^X Pneumonia Following Influenza (> died. Of the patients with influenza, therefore. 12.1 pel cent, developed pneumonia, and the mortality for pneumonia was 31 per cent. The mortality for the epidemic as a whole was 3.S per cent, of those attacked by influenza. The most striking feature of the epidemic was the extremely rapid spread of the infection throughout the camp. Starting September 22 in a regimental area situated in the extreme southwestern corner of the camp, the disease had within four days appeared throughout the camp proper, and four days later appeared in two outlying encampments, situated respectively 1 mile north and 3 miles east of the main camp. Xo single organization escaped the infection, and during the thirty days from September 20, 23.3 per cent, of the total population of the camp suffered from the disease. The figures given herewith are based on a detailed study of 11.725 cases occurring between Sep- tember 20 and October 14, and comprise practically the whole of the epidemic. The disease displayed a striking selective incidence among the men who had been in camp less than one month. During the period from August 20 to ()ctober 1 there had arrived in camp 23,216 new men, and dur- ing ()ctober 553 additional men arrived. Among these new men there developed 7,26X cases of influenza, an attack rate of 30.6 per cent. There were in camp, ( 'ctober 1, 2S.7S2 men who had been in camp prior to August 20. Of these, 4,462 developed influenza, an attack rate of 15.5 per cent., or almost exactly half the attack rate of the new men. This disparity between the attack rates of old and new men was true of both white and colored men. As between white and colored men there were marked and interesting differences in the rates of attack. The total exposed population, including those in camp October 1, and those arriving during October, wa» 52.551 men, of whom 41,77X were white and 10,773 were colored. The white men developed 10,296 cases of influenza, a rate of 24.6 per cent., while the negroes developed 1,429 cases, a rate of 13.3 per cent 3 That this difference in susceptibility is not due to any essential racial immunity is shown by the experi- ence of a draft of 1,954 negroes who arrived in camp between September 23 and 2*. Of these men, 1,860 were transferred between September 27 and Septem- ber 30 to the negro tent colony at Fort Roots. In this camp at the time of their arrival there were 2,130 negroes, most of whom had been in camp six weeks or longer. There had been no recognized cases of influenza in this camp prior to the arrival of the new draft. September 30, one of the new men developed TABLE 1.—NUMBER OF CASKS OF INFLUENZA AND PNEU- MONIA AT CAMP PIKE, ARK., FROM SEPT. 1 TO OCT. 31, 1918, BY DAYS OF ONSKT I>ay <-------Cases-------^ Day ,_______Cases_______ September Influenza Pneumonia October Influenza Pneumonia \ (8)* 1 1 864 126 2 (8) 3 2 1,214 142 3 13 2 3 857 113 * O4) 1 4 572 129 5 (8) 3 5 202 86 5 dD 3 6 252 89 I (20) 4 7 226 93 8 (16) 3 8 193 76 0 (23). 1 9 224 4" 10 (25) 3 10 132 m 11 25 4 11 141 54 12 (15) 1 12 72 34 18 (27) 4 13 36 15 14 (27) 7 14 44 13 15 (30) ]() if, 12 9 16 (10) 10 l(j 21 19 H CSV) 7 17 13 a 18 (52) 5 18 17 ,s 10 (12) 11 19 l(i 9 20 (50) 12 20 18 11 21 (57) 10 21 13 5 (SO) 13 22 5 6 214 15 11 24 436 13 24 ;; 5 25 885 25 25 s 7 26 797 16 20 0 3 27 1,037 23 27 3 1 28 992 30 28 0 29 993 40 29 3 3 30 1,166 35 30 1 0 31 3 0 Totals....................................... 12,393 1,499 Died.......................................... 2 466 » Cases in parentheses, admitted with diagnosis of acute bronchitis. influenza, followed on the next day by forty-seven cases among the new men, and on the next, October 2, by 462 cases, including a considerable number among the old men. Between September 30 and October 14 there were 972 cases of influenza in the camp, with 107 cases of pneumonia. Of the 972 cases, 810 were in the men of the new draft and 162 in the old men in camp. 4 The rate of incidence for the new men was 43.6 per cent., about as high as that of any organization in Camp Pike, while that of the old men was 7.6 per cent., approximately equal to that of negroes of similar length of service in Camp Pike. The time of development of pneumonia in relation to influenza is shown in the table. In general, the peak of the curve of pneumonia followed approxi- mately one week after that of influenza, although the highest single day for both coincided. The general incidence of pneumonia followed that of influenza, and the preponderance of new men among the pneumonia cases was even more striking than among the influenza cases. Of the pneumonia patients, 1,006, or 67.5 per cent., were new men, while 62 per cent, of tlie patients with influenza were new men. The rates of incidence of the two diseases in old and new men are given in Table 2. TABLE 2.