adjusted. 2019; https://www.bls.gov/news.release/empsit.t19.htm. Accessed October, 2019. 162. OECD. Average usual weekly hours worked on the main job. 2019; https://stats.oecd.org/index.aspx?DataSetCode=ANHRS. Accessed October, 2019. 163. Ding ZO, C.; Izanec, J.; Voelker J.; Lucas, J.; Kemp, J.; Muser, E. Work Productivity Loss and Associated Indirect Costs for Patients With Ulcerative Colitis in the US. AMCP Nexus; October 29-November 1, 2019, 2019; National Harbor, MD. 164. Ray KN, Chari AV, Engberg J, Bertolet M, Mehrotra A. Opportunity costs of ambulatory medical care in the United States. The American journal of managed care. 2015;21(8):567-574. 165. Pucheril D, Fletcher SA, Chen X, et al. Workplace absenteeism amongst patients undergoing open vs. robotic radical prostatectomy, hysterectomy, and partial colectomy. Surg Endosc. 2020. 166. Scott FI, Shah Y, Lasch K, Luo M, Lewis JD. Assessing the Optimal Position for Vedolizumab in the Treatment of Ulcerative Colitis: A Simulation Model. Inflamm Bowel Dis. 2018;24(2):286-295. 167. Beilman CL FR, Halloran BP. . Cost-effectivenesss of vedolizumab campared to infliximab for the management of moderate-to-severe ulcerative colitis. 2018. ©Institute for Clinical and Economic Review, 2020 Page 133 Final Report – Targeted Immune Modulators for UC Return to Table of Contents 168. Tappenden P, Ren S, Archer R, et al. A Model-Based Economic Evaluation of Biologic and Non- Biologic Options for the Treatment of Adults with Moderately-to-Severely Active Ulcerative Colitis after the Failure of Conventional Therapy. Pharmacoeconomics. 2016;34(10):1023-1038. 169. Wu B, Wang Z, Zhang Q. Cost-Effectiveness of Different Strategies for the Treatment of Moderate-to-Severe Ulcerative Colitis. Inflamm Bowel Dis. 2018;24(11):2291-2302. 170. Lohan C, Diamantopoulos A, LeReun C, Wright E, Bohm N, Sawyer LM. Tofacitinib for the treatment of moderately to severely active ulcerative colitis: a systematic review, network meta-analysis and economic evaluation. BMJ Open Gastroenterology. 2019;6(1):e000302. 171. Pearson SD. The ICER Value Framework: Integrating Cost Effectiveness and Affordability in the Assessment of Health Care Value. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research. 2018;21(3):258-265. 172. Silverman E. What doubts? New forecast says biosimilars will cut U.S. drug spending by $100 billion in five years. Pharmalot. 2020. https://www.statnews.com/pharmalot/2020/10/06/biosimilars-biologics-drug-prices-iqvia/. Accessed 10/13/20. 173. Welty M, Mesana L, Padhiar A, et al. Efficacy of ustekinumab vs. advanced therapies for the treatment of moderately to severely active ulcerative colitis: a systematic review and network meta-analysis. Curr Med Res Opin. 2020:1-12. 174. Ollendorf D, Pearson S. ICER Evidence Rating Matrix: A User's Guide. 2020; https://icer- review.org/methodology/icers-methods/icer-evidence-rating-matrix/. 175. Sanders GD, Neumann PJ, Basu A, et al. Recommendations for Conduct, Methodological Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on Cost-Effectiveness in Health and Medicine. Jama. 2016;316(10):1093-1103. 176. Dias S, Welton NJ, Sutton AJ, AE A. NICE DSU Technical Support Document 2: A Generalised Linear Modelling Framework for Pairwise and Network Meta-Analysis of. Randomised Controlled Trials. 2011 (updated 2014). ©Institute for Clinical and Economic Review, 2020 Page 134 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix ©Institute for Clinical and Economic Review, 2020 Page 135 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix A. Search Strategies and Results Table A1. PRISMA 2009 Checklist # Checklist Item TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. ABSTRACT Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, Structured Summary 2 participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, Objectives 4 outcomes, and study design (PICOS). METHODS Protocol and Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide 5 Registration registration information including registration number. Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, Eligibility Criteria 6 publication status) used as criteria for eligibility, giving rationale. Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional Information Sources 7 studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated. State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included Study Selection 9 in the meta-analysis). Data Collection Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for 10 Process obtaining and confirming data from investigators. List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and Data Items 11 simplifications made. Risk of Bias in Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at 12 Individual Studies the study or outcome level), and how this information is to be used in any data synthesis. Summary Measures 13 State the principal summary measures (e.g., risk ratio, difference in means). Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., Synthesis of Results 14 I2) for each meta-analysis. ©Institute for Clinical and Economic Review, 2020 Page 136 Final Report – Targeted Immune Modulators for UC Return to Table of Contents # Checklist Item Risk of Bias across Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting 15 Studies within studies). Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which Additional Analyses 16 were pre-specified. RESULTS Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each Study Selection 17 stage, ideally with a flow diagram. For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide Study Characteristics 18 the citations. Risk of Bias within 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Studies Results of Individual For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention 20 Studies group (b) effect estimates and confidence intervals, ideally with a forest plot. Synthesis of Results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. Risk of Bias across 22 Present results of any assessment of risk of bias across studies (see Item 15). Studies Additional Analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). DISCUSSION Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key Summary of Evidence 24 groups (e.g., health care providers, users, and policy makers). Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified Limitations 25 research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. FUNDING Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the Funding 27 systematic review. From: Moher D, Liberati A, Tetzlaff J, Altman DG. The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097 ©Institute for Clinical and Economic Review, 2020 Page 137 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table A2. Search Strategies for Medline (via Ovid) No. Query 1 colitis, ulcerative 2 ((ulcera* adj3 colitis) or inflammatory bowel disease* or IBD or UC).mp 3 (Infliximab or Infliximab-abda or Renflexis or Infliximab-dyyb or Inflectra or Remicade or CT P13).mp. 4 infliximab.af. 5 (Humira or Adalimumab ABTD2E7 or ABT D2E7).mp. 6 adalimumab.af. 7 (Entyvio or MLN0002 or Vedolizumab).mp. 8 vedolizumab.af. 9 (golimumab or simponi or CNTO 148).mp. 10 golimumab.af. 11 ustekinumab.af. 12 (ustekinumab or stelara or CNTO1275 or CNTO 1275).mp. 13 (tofacitinib or tofacitinib citrate or Xeljanz or CP 690?550).mp. 14 tofacitinib.af. (abstract or addresses or autobiography or bibliography or biography or clinical trial, phase I or case report or comment or congresses or consensus development conference or duplicate publication or 15 editorial or guideline or in vitro or interview or lecture or legal cases or legislation or letter or news or newspaper article or patient education handout or periodical index or personal narratives or portraits or practice guideline or review or videoaudio media).pt. 16 (animals not (humans and animals)).sh. exp cohort studies/ or comparative study.pt. or observational study.pt. or exp case-control studies/ or 17 cohort.tw. or (observational adj2 stud*).tw or prospective.tw or retrospective.tw or longitudinal.tw. or compa*.tw OR groups.tw OR case control.tw OR multivariate.tw control Groups/ or (control* adj2 (clinical or group* or trial* or study or studies or design* or arm*)).ti,ab. or ("clinical trial" or "clinical trial, phase ii" or "clinical trial, phase iii" or "clinical trial, phase 18 iv" or "controlled clinical trial" or "multicenter study" or "randomized controlled trial").pt. or (randomi?ed adj6 (study or trial* or (clinical adj2 trial*))).ti,ab. or ((single or doubl*) adj2 blind*).ti,ab. 19 1 or 2 20 or/3-14 21 19 and 20 22 21 not 15 23 22 not 16 24 17 or 18 25 23 and 24 26 limit 25 to english language 27 remove duplicates from 26 Date of Search: November 20, 2019; updated July 17, 2020. ©Institute for Clinical and Economic Review, 2020 Page 138 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table A3. Search Strategies for Embase No. Query #1 'ulcerative colitis'/exp #2 ((ulcera* NEAR/3 colitis):ab,ti) OR 'inflammatory bowel disease*':ab,ti OR uc:ti,ab OR ibd:ti,ab #3 #1 OR #2 'infliximab'/exp OR infliximab:ab,ti OR 'remicade':ab,ti OR 'renflexis':ab,ti OR 'inflectra':ab,ti OR #4 'infliximab-adba' OR 'infliximab-dyyb':ab,ti OR 'ct p13':ab,ti 'tofacitinib'/exp OR tofacitinib:ab,ti OR tasocitinib:ab,ti OR 'tofacitinib citrate':ab,ti OR xeljanz:ab,ti OR 'cp #5 690 550':ab,ti OR 'cp 690550':ab,ti #6 'adalimumab'/exp OR adalimumab:ab,ti OR humira:ab,ti OR abtd2e7:ab,ti OR 'abt d2e7':ab,ti #7 'golimumab'/exp OR 'golimumab':ab,ti OR 'simponi':ab,ti OR 'cnto 148':ab,ti #8 'ustekinumab'/exp OR ustekinumab:ab,ti OR stelara:ab,ti OR cnto1275:ab,ti OR 'cnto 1275':ab,ti #9 'vedolizumab'/exp OR vedolizumab:ab,ti OR entyvio:ab,ti OR mln0002:ab,ti #10 #4 OR #5 OR #6 OR #7 OR #8 OAR #9 #11 #3 AND #10 #11 NOT ('animal experiment'/de OR 'animal model'/de OR 'case report'/de OR 'human cell'/de OR #12 'human tissue'/de OR 'nonhuman'/de OR 'practice guideline'/de OR 'questionnaire'/de OR 'chapter'/it OR 'editorial'/it OR 'letter'/it OR 'note'/it OR 'review'/it OR 'short survey'/it) #13 #12 NOT [medline]/lim #14 #13 NOT (('animal'/exp OR 'nonhuman'/exp OR 'animal experiment'/exp) NOT 'human'/exp) #15 #14 AND [english]/lim clinical article'/exp OR 'controlled study'/exp OR 'major clinical study'/exp OR 'observational study'/exp OR 'prospective study'/exp OR 'retrospective study'/exp OR 'longitudinal study'/exp OR 'cohort #16 analysis'/exp OR 'cohort':ti,ab OR 'compa*':ti,ab OR 'groups':ti,ab OR 'case control':ti,ab OR 'multivariate':ti,ab OR retrospective:ti,ab OR prospective:ti,ab OR longitudinal:ti,ab OR ((observational NEAR/2 stud*):ti,ab) ('clinical':ti,ab AND 'trial':ti,ab) OR 'clinical trial'/exp OR 'randomized controlled trial'/exp OR 'controlled #17 clinical trial'/exp OR random*:ti,ab or control*:ti,ab OR 'control group'/exp OR 'drug therapy':lnk #18 #16 OR #17 #19 #15 AND #18 Date of Search: November 20, 2019; updated July 17, 2020. ©Institute for Clinical and Economic Review, 2020 Page 139 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure A1. PRISMA Flow Chart Showing the Results of the Literature Search for TIMs for UC 6,832 references 4 references identified identified through through other sources literature search 6,662 references after duplicate removal 6,662 references screened 5,740 citations excluded 869 citations excluded 922 references assessed 108 population for eligibility in full text 433 study design 328 outcomes 53 total references relating to 30 studies 19 RCTs in adults 9 comparative observational studies in adults 1 RCT in children 1 observational study in children 17 references included in quantitative synthesis ©Institute for Clinical and Economic Review, 2020 Page 140 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix B. Previous Systematic Reviews and Technology Assessments Canadian Agency for Drugs and Technologies in Health (CADTH) Common Drug Review: Adalimumab (Humira), 2016 Health Canada has approved adalimumab for the treatment of adult patients with moderately-to- severely active UC who have had an inadequate response to conventional therapy including corticosteroids, azathioprine and/or mercaptopurine or who are intolerant to such therapies. CADTH agrees with the labeled indication given by Health Canada but also highlights the mechanism of action as a novel way to improve patient access. Only ULTRA 2 allowed the inclusion of patients with prior TNF inhibitor use, so there is uncertainty about its use in those populations. There is also a lack of head-to-head data between adalimumab and other biologics, so a recommendation has not been given on its place in therapy. CADTH suggests that, according to patient preference, patients who achieve remission may discontinue treatment. CADTH's reimbursement recommendation was informed by a review of the manufacturer's pharmacoeconomic submission. CADTH concluded that the incremental cost-utility ratio for adalimumab plus standard of care (SOC) compared to SOC alone is between $67,000 per QALY and $130,000 per QALY. CADTH notes that the surgery costs may be overestimated, the rates of dose escalation and SOC costs may be underestimated and treatment discontinuation between weeks eight and 104 was not considered. Canadian Agency for Drugs and Technologies in Health (CADTH) Common Drug Review: Golimumab (Simponi), 2013 Health Canada has approved of golimumab 50 mg and 100 mg doses for the use of maintenance in UC, however CADTH states that the clinical benefit of golimumab 50 mg remains unclear as the PURSUIT studies have not reported any significant findings for the lower dose. CADTH's reimbursement recommendation was informed by the review of a manufacturer-provided cost- utility analysis. Based on this analysis, it was concluded that golimumab could lie in the range of $52,000 to $104,000 per QALY for patients with moderate-to-severely active UC. Canadian Agency for Drugs and Technologies in Health (CADTH): Infliximab versus Adalimumab for Patients with Moderate-to-Severe Ulcerative Colitis: Clinical and Cost-Effectiveness, 2008 CADTH collected three systematic reviews on infliximab (Lawson, Gisbert, and Rahimi) as well as five RCTs for infliximab versus placebo and two RCTs versus corticosteroids and an observational study (Peyrin-Biroulet) on adalimumab. Because of the limited evidence available on adalimumab ©Institute for Clinical and Economic Review, 2020 Page 141 Final Report – Targeted Immune Modulators for UC Return to Table of Contents at the time and that no available head-to-head trials comparing infliximab to adalimumab, CADTH was unable to draw conclusions on the efficacy of the two treatments versus each other. Canadian Agency for Drugs and Technologies in Health (CADTH): Common Drug Review Tofacitinib (Xeljanz), 2019 CADTH's review of tofacitinib aligns with Health Canada's approval of tofacitinib for treatment of moderately-to-severely active UC in adult patients with an inadequate response or intolerance to conventional UC therapy or a TNF inhibitor. Because of the high rate of infection, CADTH recommends that patients should be advised to undergo vaccination against herpes zoster infection prior to the start of treatment. No conclusions could be made about tofacitinib's effects on patients' health-related quality of life due to limited data. CADTH's reimbursement recommendation was informed using a manufacturers' cost-utility analysis. The manufacturers' analysis concluded that tofacitinib is associated with incremental cost-utility ratios is $8,897 compared to adalimumab, $145,184 compared to infliximab biosimilar, and $118,387 compared to continued conventional UC treatment for mixed populations. CADTH conducted its own economic analysis for two populations: biologic-naïve and biologic-experienced patients. Price reductions of 44% for biologic-experienced populations and 74% for biologic-naïve populations would be needed for the optimal willingness-to-pay threshold of $50,000 per QALY. Canadian Agency for Drugs and Technologies in Health (CADTH): Common Drug Review: Ustekinumab (Stelara). Expected Publication Date TBD. CADTH will be assessing ustekinumab (Stelara) for the treatment of adult patients with moderately- to-severely active UC. Ustekinumab's subcutaneous formulation was recently approved by Health Canada and is indicated for patients with moderately-to-severely active UC. Canadian Agency for Drugs and Technologies in Health (CADTH): Common Drug Review: Vedolizumab (Entyvio). Expected Publication Date TBD. CADTH will be assessing the subcutaneous formulation of vedolizumab for the treatment of UC in adult patients with moderately-to-severely active UC who have an inadequate response, loss of response to, or were intolerant to conventional therapy or infliximab. The IV formulation has already been approved by Health Canada in the same population. CADTH has also previously assessed vedolizumab's IV formulation and researchers noted a significant discontinuation rate amongst patients but said the safety profile did not reveal any significant safety concerns. CADTH also conducted an economic assessment using data provided by the manufacturer and found that the incremental cost-effectiveness ratio for vedolizumab IV compared to conventional therapy is between $60,000 to $150,000 per QALY range. ©Institute for Clinical and Economic Review, 2020 Page 142 Final Report – Targeted Immune Modulators for UC Return to Table of Contents NICE: Tofacitinib for Moderately-to-Severely Active Ulcerative Colitis (TA547), 2018 NICE recommends tofacitinib, within the marketing authorization, for treating moderately-to- severely active UC in adults with an intolerance, inadequate response, or loss of response to either conventional therapy or a biological agent only if the company provides the discount for tofacitinib as agreed upon in the commercial arrangement. Evidence from the clinical trial demonstrates tofacitinib is more effective than placebo. Indirect comparisons show tofacitinib is more effective than adalimumab and golimumab as a maintenance therapy in those who are biologic-naïve. For biologic-experienced patients, tofacitinib is more effective than adalimumab in the induction phase. A cost-effectiveness analysis was conducted by the manufacturer and reflected tofacitinib as a cost- effective treatment for the indicated population. For biologic-naïve patients, the incremental cost- effectiveness ratio as compared to conventional therapy was £8,564 per QALY gained and for biologic-experienced patients, the incremental cost-effectiveness ratio was £10,311 per QALY gained. For both groups, tofacitinib produced fewer QALYs than vedolizumab but at a lower cost. NICE: Infliximab, Adalimumab, and Golimumab for Treating Moderately-to-Severely Active Ulcerative Colitis after the Failure of Conventional Therapy (TA329), 2015 NICE recommends adalimumab, golimumab, and infliximab, within their marketing authorizations, for adults with moderately-to-severely active UC whose disease has medical contraindications, intolerance, or inadequate response to conventional therapy (e.g., corticosteroids, mercaptopurine, or azathioprine). NICE recommends golimumab only if the manufacturer provides the 100 mg dose at the same cost as the 50 mg dose (as agreed in the patient access scheme). Deciding among the three treatments is recommended to be done at an individual level between the patient and clinician where advantages and disadvantages can be discussed. If more than one treatment can be of use, the least expensive option should be chosen. Further, NICE recommends infliximab as an option of treating severely active UC in children between six and 17 years of age who respond inadequately, are intolerant, or have medical contraindications to conventional therapy. The three biologics should be given as a planned course of treatment until failure or until 12 months after starting treatment. Patients should be reassessed every 12 months. An NMA using the placebo-controlled RCTs was conducted to compare the three biologics to each other. Within the RCTs, the TNF inhibitors were clinically effective as compared with placebo. However, due to high uncertainty within the results of the NMA, no conclusion was drawn in relation to the relative effectiveness of the TNF inhibitors. Three sensitivity analyses were conducted and reported infliximab as having the greatest effect on inducing clinical response or remission. The systematic review of cost effectiveness identified three economic evaluations of TNF inhibitors for UC, but none were considered by the assessment group as the assessment group concluded they did not accurately represent the natural history of the disease. The assessment group extrapolated the results of the NMA to inform modeling. ©Institute for Clinical and Economic Review, 2020 Page 143 Final Report – Targeted Immune Modulators for UC Return to Table of Contents NICE: Ustekinumab for Treating Moderately-to-Severely Active Ulcerative Colitis [ID1511]. Expected Publication Date: 13 May 2020 NICE is currently evaluating the clinical and cost effectiveness of ustekinumab for the treatment of moderately-to-severely active UC. Proposed comparators include TNF inhibitors (adalimumab, golimumab, and infliximab), tofacitinib, vedolizumab, and conventional therapies (without biological treatments). Outcomes of interest include mortality, measures of disease activity, rates of and duration of response, relapse, and remission, rates of hospitalization and of surgical intervention, endoscopic healing, mucosal healing, corticosteroid-free remission, adverse effects of treatment, and health-related quality of life. If evidence allows, the following subgroups will be explored: people who have been previously treated with one or more biologics and people who have not received prior biologics. NICE: Vedolizumab for Treating Moderately-to-Severely Active Ulcerative Colitis (TA342), 2015 NICE recommends vedolizumab as an option to treat adults with moderately-to-severely active UC, within its marketing authorization, if the company provides the discount agreed upon in the patient access scheme. NICE recommends vedolizumab to be given until there is a loss of response/remission or surgery is needed and patients should be reassessed after 12 months on treatment. NICE did not conduct their own NMA. NMA data presented by the company was considered but the committee notes that the evidence was not powered to test for treatment effects of vedolizumab between populations (biologic-naïve and biologic-experienced patients) and the data available for effectiveness after TNF-inhibitor failure was limited to one comparison: adalimumab. The company provided deterministic base-case results for its modeled populations. For the overall population, the incremental cost-effectiveness ratio for vedolizumab as compared to conventional therapy was £33,297 per QALY gained. In the biologic-naïve population, the incremental cost-effectiveness ratio for vedolizumab was £6,634 per QALY gained when compared to adalimumab and £4,862 per QALY gained as compared to conventional therapy. In the biologic- experienced population, vedolizumab had an incremental cost-effectiveness ratio of £64,999 per QALY gained as compared with conventional therapy. The ERG carried out its own exploratory base case (combining three scenario analyses) and concluded that in the overall population, the incremental cost-effectiveness ratio for vedolizumab as compared to conventional therapy is £53,084 per QALY