Vital Signs in Accidental Hypothermia

Abstract

Pasquier, Mathieu, Evelien Cools, Ken Zafren, Pierre-Nicolas Carron, Vincent Frochaux, and Valentin Rousson. Vital signs in accidental hypothermia. High Alt Med Biol. 22: 142–147, 2021.

Background:

Clinical indicators are used to stage hypothermia and to guide management of hypothermic patients. We sought to better characterize the influence of hypothermia on vital signs, including level of consciousness, by studying cases of patients suffering from accidental hypothermia.

Materials and Methods:

We retrospectively included patients aged ≥18 years admitted to the hospital with a core temperature below 35°C. We identified the cases from a literature review and from a retrospective case series of hypothermic patients admitted to the hospital between 1994 and 2016. Patients who experienced cardiac arrest, as well as those with potential confounders such as concomitant diseases or intoxications, were excluded. Relationships between core temperature and heart rate, systolic blood pressure, respiratory rate, and level of consciousness were explored via correlations and regression.

Results:

Of the 305 cases reviewed, 216 met the criteria for inclusion. The mean temperature was 29.7°C ± 4.2°C (range 19.3°C–34.9°C). The relationships between temperature and each of the four vital signs were generally linear and significantly positive, with Spearman correlations for respiratory rate, heart rate, systolic blood pressure, and Glasgow Coma Score (GCS) of 0.29 (p = 0.024), 0.44 (p < 0.001), 0.47 (p < 0.001), and 0.78 (p < 0.001), respectively. Based on linear regression, the mean decrease of a vital sign associated with a 1°C decrease of temperature was estimated to be 0.50 minute⁻¹ for respiratory rate, 2.54 minutes⁻¹ for heart rate, 4.36 mmHg for systolic blood pressure, and 0.88 for GCS.

Conclusions:

There is a significant positive correlation between core temperature and heart rate, systolic blood pressure, respiratory rate, and GCS. The relationship between vital signs and temperature is generally linear. This knowledge might help clinicians make appropriate decisions when determining whether the clinical condition of a patient should be attributed to hypothermia. This could enhance clinical care and help to guide future research.

Introduction

The influence of accidental hypothermia on the clinical condition of a patient depends on the severity of hypothermia. Correlations between core temperature and vital signs, including level of consciousness, is used in international guidelines to guide the management of accidental hypothermia (Durrer et al., 2003; Brown et al., 2012; Paal et al., 2016; Dow et al., 2019) and have recently been studied in stratification models (Deslarzes et al., 2016). Hypothermia leads to a decrease in heart rate, blood pressure, and respiratory rate (Danzl, 2012; Dow et al., 2019). However, there are limited data published regarding these correlations. We sought to study the correlations between core temperature and vital signs (heart rate, systolic blood pressure, respiratory rate, and level of consciousness) by using a retrospective database of hypothermic patients.

Materials and Methods

This study was based on data that were collected for two retrospective studies (Deslarzes et al., 2016; Pasquier et al., 2019). The first of the two retrospective studies included cases identified by a literature review (Deslarzes et al., 2016). The second was a retrospective case series of hypothermic patients admitted to the hospital between 1994 and 2016 (Pasquier et al., 2019). Both studies included patients ≥18 years old with accidental hypothermia (core temperature below 35°C) for which the data on clinical parameters and vital signs at presentation allowed classification using the Swiss staging system (Durrer et al., 2003). To ensure that the clinical state of each patient could be confidently attributed to hypothermia alone, we specifically searched for confounding factors in the publications for the literature review (Deslarzes et al., 2016) and in the medical charts for the case series (Pasquier et al., 2019). The exclusion criteria were acute alcohol, drug or other intoxications, hypoglycemia <3 mmol/l, traumatic brain injury and medical conditions that could lead to hypothermia (myxedema, anorexia), influence vital signs, or impair consciousness (sepsis, shock, central nervous system conditions). A blood alcohol concentration up to 150 mg/dl was not an exclusion criterion, but it was considered to be the threshold at which signs of altered consciousness might occur (Finnell and McMicken, 2010). We excluded patients who had experienced cardiac arrest. We refer the reader to source articles for information about the methodology used to create the database (Deslarzes et al., 2016; Pasquier et al., 2019).