-IXCIDKXCK IN OLD MIX AND NEW MKN I'opula- /---Influenza---n ,— 1'ueuiiioiiln (ion No. PerCcnt. No. Per ('cut Old men................ 27.782 4.162 15.6 493 1.7 New men.................. 23.769 7,203 30.6 1,006 4.2 All men................... 51,551 11,725 22.7 1,499 2.9 Of the new men with influenza, 13.8 per cent, devel- oped pneumonia, as against 11 per cent, of the old men. The incidence of pneumonia in cases of influenza was much higher among the negroes than among the whites. Of 10,296 white patients with influenza, 12.7 per cent, developed pneumonia, while among 1,429 negroes with influenza, 283, or 19.8 per cent., developed pneumonia. The rate of fatality from pneumonia was slightly higher in white men than in negroes. Of 1,216 white pneumonia patients, 386, or 31.7 per cent., died, while of 283 negro pneumonia patients eighty, or 28.2 per cent., died. September 24, the base hospital contained 2,789 patients; the hospital had been planned to care for 2,009 patients. From September 24 to 30 from 188 to 486 patients were admitted daily to the hospital, so that on September 30 the hospital contained 3,587 patients. October 5, the number of patients reached the maximum, 4,233. September 24, by order of the camp surgeon the milder cases of influenza were treated in barracks. livery agency for the receiving, 5 the housing and the care of patients throughout the camp was strained by a demand far beyond that for which it had been planned. Patients were crowded in the infirmaries, in the ambulances, in the receiving room of the hospital, in the wards and even in corri- dors between the wards, where it was -essential to shelter them. It was necessary to open and organize in great haste, with inadequate equipment, wards for the receipt of patients with pneumonia. The medical personnel of the hospital was large at the time; the nursing corps was organized for the care of 2,440 patients; one quarter of them (sixty-two of 240) were attacked by influenza. The pressing need of diminishing the overcrowding of the hospital was immediately recognized fry the commanding officer, and through the authorities of the camp, barracks previously used by three replacement' regiments about a quarter of a mile distant were placed at the disposal of the hospital. Patients were moved from the overcrowded hospital into this annex as fast as the facilities available permitted. October 3, 523 patients were transferred to the annex; on the 4th, 452; on the 5th, thirty-two, and on the 6th, 387. In the main hospital, twenty wards for patients with pneumonia were added to the two that already existed. The installation of cubicles and other necessary mea- sures required some time, and during the early period arrangements for the care of patients with pneumonia. in some of the new wards were necessarily wholly unsatisfactory. With relief of overcrowding, condi- tions rapidly improved. The influence of overcrowd ing on pneumonia will be discussed later. INFLUENZA The influenza that has occurred here has been char- acterized by sudden onset with chilliness and sharp elevation of temperature, often from 103 to 105 F.' There has been extreme prostration, severe backache, suffusion of the face, and injection of the conjunc- tivae. Coryza, pharyngitis and tracheitis with a har- rassing cough have been almost invariable; epistaxis and slight hemoptysis, frequent. In the majority of cases the temperature has subsided after from two to five days, usually rather abruptly. About one third of the patients have developed purulent bronchitis. 6 Search has been made for Bacillus influenzae in a group of twenty-three patients from one to six days after the onset of the disease. From each individual a culture on blood agar ( 5 per cent, horse's blood in meat, infusion agar) has been made (a) from the nose, (b) from the throat and (e) from sputum; and (d) sputum has been injected into the peritoneal cavity of a white mouse. TABLK 3.—PRKSKXCK OF B. IXF1.UKNZAK IX TWINTV-THKKK CASKS OF INKMKNZA • Number Disease Nose Throat ( ultnre through M 1 1 + + + -1- 2 4 — + + -1 3 5 — — + — 4 4 — — + + r) 3 — — + + (i 4 _ -I + r 7 •J — + — » > 4 + + + — 9 "i — + + -V 10 2 + — — — 11 •> — + ? + 12 3 ? + + + 16 3 — — — + 14 2 — — + + 15 3 ? — — + lti 1 — + + + 17 3 .— + — + 18 4 + f ? + 19 fi — — + + 20 1 — + + + 21 1 — 4- — + 22 \ + — + + 23 3 ? — — + Tot ii Is 5 13 14 18 • The interrogation point Indicates that the plate was contaminated. Multiple cultures have demonstrated in some instances in almost pure culture the presence of B. influenzae in all of these cases of early influenza. Pas- sage of sputum through the white mouse has proved the most effective means of demonstrating the organ- ism ; cultures from the sputum or throat have been nearly as effective. There is some difficulty in demon- strating the organism in consequence of the minute size of colonies, which may be wholly overlooked by those not familiar with its cultural characters, in the presence of other organisms; for example, in the zone of hemolysis of the hemolytic streptococci the colonies become conspicuous. BRONCHITIS WITH INFLUENZA A considerable number of those attacked by influ- enza develop bronchitis. Of 103 influenza patients kept under observation, thirty-six had purulent bron- 7 chitis. The sputum is profusely mucopurulent and sometimes streaked with blood. In these cases numer- ous moist rales may be heard. The sputum on direct smear or on culture almost invariably shows B. influ- enzae, often in great numbers. In attempting to establish the relation of B. influ- enzae to influenza and its complications, we have borne in mind the observation made at Camp Funston that B. influenzae was found in the mouths of 35.1 per cent. of all healthy men examined and was present, in the absence of an epidemic of influenza, in the sputum of a very large proportion of those suffering with bron- chitis. Our observations here have shown that the organism is invariably present in the