Handbook of Clinical Neurology, Vol.

157 (3rd series)

Thermoregulation: From Basic Neuroscience to Clinical Neurology, Part II
A.A. Romanovsky, Editor
https://doi.org/10.1016/B978-0-444-64074-1.00041-0
Copyright © 2018 Elsevier B.V. All rights reserved

Chapter 41

Consequences of perioperative hypothermia

KURT RUETZLER AND ANDREA KURZ*
Departments of General Anesthesiology and Outcomes Research, Anesthesiology Institute, Cleveland Clinic,
Cleveland, OH, United States

Abstract
Perioperative hypothermia is common, with an incidence ranging between 20 and 70%, and is defined
by a body core temperature below 36.0°C. Perioperative warming was rare during the previous century,
but was subsequently identified as a significant contributor to perioperative morbidity and mortality.
Perioperative hypothermia causes impaired pharmacodynamics, surgical site infections, blood loss and
coagulopathy, transfusion requirements, thermal discomfort, prolonged recovery, and prolonged duration
of hospitalization.
Measurement of central core temperature, maintaining normothermia, and consequent warming of
patients in the perioperative period are therefore essential. Several warming devices are commercially
available, including active skin warming as the most efficient, inexpensive, easy-to-use and mostly having
a good cost/benefit ratio for the majority of patients and surgeries.

Perioperative hypothermia is broadly defined as a body
core temperature of less than 36.0oC. Induction of anes-
thesia causes several physiologic changes. In particular,
the so-called redistribution effect within 60–90 minutes
of anesthesia causes a decrease in body core temperature.
The incidence of perioperative hypothermia has been
reported to range between 20 and 70% of patients having
surgery (Forstot, 1995; Hart et al., 2011).
Most of the thermoregulation studies were performed
in the 1990s, when patient warming was rare and scien-
tific evidence about the consequences of perioperative
hypothermia was weak. However, contemporary scien-
tific evidence is mostly based on these “old” studies.
One major limitation of several studies is that authors
previously defined perioperative normothermia as core
temperature
36°C. The difficulty is that 36°C is never
a normal temperature in humans. Even at the circadian
nadir, usually about 3.00 a.m., core temperature is not
normally below 36.5°C; and at about 3.00 p.m., core
temperature is typically about 37.5°C (Sessler et al.,
1991). On average, then, normal body temperature in
humans is about 37°C, not the 36°C that is widely accepted
as suitable for perioperative patients (Castren et al., 2009;
Scott et al., 2015). At any given time of day, core temp-
erature is actively controlled by the thermoregulatory
system (Tayefeh et al., 1998). In other words, circadian
temperature changes are not passive responses to envi-
ronmental perturbations – which suggests that humans
function best at temperatures near 37°C.
PERIOPERATIVE HYPOTHERMIA
IS COMMON AFTER INDUCTION OF
ANESTHESIA, BUT NOT AT THE END
OF SURGERY
Figure 41.1 displays the distribution of core temperature
as a function of time after induction of anesthesia in a
cohort of 58,814 patients having noncardiac surgery at
the Cleveland Clinic (Sun et al., 2015). During the first
hour of anesthesia, core temperature typically drops
about 1°C, resulting in a median core temperature of
about 35.8°C. The rapid drop of core temperature after
induction of anesthesia is caused by balancing the
temperature between the “cold” peripheral thermal com-
partment and the “warm” core thermal compartment.
This period is therefore a shift of heat from the warm

*Correspondence to: Andrea Kurz, Anesthesiology Institute, Cleveland Clinic, 9500 Euclid Avenue, Cleveland OH 44195,
United States. E-mail: ak@or.org
688 K. RUETZLER AND A. KURZ
38
Core
60 Temperature
Core Temperature (°C)

Core
Temperature Threshold
37 Quantile <36.0°C
<35.5°C
Top 5% 50

Incidence of Hypothermia
<35.0°C
Top 10%
Q3
36
Median
Q1
40
Bottom 10%
35 Bottom 5%
30

0 1 2 3 4 5 6
Time After Induction (h) 20
Sample Size

60k
30k
10
0k

Fig. 41.1. Distribution of core temperature as a function of

0
time after induction. (Reproduced from Sun Z, Honar H,
Sessler DI, et al. (2015) Intraoperative core temperature pat- 0 1 2 3 4 5 6
terns, transfusion requirement, and hospital duration in patients Time After Induction (h)
warmed with forced air. Anesthesiology 122: 276-285.) Fig. 41.2. Incidence of hypothermia as a function of time after
induction, under progressive core temperature thresholds defin-
core to the cold peripheral thermal compartment and is ing hypothermia. (Reproduced from Sun Z, Honar H, Sessler
DI, et al. (2015) Intraoperative core temperature patterns, trans-
accelerated by the vasodilatation caused by anesthesia.
fusion requirement, and hospital duration in patients warmed
Therefore, this period is called redistribution. This des- with forced air. Anesthesiology 122: 276-285.)
ignation also makes clear that the actual amount of heat
is not decreased (or at least, minimally) during this
period; the overall amount of heat is just distributed
elsewhere (Chapter 37). 60
However, core temperature increased after the initial
Percent of Patients

