| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
INVITED REVIEW
HIGHLIGHTED TOPICS
A Physiological Systems Approach to Human and Mammalian Thermoregulation
Departement de physiologie, Faculté de medecine, Université Laval, Quebec, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT SET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
fever; anapyrexia
|
SET POINT |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
It was in temperature regulation that the application of the concept of physiological set point is likely to have been born, for the probable reason that temperature regulation has barely any organ of its own and works as a "pure" function that invites theorization. After the demonstration that the thresholds for heat loss and heat production responses were raised in fever (80), Hardy initiated the notion of set point in that function (45, 46). This set point is an information input that may be determined by an external signal to which the regulated variable is compared, or it may be determined by the structural characteristics of the system itself. In the case of temperature regulation the actual internal temperature is compared with the set point wanted by the organism. Hammel was the one who refined the notion with the description of adjustable set point (42–44). This minireview will therefore focus on these notions that are not yet universally accepted, perhaps precisely because they are integrative and thus not politically correct (see also Refs. 19, 20, 29). Regulatory set point and adjustable set point are not limited to temperature regulation; these concepts apply also to other physiological functions, thus showing their general usefulness to understanding physiology. Examples will be taken mostly from temperature regulation, but other functions will show that the concepts stand also in other physiological regulations where they are sometimes also object of controversy.2
|
DEFINITIONS |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
First, it is important to remember the difference between a regulated system and a simple steady state. Following Burton (15), we may define as "steady state" a system receiving passively a flow of energy or matter and returning passively an equal flow of energy or matter to its environment. Such a dynamic, but passive, exchange confers an ability to resist perturbations, even continuous perturbations within a given range, and results in stability (28). A steady state stabilizes itself without a set point because an open system in steady-state can resist perturbations, however passively. The main difference between an open system in a steady state and a regulated system lies in the responses the latter opposes to perturbations. One way to differentiate a regulated system from a simple steady-state system consists in measuring how the system responds to a global perturbation. A response opposed to the perturbation indicates that we deal with a regulated system. The existence of such a response either on the inflow or the outflow is pathognomonic of regulation. These responses modulate the inflow and/or the outflow of energy and matter exchanged by the body with its environment and oppose the perturbations whereas the steady state is passive.
"Regulation" is the maintaining constant of a variable in the milieu intérieur. The main property of a regulatory system is that a deviation of the regulated variable triggers a correcting response that opposes the deviation. The minimal deviation tolerated by the system is the error signal. All regulated variables are tensive variables, such as temperature, blood concentrations, arterial pressure, and the like. The advantage of regulating a tensive variable is that it is sufficient for the variable to be regulated at one point of a compartment for the stability to apply automatically throughout the whole compartment. This makes sense because the regulatory function is thus freed from the mass or volume in which the tensive variable is measured; e.g., a new-born infant, a morbidly obese person, and an intact adult all regulate their core temperatures, blood pressures, or Ca2+ concentrations at the same values independently of their heat contents, blood masses, and calcium contents, which vary considerably. N.B. It should be kept in mind that the unit scale on which the stability is estimated is an essential parameter. For example, the Celsius scale is more demanding than the Kelvin scale when estimating body temperature; on the other hand, the pH unit, which is logarithmic, renders problematic the concept of a regulated H+ ion concentration.
"Control" is often mistaken for regulation, yet the words are not interchangeable. Brobeck (13) pinpointed the necessity to distinguish both processes. Control refers to the action of a system on the responses that oppose perturbations. For example, body temperature is regulated through the control of heat production and heat loss. Examples of physiological regulations show that in most cases the regulation of a given variable is achieved by controlling both inflow into the system and outflow from it independently, e.g., blood pressure is regulated by the controls of both cardiac output and peripheral resistance. For a biologist, to distinguish whether a regulatory response controls the inflow (e.g., heat production, cardiac output, food-water intake) or the outflow (e.g., heat loss, peripheral vascular resistance, kidney function) is important (29). The duality of the available responses gives one degree of freedom to the system.
Many authors do not distinguish between the "set point" and a "reference." A reference is a known physical variable to which we can compare an unknown one, i.e., a physical datum against which a system or an instrument can be calibrated. For example, melting ice is a 0°C reference that can be used to calibrate a thermometer; the period of the vibration of a crystal can serve as a reference for time measurement, and so forth.
