Temperature Regulation, Fever, and Hypothermia

Temperature is both a regulated variable and a condition under which organs work. It depends on heat production, environmental exchange, internal distribution, and available responses. Fever, excessive heat accumulation, spontaneous hypothermia, and controlled cooling can produce similar readings through different mechanisms. Their interpretation requires following heat and physiological work together, while distinguishing a temperature response from improved survival or neurological recovery.

68.1 Temperature Changes When Heat Production and Loss Do Not Balance

Metabolic reactions release energy that becomes external mechanical work or heat. Resting organs continually produce heat; muscle activity, shivering, inflammation, and increased cellular work can add to production. Temperature rises when heat accumulates faster than it leaves the body, irrespective of whether the initiating mechanism is regulated fever or an overwhelming external heat load.

Heat crosses the body surface by radiation, convection, conduction, and evaporation. Radiation exchanges energy with surrounding surfaces without direct contact. Convection transfers heat to moving air or water, while conduction occurs through contact with another material. These exchanges can reverse direction when the environment is warmer than the skin. Evaporation removes heat only when water actually vaporizes; sweat that drips away does not provide its full potential evaporative cooling.

For a simplified whole-body balance, let M be metabolic power, Wext external mechanical power, L net nonmetabolic heat loss, and S the rate of heat storage, all in watts, equivalent to joules per second. L includes surface and respiratory exchange and any heat carried into or out of the body by fluids or extracorporeal systems; a net heat gain makes L negative.

Consider invented values of M = 120 W, Wext = 10 W, and L = 90 W, comprising 50 W of dry heat loss and 40 W of evaporation. Storage is 120 − 10 − 90 = 20 W. Over 3,600 seconds, this stores 72,000 J, or 72 kJ.

Let m be body mass in kg, c an assumed average specific heat in J/(kg·°C), t time in seconds, and ΔTmean the change in mean body temperature. The lumped approximation is:

For m = 70 kg and an illustrative c = 3,500 J/(kg·°C), heat capacity is 245,000 J/°C, or 245 kJ/°C. The stored 72 kJ would increase mean temperature by 72 ÷ 245 ≈ 0.29 °C. This explains storage, not an expected hourly core-temperature rise: heat capacity varies, compartments are unequal, and production and loss change during the interval.

68.2 Core and Skin Temperatures Describe Different Thermal Compartments

The body is not a uniformly heated object. Well-perfused central tissues and circulating blood form a relatively warm compartment, while skin and peripheral tissues exchange heat with the environment. The size and temperature of the peripheral compartment change with perfusion, insulation, ambient conditions, and autonomic responses. Blood transports heat internally as well as carrying oxygen and substrates.

Cutaneous vasoconstriction reduces transfer from central tissues to the surface. Vasodilation increases that transfer and can lower core temperature before much additional heat has left the whole body. Conversely, warming peripheral tissues can increase their heat content without an immediately equivalent increase in central temperature. Distribution and total heat balance are separate processes.

Matsukawa and colleagues directly examined heat content and flux in six volunteers undergoing general anesthesia. Much of the initial core-temperature fall resulted from redistribution toward peripheral tissues rather than net environmental loss alone. The experiment demonstrates the mechanism in its anesthetic setting; its proportions are not universal for shock, cooling devices, or every anesthetic.1

Temperature measurements sample this heterogeneous system. Distal esophageal measurements, central blood measurements, bladder measurements, and peripheral readings have different relationships to local perfusion and surrounding tissues. Probe placement, contact, device characteristics, and recent fluid or respiratory interventions can affect a reading. There is no constant correction that converts every peripheral temperature into central temperature.

The time course matters as much as agreement at equilibrium. A slowly responding compartment may still be cooling when another has stabilized or begun warming. In a prospective comparison of 36 post-cardiac-arrest patients, urinary bladder measurements changed more slowly during cooling than esophageal measurements. The study supports site-dependent lag, without proving one site is an infallible reference for every patient.2

A sudden change after moving a probe may be a measurement discontinuity. Conversely, a stable displayed value can conceal changing temperatures elsewhere. Interpreting the trajectory requires the site, its response time, perfusion, and thermal interventions.