gained. It also concluded that in the biologic-naïve population, adalimumab dominates vedolizumab and in the biologic-experienced population, the incremental cost- effectiveness ratio for vedolizumab as compared to conventional therapy is £48,205 per QALY gained. ©Institute for Clinical and Economic Review, 2020 Page 144 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Welty, M. et al. Efficacy of Ustekinumab versus Advanced Therapies for the Treatment of Moderately-to-Severely Active Ulcerative Colitis: a Systematic Review and Network Meta- Analysis173 Researchers conducted a systematic review and NMA to compare the efficacy of ustekinumab to other advanced therapies for the treatment of moderately-to-severely active UC. Using data taken from trials of ustekinumab, infliximab, adalimumab, golimumab, vedolizumab, and tofacitinib, they conducted two fixed-effects Bayesian NMAs: one for the induction phase of the trials alone (six to eight weeks) and another for the maintenance period (one year). They also conducted separate analyses for patients who had not been failed by a prior biologic (NBF) and patients who had (BF). The induction phases of the included trials had consistent designs and could be evaluated using a standard approach; however, the maintenance phases were either structured as treat-through or re-randomized. To conduct the analysis, trials with a re-randomized response design were re- calculated to correspond to a treat-through design to maintain the randomization used at the start of induction. Researchers cite that this approach factors in both initial and delayed responders. In trials where maintenance data for the placebo was missing for induction responders or non- responders, the data was imputed. In the maintenance phase one-year NMA in NBF patients, doses were pooled as there is no conclusive evidence of a relationship between dose and efficacy. In clinical response data analyzed from six studies for NBF patients, ustekinumab had higher odds of response versus adalimumab (OR 4.76, 95% CI: 2.25 to 10.16), golimumab (OR: 3.76, 95% CI: 1.90 to 7.57), infliximab (OR: 2.62, 95% CI: 1.2 to 5.60), and tofacitinib (OR: 2.27; 95% CI: 1.06 to 4.86). Based on the results analyzed from seven studies for clinical remission, pooled ustekinumab had higher odds of clinical remission than adalimumab (OR: 2.43, 95% CI 1.10 to 5.42) and golimumab (OR: 2.40, 95% CI: 1.40 to 5.22). Results for patients who were failed a prior biologic (BF) are not presented with pooled doses because there is a potential dose-response relationship. There were not statistical differences between ustekinumab and any TIM in the BF population. In the induction phase NMA at six to eight weeks in NBF patients, ustekinumab 6 mg/kg had higher odds of response versus adalimumab (OR: 1.94, 95% CI 1.10 to 3.45). In BF patients, ustekinumab 6 mg/kg had higher had higher odds of response versus adalimumab (OR: 2.48, 95% CI 1.17 to 5.31). Overall, in patients with moderately-to-severely active UC who have not been failed by prior biologic therapy, patients on ustekinumab have a higher probability of clinical remission and response versus other advanced therapies. Patients who have previously been failed by a biologic have a similar probability versus other therapies, but they were associated with greater uncertainty because of smaller patient counts and power placebo efficacy rates. ©Institute for Clinical and Economic Review, 2020 Page 145 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Lohan, C. et al. Tofacitinib for the Treatment of Moderately-to-Severely Active Ulcerative Colitis: A Systematic Review, Network Meta-Analysis and Economic Evaluation170 Researchers conducted a systematic review, NMA, and economic evaluation to evaluate the efficacy and safety of available treatments for patients with moderately-to-severely active UC. The treatments included in the review include TNF inhibitors (adalimumab, golimumab, and infliximab) as well as tofacitinib and vedolizumab. Varying doses and dosing regimens were seen as comparators within the review. The outcomes of interest were efficacy outcomes (clinical response and remission) and serious infection. Twenty-two RCTs were included in the review and 17 of those RCTs were included in the NMA. To be included in the NMA, RCTs needed to have data for either an induction period and/or a maintenance period, reporting on the previously mentioned outcomes. The data from three treat-through trials were recalculated to match the data from five studies that re-randomized participants after the induction phase, and the differences in placebo response rates were not adjusted for. Separate analyses for the populations of biologic-naïve and biologic-experiences were conducted. For both biologic-naïve and biologic-experienced, the results of the NMA showed no significant differences among the treatments. For the biologic-naïve population, all the treatments were more effective than placebo in the induction phase, however, no results were statistically different in the maintenance phase due to large credible intervals. For the biologic-experienced population during induction, tofacitinib was the only treatment to have statistically greater efficacy than placebo for clinical response (OR: 4.28, 95% CI to 1.27-18.59) and for clinical remission (OR: 5.61, 95% CI, 1.36 to 4.53). During maintenance, tofacitinib and vedolizumab were the only treatments to appear more efficacious than placebo. For tofacitinib, significant results were seen for clinical response and clinical remission for both the 5 mg (OR: 4.53, 95% CI: 2.1 to 22.23; OR: 4.7 95% CI: 2.12 to 26.64) dose and the 10 mg dose (OR: 8.66, 95% CI: 3.87 to 65.79; OR: 8.98, 95% CI: 3.91 to 80.19), respectively. Similarly, significant results were seen for both clinical response and clinical remission for vedolizumab 300 mg every eight weeks (OR: 6.51, 95% CI: 2.45 to 46.58; OR: 6.78, 95% CI: 2.49 to 56.15) and 300 mg every four weeks (OR: 5.74, 95% CI: 1.91 to 41.39; OR: 5.97, 95% CI: 1.94 to 49.09), respectively. No significant differences were seen among the treatments for serious infections. A cost-effectiveness analysis was conducted. For the biologic-naïve population, the incremental cost-effectiveness ratio for tofacitinib as compared with conventional treatment was £21,338 per QALY. Patients treated with tofacitinib are predicted to gain more QALYs that patients treated with infliximab. Additionally, it was suggested that a mixed strategy of conventional therapy and tofacitinib would provide more QALYs overall. For the biologic-experienced population, the incremental cost-effectiveness ratio for tofacitinib as compared with conventional treatment was £22,816 per QALY. Similarly, tofacitinib is predicted to result in more QALYs than infliximab. The QALYs between vedolizumab and tofacitinib reflected near equivalence in both populations with vedolizumab having a higher total cost over the lifetime than tofacitinib, £8,730 and £4,981, respectively. ©Institute for Clinical and Economic Review, 2020 Page 146 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix C. Ongoing Studies Estimated Title/Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Dates Infliximab Inclusion Infliximab Accelerated Intervention Ages 6-17 years with UC diagnosis Induction in Moderate to Accelerated induction of Naïve to biologics Severe Pediatric UC Randomized, Parallel IFX at 0, 1, 3 wks (5 mg/kg) Planned to initiate IFX (INDUCE) Assignment, OL Study and then at wk 7, 11, 15 PUCAI ≥35 Clinical remission on April 2022 IFX at week 20 NCT03209232 Estimated Enrollment: Active Comparator Exclusion 84 Per protocol induction of Acute severe colitis Schneider Children's IFX at 0, 2, 6 wks (5 mg/kg) Renal failure or toxic megacolon Medical Center, Israel and then at wk 14 Prior treatment with IFX or ADA Adalimumab Inclusion Percentage of Efficacy and Safety of Active UC with diagnosis for at participants who: Adalimumab in Pediatric least 12 weeks prior to screening Intervention Subjects with Moderate to Phase III, MC, DB, RCT Ages 4-17 Respond at wk 8 per ADA 0.6 mg/kg every wk Severe Ulcerative Colitis PMS and achieve September 2020 Estimated Enrollment: Exclusion clinical remission at Intervention NCT02065557 100 Subject with CD or indeterminate wk 52 ADA 0.6 mg/kg EOW colitis AbbVie Current diagnosis of fulminant Achieve clinical colitis and/or toxic megacolon remission at wk 8 Proportion of Long-term Safety and Inclusion subjects who Efficacy of Adalimumab in Phase III, MC, OL Successfully enrolled and achieve: Pediatric Subjects with extension completed M11-290 study Clinical remission Ulcerative Colitis Intervention (PMS) March 2026 ADA Estimated Enrollment: Exclusion Clinical response NCT02632175 93 Considered unsuitable candidate (PMS) by investigator PUCAI response AbbVie PUCAI remission ©Institute for Clinical and Economic Review, 2020 Page 147 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Estimated Title/Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Dates Inclusion ADA group Patients with mod-to-sev UC currently taking ADA for at least 8 wks or entering after participation A Long-Term Registry of in AbbVie or Abbott sponsored UC Humira®(Adalimumab) in study Long-term Patients with Moderately Intervention IMM group Evaluation of long- Observational to Severely Active ADA Patients with mod-to-severe UC term safety of ADA Prospective Cohort Ulcerative Colitis (UC) prescribed with or currently taking in patients with April 2027 Intervention IMM therapy for at least 12 wks mod-to-severe Estimated Enrollment: NCT01848561 IMM therapy active UC 8,250 Exclusion AbbVie Patients on IMM therapy without a concurrent biologic if they cannot continue being treated with IMM therapy Patients treated with other investigational agents Golimumab Inclusion Mod-to-sev UC Must either be currently receiving A Study to Assess the treatment with, or have a history Efficacy and Safety of of having failed to respond to, or Golimumab in Pediatric have a medical contraindication to Participants with Intervention 1 at least 1 of the following Moderately to Severely Phase III, OL RCT GOL SC through wk 50 therapies: oral or IV Active Ulcerative Colitis Clinical remission at corticosteroids, MP, and September 2024 (PURSUIT 2) Estimated Enrollment: wk 6 Intervention 2 azathioprine OR must either have 125 IFX IV through wk 46 or have has history of NCT03596645 corticosteroid dependency Janssen Research & Exclusion Development, LLC Contraindications to the use of GOL or IFX or TNFs per local prescribing information ©Institute for Clinical and Economic Review, 2020 Page 148 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Estimated Title/Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Dates History of malignancy or macrophage activation syndrome Have UC limited to rectum only or to <20% of the colon Inclusion Mod-to-sev UC Cohort 1 Currently receiving GOL or is An Observational continuing to receive after Prospective Long-term participation in a UC study, or Exposure Registry of Adult Intervention schedule to receive GOL within 30 Prospective, Cohort Patients with Moderate-to- GOL days after enrollment Study Severe Ulcerative Colitis Cohort 2 Incidence of July 2031 (OPAL) Comparator Currently receiving thiopurine lymphoma Estimated Enrollment: Participants receiving Must not be receiving approved 6000 NCT02808780 thiopurines biologics agents Janssen Biotech, Inc. Exclusion Patients who cannot be treated with GOL or thiopurines Currently receiving investigational or biologic agent other than GOL Vedolizumab Inclusion 18+ years of age Entyvio (Vedolizumab) Initiating VEDO or another biologic Long Term Safety Study Prospective agent for UC or CD Intervention (Entyvio PASS) Observational Cohort Percentage of VEDO or other biologic participants with AE July 2021 agents (ADA, certolizumab NCT02674308 Estimated Enrollment: Exclusion of special interest pegol, GOL, IFX) 5,302 Prior treatment with VEDO Takeda Enrollment in a clinical trial in which treatment for CD or UC is managed through a protocol Phase II, Randomized, Intervention Inclusion Vedolizumab IV in Pediatric Serum concentration DB, Dose-ranging High dose VEDO ≥10 kg at time of randomization December 2020 Participants With at wk 14 Study Mod-to-sev active UC ©Institute for Clinical and Economic Review, 2020 Page 149 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Estimated Title/Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Dates Ulcerative Colitis (UC) or Comparator Evidence of UC extending proximal Crohn's Disease (CD) Estimated Enrollment: Low dose VEDO to rectum 80 Inadequate response to, loss of NCT03138655 response to, or intolerance to at least one of: corticosteroids, IMM, Takeda TNF Exclusion Previous exposure to approved or investigations anti-integrins Prior exposure to VEDO Use of topical treatment with ASA or corticosteroids within 2 wks of first administration of drug dose Inclusion Long-term Safety With 2-17 years old Vedolizumab Intravenous Completed Study MLN0002-2003 (IV) in Pediatric Participants Phase IIb, randomized Intervention and at wk 22, achieved clinical With Ulcerative Colitis (UC) Percentage of extension study High dose VEDO response or Crohn's Disease participants with July 2025 treatment-emergent Estimated Enrollment: Comparator Exclusion NCT03196427 AEs 80 Low dose VEDO Hypersensitivity or allergies to VEDO Takeda Withdrew from study MLN0002- 2003 Inclusion Prior participation in study Vedolizumab Subcutaneous MLN0002SC-3027 or MLN0002SC- Intervention Long-Term Open-Label Phase IIIb non- 3031 Percentage of Group A: VEDO SC 108 mg Extension Study randomized, OL participants with Q2W Exclusion study drug related February 2022 NCT02620046 Estimated Enrollment: Surgical intervention for IBD treatment-emergent Group B: VEDO SC 108 mg 692 during or after previously AEs and SAEs QW Takeda mentioned studies Withdrawal from previous studies due to study-drug related AE ©Institute for Clinical and Economic Review, 2020 Page 150 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Estimated Title/Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Dates Tofacitinib Inclusion Subjects who completed A3921094 or A3921095 and were classified as not meeting clinical response criteria Subjects who completed Long-Term Study of CP-690, maintenance study A3921096 or 550 in Subjects with OL, long-term Intervention: CP-690 5 mg who discontinued treatment early Ulcerative Colitis extension study for 12 months in study A3921096 due to Safety measured by treatment failure the number of July 2020 NCT01470612 Estimated enrollment: Intervention: CP-690 10 mg reported AEs 944 for 12 months Exclusion Pfizer Subjects who had a major protocol violation in A3921094, A3921095 or A3921096 Presence of indeterminate colitis, or findings suggestive of CD Subjects who had surgery for UC or are likely to require surgery ADA: adalimumab, AE: adverse event, ASA: aminosalicylates, CD: Crohn's disease, DB: double blind, GOL: golimumab, IBD: inflammatory bowel disease, IFX: infliximab, IMM: immunomodulator, IV: intravenous, kg: kilogram, MC: multicenter, mg: milligram, mod-to-sev: moderate-to-severe, MP: mercaptopurine, OL: open label, PUCAI: Pediatric Ulcerative Colitis Activity Index, Q2W: every two weeks, RCT: randomized controlled trial, SAE: serious adverse event, SC: subcutaneous, TNF: tumor necrosis factor, UC: ulcerative colitis, VEDO: vedolizumab, wk: week Source: www.ClinicalTrials.gov (NOTE: studies listed on site include both clinical trials and observational studies). ©Institute for Clinical and Economic Review, 2020 Page 151 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix D. Comparative Clinical Effectiveness Supplemental Information We used criteria published by the USPSTF to assess the quality of RCTs and comparative cohort studies, using the categories "good," "fair," or "poor." Guidance for quality ratings using these criteria is presented below, as is a description of any modifications we made to these ratings specific to the purposes of this review. Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study; reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; all important outcomes are considered; and appropriate attention is paid to confounders in analysis. In addition, intention-to-treat analysis is used for RCTs. Fair: Studies were graded "fair" if any or all of the following problems occur, without the fatal flaws noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with follow-up; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are addressed. Intention-to-treat analysis is done for RCTs. Poor: Studies were graded "poor" if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied equally among groups (including not masking outcome assessment); and key confounders are given little or no attention. For RCTs, intention-to- treat analysis is lacking. Note that case series are not considered under this rating system; because of the lack of comparator, these are generally considered to be of poor quality. ICER Evidence Rating We used the ICER Evidence Rating Matrix (see Figure D1) to evaluate the evidence for a variety of outcomes. The evidence rating reflects a joint judgment of two critical components: • The magnitude of the difference between a therapeutic agent and its comparator in "net health benefit" – the balance between clinical benefits and risks and/or adverse effects • The level of certainty in the best point estimate of net health benefit.64,174 ©Institute for Clinical and Economic Review, 2020 Page 152 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure D1. ICER Evidence Rating Matrix ©Institute for Clinical and Economic Review, 2020 Page 153 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D1. Study Design of Included RCTs Trial/ Timepoint Naïve (%) Treatment Arms (n) N Trial Inclusion Criteria Trial Exclusion Criteria TT or RR* Exp (%)† Induction → Maintenance Infliximab ACT 1/ 1) IFX 5 mg/kg (n=121) 1) IFX 5 mg/kg (n=121) Mayo Score 6-12 & Rutgeerts 2005 All patients endoscopic subscore ≥2 Indeterminate colitis or 2) IFX 10 mg/kg (n=122) 2) IFX 10 mg/kg (n=122) 364 Naïve (100%) moved to CD; received rectally IND + MAINT 3) PBO (n=121) 3) PBO (n=121) MAINT Inadequate response to, administered CSs or (8/54 Wks) TT at wk 0, 2, and 6 q8w through wk 46 or had failed to tolerate, medications containing ACT 2/ 1) IFX 5 mg/kg (n=121) 1) IFX 5 mg/kg (n=121) ≥1 of the following ASAs within 2 wks of Rutgeerts 2005 All patients conventional therapies: screening; previously 2) IFX 10 mg/kg (n=120) 2) IFX 10 mg/kg (n=120) 364 Naïve (100%) moved to oral ASAs, oral CSs, AZA, exposed to IFX or other IND + MAINT 3) PBO (n=123) 3) PBO (n=123) MAINT and/or MP; or were CS TNF (8/30 Wks) TT at wk 0, 2, and 6 q8w through wk 22 dependent Mayo Score 6-12 & Recent bowel surgery endoscopic subscore ≥2 or complications; bowel complications: stricture, Kobayashi 2016 Only Inadequate response to, 1) IFX 5 mg/kg (n=104) 1) IFX 5 mg/kg (n=73) fistula, or dysplasia; responders or had failed to tolerate, IND + MAINT 208 Naïve (100%) 2) PBO (n=104) 2) PBO (n=72) treatment with other moved to ≥1 of the following (8/30 Wks) TT at wk 0, 2, and 6 q8w through wk 22 biologics, MTX, MAINT conventional therapies: calcineurin inhibitors, oral ASAs, oral CSs, AZA, or cytapheresis within and/or MP; or were CS previous 18 mo dependent Indeterminate colitis or Mayo Score 6-12 & CD; recent infection, endoscopic subscore ≥2 positive TB tests; Jiang 2015 1) IFX 3.5 mg/kg (n=41) 1) IFX 3.5 mg/kg (n=41) received rectally All patients Inadequate response to, 2) IFX 5 mg/kg (n=41) 2) IFX 5 mg/kg (n=41) administered CSs or IND + MAINT 82 Naïve (100%) moved to or had failed to tolerate, 3) PBO (n=41) 3) PBO (n=41) drugs containing ASAs (8/30 Wks) TT MAINT ≥1 of the following at wk 0, 2, and 6 q8w through wk 22 within 2 wks of conventional therapies: screening; previously oral ASAs, oral CSs, AZA, exposed to IFX or other and/or MP TNF Extensive colitis or UC Mayo Score 6-12 & limited to only rectum NCT01551290 endoscopic subscore ≥2 or less than 20 cm of 1) IFX 5 mg/kg (n=49) All patients 1) IFX 5 mg/kg (n=50) colon; treatment with IND + MAINT 99 Naïve (100%) 2) PBO (n=50) moved to 2) PBO (n=49) Active UC despite cyclosporine, (8/26 Wks) TT at wk 0, 2, and 6 MAINT q8w through wk 22 treatment with CSs, AZA, tacrolimus, sirolimus, or MP, or ASA mycophenolate mofetil within 8 wks ©Institute for Clinical and Economic Review, 2020 Page 154 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial/ Timepoint Naïve (%) Treatment Arms (n) N Trial Inclusion Criteria Trial Exclusion Criteria TT or RR* Exp (%)† Induction → Maintenance Adalimumab Ulcerative proctitis, Mayo Score 6-12 & 1) ADA 160/80 mg (n=130) previous receipt of TNF endoscopic subscore ≥2 ULTRA 1/ 160 mg at wk 0, 80 mg at wk 2, or biologic including Reinisch 2011 40 mg at wks 4 and 6 ADA, IV CS, Inadequate response to, 390 Naïve (100%) 2) ADA 80/40 mg (n=130) -- -- cyclosporine, IND only loss of response to, or 80 mg at wk 0, 40 mg at wks 2, 4, tacrolimus, (8 Wks) intolerance to at least 1, and 6 mycophenolate mofetil either oral CS, and/or 3) PBO (n=130) or MTX 30-60 days prior IMM to baseline History of subtotal Mayo Score 6-12 & colectomy, Kock pouch, endoscopic subscore ≥2 or planned bowel surgery; previous ULTRA 2/ Starting at wk Inadequate response to, 1) ADA 160/80 mg (n=248) treatment with ADA; Sandborn 2012 12, inadequate 1) ADA 40 mg EOW (n=248) or had failed to tolerate, Naïve (60%) 160 mg at wk 0, 80 mg at wk 2, receipt of IV CSs within 494 responders 2) PBO (n=246) ≥1 of the following IND + MAINT Exp (40%) then 40 mg EOW 2 wks of screening; could receive through wk 52 conventional therapies: (8/52 Wks) TT 2) PBO (n=246) receipt of therapeutic OL ADA oral ASAs, oral CSs, AZA, enema or suppository, and/or MP; TNF- other than required for experienced other than endoscopy, within 2 ADA wks of the screening Mayo Score 6-12 & Indeterminate colitis or endoscopic subscore ≥2 CD; planned bowel 1) ADA 160/80 mg (n=87) Starting at wk surgery; received CS Suzuki 2014 160 mg at wk 0, 80 mg at wk 2, 1) ADA 40 mg EOW (n=177) 8, inadequate Inadequate response to, injection, cyclosporine, then 40 mg EOW IND + MAINT 273 Naïve (100%) responders or had failed to tolerate, tacrolimus, or 2) ADA 80/40 mg (n=90) 2) PBO (n=96) (8/52 Wks) TT entered into ≥1 of following mycophenolate mofetil 80 mg at wk 0 then 40 mg EOW through wk 52 "rescue arm" conventional therapies: within 4 wks; prior 3) PBO (n=96) oral ASAs, oral CSs, AZA, treatment with TNFs or and/or MP; TNF-naïve biologic Golimumab Mayo Score 6-12 & TNF(s), natalizumab or 1) GOL 200/100 mg (n=257) 200 endoscopic subscore ≥2 other agents targeting PURSUIT SC/ mg at wk 0 and a-4 integrin, B-cell Sandborn 2014/ Naïve 100 mg at wk 2 Inadequate response to, depleting agents, or T- 773 -- -- IND Only (100%) 2) GOL 400/200 mg (n=258) 400 or had failed to tolerate, cell depleting agents (6 Wks) mg at wk 0 and 200 mg at wk 2 ≥1 of following within 12 mo of first 3) PBO (n=258) conventional therapies: study-agent injection; oral ASAs, oral CSs, AZA, cyclosporine within 8 ©Institute for Clinical and Economic Review, 2020 Page 155 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial/ Timepoint Naïve (%) Treatment Arms (n) N Trial Inclusion Criteria Trial Exclusion Criteria TT or RR* Exp (%)† Induction → Maintenance and/or MP; or CS wks before first study dependent agent injection Mayo Score 6-12 & endoscopic