upper respiratory passages of patients with influenza. PNEUMONIA WITH INFLUENZA Pneumonia has occurred in 12.1 per cent, of all influenza patients in this camp, and all pneumonia patients have been treated in the base hospital. Among 103 influenza patients selected for observation from the onset of influenza, four developed clinical evidence of bronchopneumonia and three of lobar pneumonia. The clinical features of pneumonia following influ- enza may be best described by division into three groups. It should be borne in mind, however, that the picture has been a complex one and that correct clinical interpretation is not always possible, since many cases do not conform sharply to any one type. These groups are: (1) bronchopneumonia; (2) lobar pneumonia, and (3) lobar pneumonia with purulent bronchitis. In the first group, bronchopneumonia has usually developed gradually as a sequence to influenza in which purulent bronchitis has occurred, one condition passing into the next without sharp demarcation. In the sec- ond group apparent recovery from influenza has occurred, as evidenced by fall of temperature to normal. After from one to three days of normal tem- perature, typical lobar pneumonia with characteristic rusty sputum has developed suddenly. In the third group, lobar pneumonia has developed in cases of influenza that have been complicated by purulent bron- chitis. Such cases have occasionally presented the picture of both lobar pneumonia and bronchopneu- monia in the same individual. 8 Bacteriologie study has disclosed that all these types of pneumonia have been of pneumococcus origin in most instances. As described in greater detail below, some have been further complicated by a superimposed Streptococcus hemolytic us infection. The bacteriology of the sputum early in the disease has been studied by inoculation of white mice and by direct cultures. The occurrence of pneumococci, B. influenzae and Streptococcus hcmolyticus in sixty-nine cases of lobar pneumonia has been as follows: Pneumococcus Type I, nine times (once with Type IV), or 13.1 per cent.; Type II, three times, or 4.3 per cent.; Type II atypical, fifteen times, or 21.7 per cent.; Type III, five times (once with Type IV), or 7.3 per cent; Type IV (alone), thirty-seven times, or 53.6 per cent. Streptococcus hcmolyticus has been found three times, or 4.3 per cent., in all three instances associated with pneumococ- cus Type IV. B. influenzae has been found forty-three times, or 62.2 per cent., always associated with pneumococcus (Type I, six time-; II, once; II atypical, ten times; Til, twice, and IV, twenty-four times). Pneumococcus Types I and IT, which have a pre- dominant part in the production of the pneumonia of civil life, have been present in a relatively small pro- portion of cases, whereas Type 11 atypical and Types III and IV are found with 82.6 per cent, of cases. The bacteriology of the sputum in forty-three eases of bronchopneumonia was as follows: Pneumococcus Type I, none; Type II, once, or 2.3 per cent.; Type II atypical, three times, or 7 per cent.; Type III, three times, or 7 per cent.; Type IV, thirty times, or 69.8 per cent Streptococcus hemolyticus has been found six times, or 14 per cent., twice with pneumococcus Type IV and four times with no pneumococci. B. influenzae has been found thirty-eight times, or 88.4 per cent., associated with pneumococci thirty-three times (Type II, once; II atypical, three times; III, three times; IV, twenty-six times) ; with Streptococcus hemolyticus three times, and alone twice. Here again pneumococcus Type IV has been found predominant, being present in more than two thirds of the cases. Streptococcus hcmolyticus unassociated with pneumococci has been found in 9.3 per cent, of these cases. B. influenzae unaccompanied by pneumo- cocci or hemolytic streptococci was present in two 9 cases; but in view of observations made at necropsy, it is doubtful if this organism alone was responsible for pneumonic consolidation of the lungs. PATHOLOGY AND BACTERIOLOGY OF PULMONARY LESIONS FOUND AT NECROPSY The characters of the group of pneumonias that have occurred in association with the present outbreak of influenza may be defined by the pulmonary lesions found at necropsy, described in relation to the asso- ciated bacteria. Two hundred necropsies have been performed dur- ing the outbreak of pneumonia. For the purpose of the present report it has seemed desirable to analyze seventy-nine cases which, occurring with few excep- tions in sequence at the height of the outbreak, have been subjected to careful bacteriologic study. Cultures have been made from the heart's blood, from the lung and from a bronchus. When the hemolytic strepto- coccus has been found in the lung it has usually been demonstrable in the blood. The number of instances in which B. influenzae has been obtained in cultures would have been diminished more than half if cultures from the mucosa of the bronchi had not been made. CHANGES IN THE BRONCHI Necropsies have shown the presence of purulent bronchitis in a large proportion of those who have died with pneumonia during the course of the present epi- demic of influenza. When edema of the lungs is present, the content of the small bronchi is not usually purulent, though the mucosa may be intensely injected. B. Influenzae can be grown with few exceptions from the mucopurulent material that wells out of the cut