Episode
drop (as patients were actively warmed!), resulting in Duration
normothermia in the overwhelming majority of all 40 >15 minutes
>30 minutes
patients at the end of surgery. Generally, perioperative >60 minutes
>90 minutes
hypothermia is common in the first hours after induction >120 minutes
of anesthesia, and is it is therefore clinically essential to 20 >180 minutes
>240 minutes
counteract this.
Figure 41.2 displays the incidence of hypothermia
0
under various core temperature thresholds. After
34.0 34.5 35.0 35.5 36.0
45 minutes of induction of general anesthesia, almost Threshold Defining Hypothermia
two-thirds of patients reached the hypothermia threshold
Fig. 41.3. Incidence of (any) hypothermic episodes during the
of 36°C. In addition, more than one-quarter reached
case, according to progressive core temperature thresholds
the threshold of 35.5°C within 51 minutes of induction. defining hypothermia. (Reproduced from Sun Z, Honar H,
After 71 minutes, 7% of all patients even reached 35°C. Sessler DI, et al. (2015) Intraoperative core temperature pat-
Although perioperative hypothermia is most common terns, transfusion requirement, and hospital duration in patients
in the first hour after induction of anesthesia, it neverthe- warmed with forced air. Anesthesiology 122: 276-285.)
less remains common in about 20% of all patients, even
after 6 hours of surgery.
The duration of perioperative hypothermia varies. As a consequence, perioperative hypothermia signif-
Nearly half of all patients have core temperature < 36°C, icantly contributes to perioperative morbidity and mor-
and about 20 % of all patients have a core temperature tality (Chapter 33). Specifically, several randomized
< 35.5°C for more than 1 hour (Fig. 41.3). trials have shown that mild perioperative hypothermia
Most cellular functions and enzymatic activities causes impaired pharmacodynamics, surgical site infec-
are temperature-dependent. Hypothermia also provokes tions (SSIs), blood loss and coagulopathy, transfusion
systemic responses, some of which are potentially harm- requirements, thermal discomfort, prolonged recovery,
ful for the hypothermic patient. and prolonged duration of hospitalization.
 CONSEQUENCES OF PERIOPERATIVE HYPOTHERMIA
IMPAIRED PHARMACODYNAMICS
Most cellular functions and enzyme activity are
temperature-dependent. Therefore, it is unsurprising
that even mild hypothermia prolongs the actions of var-
ious drugs. By affecting drug metabolism, perioperative
hypothermia is associated with delayed emergence from
anesthesia and prolonged stay in the postanesthesia care
unit (PACU).
Propofol and fentanyl
Propofol is one of the most commonly used intravenous
anesthetics. A 3oC decrease in core temperature results
in an approximately 28% increase in plasma propofol
concentration, largely as a result of reduced hepatic
blood flow. However, mild hypothermia does not reduce
propofol requirement during craniotomy surgery (Leslie
et al., 2002).
One study reported that the steady-state plasma
concentration of fentanyl is increased by 5%/ 1oC drop
of core temperature (Fritz et al., 2005).
Muscle relaxants
Nondepolarizing muscle relaxants are markedly affected
by perioperative hypothermia, resulting in a prolonged
duration of action. This is based on a wide range of
reasons, including changes in volume of distribution,
altered local diffusion receptor affinity, changes in
pH at the neuromuscular junction, and the net effect of
cooling on the various components of neuromuscular
transmission.
The duration of action of vecuronium is more than dou-
bled in patients experiencing perioperative hypothermia
(Heier et al., 1991). Interestingly, when neostigmine is
used, as an antagonist of vecuronium, it does not appear
to be altered by mild hypothermia (Heier et al., 2002).
Perioperative hypothermia prolongs the duration of
action of atracurium by approximately 60% in patients with
a 3oC decrease of core temperature (Leslie et al., 1995),
although atracurium appears not to be temperature-
sensitive.
Surprisingly, there is no difference in the recovery
index (time for 25–75% twitch recovery) for both atra-
curium and vecuronium during normo- and hypothermia.
The duration of action of another muscle relaxant,
rocuronium, was reported to be prolonged in car-
diac patients undergoing hypothermic cardiac bypass
(Smeulers et al., 1995).
Volatile anesthetics
Anesthesia potency is driven by the steady-state
plasma partial pressure opposed to the actual anesthetic
689
concentration in the cells. Hypothermia increases the
solubility of volatile anesthetics but does not appear
to alter the potency (Reynolds et al., 2008).
However, with volatile anesthetics, the minimum
alveolar concentration (MAC) is diminished, and thus
a lower amount of agent is needed to prevent a reac-
tion due to surgical stimulus. The MAC of halothane
and isoflurane decreases by approximately 5%/1oC as
core temperature decreases (Eger and Johnson, 1987).
Simultaneously, a decrease of 5.1%/1oC in isoflurane
MAC is observed for every 1oC reduction of core temp-
erature in children (Liu et al., 2001).
SURGICAL SITE INFECTIONS AND
COMPLICATIONS (WOUND HEALING
AND DEHISCENCE)
SSIs are the third leading cause of nosocomial infections,
accounting for 14–16% of all hospital-acquired infec-
tions, and are the leading cause of nosocomial infections
among surgical patients (Esnaola and Cole, 2011).
Considering all of the complications that may be trig-
gered by anesthesia, SSIs are likely the most significant
cause of morbidity, even greater than all other anesthetic
complications combined.
Perioperative hypothermia affects host defense
against contamination via at least three mechanisms.
First, perioperative hypothermia triggers postoperative
vasoconstriction with the goal of constraining meta-
bolic heat to the core and speeds rewarming. On the
other hand, vasoconstriction also reduces perfusion of
wounded tissues. As a consequence, tissue oxygen par-
tial pressure drops (even when blood is fully saturated),
which is especially important, as oxidative killing by
neutrophils, the primary defense against bacterial con-
tamination, mostly relies on molecular oxygen (Edwards
et al., 1984; Allen et al., 1997).
Second, perioperative hypothermia reduces systemic
immune activation, including T-cell-mediated antibody
production, and decreases motility of key cells, including
macrophages (Jonsson et al., 1988).
Third, vasoconstriction-induced tissue hypoxia impairs
tissue healing and protein metabolism, which are impor-
tant to prevent wound dehiscence and recontamination
(Carli et al., 1989).
Perioperative hypothermia is associated with a signif-
icant increase of infections in patients undergoing chole-
cystectomy (Hart et al., 2011; Madrid et al., 2016). In
a major study by Kurz et al. (1996) investigating patients
having colon resection, the incidence of SSIs was
threefold greater in hypothermic versus normothermic
patients (Fig. 41.4; Table 41.1). As a consequence,
hospitalization was prolonged in hypothermic patients.
690
SSIs increase postoperative hospitalization by an aver-
age of 4 days, resulting in an increased attributable cost
of US$8000–25,000 for each patient (Hart et al., 2011).
A recent review summarized that there was a sig-
nificant benefit of actively warmed patients over nonac-
tively warmed patients in the incidence of SSIs and
complications (risk ratio 0.36, 95% confidence interval
(CI) 0.2–0.66, p ¼ 0.0008) (Madrid et al., 2016).
BLOOD LOSS AND COAGULOPATHY
Perioperative hypothermia causes coagulopathy and
consequently increased blood loss by several routes.
25
P = 0.001
P < 0.01
20
K. RUETZLER AND A. KURZ
Platelets
Although platelet counts are stable and are not affected
by temperature, perioperative hypothermia affects plate-
let function. It is commonly suggested that inhibition of
intrinsic platelet function is not the cause of the coagu-
lopathy. In fact, it mostly results from the reversible
impairment of platelet aggregation via reduced release
of thromboxane A3, which is necessary for the formation
of an initial platelet plug (Valeri et al., 1992). Furthermore,
perioperative hypothermia has been reported to induce
morphologic changes of the platelets, and enhances the
binding of platelets to fibrinogen through activation of
GbIIb-IIa receptors.
25
Hospital Duration (days)
20