The following data were recorded: age, sex, initial vital signs (heart rate, systolic blood pressure, respiratory rate, and level of consciousness using the Glasgow Coma Score [GCS]), first recorded core temperature, cause of accidental hypothermia (immersion in water, avalanche burial, environmental exposure, or other), survival to hospital discharge, and neurological outcome using cerebral performance category (CPC: 1 = normal or slightly diminished cerebral function, 2 = moderate cerebral disability, 3 = severe cerebral disability, 4 = coma or vegetative state, 5 = brain dead) (Safar, 1981). The primary analysis was the correlation between initial vital signs, including level of consciousness, and core temperature.

The hospital data collection was approved by the institutional ethics committee (Commission cantonale [Vaud] d'éthique de la recherche sur l'être humain).

Statistical analysis

Descriptive statistics were calculated for the variables of interest and were expressed as frequencies, means, and standard deviations, depending on the characteristics of the variables (categorical or quantitative). The relationship between temperature and each of the four vital signs (respiratory rate, heart rate, systolic blood pressure, and GCS) was described by using scatterplots and summarized by Spearman correlations. To investigate the form of these relationships, a simple linear regression model, with temperature as the dependent variable and each vital sign as the independent variable, from which simple formulas to approximate the expected values and prediction intervals for the vital signs depending on temperature are available, was compared with a nonparametric fit (smoothing splines), obtained by using the statistical software R, version 3.1.2 (R Foundation for Statistical Computing). p-Values <0.05 were considered statistically significant.

Results

Of the 305 patients, 216 met the inclusion criteria, 114 (53%) from direct hospital sampling and 102 (47%) from the literature review. Of the 216 patients who were included, 126 (58%) were male. The mean age was 55 ± 24 years (range 1–93 years). The mean core temperature was 29.7°C ± 4.2°C (range 19.3°C–34.9°C). The cause of hypothermia was immersion in water in 45 (21%), avalanche burial in 12 (6%), and environmental or missing in 159 (74%) patients.

The vital signs are presented in Table 1. Scatterplots showing the relationship between temperature and each of the four vital signs (respiratory rate, heart rate, systolic blood pressure, and GCS) are shown in Figure 1. The four associations were significantly positive with Spearman correlations of 0.29 (p = 0.024), 0.44 (p < 0.001), 0.47 (p < 0.001), and 0.78 (p < 0.001), respectively. The four relationships were generally linear, as illustrated by the closeness of the linear fit and the nonparametric fit. Based on linear fit, the mean decrease of a vital sign associated with a 1°C decrease in temperature was estimated to be 0.50 minute⁻¹ (respiratory rate), 2.54 minutes⁻¹ (heart rate), 4.36 mmHg (systolic blood pressure), and 0.88 (GCS) (Table 2). The corresponding R² (interpreted as percentage of variance of a vital sign linearly predicted by temperature) were 8%, 16%, 25%, and 56%, respectively. The expected vital signs according to different hypothermia stages are presented in Table 3.

FIG. 1.

Relationship between temperature and each of the four vital signs (respiratory rate, heart rate, systolic blood pressure, and GCS). The four relationships were generally linear, as illustrated by the closeness of the linear fit (continuous line) and the nonparametric fit (dotted line). Rho: Spearman's rank correlation coefficient. GCS, Glasgow Coma Score.

Table 1.