bronchi or from material scraped with a platinum loop from the main branches. In twenty-seven of thirty instances of purulent bronchitis from which cultures were made at necropsy, B. influenzae was found. When purulent bronchitis is found, the lungs are very voluminous and preserve the shape and size of the thoracic cavity after removal; they show little tendency to collapse, even when cut. Most of the affected lungs, being from men who have come from rural districts, are pale pink and almost wholly free 10 from coal pigment. After section a small droplet of mucopurulent fluid marks the site of each bronchus or bronchiole and doubtless explains the failure of the lung to collapse. The cyanosis of the patients with the disease is doubtless referable in part at least to the same change. The mucosa of the bronchi is intensely injected. Further evidence of severe injury to the bronchi is the frequent occurrence of a zone of hemor- rhage about the smaller bronchi, particularly in the lower lobes. At times this zone of hemorrhage gives place to an encircling zone of pneumonic consolidation. Further evidence of profound injury to the bronchial wall is the dilatation that the small bronchi undergo. Bronchiectasis is most conspicuous in the basal part of the lower lobes, and is usually more advanced on the left side than on the right. Small bronchi with no car- tilage in their wall may reach a diameter of 0.5 cm More advanced bronchiectasis has been found in sev- eral necropsies performed late in the outbreak. Among seventy-nine necropsies, selected from the 200 performed, because they have been in sequence and accompanied by careful bacteriologic study, there have been thirty-six instances of lobar pneumonia, nineteen instances of bronchopneumonia, one instance of asso- ciated lobar pneumonia and bronchopneumonia, ami twenty-three instances of pneumonia with suppuration of lung tissue. LOBAR PNEUMONIA A group of thirty-six cases of pneumonia, taken at necropsy in sequence and studied bacteriologically, gives further knowledge of the bacteriology of the dis- ease as it has occurred here. The results of this study have been as follows: Pneumococcus Type I, three times; Type II, none; Type II atypical, five times; Type III, six times; Type IV, nineteen times; total, thirty-three, or 91.7 per cent. Streptococcus hemolyticus with pneumococci, ten times; with no pneumococci, three times; total, thirteen, or 36.1 per cent. B. influenzae thirty-one times, or 86.1 per cent. B. influenzae has been present with few exceptions. Of pneumococci, which are frequently found in the mouths of healthy men, Type II atypical and Types III and IV are predominant, Type IV occurring in more than half of all cases, whereas Type I has been 11 found only three times in the present series, and Type II has been found in no typical instance of lobar pneu- monia. Hemolytic streptococci have been found in approximately one third of those cases of lobar pneu- monia that have been unassociated with gross evidence of suppuration. It seems probable that the organism has invaded the lung late in the disease, and doubtless the micro-organism in many instances has had an important part in producing the fatal issue. This sub- ject will be considered later when ward infection is discussed. When the hemolytic streptococcus has been unassociated with the pneumococcus it is probable that the latter has disappeared from that part of the lung from which the culture has been taken. BRONCHOPNEUMONIA Two varieties of bronchopneumonia have been observed repeatedly. They do not include all of the instances of bronchopneumonia that occurred: (a) Bronchopneumonia may cause confluent consolidation of large parts of lobes; patches of consolidation, accu- rately limited to the secondary lobules of the lungs, occur. The larger patches of consolidation are sharply limited by lobule boundaries so that the consolidated tissue on the cut section projects conspicuously above the air-containing lung substance. Even within the large consolidated areas, which are blackish red, red oj grayish red and finely granular, the lobules are well defined; but there is no thickening of the interstitial tissue. Two instances of pneumonia due to pneumo- coccus Type II were of this character. In three other cases the lesion was caused by pneumococcus Type IV. (b) This confluent lobular consolidation is not the pre- dominant variety, and disseminated nodules or patches of reddish, gray or yellow consolidation are more com- monly seen. The bacteriology of nineteen cases of bronchopneumonia has been as follows: Pneumococcus Type I, none; Type II, twice; Type II atypical, none; Type III, none; Type IV, ten times; total, twelve, or 70.6 per cent. Streptococcus hemolyticus with pneumococci, seven times; with no pneumococci, five times; total, twelve, or 70:6 per cent. B. influenzae, seventeen times, or 89.4 per cent. 12 Pneumococci doubtless have a predominant part in the production of the disease, and hemolytic strepto- cocci act as secondary invaders. Nevertheless it im- possible that hemolytic streptococci in the presence ot influenza may independently invade the lung and pro- duce bronchopneumonia without suppuration. SUPPURATIVE PNEUMONIA Suppuration has occurred in more than one fourth of the cases that have been examined postmortem. Three varieties of suppurative lesion have been found. (a) Localized abscess formation