Enzymes of the coagulation cascade

Infections (%)

15
10
5 5
0 0
Hypothermic Normothermic Hypothermic Normothermic
34.7 ± 0.6 36.6 ± 0.5 34.7 ± 0.6 36.6 ± 0.5
Core Temperature (°C)
Fig. 41.4. Hypothermia, wound infection and duration of
hospitalization. (Reproduced with permission from Kurz A,
Sessler DI, Lenhardt R (1996) Perioperative normothermia to
reduce the incidence of surgical-wound infection and shorten
hospitalization. Study of Wound Infection and Temperature
Group. N Engl J Med 334: 1209–1215. Copyright ©
Massachusetts Medical Society.)
15
In addition, perioperative hypothermia impairs the func-
10 tion of several enzymes of the coagulation cascade and
finally reduces clot formation. This is especially impor-
tant, as this is frequently missed in the clinical setting,
as coagulation laboratory studies including prothrombin
time and partial thromboplastin time are usually per-
formed at a temperature of 37oC, rather than the patient’s
actual temperature (Rohrer and Natale, 1992). The
identical testing will be prolonged, if the testing is
done at the patient’s actual temperature, instead of at
37oC. Several in vitro studies reported prolongation
of up to 10% in hypothermic versus normothermic
tests. However, whether this prolongation is clinically
relevant remains questionable.

