Abstract
The ability of the human body to maintain normothermia via physiological reactions or behavioural changes to vastly different climatic circumstances is an evolutionarily well preserved mechanism. It is a key feature that has allowed humans to adapt to changing environmental conditions, enabling Homo sapiens worldwide evolutionary success and distribution. Humans possess a homeothermic thermoregulatory system, and all enzymatic processes in the human body are optimized to work within a tightly regulated so-called ‘normothermic’ temperature range [1].
Introduction
The ability of the human body to maintain normothermia via physiological reactions or behavioural changes to vastly different climatic circumstances is an evolutionarily well preserved mechanism. It is a key feature that has allowed humans to adapt to changing environmental conditions, enabling Homo sapiens worldwide evolutionary success and distribution. Humans possess a homeothermic thermoregulatory system, and all enzymatic processes in the human body are optimized to work within a tightly regulated so-called ‘normothermic’ temperature range [1].
The non-anaesthetized human body reacts very aggressively to even minor changes of its core temperature. The importance of core temperature regulation and how it is affected by anaesthesia only started to receive the attention it deserves in the 1990s (see Fig. 15.1). The importance of maintaining normothermia and the consequences of accidental perioperative hypothermia have been studied extensively since. More recently, the therapeutic effects of deliberately altering core temperatures are being explored, such as therapeutic hypothermia in the setting of cardiac arrest, myocardial infarction and neuropathologies, fever management (permissive, restrictive or therapeutic) and therapeutic hyperthermia as part of oncological therapies.
Fig. 15.1 Pubmed entries for the search terms ‘perioperative’ and ‘hypothermia’.
Despite these new insights, core temperature measurement is often still considered an optional parameter or ‘vital sign of minor importance’ in many hospitals and operating rooms [2, 3]. However, methods of core temperature manipulation (warming, cooling) and methods of core temperature monitoring should be considered inseparable twins.
A Definition of Core Temperature, Thermoregulatory Compartments and Heat Distribution
Heat distribution can be modelled with a two-compartment model consisting of a core and a peripheral compartment. It suffices to demonstrate the main thermoregulatory functions.
The core compartment consists of all deep body tissues that maintain the organism’s normothermic operating temperature, i.e. the organs in the trunk plus the central nervous system. The peripheral compartment consists of the limbs, the skin and peripheral tissues covering the trunk and the head. The contribution of the latter though is very small.
In a moderate climate (i.e. room temperature of 23°C) the temperature of the peripheral compartment is 2–4°C lower than that of the core compartment. This (radial) gradient depends on the vasoconstrictive or vasodilatatory state of the organism: in a colder environment, a distinct radial temperature gradient is observed, while in a warmer environment vasodilation increases blood flow to the peripheral compartment, decreasing the gradient. Adults produce around 100 Watts of heat while resting, with substantially higher rates during strenuous activity. This heat flows from the core to the peripheral compartment via two mechanisms: convection (i.e. via blood flow, a fast longitudinal heat transfer) and conduction (i.e. via adjacent tissues, a slow radial heat transfer).
To maintain a thermal steady state where heat production matches heat losses, the heat generated by the body must be able to leave and dissipate to the environment. This heat is dissipated mostly via skin, with only 5% of heat being lost via respiration. The most efficient way to lose heat via the skin is by sweating: sweating can dissipate ten times the heat generated by the basal metabolism [4].
Numerous formulas and models with more or less elaborate assumptions and simplifications have been developed to calculate heat content, heat production and heat transfer rates of the human body. These models could be used in clinical practice to predict which patients are prone to develop hypothermia (see below), or to individualize patient warming options.
Since there are (at least, simplistically) two compartments present in human thermoregulation, measuring core temperature alone is not sufficient to assess the full thermal state of the human body. The concept of ‘mean body temperature’ recognizes this. Mean body temperature is the mass-weighted average of tissue temperatures throughout the body. While core temperature is relatively easy to measure (as demonstrated in the next sections), the exact measurement of temperatures of the peripheral compartment requires the insertion of multiple needle thermometers into the limbs at different depths, and both proximally and distally. However, sufficiently accurate estimates of mean body temperature can also be obtained in a much simpler and less invasive manner by using core temperature and non-invasive skin temperatures [5].