68.3 Thermoregulation Uses Integrated Signals and Different Response Thresholds

Thermoreceptors in the skin provide information about environmental exposure, while central and deep-tissue signals reflect internal thermal conditions. These inputs are integrated through hypothalamic and other neural pathways that coordinate behavior, skin blood flow, sweating, and heat production. Skin cooling can initiate defenses before core temperature has fallen substantially.

The familiar thermostat analogy is useful only up to a point. Different effectors have different activation thresholds, gains, and limits. The temperature provoking vasoconstriction need not equal the temperature provoking shivering, and sweating has its own control characteristics. The regulated state emerges from interacting circuits rather than one literal sensor comparing the entire body with a fixed number.

Behavior is a major effector in an awake person: seeking warmth or shade, altering clothing, changing posture, or drinking. Critical illness can remove these options through weakness, impaired consciousness, restraint, or mechanical support. An immobile patient may consequently depend on autonomic responses and environmental care for functions that a healthy person supplies behaviorally.

Autonomic capacity can also change. Sedatives, anesthetics, neural injury, age, metabolic disease, and severe illness influence thresholds or the ability to carry out a response. In a small human experiment, propofol substantially reduced vasoconstriction and shivering thresholds while having much less effect on the sweating threshold. The unequal effects illustrate why pharmacological impairment cannot be described simply as turning all thermoregulation off.3

Effector capacity is distinct from the signal requesting it. A weak patient may be unable to generate substantial shivering despite a strong cold-defense drive. A vasoplegic circulation may not achieve effective cutaneous vasoconstriction. Sweat production can fail, while high ambient humidity limits evaporation even when sweating continues.

Absence of shivering does not prove absence of a cold signal, and high core temperature does not prove that sweating should be active. Interpretation requires operating thresholds, sensed skin conditions, available effectors, and the disease mechanism.

68.4 Fever Couples Immune Signals to Heat-Conserving Responses

Fever involves a regulated change in thermoregulatory responses during inflammation. Microbial products and host inflammatory mediators influence signaling at the brain's vascular and neural interfaces. Prostaglandin E₂ is an important link between immune activation and the circuits controlling heat conservation and production. The detailed inflammatory pathways are developed in Chapter 66.

Brain endothelial cells are an active signaling interface, not simply a route through which cytokines must diffuse. Induction of cyclooxygenase-2 and microsomal prostaglandin E synthase-1 can support local prostaglandin production. In mice, selective deletion of these synthetic pathways in brain endothelial cells attenuated inflammatory fever. This cell-specific experiment supports a causal mechanism in that model without establishing an identical contribution at every stage of human sepsis.4

Prostaglandin signaling modifies central output. Rat experiments by Nakamura and colleagues identified preoptic EP3-receptor-expressing neurons that provide inhibitory signaling to downstream thermogenic pathways. Prostaglandin E₂ inhibited these neurons, releasing heat-producing responses from inhibition. Their manipulations also demonstrated heat-dissipating responses when the neurons were activated. This is a circuit mechanism from an animal model, rather than a direct recording of the full human febrile response.5

During the rising phase, the person's current temperature can be below the newly defended operating range. Vasoconstriction, warmth-seeking behavior, and sometimes shivering follow. The patient can feel cold while the measured temperature is already elevated. This is a coherent regulatory response rather than evidence that the thermometer must be wrong.

When pyrogenic signaling diminishes, the relationship reverses. Heat-loss responses can become active at the existing temperature, producing vasodilation and sweating. The same temperature can therefore accompany chills during one phase and heat dissipation during another.

This temporal pattern also explains why removing heat from the skin and changing the central immune signal are physiologically different interventions. Surface cooling can oppose a defended temperature and provoke compensatory work. Altering pyrogenic signaling can change the operating thresholds themselves. Neither mechanism alone determines the net clinical benefit.

