THE REGULATION OF BODY DURING FEVERS by G* H. Park, Capt., M.C., and K, D, Paines, Ph.D., Physiologist J.XUL liedical Departeent Field Research Laboratory Fort Knocx, Kentucky ' 1 October 1%S *Sub—project under Study of Body Reactions and Requirements under Varied Snvironnental and Climatic Conditions. Approved 31 May 1946. MDFHL Project Mo* 6-64-12-06-(6). Project Ko. 6-64-12-06 Sub-projact HDFKL 06-(6) USD.BA 1 October 1%8 ABSTRACT m REGULATION OF BODY TEMPERATURE DURING FEVER OB JSfl Fever in man was studied calorimetrically to determine, first, the heat flows which cause changes in body temperature and, second, the physiological regulations which directly control these flows* Seventeen reactions induced by typhoid vaccine were observed in environments rang- ing in ambient temperature from 27° to A-3°C* RESULTS AND CONCLUSIONS Changes in body temperature are due almost entirely to autonomic regulation of peripheral blood flew, sweat secretion, and muscular activity. Peripheral blood flow, by altering the skin temperature, is a control of heat exchange by convection and radiation; sweating is a control of evaporative cooling; and muscular activity, in the form of shaking chills, controls metabolic heat production. Sweating and muscular activity secondarily cause changes in convection and radiation. These changes are opposed to the primary thermal effects of sweat secre- tion and chills but are smaller in magnitude. Changes in body temperature are due almost entirely to autonomic regulation of peripheral blood flew, sweat secretion, and muscular activity. Peripheral blood flow, by altering the skin temperature, is a control of heat exchange by convection and radiation; sweating is a control of evaporative cooling; and muscular activity, in the form of shaking chills, controls metabolic heat production. Sweating and muscular activity secondarily cause changes in convection and radiation. These changes are opposed to the primary thermal effects of sweat secre- tion and chills but are smaller in magnitude. To raise the body teinperature, heat production can be increased, leat loss restricted, or both changes can be induced# Peripheral blood flow and sweating must be restricted in order to maintain a normal body temperature in a cool environment. Since heat loss cannot be reduced appreciably a febrile rise can be accomplished only by increasing heat production, i.e., by muscular activity, and the degree of fever will be proportional to the severity of the chill. When the environment is warm (above fever can be produced by de- creasing heat loss, i.e,, restriction of peripheral blood flow and sweat secretion. Unless the environment is very hot, however, only gradual temperature rises can be effected. If the pyrogenic stimulus demands a sharp rise in temperature, a chill is superimposed on restricted heat loss even in very hot environments. The lysis of fever my be a passive or an active process* Passive cooling occurs when there is no increase in peripheral blood flow or sweat secretion. The skin tenperature is elevated at tlie height of fever and the greater loss of heat by convection and radiation exceeds heat production. Passive cooling can occur only in a cool environment after fever induced by a chill, or after a fall in environmental temperature. When peripheral blood flow and/or sweating increases, as is usually the case, cooling is active. Active cooling is necessary in hot environ- ments, or to produce rapid defervescence in cooler surroundings. RE GQMP1H) AT I ON S None Submitted by? C. R. Park, Capt., H.C. S. D. Palmes, Ph.D., Physiologist Approved RAY A I}pGS (J (J Director of i?,csearc \pproved FREDERICK # KRpBLAUi Lt. Col., Command ini! THE REGULATION OF BODY TEMPERATURE DURING FEVER I* INTRODUCTION A change in the temperature, or neat content, oi man is produced by an imbalance between the rates of gain and loss of heat by the body. The magnitude of change depends on the degree of imbalance and the tiias for wfriich it persists, as expressed in the equation: A H = (H + E 4 C 4 R) t where A H r the change in body heat content U = the rate of heat production by metabolism S = the rate of cooling by evaporation C - the rate of heat exchange with the ambient air by convection R - the rate of heat exchange with the surfaces of the environment by radiation t = time If, in this equation, heat gains ere considered positive and heat losses negative, metabolism will always be a positive rate, evaporation a negative rate, while convection and radiation will be positive if the environmental temperature is above that of the skin and negative if the ambient temperature is below the skin temperature. In this report, a description is given of the changes in the flow of heat between man and his environment which occur in the production and lysis of fever. Some of the physiological regulations directly concerned in the control of these flows are discussed. Thermal balance during fever has been studied most extensively at the Russell Sage Institute in New York, and the findings of these and previous investigators have been reviewed by Du Bois (1,2,3), It was shown that fever could be produced by an increase in metabolic heat production if a shaking chill occurred. Defervescence was caused by a rise in the rate of