Vital Signs of 216 Non-Arrested Hypothermic Patients

Vital signs
Respiratory rate, breaths/min (n = 60; missing = 156)
Mean (SD)	17 (6)
Range	5–30
Heart rate, beats/min (n = 174; missing = 42)
Mean (SD)	69 (26)
Range	18–179
Systolic blood pressure, mmHg (n = 157; missing = 59)
Mean (SD)	121 (31)
Range	40–188
GCS (n = 185; missing = 31)
Mean (SD)	11 (5)
Range	3–15
Clinical stage according to the Swiss staging system (3)
Stage 1, n (%)	89 (41)
Stage 2, n (%)	47 (22)
Stage 3, n (%)	80 (37)

GCS, Glasgow Coma Score; SD, standard deviation.

Table 2.

Results of the Linear Regression

Vital sign	n	Intercept	Slope	p	R² (%)	SD
Respiratory rate, breaths/min	60	2.256	0.501	0.028	8	6.249
Heart rate, beats/min	174	−6.876	2.538	<0.001	16	23.743
Systolic blood pressure, mmHg	157	−13.274	4.361	<0.001	25	27.120
GCS	185	−15.294	0.876	<0.001	56	3.221

Table 3.

Expected Vital Signs According to Hypothermia Stage

Vital sign	35°C	32°C	30°C	28°C	26°C	24°C	22°C	20°C
Respiratory rate, breaths/min	20	18	17	16	15	14	13	12
Heart rate, beats/min	82	74	69	64	59	54	49	44
Systolic blood pressure, mmHg	139	126	118	109	100	91	83	74
GCS	15	13	11	9	8	6	4	3

The expected values can be calculated as intercept + slope × temperature, using the intercepts and slopes provided in Table 2. To calculate prediction intervals containing ∼95% of individual values, one may add and subtract twice the SD from Table 2. Calculated values below 0 should be set to 0. For GCS, calculated values higher than 15 should be set to 15 and calculated values lower than 3 should be set to 3.

Hospital outcomes were available for 200 patients, of whom 7 patients (3.5%) died and 193 (96.5%) survived. CPC scores were available for 189 patients, of whom 188 patients (99.5%) had a CPC of 1.

Discussion

In our study, core temperature was positively and significantly associated with each of the four vital signs. All these relationships were generally linear, although with some variability around the linear trend. The strongest association was between temperature and GCS (R² = 56%). These results allow estimation of the vital signs to be expected for a given level of hypothermia when the core temperature is known. These associations could also be exploited to predict the core temperature when the vital signs are known.

Previous studies have shown that mild hypothermia induces tachycardia. Bradycardia develops as the temperature decreases further (Darocha et al., 2015; Loppnow and Wilson, 2015; Truhlář et al., 2015; Rischall and Rowland-Fisher, 2016). In the middle of the 20th century, dog studies demonstrated that heart rate decreased linearly during induced hypothermia (Hamilton et al., 1937; Prec et al., 1949; Rittenhouse et al., 1971). Heart rate dropped to about 20% of normal levels with induced hypothermia to 20°C in dogs (Rittenhouse et al., 1971, 1974). These analyses were subject to many confounders, especially because anesthetic agents known to reduce heart rate, respiratory rate, systolic blood pressure, and GCS were used to sedate the animals. In clinical practice, bradycardia generally occurs at core temperatures below 30°C (Dow et al., 2019).

A linear relationship between the respiratory rate and core temperature was first described in hypothermic dogs (Prec et al., 1949). Prior research in hypothermic individuals revealed that the respiratory rate increased in mild hypothermia, but it decreased in moderate and severe hypothermia (Danzl, 2012; Rischall and Rowland-Fisher, 2016). The reduced respiratory rate is likely due to a decrease in CO₂ production, which drops to 50% for each 8°C fall in temperature (Danzl, 2012).