The Physiology of Core Temperature
While the body’s core temperature set point itself can vary by approximately 1°C individually due to circadian variation or hormonal changes (menstrual cycle), temperature deviations from this set point as small as ± 0.1°C activate regulatory processes (physiological and behavioural reactions) to counteract the temperature change [6]. This makes sense from an evolutionary point of view, because the enzymes of the body optimally function within the normothermic temperature range (usually defined as a range from 36.5 to 37.5°C).
Physiological thermoregulatory control consists of the three elements: temperature measurement, temperature regulation, and the reactions that core temperature changes trigger. Temperature measurement integrates the input of sensors located in the skin (C-fibres and Aδ-fibres), in deep tissues (i.e. the organs in the trunk and the brain) and in the spinal cord and various regions in the brain (hypothalamus). Most input comes from the skin (50% of all temperature information input) and the hypothalamus [7]. Temperature regulation mostly occurs in the hypothalamus. The third component is the sum of the reactions that core temperature changes trigger. Autonomic responses include vasoconstriction and vasodilation, shivering and sweating; in the non-anaesthetized patient, these reactions also include behavioural changes (e.g. putting on clothes, moving into the sunlight, etc.).
Non-shivering thermogenesis by brown fat occurs to a relevant degree in infants only [8].
The aforementioned autonomic responses (vasoconstriction and vasodilation, shivering and sweating) are very tightly regulated. From a physiological standpoint the properties of these responses can be described by the thresholds at which they are activated, their gain (relationship between the degree of core temperature change and the intensity of autonomic reaction) and their maximum response.
While vasodilation and sweating share the same threshold temperature, shivering occurs up to 1°C below the vasoconstriction threshold. The interval between the sweating and the vasoconstriction threshold includes the so called ‘normothermic range’, also known as the ‘interthreshold range’ (see Fig. 15.2). This interthreshold range is increased by many anaesthetic drugs used and altered by several other medications [9].
There is no general consensus on what exactly constitutes the normal normothermic core temperature range, but most authors agree this is 36.5°C to 37.5°C. Most physicians agree that a core temperature below 36°C definitely constitutes hypothermia, but a grey area remains between 36.0°C and 36.5°C. Reported ‘fever’ thresholds range from 37.2°C to 38.0°C, below which some authors define a ‘subfebrile’ range to describe moderately elevated core temperatures or ‘low-grade fever’.
How Anaesthetic Drugs Affect Core Temperature
General Anaesthesia
Almost all anaesthetic drugs alter thermoregulatory homeostasis in numerous ways. This phenomenon applies to both intravenous drugs (propofol [10], opioids [11]) and inhaled anaesthetics (desflurane [12], isoflurane [13], N2O) [14]. Their effects are mostly comparable: increased sweating threshold, decreased vasoconstriction and shivering thresholds, and near-abolition of non-shivering thermogenesis in infants. The reductions of the thermoregulatory thresholds for shivering and vasoconstriction are typically linearly dose dependent for intravenous drugs, while for volatile anaesthetics the vasoconstriction and shivering thresholds are typically nonlinearly decreased with higher volatile gas concentrations (see Fig. 15.3). Also the gain of vasoconstriction is reduced by volatile anaesthetics, as is maximum shivering intensity. Interestingly, opioids and particularly midazolam have no clinically relevant impact on thermoregulatory thresholds, gain and maximum intensity of shivering [15]. Even though vasoconstriction and shivering thresholds are lowered during general anaesthesia, these compensatory mechanisms are still active beyond these (altered) thresholds. Despite numerous advances in the basic understanding of the mechanisms of anaesthesia, the exact mechanism why and how anaesthetics change thermoregulatory thresholds remains mostly unclear. A potential hypothesis for the understanding of the mechanism is the effect of volatile anaesthetics on the TRPV1 receptor, possibly contributing to both the analgesic effect of volatile anaesthetics and their modification of thermoregulatory input [16].