subscore ≥2 PURSUIT M/ Responders at 1) GOL 50 mg (n=154 ) Inadequate response to, Sandborn 2014 Patients with isolated Naïve wk 6 were 2) GOL 100 mg (n=154 ) or had failed to tolerate, 464 PURSUIT-SC and PURSUIT-IV proctitis, patients with MAINT (100%) randomized in 3) PBO (n=156) ≥1 of following TB (54 Wks) RR MAINT q4w through wk 52 conventional therapies: oral ASAs, oral CSs, AZA, and/or MP; CS dependent Mayo Score 6-12 & endoscopic subscore ≥2 Severe and extensive PURSUIT J/ Responders at Inadequate response to, colitis requiring Hibi 2017 1) GOL 100 mg (n=32) Naïve 1) OL GOL 200/100 mg (n=144) wk 6 were or had failed to tolerate, colectomy; colitis 144 2) PBO (n=31) MAINT (100%) 200 mg at wk 0, 100 mg at wk 2 randomized in ≥1 of following limited to 20 cm of q4w through wk 52 (54 Wks) RR MAINT conventional therapies: colon; any prior oral ASAs, oral CSs, AZA, abdominal surgery and/or MP; CS dependent Vedolizumab Received TNF within 60 Mayo Score 6-12 & days prior to endoscopic subscore of enrollment; Cohort 1 ≥2 GEMINI 1/ cyclosporine, 1) VEDO 300 mg (n=225) Responders at Feagan 2013 Naïve (52%) 1) VEDO q8w (n=122) thalidomide, or 2) PBO (n=149) wk 6 were Inadequate response to, 895 Experienced 2) VEDO q4w (n=125) investigational drugs IND + MAINT Cohort 2 randomized in or had failed to tolerate, (48%) 3) PBO (n=126) through wk 52 within 30 days of (6/52 Wks) RR 1) OL VEDO 300 mg (n=521) at MAINT ≥1 of following enrollment; previous wk 0 and 2 conventional therapies: VEDO, natalizumab, glucocorticoids, AZA, MP, efalizumab, or or TNFs rituximab Mayo Score 6-12 & Exposure to any VISIBLE 1‡/ 1) VEDO 108 mg (SC) q2w endoscopic subscore ≥2 biologics within 60 days Responders at Sandborn 2019 (n=106) or 5 half-lives of Naïve (61%) 1) OL VEDO 300 mg (n=383) at wk 6 were 383 2) VEDO 300 mg (IV) q8w Inadequate response to, screening; exposure to IND + MAINT Exp (39%) wk 0 and 2 randomized in (n=154) loss of response to, or any nonbiologic (6/52 Wks) RR MAINT 3) PBO (n=56) through wk 52 intolerance to at least 1, therapies such as either a CS, IMM, or TNF cyclosporine, ©Institute for Clinical and Economic Review, 2020 Page 156 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial/ Timepoint Naïve (%) Treatment Arms (n) N Trial Inclusion Criteria Trial Exclusion Criteria TT or RR* Exp (%)† Induction → Maintenance tacrolimus, thalidomide, MTX, or TOF within 30 days or 5 half-lives of screening was also not permitted Suspected abdominal abscess or toxic Mayo Score 6-12 & Cohort 1 megacolon; history of endoscopic subscore ≥2 Motoya 2019 1) VEDO 300 mg (n=164) Responders at colectomy or recent Naïve (49%) 2) Placebo (n=82) wk 6 were 1) VEDO 300 mg q8w (n=41) enterectomy or IND + MAINT 292 Total or left-side Exp (51%) Cohort 2 randomized in 2) PBO (n=42) through wk 52 previously treatment (10/60 Wks) RR diagnosis with treatment 1) OL VEDO 300 mg (n=46) at wk MAINT with VEDO, failure with CSs, IMMs, 0, 2, and 6 natalizumab, or TNF efalizumab, or rituximab Head-to-Head Mayo Score 6-12 & Crohn's colitis, or endoscopic subscore ≥2 indeterminate colitis; VARSITY/ 1) VEDO 300 mg (n=383) subtotal or total Sands 2019 at wk 0, 2, and 6 All patients 1) VEDO 300 mg q8w (n=383) No response or loss of Naïve (79%) colectomy; active 769 2) ADA 160/80mg (n=386) moved to 2) ADA 40 mg EOW (n=386) response to conventional IND + MAINT Exp (21%) infection, cyclosporine, 160 mg at wk 0, 80 mg at wk 2, MAINT through wk 50 treatments or (14/52 Wks) TT tacrolimus, thalidomide and 40 mg at wk 4 and 6 discontinued treatment within 30 days; history with a TNF (except ADA) of malignancy or TNF-naïve Tofacitinib OCTAVE 1/ Mayo Score 6-12 & rectal Sandborn 2017 1) TOF 10 mg (n=476) bleeding subscore of 1 to Naïve (47%) 598 2) PBO (n=122) -- -- 3 & endoscopic subscore Presence of CD, UC IND only Exp (53%) twice daily for 8 wks of ≥2 limited to distal 15 cm (8 Wks) of colon, signs of Inadequate response to, fulminant colitis, toxic OCTAVE 2/ or had failed to tolerate, megacolon, or Sandborn 2017 1) TOF 10 mg (n=476) ≥1 of following indeterminate, Naïve (45%) 588 2) PBO (n=112) -- -- conventional therapies: microscopic, ischemic, IND only Exp (55%) twice daily for 8 wks oral or IV glucocorticoids, or infectious colitis (8 Weeks) AZA, and/or MP, IFX, or ADA ©Institute for Clinical and Economic Review, 2020 Page 157 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Trial/ Timepoint Naïve (%) Treatment Arms (n) N Trial Inclusion Criteria Trial Exclusion Criteria TT or RR* Exp (%)† Induction → Maintenance OCTAVE Patients who met SUSTAIN/ Responders at 1) TOF 5 mg (n=198) Adults who completed criteria for treatment Naïve (52%) wk 8 were 2) TOF 10 mg (n=197) OCTAVE induction 1 or 2 Sandborn 2017 593 OCTAVE 1 and 2 failure and received Exp (48%) randomized in 3) PBO (n=198) and had clinical response MAINT rescue therapy during MAINT twice daily for 52 wks during induction trial (52 Wks), RR IND trial Ustekinumab Severe extensive colitis Mayo Score 6-12 & and at imminent risk of endoscopic subscore ≥2 colectomy, have UC UNIFI/ 1) UST 130 mg (n=320) Responders at limited to rectum, Sands 2019 1) UST 90 mg SC q12w (N=172) Inadequate response to Naïve (49%) 2) UST 6 mg/kg (n=322) wk 8 were presence of a stoma or 961 2) UST 90 mg SC q8w (N=176) or unacceptable side IND + MAINT Exp (51%) 3) PBO (n=319) randomized in history of a fistula, through wk 52 effects from TNFs, VEDO, (8/52 Wks), RR single dose MAINT history of extensive or conventional colonic resection, or (i.e., nonbiologic) history of mucosal therapy dysplasia ADA: adalimumab, ASA: acetylsalicylic acid, AZA: azathioprine, CD: Crohn's disease, cm: centimeter, CS: corticosteroid, EOW: every other week, Exp: Experienced, GOL: golimumab, IFX: infliximab, IMM: immunomodulator, IND: induction, IV: intravenous, MAINT: maintenance, mg: milligram, mg/kg: milligram per kilogram, MP: mercaptopurine, mo.: month, MTX: methotrexate, n: number, N: total number, OL: open label, PBO: placebo, q12w: every 12 weeks, q8w: every 8 weeks, q4w: every 4 weeks, RR: re-randomized, SC: subcutaneous, TB: tuberculosis, TNF: tumor necrosis factor, TOF: tofacitinib, TT: treat through, UC: ulcerative colitis, UST: ustekinumab, VEDO: vedolizumab, Wk: week *Treat-through or re-randomized trial for trials with maintenance phase. †Reported are the proportions of biologic-naïve and biologic-experienced populations. Note that trials used different criteria regarding prior exposure to biologics to define their strata when reporting subgroup results. ULTRA 2, Motoya 2019, and VARSITY defined subgroups as "TNF-naïve" and "TNF-experienced." GEMINI 1 and VISIBLE 1 defined subgroups as "TNF-naïve" and "TNF-failure." The OCTAVE trials defined subgroups as "no TNF-failure" and "TNF-failure." UNIFI defined subgroups as "biologic failure" and "biologic non-failure." Of note, UNIFI allowed patients who were failed by vedolizumab to enroll. ‡For VISIBLE 1, subcutaneous vedolizumab was not an intervention of interest, but we have presented data in the appendices. ©Institute for Clinical and Economic Review, 2020 Page 158 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D2. Study Quality of Included RCTs Non- Clear Clear Patient/Investigator Intention Approach Comparable Differential Definition Definition Measurements USPSTF Trial Blinding (Double- to Treat to Missing Groups Lost to of of Valid Rating Blind) Analysis Data† Follow-Up Intervention Outcomes ACT 1 Yes Yes Yes Yes Yes Yes ITT NRI Fair ACT 2 Yes No Yes Yes Yes Yes ITT NRI Fair Jiang 2015 Yes No Yes Yes Yes Yes ITT NRI Fair Kobayashi 2016 Yes Yes Yes Yes Yes Yes ITT NRI/LOCF Fair NCT01551290 Unclear Yes Yes Yes Yes Yes mITT NRI * ULTRA 1 Yes Yes Yes Yes Yes Yes mITT NRI Good ULTRA 2 Yes Yes Yes Yes Yes Yes ITT NRI Good Suzuki 2014 Yes Yes Yes Yes Yes Yes mITT NRI Good PURSUIT-SC Yes Yes Yes Yes Yes Yes ITT NRI Good PURSUIT-M Yes Yes Yes Yes Yes Yes ITT NRI Good PURSUIT-J Yes No Yes Yes Yes Yes ITT NRI/LOCF Fair OCTAVE 1 Yes Yes Yes Yes Yes Yes ITT NRI Good OCTAVE 2 Yes Yes Yes Yes Yes Yes ITT NRI Good OCTAVE SUSTAIN Yes Yes Yes Yes Yes Yes ITT NRI Good UNIFI Yes Yes Yes Yes Yes Yes ITT NRI Good GEMINI 1 Yes Yes Yes Yes Yes Yes ITT NRI Good Motoya 2019 Yes Yes Yes Yes Yes Yes mITT NRI Good VISIBLE 1 Yes Yes Yes Yes Yes Yes mITT NRI Good VARSITY Yes Yes Yes Yes Yes Yes mITT NRI Good ITT: intention-to-treat, LOCF: last observation carried forward, mITT: modified intention-to-treat, NRI: non-responder imputation, USPSTF: United States Preventive Services Task Force *Data was only available in grey literature. Due to this, we did not assign an overall quality rating for the trial. †For response and remission outcomes. ©Institute for Clinical and Economic Review, 2020 Page 159 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D3. Baseline Characteristics in Included RCTs Disease Disease Severity Prior TNF Study Name/ Age (Yrs), Female, Weight (kg), Duration Classification (Mayo), n Mayo Score, Disease Localization, CRP (mg/dL), Arms N Use, Trial Identifier Mean (SD) n (%) Mean (SD) (Yrs), (%) Mean (SD) n (%) Mean (SD) n (%) Mean (SD) Moderate Severe Left Side: 63 (52.9) IFX 5 mg/kg 121 42.4 (14.3) 43 (35.5) 80 (17.8) 5.9 (5.4) N/A NR NR 8.5 (1.7) 1.4 (1.9) Extensive: 56 (47.1) Left Side: 67 (55.4) ACT 1 IFX 10 mg/kg 122 41.8 (14.6) 50 (41.0) 76.9 (17.1) 8.4 (8.1) N/A NR NR 8.4 (1.4) 1.6 (2.3) Extensive: 54 (44.6) Left Side: 66 (55) PBO 121 41.4 (13.7) 49 (40.5) 76.8 (16.2) 6.2 (5.9) N/A NR NR 8.4 (1.8) 1.7 (2.7) Extensive: 54 (45) Left Side: 70 (59.3) IFX 5 mg/kg 121 40.5 (13.1) 45 (37.2) 78.4 (17.8) 6.7 (5.3) N/A NR NR 8.3 (1.5) 1.3 (2.3) Extensive: 48 (40.7) Left Side: 75 (62.5) ACT 2 IFX 10 mg/kg 120 40.3 (13.3) 52 (43.3) 40.3 (13.3) 6.5 (5.8) N/A NR NR 8.3 (1.6) 1.4 (2.2) Extensive: 45 (37.5) Left Side: 70 (58.3) PBO 123 39.3 (13.5) 52 (42.3) 39.3 (13.5) 6.5 (6.7) N/A NR NR 8.5 (1.5) 1.6 (2.9) Extensive: 50 (41.7) Left Side: 21 (20.3) IFX 5 mg/kg 104 40 (12.7) 38 (36.5) 57.6 (12.7) 8.1 (7.2) N/A NR NR 8.6 (1.4) 1.0 (1.5) Extensive: 83 (79.8) Kobayashi 2016 Left Side: 20 (19.2) PBO 104 37.8 (12.9) 37 (35.6) 60.3 (11.6) 7.1 (6.6) N/A NR NR 8.5 (1.4) 0.7 (1.1) Extensive: 84 (80.8) Left Side: 16 (39.1) IFX 5 mg/kg 41 34.3 (14.3) 15 (36.6) 62.8 (14.9) 4.4 (2.8) N/A 15 (36.6) 26 (63.4) NR 3.6 (2.26) Pancolitis: 25 (60.9) Jiang 2015 Left Side: 17 (41.5) PBO 41 34.5 (14.9) 16 (39.1) 61.2 (15.7) 4.4 (2.6) N/A 16 (31.9) 25 (60.9) NR 3.5 (1.8) Pancolitis: 24 (58.5) IFX 5 mg/kg 50 NR NR 3.7 N/A NR NR NR NR NCT01551290 37 (Median) 8.0 (Median) PBO 49 NR NR (Median) N/A NR NR NR NR Rectum and Sigmoid Colon Only: 25 (11.1); VEDO 300 Left Side: 92 (40.9); mg (Cohort 225 40.1 (13.1) 93 (41.3) 72.4 (17.1) 6.1 (5.1) 95 (42.2) N/A NR 8.5 (1.8) NR Proximal to the splenic 1) flexure: 25 (11.1); All: 83 (36.9) GEMINI 1 Rectum and Sigmoid Colon Only: 22 (14.8); Left Side: 59 (39.6); PBO 149 41.2 (12.5) 57 (38.3) 72.4 (17.6) 7.1 (7.2) 73 (49) NR NR 8.6 (1.7) NR Proximal to the splenic flexure: 18 (12.1); All: 50 (33.6) Proctosigmoiditis: 15 (14.2) VEDO 108 VISIBLE 1 106 38.1 (13.1) 41 (38.7) 71.6 (17.2) 9 (6.2) 40 (37.7) 46 (43.4) 60 (56.6) 9.0 (Median) Left Side: 46 (43.4) NR mg (SC) Extensive: 7 (6.6) Pancolitis: 37 (34.9) ©Institute for Clinical and Economic Review, 2020 Page 160 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Disease Disease Severity Prior TNF Study Name/ Age (Yrs), Female, Weight (kg), Duration Classification (Mayo), n Mayo Score, Disease Localization, CRP (mg/dL), Arms N Use, Trial Identifier Mean (SD) n (%) Mean (SD) (Yrs), (%) Mean (SD) n (%) Mean (SD) n (%) Mean (SD) Moderate Severe Proctosigmoiditis: 7 (13) VEDO 300 Left Side: 21 (38.9) 54 41.6 (14.1) 23 (42.6) 77 (16.9) 8.2 (5.9) 24 (44.4) 17 (31.5) 37 (68.5) 9.0 (Median) NR mg (IV) Extensive: 7 (13) Pancolitis:19 (35.2) Proctosigmoiditis: 7 (12.5) PBO 56 39.4 (11.7) 22 (39.3) 74 (20.9) 7.4 (7.1) 20 (35.7) 20 (35.7) 36 (64.3) 9.0 (Median) Left Side: 24 (42.9) NR Extensive: 4 (7.1) Pancolitis: 21 (37.5) <3 mg/L: 76 VEDO 300 Total Colitis: 101 (61.6) (46.3) mg (Cohort 164 42.3 (14.4) 65 (39.6) NR 7.2 (6.2) 85 (51.8) NR NR 8.3 (1.5) Left Side: 63 (38.4) ≥3 mg/L: 50 1) (61.0) Motoya 2019 <3 mg/L: 88 Total Colitis: 51 (62.2) (53.7) PBO 82 44.0 (16.0) 27 (32.9) NR 8.6 (8.0) 41 (50.0) NR NR 8.1(1.5) Left Side: 31 (37.8) ≥3 mg/L: 32 (39.0) ADA 160/80/40 386 40.5 (13.4) 170 (44.0) 73.4 (18.4) 6.4 (6.0) 80 (21.0) 169 (43.8) 210 (54.4) 8.7 (1.5) NR NR VARSITY mg VEDO 300 385 40.8 (13.7) 111 (39.2) 72.7 (17) 7.3 (7.2) 80 (20.8) 154 (40.0) 217 (56.4) 8.7 (1.6) NR NR mg Left Side: 36.9 6.91 ADA 80/40 130 40 (Median) 52 (40) 76.8 (15) N/A NR NR 9 (1.62) Extensive: 53.8 0.64 (Median) (Median) Other: 9.2 ULTRA 1 Left Side: 46.9 36.5 6.06 ADA 160/80 130 47 (36.2) 75.5 (14.2) N/A NR NR 8.8 (1.61) Extensive: 46.2 0.33 (Median) (Median) (Median) Other: 6.9 Left Side: 32.3 5.4 PBO 130 37 (Median) 48 (36.9) 78.7 (17.4) N/A NR NR 8.7 (1.56) Extensive: 56.2 0.32 (Median) (Median) Other: 11.5 Pancolitis: 120 (48.4) ADA Descending Colon: 96 248 39.6 (12.47) 106 (42.7) 75.3 (17.71) 8.1 (7.09) 98 (39.1) NR NR 8.9 (1.5) 1.5 (3.2) 160/80/40 (38.7) Other: 32 (12.9) ULTRA 2 Pancolitis: 120 (48.8) Descending Colon: 96 PBO 246 41.3 (13.22) 94 (38.2) 77.1 (17.31) 8.5 (7.37) 101 (41.1) NR NR 8.9 (1.8) 1.3 (3.7) (39) Other: 30 (12.2) Pancolitis: 63 (70) Suzuki 2014 ADA 160/80 90 42.5 (14.6) 29 (32.2) 60.1 (12.3) 7.8 (7.1) N/A NR NR 8.6 (1.4) Descending colon: 27 0.22 (Median) (30) ©Institute for Clinical and Economic Review, 2020 Page 161 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Disease Disease Severity Prior TNF Study Name/ Age (Yrs), Female, Weight (kg), Duration Classification (Mayo), n Mayo Score, Disease Localization, CRP (mg/dL), Arms N Use, Trial Identifier Mean (SD) n (%) Mean (SD) (Yrs), (%) Mean (SD) n (%) Mean (SD) n (%) Mean (SD) Moderate Severe Pancolitis: 54 (62.1) ADA 80/40 87 44.4 (15) 37 (42.5) 58.7 (11.1) 8.3 (7.7) N/A NR NR 8.5 (1.4) Descending Colon: 32 0.31 (Median) (36.8) Pancolitis: 59 (61.5) PBO 96 41.3 (13.6) 26 (27.1) 60.8 (14.1) 7.8 (6.6) N/A NR NR 8.5 (1.6) Descending Colon: 35 0.34 (Median) (36.5) GOL 100/50 Left Side: 43 (59.7) 0.8 72 40.9 (12.19) 32 (44.4) NR 6.6 (7.33) N/A 49 (68.1) 23 (31.9) 8.2 (1.36) mg Extensive: 29 (40.3) (1.0) GOL 200/100 40.0 Left Side: 193 (58.3) 1.1 331 151 (45.6) NR 6.4 (6.17) N/A 182 (55.2) 148 (44.8) 8.6 (1.53) PURSUIT-SC mg (13.54) Extensive: 138 (41.7) (1.5) GOL 400/ Left Side: 191 (57.7) 331 40.7 (13.75) 130 (39.3) NR 6.4 (6.27) N/A 195 (59.5) 133 (40.5) 8.5 (1.47) 1.3 (2.6) 200 mg Extensive: 140 (42.3) Left Side: 188 (57.0) PBO 331 39 (13.04) 146 (47.1) NR 6.0 (6.65) N/A 209 (63.7) 119 (36.3) 8.3 (1.50) 1.1 (1.7) Extensive: 142 (43.0) GOL 50 mg 154 41.4 (13.84) 77 (50.0) NR 6.8 (6.93) N/A 145 (94.2) 143 (92.9) 8.1 (1.38) NR 0.9 (1.3) PURSUIT-M GOL 100 mg 154 39.1 (13.11) 65 (42.2) NR 7.2 (7.04) N/A 9 (5.8) 11 (7.1) 8.5 (1.34) NR 0.9 (1.5) PBO 156 40.2 (14.05) 81 (51.9) NR 6.9 (6.96) N/A 145 (92.9) 11(7.1) 8.3 (1.37) NR 1.0 (1.5) OL-IND: GOL 42.40 61.51 5.08 Left Side: 89 (62) 144 46 (32) N/A 141 (98) N/A 8.0 (Median) 0.4 (1.1) SC 200 mg (14.74) (11.18) (Median) Extensive: 55 (38) 39.30 64.59 5.35 Left Side: 20 (63) PURSUIT-J GOL 100 mg 32 13 (41) N/A 31(97) N/A 8.0 (Median) 0.5 (1.5) (12.00) (14.73) (Median) Extensive: 12 (38) 42.90 5.74 Left Side: 19 (61) PBO 31 12 (39) 59.48 (9.73) N/A 30(97) N/A 8.0 (Median) 0.4 (0.8) (14.41) (Median) Extensive: 12 (39) Proctosigmoiditis: 65 (13.7) 6.5 TOF 10 mg 476 41.3 (14.1) 199 (41.8) 72.9 (16.8) 254 (53.4) NR NR 9.0 (1.4) Left Side: 158 (33.3) 0.4 (Median) (Median) Extensive Colitis or Pancolitis: 252 (53.1) OCTAVE 1 Proctosigmoiditis: 19 (15.6) 6.0 PBO 122 41.8 (15.3) 45 (36.9) 72.7 (16.7) 65 (53.3) NR NR 9.1 (1.4) Left Side: 37 (30.3) 0.5 (Median) (Median) Extensive Colitis or Pancolitis: 66 (54.1) Proctosigmoiditis: 67 (15.7) 6.0 TOF 10 mg 429 41.1 (13.5) 170 (39.6) 74.4 (16.8) 234 (54.5) NR NR 9.0 (1.5) Left Side: 149 (34.8) 0.5 (Median) (Median) Extensive Colitis or OCTAVE 2 Pancolitis: 211 (49.3) Proctosigmoiditis: 16 6.2 PBO 112 40.4 (13.2) 61 (50.9) 73.2 (16.2) 65 (58.0) NR NR 8.9 (1.5) (14.4) 0.5 (Median) (Median) Left Side: 39 (35.1) ©Institute for Clinical and Economic Review, 2020 Page 162 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Disease Disease Severity Prior TNF Study Name/ Age (Yrs), Female, Weight (kg), Duration Classification (Mayo), n Mayo Score, Disease Localization, CRP (mg/dL), Arms N Use, Trial Identifier Mean (SD) n (%) Mean (SD) (Yrs), (%) Mean (SD) n (%) Mean (SD) n (%) Mean (SD) Moderate Severe Extensive Colitis or Pancolitis: 56 (50.5) Proctosigmoiditis: 33 (16.8) 6.8 TOF 10 mg 197 42.9 (14.4) 87 (44.2) 74.6 (15.1) 101 (51.3) NR NR 3.4 (1.8) Left Side: 60 (30.6) 0.09 (Median) (Median) Extensive Colitis or Pancolitis: 103 (52.6) Proctosigmoiditis: 28 OCTAVE (14.3) 6.5 SUSTAIN TOF 5 mg 198 41.9 (13.7) 95 (48) 73.4 (17.8) 90 (45.5) NR NR 3.3 (1.8) Left Side: 66 (33.7) 0.07 (Median) (Median) Extensive Colitis or Pancolitis: 102 (52.0) Proctosigmoiditis: 21 (10.6) 7.2 PBO 198 43.4 (14.0) 82 (41.4) 76.2 (16.7) 92 (46.5) NR NR 3.3 (1.8) Left Side: 68 (34.3) 0.1 (Median) (Median) Extensive Colitis or Pancolitis: 108 (54.5) UST 6 mg/kg 322 41.7 (13.7) 39.4 73 (19.3) 8.2 (7.8) NR 276 (86.0) 46 (14.0) 8.9 (1.5) Left Side: 168 (52.5) 0.5 (0.2-1.3) UNIFI UST 130 mg 320 42.2 (13.9) 40.6 73.7 (16.8) 8.1 (7.2) NR 271 (84.7) 49 (15.3) 8.9 (1.6) Left Side: 183 (57.5) 0.5 (0.2-0.9) PBO 319 41.2 (13.5) 38.2 72.9 (16.8) 8 (7.2) NR 263 (82.4) 56 (17.6) 8.9 (1.6) Left Side: 167 (52.8) 0.5 (0.2-1.0) ADA: adalimumab, AZA: azathioprine, DB: double blind GOL: golimumab, IFX: infliximab, IV: intravenous, kg: kilogram, mg/L: milligram per liter, mg/dL: milligrams per deciliter, N: total number, N/A: not applicable, NR: not reported, OL: open label, PBO: placebo, SC: subcutaneous, SD: standard deviation, TNF: tumor necrosis factor, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, Yrs: years ©Institute for Clinical and Economic Review, 2020 Page 163 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D4. Baseline Medication Use in Included RCTs Study Name/ Concomitant Medication, n (%) Arms N Trial Identifier CS CS ≥20 mg/Day ASA IMM AZA MERC CS + IMM Infliximab IFX 5 mg 121 70 (57.9) 45 (37.2) 82 (67.8) 66 (54.5) 45 (37.2) 15 (12.3) NR ACT 1 IFX 10 mg 122 73 (59.8) 46 (37.7) 86 (70.5) 59 (48.4) 44 (36.1) 21 (17.4) NR PBO 121 79 (65.3) 54 (44.6) 85 (70.2) 53 (43.8) 36 (29.8) 17 (14.0) NR IFX 5 mg 121 60 (49.6) 40 (33.1) 92 (76) 52 (43) 41 (33.9) 11 (9.1) NR ACT 2 IFX 10 mg 120 66 (55) 47 (39.2) 47 (39.2) 50 (41.7) 37 (30.8) 13 (10.8) NR PBO 123 60 (48.8) 43 (35) 43 (35) 54 (43.9) 35 (28.5) 19 (15.4) NR Kobayashi IFX 5 mg 104 68 (65.4) NR 77 (74.0) 50 (48.1) 38 (36.5) 12 (11.5) NR 2016 PBO 104 69 (66.3) NR 70 (67.3) 49 (47.1) 34 (32.7) 15 (14.4) NR IFX 5 mg 41 22 (53.7) 14 (34.1) 34 (82.9) NR 12 (29.3) NR NR Jiang 2015 PBO 41 21 (51.2) 14 (34.1) 35 (85.4) NR 13 (31.7) NR NR IFX 5 mg 50 30 (60) NR NR NR NR NR NR NCT01551290 Placebo 49 39 (80) NR NR NR NR NR NR Vedolizumab VEDO Cohort 1 225 79 (35.1)* NR NR 28 (12.4)* NR NR NR GEMINI 1 PBO 149 58 (38.9)* NR NR 18 (12.1)* NR NR NR VEDO SC 106 45 (42.5) NR NR NR NR NR NR VISIBLE 1 VEDO IV 54 21 (38.9) NR NR NR NR NR NR PBO 56 32 (57.1) NR NR NR NR NR NR VEDO 164 31 (18.9)* NR 145 (88.4) 59 (36.0)* NR NR 21 (12.8) Motoya 2019 PBO 82 11 (13.4)* NR 75 (91.5) 29 (35.4)* NR NR 14 (17.1) Head-to-Head ADA 386 140 (36.3)* NR NR 100 (25.9)* NR NR NR VARSITY VEDO 385 139 (36.1)* NR NR 101 (26.2)* NR NR NR Adalimumab ADA 80/40 130 48 (36.9)* NR 99 (76.2) 25 (19.2)* NR NR 26 (20.0) ULTRA 1 ADA 160/80 130 48 (36.9)* NR 105 (80.8) 28 (21.5)* NR NR 23 (17.7) PBO 130 55 (41.5)* NR 98 (75.4) 18 (13.8)* NR NR 34 (26.1) ADA 248 150 (60.5) NR 146 (58.9) 93 (37.5) NR NR NR ULTRA 2 PBO 246 140 (56.9) NR 155 (63) 80 (32.5) NR NR NR ADA 160/80 90 57 (63.3) NR 83 (92.2) 41 (45.6) NR NR NR Suzuki 2014 ADA 80/40 87 63 (72.4) NR 84 (96.6) 38 (43.7) NR NR NR PBO 96 58 (60.4) NR 89 (92.7) 52 (54.2) NR NR NR ©Institute for Clinical and Economic Review, 2020 Page 164 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Study Name/ Concomitant Medication, n (%) Arms N Trial Identifier CS CS ≥20 mg/Day ASA IMM AZA MERC CS + IMM Golimumab GOL 100/50 mg 72 35 (48.6) 25 (34.7) 59 (81.9) 27 (37.5) 27 (37.5) NR GOL 200/100 mg 331 142 (42.9) 85 (25.7) 27 (81.6) 105 (31.7) 100 (30.2) NR PURSUIT SC GOL 400/200 mg 331 145 (43.8) 93 (28.1) 267 (80.7) 107 (32.3) 103 (31.1) NR PBO 331 134 (40.5) 78 (23.6) 276 (83.4) 106 (32.0) 102 (30.8) NR GOL 50 mg 154 77 (50.0) 52 (33.8) 128 (83.1) 47 (30.5) 45 (29.2) NR PURSUIT-M GOL 100 mg 154 79 (51.3) 55 (35.7) 119 (77.3) 48 (31.2) 48 (31.2) NR PBO 156 83 (53.2) 59 (37.8) 125 (80.1) 52 (33.3) 51 (32.7) NR Induction: GOL 144 42 (29) 12 (8) 128 (89) NR 64 (44) NR SC 200 mg DB MAINT: GOL PURSUIT-J 32 9 (28) 4 (13) 29 (91) NR 16 (50) NR SC 100 mg DB MAINT: PBO 31 9 (29) 5 (16) 27 (87) NR 13 (42) NR 100 mg Tofacitinib TOF 10 mg 476 214 (45.0) NR NR NR NR NR NR OCTAVE 1 PBO 122 58 (47.5) NR NR NR NR NR NR TOF 10 mg 429 198 (46.2) NR NR NR NR NR NR OCTAVE 2 PBO 112 55 (49.1) NR NR NR NR NR NR OCTAVE TOF 10 mg 197 87 (44.2) NR NR NR NR NR NR TOF 5 mg 198 101 (51.0) NR NR NR NR NR NR SUSTAIN PBO 198 100 (50.5) NR NR NR NR NR NR Ustekinumab UST 6 mg/kg 322 168 (52.2) NR 238 (73.9) 89 (27.6) NR NR NR UNIFI UST 130 mg 320 173 (54.1) NR 215 (67.2) 93 (29.1) NR NR NR PBO 319 157 (49.2) NR 207 (64.9) 89 (27.9) NR NR NR ADA: adalimumab, ASA: aminosalicylates, AZA: azathioprine, CS: corticosteroids, DB: double blind, GOL: golimumab, IFX: infliximab, IMM: immunomodulator, IV: intravenous, MERC: mercaptopurine, mg: milligram, mg/kg: milligram per kilogram, N: total number, N/A: not applicable, NR: not reported, OL: open label, PBO: placebo, SC: subcutaneous, SD: standard deviation, TNF: tumor necrosis factor, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab *Reported as the proportion of patients taking IMM or CS alone. ©Institute for Clinical and Economic Review, 2020 Page 165 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D5. Response, Remission, and Endoscopic Improvement in the Induction Phase (Week Six to 14)* Induction Phase (Week 6-14) Trial Arm Response Remission Endoscopic Improvement n N % Significance n N % Significance n N % Significance Week 8 Overall – Biologic-Naïve ACT 1 IFX 5 mg/kg 84 121 69.4 p<0.001 47 121 38.8 p<0.001 75 121 62 p<0.001 IFX IFX 10 mg/kg 75 122 61.5 p<0.001 39 122 32.0 p=0.002 72 122 59 p<0.001 PBO 45 121 37.2 -- 18 121 14.9 -- 41 121 33.9 -- Biologic-Experienced (Population Not Studied) Week 8 Overall – Biologic-Naïve ACT 2 IFX 5 mg/kg 78 121 64.5 p<0.001 41 121 33.9 p<0.001 73 121 60.3 p<0.001 IFX IFX 10 mg/kg 83 120 69.2 p<0.001 33 120 27.5 p<0.001 74 120 61.7 p<0.001 PBO 36 123 29.3 -- 7 123 5.7 -- 38 123 30.9 -- Biologic-Experienced (Population Not Studied) Week 8 Overall – Biologic-Naïve Jiang 2015 IFX 3.5 mg/kg 30 41 73.1 p=0.001 21 41 51.2 p=0.006 23 41 56.1 p=0.003 IFX IFX 5 mg/kg 32 41 78.1 p=0.00 22 41 53.7 p=0.003 24 41 58.5 p=0.002 PBO 15 41 36.6 -- 9 41 21.9 -- 10 41 24.4 -- Biologic-Experienced (Population Not Studied) Week 8 Kobayashi Overall – Biologic-Naïve 2016 IFX 5 mg/kg 57 104 54.8 p=0.005 21 104 20.2 NS 48 104 46.2 p=0.006 IFX PBO 37 104 35.6 -- 11 104 10.6 -- 29 104 27.9 -- Biologic-Experienced (Population Not Studied) Week 8 NCT01551290 Overall – Biologic-Naïve IFX 5 mg/kg 32 50 64.0 p=0.0021 11 50 22.0 NS 17 50 34.0 p=0.045 IFX PBO 16 49 32.7 -- 5 49 10.2 -- 8 49 16.3 -- Biologic-Experienced (Population Not Studied) Week 8 Overall – Biologic-Naïve ULTRA 1 ADA 80/40 mg 67 130 51.5 NS 13 130 10 NS 49 130 37.7 NR ADA ADA 160/80 mg 71 130 54.6 NS 24 130 18.5 P=0.031 61 130 46.9 NR PBO 58 130 44.6 -- 12 130 9.2 -- 54 130 41.5 -- Biologic-Experienced (Population Not Studied) Week 8 ©Institute for Clinical and Economic Review, 2020 Page 166 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Induction Phase (Week 6-14) Trial Arm Response Remission Endoscopic Improvement n N % Significance n N % Significance n N % Significance Overall ADA 160/80 mg 125 248 50.4 p<0.005 41 248 16.5 p<0.05 102 248 41.1 p<0.05 PBO 85 246 34.6 -- 23 246 9.3 -- 78 246 31.7 -- ULTRA 2 Biologic-Naïve ADA 160/80 mg 89 150 59.3 p<0.001 32 150 21.3 p=0.017 74 150 49.3 p=0.014 ADA PBO 56 145 38.6 -- 16 145 11 -- 51 145 35.2 -- Biologic-Experienced ADA 160/80 mg 36 98 36.7 NS 9 98 9.2 NS 28 98 28.5 p=0.77 PBO 29 101 28.7 -- 7 101 6.9 -- 27 101 26.7 -- Week 8 Overall – Biologic-Naïve Suzuki 2014 ADA 80/40 mg 37 87 43 NS 12 87 14 NS 34 87 39 NS ADA ADA 160/80 mg 45 90 50 p=0.044 9 90 10 NS 40 90 44 p=0.045 PBO 34 96 35.4 -- 11 96 11.5 -- 29 96 30 - Biologic-Experienced (Population Not Studied) Week 6 Overall – Biologic-Naïve PURSUIT-SC GOL 200/100 mg 129 253 51.0 p<0.0001 45 253 17.8 p<0.0001 107 253 42.3 p<0.001 GOL GOL 400/200 mg 141 257 54.9 p<0.0001 46 257 17.9 p<0.0001 116 257 45.1 p=0.001 PBO 76 251 30.3 -- 16 251 6.4 -- 72 251 28.7 -- Biologic-Experienced (Population Not Studied) Week 6 Overall VEDO 300mg 106 225 47.1 p<0.001 38 225 16.9 p=0.001 92 225 40.9 p=0.001 PBO 38 149 25.5 -- 8 149 5.4 -- 37 149 24.8 -- GEMINI 1 Biologic-Naïve VEDO 300 mg 69 130 53.1 Diff. (95% CI): 30 130 23.1 Diff. (95% CI): 64 130 49.2 VEDO Diff. (95% CI): 26.4 (12.4 to 15.5 (5.1 to PBO 20 76 26.3 5 76 6.6 19 76 25.0 23.9 (10.0, 37.7) 40.4) 25.9) Biologic-Experienced VEDO 300 mg 32 82 39 Diff. (95% CI): 8 82 9.8 Diff. (95% CI): 25 82 30.5 Diff. (95% CI): PBO 13 63 20.6 18.1 (2.8, 33.5) 2 63 3.2 7.0 (-1.3, 15.2) 13 63 20.6 9.9 (-4.7,24.4) Week 10 Motoya 2019 Overall VEDO 300 mg 65 164 39.6 p=0.2722 30 164 18.3 p=0.1980 60 164 36.6 0.3168 VEDO PBO 27 82 32.9 -- 10 82 12.2 -- 25 82 30.5 -- Biologic-Naïve ©Institute for Clinical and Economic Review, 2020 Page 167 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Induction Phase (Week 6-14) Trial Arm Response Remission Endoscopic Improvement n N % Significance n N % Significance n N % Significance VEDO 300 mg 42 79 53.2 Diff. (95% CI): 22 79 27.8 Diff. (95% CI): 38 79 48.1 Diff. (95% CI): 16.6 (-1.8 to 13.2 (-1.4 to PBO 15 41 36.6 6 41 14.6 13 41 31.7 16.4 (-1.6, 34.4) 35.0) 27.9) Biologic-Experienced VEDO 300 mg 23 85 27.1 8 85 9.4 Diff. (95% CI): 22 85 25.9 Diff. (95% CI): Diff. (95% CI): -0.3 (-11.3, -3.4 (-20.14, PBO 12 41 29.3 -2.2 (-19.0, 41.6) 4 41 9.8 12 41 29.3 10.7) 13.37) Week 14 Overall VEDO 300 mg 257 383 67.1 Diff. (95% CI): 102 383 26.6 Diff. (95% CI): NR NR NR NR ADA 40 mg 177 386 45.9 21.2 (14.4, 28.0) 82 386 21.2 5.4 (-0.7, 11.4) NR NR NR NR VARSITY Biologic-Naïve VEDO vs. VEDO 300 mg 213 304 70.1 Diff. (95% CI): 84 304 27.6 Diff. (95% CI): NR NR NR NR ADA ADA 40 mg 151 305 49.5 20.6 (12.9, 28.2) 72 305 23.6 4.0 (-2.9, 10.9) NR NR NR NR Biologic-Experienced VEDO 300 mg 44 79 55.7 18 79 22.8 Diff. (95% CI): NR NR NR NR Diff. (95% CI): 10.3 (-1.5, ADA 40 mg 26 81 32.1 23.6 (8.5, 38.7) 10 81 12.3 NR NR NR NR 22.2) Week 8 Overall TOF 10 mg 285 476 59.9 p<0.001 88 476 18.5 p=0.007 149 476 31.3 p<0.001 PBO 40 122 32.8 -- 10 122 8.2 -- 19 122 15.6 -- OCTAVE 1 Biologic-Naïve TOF TOF 10 mg NR NR NR NR NR NR NR NR 88 222 39.6 p=0.06 PBO NR NR NR NR NR NR NR NR 15 57 26.3 -- Biologic-Experienced TOF 10 mg NR NR NR NR NR NR NR NR 61 254 24 p=0.001 PBO NR NR NR NR NR NR NR NR 4 65 6.2 -- Week 8 Overall TOF 10 mg 236 429 55 p<0.001 72 429 16.6 p<0.001 122 429 28.4 p<0.001 OCTAVE 2 PBO 32 112 28.6 -- 4 112 3.6 -- 13 112 11.6 -- Biologic-Naïve TOF TOF 10 mg NR NR NR NR NR NR NR NR 71 195 36.4 p=0.02 PBO NR NR NR NR NR NR NR NR 9 47 19.1 -- Biologic-Experienced TOF 10 mg NR NR NR NR NR NR NR NR 51 234 21.8 p=0.004 ©Institute for Clinical and Economic Review, 2020 Page 168 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Induction Phase (Week 6-14) Trial Arm Response Remission Endoscopic Improvement n N % Significance n N % Significance n N % Significance PBO NR NR NR NR NR NR NR NR 4 65 6.2 -- Week 8 Overall TOF 10 mg 521 905 57.6 -- 159 905 17.6 -- 271 905 29.9 p<0.001 OCTAVE 1 PBO 72 234 30.8 -- 14 234 6 -- 32 234 13.7 -- Biologic-Naïve and 2 TOF 10 mg 284 440 64.5 p<0.0001 106 440 24.1 p<0.01 168 440 38.2 p<0.01 Pooled PBO 43 110 39.1 -- 13 110 11.8 24 110 21.8 -- Biologic-Experienced TOF 10 mg 237 465 51 p<0.0001 53 465 11.4 p<0.01 103 465 22.0 p<0.001 PBO 29 124 23.4 -- 1 124 0.8 -- 8 124 6.1 -- Week 8 Overall UST 130 mg 164 320 51.3 p<0.001 50 320 15.6 p<0.001 65 320 26.3 p<0.001 UST 6 mg/kg 199 322 61.8 p<0.001 50 322 15.5 p<0.001 59 322 27.0 p<0.001 PBO 100 319 31.3 -- 17 319 5.3 -- 28 319 13.8 - UNIFI Biologic-Naïve UST 130 mg 90 156 57.7 p<0.001 31 156 19.9 p=0.009 54 156 34.6 p=0.006 UST UST 6 mg/kg 104 156 66.7 p<0.001 29 156 18.6 p=0.022 52 156 33.3 p=0.01 PBO 56 158 35.4 -- 15 158 9.5 -- 33 158 20.9 -- Biologic-Experienced UST 130 mg 74 164 45.1 p<0.001 19 164 11.6 p<0.001 30 164 18.3 p=0.002 UST 6 mg/kg 95 166 57.2 p<0.001 21 166 12.7 p<0.001 35 166 21.1 p<0.001 PBO 44 161 27.3 -- 2 161 1.2 -- 11 161 6.8 -- ADA: adalimumab, GOL: golimumab, IFX: infliximab, mg: milligram, mg/kg: milligram per kilogram, n: number, N: total number, NR: not reported, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab *Available results reported in published trials; for data used in our NMA refer to Appendix Tables F3-F5. ©Institute for Clinical and Economic Review, 2020 Page 169 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D6. Response and Remission in the Maintenance Phase* Maintenance Phase (Week 30-60) Response Sustained Response† Remission Sustained Remission Trial Arm n N % n N % n N % n N % Week 54 Overall – Biologic-Naïve ACT 1 IFX 5 mg/kg 55 121 45.5 47 121 38.8 42 121 34.7 24 121 19.8 IFX IFX 10 mg/kg 54 122 44.3 45 122 36.9 42 122 34.4 25 122 20.5 PBO 24 121 19.8 17 121 14.0 20 121 16.5 8 121 6.6 Biologic-Experienced (Population Not Studied) Week 30 Overall – Biologic-Naïve ACT 2 IFX 5 mg/kg 57 121 47.1 50 121 41.3 31 121 25.6 18 121 14.9 IFX IFX 10 mg/kg 72 120 60 64 120 53.3 43 120 35.8 27 120 22.5 PBO 32 123 26 19 123 15.4 13 123 10.6 3 123 2.4 Biologic-Experienced (Population Not Studied) Week 30 Overall – Biologic-Naïve Jiang 2015 IFX 3.5 mg/kg 26 41 63.4 NR 20 41 48.8 NR IFX IFX 5 mg/kg 27 41 65.8 NR 21 41 51.2 NR PBO 11 41 26.8 NR 10 41 24.4 NR Biologic-Experienced (Population Not Studied) Week 30 Kobayashi Overall – Biologic-Naïve 2016 IFX 5 mg/kg 48 104 46.2 NR 22 104 21.2 NR IFX PBO 33 104 31.7 NR 17 104 16.3 NR Biologic-Experienced (Population Not Studied) Week 26 NCT01551290 Overall – Biologic-Naïve IFX 5 mg/kg 29 50 58.0 27 50 54.0 14 50 28.0 7 50 14.0 IFX PBO 26 49 53.1 12 49 24.5 5 49 10.2 2 49 4.1 Biologic-Experienced (Population Not Studied) Week 52 Overall ULTRA 2 ADA 40 mg 75 248 30.2 59 248 23.8 43 248 17.3 21 248 8.5 ADA PBO 45 246 18.3 30 246 12.2 21 246 8.5 10 246 4.1 Biologic-Naïve ADA 40 mg 55 150 36.7 44 150 29.3 33 150 22.0 16 150 10.7 ©Institute for Clinical and Economic Review, 2020 Page 170 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance Phase (Week 30-60) PBO 35 145 24.1 24 145 16.6 18 145 12.4 9 145 6.2 Biologic-Experienced ADA 40 mg 20 98 20.4 15 98 15.3 10 98 10.2 5 98 5.1 PBO 10 101 9.9 6 101 5.9 3 101 3.0 1 101 1.0 Week 52 Suzuki 2014 Overall – Biologic-Naive ADA 40 mg 55 177 31 50 82 61 41 177 23 NR NR NR ADA PBO 17 96 18 NR NR NR 7 96 7 NR NR NR Biologic-Experienced (Population Not Studied) Week 54 Overall – Biologic-Naive PURSUIT-M GOL 50 mg 71 151 47 71 151 47 35 151 23.2 19 151 12.6 GOL GOL 100 mg 75 151 49.7 75 151 49.7 42 151 27.8 21 151 13.9 PBO 48 154 31.2 48 154 31.2 24 154 15.6 13 154 8.4 Biologic-Experienced (Population Not Studied) Week 54 PURSUIT-J Overall – Biologic-Naive GOL 100 mg 18 32 56.3 18 32 56.3 16 32 50.0 9 32 28.1 GOL PBO 6 31 19.4 6 31 19.4 2 31 6.5 2 31 6.5 Biologic-Experienced (Population Not Studied) Week 52 Overall VEDO 108 mg (SC) 68 106 64.2 68 106 64.2 49 106 46.2 16 106 15.1 VEDO 300 mg (iv) 39 54 72.2 39 54 72.2 23 54 42.6 10 54 16.7 PBO 17 56 28.6 17 56 28.6 8 56 14.3 3 56 5.4 VISIBLE 1 Biologic-Naive VEDO 108 mg (SC) NR NR 36 67 53.7 NR VEDO VEDO 300 mg (iv) NR NR 17 32 53.1 NR PBO NR NR 7 37 18.9 NR Biologic-Experienced VEDO 108 mg (SC) NR NR 13 39 33.3 NR VEDO 300 mg (iv) NR NR 6 22 27.3 NR PBO NR NR 1 19 5.3 NR Week 52 Overall GEMINI 1 VEDO 300 mg q4w 65 125 52 65 125 52 56 125 44.8 30 125 24 VEDO VEDO 300 mg q8w 69 122 56.6 69 122 56.6 51 122 41.8 25 122 20.5 PBO 30 126 23.8 30 126 23.8 20 126 15.9 11 126 8.7 Biologic-Naive ©Institute for Clinical and Economic Review, 2020 Page 171 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance Phase (Week 30-60) VEDO 300 mg q4w 41 73 56.2 41 73 56.2 35 73 47.9 21 73 28.8 VEDO 300 mg q8w 47 72 65.3 47 72 65.3 33 72 45.8 16 72 22.2 PBO 21 79 26.6 21 79 26.6 15 79 19.0 10 79 12.7 Biologic-Experienced VEDO 300 mg q4w 17 40 42.5 17 40 42.5 14 40 35.0 5 40 12.5 VEDO 300 mg q8w 20 43 46.5 20 43 46.5 16 43 37.2 9 43 20.9 PBO 6 38 15.8 6 38 15.8 2 38 5.3 1 38 2.6 Week 60 Overall VEDO 300 mg 27 41 65.9 27 41 65.9 23 41 56.1 11 41 26.8 PBO 15 42 35.7 15 42 35.7 13 42 31 7 42 16.7 Motoya 2019 Biologic-Naive VEDO VEDO 300 mg 16 24 66.7 16 24 66.7 13 24 54.2 3 17 17.6 PBO 10 28 35.7 10 28 35.7 10 28 35.7 1 14 7.1 Biologic-Experienced VEDO 300 mg 11 17 64.7 11 17 64.7 10 17 58.8 3 17 17.6 PBO 5 14 35.7 5 14 35.7 3 14 21.4 1 14 7.1 Week 52 Overall VEDO 300 mg 211 383 55.1 NR NR NR 120 383 31.3 70 383 18.3 VARSITY ADA 40 mg 166 386 43.0 NR NR NR 87 386 22.5 46 386 11.9 Biologic-Naive VEDO vs. VEDO 300 mg NR NR 104 304 34.2 NR ADA ADA 40 mg NR NR 74 305 24.3 NR Biologic-Experienced VEDO 300 mg NR NR 16 79 20.3 NR ADA 40 mg NR NR 13 81 16.0 NR Week 52 Overall TOF 5 mg 102 198 51.5 97 198 49 68 198 34.3 44 198 22.2 TOF 10 mg 122 197 61.9 117 197 59.4 80 197 40.6 50 197 25.4 OCTAVE PBO 40 198 20.2 38 198 19.2 22 198 11.1 10 198 5.1 Biologic-Naive SUSTAIN TOF 5 mg 65 115 56.5 65 115 56.5 48 115 41.7 NR NR NR TOF TOF 10 mg 67 104 64.4 67 104 64.4 46 104 44.2 NR NR NR PBO 27 109 24.8 27 109 24.8 12 109 11.0 NR NR NR Biologic-Experienced TOF 5 mg 37 83 44.6 37 83 44.6 20 83 24.1 NR NR NR TOF 10 mg 55 93 59.1 55 93 59.1 34 93 36.6 NR NR NR ©Institute for Clinical and Economic Review, 2020 Page 172 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance Phase (Week 30-60) PBO 13 89 14.6 13 89 14.6 10 89 11.2 NR NR NR Week 52 Overall UST 90 mg q12w 117 172 68.0 117 172 68.0 66 172 38.4 26 40 65 UST 90 mg q8w 125 176 71.0 125 176 71.0 77 176 43.8 22 38 58 PBO 78 175 44.6 78 175 44.6 42 175 24 17 45 38 Biologic-Naïve UST 90 mg q12w 78 102 76.5 78 102 76.5 50 102 49.0 21 30 70.0 UNIFI UST 90 mg q8w 66 85 77.6 66 85 77.6 41 85 48.2 12 16 75.0 UST PBO 44 87 50.6 44 87 50.6 27 87 31.0 9 25 36.0 Biologic-Experienced UST 90 mg q12w 39 70 55.7 39 70 55.7 16 70 22.9 23 32 71.9 UST 90 mg q8w 59 91 64.8 59 91 64.8 36 91 39.6 12 18 66.7 PBO 34 88 38.6 34 88 38.6 15 88 17.0 9 25 36.0 ADA: adalimumab, GOL: golimumab, IFX: infliximab, mg: milligram, mg/kg: milligram per kilogram, n: number, N: total number, NR: not reported, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab *Available results reported in published trials; for data used in our NMA refer to Appendix Table F6. †Sustained response in treat-through trials is defined as having response at end of induction and end of maintenance. Note that for re-randomized trials, the rates of response are among induction responders, so "response" and "sustained response" rates in this table are equivalent. ©Institute for Clinical and Economic Review, 2020 Page 173 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D7. Delayed Response Among Non-Responders to Induction Therapy Prior Biologic Response Remission Trial Treatment Arms Week Exposure % % ULTRA 2 Overall Adalimumab 16 NR 4.1 Biologic-naïve 56.2 18.8 OCTAVE 1 and 2 Biologic-experienced Tofacitinib 16 42.6 39.7 Overall 50.1 13.9 Biologic-naïve 79.1 18.6 UNIFI Ustekinumab 6 mg/kg 16 Biologic-experienced 43.1 1.7 Vedolizumab 39.0 NR GEMINI 1 Overall 14 Placebo 20.7 NR PURSUIT-SC Naïve Golimumab 16 28.1 NR kg: kilogram, mg: milligram, NR: not reported ©Institute for Clinical and Economic Review, 2020 Page 174 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D8. IBDQ Outcomes in the Induction Phase Induction Phase (Week 6-14) IBDQ Score IBDQ Response / Remission* Trial Arm Data Data Type Value Type Value p-value n N % p-value (Estimate) (Error) Week 8 ACT 1-2 Overall IFX 5 mg/kg Mean ∆ 40 SD 34 <0.05 169 242 69.7 <0.05 POOLED IFX 10 mg/kg Mean ∆ 36 SD 34 <0.05 164 242 67.8 <0.05 Infliximab PBO Mean ∆ 21 SD 28 -- 121 244 49.6 -- Stratified Biologic-Naïve and Experienced Data Not Reported Week 8 Overall ADA 160/80 mg NR 144 248 58.1 0.006 PBO NR 112 246 45.5 -- ULTRA 2 Biologic-Naïve Adalimumab ADA 160/80 mg NR 102 150 68 0.004 PBO NR 75 145 51.7 -- Biologic-Experienced ADA 160/80 mg NR 42 98 42.9 0.370 PBO NR 37 101 36.6 -- Week 8 Overall – Biologic-Naive Suzuki 2014 ADA 80/40 mg NR 42 87 48.3 -- Adalimumab ADA 160/80 mg NR 38 90 42.2 -- PBO NR 38 96 39.6 -- Biologic-Experienced (Population Not Studied) Week 6 Overall – Biologic-Naive PURSUIT-SC GOL 200/100 mg Mean ∆ 27.0 SD 33.72 <0.001 NR Golimumab GOL 400/200 mg Mean ∆ 26.9 SD 34.28 <0.001 NR PBO Mean ∆ 14.8 SD 31.25 -- NR Biologic-Experienced (Population Not Studied) Week 8 OCTAVE 1 Overall Tofacitinib Remission TOF 10 mg Mean ∆ 40.7 SE 1.7 <0.0001 206 476 43.3 <0.001 ©Institute for Clinical and Economic Review, 2020 Page 175 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Induction Phase (Week 6-14) IBDQ Score IBDQ Response / Remission* Trial Arm Data Data Type Value Type Value p-value n N % p-value (Estimate) (Error) PBO Mean ∆ 21 SE 2.9 -- 32 122 26.2 -- Response TOF 10 mg 307 476 64.5 <0.001 -- PBO 56 122 45.9 -- Stratified Biologic-Naïve and Experienced Data Not Reported Week 8 Overall Remission OCTAVE 2 TOF 10 mg Mean ∆ 44.6 SE 1.8 <0.0001 173 429 40.3 <0.001 PBO Mean ∆ 25 SE 3.3 - 20 112 17.9 -- Tofacitinib Response TOF 10 mg 288 429 67.1 <0.001 -- PBO 54 112 48.2 -- Stratified Biologic-Naïve and Experienced Data Not Reported Week 8 Overall Response UST 130 mg Mean ∆ 33.4 SD 32.5 <0.001 213 320 66.6 <0.001 UNIFI UST 6 mg/kg Mean ∆ 35.0 SD 31.9 <0.001 221 322 68.6 <0.001 PBO Mean ∆ 16.1 SD 31.4 -- 141 319 44.2 - Ustekinumab UST 130 mg 141 320 44.2 <0.001 UST 6 mg/kg -- 150 322 46.6 <0.001 PBO 101 319 31.7 -- Stratified Biologic-Naïve and Experienced Data Not Reported ADA: adalimumab, GOL: golimumab, IBDQ: inflammatory bowel disease questionnaire, IFX: infliximab, IQR: interquartile range, LS: least squares, mg: milligram, n: number, N: total number, NR: not reported, PBO: placebo, QoL: quality of life, SD: standard deviation, SE: standard error, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab *Rates of IBDQ response (change ≥ 16 points) are reported unless otherwise noted that rates are for IBDQ remission (score ≥ 170). ©Institute for Clinical and Economic Review, 2020 Page 176 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D9. 36-Item Short Form Survey Outcomes in the Induction Phase Induction (Week 6-14) SF-36 (Mental) SF-36 (Physical) Trial Arm Data Data Data Type Data Type Value Type Value p-value Value Type Value p-value (Estimate) (Estimate) (Error) (Error) Week 8 ACT 1-2 Overall – Biologic-Naive IFX 5 mg/kg Mean ∆ 3.0 SD 9.6 <0.05 Mean ∆ 3.7 SD 6.5 <0.05 POOLED QoL IFX 10 mg/kg Mean ∆ 5.9 SD 10.5 <0.05 Mean ∆ 6.8 SD 7.6 <0.05 Infliximab PBO Mean ∆ 6.1 SD 10.9 --- Mean ∆ 6.4 SD 7.7 --- Biologic-Experienced (Population not Studied) Week 8 OCTAVE 1 Overall TOF 10 mg Mean ∆ 6.8 SE 0.5 Mean ∆ 6.8 SE 0.2 Tofacitinib <0.0001 <0.0001 PBO Mean ∆ 3.5 SE 0.9 Mean ∆ 2.5 SE 0.5 Stratified Biologic-Naïve and Experienced Data Not Reported Week 8 OCTAVE 2 Overall TOF 10 mg Mean ∆ 7.6 SE 0.5 Mean ∆ 6.8 SE 0.3 Tofacitinib PBO Mean ∆ 4.4 SE 0.9 <0.01 Mean ∆ 4.6 SE 0.5 <0.01 Stratified Biologic-Naïve and Experienced Data Not Reported Overall NR OCTAVE 1 Biologic-Naive TOF 10 mg Mean ∆ 8.3 SE 0.6 <0.001 Mean ∆ 7.3 SE 0.4 <0.001 and 2 PBO Mean ∆ 4.1 SE 1 -- Mean ∆ 4.7 SE 0.7 -- POOLED Biologic-Experienced TOF 10 mg Mean ∆ 6 SE 0.5 <0.001 Mean ∆ 6.2 SE 0.3 <0.001 PBO Mean ∆ 3.4 SE 0.9 -- Mean ∆ 2.3 SE 0.6 -- ADA: adalimumab, GOL: golimumab, IFX: infliximab, IQR: interquartile range, LS: least squares, mg: milligram, n: number, N: total number, NR: not reported, NS: not significant, PBO: placebo, QoL: quality of life, SD: standard deviation, SE: standard error, SF-36: 36-item short form survey, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 177 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D10. Endoscopic Improvement and Corticosteroid-Free Remission in the Maintenance Phase Maintenance (Week 30-60) Trial Arm Endoscopic Improvement Corticosteroid-Free Remission n N % Significance n N % Significance Week 54 Overall – Biologic-Naïve ACT 1 IFX 5 mg/kg 55 121 45.5 <0.001 18 70 25.7 0.006 Infliximab IFX 10 mg/kg 57 122 46.7 <0.001 12 73 16.4 0.15 PBO 22 121 18.2 -- 7 79 8.9 -- Biologic-Experienced (Population Not Studied) Week 30 Overall – Biologic-Naïve ACT 2 IFX 5 mg/kg 56 121 46.3 0.009 11 60 18.3 0.010 Infliximab IFX 10 mg/kg 68 120 56.7 <0.001 18 66 27.3 <0.001 PBO 37 123 30.1 -- 2 60 3.3 -- Biologic-Experienced (Population Not Studied) Week 30 Overall – Biologic-Naïve Jiang 2015 IFX 3.5 mg/kg 21 41 51.2 0.006 NR NR NR NR Infliximab IFX 5 mg/kg 22 41 53.7 0.003 NR NR NR NR PBO 9 41 21.9 -- NR NR NR NR Biologic-Experienced (Population Not Studied) Week 30 Kobayashi 2016 Overall – Biologic-Naïve IFX 5 mg/kg 43 104 41.3 0.057 NR NR NR NR Infliximab PBO 30 104 28.8 -- NR NR NR NR Biologic-Experienced (Population Not Studied) Week 26 NCT01551290 Overall – Biologic-Naïve IFX 5 mg/kg 20 50 40.0 0.1781 5 30 16.7 0.0428 Infliximab PBO 13 49 26.5 -- 1 39 2.6 -- Biologic-Experienced (Population Not Studied) Week 52 Overall ULTRA 2 ADA 40 mg 71 248 25 <0.05 20 248 13.3 0.04 Adalimumab PBO 38 246 15.4 -- 8 246 5.7 -- Overall – Biologic-Naïve ADA 40 mg 47 150 31.3 0.02 15 150 13.6 0.09 ©Institute for Clinical and Economic Review, 2020 Page 178 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance (Week 30-60) Trial Arm Endoscopic Improvement Corticosteroid-Free Remission n N % Significance n N % Significance PBO 28 145 19.3 -- 5 145 6.2 -- Biologic-Experienced ADA 40 mg 15 98 15.3 0.25 5 98 12.5 0.26 PBO 10 101 9.9 -- 3 101 5.1 -- Week 52 Overall – Biologic-Naïve Suzuki 2014 ADA 40 mg 51 177 29 0.02 17 177 14.2 -- Adalimumab PBO 15 96 16 -- 4 96 6.9 -- Biologic-Experienced (Population Not Studied) Week 54 Overall – Biologic-Naïve PURSUIT-M GOL 50 mg 63 151 41.7 0.011 30 78 28.2 0.003 Golimumab GOL 100 mg 64 151 42.4 0.002 25 82 23.2 0.14 PBO 41 154 26.6 -- 18 87 18.4 -- Biologic-Experienced (Population Not Studied) Week 54* PURSUIT-J Overall – Biologic-Naïve GOL 100 mg 20 32 62.5 NR 5 9 55.6 NR Golimumab PBO 5 16 31.3 -- 1 9 11.1 -- Biologic-Experienced (Population Not Studied) Week 52 Overall VEDO 108 mg VISIBLE 1 60 106 56.6 p<0.001 31 106 28.9 p=0.067 (SC) Vedolizumab VEDO 300 mg 29 54 53.7 NR 6 21 28.6 NR (IV) PBO 41 56 21.4 -- 2 24 8.3 -- Stratified Biologic-Naïve and Experienced Data Not Reported Week 52 Overall VEDO 300 mg GEMINI 1 70 125 56 p<0.001 33 73 45.2 p<0.001 q4w Vedolizumab VEDO 300 mg 63 122 51.6 p<0.001 22 70 31.4 p=0.0100 q8w PBO 25 126 19.8 -- 10 72 13.9 -- Biologic-Naïve ©Institute for Clinical and Economic Review, 2020 Page 179 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance (Week 30-60) Trial Arm Endoscopic Improvement Corticosteroid-Free Remission n N % Significance n N % Significance Diff (95% CI): Diff (95% CI): VEDO 300 mg 44 73 60.3 35.5 (19.9, 23 73 52.3 33.7 (13.8, q4w 51.0) 53.6) Diff (95% CI): Diff (95% CI): VEDO 300 mg 43 72 59.7 35.4 (19.8, 14 72 35.9 17.0 q8w 51.1) (-2.0, 36.0) PBO 19 79 21.1 -- 8 43 18.6 -- Biologic-Experienced Diff (95% CI): Diff (95% CI): VEDO 300 mg 19 40 47.5 39.3 (19.3, 23 73 52.3 33.7 (13.8, q4w 59.3) 53.6) Diff (95% CI): VEDO 300 mg Diff (95% CI): 29.8 18 43 41.9 14 72 35.9 17.0 q8w (11.6, 48.1) (-2.0, 36.0) PBO 3 381 7.9 -- 8 43 18.6 -- Week 60 Overall VEDO 300 mg 26 41 63.4 0.006 NR NR NR NR PBO 14 42 33.3 -- NR NR NR NR Motoya 2019 Biologic-Naïve Vedolizumab VEDO 300 mg 15 24 62.5 Diff. (95% CI): 4 9 44.4 Diff (95% CI): 22.2 PBO 10 28 35.7 26.8 (0.52, 53.1) 2 9 22.2 (-20.1, 64.6) Biologic-Experienced VEDO 300 mg 11 17 64.7 Diff. (95% CI): 2 4 50 Diff (95% CI): 33.3 PBO 4 14 28.6 36.1 (3.3, 68.9) 1 2 16.7 (-24.0, 90.7) Week 52 Overall VEDO 300 mg 152 383 39.7 Diff. (95% CI): 11.9 14 383 12.6 Diff. (95% CI): ADA 40 mg 107 386 27.7 (5.3, 18.5) 26 386 21.8 -9.3 (-18.9, 0.4) VARSITY Biologic-Naïve VEDO vs. ADA VEDO 300 mg 131 304 43.1 Diff. (95% CI): 13.6 13 87 14.9 Diff (95% CI): -6.8 ADA 40 mg 90 305 29.5 (6.0, 21.2) 20 92 21.7 (-18.1, 4.5) Biologic-Experienced VEDO 300 mg 21 79 26.6 Diff. (95% CI): 5.5 1 24 4.2 Diff (95% CI): -18.1 ADA 40 mg 17 81 21 (-7.7, 18.8) 6 27 22.2 (-44.2, 10.0) Week 52 Overall ©Institute for Clinical and Economic Review, 2020 Page 180 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Maintenance (Week 30-60) Trial Arm Endoscopic Improvement Corticosteroid-Free Remission n N % Significance n N % Significance TOF 5 mg 74 198 37.4 <0.001 23 65 35.4 <0.001 TOF 10 mg 90 196 45.7 <0.001 26 55 47.3 <0.001 PBO 26 198 13.1 -- 3 59 5.1 -- Biologic-Naïve OCTAVE SUSTAIN TOF 5 mg 49 115 42.6 <0.001 19 47 40.4 <0.01 TOF 10 mg 53 104 51 <0.001 19 37 51.4 <0.001 Tofacitinib PBO 15 109 13.8 -- 2 38 5.3 -- Biologic-Experienced TOF 5 mg 25 83 30.1 <0.01 4 18 22.2 -- TOF 10 mg 37 93 39.8 <0.001 7 18 38.9 <0.01 PBO 11 89 12.4 -- 1 21 4.8 -- Week 52 Overall UST 90 mg 88 172 51.1 <0.001 65 172 37.8 0.0020 q12w UST 90 mg 77 176 43.6 <0.001 74 176 42 <0.001 q8w PBO 50 175 28.6 -- 41 175 23.4 -- Biologic-Naïve UST 90 mg UNIFI 57 102 55.9 0.007 49 102 48.0 0.028 q12w Ustekinumab UST 90 mg 49 85 57.6 0.002 40 85 47.1 0.034 q8w PBO 30 87 34.5 -- 27 87 31.0 -- Biologic-Experienced UST 90 mg 18 70 25.7 p=0.163 16 70 22.9 0.026 q12w UST 90 mg 41 91 45.1 p<0.001 34 91 37.4 <0.001 q8w PBO 20 88 22.7 -- 14 88 15.9 -- ADA: adalimumab, GOL: golimumab, IFX: infliximab, IQR: interquartile range, IV: intravenous, LS: least squares, mg: milligram, n: number, N: total number, NR: not reported, NS: not significant, PBO: placebo, QoL: quality of life, SC: subcutaneous, SD: standard deviation, SE: standard error, SF-36: 36-item short form survey, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, ∆: difference ©Institute for Clinical and Economic Review, 2020 Page 181 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D11. IBDQ and EQ-5D Outcomes in the Maintenance Phase IBDQ Score IBDQ Response/Remission* EQ-5D (Visual Analogue Scale) Trial Arm Data Data p- p- Data Data p- Value Value n N % Value Value Type Type value value Type Type value Week 54 Overall – Biologic-Naïve ACT 1 IFX 5 mg/kg Mean ∆ 32.8 95% CI 25.8-38.8 <0.05 NR NR Infliximab IFX 10 mg/kg Mean ∆ 31.8 95% CI 24.8-38.8 <0.05 NR NR PBO Mean ∆ 12.8 95% CI 6.8-17.9 -- NR NR Biologic-Experienced (Population Not Studied) Week 30 Overall – Biologic-Naïve ACT 2 IFX 5 mg/kg Mean ∆ 31.9 95% CI 24.9-37.8 <0.05 NR NR Infliximab IFX 10 mg/kg Mean ∆ 36.1 95% CI 28.8-42.7 <0.05 NR NR PBO Mean ∆ 17.9 95% CI 11.8-22.8 -- NR NR