Set point should not be mistaken, therefore, for reference. Set point is the value defended by a regulation. In the absence of external perturbation the regulated variable stabilizes at the set point of the system. This set point is an information input that may be determined by an external signal to which the regulated variable is compared or may be determined by the structural characteristics of the system itself. Thus the set point is an information input, e.g., the temperature at which we adjust any central heating thermostat to regulate the temperature of a building. In that case, this set point is an input that is determined by an external signal to which the regulated variable is compared. For example, temperature regulation defends a core temperature close to 37°C; glucose regulation defends a blood glucose concentration close to 1 g/l, and so forth.
The difference between the actual value of a regulated variable and its set point is the "error signal." The error signal triggers correcting defense responses opposing it. For example, when core body temperature is below set point, temperature regulation opposes such hypothermia with behavior and shivering (the latter only in endothermic animals). The concept of set point bears upon the nonlinearity of the responses that oppose any deviation of the defended tensive variable, which thus remains constant.
|
TEMPERATURE REGULATION OR HEAT REGULATION? |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
The concept of heat regulation, however, opposes totally not only the concept of set point but that of regulation itself because heat is not a tensive variable. It bears upon the idea that heat flows in and out the body are sensed. Besides the fact that no heat-flow sensor has ever been discovered in the body, the mere fact that ectotherms regulate their body temperature behaviorally, with virtually no heat flow from their body to the environment, is sufficient to rule out the concept. The last time heat regulation was defended, as far as I know, was in 1997 (96). Because the argument was not new, its previous rebuttal still holds and I send to these theoretical papers (19, 22) the readers interested in that age-old discussion.
|
NO SET POINT? |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
If a stable variable in the body is not regulated, then its stability must result from the properties of a simple steady state. The main difference between steady state and regulation lies in the responses that oppose changes to the regulated variable in the latter. The student of any regulated system can recognize a set point from the threshold at which these correcting responses, when plotted against the regulated variable, start to take place to oppose a deviation that perturbs the system. Such a threshold for an opposing response is the nonlinear signal that indicates the set point. Reciprocally, the set point is the value of the regulated variable at which defense responses become nil. Thus, to answer the set point question comes to detect a threshold, indicative of a nonlinear signal, in the response that opposes deviations. The fact that such a general description holds true for different functions is not trivial because such similarity might reflect a profound unity in physiological functioning.
The concept of set point does not prejudge the underlying mechanism that may be nervous or endocrine. (See Refs. 10a, 41, and 42 for such an analysis in temperature regulation.) In the case of calcemia, the set point is so strict that it may be considered as a mere structural characteristic of the system itself.
In the case of body weight regulation,3 the set point can be discerned not from anatomy, at Romanovsky's regret (33), but from the threshold of opposing responses to a body weight drop. A reductionist approach, seeking the organ to explain the function, does not help when a function involves the body in its totality. When body weight is lower than its set point, the inflow of food raises and the outflow of energy expenditure drops. Examples of such response thresholds can be found with human gustatory pleasure, rat facial responses to sweet stimuli, rat food hoarding behavior, human saliva secretion, and metabolic responses to changes in body weight. All these responses oppose deviations from body weight and imply that there is a set point (20, 24, 57, 74, 81).
In the case of temperature regulation, the plot of responses to core temperature changes shows clear thresholds of both heat production and heat loss (Fig. 1). Further analysis may reveal that various responses have different thresholds. For example, Romanovsky (83) analyzed that there are cases where the thresholds for heat production and heat loss are separated with a dead band. This led him to question the notion of set point and its implications: fever and anapyrexia.
|
|
SET POINT OR NULL ZONE? |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
0.5°C. However, such a null zone disappears completely when group results are plotted and both heat loss, and, moreover, heat production responses tend even to overlap [Mekjavic et al.'s (71) Fig. 1] and no null zone is visible. I incline to believe that if a null zone exists it is of little physiological significance for the following reasons: 1) the breadth of the null zone, in Mekjavic et al.'s experiment, was smaller when plotted against esophageal temperature (Tes) than when plotted against rectal temperature (Tre; 0.47 vs. 0.57°C). Tre is known to be a slow-responding temperature that lags more than Tes during transients. Considering that neither Tre nor Tes is brain temperature, I strongly suspect that the already narrow null zone might become virtual if the responses were plotted against tympanic temperature (21). I shall return to the influence of selective brain cooling below, when discussing fever. 2) When the perception of comfort and discomfort is explored, the main signal is core temperature, but the null zone between feeling cold and feeling warm is virtual (3, 27).