Different anaesthetic drugs have different effects on the thermoregulatory thresholds of the three key thermoregulatory mechanisms, as displayed in Table 15.1.
| Medication | Vasoconstriction | Shivering | Sweating |
|---|---|---|---|
| Desflurane [12] | pronounced, nonlinear decrease | pronounced, nonlinear decrease | minor linear increase |
| Isoflurane [13] | pronounced, nonlinear decrease | pronounced, nonlinear decrease | minor linear increase |
| Propofol [10] | pronounced, linear decrease | pronounced, linear decrease | slight linear increase |
| Midazolam [15] | minimal, linear decrease | minimal, linear decrease | minimal linear increase |
| Meperidine [17] | pronounced, linear decrease | marked effect on vasoconstriction, linear decrease | slight linear increase |
| Alfentanil [18] | pronounced, linear decrease | pronounced, linear decrease | slight linear increase |
Neuraxial Anaesthesia (Spinal or Epidural Anaesthesia)
The effect of neuraxial anaesthesia on thermoregulation is remarkably similar to that of anaesthetic drugs, i.e. an increased interthreshold range and linear decrease of the vasoconstriction and shivering thresholds (dependent on the number of dermatomes blocked) [19]. Gain and maximum intensity of shivering are also decreased [20]. An accepted hypothesis for this phenomenon is that the thermometric ‘input signal’ is being modified by blocking of the cold sensation from the blocked dermatomes [21]. Blocking the vasomotor nerves causes the vessels in the blocked dermatomes to vasodilate, which accelerates heat loss. After neuraxial anaesthesia, patients have a sensation of warmth, even if they become hypothermic [22, 23, 24]. The effect of epidural and general anaesthesia on thermoregulatory thresholds is additive, causing these patients to be at particular risk of perioperative hypothermia [21].
How Core Temperature Affects Anaesthetic Drugs
If core temperature is lowered, enzymatic activity decreases and the half-life (elimination half-life) of almost all drugs used perioperatively is consequently prolonged. Plasma clearance of vecuronium is decreased by 11% for every 1°C [25]. In the recovery room, propofol plasma concentrations are 28% higher in hypothermic individuals (≈ 34°C) [26]. These factors cause emergence from anaesthesia and recovery room stay of accidentally hypothermic patients to be prolonged [27](also see below).
Complications of Accidental Hypothermia
Thermal Discomfort
Patients experience thermal discomfort if they are hypothermic after anaesthesia [28]. This can be readily alleviated by active skin warming: because the brain integrates the thermometric input of both the skin and the core into a single input to the hypothalamus, active skin warming will trick the hypothalamus into calculating a higher overall mean body temperature [29] closer to the normothermic ‘comfort-zone’.
Impairment of Coagulation
The effects of hypothermia on coagulation are well documented. Impairment of platelet aggregation and reduced clot formation (due to impairment of the enzymatic coagulation cascades) may increase perioperative blood losses: a temperature difference of just 1°C increases the relative risk of transfusion by 22% [30].
Wound Infection and Length of Stay in the Postoperative Care Unit and in the Hospital
Hypothermia can increase the incidence of wound infection in several ways. First, hypothermia-induced vasoconstriction decreases wound perfusion and thus oxygen delivery. The resulting lower tissue O2 partial pressures negatively affect oxidative killing, which typically acts as a first barrier against bacteria [31]. Second, the motility of immune cells vital for bacterial defence is decreased [32]. Finally, impaired formation of collagen fibres compromises wound healing, facilitating wound dehiscence [33]. The clinical relevance for surgical wound healing has been documented in two studies; one in patients undergoing colorectal surgery [34], and one in patients undergoing ‘clean’ surgery after prewarming [35].
Cardiac Complications
Hypothermia causes a stress response, with increased catecholamine levels and increased oxygen consumption [36]. While common sense dictates this might increase cardiac morbidity in patients at risk for cardiac ischaemia, clinical evidence for this complication relies on only a single study with rather coarse methods for the detection of cardiac complications [37].