Our results showed that systolic blood pressure decreased linearly with a decreasing core temperature. These findings are in contrast to previous research, which showed an increase in blood pressure in mild hypothermia due to catecholamine release, leading to peripheral vasoconstriction and increased cardiac output (Rischall and Rowland-Fisher, 2016). In moderate and severe hypothermia, heart rate and cardiac output decrease and blood pressure drops. Hypotension is further exacerbated by fluid shifts and hypovolemia secondary to cold diuresis. Significant hypotension can be expected once the core temperature reaches 24°C (Rischall and Rowland-Fisher, 2016). In dogs, mean aortic pressure reaches a plateau at 30°C and then it gradually decreases (Rittenhouse et al., 1971). Systolic blood pressure is an important parameter, because in-hospital deaths have been associated with hypotension independent of core temperature (Matsuyama et al., 2018). We demonstrated that GCS decreases proportionately to the degree of hypothermia, as shown in previous studies (Danzl, 2012; Matsuyama et al., 2018), although individual responses to hypothermia differ significantly (Rischall and Rowland-Fisher, 2016). Our setting of uncontrolled patients may explain the differences found between our findings and some of these experimental results.

It is helpful to be able to estimate the expected vital signs for a given core temperature. Because hypothermia, especially deep hypothermia, is rare, most clinicians do not have the experience to estimate expected vital signs for a given core temperature. Our study gives an estimate of expected vital signs based on the core temperature, from where prediction intervals containing (∼) 95% of the individual values can also be calculated (Table 3). A dedicated calculator is available on the Hypothermia Score website. The comparison between the expected and actual vital signs of a given patient may have significant practical consequences. To illustrate this point, we present two hypothetical scenarios. In the first scenario, a hypothermic patient has vital signs that are close to those predicted by the core temperature. In this case, vital signs may help a clinician to manage hypothermia effectively. Bradycardia and hypotension associated with hypothermia should not be treated provided that they are directly related to hypothermia. This suggests that perfusion is adequate (Danzl, 2012; Paal et al., 2016). Confidence that a patient's vital signs are compatible with the level of hypothermia may help clinicians avoid unnecessary use of vasopressors or inotropes, which can be toxic in hypothermic patients due to decreased metabolism (Truhlář et al., 2015; Dow et al., 2019).

In the second scenario, a patient has vital signs that are unexpectedly low or high compared with those expected for a given core temperature. We showed that hypothermia almost linearly reduces heart rate, respiratory rate, systolic blood pressure, and GCS. A significant deviation from this linear relationship should lead to a broadening of the differential diagnosis (Dow et al., 2019). For example, if the heart rate is abnormally high for the level of hypothermia, hypovolemia or hemorrhage may be suspected as the cause of a relative sinus tachycardia (Pasquier et al., 2011). The same reasoning might apply for an unexpectedly low systolic blood pressure. A heart rate slower than expected should raise suspicion for another cause of bradycardia, such as drug intoxication. Drug intoxication might also be the cause for an unexpectedly slow respiratory rate or a decreased level of consciousness. Hypoglycemia, hypovolemia, alcohol, and other intoxications are the most common conditions that may interfere with vital signs (Table 4) (Rischall and Rowland-Fisher, 2016).

Table 4.

Partial List of Underlying Medical Condition That May Explain Disproportion Between the Clinical Findings and the Hypothermia Stage

Vital sign	Higher than expected	Lower than expected
Respiratory rate	Metabolic acidosis	Drug intoxication (opioids)
Heart rate	Arrhythmias (AF); shock (hypovolemic/septic); hypoglycemia; drug intoxication (anticholinergic and serotonergic drugs, amphetamines, cocaine, TCAs); drug side effects (ketamine)	Drug intoxication (opioids, alcohol)
Systolic blood pressure	Drug intoxication (anticholinergic and serotonergic drugs, amphetamines, cocaine, TCAs); drug side effects (ketamine)	Arrhythmias (AF); shock (hypovolemic/septic); intoxication (opioids, alcohol)
GCS		Intoxication (alcohol, drugs); hypoglycemia; sepsis

Normal and disproportionate heart rates according to the hypothermia stage.

AF, atrial fibrillation; TCA, tricyclic antidepressant.