68.5 Fever and Excessive Heat Accumulation Have Different Causes and Costs

An elevated temperature does not identify its mechanism. Fever is associated with changed thermoregulatory signaling. In hyperthermic states, heat production or environmental gain can exceed heat-dissipating capacity without the same immune-mediated upward shift. Thermoregulatory defenses may remain active but be overwhelmed; the term should not imply that every neural feedback pathway has disappeared.

Environmental exposure, intense muscle activity, impaired evaporation, and some drug-related syndromes can generate this imbalance. Excessive skeletal-muscle metabolism and inability to transfer heat outward are different causes, even if their temperatures overlap. A measurement alone cannot distinguish them, and an inflammatory illness can coexist with external heat stress or impaired heat loss.

Fever has potentially useful biological effects, including changes in immune-cell activity and conditions for pathogen growth. These effects depend on organism, tissue, temperature, and timing. They do not prove that every degree of fever is beneficial or that allowing a higher temperature improves outcomes in all patients.

The costs are also contextual. Increased metabolic activity, tachycardia, respiratory work, discomfort, and fluid loss can burden a patient with limited reserve. A healthy circulation may increase flow to support greater oxygen use; a failing heart or injured lung may be unable to do so. The balance depends on the additional demand and the capacity to meet it.

Excessive heating can itself disrupt cellular proteins, membranes, endothelial function, and organ performance. In this setting, inflammatory and coagulation abnormalities can become consequences as well as contributors. Describing all elevated temperatures as a helpful immune response would therefore miss an important causal distinction.

Low temperature is similarly ambiguous. Spontaneous hypothermia can accompany exposure, impaired production, endocrine disease, drugs, or severe systemic illness. A low reading during shock is not physiologically equivalent to controlled temperature reduction in a selected, monitored trial population. Its association with illness severity cannot establish that cold itself explains the entire outcome.

Classification combines trajectory, exposures, context, heat production, and responses. One temperature cannot supply a universal diagnostic boundary or the same interpretation across infection, environmental illness, anesthesia, and brain injury.

68.6 Cooling Can Increase Metabolic Demand When Cold Defenses Remain Active

Cooling transfers heat out of the body, but its net effect depends on the response it provokes. Skin vasoconstriction can reduce delivery of central heat to the surface. Shivering increases muscle activity and metabolic heat production. Adrenergic responses can increase cardiovascular work and redistribute flow. Device heat extraction and patient heat production therefore need not move in parallel.

Lenhardt and colleagues studied nine people with interleukin-2-induced fever under different physical conditions. Active surface cooling commonly provoked shivering and increased oxygen consumption and autonomic responses without significantly reducing the observed fever burden. This induced-fever experiment demonstrates a mechanism; its conclusions are not a universal rule for sedated, paralyzed, or otherwise critically ill patients.6

The oxygen consequence can be illustrated without assuming a fixed effect of temperature. Let V̇O₂ be whole-body oxygen consumption in mL O₂/min, CO cardiac output in L/min, and CaO₂ and CvO₂ arterial and mixed venous oxygen contents in mL O₂/dL. At steady state, the Fick relationship with this unit convention is:

For an invented CO of 5 L/min and V̇O₂ of 250 mL/min, the required content difference is 250 ÷ (5 × 10) = 5 mL/dL. If CaO₂ is 18 mL/dL, CvO₂ is 13 mL/dL. Suppose measured demand instead becomes 350 mL/min while output and arterial content remain unchanged. The difference must become 350 ÷ 50 = 7 mL/dL, leaving CvO₂ at 11 mL/dL.

The second state requires greater extraction; the chosen increase is illustrative, not a predicted percentage per degree or a safety threshold. Whether tissues can achieve it depends on regional flow, diffusion, extraction reserve, and mitochondrial use. A whole-body mixed venous value can conceal regional inadequacy.

Reduced muscle activity or changed thresholds may allow heat removal without the same demand increase. Nevertheless, falling core temperature does not independently establish that cardiovascular and respiratory work have fallen.