cooling, chiefly by evaporation. In the present study, among the many further aspects of temperature regulation in fever, the following questions received particular attention, uhat are the patterns of thermal exchange and regulation in the more commonly observed type of febrile reaction in which no chill occurs? .Vhat effect does the thermal nature of the environment have on the pattern of febrile reaction? (In this regard, it should be noted that heat exchanges by convection and radiation depend as much on the thermal properties of Uie environment as on the temperature of the skin. The wind velocity and relative humidity affect the rate of cooling by evaporation and it has been postulated that the metabolic rate ioay be affected by the environmental temperature* Thus, it might be expected that a different pattern of heat exchange and regulation would be seen in warn and hot environments when compared to the cool environments in which all previous studies have bean carried out.) Finally, what are the autonomic physiological regulations which directly influence thermal flows in fever? Thus convection and radiation are partial functions oi the skin temperature which, in turn, is in part a fimetion of the peripheral blood flovf*. Evaporative cooling is dependent to a considerable degree on the rate of sweating, pulmonary ventilation, and other physiological factors. Metabolism is largely regulated by muscular activity but a "ciiemical regulation,” that is, a controlled change in heat production without a change in muscular movement, lias been postulated (4) and may be operative in fever. Fever can be defined at present only in terms of the level of body temperature and, since the rectal and skin temperatures can vary over, a considerable range under normal circumstances, any definition is neces- sarily arbitrary and inexact. In these studies, the experimental condi- tions were so controlled that fever was defined as any rise in rectal temperature greater than 0.2°C, or any rise in body heat content of more than 10 Calories per square meter of body surface. II. EXPRRIlXhN TAL A. Methods and Procedures Ten young soldiers volunteered as subjects, bach was studied In a constant environment during one control and two fever experiments. The data obtained in the control studies were used as base line values for comparison with measurements made during, febrile reactions following the intravenous injection of U. S. Army triple typhoid vaccine. A new method ol human calorimetry was employed, details oi which have been presented separately (?). The nude subject was observed for a period of 4 to <3 hours while reclining on a netting in the closely regu- lated environment of a small experimental chamber. The air temperature and wall (radiant) temperatures inside the chamber were held at essen- tially the same constant values throughout a given run. Different environ- ments u’ere created from one experiment to the next over the temperature range from 2?° to 43°C. The wind velocity was constant at 15 ft./min. (equivalent to a well-ventHated but not windy room), and the humidity was always low (vapor pressure of water, 7-12 mm. Hg). A constant flow of air was maintained over the subject, and evaporation was measured continuously and nearly instantaneously by a special modification of an N.D.R.C. infra-red gas analyzer (£•) which recorded directly the difference between the water vapor concentrations at the inlet and outlet of the chamber. Oxygen consumption was measured by on "adaptation of a closed circuit system and heat production was computed therefrom as a continuous series of average values, each of which was based on the time required for the utilization of 2.5 liters of oxygen. * The term peripheral blood flow" refers throughout this report to the flow of blood through the outermost tissues of the body to the depth of 2 cm. (5, 6). Rectal, skin (9), and environmental temperatures were determined by thermo- couples at 15-minute intervals, and the mean skin temperature was calculated by weighting each of the 10 skin readings by the fraction of body surface represented (7)• Peripheral blood flow was computed from the calorimetric values by the procedure of Hardy and Soderstrom (6). B, Results All results have been calculated and graphed in the same way. In each graph (Figures 1 through 6) the rectal and mean skin temperatures are shown in the top panel with the scale on the right. The cumulative change in heat content of the body, A H, appears in the same panel as a shaded area with the scale on the left. The rates of heat flow are shown in the middle three panels, Metabolism, radiation, and convection are plotted S3 average values for intervals of fifteen minutes or less. Evaporation is shown by two smooth curves, the upper of which is total evaporation and the lower from the skin only. The separation of the lines indicates the rate of vaporization from the lungs. Peripheral blood flow was calculated for successive fifteen-minute intervals. 