Biologic-Experienced (Population Not Studied) Week 52 Overall ADA 40 mg NR 65 248 26.2 0.0070 NR ULTRA 2 PBO NR 40 246 16.3 -- Biologic-Naive Adalimumab ADA 40 mg NR 48 150 32.0 0.0390 NR PBO NR 31 145 21.4 -- NR Biologic-Experienced ADA 40 mg NR 17 98 17.3 0.078 NR PBO NR 9 101 8.9 -- NR Week 52 Suzuki 2014 Overall – Biologic-Naïve ADA 40 mg NR 45 177 25.4 <0.01 NR Adalimumab PBO NR 12 96 12.5 -- NR Biologic-Experienced (Population Not Studied) Week 54 PURSUIT J Overall – Biologic-Naïve GOL 100 mg NR 11 20 55.0 NR NR Golimumab PBO NR 6 27 22.2 -- NR Biologic-Experienced (Population Not Studied) Week 52 Overall ©Institute for Clinical and Economic Review, 2020 Page 182 Final Report – Targeted Immune Modulators for UC Return to Table of Contents IBDQ Score IBDQ Response/Remission* EQ-5D (Visual Analogue Scale) Trial Arm Data Data p- p- Data Data p- Value Value n N % Value Value Type Type value value Type Type value VEDO 108 Diff vs. Diff vs. 95% 11- < 43.9 95% CI 30.6-57.1 <.001 NR 17.6 mg (SC) PBO PBO CI 24.3 .001 VISIBLE 1 VEDO 300 Diff vs. Diff vs. 95% 5.5- 37.1 95% CI 21.9-52.4 <.001 NR 13.1 .001 Vedolizumab mg (iv) PBO PBO CI 20.8 PBO ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ Stratified Biologic-Naïve and Biologic-Experienced Data Not Reported Week 52 Overall Remission VEDO 300 Mean ∆ 49 SE 3.3 NR 85 125 68 NR Mean 19.4 SE 1.7 NR mg q4w VEDO 300 Mean Mean ∆ 48.4 SE 3.4 NR 72 122 59 NR 19 SE 1.7 NR mg q8w ∆ PBO Mean ∆ 27.3 SE 3.3 NR 48 126 38 NR Mean ∆ 9.7 SE 1.7 NR Biologic-Naive GEMINI 1 VEDO 300 Diff. vs Diff. vs 95% 5.5- 25.8 95% CI 14.7-36.9 NR NR 11.1 NR Vedolizumab mg q4w PBO PBO CI 16.7 VEDO 300 Diff. vs Diff. vs 95% 4.9- 25.9 95% CI 14.6-37.3 NR NR 10.6 NR mg q8w PBO PBO CI 16.3 PBO ─ ─ ─ ─ ─ NR ─ ─ ─ ─ ─ Biologic-Experienced VEDO 300 Diff. vs Diff. vs 95% -2.0- 13.4 95% CI -3.4-30.2 NR NR 6.9 NR mg q4w PBO PBO CI 15.7 VEDO 300 Diff. vs Diff. vs 95% -1.8- 14.1 95% CI -2.5-30.5 NR NR 6.8 NR mg q8w PBO PBO CI 15.5 PBO ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ Week 52 Overall Remission VEDO 300 NR 192 383 50.1 NR NR VARSITY mg ADA 40 mg NR 156 386 40.4 -- NR VEDO vs. ADA Response VEDO 300 ─ 199 383 52.0 NR NR mg ADA 40 mg 164 386 42.2 -- NR Stratified Biologic-Naïve and Biologic-Experienced Data Not Reported ©Institute for Clinical and Economic Review, 2020 Page 183 Final Report – Targeted Immune Modulators for UC Return to Table of Contents IBDQ Score IBDQ Response/Remission* EQ-5D (Visual Analogue Scale) Trial Arm Data Data p- p- Data Data p- Value Value n N % Value Value Type Type value value Type Type value Week 52 Overall Remission TOF 5 mg Mean ∆ 3.7 SE 3.4 <.0001 76 198 38.4 <.001 NR OCTAVE TOF 10 mg Mean ∆ 4.8 SE 3.2 <.0001 95 197 48.2 <.001 NR SUSTAIN PBO Mean ∆ -26.5 SE 3.8 ─ 29 198 14.6 ─ NR Tofacitinib Response TOF 5 mg 92 198 46.5 <.001 NR TOF 10 mg ─ 106 197 53.8 <.001 NR PBO 38 198 19.2 ─ NR Stratified Biologic-Naive and Experienced Not Reported Week 52 Overall Response UST 90 mg Mean ∆ -3.0 SD 32.9 NR 118 172 68.6 <.001 NR q12w UST 90 mg Mean ∆ 3.9 SD 31.5 NR 129 176 73.3 <.001 NR UNIFI q8w PBO Mean ∆ -15.1 SD 35.4 ─ 83 175 47.4 ─ NR Ustekinumab Remission UST 90 mg 97 172 56.4 .004 NR q12w UST 90 mg ─ 111 176 63.1 <.001 NR q8w PBO 72 175 41.1 ─ NR Stratified Biologic-Naïve and Experienced Data Not Reported ADA: adalimumab, EQ-5D: EuroQol - 5 dimension, GOL: golimumab, IBDQ: inflammatory bowel disease questionnaire, IFX: infliximab, IQR: interquartile range, IV: intravenous, LS: least squares, mg: milligram, n: number, N: total number, NR: not reported, NS: not significant, PBO: placebo, QoL: quality of life, SC: subcutaneous, SD: standard deviation, SE: standard error, SF-36: 36-item short form survey, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, 95% CI: 95% confidence interval, ∆: difference *Rates of IBDQ response (change ≥ 16 points, except for PURSUIT-J which used change >20 points) are reported unless otherwise noted that rates are for IBDQ remission (score ≥ 170). ©Institute for Clinical and Economic Review, 2020 Page 184 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D12. 36-Item Short Form Survey Outcomes in the Maintenance Phase SF-36 (Mental) SF-36 (Physical) Trial Arm Data Value Data Type Value p-value Data Type Value Data Type Value p-value Type Week 52 Biologic-Naive Diff. vs VEDO 300 mg q4w 6.2 95% CI 3.2-9.1 NR Diff. vs PBO 3.6 95% CI 1.4-5.8 NR PBO Diff. vs VEDO 300 mg q8w 6 95% CI 2.9-9.0 NR Diff. vs PBO 3.9 95% CI 1.7-6.2 NR GEMINI 1 PBO PBO -- -- -- -- -- -- -- -- -- -- Vedolizumab Biologic- Experienced Diff. vs VEDO 300 mg q4w 2.2 95% CI -2.3-6.7 NR Diff. vs PBO 1.1 95% CI -2.1-4.4 NR PBO Diff. vs VEDO 300 mg q8w 3.3 95% CI -1.2-7.8 NR Diff. vs PBO 2.2 95% CI -1.0-5.4 NR PBO PBO -- -- -- -- -- -- -- -- -- -- Week 52 Overall TOF 5 mg Mean ∆ -1 SE 0.9 <0.0001 Mean ∆ 0 SE 0.8000 <0.0001 TOF 10 mg Mean ∆ 0.1 SE 1.9 <0.0001 Mean ∆ 0.3 SE 0.7000 <0.0001 PBO Mean ∆ -6.7 SE 1.2 ─ Mean ∆ -5.2 SE 0.9000 ─ OCTAVE Biologic-Naive SUSTAIN TOF 5 mg Mean ∆ -1.7 SE 1.1 <0.001 Mean ∆ -1.2 SE 0.9 <0.001 Tofacitinib TOF 10 mg Mean ∆ -1 SE 1.1 <0.0001 Mean ∆ -0.2 SE 0.9 <0.001 PBO Mean ∆ -8.5 SE 1.5 -- Mean ∆ -5.9 SE 1.1 -- Biologic-Experienced TOF 5 mg Mean ∆ -2.2 SE 1.2 <0.01 Mean ∆ -0.9 SE 1 <0.001 TOF 10 mg Mean ∆ -0.8 SE 1.1 <0.001 Mean ∆ -1.2 SE 0.9 <0.001 PBO Mean ∆ -6.7 SE 1.5 -- Mean ∆ -6.1 SE 1.2 -- ADA: adalimumab, GOL: golimumab, IFX: infliximab, IQR: interquartile range, IV: intravenous, LS: least squares, mg: milligram, n: number, N: total number, NR: not reported, NS: not significant, PBO: placebo, QoL: quality of life, SC: subcutaneous, SD: standard deviation, SE: standard error, SF-36: 36-item short form survey, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, ∆: difference ©Institute for Clinical and Economic Review, 2020 Page 185 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D13. Induction and Maintenance WPAI Questionnaire I* WPAI (Absenteeism) WPAI (Presenteeism) Trial Arm Data Data Data Value Value p-value Data Type Value Value p-value Type Type Type Induction – Week 8 (Data Not Reported) ULTRA 2 Maintenance – Week 52 Adalimumab ADA 160/80 mg Mean 7 NR NR NR Mean 21 NR NR NR PBO NR NR NR NR NR NR NR NR NR NR OCTAVE 1 Induction only – Week 8 TOF 10 mg Mean ∆ -11.2 95% CI -9.9, 1.6 0.16 Mean -22.1 95% CI -19.8, -6.0 0.003 Tofacitinib PBO Mean ∆ -7.1 -- -- -- Mean -9.2 -- -- -- OCTAVE 2 Induction only – Week 8 TOF 10 mg Mean ∆ -7.3 95% CI -4.4, 8.5 0.53 Mean -18.6 95% CI -12.0, 2.2 0.18 Tofacitinib PBO Mean ∆ -9.3 -- -- -- Mean -13.7 -- -- -- OCTAVE Maintenance only – Week 52 SUSTAIN TOF 5 mg Mean ∆ -4.5 95% CI -12.2, 1.0 0.09 Mean ∆ -3.6 95% CI -18.9, -2.8 0.008 TOF 10 mg Mean ∆ -3.1 95% CI -10.7, 2.4 0.21 Mean ∆ -4.3 95% CI -19.5, -3.4 0.005 Tofacitinib PBO Mean ∆ 1.1 -- -- -- Mean ∆ 7.2 -- -- -- Induction – Week 8 UST 130 mg Mean ∆ -5.9 SD 31.39 <0.05 Mean ∆ -15.1 SD 29.17 <0.05 UST 6 mg/kg Mean ∆ -9.1 SD 23.84 <0.01 Mean ∆ -20.4 SD 24.11 <0.001 UNIFI PBO Mean ∆ -3.7 SD 30.41 -- Mean ∆ -6.9 SD 21.89 -- Ustekinumab Maintenance – Week 52 UST 90 mg q12w Mean ∆ -2 NR NR 0.133 Mean ∆ -1.6 NR NR 0.017 UST 90 mg q8w Mean ∆ -2.1 NR NR 0.172 Mean ∆ -6.4 NR NR <0.001 PBO Mean ∆ 4.7 NR NR -- Mean ∆ 7.4 NR NR -- ADA: adalimumab, diff: difference, IV: intravenous, kg: kilogram, mg: milligram, n: number, N: total number, NA: not applicable, NR: not reported, PBO: placebo, q8w: every 8 weeks, q12w: every 12 weeks, SC: subcutaneous, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, WPAI: work productivity and activity impairment questionnaire All other trials did not report WPAI. *Stratified biologic-naïve and experienced data not reported. ©Institute for Clinical and Economic Review, 2020 Page 186 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D14. Induction and Maintenance WPAI Questionnaire II* WPAI (Work Productivity Loss) WPAI (Non-Work Productivity Loss) Trial Arm Data Type Value Data Type Value p-value Data Type Value Data Type Value p-value Induction – Week 8 (Data Not Reported) ULTRA 2 Maintenance – Week 52 Adalimumab ADA 160/80 mg Mean 24 NR NR NR Mean 23 NR NR NR PBO NR NR NR NR NR NR NR NR NR NR OCTAVE 1 Induction only – Week 8 TOF 10 mg Mean ∆ -19.1 95% CI -19.1, -2.1 0.0143 Mean ∆ -25.4 95% CI -19, -8.9 <0.0001 Tofacitinib PBO Mean ∆ -8.5 - -- -- Mean ∆ -11.5 -- -- -- OCTAVE 2 Induction only – Week 8 TOF 10 mg Mean ∆ -14.7 95% CI -11.9, 4.9 0.412 Mean -24.0 95% CI -17.2, -6.4 <0.0001 Tofacitinib PBO Mean ∆ -11.2 -- -- -- Mean -12.2 -- -- -- OCTAVE Maintenance only – Week 52 TOF 5 mg Mean ∆ -3.4 95% CI -17.8, 9.0 0.519 Mean ∆ -2.8 95% CI -20.6, -7.5 <0.0001 SUSTAIN TOF 10 mg Mean ∆ -6.6 95% CI -20.6, -5.4 0.253 Mean ∆ -3.1 95% CI -20.8, -5.4 <0.0001 Tofacitinib PBO Mean ∆ 1.0 -- -17.8, 9.0 0.519 Mean ∆ 11.3 -- -- -- Induction – Week 8 UST 130 mg Mean ∆ -17.2 SD 30.36 <0.01 Mean ∆ -17.7 SD 29.45 <0.01 UST 6 mg/kg Mean ∆ -21.8 SD 26.26 <0.001 Mean ∆ -20.8 SD 26.27 <0.001 UNIFI PBO Mean ∆ -8 SD 24.83 -- Mean ∆ -10.9 SD 28.66 -- Ustekinumab Maintenance – Week 52 UST 90 mg q12w Mean ∆ -2.2 NR NR 0.013 Mean ∆ 0.8 NR NR 0.002 UST 90 mg q8w Mean ∆ -6.1 NR NR <0.001 Mean ∆ -4.2 NR NR <0.001 PBO Mean ∆ 7.7 NR NR -- Mean ∆ 9.3 NR NR -- Maintenance only – Week 52 VISIBLE -27.9, -33.2, VEDO 300 mg (IV) Mean 27 95% CI 0.04 Mean 22 95% CI <0.001 Vedolizumab -0.6 -13.2 PBO Mean 39 -- -- Mean 45 -- -- -- ADA: adalimumab, diff: difference, IV: intravenous, kg: kilogram, mg: milligram, n: number, N: total number, NA: not applicable, NR: not reported, PBO: placebo, q8w: every 8 weeks, q12w: every 12 weeks, SC: subcutaneous, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, WPAI: work productivity and activity impairment questionnaire All other trials did not report WPAI. *Stratified biologic-naïve and experienced data not reported. ©Institute for Clinical and Economic Review, 2020 Page 187 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D15. Safety in the Induction Phase Infusion/ Any Related D/C due Serious Severe Anti- Serious Trial Arm Death Injection Site Infections TB AE AE to AE AE AE bodies Infections Reaction Kobayashi IFX 5 mg/kg 81.7 NR 4.8 NR 8.7 NR NR 10.6 31.7 1 NR 2016 PBO 82.7 NR 7.7 NR 12.5 NR NR 8.7 33.7 1.9 NR ADA 80/40 mg 53.8 NR 6.2 0 3.8 6.9 NR 5.4 20 1.5 NR ULTRA 1 ADA 160/80 mg 50.2 NR 5.4 0 4 8.5 NR 5.8 14.3 0 NR PBO 48.4 NR 5.4 0 7.6 7.6 NR 3.1 15.7 1.3 NR Suzuki ADA 80/40 mg 56.3 16.1 0 NR 2.3 NR -- 5.7 12.6 0 0 ADA 160/80 mg 44.4 13.3 6.7 NR 4.4 NR -- 7.8 18.9 3.3 1.1 2014 PBO 46.9 10.4 4.2 NR 7.3 NR -- 2.1 15.6 0 0 PURSUIT- GOL 200/100 mg 37.5 NR 0.3 0 2.7 NR 0.3 3.3 11.8 0.3 0 GOL 400/200 mg 38.9 NR 0.3 0.3 3.3 NR 0.6 3 12.3 0.9 0 SC PBO 38.2 NR 0.9 0 3.9 NR NR 1.5 12.1 1.8 0 VEDO 300 mg 45 NR 6.7 2.2 3 NR NR <1 14 <1 NR GEMINI 1 PBO 46 NR 0 0 7 NR NR <1 15 2 NR Motoya VEDO 300 mg 50 10.4 4.9 0 6.1 NR NR 3 NR 0.6 NR 2019 PBO 52.4 14.6 2.4 0 2.4 NR NR 2.4 NR 2.4 NR TOF 10 mg 56.5 NR 3.8 0.21 3.4 NR NR NR 23.3 1.3 0 OCTAVE 1 PBO 59.8 NR 1.6 0 4.1 NR NR NR 15.6 0 0 TOF 10 mg 54.1 NR 4 0 4.2 NR NR NR 18.2 0.2 0 OCTAVE 2 PBO 52.7 NR 7.1 0 8 NR NR NR 15.2 0 0 UST 130 mg 41.4 NR NA 0 3.7 NR NR 2.2 15.9 0.6 NR UNIFI UST 6 mg/kg 50.6 NR NA 0.3 3.4 NR NR 0.9 15.9 0.3 NR PBO 48 NR NA 0 6.9 NR NR 1.9 15.4 1.6 NR ADA: adalimumab, AE: adverse event, D/C: discontinuation, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, NA: not applicable, NR: not reported, PBO: placebo, TB: tuberculosis, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 188 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table D16. Treatment Emergent Adverse Events from Open-Label Extensions Intervention Arms Duration N Discontinuations SAE Any AE Serious Infections Suzuki OLE Any ADA 4 years 266 72 (12.4) 129 (22.3) 2900 (431.5) 23 (4.0) Per 100-PY ULTRA OLE ADA 4 years 1010 249 (10.7) 414 (17.7) 8,057 (344.6) 79 (3.4) Per 100-PY 160/80/40 mg PURSUIT M GOL 50 mg 94 2.89 (1.16, 5.95) 7.84 (4.72, 12.24) 187.68 (170.83, 205.74) 1.24 (0.26, 3.62) 139 Extension GOL 100 mg 524 6.18 (4.93, 7.66) 10.23 (8.6, 12.08) 211.45 (203.78, 219.32) 2.65 (1.86, 3.67) weeks Per 100-PY Combined 599 5.69 (4.58, 6.98) 9.87 (8.39, 11.54) 207.85 (200.84, 215.03) 2.44 (1.73, 3.33) ACT 1 & 2 Extn 152 IFX 230 4.63 21 506 3.4 Per 100-PY weeks OCTAVE OLE TOFA 5/10 mg 6.1 years 1157 NR 109 (11.5) 764 (80.9) 32 (3.4) N (%) GEMINI LTS 152 VEDO 300 mg 894 87 (10) 183 (20) 789 (88) 45 (5) N (%) weeks ADA: adalimumab, AE: adverse event, GOL: golimumab, LTS: long term study, mg: milligram, N: number, OLE: open-label extension, PY: patient year, SAE: serious adverse event, TOFA: tofacitinib, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 189 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix E. Comparative Value Supplemental Information Table E1. Impact Inventory Included in This Analysis from… Sector Type of Impact Perspective? Notes on Sources Health Societal Care Formal Health Care Sector Assume no direct impact on mortality outside of potential Longevity effects   reduction in colectomy- Health Outcomes associated mortality Health-related quality of life effects X X Adverse events X X Paid by third-party payers X X Paid by patients out-of-pocket   Medical Costs Future related medical costs X X Direct health state costs Future unrelated medical costs   Informal Health Care Sector Health-Related Patient time costs NA  Unpaid caregiver-time costs NA  Costs Transportation costs NA  Non-Health Care Sectors Labor market earnings lost NA X Cost of unpaid lost productivity due NA X Productivity to illness Cost of uncompensated household NA  production Future consumption unrelated to Consumption NA  health Cost of social services as part of Social Services NA  intervention Number of crimes related to Legal/Criminal NA  intervention Justice Cost of crimes related to NA  intervention Impact of intervention on Education educational achievement of NA  population Cost of home improvements, Housing NA  remediation Production of toxic waste pollution Environment NA  by intervention NA: not applicable Adapted from Sanders et al.175 ©Institute for Clinical and Economic Review, 2020 Page 190 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Description of evLYG Calculations The cost per evLYG considers any extension of life at the same "weight" no matter what treatment is being evaluated. Below are the stepwise calculations used to derive the evLYG. 1. First, we attribute a utility of 0.851, the age- and gender-adjusted utility of the general population in the US that are considered healthy.127 2. For each cycle (Cycle I) in the model where using the intervention results in additional years of life gained, we multiply this general population utility with the additional life years gained (ΔLYG). 3. We sum the product of the life years and average utility (cumulative LYs/cumulative QALYs) for Cycle I in the comparator arm with the value derived in Step 2 to derive the equal value of life years (evLY) for that cycle. 4. If no life years were gained using the intervention versus the comparator, we use the conventional utility estimate for that Cycle I. 5. The total evLY is then calculated as the cumulative sum of QALYs gained using the above calculations for each arm. 6. We use the same calculations in the comparator arm to derive its evLY. Finally, the evLYG is the incremental difference in evLY between the intervention and the comparator arms. Extended Induction Calculations Additional efficacy benefit from extended induction was calculated for adalimumab, golimumab, tofacitinib, ustekinumab, and vedolizumab. For adalimumab, 28.6% of patients achieved remission at eight weeks and 32.7% did at 16 weeks, an absolute gain of 4.1%. If the 71.4% who did not achieve remission at eight weeks continued to be treated, 5.7% of these patients (4.1%/71.4%=5.7%) would be expected to achieve remission at 16 weeks to equal a total of 32.7% at 16 weeks.93 For golimumab, 112 of 398 clinical non-responders at eight weeks in the PURSUIT trial who continued to receive golimumab achieved partial Mayo response at week eight of maintenance (week 16 since treatment initiation), 28.1%.70 No information was available for clinical response without remission. For tofacitinib, of 295 non-responders at end of induction in OCTAVE, 148 achieved response (with and without remission) at 16 weeks. Of these, 41 achieved remission (13.9%) and 107 (148-41) were in response without remission. Therefore, 36.3% (107/295) of patients with active UC at eight weeks will move to the response without remission health state at week 16.118 For ustekinumab in the biologic-naïve population, of 43 non-responders at end of induction in OCTAVE, 34 achieved response (with and without remission) at 16 weeks at the 6 mg/kg dosing.71 Of these, eight achieved remission (18.6%) and 26 (34 minus eight) were in response without remission. Therefore 60.5% (26/43) of patients with active UC at eight weeks will ©Institute for Clinical and Economic Review, 2020 Page 191 Final Report – Targeted Immune Modulators for UC Return to Table of Contents move to the response without remission health state at week 16. For the biologic-experienced population, one of 58 (1.7%) non-responders at the end of induction achieved remission at week 16 and 25 achieved response (with or without remission).71 For vedolizumab, 39% of non-responders at the end of six weeks had response (with or without remission) at 14 weeks. The proportion achieving remission at 14 weeks was not reported.56 One-Way Sensitivity Analysis Tornado Charts Figure E1. Tornado Diagram for One-Way Sensitivity Analyses of Adalimumab versus Conventional Treatment (Cost per QALY) Biologic-Naive Biologic-Experienced ada: adalimumab, bio: biologic, exp: experienced, IV: intravenous, resp: response without remission, RR: risk ratio, tx: treatment, UC: ulcerative colitis ©Institute for Clinical and Economic Review, 2020 Page 192 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E2. Tornado Diagram for One-Way Sensitivity Analyses of Golimumab versus Conventional Treatment (Cost per QALY) Biologic-Naive IV: intravenous, resp: response without remission, RR: risk ratio, UC: ulcerative colitis Figure E3. Tornado Diagram for One-Way Sensitivity Analyses of Infliximab versus Conventional Treatment (Cost per QALY) Biologic-Naive bio: biologic, infx: infliximab, resp: response without remission, RR: risk ratio, UC: ulcerative colitis Figure E4. Tornado Diagram for One-Way Sensitivity Analyses of Infliximab-dyyb versus Conventional Treatment (Cost per QALY) Biologic-Naive bio: biologic, infx: infliximab, resp: response without remission, RR: risk ratio, tx: treatment, UC: ulcerative colitis, ust: ustekinumab ©Institute for Clinical and Economic Review, 2020 Page 193 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E5. Tornado Diagram for One-Way Sensitivity Analyses of Infliximab-abda versus Conventional Treatment (Cost per QALY) Biologic-Naive bio: biologic, infx: infliximab, resp: response without remission, RR: risk ratio, UC: ulcerative colitis Figure E6. Tornado Diagram for One-Way Sensitivity Analyses of Tofacitinib versus Conventional Treatment (Cost per QALY) Biologic-Experienced bio: biologic, exp: experienced, resp: response without remission, tofa: tofacitinib, UC: ulcerative colitis ©Institute for Clinical and Economic Review, 2020 Page 194 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E7. Tornado Diagram for One-Way Sensitivity Analyses of Ustekinumab versus Conventional Treatment (Cost per QALY) Biologic-Naive Biologic-Experienced bio: biologic, exp: experienced, resp: response without remission, RR: risk ratio, tx: treatment, UC: ulcerative colitis, ust: ustekinumab ©Institute for Clinical and Economic Review, 2020 Page 195 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E8. Tornado Diagram for One-Way Sensitivity Analyses of Vedolizumab versus Conventional Treatment (Cost per QALY) Biologic-Naive Biologic-experienced bio: biologic, exp: experienced, IV: intravenous, resp: response without remission, RR: risk ratio, UC: ulcerative colitis, vedo: vedolizumab Undiscounted Results, Biologic-Naïve Population Table E2. Undiscounted Results for the Base-Case for TIMs and Conventional Treatment: Biologic- Naïve Initial TIM Drug Parameter Total Cost LY QALYs evLYs Cost Adalimumab $45,000 $734,000 38.879 27.436 27.443 Golimumab $42,000 $731,000 38.883 27.442 27.450 Infliximab $25,000 $709,000 38.888 27.490 27.498 Infliximab- $23,000 $707,000 38.888 27.490 27.498 dyyb Infliximab- $23,000 $708,000 38.888 27.490 27.498 abda Ustekinumab $148,000 $828,000 38.886 27.530 27.539 Vedolizumab $71,000 $756,000 38.886 27.486 27.495 Conventional $600 $692,000 38.874 27.412 27.419 Treatment evLY: equal value of life years, LY: life year, N/A: not applicable, QALY: quality-adjusted life year, TIM: targeted immune modulator Costs rounded to nearest $1,000. ©Institute for Clinical and Economic Review, 2020 Page 196 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table E3. Undiscounted Incremental Cost-Effectiveness Ratios for the Base Case Compared to Conventional Treatment: Biologic-Naïve Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab $7,417,000 $1,810,000 $1,749,000 Golimumab $4,209,000 $1,347,000 $1,288,000 Infliximab $1,222,000 $224,000 $218,000 Infliximab-dyyb $1,089,000 $199,000 $194,000 Infliximab-abda $1,138,000 $208,000 $203,000 Ustekinumab $10,562,000 $1,157,000 $1,139,000 Vedolizumab $5,249,000 $862,000 $842,000 Conventional Treatment Reference Reference Reference evLY: equal value of life years, LY: life year, QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000 or $10,000, if over $1,000,000. Table E4. Undiscounted Incremental Cost-Effectiveness Ratios for the Base Case Compared to Infliximab: Biologic-Naïve Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab Higher cost and fewer LYs Higher cost and fewer QALYs Higher cost and fewer evLYs Golimumab Higher cost and fewer LYs Higher cost and fewer QALYs Higher cost and fewer evLYs Infliximab Reference Reference Reference Infliximab-dyyb Lower cost and equal LYs Lower cost and equal QALYs Lower cost and equal evLYs Infliximab-abda Lower cost and equal LYs Lower cost and equal QALYs Lower cost and equal evLYs Ustekinumab Higher cost and fewer LYs $2,930,000 $2,950,000 Vedolizumab Higher cost and fewer LYs Higher cost and fewer QALYs Higher cost and fewer evLYs evLY: equal value of life years, LY: life year, QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000 or $10,000, if over $1,000,000. ©Institute for Clinical and Economic Review, 2020 Page 197 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Undiscounted Results, Biologic-Experienced Population Table E5. Undiscounted Results for the Base-Case for TIMs and Conventional Treatment: Biologic- Experienced Initial TIM Drug Parameter Total Cost LY QALYs evLYs Cost Adalimumab $34,000 $742,000 38.863 27.142 27.149 Tofacitinib $33,000 $739,000 38.876 27.184 27.192 Ustekinumab $80,000 $786,000 38.876 27.188 27.197 Vedolizumab $55,000 $761,000 38.870 27.183 27.190 Conventional $300 $711,000 38.863 27.127 27.134 Treatment evLY: equal value of life years, LY: life year, N/A: not applicable, QALY: quality-adjusted life year, targeted immune modulator Costs rounded to nearest $1,000. Table E6. Undiscounted Incremental Cost-Effectiveness Ratios for the Base Case Compared to Conventional Treatment: Biologic-Experienced Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab $76,500,000 $2,104,000 $2,094,000 Tofacitinib $2,224,000 $498,000 $482,000 Ustekinumab $5,767,000 $1,225,000 $1,188,000 Vedolizumab $7,110,000 $910,000 $893,000 Conventional Treatment