It is likely that when the constancy of any regulated variable is vital, the thresholds for the defense responses against positive or negative error signals tend to coincide. They may even overlap as is the case of the regulation of blood calcium: the threshold for stopping the secretion of parathormone, a hypercalcemic hormone, is 12 mg/100 ml blood, while simultaneously the threshold for starting the secretion of thyrocalcitonin, a powerful hypocalcemic hormone, is 10 mg/100 ml blood. It follows that blood calcium is narrowly guarded at 10 mg/100 ml. The best way to describe such a system is as 10 mg/100 ml being the set point for blood calcium regulation (9)
The existence of a hypothetical narrow null zone in temperature regulation would just mean that temperature stability is less immediately vital than calcium concentration.
|
ADJUSTABLE SET POINT |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
A convincing argument in favor of a set point of core temperature in health, as well as during fever and anapyrexia, is the narrow control on the behavioral response. As already underlined by Caputa and Romanovsky (33), behavior is a whole body response adapted to the defense of homeostasis. In the experiments mentioned above, the behaviors, food intake, and temperature regulation that were triggered by sensory pleasure were not limited to correcting immediate needs but also anticipated future needs for chemical and thermal energy. Such anticipation is an especially beneficial property of adjustable set points. The set points of the various regulations taking place in physiology are not fixed constants. On the contrary, they tend to be permanently adjusted according to peripheral signals or genetically programmed signals. The mechanism involved is a simple resetting with time.
Set points may be constant as is the case with calcemia regulation, but they are most often adjustable from inherent internal signals, cyclically, or unidirectionally during aging, and under the influence of peripheral sensory signals.
Short term. The thermoregulatory set point can be adjusted acutely, as proposed by Hammel (43), under the influence of peripheral temperature signals. Such a mechanism has anticipatory properties because a cold skin raises the set point and a warm skin lowers it before ambient temperature has a chance to modify core temperature.
In the nycthemeral cycle, the core temperature set point oscillates over a period of 24 h, by a mechanism probably due to melatonin secretion (2). During this cycle, the pleasure of thermal sensation is precisely adjusted to trigger behaviors that defend the oscillating set point (4, 25, 93). The preferred ambient temperature and dressing behavior will anticipate the oscillation (87). Through this hedonic experience, behavior is in turn adapted to maintaining of core temperature close to the oscillating set point. Oscillations of alliesthesia and dressing behavior that occur during the ovarian cycle indicate the 28-day hormonal resetting of the set point (37, 49, 50, 58, 62).
Similar mechanisms are likely to take place in the ponderostat under sensory and endocrine influences (18).
Long term. Adjustments of the body set points of various functions may adapt physiology and its behavioral response to changing needs, such as circannual environmental changes (67, 84). This phenomenon has been called "homeorhesis" by Nicolaïdis (77) and "rheostasis" by Mrosovsky (73).
In temperature regulation, long-term adjustments of the set point are limited normally to hibernators, but some have been described in humans too (14, 63, 67). They may be predominant in other functions, e.g., arterial pressure rise or weight gain with aging (47) could be regarded as an adaptation of homeostasis, resulting in an improved adaptation of living organisms to their environment.
Mrosovsky (73) has superbly theorized on these long-term adjustments of the various set points of the physiological regulations. He reached the concept of rheostasis by taking into account that 1) nocturnal short-term hypothermia anticipates the next day's heat; 2) the longer term fluctuations of the various physiological "constants" are indeed defended; and 3) the adjustable set points of the various constants in the milieu intérieur are a specialized adaptation. This mechanism is close to the concept of "allostasis" created by Sterling and Eyer (89) in that it accounts for the long term. However, rheostasis/homeorhesis reinforces strongly the concept of set point, contrary to allostasis, which, in my opinion, blurs the understanding of regulatory process.