1, Control Experiments. A description of certain features of thermal regulation in the resting normal subject is necessary for an appreciation of the changes which occur during fever. The experiment in an environment of 29°C,, shown in Figure 1, illustrates several of these characteristics which were common to all control runs. A slow fall in rectal temperature, skin temperature, and heat content occurred during the first three hours. Such a fall ivas a consistent finding* in all environments and was ofton two or three times larger than shown here. In the last hours of this run a rise of C.2°C. in rectal temperature and B Calories per square meter of body surface in heat content was measured and, since this was the largest rise in any control run, it was taken as the arbitrary limit separating normal and febrile reactions. Fluctuations in metabolic rate of about the magnitude shown here were seen in all subjects. They could not be correlated with any visible changes in muscular activity and may reflect in part instru- mental error. Heat exchanges by convection and radiation paralleled each other since both were nearly equivalent functions of the difference between the environmental and mean skin temperatures. Evaporation from the skin fluctuated irregularly during the first three hours. This was the char- acteristic pattern of evaporation when sweating was active and probably reflected closely the irregular rate of sweat secretion from the body as a whole, since water was evaporated very rapidly at all times, With heavy sweating, variations in rate xvere more pronounced. During the last tliree hours, evaporation fell to a constant level, A low, flat tracing of tills * in subsequent unpublished studies in which subjects in the basal state were employed, this fall was less or absent. It seems probable that the physiological thermostat slowly adjusts to a lotver level of body tempera- tore with the cessation of physical activity and the decline in specific dynamic action. FIG. 1. THE TEM PERATURE , HEAT E XC H ANGES , AN D P E RI PHE R AL BLOOD FLOW 0 A NUDE MAN IN AN ENVIRONMENT OF 29°C. SUBJECT TEL. type was typical of “insensible water loss” when sweating was absent and the only water vaporized was that diffusing through the skin. Evaporation from the lungs did not change since there was no change in pulmonary ventilation. The range of experimental conditions and calorimetric values have been summarized for all control experiments in Table I, There was no correlation between the rectal and environmental temperatures, and only slight tendency for the skin temperature to be higher in the heat. Letab- olisni was only mildly influenced by the ambient temperat'ore, although the subjects felt close to shivering after several hours in the coolest environments and were uncomfortably warm in the hottest environment. Heat exchanges by convection, radiation, and evaporation were greatly altered by changes in the environment. Convection and radiation, which were paths of heat loss in the cool, became routes of gain in the heat. In cool environments with subjects I, II, and III evaporation was near the level of invisible water loss, but in hot environments, with subjects IX and X, the rates of sweat secretion averaged more than 300 grains per hour. Peripheral blood flow tended to be higher in the heat. The autonomic physiological mechanisms for temperature regu- lation ,were apparently adequate throughout this range of environmental conditions. This was suggested, first, by the fact that the rectal temp- erature was not influenced by environmental differences, and second, the fall in heat content was as great in the hot environments as in the cool ones. The mechanisms for temperature control proved to be similar in the normal and febrile state differing only in degree and in temperature relations. 2. Fever Experiments, These studies of febrile reactions have been grouped according to the environmental temperature, beginning with the experiments at 27-29°C, and ending with the experiments at 43 Each run hes been subdivided into arbitrary periods based on the upward or downward trend of the change in body heat content. Control periods are not included in the charts. Typhoid vaccine was injected 15 minutes before the first experimental point on each graph. a. Environmental Temperature 27-29°0. Alien the environment is sufficiently cool, all channels of heat loss from the body are reduced essentially to absolutely minimal levels to conserve body heat. In such circumstances a febrile rise in temperature, even if it is slight, can only be induced by an increase in heat production. In our experience, such an increase must be accomplished by a shaking chill. Situations of this type are approximated in the experiments of Figures 2 and 3. Of the four febrile reactions studied in this environment, the most severe has been charted in Figure 2. Fifty million organisms were injected intravenously at zero tine, and the first measurements were made at 15 minutes. So divergence in the pattern of thermal regulation from that in the control experiments could be noticed during the first or ‘’normal*1 period lasting for 1 hour, 15 minutes. All thermal exchanges were near the average control level except for evaporation which was slightly SUBJECT ENVIRONMENT PH20 = 7-12 nan Hg Wind s 15 ft/min BODY TEMPERATURE °C HEAT FLOWS IN caiA,2Air AH cai/e2 PERI- PHERAL FLOW L/R2/Vin. Air °C Wall °C Rectal Mean Skin M xR C E I 27.2 28.9 36.8 33.7 12.2 -22.0 - 9.1 - 11.9 - 1.5 0.08 II 27.8 23.9 36.3 34.3 15.3 -20.1 -10.6 - 16.1 - 6.8 0.25 III 22.9 30.6 36.7 31.2 10.0 -18.1 -10.6 - 11.1 - 3.1 o.n IV 32.2 32.8 36.3 31*6 10.1 - 7.8 - 3.8 - 33.9 -23.1 0.19 V 32.2 32.8 36.7 35.3 10.5 -11.1 - 7.0 - 23.5 - 7.7 0.39 VI 37.8 37.2 36.9 31.8 15.3 10.9 8.6 - 66.1 - 7.2 0.21 VII 37.2 37.2 36.8 35.6 4 '..6 7.1 10.7 - 63.3 -13.2 0.52 VIII 33.3 37.8 36.7 31.9 13.8 12.5 r* n 0 • vy - 68,6 -19.1 0.27 IX 13.3 11.1 37.1 35.8 19.8 25.8 21.7 -101.8 -11.1 0.52 X 13.3 40.6 36.2 31.8 r*-. 