Shivering
Postoperative shivering can be a response of the autonomic nervous system to accidental perioperative hypothermia. But not all postoperative shivering is caused by hypothermia – the differential diagnosis should include pain, stress, over-activity of the sympathetic nervous system, and unknown causes [38].
Shivering can be quantified by one of several scales, the most widely used being that by Crossley and Mahajan [39].
Table 15.2. Quantification of shivering, scale by Crossley and Mahajan [39].
| 0 | No shivering |
| 1 | No visible muscle activity, but one or more of the following: piloerection, peripheral vasoconstriction, or peripheral cyanosis (after exclusion of other causes). |
| 2 | Muscular activity in one muscle group |
| 3 | Moderate muscular activity in more than one muscle group but no generalized shaking |
| 4 | Violent muscular activity involving the whole body |
Shivering has to be treated rapidly and effectively because it is very uncomfortable for the patient and increases the patient’s oxygen consumption and metabolic rate [40, 41]. There are several drugs with potent anti-shivering properties, which share one common feature: they lower the thermoregulatory threshold for shivering. A potent non-pharmacological therapy is ‘counter’-warming, i.e. actively warming the hypothermic patient’s skin, the mechanism of which has been explained above.
| Drug | Dose (in Respective Study) | Effect on Shivering |
|---|---|---|
| Pethidine [42] | 25 mg | Very strong; can be used in combination with other drugs |
| Clonidine [43] | 75 µg | Good |
| Dexmedetomidine [44] | 0.5 µg/kg | Stronger than clonidine but with more sedation |
| Ketamine [45] | 0.5 mg/kg | Low dose effective |
| Nefopam [46] | 20 mg | Moderate; only on shivering threshold |
| Magnesium [47] | 80 mg/kg + 2 g/h | No clinically relevant effect |
| Doxapram [48] | Titrated to 2.5 µg/ml plasma level | No clinically relevant effect |
| Odansetron [49, 50] | 4–8 mg | Effect controversial |
| Buspirone [51] | 60 mg | Synergistic effect with pethidine |
| Dantrolene [52] | 2.5 mg/kg | Primary effect on gain of shivering |
| Other methods | ||
| Counterwarming/Skin warming [53] | n/a | Reduces shivering threshold; additive effect with pethidine |
In summary, over the last decades scientific evidence has led perioperative normothermia to become a standard-of-care in anaesthesia. Even though many of the studies were published almost 20 years ago, and even though some of them are considered to provide only ‘low to moderate’ evidence by today’s scientific standards, they are unlikely to be repeated using a ‘no-active-warming’ control group because not actively warming perioperatively can no longer be accepted and would be considered unethical [54, 55].
How to Measure Core Temperature
While properly measuring the temperature of a patient might seem a pretty straightforward task, it is not, because there is no consensus on where exactly the patient’s core temperature is actually ‘located’ (in which tissue or body cavity). The problem of definition becomes even more difficult when a patient is undergoing surgery: while cerebral tissue temperature and pulmonary artery temperature are both supposed to be gold standards of core thermometry, they will become invalid e.g. during brain surgery or cardiac/thoracic surgery, respectively. And even though pulmonary artery temperature and intracerebral temperature are both considered ‘gold standards’ of core temperature measurement, even these ‘core temperatures’ are not always in agreement, e.g. after brain injury, with higher temperatures in the cerebral tissue [56–58]. In addition, intracerebral temperature probes and pulmonary artery catheters are too invasive to be used just for core temperature measurement. Thus less invasive surrogate measurements are needed. These surrogate core temperature measurement methods are presented below, from most invasive to least invasive method.
Distal Oesophageal Temperature
During general anaesthesia, distal oesophageal temperature measurement is one of the best substitutes according to numerous studies. Furthermore it is highly resistant to artefacts [59] and dislocation [60]. The text-book recommended placement of the probe is the region of the oesophagus bounded by the left ventricle and aorta, corresponding to the level of the eighth and ninth thoracic vertebrae, but exactly positioning it there can be difficult (particularly in paediatric patients) [61, 62] (see Fig. 15.4a).