Pathology. Many studies have evidenced that anesthesia lowers the thermoregulatory set point (e.g., Refs. 70, 76, 90). However, anesthesia may also raise the set point in such a way as to reach lethal temperatures (12, 53). Paradoxical shifts of the set point have been described in the late stages of hyperthermia and hypothermia. When temperature regulation is overwhelmed, suddenly the thermostat shifts drastically toward higher values in heat stroke (5) and lower values in the late state of hypothermia (39). In both cases, thermal discomfort and behavioral responses defend the new set points. Such a resetting of the set point on the border of death leads the body to abandon all autonomic responses in subjects whose physiological capacity is already overtaxed. Temperature regulation thus ranks lower than heart and brain survival when the available resources become minimal. These symptoms are bad prognoses as they are the last defenses of dying patients, but they give additional evidence of set point in temperature regulation.
|
NO FEVER? |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
As mentioned above, behavior is the paradigmatic integrated response displayed by whole body organisms. There are innumerable examples of a close control of animal behavior and human thermal comfort to defend core temperature in coordination with autonomic responses in endotherms. For examples of such examples during fever, see Refs. 11, 38, and 88. Based on Vybiral et al.'s (94) results in rabbits, Romanovsky (83) acknowledges that point but argues that this is true only during the onset of fever but that in later stages the thresholds for panting and shivering separate from one another. The threshold for panting rises and the threshold for shivering drops, while the gains of both responses remain unchanged. Such a pattern is similar to that which is produced in that species by a suppression of fever with an antipyretic drug (95). Thus a large null zone takes place. In my opinion, such an argument is not sufficient to accept that there is no set point and that the concept of fever is obsolete, for the following reasons: 1) During the first stages of fever thresholds for both heat gain and heat loss responses are moved upward simultaneously. Such an integrated response thus confirms the notions of set point and fever. 2) The null zone that took place between the thresholds for warm and cold defense responses was observed only in rabbits, a special species, presumably under restrained conditions. Restriction is known to inhibit rabbits' temperature regulation. In humans and in dogs, coordinated responses have been repeatedly recorded after bacterial or other pyrogenic substances, thus confirming both set point and fever without null zone. 3) The null zone that took place between the thresholds for warm and cold defense responses were observed only when core temperature was already high and the rabbits' response in that case was identical to that after antipyretics; therefore that was perhaps the beginning of defervescence.
Therefore, even if both thresholds are shifted unequally upward during fever, as postulated by Romanovsky (83), the very fact that both moved upward during the early stage of fever would confirm the notion of set point. The only piece of information provided by a later unequal shift would be a confirmation of the different networks involved in heat production, heat conservation, and heat loss networks.
Fever and selective brain cooling. The case of dehydration is interesting from this point of view. During dehydration core temperature at rest is raised by 0.3–0.5°C or more, depending on the magnitude of dehydration (1), and sweat secretion is inhibited on the trunk and limbs in sweating mammals (75, 92) as well as humans (78, 91). When sweat rate is plotted against trunk temperature, the response of dehydrated subjects is the same as that of the subjects in control conditions but the response starts at a higher core temperature (31).
Such a change, similar to a fever, is accurately described as a resetting of temperature regulation to a higher value. However, the knowledge of selective brain cooling leads to a reexamination of this concept. If the tympanic temperature of dehydrated subjects is recorded together with their trunk temperature, it is confirmed that trunk temperature is elevated; however, it also becomes immediately obvious that tympanic temperature is not elevated (31). In addition, if sweat secretion is indeed inhibited on the trunk during dehydration, the face and the head tend to sweat normally in response to hyperthermia. Thus, if one considers that intracranial temperature is the regulated variable, there is no resetting of the thermostat (Fig. 2). Selective brain cooling puts the trunk in a somewhat open-loop situation that may lead to erroneous conclusions regarding set point. The questioning of the concept of fever validity should, in turn, be itself questioned taking into account not trunk temperature recordings but brain temperature in animals or tympanic temperature in humans.