47.0 27.7 20.9 -102.1 -2017 0.44 EKVIHOMKaMTAL CONDITIONS AND AVERAGE CALORIMETRIC '/.'LUES FOE THE CONTROL EXPERIMENTS TABLE I FIG. 2 A SEVER FEBRILE REACTION !N A N E N VlRONM ENT OF 29°C. SUBJECT HI. elevated because of low grade active sweating. Total heat loss exceeded gain causing a slight fall in heat content and rectal temperature. The second period, during which there was a rapid climb in body heat content, lasted from 1 hour 15 minutes to 2 hours 15 minutes. Evaporation had been falling and reached the level of insensible -water loss when sweating ceased at the beginning of this period. A few minutes later a severe shaking chili occurred producing a high spike in metabolic heat production. Pronounced hyperventilation reaching 200 liters per minute accompanied the chill, but the increase in pulmonary evaporation was small in relation to other thenna1 changes. The excess heat produc- tion of the chill and the alight restriction of heat loss by evaporation led to a steeply climbing body temperature. In terms of the overall thermal balance, heat gain far exceeded lieat loss during this interval, but the rate of heat loss was nevertheless slightly greater than in the "normal11 period when temperatures were falling. This greater heat loss was caused by the elevated skin temperature which increased thermal outflow by convection and radiation. This observation is similar to the findings of Barr, Cecil and Du Bois in a comparable environment (10)• The period of falling heat content, beginning at 2 hours and 15 minutes, could be divided into two phases. In the first, lasting to 3 hours and 30 minutes, cooling was passive and the fall in heat content was slow. There was no sweating; peripheral blood flow was low; and cool- ing by convection and radiation remained above the control level by virtue of the high skin temperature persisting from the previous period. The body lost heat in the same manner as an inanimate object, an analogy suggested by Hardy and Soderstrom in their studies of normal thermal exchanges (6), In the second phase, cooling became active and the fall in heat content was accelerated. Peripheral blood flow increased and, with the onset of sweating, evaporative heat loss rose sharply. Toward the end of the period sweating and peripheral flow were reduced again and, since the skin temper- ature had now fallen, no further loss in heat content occurred. Through- out the time of falling temperat*irc, heat production was above the base line level. There was no visible muscular movement, the subject felt relaxed and comfortable and the elevated metabolic rate was very probably a function of the elevated body temperature according to van't Hoff's principle (11), A milder chill and febrile reaction is illustrated by the experiment shown in Figure 3, in which 50 million killed typhoid organisms were injected intravenously. The rise in heat content lasted over a four- hour period and the highest rectal temperature reached was only 3&°C. Peripheral blood flow was below the control level on the average and no significant degree of active sweating was measured during the time of rising heat content. Despite these restrictions, the sun of heat losses in this environment very nearly equalled heat gain by metabolism and, up to 2 hours and 45 minutes, the climb in body temperature was very slow. At this time the pyrogenic stimulus apparently necessitated a slightly higher temperature, and a very gentle shaking chill occurred, increasing heat production and accelerating the temperature rise. FIG. 3. A MILO FEBRILE REACTION IN AN ENVIRONMENT OF 29 °C. SUBJECT XE . There was no phase of passive cooling in this run. Sweating began at h hours and was accompanied by a rise in peripheral blood flaw. A fall in heat content ensued. The two other experiments in this environment closely followed the pattern of Figure 3. The environment prohibited effective restriction of heat losses and fevers were induced by shaking chills. b, environmental Temperature Jdu. in this warmer environ- ment, the temperature gradient between the body surface and the environment was reduced and, hence, heat loss by convection and radiation was dimin- ished, To maintain a normal temperature, therefore, a compensatory rise in evaporative cooling was necessary, and mild sweating occurred throughout all base line runs. In the presence of active cooling, the situation was such that a febrile reaction could be produced by restriction of heat losses alone. Such a regulation of the rate of cooling appeared to be responsible for all the changes in heat content in the experiment of Figure /*, which illustrates a mild cyclical reaction to the intravenous injection of 25 million killed typhoid organisms. Ko chill occurred and the overall changes in heat production were merely secondary to variations in body temperature. From