Reference Reference Reference evLY: equal value of life years, LY: life year, QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. Table E7. Undiscounted Incremental Cost-Effectiveness Ratios for the Base Case Compared to Adalimumab: Biologic-Experienced Treatment Cost per LY Gained Cost per QALY Gained Cost per evLY Gained Adalimumab Reference Reference Reference Lower cost and greater Tofacitinib Lower cost and greater LYs Lower cost and greater QALYs evLYs Ustekinumab $3,434,000 $937,000 $901,000 Vedolizumab $2,807,000 $464,000 $453,000 evLY: equal value of life years, LY: life year, QALY: quality-adjusted life year Incremental cost-effectiveness ratios rounded to nearest $1,000. ©Institute for Clinical and Economic Review, 2020 Page 198 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table E8. League Table of Incremental Cost-Effectiveness Ratios for TIMs Against Conventional Treatment and Each Other, Biologic-Naïve Population Total Costs Total QALYs ICER Conventional with Subsequent tx $421,000 15.57 Reference Adalimumab $461,000 15.60 Dominated Golimumab $458,000 15.60 Dominated Vedolizumab IV $480,000 15.64 Dominated Infliximab $435,000 15.64 Dominated by extension Infliximab-abda $434,000 15.64 Dominated by extension Infliximab-dyyb $434,000 15.64 $186,000 Ustekinumab $545,000 15.68 $3,010,000 ICER: incremental cost-effectiveness ratio, IV: intravenous, QALY: quality-adjusted life year, tx: treatment Incremental cost-effectiveness ratios rounded to nearest $1,000 or $10,000, if over $1,000,000. Table E9. League Table of Incremental Cost-Effectiveness Ratios for TIMs Against Conventional Treatment and Each Other, Biologic-Experienced Population Total Costs Total QALYs ICER Conventional with Subsequent tx $434,000 15.39 -- Adalimumab $465,000 15.41 Dominated Tofacitinib $460,000 15.45 $495,000 Vedolizumab IV $482,000 15.45 $35,130,000 Ustekinumab $504,000 15.45 $10,660,000 ICER: incremental cost-effectiveness ratio, IV: intravenous, QALY: quality-adjusted life year, tx: treatment ©Institute for Clinical and Economic Review, 2020 Page 199 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table E10. Results of Probabilistic Sensitivity Analysis for TIMs versus Conventional Treatment: Biologic-Naïve TIM Conventional Treatment Incremental Mean 95% Credible Range Mean 95% Credible Range Mean 95% Credible Range Adalimumab Total Costs $467,492 ($418,174, $520,639) $383,551 ($338,041, $433,003) $43,231 ($34,129, $56,589) Total QALYs 15.7 (14.1, 17.4) 15.6 (13.9, 17.3) 0.03 (0.00, 0.06) ICER -- -- -- -- $1,536,040 ($918,951, $7,192,797) Golimumab Total Costs $462,452 ($412,330, $517,893) $383,551 ($338,041, $433,003) $38,191 ($30,034, $48,979) Total QALYs 15.7 (14.1, 17.4) 15.6 (13.9, 17.3) 0.03 (0.00, 0.06) ICER -- -- -- -- $1,375,614 ($761,547, $8,322,842) Infliximab Total Costs $439,069 ($391,499, $489,736) $383,551 ($338,041, $433,003) $14,808 ($11,395, $18,669) Total QALYs 15.8 (14.1, 17.4) 15.6 (13.9, 17.3) 0.08 (0.03, 0.17) ICER -- -- -- -- $192,451 ($111,199, $404,398) Infliximab-dyyb Total Costs $437,113 ($389,159, $488,141) $383,551 ($338,041, $433,003) $12,852 ($9,717, $16,244) Total QALYs 15.8 (14.1, 17.4) 15.6 (13.9, 17.3) 0.08 (0.03, 0.17) ICER -- -- -- -- $166,730 ($98,353, $345,518) Infliximab-abda Total Costs $437,913 ($390,243, $489,800) $383,551 ($338,041, $433,003) $13,653 ($10,410, $17,233) Total QALYs 15.8 (14.1, 17.5) 15.6 (13.9, 17.3) 0.08 (0.03, 0.17) ICER -- -- -- -- $177,482 ($103,715, $371,439) Ustekinumab Total Costs $562,291 ($470,587, $707,054) $383,551 ($338,041, $433,003) $138,030 ($58,199, $281,954) Total QALYs 15.8 (14.2, 17.5) 15.6 (13.9, 17.3) 0.13 (0.02, 0.35) ICER -- -- -- -- $1,093,362 ($811,762, $2,414,141) Vedolizumab Total Costs $484,525 ($431,282, $541,474) $383,551 ($338,041, $433,003) $60,265 ($45,732, $78,911) Total QALYs 15.8 (14.1, 17.4) 15.6 (13.9, 17.3) 0.07 (0.02, 0.14) ICER -- -- -- -- $863,528 ($579,730, $1,893,411) ICER: incremental cost-effectiveness ratio, QALY: quality-adjusted life year, TIM: targeted immune modulator ©Institute for Clinical and Economic Review, 2020 Page 200 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E9. Probabilistic Sensitivity Analysis Results: Cost-Effectiveness Clouds: Biologic-Naïve IV: intravenous, QALY: quality-adjusted life year Table E11. Results of Probabilistic Sensitivity Analysis for TIMs versus Conventional Treatment: Biologic-Experienced TIM Conventional Treatment Incremental Mean 95% Credible Range Mean 95% Credible Range Mean 95% Credible Range Adalimumab Total Costs $469,422 ($416,569, $535,411) $396,179 ($351,693, $448,136) $31,669 ($22,189, $49,881) Total QALYs 15.5 (13.9, 17.2) 15.4 (13.8, 17.0) 0.02 (0.00, 0.08) ICER -- -- -- -- $1,590,608 ($643,519, -$11,824,398) Tofacitinib Total Costs $464,295 ($412,153, $524,222) $396,179 ($351,693, $448,136) $26,543 ($17,516, $42,404) Total QALYs 15.5 (14.0, 17.2) 15.4 (13.8, 17.0) 0.1 (0.02, 0.13) ICER -- -- -- -- $473,341 ($338,730, $816,525) Ustekinumab Total Costs $522,911 ($447,096, $611,399) $396,179 ($351,693, $448,136) $85,159 ($45,316, $147,519) Total QALYs 15.6 (14.0, 17.2) 15.4 (13.8, 17.0) 0.08 (0.02, 0.16) ICER -- -- -- -- $1,135,016 ($947,554, $1,956,180) Vedolizumab Total Costs $488,633 ($431,764, $563,913) $396,179 ($351,693, $448,136) $50,881 ($34,891, $89,064) Total QALYs 15.6 (14.0, 17.2) 15.4 (13.8, 17.0) 0.07 (0.02, 0.21) ICER -- -- -- -- $775,645 ($421,194, $1,949,335) ICER: incremental cost-effectiveness ratio, QALY: quality-adjusted life year, TIM: targeted immune modulator ©Institute for Clinical and Economic Review, 2020 Page 201 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Figure E10. Probabilistic Sensitivity Analysis Results: Cost-Effectiveness Clouds: Biologic- Experienced IV: intravenous, QALY: quality-adjusted life year ©Institute for Clinical and Economic Review, 2020 Page 202 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table E12. Proportion of Patients in Each Health State at One, Five, and 10 Years: Biologic-Naïve Adalimumab Active UC Response Remission Post-Colectomy Dead 1 Year 35% 38% 26% 0% 0% 5 Years 76% 10% 9% 4% 1% 10 Years 76% 7% 5% 9% 3% Golimumab Active UC Response Remission Post-Colectomy Dead 1 Year 34% 38% 27% 0% 0% 5 Years 76% 10% 9% 4% 1% 10 Years 76% 7% 5% 9% 3% Infliximab, Infliximab- Active UC Response Remission Post-Colectomy Dead dyyb, and Infliximab-dyyb 1 Year 28% 36% 35% 0% 0% 5 Years 72% 11% 12% 3% 1% 10 Years 76% 7% 5% 9% 3% Ustekinumab Active UC Response Remission Post-Colectomy Dead 1 Year 27% 35% 38% 0% 0% 5 Years 68% 12% 16% 3% 1% 10 Years 75% 8% 6% 8% 3% Vedolizumab Active UC Response Remission Post-Colectomy Dead 1 Year 29% 36% 34% 0% 0% 5 Years 72% 11% 12% 3% 1% 10 Years 76% 7% 5% 9% 3% Conventional Treatment Active UC Response Remission Post-Colectomy Dead 1 Year 40% 39% 21% 0% 0% 5 Years 78% 9% 8% 4% 1% 10 Years 76% 7% 4% 10% 3% UC: ulcerative colitis ©Institute for Clinical and Economic Review, 2020 Page 203 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table E12. Proportion of Patients in Each Health State at One, Five, and 10 Years: Biologic- Experienced Adalimumab Active UC Response Remission Post-Colectomy Dead 1 Year 45% 36% 18% 0% 0% 5 Years 85% 5% 4% 5% 1% 10 Years 82% 3% 1% 11% 3% Tofacitinib Active UC Response Remission Post-Colectomy Dead 1 Year 37% 39% 23% 0% 0% 5 Years 83% 6% 5% 4% 1% 10 Years 83% 3% 1% 10% 3% Ustekinumab Active UC Response Remission Post-Colectomy Dead 1 Year 36% 40% 24% 0% 0% 5 Years 83% 6% 5% 4% 1% 10 Years 83% 3% 1% 10% 3% Vedolizumab Active UC Response Remission Post-Colectomy Dead 1 Year 39% 37% 23% 0% 0% 5 Years 82% 6% 6% 4% 1% 10 Years 82% 3% 1% 10% 3% Conventional Treatment Active UC Response Remission Post-Colectomy Dead 1 Year 49% 36% 15% 0% 0% 5 Years 86% 4% 3% 5% 1% 10 Years 82% 3% 1% 11% 3% UC: ulcerative colitis ©Institute for Clinical and Economic Review, 2020 Page 204 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix F. Network Meta-Analysis Supplemental Information Network Meta-Analysis Methods The comparative efficacy of the TIMs for patients living with moderate-to-severe UC was assessed by means of NMA, where feasible. Trials that were deemed sufficiently similar in terms of population, intervention type, duration, and outcome definitions were included in the NMAs. NMAs focused on clinical response, clinical remission, and endoscopic improvement were conducted. Given the expected differences in the clinical efficacy of treatment in patients with and without prior biologic exposure, separate networks were developed for biologic-naïve and biologic- experienced populations. Clinical response and remission were analyzed as ordered categorical outcomes ("no response," "response without remission," and "remission") with a multinomial likelihood and a probit link. Endoscopic improvement was analyzed as a dichotomous outcome ("yes" or "no") with a binomial likelihood and log link. Outcomes were analyzed separately for the induction phase (six to 14 weeks) and maintenance phase (52-60 weeks). All efficacy outcomes were analyzed in the induction phase (six to 14 weeks). In addition, clinical response and remission were analyzed in the maintenance phase (52-60 weeks). Trials included in the maintenance NMAs had a maintenance phase of at least 52 weeks. Therefore, some trials with a shorter maintenance phase (ACT 2, Jiang 2015, and Kobayashi 2016) were excluded from the maintenance evidence network. Endoscopic improvement was analyzed only during the induction phase due to limited data availability and trial design differences. The evidence base for the maintenance phase in the included trials is a combination of "treat- through" designs, where patients were randomized only at baseline and followed through until the end of maintenance, and "re-randomized" designs, where responders to treatment from one or two induction trials were re-randomized in the maintenance phase. In order to analyze all trials in comparable fashion in one network, results from treat-through trials were adjusted to more closely resemble results from re-randomized trials. Three maintenance trials included in the NMAs had a "treat-through" study design (ACT 1, ULTRA 2, and VARSITY);8,90,93 of note, another one of the available trials also had a treat-through design (Suzuki 2014), but was not included in the maintenance NMA due to lack of data. Six maintenance trials had a "re-randomized" study design (PURSUIT-M, GEMINI 1, Motoya 2019, VISIBLE 1, OCTAVE SUSTAIN, and UNIFI);9,77,92,94,99 of note, another available trial also had a re-randomized design (PURSUIT-J), but was not included in the maintenance NMA due to a lack of data. ©Institute for Clinical and Economic Review, 2020 Page 205 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Data were available for all TIMs (adalimumab, golimumab, infliximab, tofacitinib, ustekinumab, and vedolizumab) in the biologic-naïve population. We note, however, that use of tofacitinib is no longer feasible in a biologic-naïve population, based on an FDA-enforced label change (July 2019) that now requires that tofacitinib use be reserved for "…patients who have failed or do not tolerate tumor necrosis factor (TNF) blockers."65 Based on this information, tofacitinib was not included in the NMAs for induction or maintenance within the biologic-naïve population. Data were available for adalimumab, tofacitinib, ustekinumab, and vedolizumab in the biologic-experienced population; data were not available for infliximab or golimumab, so we were unable to generate comparative efficacy estimates for these drugs. Both random- and fixed-effects models were explored. We used fixed-effects models given the limited data available for each network. In addition, we explored adjusting for baseline risk given the differences in placebo response rates across trials. The population of focus for this review included patients with moderate-to-severe UC who had inadequate response to conventional treatment. Despite this relatively narrow definition, trial populations may differ in terms of prior conventional therapies used, other demographic or clinical risk factors, timing of trial assessments, and other concerns. Adjusting for placebo response in an NMA design is frequently performed as means of controlling for differences in population characteristics and baseline risk; we considered placebo adjustment in situations where a) model fit and convergence was not compromised; and b) inclusion of such an adjustment materially changed model findings. We attempted placebo adjustment for all four populations of interest in our NMAs. Material changes in findings were observed in both maintenance populations (i.e., biologic-naïve and biologic-experienced populations). However, results were highly unstable in the biologic- experienced maintenance NMA, regardless of the number of iterations attempted, in all likelihood due to the sparsity of the available network. We therefore included a placebo adjustment only in the maintenance NMA for the biologic-naïve population. Table F1 lists the NMAs we conducted and the details of the model, and Table F2 lists the trials included in our NMAs as well as reasons for exclusion of trials. Finally, while different doses of some of the TIMs were studied in available trials, we found no statistically significant differences in rates of response or remission between doses in any relevant trial, so these data were pooled at the drug level for our NMAs. We conducted inconsistency tests for each of the NMA populations using the node-splitting approach. We found no statistical differences between direct and indirect estimates in any of our populations of interest (comparing indirect and direct evidence for vedolizumab and adalimumab: induction biologic-naive, p=0.67; induction biologic-experienced, p=0.08; maintenance biologic-naive, p=0.63 maintenance biologic-experienced, p=0.59). ©Institute for Clinical and Economic Review, 2020 Page 206 Final Report – Targeted Immune Modulators for UC Return to Table of Contents All NMAs were conducted in a Bayesian framework in R. NMAs on clinical response and remission were conducted with JAGS using the R2jags package.67 For our NMAs on clinical response and remission, we based out analysis on existing code.176 NMAs on endoscopic improvement were conducted using the gemtc package.66 Table F1. NMAs Conducted to Inform Comparative Efficacy of TIMs Population Model Number of Trials Induction Outcomes Clinical Response and a) Biologic-naïve Multinomial with probit a) 12 Remission b) Biologic-experienced link b) 7 Endoscopic a) Biologic-naïve a) 11 Binomial with log link Improvement b) Biologic-experienced b) 6 Maintenance Outcomes Clinical Response and a) Biologic-naïve Multinomial with probit a) 8 Remission b) Biologic-experienced link b) 7 Table F2. Trials Included in NMAs Conducted to Inform Comparative Efficacy of TIMs Induction NMAs Maintenance NMAs Trial Response and Remission Endoscopic Improvement Response and Remission Naïve Experienced Naïve Experienced Naïve Experienced VARSITY ✓ ✓ ✗ ✗ ✓ ✓ ULTRA 1 ✓ - ✓ - - - ULTRA 2 ✓ ✓ ✓ ✓ ✓ ✓ Suzuki 2014 ✓ - ✓ - ✗ - PURSUIT-SC ✓ - ✓ - - - PURSUIT-M - - - - ✓ - PURSUIT-J - - - - ✗ - ACT 1 ✓ - ✓ - ✓ - ACT 2 ✓ - ✓ - ✗ - Kobayashi 2016 ✓ - ✓ - ✗ - Jiang 2015 ✓ - ✓ - ✗ - NCT01551290 ✗ - ✗ - ✗ - OCTAVE 1 ✗ ✓ ✗ ✓ - - OCTAVE 2 ✗ ✓ ✗ ✓ - - OCTAVE - - - - ✗ ✓ SUSTAIN UNIFI ✓ ✓ ✓ ✓ ✓ ✓ GEMINI 1 ✓ ✓ ✓ ✓ ✓ ✓ Motoya 2019 ✓ ✓ ✓ ✓ ✓ ✓ VISIBLE 1 - - - - ✓ ✓ NMA: network meta-analysis ©Institute for Clinical and Economic Review, 2020 Page 207 Final Report – Targeted Immune Modulators for UC Return to Table of Contents In the table above, a check ("✓") indicates the trial was included in the specified NMA, while a cross mark ("✗") indicates the trial was excluded from the NMA. A dash ("-") indicates that the trial was not designed to provide data for the specified NMA (i.e., the trial only included biologic-naïve patients [all infliximab and golimumab trials, ULTRA 1, and Suzuki 2014], or the trial included only a randomized induction phase [ULTRA 1, PURSUIT-SC, OCTAVE 1 & 2] or only a randomized maintenance phase [PURSUIT-M, PURSUIT-J, OCTAVE SUSTAIN, VISIBLE 1]). Below are the following reasons we excluded trials from our NMAs. • VARSITY was excluded from the endoscopic improvement NMAs as data stratified by biologic exposure was not available. • Suzuki 2014 and PURSUIT-J were excluded from the response and remission maintenance NMA due to a lack of data to achieve outcomes comparable to those in the other trials included in the network (more detailed provided in 'Inputs used in NMAs of Maintenance Outcomes'). • ACT 2, Kobayashi 2016, and Jiang 2015 were excluded from response and remission maintenance NMA as their maintenance phase was <52 weeks (as mentioned earlier). • NCT01551290 was excluded from the NMAs as it was available only in grey literature. • OCTAVE 1, 2, and SUSTAIN were excluded from the biologic-naïve NMAs due to tofacitinib's FDA-enforced label change. ©Institute for Clinical and Economic Review, 2020 Page 208 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Network Meta-Analysis Inputs Inputs Used in Network Meta-Analyses of Induction Outcomes Table F3. Inputs Used in NMA of Response and Remission in the Induction Phase, Biologic-Naïve Response Remission Trial Wk Arm* n N % n N % IFX 5 mg/kg 84 121 69.4 47 121 38.8 IFX 10 mg/kg 75 122 61.5 39 122 32.0 ACT 1 8 IFX pooled 159 243 65.4 86 243 35.4 PBO 45 121 37.2 18 121 14.9 IFX 5 mg/kg 78 121 64.5 41 121 33.9 IFX 10 mg/kg 83 120 69.2 33 120 27.5 ACT 2 8 IFX pooled 161 241 66.8 74 241 30.7 PBO 36 123 29.3 7 123 5.7 IFX 5 mg/kg 32 41 78.1 22 41 53.7 Jiang 2015 8 PBO 15 41 36.6 9 41 21.9 IFX 5 mg/kg 57 104 54.8 21 104 20.2 Kobayashi 2016 8 PBO 37 104 35.6 11 104 10.6 ADA 160/80 mg 71 130 54.6 24 130 18.5 ULTRA 1 8 PBO 58 130 44.6 12 130 9.2 ADA 160/80 mg 89 150 59.3 32 150 21.3 ULTRA 2 8 PBO 56 145 38.6 16 145 11.0 ADA 160/80 mg 45 90 50.0 9 90 10.0 Suzuki 2014 8 PBO 34 96 35.4 11 96 11.5 GOL 200/100 147 294 50.0 52 294 17.7 mg PURSUIT-SC GOL 400/200 (Phase II & III 6 163 298 54.7 56 298 18.8 mg Pooled) GOL pooled 310 592 52.4 108 592 18.2 PBO 89 292 30.5 20 292 6.8 VEDO 300 mg 69 130 53.1 30 130 23.1 GEMINI I 6 PBO 20 76 26.3 5 76 6.6 VEDO 300 mg 42 79 53.2 22 79 27.8 Motoya 2019 10 PBO 15 41 36.6 6 41 14.6 VEDO 300 mg 213 304 70.1 84 304 27.6 VARSITY 14 ADA 160/80 mg 151 305 49.5 72 305 23.6 UST 6 mg/kg 104 156 66.7 29 156 18.6 UNIFI 8 PBO 56 158 35.4 15 158 9.5 ADA: adalimumab, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, PBO: placebo, UST: ustekinumab, VEDO: vedolizumab, Wk: week *Pooled doses used in primary NMA. Unpooled doses used in sensitivity analysis. ©Institute for Clinical and Economic Review, 2020 Page 209 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table F4. Inputs Used in NMA of Response and Remission in the Induction Phase, Biologic- Experienced Response Remission Trial Wk Arm* n N % n N % ADA 160/80 36 98 36.7 9 98 9.2 ULTRA 2 8 mg PBO 29 101 28.7 7 101 6.9 VEDO 300 mg 32 82 39 8 82 9.8 GEMINI 1 6 PBO 13 63 20.6 2 63 3.2 VEDO 300 mg 23 85 27.1 8 85 9.4 Motoya 2019 10 PBO 12 41 29.3 4 41 9.8 VEDO 300 mg 44 79 55.7 18 79 22.8 VARSITY 14 ADA 160/80 26 81 32.1 10 81 12.3 mg OCTAVE TOF 10 mg 237 465 51.0 53 465 11.4 1 and 2 8 PBO 29 124 23.4 1 124 0.8 (Pooled) UST 6 mg/kg 95 166 57.2 21 166 12.7 UNIFI 8 PBO 44 161 27.3 2 161 1.2 ADA: adalimumab, kg: kilogram, mg: milligram, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, Wk: week ©Institute for Clinical and Economic Review, 2020 Page 210 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table F5. Inputs Used in NMA of Endoscopic Improvement in the Induction Phase, Biologic-Naïve and Biologic-Experienced Biologic-Naïve Biologic-Experienced Trial Wk Arm n N % n N % IFX pooled 147 243 60 ACT 1 8 PBO 41 121 34 IFX pooled 147 241 61 ACT 2 8 PBO 38 123 31 IFX pooled 24 41 59 Not studied in biologic- Jiang 2015 8 PBO 10 41 24 experienced population Kobayashi IFX pooled 48 104 46 8 2016 PBO 29 104 28 ADA 160 mg 61 130 47 ULTRA 1 8 PBO 54 130 42 ADA 160 mg 74 150 49 28 98 28.6 ULTRA 2 8 PBO 51 145 35 27 101 26.7 ADA 160 mg 40 90 44 Not studied in biologic- Suzuki 2014 8 PBO 29 96 30 experienced population VEDO 300 mg 64 130 49 25 82 30.5 GEMINI 1 6 PBO 19 76 25 13 63 20.6 PURSUIT-SC GOL pooled 256 592 43 Not studied in biologic- (Phase II & 6 PBO 82 292 28 experienced population III Pooled) Motoya VEDO 300 mg 38 79 48 22 85 25.8 10 2019 PBO 13 41 32 12 41 29.2 UST 6 mg/kg 52 156 33 35 166 21.1 UNIFI 8 PBO 33 158 21 11 161 6.8 OCTAVE TOF 10 mg 112 488 22.9 Data not included in biologic- 1&2 8 PBO naive population 8 130 6.2 (Pooled) ADA: adalimumab, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab, Wk: week Inputs Used in Network Meta-Analyses of Maintenance Outcomes In the section that follows, we describe the inputs used in our NMAs of response and remission in the maintenance phase. Throughout our review, we describe results for the key outcomes in the biologic-naïve and biologic-experienced populations as reported in the published trials (i.e., measures designated as primary and key secondary outcomes for the overall population). As discussed earlier, we adjusted the rates from the treat-through trials to more closely resemble results from re-randomized trials to enable comparisons in our NMA. Additionally, for some re- randomized trials, we preferred manufacturer-submitted inputs or other published secondary outcomes over published and key secondary outcomes as they provided us with more comparable ©Institute for Clinical and Economic Review, 2020 Page 211 Final Report – Targeted Immune Modulators for UC Return to Table of Contents outcomes for our NMA (e.g., response at end of maintenance vs. response through the end of maintenance). The data inputs used in our response and remission maintenance NMA are provided in Table F6. Adjusted Rates from Treat-Through Trials As noted earlier, we adjusted the rates from the treat-through trials to more closely resemble results from re-randomized trials to enable comparisons in our NMA. Specifically, we adjusted rates from treat-through trials to reflect maintenance outcomes among induction responders. Responders to induction treatment (week six to 14) in the treat-through trials were assumed to enter the maintenance phase. The assumed denominator for the treat-through trials was the number of responders at the end of the induction phase. The numerator for clinical response was the number of sustained responders (i.e., having response at both beginning of induction and end of maintenance). The numerator for remission was the number of patients with remission at the end of maintenance among induction responders. To assist in these calculations, we have received adjusted rates from treat-through trials primarily from manufacturer data submissions. Trial specific adjustments included: • VARSITY: Rate of response at week 52 among week-six responders (based on partial Mayo Score) submitted by manufacturer • ULTRA 2: Rate of response at maintenance week 52 among week-eight responders reported in trial; rate of remission at week 52 among week-eight responders reported in conference abstract102 • ACT 1: Rate of response and remission at week 54 among eight-week responders submitted by manufacturer Of note, in Suzuki 2014, the rate of response and remission among induction responders was available for the treatment (adalimumab) arm only; consequently, the lack of comparator arm data did not allow for inclusion in the maintenance NMA. Alternative