68.7 Hypothermia Changes Metabolism Without Abolishing Cellular Requirements

Lower temperature slows many enzyme-mediated processes, membrane transport systems, and electrical events. Oxygen consumption and carbon dioxide production can fall when these effects dominate. However, shivering, adrenergic stimulation, seizures, inflammation, and ongoing work can oppose that reduction. The net metabolic response belongs to the whole clinical state, not to temperature in isolation.

Temperature sensitivity is often summarized by Q₁₀, the rate at a higher temperature divided by the rate 10 °C lower. This coefficient is specific to the process and conditions studied. It need not remain constant across temperatures, organs, disease states, or measurement methods. A single universal percentage reduction in oxygen use per degree is therefore inappropriate.

Human cardiac-surgical observations illustrate both the phenomenon and its limits. McCullough and colleagues estimated cerebral metabolism during cooling in adults undergoing operations involving hypothermic circulatory arrest. Oxygen use was suppressed but remained detectable even at deep hypothermia. Their flow estimates and surgical conditions do not define a universal cerebral Q₁₀ or a safe duration of interrupted circulation for other patients.7

Lower ATP demand is one reason hypothermia can be protective in experimental ischemia. Temperature also influences excitatory neurotransmission, ion movements, inflammation, and injury pathways. These mechanisms do not imply that all tissue injury stops or that systemic complications disappear. Existing ischemic damage, regional flow failure, and loss of cellular integrity can continue despite reduced average demand.

Supply and demand may change together. A lower cardiac output can accompany lower metabolism, while vasoconstriction or microvascular abnormalities alter regional supply. A smaller whole-body oxygen consumption is not automatically evidence of improved oxygen balance: it can also accompany loss of functioning tissue or impaired utilization.

Metabolic measurements likewise require context. Indirect calorimetry depends on gas exchange, stable conditions, and technical assumptions. Changing body stores and nonsteady elimination can separate measured pulmonary exchange from instantaneous tissue metabolism. The physiological rationale for cooling therefore remains distinct from the randomized evidence about neurological function and survival in a particular population.

68.8 Cardiovascular Effects Depend on Temperature, Phase, and Reserve

Cold defense increases cutaneous vascular resistance and reduces heat transfer to the skin. This redistributes blood and can alter central filling pressures. Pressure changes do not necessarily represent an equivalent change in total blood volume or useful ventricular preload. Myocardial compliance, intrathoracic pressure, vascular tone, and the state of the heart modify the relationship.

Cooling also affects pacemaker activity, conduction, repolarization, and contraction. Heart rate often falls as temperature decreases, while vascular resistance can rise. Cardiac output depends on the combined changes in rate, stroke volume, loading, and myocardial performance. Bradycardia cannot be interpreted independently of perfusion and demand, nor does an increased pressure guarantee increased flow.

The response is not monotonic. Early cold stress and shivering can increase adrenergic activity and cardiac work, whereas deeper cooling can suppress electrical and mechanical function. Electrolyte abnormalities, ischemia, acid–base disturbances, and drugs interact with direct temperature effects. This combination can increase susceptibility to clinically consequential arrhythmias.

Vascular resistance and capacitance also change during warming. Relaxation of cold-induced vasoconstriction increases peripheral perfusion and can lower pressure when cardiac output or circulating volume cannot compensate. Fluid losses that developed during cooling may become more evident as the vascular space redistributes. These mechanisms do not justify inferring volume deficiency from every episode of warming-associated hypotension.

Local temperatures can lag behind central measurements. Blood returning from colder tissues and continuing heat redistribution can influence the central trajectory during transitions. The observed response depends on how heat is transferred and on ongoing environmental losses, rather than on the displayed central temperature alone.

These effects distinguish accidental exposure, spontaneous hypothermia in severe illness, and controlled cooling after resuscitation. Their underlying myocardial and vascular conditions differ, as do the opportunities for feedback and support. A physiological change that is tolerated in one setting may become harmful in another.