the injection of vaccine at 0 time until 1 l/4 hours, there was no evidence of any disturbance in thermal regulation and a fall in heat content typical of a control experiment took place. At the beginning of the second period sweating ceased, peripheral blood flow reached a low level, and a positive thermal balance was established. The rectal temperature began to climb, Within one half hour, however, per- ipheral blood flow rose, and within i l/U hours, sweating began again at a low level. Thus it appeared that complete restriction of active cooling caused a response that exceeded the pyrogenic stimulus to the temperature regulating, center, and slight cooling was required to slow the reaction. Changes in peripheral blood flow and sweat rate in tills experiment occurred at different tiiaes and thus appeared to be independent regulatory mechanisms. In the second period, for example, peripheral flow rose at a time when sweat- ing was cut off; in the third period, flow reached a low level while sweating was elevated; and in the fourth period, peripheral flow rose while sweating declined. In other experiments, changes in peripheral flow and sweating rate were nearly synchronous. The differences among the curves of rectal temperature, skin temperature and heat content were marked in this experi- ment and will be discussed subsequently. Only mild reactions were obtained in 3 other runs at 32°C. These followed the pattern shown in Figure U, Ko chills occurred and body temperature was controlled entirely by regulation of the rate of cooling. c. Environmental Temperature 3S°C, The most severe reaction studied in this environment has been charted in Figure 5, in which 100 million killed typhoid organisms were Injected intravenously1'. In this instance (fever was produced by a combination of the principal changes described in the earlier experiments) heat elimination was restricted and heat production was increased. FIG. 4. LOW GRADE CYCLICAL FEVER IN AN ENVIRONMENT OF 32° C. SUBJECT m In the control runs, convection arid radiation were positive values because the temperature of the surroundings was above skin tempera- ture. Sweating was marked since evaporation was the only channel for heat loss. Thus a reduction in sweating ccild cause the retention of large amounts of heat Tor the genesis of fever. There was again a h5~n±mxte "normal” period following the injection of vaccine during which time the rectal temperature and heat content fell. Beginning at 45 minutes, sweating and peripheral blood flow vere simultaneously and progressively reduced but, before the maximal restriction of heat elimination was reached, a sudden rise in heat produc- tion occurred as the result of a shaking chill. A rapid rise of temperature ensued. The elevated skin temperature, however, reduced beat gain by con- vection and radiation. This effect was thermally equivalent to the increased neat loss noted at the time of rising temperatures in the cool environments (Fig. 2) and, in both instances, the change was thermally opposed to the general trend of temperature regulation. The period of cooling began abruptly at 2 hours, with a rise in peripheral blood flow and the onset of sweating. The wide fluctuations subsequently occurring in evaporation were consistent with the general observation that the sweat rate was much more variable in the febrile than in the normal subject. These fluctuations diminished toward the end of the run as normal body temperatures were approached. In four other experiments in this environment, the reactions were mild. Ho chills occurred and temperature control was achieved by regulation of the rate of cooling alone. d. Environmental Temperature 13 . The experiment shown in Figure 6 illustrates a severe febrile reaction to the Intravenous injectlor of 100 million killed typhoid organisms in a very hot environment, at the time the typhoid vaccine was administered the subject was slightly febrile as the result of a vaccine injection on the previous day, as evidenced by the initially elevated values of the skin and rectal temperatures. For this reason, also, the first measurements of heat exchanges did not fall close to control lines. The pyrogenic stimulus from the typhoid injection was apparent at 30 minutes when a steep fall in sweat rate and peripheral blood flow* was observed. Although evaporation from the skin was reduced to the level of insensible water loss, a shaking chill occurred in addition and body temperatures climbed rapidly, during this time, the skin tempera- ture exceeded the rectal temperature since heat was gained rapidly at the hodv surface from the environment bv convection and radiation, and evanora- * The calculation of peripheral blood flew in this period gave negative values that were obviously incorrect. The error arose chiefly from two sources: (1) the rectal and me.an skin temperatures are poor indices of aver- age deep and superficial temperatures when these are changing rapidly, and (2) the depth of the thermal gradient arbitrarily assumed for conduction is increased with marked vasoconstriction. It seems, however, permissible to accept the chan pies in peripheral blood flow as qualitatively significant. FIG. 5. A SEVERE FEBRILE REACTION IN AN ENVIRONMENT OF 38°C. SUBJECT 3ZI FIG. 6. A