Endpoints Used for Re-Randomized Trials For some re-randomized trials, we preferred manufacturer-submitted inputs or other published secondary outcomes over key outcomes as they provided us with more comparable outcomes for our NMA. Specifically, while many trials measured the rates of response at the end of maintenance, some trials used stricter criteria and measured the rates of response through the end of maintenance (PURSUIT-M, PURSUIT-J, and UNIFI). In addition, while most trials re-randomized patients who initially responded to six to ten-week induction treatment with the active agent, one trial also re-randomized patients who responded to placebo (OCTAVE SUSTAIN), and another trial re-randomized patients who did not respond to placebo but subsequently responded to the active agent after another eight weeks (UNIFI). Given the variation, we used manufacturer-submitted ©Institute for Clinical and Economic Review, 2020 Page 212 Final Report – Targeted Immune Modulators for UC Return to Table of Contents inputs or other published secondary outcomes if they provided us with more comparable outcomes, where feasible. Below, we provide details for each trial. PURSUIT-M PURSUIT-M and PURSUIT-J measured the rates of maintaining response through the end of maintenance rather than measuring response at the end of maintenance. For PURSUIT-M, we included the rates of remission at the end of maintenance (i.e., at week 54) in our NMA. However, for PURSUIT-J, only the rate of remission at both weeks 30 and 54 was reported; therefore, we did not include PURSUIT-J in our analysis. OCTAVE SUSTAIN OCTAVE SUSTAIN re-randomized both tofacitinib- and placebo-induction responders in the maintenance phase. However, all other re-randomized trials included in our review re-randomized responders to the active agent only. Therefore, we included results from the modified full analysis set (mFAS) that only included tofacitinib induction responders in NMA. We obtained the rates of remission at week 52 in the mFAS from a conference abstract that reported the rates of remission among baseline remitters and baseline responders in the mFAS;85 in order to obtain the rates among all responders (both remitters and responders without remission), we added the reported rates. The rates of response were not available from the mFAS. UNFI UNIFI measured the rate of response through the end of maintenance rather than measuring response at the end of maintenance. We received the rates of response at week 52 from manufacturer submissions. In addition, UNIFI re-randomized patients that did not respond to induction therapy with placebo but then responded to induction therapy with ustekinumab. Since no other trial re-randomized patients following a similar path, we used rates from analyses submitted by the manufacturer that excluded these patients. VISIBLE 1 For VISIBLE 1, the rates of response at week 52 were submitted by the manufacturer. ©Institute for Clinical and Economic Review, 2020 Page 213 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table F6. Inputs Used in the NMA of Response and Remission in Maintenance Phase* Maintenance Outcomes Among Induction Responders Biologic-Naïve Biologic-Experienced Trial Week Arm* Response Remission Response Remission n N % n N % n N % n N % Treat-Through Trials VEDO 300 mg q8w VARSITY1 52 NR NR ADA 40 mg ADA 40 mg 44 89 49.4 34 89 38.2 15 36 41.7 10 36 27.8 ULTRA 22 52 PBO 24 56 42.9 15 56 26.8 6 29 20.7 2 29 6.9 IFX 5 mg/kg IFX 10 mg/kg ACT 13 54 N/A IFX pooled PBO Re-Randomized Trials GOL 100 mg 51 151 33.8 PURSUIT-M4 54 NR N/A PBO 34 154 22.1 TOF 5 mg 16 76 21.1 OCTAVE TOF 10 mg 28 78 35.9 52 N/A NR SUSTAIN5 TOF pooled 44 154 28.6 PBO 9 80 11.3 UST 90 mg q8w UNIFI6 52 PBO VEDO 300 mg q4w 41 73 56.2 35 73 47.9 17 40 42.5 14 40 35 VEDO 300 mg q8w 47 72 65.3 33 72 45.8 20 43 46.5 16 43 37.2 GEMINI 1 52 VEDO pooled 88 145 60.7 68 145 46.9 37 83 44.6 30 83 36.1 PBO 21 79 26.6 15 79 19.0 6 38 15.8 2 38 5.3 VEDO 300 mg q8w 16 24 66.7 13 24 54.2 11 17 64.7 10 17 58.8 Motoya 2019 60 PBO 10 28 35.7 10 28 35.7 5 14 35.7 3 14 21.4 VEDO 300 mg q8w 17 32 53.1 6 22 27.3 VISIBLE 17 52 PBO 7 37 18.9 1 19 5.3 ADA: adalimumab, GOL: golimumab, IFX: infliximab, kg: kilogram, mg: milligram, NR: not reported, N/A: not applicable, PBO: placebo, q8w: every 4 weeks, q8w: every 8 weeks, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab *Pooled doses used in primary NMA. Unpooled doses used in sensitivity analysis. 1 VARSITY: Rate of response at week 52 among week-six responders (based on partial Mayo Score) submitted by manufacturer. 2 ULTRA 2: Rate of response at week 52 among week-eight responders reported in trial; rate of remission at week 52 among week-eight responders reported in conference abstract. 3 ACT 1: Rate of response and remission at week 54 among eight-week responders submitted by manufacturer. 4 PURSUIT-M: Rate of remission at week 54 was reported in trial. 5 OCTAVE SUSTAIN: Rate of remission in the modified full analysis set obtained from conference abstract. 6 UNIFI: Rate of response and remission at maintenance week 52 among patients who initially responded to ustekinumab were submitted by the manufacturer. 7 VISIBLE 1: Rate of response submitted by the manufacturer. ©Institute for Clinical and Economic Review, 2020 Page 214 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Supplemental Network Meta-Analysis Results Network Diagrams for the Primary Network Meta-Analyses ©Institute for Clinical and Economic Review, 2020 Page 215 Final Report – Targeted Immune Modulators for UC Return to Table of Contents ©Institute for Clinical and Economic Review, 2020 Page 216 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Sensitivity Analyses Unpooled Doses We conducted a sensitivity analysis using unpooled doses. Results were generally consistent with the primary analysis results, with the exception of results for tofacitinib in the maintenance biologic-experienced NMA. Results from the sensitivity analysis showed tofacitinib 10 mg had statistically higher rates of response and remission compared to placebo, but tofacitinib 5 mg did not differ from placebo. When the doses were pooled in the primary analysis, tofacitinib had statistically higher rates of response and remission compared to placebo. Table F7. Sensitivity Analysis Using Unpooled Doses, Induction Biologic-Naïve: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO -- -- -- VEDO 300 1.76 (1.54, 2) 1.37 (1.23, 1.54) 2.8 (2.18, 3.56) IFX 5 1.9 (1.68, 2.17) 1.39 (1.22, 1.6) 3.3 (2.64, 4.13) IFX 10 1.83 (1.58, 2.11) 1.38 (1.22, 1.58) 3.04 (2.32, 3.94) ADA 160/80 1.38 (1.21, 1.57) 1.24 (1.14, 1.36) 1.75 (1.38, 2.21) GOL 200/100 1.58 (1.36, 1.84) 1.32 (1.2, 1.48) 2.28 (1.7, 2.99) GOL 400/200 1.68 (1.46, 1.94) 1.35 (1.22, 1.52) 2.55 (1.97, 3.31) UST 6 1.71 (1.43, 2.04) 1.36 (1.21, 1.54) 2.65 (1.88, 3.68) ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, UST: ustekinumab, VEDO: vedolizumab Table F8. Sensitivity Analysis Using Unpooled Doses, Maintenance Biologic-Naïve: Risk Ratios versus Placebo Response without Response Remission Remission PBO --- --- --- VEDO 300 q8w 1.7 (1.25, 2.12) 1.1 (0.93, 1.23) 2.04 (1.35, 2.77) VEDO 300 q4w 1.58 (1.04, 2.07) 1.1 (0.9, 1.24) 1.84 (1.05, 2.68) IFX 5 1.56 (1.03, 2.27) 1.1 (0.86, 1.22) 1.8 (1.04, 3.08) IFX 10 1.67 (1.14, 2.35) 1.09 (0.82, 1.22) 1.99 (1.2, 3.25) ADA 40 1.39 (0.98, 1.81) 1.11 (0.98, 1.23) 1.54 (0.98, 2.2) GOL 100 1.28 (0.83, 1.74) 1.09 (0.89, 1.22) 1.38 (0.79, 2.09) UST 90 q8w 1.78 (1.24, 2.49) 1.06 (0.69, 1.21) 2.18 (1.32, 3.56) ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, q4w: every 4 weeks, q8w: every 8 weeks, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 217 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table F9. Sensitivity Analysis Using Unpooled Doses, Maintenance Biologic-Experienced: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO --- --- --- VEDO 300 q8w 2.5 (1.91, 3.17) 1.49 (1.21, 1.78) 3.64 (2.45, 5.21) VEDO 300 q4w 2.39 (1.58, 3.25) 1.47 (1.18, 1.76) 3.37 (1.85, 5.45) ADA 40 2.2 (1.48, 2.98) 1.47 (1.21, 1.74) 3 (1.69, 4.75) TOF 10 2.27 (1.57, 3.05) 1.47 (1.24, 1.76) 3.14 (1.84, 4.86) TOF 5 1.57 (0.9, 2.38) 1.32 (0.93, 1.61) 1.84 (0.88, 3.35) UST 90 q8w 1.95 (1.44, 2.59) 1.44 (1.23, 1.69) 2.51 (1.65, 3.76) ADA: adalimumab, GOL: golimumab, PBO: placebo, q4w: every 4 weeks, q8w: every 8 weeks, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Multinational Trials Only We conducted another sensitivity analysis assessing multinational trials separately by excluding trials conducted exclusively in Asian populations (Suzuki 2014, Jiang 2015, Kobayashi 2016, and Motoya 2019). We were unable to conduct this sensitivity analysis for our maintenance biologic- naïve NMA as removal of Motoya 2019 led to a sparser network and the results from the model with placebo adjustment were unstable. We noted that risk ratios were higher when removing the Asian studies, with the largest effect on risk ratios observed for vedolizumab in the induction biologic-experienced NMA. Table F10. Sensitivity Analysis with Multinational Trials Only, Induction Biologic-Naïve: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO -- -- -- VEDO 300 1.84 (1.59, 2.13) 1.4 (1.24, 1.6) 3.1 (2.33, 4.03) IFX Pooled 1.96 (1.72, 2.25) 1.41 (1.23, 1.63) 3.54 (2.77, 4.52) ADA 160/80 1.44 (1.24, 1.66) 1.27 (1.16, 1.42) 1.91 (1.47, 2.46) GOL Pooled 1.64 (1.42, 1.88) 1.35 (1.23, 1.52) 2.43 (1.89, 3.12) UST 6 1.73 (1.43, 2.07) 1.37 (1.23, 1.57) 2.71 (1.89, 3.79) ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 218 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Table F11. Sensitivity Analysis with Multinational Trials Only, Induction Biologic-Experienced: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO -- -- -- VEDO Pooled 1.99 (1.46, 2.56) 1.72 (1.37, 2.09) 3.66 (2.01, 6.21) ADA 160/80 1.2 (0.84, 1.64) 1.17 (0.85, 1.51) 1.38 (0.74, 2.46) TOF 10 2.16 (1.74, 2.67) 1.81 (1.53, 2.17) 4.35 (2.79, 6.63) UST 6 2.17 (1.75, 2.68) 1.82 (1.54, 2.16) 4.42 (2.83, 6.73) ADA: adalimumab, PBO: placebo, TOF: tofacitinib, UST: ustekinumab, VEDO: vedolizumab Table F12. Sensitivity Analysis with Multinational Trials Only, Maintenance Biologic-Experienced: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO --- --- --- VEDO Pooled 2.56 (1.92, 3.29) 1.55 (1.26, 1.87) 3.86 (2.48, 5.76) ADA 40 2.25 (1.49, 3.09) 1.52 (1.26, 1.82) 3.16 (1.73, 5.12) TOF 1.97 (1.34, 2.73) 1.48 (1.22, 1.76) 2.6 (1.49, 4.24) UST 90 q8w 2 (1.4, 2.73) 1.49 (1.24, 1.77) 2.66 (1.59, 4.18) ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, q8w: every 8 weeks, UST: ustekinumab, VEDO: vedolizumab Additional Analyses As noted earlier, we included placebo adjustment in the maintenance biologic-naïve NMA. Below, we provide results without placebo adjustment, as previously reported as our primary analysis in our Draft Evidence Report. Table F13. Sensitivity Analysis without Placebo Adjustment, Maintenance Biologic-Naïve: Risk Ratios versus Placebo Response without Treatment Response Remission Remission PBO --- --- --- VEDO Pooled 1.76 (1.52, 2.02) 1.05 (0.87, 1.2) 2.14 (1.74, 2.58) IFX Pooled 1.46 (1.06, 1.87) 1.09 (0.96, 1.2) 1.66 (1.08, 2.31) ADA 40 1.43 (1.15, 1.71) 1.1 (1.01, 1.19) 1.6 (1.2, 2.04) GOL 100 1.35 (1.05, 1.65) 1.09 (1, 1.18) 1.48 (1.07, 1.97) UST 90 q8w 1.58 (1.16, 1.98) 1.08 (0.91, 1.2) 1.84 (1.21, 2.52) ADA: adalimumab, GOL: golimumab, IFX: infliximab, PBO: placebo, q8w: every 8 weeks, UST: ustekinumab, VEDO: vedolizumab ©Institute for Clinical and Economic Review, 2020 Page 219 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Network Meta-Analysis Code Model (Fixed Effects) model <- function() { # *** PROGRAM STARTS for(i in 1:ns){ # LOOP THROUGH STUDIES mu[i] ~ dnorm(0,.0001) # vague priors for all trial baselines for (k in 1:na[i]) { # LOOP THROUGH ARMS p[i,k,1] <- 1 # Pr(eff>0) for (j in 1:(nc[i]-1)) { # LOOP THROUGH CATEGORIES r[i,k,j] ~ dbin(q[i,k,j],n[i,k,j]) # binomial likelihood q[i,k,j] <- 1-(p[i,k,C[i,(j+1)]]/p[i,k,C[i,j]]) # conditional probabilities z.index[i,j,k]<- C[i,(j+1)]-1 # index the cut point theta[i,k,j] <- mu[i] + d[t[i,k]] - d[t[i,1]] + z[z.index[i,j,k]] # linear predictor rhat[i,k,j] <- q[i,k,j] * n[i,k,j] # predicted number events dv[i,k,j] <- 2 * (r[i,k,j]*(log(r[i,k,j])-log(rhat[i,k,j])) #Deviance contribution of each category +(n[i,k,j]-r[i,k,j])*(log(n[i,k,j]-r[i,k,j]) - log(n[i,k,j]-rhat[i,k,j]))) } dev[i,k] <- sum(dv[i,k,1:(nc[i]-1)]) # deviance contribution of each arm for (j in 2:nc[i]) { # LOOP THROUGH CATEGORIES p[i,k,C[i,j]] <- 1 - phi.adj[i,k,j] # link function # adjust link function phi(x) for extreme values that can give numerical errors # when x< -5, phi(x)=0, when x> 5, phi(x)=1 phi.adj[i,k,j] <- step(5+theta[i,k,(j-1)])*(step(theta[i,k,(j-1)]-5) + step(5-theta[i,k,(j-1)])*phi(theta[i,k,(j-1)]) ) } } resdev[i] <- sum(dev[i,(1:na[i])]) # summed residual deviance contribution for this trial } z[1] <- 0 # set zRs=0 for (j in 2:(Cmax-1)) { # Set priors for z, for any number of categories z.aux[j] ~ dunif(0,5) # priors z[j] <- z[j-1] + z.aux[j] # ensures z[j]~Uniform(z[j-1], z[j-1]+5) } totresdev <- sum(resdev[]) #Total Residual Deviance d[1] <- 0 # treatment effect is zero for reference treatment for (k in 2:nt){ d[k] ~ dnorm(0,.0001) # vague priors for treatment effects } ©Institute for Clinical and Economic Review, 2020 Page 220 Final Report – Targeted Immune Modulators for UC Return to Table of Contents A ~ dnorm(meanA,precA) for (k in 1:nt) { # calculate prob of achieving >=2, >=3 on treat k for (j in 1: (Cmax-1)) { T[j,k] <- 1 - phi(A + d[k] + z[j])} # calculate prob of achieving ranges [1,2), [2,3), [3, Inf) on treat k T1[k] <- phi(A + d[k] + z[1]) T2[k] <- phi(A + d[k] + z[2])-T1[k] T3[k] <- 1-phi(A + d[k] + z[2]) } # calculate risk ratios for (k in 1:(nt-1)){ for (kk in (k+1):nt){ rr1[kk,k] <- T[1,kk]/T[1,k] rr2[kk,k] <- T[2,kk]/T[2,k] rr1_exc[kk,k] <- T2[kk]/T2[k] rr1[k,kk] <- T[1,k]/T[1,kk] rr2[k,kk] <- T[2,k]/T[2,kk] rr1_exc[k,kk] <- T2[k]/T2[kk] } } } Analysis NMAresults <- jags(data=datalist, inits=jaginits, parameters.to.save = c("d", "z", "T1", "T2", "T3", "rr1", "rr2", "rr1_exc"), model.file = model, n.iter = 100000) ©Institute for Clinical and Economic Review, 2020 Page 221 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix G. Public Comments This section includes summaries of the public comments prepared for the CTAF Public Meeting on September 24. These summaries were prepared by those who delivered the public comments at the meeting and are presented in order of delivery. Three speakers did not submit summaries of their public comments. A video recording of all comments can be found here. Conflict of interest disclosures are included at the bottom of each statement for each speaker who is not employed by a pharmaceutical manufacturer. Maria T. Abreu, MD, University of Miami Miller School of Medicine Professor of Medicine Thank you for the opportunity to provide my comments during the California Technology Assessment Forum (CTAF) meeting on September 24, 2020 regarding the Institute for Clinical and Economic Review's (ICER) Draft Evidence Report assessing the comparative clinical effectiveness and value of treatments for ulcerative colitis (UC). The AGA urges CTAF to consider the unique needs of UC patients and the nature of the disease when making decisions regarding coverage and treatment. Because each patient responds differently to UC treatments, it is important that physicians can tailor the best treatment for each patient. Like many other experts who delivered comments, we urge CTAF not to incorporate step therapy treatment into its model. There is no evidence that step therapy improves health outcomes and it is not supported by guidelines from AGA, ACG, or ASGE. We would like to comment on the Long-Term Cost Effectiveness section of the report. Although we respect the financial analysis performed and accept the data as presented, we question the use of discounted drug prices as shown in the Drug Acquisition Costs section on page 79. The report states: "For all TIMs, we obtained net pricing estimates from SSR Health, LLC, which combines data on unit sales with publicly-disclosed US sales figures that are net of discounts, rebates, concessions to wholesalers and distributor, and patient assistance programs to derive a net price." As shown in Table 5.16, Drug Unit Costs of the report, discounts from WAC range from 18.3% to 62%. The significant range of discounts significantly favors the use of infliximab against use of all of its competitors. Furthermore, since these discounts are not transparent to either the provider space or the patient space, the following results occur: • Value-based care contracts with providers, which are based on shared savings, do not reflect these nontransparent discounts provided to the payer by the manufacturer. ©Institute for Clinical and Economic Review, 2020 Page 222 Final Report – Targeted Immune Modulators for UC Return to Table of Contents • Even in fee for service contracts, nontransparent discounts provided from manufacturers to payers result in the use of brand-named drugs over the use of biosimilars. Although this is financially favorable for the payer, patients suffer. As a result of use of a more expensive drug, patients are paying higher copays. • Patient copays and deductibles do not take these rebates into consideration. Therefore, patients are paying copays and deductibles against the "non-net" payment. • These discounts vary from payer to payer depending on their contractual relationship and are not transparent. • Self-administered drug discounts and rebates vary by the Pharmacy Benefit Manager We would have preferred a standardized payment to be used across all TIMs which would normalize these discounts. Medicare payment rates could be used as a proxy. A second issue is site of service. Due to contractual agreements, TIMs provided in an office setting are significantly less expensive than those administered in hospital outpatient departments. The current analysis does not incorporate site of service contractual differences. Again, a standardized payment would correct for this. A third issue concerns patient preference, which is trending away from infused TIMs to the self- administered drugs. This is due both to drug development as well as patient preference. The administration costs of infused drugs as well as the lost time from work/school must be considered. In conclusion, I urge CTAF to consider the unique needs of UC patients and the nature of the disease when making decisions regarding coverage and treatment, to reject step therapy as it conflicts with AGA, ACG, and ASGE guidelines on treatment of patients with UC, and to consider data from payor sources as you make your determination. Dr. Abreu is a grant/research support consultant. Ed Barnes, MD, MPH, UNC School of Medicine Assistant Professor of Medicine When considering the care of patients with ulcerative colitis (UC) and in particular treatment decisions for patients with UC, I think that it is critical to remember the heterogeneity that is present in both patients with UC and their individual disease presentations. This heterogeneity and diversity may present in multiple different ways including disease severity (and prognosis), comorbidities, and patient preferences, each of which can have direct impacts on the therapy decisions that a patient and their physician or other providers ultimately consider. Given the diversity that is present in these treatment choices, it is critical to avoid prescribed or forced "step therapy" where providers and patients are forced into algorithmic programs. ©Institute for Clinical and Economic Review, 2020 Page 223 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Certainly, as physicians and particularly inflammatory bowel disease specialists, we strive to practice evidence-based medicine. We seek to incorporate all best available evidence into treatment decisions. However, the forced use of sub-standard or non-advised therapies because of payer or other entity recommendations should not be factored into the ultimate shared decision- making process of a patient and their provider. This shared decision-making process is the backbone of the current practice of medicine in the care of patients with UC. Future treatment decisions will always need to incorporate the heterogeneity present in the disease processes, as well as our own knowledge that there is no universally outstanding agent for all patients with UC (to date). In reviewing the current analyses presented by ICER, it is also notable that while based in much of the available evidence, there are gaps in our published literature that do not reflect real-world clinical trial practice. For example, infliximab has not been included in randomized controlled trials after failure of other biologic (or specifically anti-TNF) therapies. This leaves large knowledge gaps in modeling, particularly with relatively common scenarios such as the use of vedolizumab as a first- line therapy (as recommended by the American Gastroenterological Association and other societies) followed by infliximab in the setting of non-response or loss of response to vedolizumab. Additionally, many of the modeled colectomy outcomes may not match patient preferences or the expected outcomes over the lifetime/disease course after an ileal pouch-anal anastomosis. Although it is important to include outcomes such as chronic pouchitis, it may be difficult to fully capture the impact of having a chronic inflammatory condition of the pouch with respect to quality of life as this is a major area of ongoing research and work. This is particularly problematic with a clinical condition that was not modeled such as Crohn's-like disease of the pouch, which affects 10% of patients after IPAA for UC. The impact of a chronic inflammatory condition of the pouch, or even a new Crohn's-like disease process despite the preoperative diagnosis of UC, is difficult to quantify in these models. Similarly, future use of biologics or small molecule therapies for chronic inflammatory conditions of the pouch would further complicate these models. I appreciate the opportunity to participate in this process, and hope that there will be a continued focus on the heterogeneity and diversity in presentations of UC among individual patients with UC that ultimately affect treatment decisions. Dr. Ed Barnes is a consultant for AbbVie, Takeda, Gilead, Pfizer, and Target RWE. Sarah Buchanan, Crohn's & Colitis Foundation Director of Advocacy The Crohn's & Colitis Foundation (Foundation) is concerned about patient-borne medical costs and supports efforts to develop value-based assessments for Crohn's disease and ulcerative colitis (UC) ©Institute for Clinical and Economic Review, 2020 Page 224 Final Report – Targeted Immune Modulators for UC Return to Table of Contents treatments. The Foundation recognizes the efforts by the Institute for Clinical and Economic Review (ICER) to embark on a UC treatment cost-effective assessment. Crohn's disease and UC are chronic conditions of the gastrointestinal tract and impact as many as 3.1 million Americans. These conditions are collectively known as inflammatory bowel disease (IBD). IBD is complex and disease progression is irreversible. As such, patients require individualized treatment and timely access to care. Treatment seeks to improve symptoms, induce clinical remission, and ultimately lead to mucosal healing (that is, no evidence of disease activity seen during a