within a patch of pneumonic consolidation. A patch of consolidation about one abscess is occasionally the only evidence of pneumonia, but more frequently there are multiple patches of bronchopneumonia. The abscess or abscesses are usually situated immediately below the pleura, often separated from the cavity by remains of the membrane no thicker than tissue paper, and the cavity is the site of purulent pleurisy with effusion varying from several hundred up to 1.700 c.c. In all but one of eleven cases of abscess with bronchopneu- monia, hemolytic streptococci were found both in the blood of the heart and in the affected lung. In the one exceptional case, pneumococcus Type IV was obtained from the blood; no growth was obtained on the plate inoculated from the lung, and hemolytic streptococci with B. influenzae were found in the bronchus. B. influenzae was found in the bronchi (six) or in the lung (two) in all but three of these eleven cases. There can be no doubt that the sup- purative process due to hemolytic streptococci was in some instances superimposed on pneumonia caused by pneumococci. In one instance, associated pneumo- coccus Type II was present in blood, lungs and bron- chus; in one instance, pneumococcus Type IV was present in blood and bronchus. Of less significance has been the demonstration of associated pneumococci, Type II atypical or Type IV, in the bronchi or lungs (three instances). (&) Pneumonia with suppuration of the interstitial tissue. In association with bronchopneumonia (three times), with typical lobar pneumonia (twice) or with all defined patches of consolidation (three times), probably best classified as bronchopneumonia, suppu- 13 ration of the interstitial tissue of the lung has occurred eight times among our cases with completed bacterio- logic examination. The interstitial septums appear as conspicuous yellow lines from which purulent fluid may be scraped. The septums are usually swollen to a thickness of about 1 or 2 mm., but wide lines of sup- puration 0.5 cm. across are seen. Edematous swelling of septums outside of the consolidated tissue is not infrequently found. These suppurating septums extend up to the pleura, and the overlying lymphatics are often widely distended with fluid. Purulent pleu- risy usually with copious effusion up to 1,900 c.c. has been found. In seven of these eight cases, hemolytic streptococci were found both in the heart's blood and in the affected lung. In one instance of lobar pneu- monia with interstitial suppuration, we found pneumo- coccus Type IV with no hemolytic streptococci in the heart's blood and lungs. By unfortunate error, no culture from the affected lung is recorded. The rela- tion of the lesion to associated infection with ] neumo- cocci is important. Pneumococcus Type II was found in one instance associated with hemolytic streptococci in the heart's blood, lung and bronchus. In one instance, pneumococcus Type II atypical with B. influ- enzae was found in the sputum during life unassociated with hemolytic streptococci, whereas three days later pneumococci were not demonstrable in the lungs or bronchi, and hemolytic streptococci were found in blood, lungs and bronchus. In another case, pneumo- coccus Type IV and hemolytic streptococci were found in the sputum during life, whereas after death five days later pneumococci had apparently disappeared and hemolytic streptococci were found in the blood and the bronchus, with B. influenzae in the latter situation. (c) Multiple abscesses clustered about bronchi. In four instances in the group of cases under considera- tion, suppuration has occurred within bronchopneu- monic patches clustered about a medium sized bronchus. These abscess cavities in communication with the bronchus are several millimeters up to 0.5 cm. in diameter, and are surrounded by gray, consoli- dated tissue. The patch of consolidation studded with abscesses may be scarcely more than 5 cm. across, but a much larger area may be involved. There has been 14 no empyema nor plural effusion in these cases. The bacteriology of the four cases cited has been as follows: Necropsy 280.—Hemolytic streptococci were found in the heart's blood, in consolidated lung tissue and in the bronchus. In the bronchus were found B. influenzae and a few staphy- lococci. Culture from the abscess was contaminated. Necropsy 322.—The blood was sterile. Staphylococcus aureus was obtained from consolidated lung; Staphylococcus aureus and pneumococcus Type III were obtained from the abscess. Necropsy 329.—The blood was sterile. Staphylococcus aureus and pneumococcus Type IV wore obtained from the abscess; B. influenzae, Staphylococcus aureus and pneumococ- cus Type IV from the bronchus Necropsy 333.—The blood contained pneumococcus Type II atypical; Staphylococcus aureus and pneumococcus Type II atypical were obtained from the abscess; Streptococcus hemolyticus from the lung on the opposite side; Staphylococ- cus aureus, B. influenzae and a few hemolytic streptococci from the bronchus. The foregoing observations indicate that Staphylo- coccus aureus may be engrafted on a pneumococcus pneumonia and cause suppuration in clustered foci, unaccompanied by empyema. WARD INFECTION Transmission of Streptococcus Pneumonia.