SEVERE FEBRILE REACTION IN AN ENVIRONMENT OF 43°C. SUBJECT X tive cooling was slight. Defervescence began abruptly at 1 hour and 30 filnutes, with a steep rise in peripheral blood flow and the simultaneous onset of sweating. Evaporation reached the rate of 600 gm./hr. within a few minutes. In two other experiments at 43 °C., mild reactions were obtained. These were caused by regulation of the cooling rate. Ho chills occurred. III. DI3CTIb3I0H A. Physiological IWuluti on of Heat Content Three physiological reactions, muscular activity, sweat secre- tion, and peripheral blood flow, all functioning autonordically, are the principal controls of body heat content in fever. wiuscular activity is presumably the only mechanism for the regu- lation of l:ieat production. Theoretically, elevation of heat production can arise from chills, from an increased rauscle tonus, or may follow elevations of body temperature according to van*t Hoff1s law (ll). In these experi- ments, all augmentations of heat production were caused, by chills or were secondary to elevations in body temper attire. Ho data were obtained in favor of a "chemical regulation" of the metabolic rate in the febrile or resting subjects. The rate of sweat secretion is the chief control of evaporative cooling. Evaporation is dependent in part on the wetted area of the body (12) and this, in turn, is a function of the rate of sweat secretion. Hie environments employed were sufficiently dry to allow rapid evaporation, and the fluctuating tracings obtained indicated closely the pattern of sweat secretion from the body as a whole. Peripheral blood flow serves as a con rol of heat exchange by convection and radiation, since it is one of the factors determining skin temperature. A rise in flow increases the transport of heat from the deep tissues to the skin, raises the surface temperature, and alters the thermal gradients for convection and radiation. In addition, muscular activity and sweat secretion have secondary effects on convection and radiation, A chill raises the deep tissue tem- perature and, as heat flews to the periphery, the skin temperature rises. Increased sweating allows greater evaporative cooling and, as lieat is sup- plied directly at the body surface, the skin temperature falls. Muscular activity, therefore, promotes heat loss by convection and radiation, and increased sweating promotes heat gam by these routes. Hie secondary effects on convection and radiation of these regulatory mechanisms are thus opposed to their primary effects but are smaller in magnitude. The data show that heat exchangee in the resting and the febrile subject fluctuate constantly and often widely. It thus appears that ther- mal regulation proceeds as a continuous sequence of adiustinents. often of a gross nature, which, however, have little immediate effect on deep tissue temperature because of the large heat capacity of the body and the thermal insulation of the peripheral tissues. These fluctuations in heat exchanges have an easily seen effect on the skin temperature and, hence, on convection and radiation. From the present data these effects can be presented only in a qualitative manner. Thus in Figure 2, during the time of climbing heat content, the rise in sldn temperature and C 4 H exchange could be ascribed chiefly to the heat production of muscular activity, since only small changes occurred in sweat secretion and peripheral flow. In all charts, gross changes in sweat rate caused inflections in the curves of mean skin temperature. The effect of variations in peripheral flow was usually masked by the secondary effect of muscular activity or sweating, but its important role can be shown in a sample rough calculation. In Figure 3, in the interval from 1 hour and A5 isinutes to 2 hours, peripheral flow at a value of 0.02 carried 1 to the skin. In the succeeding 15 minutes, the flow rose to 0.37 I/hymin, and carried 51 Cal/Li2/hr. If the flow had not changed, only 2 would have readied the skin in this time; the surface ature would have stabilized lower, and an increased heat gain by C 4 R of 12 Cal/ul2/hr. would have resulted. Changes in peripheral flow, which usually initiated and were always in accord with the general trend of temperature regulation, were always opposed to the effects of changes in □uscilar activity and sweat rate on the basis of C 4 R exchange. Several other physiological changes had minor effects. Pulmonar; ventilation increased slightly as the body temperature rose, and marked hyperventilation was seen during chills. A proportionate rise in evap- orative cooling from the lungs resulted, but this was never large or sustained. Heat exchange by pulmonary convection was always negligible. Kuscular movement during chills increased wind velocity and promoted heat exchange by convection. This increase could not be measured accurately, but, by calc;ilation from the data of Hardy, Milhorat, and Du Bois (13), it would have little overall significance. PiXar erection, seen regularly with chills, was assumed to have no important effect. B. The Patterns of Regulation These experiments show that a definite sequence of physiological reactions occur in the production and lysis of fever. In general, these are directed first, toward reducing the loss of body heat