colonoscopy) and histologic remission (that is, no disease activity at the tissue-biopsy level). Because UC is a complex, heterogeneous condition, a one-size approach to treatment will not work. For this reason, the Foundation opposes the application of insurance-mandated step therapy for UC patients. Insurance-mandated step therapy has led to severe and irreversible negative health consequences for patients. There are far too many stories of patients who have required urgent surgery, including total colectomies, because the treatment selected by the patient and their provider was denied, and the patient's disease progressed during the time it took to prove failure of the insurer-preferred treatment. Step therapy interrupts the patient-provider shared decision-making relationship, which is critical for ensuring the unique patient's needs are addressed in the treatment plan. Further, insurance-mandated step therapy protocols are often based on rebate deals or other cost considerations, and not based on clinical guidelines and available medical science. The Foundation urges ICER to include in its public policy report a robust discussion on the harms of insurance-mandated step therapy for UC patients and to recommend against its use in the UC population. In addition, the Foundation is concerned that ICER's cost-effectiveness model does not include real- world evidence (RWE) and is therefore inadequate to reflect the value of UC treatments accurately for the range of patients affected. In order to be taken up by the UC community, value-based assessments must reflect the full range of patient experience. The Foundation has built an infrastructure and resources to gather real-world data and evidence. We welcome the opportunity to continue to engage with ICER to design value-based assessments that can address cost- effectiveness and patient-centered questions in a meaningful way for the IBD community. The Foundation thanks ICER for its consideration of our views. We look forward to engaging with ICER and other stakeholders on these critical discussions related to treatment costs and value in order to improve the ability of patients to access and afford appropriate care. No financial conflicts to disclose. ©Institute for Clinical and Economic Review, 2020 Page 225 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Emily Morgan Patient Expert • Diagnosed at age 13. • Days were consumed by school and competitive swimming. First, I battled mono for a few months and then it seemed like my energy never got back to how it was beforehand. Severe cramping after meals and urgent bathroom visits with blood resulted in me getting a colonoscopy and there was no question as to what was going on- it was UC. Fast forward three years- I had been on four different medications, each of them failing, which meant I had to take prednisone to get the inflammation back under control. Finally, after being hospitalized the two weeks leading up to my 16th birthday, my doctor and I (along w my family) decided it was time to speak to a surgeon. Knowing that there was a procedure that could make it all go away made me excited and nervous. Being 16, I'm not sure that I truly knew the magnitude of the two-step procedure. But what I did know was that I would NOT miss my colon. • After a bit more time and consult, I ultimately decided to have a colectomy- a two-step surgery- during the summer between 10th and 11th grade. It wasn't easy but, for me, having surgery was the best decision I have ever made. Since that summer, I have never been hospitalized for my UC. It hasn't been a totally smooth ride but my overall quality of life is significantly better than it ever was, I have found balance in my disease, and it's important to me to use my voice and experience to advocate for others. That's just the thing…. Others haven't had the same experience as me. As you all have heard and been researching- UC is heterogeneous- no two patients (that I know of) are the same. We even have different symptoms sometimes and different food triggers is VERY common in my experience. For example, one of my very close UC friends had surgery a few years back and is unfortunately still struggling to reach her desired quality of life. It's frustrating and confusing. And so with that- I really want to focus on the topic of colectomy. ICER's report seems to indicate that a colectomy is more cost- effective than the studied treatments. That's great; however, I know for a fact that each patient's goals and preferences are different. Not everyone bounces back after surgery and lives their "best life." As you know, I did opt for a colectomy, but for others, it's the last resort. And not to mention this was over 10 years ago. Now- there are even more approved therapies on the market for UC patients. I sometimes wonder if I would've been that ready and willing had there been more options to try. ©Institute for Clinical and Economic Review, 2020 Page 226 Final Report – Targeted Immune Modulators for UC Return to Table of Contents I wouldn't go back and change my decision, but surgery brings its own set of barriers and considerations. For me- it's scars and body image battles and then, most recently- the topic of conception and pregnancy. I speak to about four to five patients/year- connected via my gastroenterologist to discuss my experience with a colectomy. I've heard firsthand about other patient's fears, concerns, and treatment goals. No two patients are ever the same. The obvious and ugly truth is that there is no cure for UC. So as a result, patients are having to figure out what is most important to them- how they can best manage their disease and still live a life as close as possible to the one they desire. For some, that may be completing an Iron Man, some it's being able to build a family and have children, and for others it's being able to attend their niece and nephews sporting events on an 85- degree weekend without having to run to the bathroom or take a nap afterward. To conclude, as a UC patient I think that it is imperative that health plans should not steer patients towards treatment strategies based on cost effectiveness data--- much less data that isn't nuanced enough to reflect the patient experience or perspective, but rather a patient's plan should allow patients to make treatment decisions with their provider and loved ones. Emily Morgan is a full-time health behavior medical writer for IQVIA. Rosanne Mottola, Crohn & Colitis Foundation Volunteer Patient Expert Thank you to the ICER for inviting me to speak today and for ensuring that ulcerative colitis patients are heard during this critical meeting. My name is Rosanne Mottola, and I am 35 years old. I am a public affairs director for a NYC hospital, I'm a volunteer with the Crohn's and Colitis foundation and – most importantly – I am a wife and a mother to my two young children Ryan and Abigail. 14 years ago, I did not think any of these things would've been possible. I was 21 years old when I was diagnosed with Ulcerative Colitis. I was a few months away from graduating college and I was deathly ill and incredibly scared. I was housebound and in excruciating pain at one of the most important times of my life. My doctor at the time, who claimed I had only a mild case, would put me on high doses of steroids every few months which had devastating side effects and impacted both my physical and mental well-being. My health quickly deteriorated and I was hospitalized for the entirety of fall 2006. I had to medically withdraw from my first semester of graduate school at NYU. At this time, I met my current gastroenterologist who worked alongside me in exploring treatment options. The most ©Institute for Clinical and Economic Review, 2020 Page 227 Final Report – Targeted Immune Modulators for UC Return to Table of Contents frustrating part about fighting UC is the fact that no two patients are the same. There is no roadmap. There is trial and error, and time is critical. No two patients react to medications the same way, have the same food sensitivities, and no two patients' journeys are identical. We must treat ulcerative colitis patients individually. Therefore, I was put on a number of different biologics and dealt with the pain and stress of trying new therapies. At the same time, I was planning my wedding and trying to hold down a job to remain on my health insurance. That's when my health took a turn for the worse and I was facing a colectomy. There's a lot of differing opinions on colectomies among the members of the UC patient community. Through my advocacy work with the CCF I know a lot of patients who are leading healthy lives post colectomy. However, I also know many patients who have had difficulties post- colectomy and whose symptoms have manifested elsewhere. All of these patients are warriors and I'm proud to call them my friends. However, cost of treatments should not be the determining factor in an UC patient's decision to have a colectomy. Almost every UC patient I've met were diagnosed during their formative years. If there are medications that can be tried and a patient is willing to try them, then they shouldn't be forced in any way to choose a colectomy because of cost. Plus, we shouldn't pretend that colectomies are cheap either: the surgery itself is outrageously expensive, and lifelong ostomy supplies are too. Personally, I was in my early 20s and was terrified at the thought of a colectomy and wanted to exhaust all of my options first. While I met with surgeons, my gastroenterologist - who is my partner in my care - put in a last-ditch effort. By prescribing an immunomodulator and a double dosage of an infusion biologic, I finally found the mix of medications that worked for me and kept me in remission until this day. Through this magical mix that has worked so well for me, I was able to finish graduate school, travel with my husband, advance my career, and deliver two healthy babies. All on the same biologic I started on a decade ago. I often think of what my life would be like if my doctor had continued to treat me with courses of steroids or jumped the gun on performing a colectomy. Today, please let me leave you with these two thoughts: 1. When considering the voting questions, please remember the age of diagnosis of most ulcerative colitis patients and how these decisions impact the quality of rest of their lives. 2. In a lot of cases (like mine) finding the right mix of medications is key to proper treatment, and with proper treatment ulcerative colitis patients can avoid costly surgeries, hospitalizations and future complications. ©Institute for Clinical and Economic Review, 2020 Page 228 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Thank you to the Crohn's and Colitis Foundation and ICER for giving me the opportunity to speak today once again and for your time and attention. No financial conflicts of interest to disclose. Phil Naughten, PharmD, Takeda Pharmaceuticals Vice President, HEOR Takeda is committed to bringing better health and a brighter future to patients by translating science into highly innovative medicines. We are dedicated to achieving this objective through pricing responsibly and working with stakeholders to ensure patient access. Recognizing the unmet need for managing chronic gastrointestinal diseases, Takeda has developed vedolizumab, a humanized monoclonal antibody, α4β7 integrin receptor antagonist, indicated for the treatment of adult patients with moderately to severely active ulcerative colitis (UC) and Crohn's disease (CD). It is the only gut-selective treatment option that provides a favorable benefit-risk profile for moderate to severe UC patients, as well as unique benefits to the US health care system. I would like to underscore the broad evidence from both RCT and RWE that have demonstrated the efficacy, safety, and real-world effectiveness of vedolizumab IV. The efficacy and safety of vedolizumab is supported by the results of the pivotal GEMINI clinical trial program, where significant and durable improvements in response and remission, and better mucosal healing were observed among bio-naive and bio-experienced patients. Data from the GEMINI long-term safety (LTS) study (up to nine years follow-up) demonstrated that vedolizumab was safe, well-tolerated, and continued to achieve favorable clinical outcomes for long-term IBD treatment. In the first and only head-to-head IBD clinical trial of biologics, VARSITY, which was published in the NEJM in September 2019, vedolizumab was superior to adalimumab in achieving clinical remission and endoscopic mucosal healing in the overall and anti-TNF Naïve UC Populations. In addition to the clinical trial programs, substantial real-world data documents the effectiveness and safety of vedolizumab through both the VICTORY consortium and EVOLVE study. VICTORY is a multicenter collaborative effort among US institutions to prospectively maintain a registry of real- world outcomes of IBD patients treated with biologics. Results from VICTORY demonstrated effectiveness in achieving clinical remission, endoscopic healing, and favorable safety profile of vedolizumab vs. anti-TNF therapies. EVOLVE, a real-world retrospective medical chart review study evaluating the long-term outcomes (up to 24 months) in biologic-naïve vedolizumab patients with IBD, also demonstrated real-world effectiveness, persistence, and safety of vedolizumab. Taken together, these data not only complement recent published literature, but also have been recognized and incorporated into guideline recommendations. Vedolizumab's gut-selective ©Institute for Clinical and Economic Review, 2020 Page 229 Final Report – Targeted Immune Modulators for UC Return to Table of Contents mechanism of action is well aligned with the 2019 ACG recommendation of starting with organ- selective biologic treatments before the use of systemic therapies, and the 2020 AGA guidelines suggested using vedolizumab as one of the first-line biologic therapies over adalimumab among adult biologic-naïve patients. Additionally, I would like to make a few comments on the analysis itself. Several limitations of this analysis impact interpretation of value, and we hope that CTAF will be able to take these details into consideration. 1. The base-case comparator considered in this assessment was conventional therapy, which may include immunomodulators and/or immunosuppressants. The assumption that conventional therapy is a potential option for treating moderate-to-severe UC patients contradicts 2020 AGA and 2019 ACG clinical guidelines. Additionally, payers considering formulary placement of new biologics usually compare their value against existing biologics. Representing the current state of clinical management for UC is a critical part of assessing the value of therapies. 2. Health state utilities are important variables in the model, and despite a change to ICER's original data sources, the model now relies on data from one single Australian study that does not align with values used in prior HTA submissions. This data source now minimizes the differences between active UC, response, and remission health states. In addition, with no adjustment to the post-colectomy utility source, the model input suggested that the quality of life among patients living with remission is nearly equivalent compared to those with colectomy. The revised model inputs therefore continue to underrepresent the importance of improving the health of patients living with UC, including avoiding the need for surgery. 3. ICER's review was solely based on RCT data. In UC management with biologics, outcomes such as mucosal healing and its associated benefits are well captured in real-world data. Indeed, due to the long-term nature of therapy in UC as well as change in effects of therapies with long- term use, inclusion of real-world data would likely improve modeled insights. I urge CTAF to keep these limitations of ICER's assumptions and inputs in mind while deliberating on the value of biologics such as vedolizumab IV. Ultimately, we seek to see all products assessed according to their full holistic value to patients and society, and Takeda supports appropriate analyses that incorporate elements that are important to patients and reflect real-world clinical practice. Phil Naughten, PharmD, is a full-time employee of Takeda. ©Institute for Clinical and Economic Review, 2020 Page 230 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Jordan Wilson Patient Expert I'm a 37-year-old Realtor out here in southern California, indoor cycling instructor and most importantly, a patient advocate for the IBD community. My journey with ulcerative colitis started just over ten years ago. After showing minor symptoms I was diagnosed with UC back in 2010. Luckily, I was put into remission rather quickly following my diagnosis. That all changed about 18 months later when my symptoms came back and I went into a flare, which my gastroenterologist said was the worst he'd seen in five years. I went from living a "normal life" to not being able to eat, not being able to sleep, exercise, or work. My colon went from having only a minor case of ulcers, to the entire organ being diseased and inflamed, all in less than two years. For a year and a half, I was on and off steroids, biologics, immunomodulators, and unfortunately nothing seemed to work for me. I was hospitalized several times and actually almost died from complications of ulcerative colitis twice. I was referred to a second GI and she suggested surgery. I was explained what a colectomy was and how a j-pouch is constructed. With my back against the wall and viewing this option as truly a last resort, I decided to go through with the procedure. I'm beyond grateful to say that I am one of the lucky ones, even after going through everything I just mentioned. I have been living with a relatively healthy j-pouch since 2013. Unfortunately, that can't be said for everyone. Through my advocacy work I have met numerous patients who, while having the same procedure as me, aren't having the same success. These varied responses to medication and surgeries just confirm what we already know, ulcerative colitis is such an individual disease and no two patients react to treatments the same way. No matter how positive I try to remain, there are very different opinions on colectomies throughout the patient community. There are stories of the disease reappearing, chronic pouchitis, or having to remove the j-pouch and go back to living with an ostomy. Quality of life and the efficacy of treatments should be the main consideration when treating an ulcerative colitis patient, not cost. Since each patient is different, each treatment will have to be as well. No financial conflicts of interest to disclose. ©Institute for Clinical and Economic Review, 2020 Page 231 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Appendix H. Conflict of Interest Disclosures Tables H1 through H3 contain conflict of interest (COI) disclosures for all participants at the CTAF September 24 public meeting. Table H1. ICER Staff and Consultants and COI Disclosures Name Organization Disclosures Lisa Bloudek, PharmD, MS Senior Research Scientist, University of Washington * Pamela Bradt, MD, MPH Chief Scientific Officer, ICER * Associate Professor, Department of Pharmacy, University Josh J. Carlson, PhD, MPH * of Washington Rick Chapman, PhD, MS Director of Health Economics, ICER * Laura Cianciolo Program Manager, ICER * Katherine Fazioli (Former) Research Lead, Evidence Synthesis, ICER * Monica Frederick Program and Event Coordinator, ICER * Director, Value Measurement & Global Health Initiatives, Daniel A. Ollendorf, PhD Center for the Evaluation of Value and Risk in Health, Tufts * University Medical Cent Rajshree Pandey, PhD, MPH Research Lead, Evidence Synthesis, ICER * Steven D. Pearson, MD, MSc President, ICER * *No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. Table H2. CTAF Panel Member Participants and COI Disclosures Name Organization Disclosures Ralph G. Brindis, MD, MPH, Clinical Professor of Medicine, UCSF * MACC, FSCAI, FAHA Felicia Cohn, PhD Bioethics Director, Kaiser Permanente, Orange County * Rena K. Fox, MD Professor of Medicine, UCSF * Paul Heidenreich, MD, MS Professor and Vice-Chair for Quality, Clinical Affairs, and * (Chair) Analytics, Stanford University Associate Director, Center for Healthcare Policy and Research, Jeffrey Hoch, PhD * UC Davis Jeff Klingman, MD Chief of Neurology, The Permanente Medical Group * Regional Lead for Clinical and Translational Neuroscience, Annette Langer-Gould, MD, Southern California Permanente Medical Group, Kaiser * PhD Permanente Professor of Clinical Medicine, UCSF; Chief, Division of Elizabeth J. Murphy, MD, DPhil Endocrinology and Metabolism, Zuckerberg San Francisco * General Hospital Professor of Health Economics and Health Services Research; Director and Founder, UCSF Center for Translational and Policy Kathryn A. Phillips, PhD Research on Personalized Medicine; Department of Clinical * Pharmacy/School of Pharmacy, UCSF Institute for Health Policy Studies, and UCSF Comprehensive Cancer Center ©Institute for Clinical and Economic Review, 2020 Page 232 Final Report – Targeted Immune Modulators for UC Return to Table of Contents Assistant Professor, Department of Radiation Oncology, UCLA Ann Raldow, MD, MPH * David Geffen School of Medicine Rita F. Redberg, MD, MSc, Professor, UCSF School of Medicine * FACC Associate Professor of Medicine, University of California San Sei Lee, MD * Francisco Alexander Smith, MD, MPH Associate Professor of Medicine, UCSF * Joanna Smith, LCSW, MPH, Chief Executive Officer, Healthcare Liaison, Inc. * CHA Patient Advocate and Lead Volunteer, California, National Anthony Sowry Patient Advocate Foundation; Senior Vice President, Maritime * Container Shipping (Retired) *No relevant conflicts of interest to disclose, defined as more than $10,000 in healthcare company stock or more than $5,000 in honoraria or consultancies during the previous year from health care manufacturers or insurers. Table H3. Policy Roundtable Participants and COI Disclosures Name Organization Disclosures Thomas Brownlie, PhD, MS Senior Director, US Payer Policy, Pfizer Full-time employee of Pfizer. Leader, US Payer Marketing, Primary Deborah Gan, MBA Full-time employee of Merck. Care, Merck Assistant Vice President, Health Full-time employee of Prime Patrick Gleason, PharmD Outcomes, Prime Therapeutics Therapeutics. Dr. Sands has consulted for AbbVie, Janssen (service on a Professor of Medicine and Chief, scientific advisory board), Pfizer Bruce Sands, MD, MS Division of Gastroenterology, Mount (service on a scientific advisory Sinai board), and Takeda (honoraria for speaking in a CME program). Dr. Singh has received research Assistant Professor of Medicine, UC grants from AbbVie, and San Diego School of Medicine; Siddharth Singh, MD consulting fees from AbbVie, Gastroenterologist, UC San Diego Takeda, Pfizer, AMAG Health Pharmaceuticals. No conflicts of interest to Megan Starshak Patient Advocate disclose. Director, Inflammatory Bowel Disease Dr. Velayos' spouse has received Fernando Velayos, MD Program, Kaiser Permanente Northern honoraria in consulting fees. California Senior Vice President, Education, No conflicts of interest to Laura Wingate Support, and Advocacy, Crohn's & disclose. Colitis Foundation ©Institute for Clinical and Economic Review, 2020 Page 233 Final Report – Targeted Immune Modulators for UC Return to Table of Contents