—The fatal effect of the necessary overcrowding of this hos- pital is well shown by a comparison of the death rate from pneumonia in wards, on the one hand, which had long been organized for the care of patients with pneumonia, and in wards, on the other hand, which were opened for patients with pneumonia and imme- diately overfilled at a time when the hospital was overwhelmed by the epidemic of influenza. Ward 3 was provided with cubicles and conducted by medical officers, nurses and enlisted corps men accustomed to the care of patients with pneumonia. The figures in Table 4 show the number of patients in the ward each day from September 6 to October 5, the number of patients admitted each day and the number of deaths among patients admitted on the cor- responding day. The first period of ten days imme- diately preceded the present outbreak of pneumonia (Table 1). During the second period of ten days the 15 incidence of pneumonia gradually increased, and dur- ing the third period maintained its maximum. It is noteworthy that the death rate from pneumonia among those admitted during the first ten days was 27.2 per cent, at a time when the ward contained from fifteen to twenty-one patients. It increased during the second ten days, when the ward contained from TABLE 4.—PNEUMONIA IN WARD 3 Number Number Number of of of Patients Patients Patients Admitted on Admitted Corresponding Sept. 6 Ward 19 20 18 17 15 17 19 20 20 21 Each Day 1 2 0 1 0 2 2 1 1 1 11 Day Who Sub- sequently Died 0 ' 0 0 1 0 2 0 0 0 0 Total Deaths Percentage of Among Deaths Among Patients Patients Admitted Admitted During Period During Period of Ten Days of Ten Days Necropsy: Pneumococcus Strep, hem. 16 25 6 2 17 31 8 3 18 38 6 2 19 42 6 1 20 45 3 1 21 51 7 3 22 55 9 1 23 60 5 3 24 57 0 0 20 07 2 0 . Total. .. 52 26 60 3 0 27 57 0 0 L'8 60 3 2 2y 60 2 0 30 60 3 1 1 55 2 1 2 59 4 2 3 58 0 0 4 51) 1 0 5 61 5 2 .. 23 Necropsy: Pneumococcus 13 16 Strep, hem. 1 Undetermined 2 Necropsy: Pneumococcus 5 8 Strep, hem. 1 Undetermined 2 * In this and in the following three tables, these figures do not indicate the number of deaths on the corresponding dates but represent deaths which eventually occurred among patients admitted on these dai<-!>. twenty-five to sixty patients, to 30.7 per cent. With further increase of the ward population to from fifty- five to sixty-one, the death rate among those admitted during this period rose to 34.7 per cent. If the case of pneumonia due to infection with hemolytic strepto- coccus which occurred in each of the second and third periods be omitted, the figures for the three periods 16 will be, respectively. 27.2, 2X.X and 30.4 per cent. In -jute of the overcrowding of the ward, the introduc- tion of two cases of streptococcus pneumonia under the favorable conditions existing in the ward did not cause an outbreak of the infection. Ward S had long been used for the care of colored patients with pneumonia. Cubicles were in use and precautions were taken against the spread of infection. During the period from September 1 to September 20, all new patients with pneumonia admitted were treated on the porch of the ward, the remaining popu- lation of the ward being largelv convalescents who were placed inside. During the period from Septem- ber 21 to October 5. when the influenza epidemic was at its height, the ward rapidly filled with active cases of pneumonia and became distinctly crowded. Never- theless, no outbreak of secondary infection with the hemolytic streptococcus occurred. Ward 2 was opened, September 28, during the period in which twenty new wards for pneumonia were organized. From September 2(> to October 1, the cubicle system was not in use, the ward was crowded, organization was imperfect, and few precautions were taken to prevent the transfer of infection from one patient to another. October 2, the cubicle system was introduced, and strict and efficient precautious for the prevention of infection were instituted. Among patients admitted between September 26 and 2^ there was a ward epidemic of streptococcus pneumonia. Nine of the fifty-seven patients admitted to the ward between September 26 and October 1 were shown by examination of the sputum to have pneumococcus pneumonia free from hemolytic strep- tococci. All died, seven with infection by the hemo- lytic streptococcus, one with what was presumably an acquired infection with pneumococcus Type III (pneu- mococcus Type IV was found in the sputum on admis- sion), and one with undetermined cause (no necropsy was held). The sputum of no other patients in this ward was examined on admission. September 29 (the beginning of the second period), the porch of the ward was put into use, and of twelve patients admitted on this date eleven were put on the porch. Of these eleven patients, eight recovered and three died, of whom two were transferred to beds 17 within the ward on the day after admission. The use of the porch doubtless explains the diminished mortal- ity of this period. From October 5 to 7 the ward was closed to further admissions, save six patients transferred from Ward 1 seriously ill; of these all died. After October 7, admissions were limited to patients with pneumonia in TABLE 5.—PNEUMONIA IN WARD 8 (COLORED) Number Number Numbei Of Total Deaths Percentage of of of Patients Among Deaths Among Patients Patients Admitted on Patients Patients in Admitted Corresponding Admitted Admitted Ward Each Day Who Sub- During Two During Two Day sequently Died Periods Periods Sept. 1 48 1 0 2 47 2 0 3 45 0 0 4 47 4 0 5 40 0 0 6 29 3 0 7 25 2 0 8 28 2 0 * 9 27 0 0 Necropsy: 10 26 0 0 2 Pneumococcus 2 S 11 26 0 0 Strep, hem. 0 12 26 1 0 13 27 1 0 14 27 0 0 15 31 4 0 16 32 1 1 17 19 2 0 18 19 0 0 19 20 2 1 20 20 Total. . 