and second, toward increasing internal thermogenesis. Xore specifically, the earliest change is a restriction of peripheral blood flow, the next a decrease in sweat rate, and the last an increase in muscular activity in the form of a shaking chill. Two factors influence the- pattern of these reactions— the nature of the environment and the strength of the pyrogenic stimulus, in a cool environment peripheral blood flow and sweating are restricted. «hen a pyrogenic stimulus is felt, no farther reduction in those controls is possible if the environment is sufficiently cool, hence heat loss cannot be reduced, and heat production must be increased by a chill. The rapidity of the febrile rise and the height reached are proportional to the severity and duration of muscular activity. In warmer environments, peripheral flow and sweating are normally considerably higher and can, therefore, be ef- factively reduced. When heat loss is diminished a rise of body temperature begins. Unless the enviroruaent is very hot, however, only a small sparing of heat loss can be made, and a gradual rise in body temperature results. If a strong pyrogenic stimulus demands a sharp rise, heat production is increased by a cliill in addition, even in very hot surroundings. The lysis of fever may be a passive or active process0 If a chill has occurred, the skin temperature nay be so high at the beginning of defervescence that the increased loss of heat by convection and radia- tion plus that of evaporative cooling from insensible water loss, exceeds the rate of heat production, and the body cools. More commonly, however, defervescence is a process in which heat loss is actively increased, first, by a rise in peripheral blood flow, and second, by sweat secretion. Active cooling must be induced if no chill has taken place, if the environmental temperature is above the skin temperature or if rapid cooling is required. Changes in peripheral blood flow and sweat rate were closely associated and nearly synchronous except in a few instances where the rise and fall of heat content proceeded very slowly and changes in peripheral flow were noted first. ?foile it should be possible theoretically to reach a febrile state by restriction of peripheral flow alone, such a case was not observed. ho tendency was noted for fevers to be more severe in hot sur- roundings, giving further evidence of the adaptation of the pattern of tem- pera ture control to the thermal properties of the environment. It seems probable that fever caused by natural infections and many other pyrogenic agents follow the same sequence of reactions as observed In these studies although this has no experimental confirmation. Barr, Cecil, and Du Bois (10, 1A) found no significant difference in regulation during fevers due to typhoid vaccine and to malaria. The occurrence of a latent period between the time of typhoid injection and the first evidence of any dis- turbance in thermal regulation is consonant with the thesis that the typhoid organism must undergo alteration within the body or that a "pyrexia" (kenkin (15)} must be formed before the temperature center is stimulated. G, Other Considerations ihe skin and rectal temperatures proved imperfect indices of heat content at average body temperature, as noted previously by Barr and Du Bois (lij. The rectal temperatures lagged behind changes in heat content and, following chills, the maximum rectal temperature might not be attained until A5 minutes after the peak heat content was reached. Changes in skin temper- ature wore more sensitive to changes in heat content but the absolute value of skin temperature was a poor index of average body temperatures. In Table II, the rates of evaporative water loss during a cycle of fever are compared with water loss at rest in the same environment. The data indicate that if evaporation is not impeded, excessive sweating con- tributes little to the dehydration sometimes seen in fever. TABUS II EVAPORATIVE VATER LQSS DURING CONTROL AM) FEVER EXPERIMENTS Subject Envir onmenta1 Tsnperature °C Evaporative ..ater Loss in Control Hun . Whr Increase in Evaporative Water Loss in Fever gfn/hr. Run 1 Ron 2 I 29 39 30 32 II 29 28 II 29* 32 39 III 29 50 28 30 IV 32 65 6 30 V ■32 109 -1 -21 VI 38 223 -2 6 VII 38 2JU -3 VIII 38 242 -5 15 IX 43 319 2 11 X 43 341 5 ♦Vaccine injected but no febrile reaction. In fevers seen in clinical practice, the patient makes use of other controls of thermal regulation in addition to the autonomic mechan- isms which have been described. These are voluntary and include regu- lation of hiis personal environment by the putting on and off of clothing, alterations in posture, and possibly clianges in muscular activity. The addition of clothing is common during the febrile rise and creates a warmer environment. This may allow fever to occur without a chill by the pattern of regulation described for warn environments. In this connection it was noted in cur studies that severe discomfort was experienced by the subjects only during chills. The patient daring the rise frequently assumes a hunched position which reduces heat loss by convection, radi- ation, and evaporation. Restlessness may increase heat production very slightly. In the lyses of fever, if clothing is removed, the cooler environment