Clinical findings or vital signs that are not consistent with the level of hypothermia should be addressed, and other causes of decreased consciousness and depressed vital signs should be considered (Danzl, 2012; Rischall and Rowland-Fisher, 2016). The same reasoning would also apply to therapeutic decision making. For example, a blood pressure that is lower than expected for the measured core temperature may require active treatment and additional testing (Oung et al., 1992; Danzl, 2012). As far as we know, our results are the first to provide concrete data on the vital signs to be expected for a given hypothermic core temperature.

Our study may also give additional insights about the clinical staging of hypothermia. Low-reading thermometers are not available in some settings (Pasquier et al., 2012; Karlsen et al., 2013; Strapazzon et al., 2014; Niven et al., 2015; Darocha et al., 2017; Henriksson et al., 2017; Podsiadło et al., 2017). It can be difficult to obtain an accurate temperature measurement in the prehospital setting, especially in a conscious patient (Strapazzon et al., 2014, 2015; Freeman et al., 2018; Dow et al., 2019). Development of new types of thermometers is ongoing and might provide a solution for the future (Darocha et al., 2017). However, even if an appropriate thermometer is developed, it might not always be available. It would seem prudent to have a reliable method of estimating core temperature based on clinical findings. The Swiss staging model is widely used in prehospital settings, due to the difficulty of measuring core temperature in the field (Brown et al., 2012; Zafren et al., 2014; Truhlář et al., 2015; Paal et al., 2016). In patients with accidental hypothermia that is not associated with trauma or other medical conditions, use of the vital sign levels, in conjunction with the Swiss clinical staging system (Deslarzes et al., 2016; Pasquier et al., 2019), might help to estimate temperature more reliably, using only clinical observations.

Limitations

The main limitation of our study is that it was retrospective and used a non-random case sample. We cannot be certain that proper methods were used to measure the core temperatures of the hypothermic patients that we included. Inappropriate temperature measurement or delays between measurement and clinical evaluations might have biased our results, although we think that these problems, if they occurred, would not lead to systematic bias. Additional prospective studies or the analysis of good-quality prospectively collected data would help to evaluate our findings.

The original studies excluded patients with most potential confounding factors that might have influenced the clinical findings. However, we were not able to control for all possible confounders. For example, daily medication was allowed, but may have influenced the heart rate, in the case of a beta blocker and the blood pressure, with use of any antihypertensive medication. Also, pain may have influenced the vital signs, although heart rate does not correlate well with acute pain (Bossart et al., 2007). Arrhythmias may also have influenced heart rates, although most hypothermic patients remain in sinus rhythm (Vassallo et al., 1999). Finally, our study only included adults. Our results may not apply to children.

Conclusion

Our results may help to quantify the relationships between vital signs and measured hypothermic core temperatures. To our knowledge, our results are the first to provide concrete data on the vital signs to be expected for a given hypothermic core temperature. Our data may allow more accurate diagnosis and management of patients suffering from accidental hypothermia.

Footnotes

Authors' Contributions

M.P., V.R., P.-N.C., and V.F. conceived and designed the study. M.P., E.C., and V.R. drafted the article. M.P., E.C., and V.R. conducted the data analysis. M.P. and V.R. performed the statistical analysis. M.P., V.R., E.C., K.Z., P.-N.C., and V.F. critically revised the article. All authors have reviewed and approved the final article.

Acknowledgment

For the creation of the internet website: Dr. Alexandre Gnaegi, Switzerland.

Author Disclosure Statement

The authors declare that they have no competing interests.

Funding Information

Emergency Department, CHUV, Lausanne, Switzerland.

References

Bossart

, Fosnocht

, and Swanson

. (2007). Changes in heart rate do not correlate with changes in pain intensity in emergency department patients. J Emerg Med, 32:19–22.

Brown

, Brugger

, Boyd

, and Paal

. (2012). Accidental hypothermia. N Engl J Med, 367:1930–1938.