0 . . 25 0 21 26 0 3 22 31 5 3 23 36 5 1 24 40 4 1 25 43 5 1 26 44 3 1 Necropsy: 27 45 1 0 Pneumococcus 10 28 47 2 1 20 Strep, hem. 1 33.9 29 48 1 1 Undetermined 10 30 46 0 0 Oct. 1 50 ■4 0 2 52 4 1 3 58 6 3 4 62 5 1 5 64 Total.. 8 59 3 whom throat swab had demonstrated Streptococcus hemolyticus. Ward 1 was opened, September 24. During the period from September 24 to October 1, the cubicle system was not in use and there were few precautions to prevent transmission of infection. October 2, the cubicle system was installed, and somewhat insufficient precautions against the spread of contagion were insti- 18 tuted. Beginning October o, the ward was closed to further admission. Table 7 is divided into two peri- ods, the first ending September 29, because on Septem- ber 29 and 30 four patients infected with hemolytic L'pt. 2(1 29 30 Oct. 1 TABLE 6.— PNKUMONIA K. WARD 2 Number Number Number of Total Deaths Percentage of of of Patients Among Deaths Among Patients Patients Admitted on Patients Patients in Admitted Corresponding Admitted Admitted Ward Each Day Who Sub- During Period During Period Day sequently Died of Three Days of Three Dayx 10 10 7 1 Necropsy: 27 17 10 \ 27 Pneumococcus 0 67.5 40 13 10 J Strep, hem. 23 --- Undetermined 4 Total..... 40 51 12 3 1 Necropsy: 49 1 1 6 Pneumococcus 2 35.8 43 4 2 J Strep, hem. 2 --- Undetermined 2 Total..... 17 47 6 2 1 Necropsy: 42 0 0 \ 4 Pneumococcus 2 40.0 41 4 2 J Strep, hem. 1 * --- Undetermined 1 Total..... 10 streptococci were admitted to the ward. Deaths that occurred (with no reference to date of admission) prior to September 30 were all due to pneumococcus with one exception, namely, a patient entering the TABLE 7.—PNF.r.MONTA IN WARD 1 Oct. Number Number Numb* rof Total Deaths Percentage of of of Patients Among Deaths Among Patients Patients Admitted on Patients Patients in Admitted Corresponding Admitted Admitted Ward Each Day Who Sub- During Period During Period Day sequent 1 t Died of Three Days of Three Days -') 13 7 2 I'l 35 15 4 Necropsy: ti 44 6 3 Pneumococcus 5 7 42 2 0 11 Strep, hem. 3 32.8 s 39 1 0 Undetermined 8 !S 42 3 2 Total.. .. 34 10 56 15 10 1 54 2 1 Necropsy: 2 59 6 8 Pneumococcus 6 3 51 3 2 24 Strep, hem. 14 50.0 4 56 7 5 Undetermined 4 6 56 Total, 5 .. 40 3 ward on September 26 and dying several hours later. Hemolytic streptococci introduced into Ward 1 con- taining patients suffering with pneumonia due, as our experience here has shown, to pneumococci, has pro- 19 duced an epidemic of secondary infection with strep- tococci. Though the ward was evidently in great danger of infection from September 24 to 29, as the result of unfavorable conditions, none appears to have occurred; but with the introduction of the micro- organism, September 29 and 30, widespread dissemi- nation of fatal infection was immediate. The occurrence of ward infection is well shown by the bacteriologic examination of the heart's blood and lungs at necropsy. Up to October 6, necropsies were performed on nearly all patients dying with pneu- monia. The results of bacteriologic examination at necropsy on patients dying in pneumonia wards is given in Table 8. TABLK 8.—BACTERIOLOGIC FINDINGS IN PATIENTS DYING IN PNEUMONIA WARDS* No Hemolytic Hemolytic Streptococci Streptococci Found Found Ward 1........................................ 12 7 2........................................ 4 24 3....................................... 18 1 4........................................ 0 1 5........................................ 1 0 6........................................ 1 0 7........................................ 0 1 8........................................ 11 0 19....................................... 6 2f 20........................................ 0 1 22........................................ 2 0 26....................................... 0 1 33........................................ 0 1 49........................................ 1 0 Total...................................... 56 39 ' With the exceptions of Wards 1, 2, 3 and 8, these wards had been opened less than one week. t One of these cases of streptococcus pneumonia was the wardmaster of Ward 2, whose illness began, October 1, during the period of severe infection of this ward. The figures demonstrate that routine necropsies with bacteriologic examination of the blood and lungs of every patient that has died with pneumonia furnishes a certain means of recognizing the occurrence of ward infection with hemolytic streptococci. This means of checking the results of bacteriologic examination of the throat and sputum should never be neglected. The recognition of suppurative pneumonia, even when no cultures are made at necropsy, will indicate that pul- monary infection with hemolytic streptococci exists. It is obviously essential to separate pneumonia patients harboring hemolytic streptococci in the mouth from those that are free from this organism; but the exis- 20 tence of hemolytic streptococci in the mouth of patients with pneumonia is not conclusive evidence that the lung has been invaded. These necropsies, with the accompanying bacterio- logic studies, gave us the first conclusive evidence that the pneumonia in this hospital following influenza was subject to secondary invasion with hemolytic strepto- cocci. They showed that Ward 3 had been maintained almost free from pyogenic infection; they forced the conviction that all patients who might subsequently be admitted to Ward 2 would with relatively few excep- tions die with it. The ward was closed. The grave danger to patients entering Ward 1 was recognized, and this ward also was closed to subsequent admis- sions. The throats of all the 900 patients in the hos- pital were swabbed as rapidly as accessible facilities permitted, and patients harboring hemolytic strepto- cocci were separated from those that were free from the micro-organism. TABLE 9.