Interpretation integrates rhythm, flow, pressure, regional perfusion, and oxygen demand during the current phase. No single heart rate, pressure, or temperature supplies that account.

68.9 Blood Gases and Electrolytes Change Through Several Mechanisms

Temperature affects gas solubility and chemical equilibria as well as metabolism. For the same blood gas content, cooling generally lowers gas partial pressures and raises pH relative to the values obtained after warming the sample. Routine blood gas analysis is performed at a standardized temperature, commonly 37 °C; temperature-adjusted results estimate values under the patient's thermal conditions.

Alpha-stat uses values reported at standard temperature while allowing in situ pH and partial pressures to change with cooling. In contrast, pH-stat uses temperature-adjusted values to guide gas management, generally requiring more carbon dioxide during cooling relative to alpha-stat. Their implications depend on population and setting. Serial interpretation must retain the reporting convention; neither approach supplies universal targets for every patient.

Experiments by Bradley and colleagues showed that blood gas tensions change with temperature in anaerobically enclosed samples. This supports distinguishing measurement-temperature values from patient-temperature estimates; it does not establish which management convention improves clinical outcomes. Derived gas-exchange variables inherit the same conventions.8

Cooling shifts hemoglobin's oxygen affinity, tending to favor binding at a given partial pressure. Tissue unloading still depends on local temperature, pH, carbon dioxide, flow, and demand. A partial pressure cannot therefore replace oxygen content, and a normal-looking arterial value cannot establish adequate regional oxygen delivery.

Electrolytes change through transcellular redistribution, renal losses, altered intake, and treatment. Cooling can accompany increased urine production and urinary electrolyte loss, even while renal perfusion or filtration is impaired. A human severe-head-injury comparison documented falls in several electrolyte concentrations and increased urinary losses during induced cooling. Its small, selected population does not establish a universal depletion pattern or a prophylactic replacement prescription.9

Potassium illustrates the importance of phase: movement into cells during cooling can contribute to a lower plasma concentration, while warming can reverse part of the redistribution. Tissue injury, kidney function, acid–base changes, and previously administered potassium modify the result. Plasma concentration is not a direct measure of total-body stores, and neither cooling nor rewarming predicts one fixed electrolyte response.

68.10 Coagulation and Drug Handling Are Temperature Dependent

Hemostasis depends on enzyme reactions, platelet behavior, blood flow, endothelial function, and the available cellular and plasma components. Hypothermia can slow clot-generating reactions without reducing the measured concentration of every clotting factor. In pooled normal plasma, Rohrer and Natale demonstrated prolonged clotting times when testing was performed at lower temperatures. This isolates a thermal effect that can be missed when a patient's sample is tested after warming to standard laboratory conditions.10

Platelet responses also depend on temperature and experimental conditions. Human in vitro and local cooling experiments demonstrated reversible inhibition of several activation and aggregation responses. These observations support a mechanistic contribution to impaired hemostasis, but are not a complete model of trauma, extracorporeal circulation, or post-arrest coagulopathy.11

The net clinical phenotype remains multifactorial. Acidosis, dilution, tissue injury, consumption, anticoagulant exposure, and endothelial activation can coexist with cooling. Rewarming a sample does not restore missing factors, and warming a patient does not automatically reverse every cause of bleeding or thrombosis. Conversely, a relatively preserved test obtained at standard temperature cannot exclude a temperature-dependent in vivo impairment.

Drug disposition changes for similar reasons. Temperature can influence enzymatic metabolism, transport, organ blood flow, distribution, and binding. The effect varies by drug and elimination pathway. Hypothermia, inflammation, and organ failure may combine, making a concentration or duration of effect difficult to attribute to temperature alone.

Hostler and colleagues used midazolam as a metabolic probe in a small crossover experiment in healthy volunteers. Mild, short-duration cooling altered metabolism and modeled clearance. This provides human proof of a kinetic effect, without supplying a universal per-degree correction for critically ill patients or for other drugs.12

Pharmacodynamics can change as well as concentration. Sedation, neuromuscular function, and cardiovascular responsiveness may evolve during cooling and recovery. A delayed clinical response after warming can reflect persistent drug exposure, tissue effects, underlying neurological injury, or several mechanisms together. Temperature history therefore belongs alongside dosing history and organ function when interpreting drug effects.