Danzl

. (2012). Accidental hypothermia. In: Wilderness Medicine, 6th ed. PS Auerbach, ed. Elsevier, Mosby, Philadelphia, PA. pp. 116–142.

Darocha

, Majkowski

, Sanak

, Podsiadlo

, Kosinski

, Salapa

, Mazur

, Zietkiewicz

, Galazkowski

, Krzych

, and Drwila

. (2017). Measuring core temperature using the proprietary application and thermo-smartphone adapter. J Clin Monit Comput, 31:1299–1304.

Darocha

, Sobczyk

, Kosinski

, Jarosz

, Galazkowski

, Nycz

, and Drwila

. (2015). Electrocardiographic changes caused by severe accidental hypothermia. J Cardiothorac Vasc Anesth, 29:e83–e86.

Deslarzes

, Rousson

, Yersin

, Durrer

, and Pasquier

. (2016). An evaluation of the Swiss staging model for hypothermia using case reports from the literature. Scand J Trauma Resusc Emerg Med, 24:16.

Dow

, Giesbrecht

, Danzl

, Brugger

, Sagalyn

, Walpoth

, Auerbach

, McIntosh

, Nemethy

, McDevitt

, Schoene

, Rodway

, Hackett

, Zafren

, Bennett

, and Grissom

. (2019). Wilderness Medical Society clinical practice guidelines for the out-of-hospital evaluation and treatment of accidental hypothermia: 2019 update. Wilderness Environ Med, 30:S47–S69.

Durrer

, Brugger

, and Syme

. (2003). International Commission for Mountain Emergency Medicine. The medical on-site treatment of hypothermia: ICAR-MEDCOM recommendation. High Alt Med Biol, 4:99–103.

Finnell

, and McMicken

. (2010). Alcohol-related disease. In: Rosen's Emergency Medicine: Concepts and Clinical Practice. 7th ed. RS Hockberger, JA Marx, and RM Walls, eds. Mosby, Elsevier, Philadelphia, PA. pp. 2375–2392.

10.

Freeman

, Deakin

, Nelson

, and Bootland

. (2018). Managing accidental hypothermia: a UK-wide survey of prehospital and search and rescue providers. Emerg Med J, 35:652–656.

11.

Hamilton

, Dresbach

, and Hamilton

. (1937). Cardiac changes during progressive hypothermia. Am J Physiol, 118:71–76.

12.

Henriksson

, Björnstig

, Saveman

, and Lundgren

. (2017). Protection against cold—a survey of available equipment in Swedish pre-hospital services. Acta Anaesthesiol Scand, 61:1354–1360.

13.

Karlsen

, Thomassen

, Vikenes

, and Brattebø

. (2013). Equipment to prevent, diagnose, and treat hypothermia: a survey of Norwegian pre-hospital services. Scand J Trauma Resusc Emerg Med, 12:63.

14.

Loppnow

, and Wilson

. (2015). Effects of ethanol on systemic hemodynamics in a porcine model of accidental hypothermia. Am J Emerg Med, 33:1414–1419.

15.

Matsuyama

, Morita

, Ehara

, Miyamae

, Okada

, Jo

, Sumida

, Okada

, Watanabe

, Nozawa

, Tsuruoka

, Fujimoto

, Okumura

, Kitamura

, and Ohta

. (2018). Characteristics and outcomes of accidental hypothermia in Japan: the J-Point registry. Emerg Med J, 35:659–666.

16.

Niven

, Gaudet

, Laupland

, Mrklas

, Roberts

, and Stelfox

. (2015). Accuracy of peripheral thermometers for estimating temperature: a systematic review and meta-analysis. Ann Intern Med, 163:768–777.

17.

Oung

, English

, Chiu

, and Hinchey

. (1992). Effects of hypothermia on hemodynamic responses to dopamine and dobutamine. J Trauma, 33:671–678.

18.