—REPORT OF TWO CASKS OF PNEUMONIA Pncuniococcus Pncumo- Bed in Sputum coccus Ocou- on Ad- Course of Found at Name Admitted pied mission Disease Necropsy Private Sw. Sept. 16 :; Type II Crisis, Sept. 21 Private P. Sept. 9 0 Type IV Crisis, Sept. 14; Type II second at lack of pneumonia, Sept. 24; death, Sept. 30; lobar pneumonia Transmission of Pneumococci.—The foregoing observations have demonstrated that streptococcus pneumonia may spread by contagion throughout an entire ward. Subsequent observations have demon- strated that pneumococci may be transmitted in the same way. Opportunity to obtain this evidence has been favored by the predominance of Types IV, III and II atypical among the pneumonias occurring here. Among cases of pneumonia of which the type has been determined between September 6 and October 1, only six instances of Type II pneumonia were found. Two of these cases occurred in Ward 3, and there is evi- dence that the disease was transmitted from these patients to three others in adjacent beds. Infection with pneumococcus Type II occurred in a patient, Private P., recovering from pneumonia caused by pneumococcus Type IV. This infection with pneu- 21 mococcus Type II was apparently acquired from a patient, Private Sw., occupying an adjacent bed. Private Sw. was admitted to Ward 3 and placed in Bed 3, September 16, one day after the onset of pneu- monia. Pneumococcus Type II was found in his sputum. His crisis occurred, September 21, on the sixth day of his illness. At the time Sw. was admit- ted, Private P. occupied Bed 5, separated by one bed from Sw.; he had been admitted, September 9, and had bronchopneumonia with signs of consolidation in the left lower lobe. Pneumococcus Type IV was found in his sputum. A crisis occurred, September 14, and the temperature remained normal until Sep- tember 24, save for a sharp rise, September 20 to 21, TABLE 10.—REPORT OF THREE CASES OF PNEUMONIA Pneumococcus Pneumo- Bed in Sputum coccus Occu- on Ad- Course of Found at Name Admitted pied mission Disease Necropsy Private Sm. Sept. 18 26 Type II Died, Sept. 27; Type II bronchopneumonia Private T. Sept. 17 28 Atypical II Crisis, Sept. 19; Type II temperature normal until Sept. 21, when it rose to 104.4; death, Sept. 29; lobar pneumonia Private L. Sept. 16 30 Type IV Some improvement Type II occurred, Sept. 24 and 25, but temper- ature did not become normal; death, Sept. 30; bronchopneumonia lasting eighteen hours. A second attack of pneumonia began, September 24, and death occurred, September 30. Necropsy disclosed the presence of lobar pneu- monia with gray hepatization of the greater part of the right lower lobe and of the left lung. Pneumo- coccus Type II was found in the heart's blood and in the lung, and with B. influenzae in the right bronchus. Private P. had the opportunity to acquire his second infection from Private Sw. Infection with pneumococcus Type II occurred in two patients apparently recovering from other types of infection, in one instance caused by pneumococcus Type II atypical, and in another by Type IV. In both instances pneumococcus Type II has apparently been acquired from a third patient occupying an adja- cent bed. 22 Private Sm. was admitted with pneumonia to Ward 3 and placed in Bed 2(>. September 18. Pneumococcus Type II was found in the sputum. The patient died, September 27, with bronchopneumonia. There were scattered patches of lobular consolidation and diffuse consolidation of the greater part of the right upper lobe. Pneumococcus Type II was found in the heart's blood. Private T. had been admitted to Ward 3, Septem- ber 17, with a diagnosis of bronchopneumonia and signs of consolidation in the right lower lobe. Pneu- mococcus Type II atypical was found in the sputum. Crisis occurred, September 19, and the temperature was normal until 8 p. m., September 21, when there was a sudden rise to 104.4; death occurred, September 29. At necropsy there was firm hepatization with TABLE 11.—REPORT OF FIVE CASKS OF PNEUMONIA Bed Pneumococcus in Pneumococcus Occu- Sputum on Found at Name pied Admission Course of Disease Necropsy Pg. is Atypical II Returned to duty, Oct. 8 .1. 49 Atypical II Died, October 4 Atypical II l.«. oo Type I Crisis; recurrence; died, Atypical II October 8 S(. 01 Atypical II Returned to duty, Oct. 25 (' 53 Atypical II Died, September 25 Atypical II yellow color and characters of lobar pneumonia at posterior parts of both lower lobes; but throughout the remainder of the lungs there were small scattered patches of bronchopneumonic consolidation. Pneu- mococcus Type II was obtained from the heart's blood and lungs; B. influenzae was found in the consolidated part of the lung and in the right bronchus. At the time T. developed his second attack of pneumonia he occupied Bed 2