68.11 Outcome Trials Test Specific Temperature Strategies

Reducing a temperature or metabolic variable demonstrates a physiological effect. Trials must establish whether a particular strategy improves outcomes in its enrolled population. Fever treatment during infection, hypothermia after cardiac arrest, and the duration of an already established cooling strategy are different questions.

HEAT randomized 700 febrile ICU patients with known or suspected infection to acetaminophen or placebo. It found no significant difference in the primary outcome of ICU-free days through day 28. This does not establish that fever is always beneficial or that symptom relief has no value; it evaluates the tested pharmacological strategy and endpoint.13

TTM2 compared hypothermia with targeted normothermia and early fever treatment after out-of-hospital cardiac arrest of presumed cardiac or unknown cause. It found no reduction in six-month mortality or poor functional outcome with hypothermia, while arrhythmias causing hemodynamic compromise were more frequent. Its comparator was active temperature management, not unrestricted fever.14

HYPERION studied a different population: comatose survivors of cardiac arrest with nonshockable rhythms, including in-hospital arrests. Its primary favorable neurological-survival outcome at day 90 was more frequent with hypothermia, with a confidence interval only narrowly excluding no difference; mortality did not significantly differ. The population, endpoint, and uncertainty should remain visible when comparing its result with TTM2.15

ICECAP, published in 2026, addressed duration within a hypothermic strategy. In 1,158 selected out-of-hospital cardiac-arrest survivors already cooled rapidly, adaptive randomization compared durations of treatment at 33 °C. Increasing duration did not improve the modeled 90-day neurological outcome, and the trial met a prespecified stopping rule. It did not compare hypothermia against no cooling or prove that its shortest duration is appropriate for every patient.16

These studies cannot be combined into a claim that temperature never matters, that lower is always better, or that one duration or temperature is universally optimal. Their relevance comes from retaining the exact physiological intervention, comparator, population, and patient-centered outcome rather than treating all temperature management as one treatment.

68.12 Rewarming and Clinical Interpretation Require the Whole Trajectory

Rewarming is an active physiological transition rather than simple reversal of a number. Metabolic demand can increase as tissues warm; vasodilation changes vascular distribution; electrolytes redistribute; and drug clearance and responsiveness evolve. The direction and timing of these changes depend on prior cooling, illness, organ function, and ongoing treatment.

Thermal compartments need not warm together. A central reading can improve while peripheral tissues remain cold and residual gradients redistribute heat. Changes in probe site or perfusion add apparent discontinuities. Interpretation connects measurements with heat-transfer methods and the sequence of interventions.

Consider a constructed patient whose central temperature falls during surface cooling while oxygen consumption and vasopressor requirement rise. The thermal result alone cannot establish reduced physiological stress; persistent cold defenses are one possible explanation. If pressure later falls during warming, vasodilation, cardiac reserve, fluid losses, drugs, and the underlying illness all remain relevant rather than one automatic diagnosis of hypovolemia.

Current guidance also preserves population and purpose. The 2026 Surviving Sepsis Campaign conditionally suggests against pharmacological or surface antipyretic treatment solely to improve clinical outcomes in adults with sepsis or septic shock, citing very low-certainty evidence. It explicitly excludes symptom or pain control and other indications such as neurocritical care and post-arrest temperature management from that statement.17

The 2025 American Heart Association guidance recommends deliberate, protocolized temperature control for adults who remain unresponsive to verbal commands after return of spontaneous circulation. That population-specific recommendation is not a claim that one hypothermic temperature is superior for every patient, and the document predates the 2026 ICECAP publication.18

Temperature physiology helps explain why these questions remain distinct. Interpretation combines cause, measurement, heat balance, compensatory work, perfusion, and the changing demands of recovery. A temperature trajectory is one component of that account; it cannot alone establish tissue protection, adequate circulation, or the net value of an intervention.