Paal

, Gordon

, Strapazzon

, Brodmann Maeder

, Putzer

, Walpoth

, Wanscher

, Brown

, Holzer

, Broessner

, and Brugger

. (2016). Accidental hypothermia-an update: the content of this review is endorsed by the International Commission for Mountain Emergency Medicine (ICAR MEDCOM). Scand J Trauma Resusc Emerg Med, 24:111.

19.

Pasquier

, Carron

, Rodrigues

, Dami

, Frochaux

, Sartori

, Deslarzes

, and Rousson

. (2019). An evaluation of the Swiss staging model for hypothermia using hospital cases and case reports from the literature. Scand J Trauma Resusc Emerg Med, 27:60.

20.

Pasquier

, Pasquier

, and Vallotton

. (2011). J wave in hypothermia. J Trauma, 71:e99.

21.

Pasquier

, Rousson

, Zen Ruffinen

, and Hugli

. (2012). Homemade thermometry instruments in the field. Wilderness Environ Med, 23:70–74.

22.

Podsiadło

, Darocha

, Kosiński

, Sałapa

, Ziętkiewicz

, Sanak

, Turner

, and Brugger

. (2017). Severe hypothermia management in mountain rescue: a survey study. High Alt Med Biol, 18:411–416.

23.

Prec

, Rosenman

, Braun

, Rodbard

, and Katz

. (1949). The cardiovascular effects of acutely induced hypothermia. J Clin Invest, 28:293–300.

24.

Rischall

, and Rowland-Fisher

. (2016). Evidence-based management of accidental hypothermia in the emergency department. Emerg Med Pract, 18:1–18.

25.

Rittenhouse

, Ito

, Mohri

, and Merendino

. (1971). Circulatory dynamics during surface-induced deep hypothermia and after cardiac arrest for one hour. J Thorac Cardiovasc Surg, 61:359–369.

26.

Rittenhouse

, Mori

, Dillard

, and Merendino

. (1974). Deep hypothermia in cardiovascular surgery. Ann Thorac Surg, 17:63–98.

27.

Safar

. (1981). Resuscitation after brain ischemia. In: Brain Failure and Resuscitation. A Grenvik and P Safar, eds. Churchill Livingstone, New. York.

28.

Strapazzon

, Procter

, Paal

, and Brugger

. (2014). Pre-hospital core temperature measurement in accidental and therapeutic hypothermia. High Alt Med Biol, 15:104–111.

29.

Strapazzon

, Procter

, Putzer

, Avancini

, Dal Cappello

, Überbacher

, Hofer

, Rainer

, Rammlmair

, and Brugger

. (2015). Influence of low ambient temperature on epitympanic temperature measurement: a prospective randomized clinical study. Scand J Trauma Resusc Emerg Med, 23:90.

30.

Truhlář

, Deakin

, Soar

, Khalifa

, Alfonzo

, Bierens

, Brattebø

, Brugger

, Dunning

, Hunyadi-Antičević

, Koster

, Lockey

, Lott

, Paal

, Perkins

, Sandroni

, Thies

, Zideman

, and Nolan JP; Cardiac Arrest in Special Circumstances Section

Collaborators

. (2015). European Resuscitation Council Guidelines for Resuscitation 2015: Section 4. Cardiac arrest in special circumstances. Resuscitation, 95:148–201.

31.

Vassallo

, Delaney

, Hoffman

, Slater

, and Goldfrank

. (1999). A prospective evaluation of the electrocardiographic manifestations of hypothermia. Acad Emerg Med, 6:1121–1126.

32.

Zafren

, Giesbrecht

, Danzl

, Brugger

, Sagalyn

, Walpoth

, Weiss

, Auerbach

, McIntosh

, Némethy

, McDevitt

, Dow

, Schoene

, Rodway

, Hackett

, Bennett

, and Grissom CK; Wilderness Medical

Society

. (2014). Wilderness Medical Society practice guidelines for the out-of-hospital evaluation and treatment of accidental hypothermia. Wilderness Environ Med, 25:425–445.