reduces the requirement for active cooling. Occasionally, a spread-out position is assumed which increases heat losses by convec- tion, radiation, and evaporation. The patient in defervescence charac- teristically is auiet. If enough clothing is not removed during lysis, interference with evaporation leads to excessive sweating. Vftien a normal temperature is reached, residual moisture in the clothing continues to evaporate and cools the patient excessively. This imsy account for the occasional chills seen after rapid defervescence. IV. SITM1ARY AND CONCLUSIONS Fever in man was studied calorimetrically to determine, first, the heat flows which cause changes in body temperature and, second, the physiological regulations which directly control those flows. Seventeen reactions induced by typhoid vaccine were observed in environments ranging in ambient temperature frcaa 27 to 13" C. Changes in body temperature are due almost entirely to autonomic regulation of peripheral blood flow, sweat secretion, and muscular activity. Peripheral blood flow, by altering the skin temperature, is a control of heat exchange by convection and radiation; sweating is a control of evap- orative cooling; and muscular activity, in the form of chills, controls metabolic heat production. Sweating and muscular activity secondarily cause changes in convection and radiation. These changes are opposed to the primary thermal effects of sweat secretion and chills, but are smaller in magnitude. The pattern of temperature regulation in fever is governed by the thermal nature of the environment and the strength of the pyrogenic stimulus, whenever possible, temperature changes are bro'ught about by altering the rate of heat loss from the body rather than by changing the rate of heat production. To raise the body temperature, heat production can be increased, heat loss restricted, or both changes can be induced. Peripheral blood flow and sweating '.oust be restricted in order to maintain a normal body temperature in a cool environment. Since heat loss cannot be reduced further, a febrile rise can be accomplished only by increasing heat production, i.e., muscular activity, and the degree of fever will be proportional to the severity of the chill. «>ien the environment is warn (above 30°C.) fever can be produced by decreasing heat loss, i,e,, restriction of peripheral blood flow and sweat secretion. Unless the environment is very hot, however, only gradual temperatore rises can be effected. If the pyrogenic stimulus demands a sharp rise in temperature, a chill is superimposed on restricted heat loss even in very hoi environments. The lysis of fever may be a passive or an active process. Passive cooling occurs when there is no increase in peripheral blood flov/ or sweat secretion. The skin temperature is elevated at the height of fever and the greater loss of heat by convection and radiation exceeds heat production. Passive cooling can occur only in a cool environment after fever induced by a chill, or after a fall in environmental temperature, $hen peripheral blood flow and/or sweating increases, as is usually the case, cooling is active. Active cooling is necessary in hot environments, or in cooler surroundings to produce rapid defervescence. V. R7Xg-&gn)ATIQ!;S None, VI. BIBLIOGRAPHY I, Du Bois, E. F.: Mechanism of heat loss and temperature regulation, Stanford University Publication. Medical Sciences 5: 315, 1937. 2. Du Bois, E. F,: Heat loss from the human body, Bullc H. Y. Academy Bed. 15: 143, 1939. 3, Du Bois, E. F.: Basal metabolism in health and disease, ed. 3, Lea and Febiger, Philadelphia, 1936. 1. Lusk, G, t The Science of nutrition, ed. 1, . 3, Saunders Go,, Philadelphia, 1923. 5. Herrington, L, P., »Vinalov?, G.-L, A,, and Gagge, A, P,: The relative influence of radiation and convection upon vasomotor temperature regulation. Am. J, Physiol. 120: 133, 193?• 6. Hardy, J. D., and Sodcrstrom, G. P.; Heat loss from the nude body and peripheral blood flow at temperatures of 22®C. to 35°6«, J. Kutrition 16; 193, 193-6. 7. Palmes, L, D,, and Par’:, C, R,: A method oi human calorimetry. Report 55-3, M.D.F.H.L., Fort Knee;, Kentucky, April, 1947- 6, Palmes, E, D.: An apparatus and net. od lor tna continuous aeter- juination of evaporative water loss from human subjects. To be noblished. Review of Scientific Instruments, October, 1943. 9. Palmes, E. D., and Park, C. K,; An improved mounting for thermo- couples for the measurement of the surface temperature of the body, J. Lab. and Clin, Led. 33.: 1014, 1946, 10. Barr, D. P., Cecil, R. L, and Du Bois, E. F.i Clinical calorimetry. XXXII, Temperature regulation after the intravenous injection of proteose and typhoid vaccine, Arch. Int, led, 606, 1922. 11, Du Bois. S. F0: Metabolism in fever, J. A. h. A, XL1 352, 1921. 12, Gagge, A. P.: A new physiological variable associated with sensible and insensible perspiration. An. J. Physiol. 120i 277, 1937. 13. Hardy, J. D,, Milhorat, A. T., and Du Bois, a. r.s oi exercise and chills on heat loss from the body, 1. hutrition 15* 563, 1933. H. Barr, D. P. and Du Bois, 2. F.j Clinical calorimetry. 28th paper. The metabolism in malarial fever. Arch. Int. -led. Zlt 627, 1916. 15. Menkin, V.; Chemical basis of fever with inflammation. .rch. Path. 22; 23, 1945.