References

1. Matsukawa T, Sessler DI, Sessler AM, et al. Heat flow and distribution during induction of general anesthesia. Anesthesiology. 1995;82:662–673. Original human heat-balance experiment.

2. Umińska JM, Buszko K, Ratajczak J, et al. Comparison of temperature measurements in esophagus and urinary bladder in comatose patients after cardiac arrest undergoing mild therapeutic hypothermia. Cardiol J. 2020;27:735–741. Original prospective measurement study.

3. Matsukawa T, Kurz A, Sessler DI, et al. Propofol linearly reduces the vasoconstriction and shivering thresholds. Anesthesiology. 1995;82:1169–1180. Original human threshold experiment.

4. Wilhelms DB, Kirilov M, Mirrasekhian E, et al. Deletion of prostaglandin E₂ synthesizing enzymes in brain endothelial cells attenuates inflammatory fever. J Neurosci. 2014;34:11684–11690. Original mouse experiments.

5. Nakamura Y, Yahiro T, Fukushima A, et al. Prostaglandin EP3 receptor-expressing preoptic neurons bidirectionally control body temperature via tonic GABAergic signaling. Sci Adv. 2022;8:eadd5463. Original rat circuit experiments.

6. Lenhardt R, Negishi C, Sessler DI, et al. The effects of physical treatment on induced fever in humans. Am J Med. 1999;106:550–555. Original induced-fever experiment.

7. McCullough JN, Zhang N, Reich DL, et al. Cerebral metabolic suppression during hypothermic circulatory arrest in humans. Ann Thorac Surg. 1999;67:1895–1899. Original human surgical physiology study.

8. Bradley AF, Stupfel M, Severinghaus JW. Effect of temperature on PCO₂ and PO₂ of blood in vitro. J Appl Physiol. 1956;9:201–204. Original blood gas experiments.

9. Polderman KH, Peerdeman SM, Girbes ARJ. Hypophosphatemia and hypomagnesemia induced by cooling in patients with severe head injury. J Neurosurg. 2001;94:697–705. Original clinical physiological comparison.

10. Rohrer MJ, Natale AM. Effect of hypothermia on the coagulation cascade. Crit Care Med. 1992;20:1402–1405. Original plasma experiments.

11. Michelson AD, MacGregor H, Barnard MR, et al. Reversible inhibition of human platelet activation by hypothermia in vivo and in vitro. Thromb Haemost. 1994;71:633–640. Original human and laboratory experiments.

12. Hostler D, Zhou J, Tortorici MA, et al. Mild hypothermia alters midazolam pharmacokinetics in normal healthy volunteers. Drug Metab Dispos. 2010;38:781–788. Original human crossover experiment.

13. Young P, Saxena M, Bellomo R, et al. Acetaminophen for fever in critically ill patients with suspected infection. N Engl J Med. 2015;373:2215–2224. HEAT randomized trial.

14. Dankiewicz J, Cronberg T, Lilja G, et al. Hypothermia versus normothermia after out-of-hospital cardiac arrest. N Engl J Med. 2021;384:2283–2294. TTM2 randomized trial.

15. Lascarrou JB, Merdji H, Le Gouge A, et al. Targeted temperature management for cardiac arrest with nonshockable rhythm. N Engl J Med. 2019;381:2327–2337. HYPERION randomized trial.

16. Meurer WJ, Yeatts SD, Geocadin RG, et al. Duration of therapeutic hypothermia after out-of-hospital cardiac arrest: the ICECAP randomized clinical trial. JAMA. 2026;336:857–867. Original adaptive randomized trial.

17. Prescott HC, Antonelli M, Alhazzani W, et al. Surviving Sepsis Campaign: international guidelines for management of sepsis and septic shock 2026. Crit Care Med. 2026. Official SCCM recommendations and remarks.

18. American Heart Association. Part 11: post–cardiac arrest care. 2025 guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Official temperature-control recommendations.