Contact Tympanic Temperature
A contact tympanic thermometer very accurately measures brain temperature via a specialized cotton swab probe placed in direct contact with the eardrum (tympanic membrane). The probe may be poorly tolerated by the awake patient, with the sensation ranging from negligible to uncomfortable and even painful. Perforation of the tympanic membrane is a possible complication [63] (see Fig. 15.4b).
Nasopharyngeal Temperature
Nasopharyngeal measurement of core temperature is also often used in the patient undergoing general anaesthesia. Measurements can be affected by airflow over the nasopharyngeal airways and by failing to insert the probe deep enough (10–20 cm) [64]. Care has to be taken so as not to provoke epistaxis during insertion [65] (see Fig. 15.4).
Sublingual Temperature
To measure sublingual temperature accurately, the thermometer has to remain in the ‘sublingual pocket’ during the entire measurement period, which makes continuous measurement difficult and dependent on patient compliance. Oral temperature fluctuates with mouth opening, oral food or fluid intake, and mucosal inflammation [66]. Nevertheless, if used correctly, the sublingual measurement method is a good trade-off between noninvasiveness and accuracy [67] (see Fig. 15.4c).
Fig. 15.4c Standard thermometer for oral, axillary, rectal use.
Bladder Temperature
Bladder temperature is an adequate substitute for core temperature (if the surgical field does not involve the lower abdomen). Complications do not differ from those of regular bladder catheters (i.e perforation, infection, urethral lesions, etc.), since there is no clinically significant difference in size between bladder catheters with and without built-in thermometers [68]. Urine flow may have some influence on the accuracy of the measurement [69] (see Fig. 15.4d).
Rectal Temperature
While some authors claim rectal temperature to be an accurate and relatively noninvasive substitute for core temperature, its time lag of up to one hour prevents it from detecting the rapid onset of fever and hypothermia [70, 71]. Intestinal perforation has been described after the probe has been introduced too deeply or vigorously, which is a particular risk in paediatric patients [72] (see Fig. 15.4c for spot measurements or 15.4d for continuous rectal thermometry).
Temporal Artery Thermometer
Temporal artery thermometers (TATs) – theoretically – measure the temperature above the temporal artery. The actual measurement involves moving the thermometer in a slow, continuous and sweeping motion from the centre of the forehead to the retroauricular area. TATs are less accurate than the aforementioned more invasive measurements and are more affected by environmental factors (cold or heat exposure) and patient factors (vasoconstriction, sweating) [73]. Measurements cannot be made continuously (see Fig. 15.4e).
Axillary Temperature
Despite ample evidence of its low accuracy and poor precision, axillary temperature measurement is still one of the most widely used thermometry methods. Axillary temperature is typically 1–2°C lower than actual core temperature [74, 75]. Accuracy may be decreased even further if the measurement period is too short – at least 4 minutes are required if a mercury-in-glass thermometer is used) [76] or if the probe is not adjacent to the axillary artery (see Fig. 15.4c).
Skin Temperature
Another popular method of thermometry is the measurement of skin temperature, typically on the forehead (via a liquid crystal thermometer or via the human hand). To derive an estimation of core temperature, 2°C have to be added to the measurement. Still, the resulting estimate of core temperature is so inaccurate [77] as to render electronic and liquid crystal forehead thermometers unsuitable for clinical use [78]. Surprisingly, the human hand performs as well as less accurate, yet more high-tech methods in detecting fever, such as liquid crystal forehead thermometers [79] (see Fig. 15.4f).
Infra-red Tympanic
Another very common core thermometry method is infra-red (IR) tympanic thermometry via the external aural canal. However, the method typically measures the temperature of the external ear canal rather than that of the tympanic membrane. This is a result of its shape, designed to prevent deep insertion to minimize the likelihood of tympanic membrane perforation. Cerumen may deteriorate its accuracy. The method is not accurate enough to obtain a reasonable estimate of core temperature in clinical practice [80, 81].