|
Yet it is of interest to look at the resetting from the point of view of selective brain cooling. Figure 3 shows that during fever when Tes was elevated up to 38.7°C, brain temperature, although elevated, was about 0.3°C lower than Tes. This means that again some open-loop rise of trunk temperature takes place during fever, and this as soon as fever manifests itself as can be seen on the recording before exercise started. The left side of the figure illustrates the vasomotor factor of selective brain cooling, the reversal of blood flow when core temperature evolves from hypo- to hyperthermia. During hypothermia the venous blood through the cranium tends to flow from brain to skin and during hyperthermia from skin to brain. The right side of the figure shows that this is also the case during fever, but at a higher core temperature. The hypothalamic set point is raised less than one would expect looking at Tre as in done for clinical purpose, but it is raised nevertheless. There is a resetting of the thermostat during fever, and intracranial temperature is defended at a higher level even though selective brain cooling takes place (32). Thus selective brain cooling does not fundamentally change our view of fever; if anything, such results give another example of raised threshold in a given response.
|
|
NO ANAPYREXIA? |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
Romanovsky's arguments against the concept of anapyrexia were drawn from "models" of anapyrexia and from pharmacology.
Animal models were pathological situations such as toxic shock in rat and guinea pig. Because they do not reproduce what takes place in humans, for whom the concept of anapyrexia was devised, the logical conclusion is that those situations are poor models of anapyrexia.
In humans, nalbuphine lowered the shivering threshold more than the vasodilatation threshold (40), a pattern similar to that obtained with various pharmacological influences in cats' (10) and dogs' (17) shivering and panting thresholds. This again, apparently contradicts the notion of a coordinated lower set point. However, the fact that a drug inhibits one type of response more than the other simply confirms that the networks controlling the heat gain and heat loss responses are different, not that there is no coordination between them in the intact animal or person. These results do not reproduce the typical anapyrexia that takes place spontaneously in humans, either in acute attacks (48, 52, 86), chronic episodes (51), or, as mentioned above, in deep irreversible hypothermia (39). During anapyrexia the shivering threshold is lowered as well as the sweating, vasodilatation, and warm discomfort thresholds. Indeed, at the end of the attack, shivering immediately takes place and core temperature returns to normal.
|
SUMMARY AND CONCLUSION |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
The observation that it is possible to dissociate experimentally the various thermoregulatory defense responses led recently to question the set point concept and its modifications, fever and anapyrexia. These responses are thus relatively independent from one another, especially in artificial experimental conditions. That was already known (17, 82, 85), but does not mean that they are not physiologically synchronous when the organism operates normally. On the contrary, all the available evidence tends to show that, although independent, all thermoregulatory responses take place synchronously. There may be independent set points for the various defense responses, but in normal life they are all synchronous.
The proposal to abandon the concept of set point in temperature regulation is interesting and stimulating. However, close analysis of the arguments put forward by its proponent simply confirms that the various networks controlling the multiple thermoregulatory responses can operate independently from one another under pharmacological influence or brain lesions but does not invalidate the understanding of temperature regulation acquired on older experiments. I believe that Hammel's concept of adjustable set point is still valid and that to abandon set point, fever, and anapyrexia is, at best, premature.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 In scientific psychology the same trend is also observed. 
2 In my opinion, because a lack of knowledge of what is a regulation.
3 N.B. Body mass is an extensive variable that cannot, as such, be regulated but is used here to make the description simpler.
4 Dublin, July 1971, I believe.
|
REFERENCES |
|---|
|
TOPABSTRACTSET POINTDEFINITIONSTEMPERATURE REGULATION OR HEAT...NO SET POINT?SET POINT OR NULL...ADJUSTABLE SET POINTNO FEVER?NO ANAPYREXIA?SUMMARY AND CONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
O2), left y-axis and open symbols] start rising as soon as esophageal temperature (Tes) drops below +0.1°C. The threshold for evaporative heat loss [a correlate of sweat rate (Esw), right y-axis and solid symbols] is practically superimposed with that of thermogenesis. It follows that there is actually no "null zone." Reprinted from Ref. 
, metabolic heat production (indirect respiratory calorimetry); 
, direct recording of evaporative water loss from the forehead; 
, direct calorimetry from one hand indicative of peripheral vasomotor response. There was no dead zone. Reprinted from Ref. 