Zero Heat Flux
Zero heat flux technology combines thermal isolation of a skin area (most often the forehead) with a heating element above this area. A thermal equilibrium is established between the heating element and the patient’s core. This allows the core temperature to be derived from the measured skin temperature in the isolated area. This technology is noninvasive, continuous and well tolerated in awake patients. Due to its high accuracy, the technology can provide an alternative to more invasive thermometry [82–84] (see Fig. 15.4g).
Heat Flux/Double Sensor Technology
While heat flux/double sensor technology does not need an active warming element, the double-sensor technology similarly relies on an isolated skin area on the forehead where the effluent heat flow is measured between two thermistors with a standardized insulator in between. Again, the sensor has a high accuracy and can be used as an alternative to more invasive thermometry [85–88] (see Fig. 15.4h).
Experimental Temperature Measurement Methods
Experimental methods used to measure or estimate core temperature include ultrasound [89], MRI [90], measurement via the inner canthus of the eye [91] and simulation (see further) [92–95]. Thermography can be used for fever screening, but is not accurate enough to derive core temperature [96].
How to Model Core Temperature
Modelling and simulation of human core temperature have been facilitated by modern computer technology. Several programmes are available to simulate perioperative patient temperature curves, which can be used as a teaching aid on physiology and thermoregulation or to predict perioperative hypothermia risk and the need for active warming.
Mechanical Models
The most basic mechanical model for human thermoregulation is a simple, passive cylinder.
However, one cylinder alone cannot adequately simulate heat loss because it cannot account for the different shapes of the head, arms, legs and torso that all differ in their surface/mass-ratio. More elaborate models thus do include multiple cylinders [97], and more elaborate models also simulate the effect of sweating and heat distribution between the core and periphery via circulation. Despite several simplifications, some of these models have been successfully used to test patient warming devices [98, 99], and to address other clinical questions that cannot be addressed by clinical studies for practical and/or ethical reasons [100].
Numerical Models
The complexity of numerical models is only limited by the ingenuity of the researcher and computing power. Numerical models divide the human body into several segments that each have their individual thermal capacity, heat production and heat balance. The different segments are connected via blood flow. The most basic model consists of a peripheral and a central compartment, which is able to simulate the main mechanisms of thermoregulation and to predict core and peripheral temperature trends under different ambient and physiological conditions.
Vasoconstriction, Blood Flow and Metabolism in a Basic Two Compartment Model
Blood, circulating in the central compartment, has a high thermal conductivity. A resistor models vasomotion: the resistance changes depending on the difference between the core temperature’s set point and the ‘actual’ core temperature. If the core is too warm, the resistance is decreased and blood flow from the central to peripheral compartment rises to a certain maximum; if the core is too cold, the resistance increases and blood flow from the peripheral to central compartment decreases to a certain minimum.
Heat production generated by metabolism is modelled in an analogue manner, i.e. it will adjust itself depending on the difference between core temperature set point and actual core temperature. When the core temperature is close to its set point, the metabolic rate is typically at its baseline value. When core temperature drops 2°C below the set point temperature, metabolic rate becomes maximal. If core temperature decreases even more, metabolic rate decreases as well.
Example of Model-based Simulation of Thermoregulation
In the following section, the Bussman model is used [101].
This two compartment model incorporates heat losses to the environment by convection, radiation and evaporation and simulates blood flow between the core and peripheral compartment as described above. Body surface area is calculated by an empirical equation [102]. The equation for convection includes ambient air temperature and humidity as well as air velocity between patient and environment or active warming therapy device. Heat exchange via radiation is calculated via skin temperature of the patient and wall temperature of the room. For conduction, heat transfer to or from the insulated blanket or the actively heated device are added to the calculation. The evaporation rate is described by an empirical equation [103]. The core and peripheral temperatures represent the mean temperatures of these body compartments that are considered homogeneously mixed. Blood flows freely between the central organs, but blood flow to the periphery is regulated by vasoconstriction and vasodilation. When the patient’s core temperature and set point differ, vasoconstriction decreases the core-to-peripheral blood flow from approximately 20 mL/kg tissue/min to 1 mL/kg/min if the core temperature is too low or vasodilation increases the core-to-peripheral blood flow up to 50 mL/kg/min if the core temperature is too high.