Table 41.1
Postoperative findings in the two study groupsa

Normothermia Hypothermia
Variable (n ¼ 104) (n ¼ 96) p-value

All patients
Infection: number of patients (%) 6 (6) 18 (19) 0.009
ASEPSIS score 7  10 13  16 0.002
Collagen deposition (mg/cm) 328  135 254  114 0.04
Days to first solid food 5.6  2.5 6.5  2.0 0.006
Days to suture removal 9.8  2.9 10.9  1.9 0.002
Days of hospitalization 12.1  4.4 14.7  6.5 0.001
Uninfected patients
Number of patients 98 78
Days to first solid food 5.2  1.6 6.1  1.6 <0.001
Days to suture removal 9.6  2.6 10.6  1.6 0.003
Days of hospitalization 11.8  4.1 13.5  4.5 0.01

Reproduced with permission from Kurz A, Sessler DI, Lenhardt R (1996) Perioperative normothermia to reduce the incidence of surgical-wound
infection and shorten hospitalization. Study of Wound Infection and Temperature Group. N Engl J Med 334: 1209–1215. Copyright © Massachu-
setts Medical Society.
a
Plus and minus values are means  standard deviation.
 CONSEQUENCES OF PERIOPERATIVE HYPOTHERMIA
Fibrinolysis
Generally, the fibrinolytic system regulates the balance
between the formation of hemostatic plugs and restora-
tion of blood flow after clot formation (Reynolds et al.,
2008). Excessive fibrinolysis predisposes patients to
hemorrhage, while inadequate fibrinolysis predisposes
patients to thrombosis.
The most important structural element in formed
clots is fibrin, but it is subject to degradation by plasmin,
which represents the activated from of plasminogen.
Therefore, activation of plasminogen and conversion
to plasmin represent the core steps of the fibrinolytic
system and are mostly mediated by tissue-type plasmin-
ogen. Findings of several studies suggest that fibrinolysis
is not affected by mild perioperative hypothermia,
while hyperthermia intensively affects fibrinolysis.
In conclusion, this might result in the assumption that
hypothermia-induced coagulopathy does not result from
excessive clot lysis. Data from thromboelastograms
suggest that hypothermia impairs clot formation, rather
than facilitating clot degeneration.
691
Although there are many divergent results from
several retrospective studies, two randomized trials have
confirmed that a reduction of only approximately 0.5oC
core temperature increases blood loss by 200–300 mL in
patients undergoing hip arthroplasty (Winkler et al., 2000).
A comprehensive meta-analysis of randomized
controlled trials by Rajagopalan et al. (2008) compared
normothermic patients with patients experiencing peri-
operative hypothermia (34–36oC). Fourteen studies were
included in order to assess the relationship between
hypothermia and blood loss: mild hypothermia increased
blood loss by approximately 16% (Table 41.2).
TRANSFUSION REQUIREMENT
There is no linear relationship between blood loss and
transfusion requirements. It seems to be obvious that
increased blood loss increases the individual need for
transfusion requirements. Blood transfusion was expected
to be safe for decades (taking into account the risk of
human immunodeficiency virus (HIV) and hepatitis C),
but blood transfusions are more toxic than previously

Table 41.2
Total blood loss meta-analysis and forest plot

Sample Normothermic Hypothermic Outcome

Size (N) (N) (H) (N/H)
Study N:H mean (sd) mean (sd) mean (95% Cl)

Schmied 30 : 30 1670 (320) 2150 (550) 0.79 (0.70, 0.88)

Winkler 75 : 75 1531(1055, 1746) 1678(1366, 1965) 0.90 (0.82, 1.00)
Widman 22 : 24 923 (410) 1068 (482) 0.87 (0.68, 1.11)
Persson 29 : 30 186 (145) 308 (257) 0.62 (0.43, 0.89)
Hofer 29 : 29 1497 (497) 2300 (788) 0.65 (0.55, 0.77)
Bock 20 : 20 635 (507) 1070 (803) 0.58 (0.38, 0.89)
Johansson 25 : 25 1047 (413) 1066 (441) 0.99 (0.80, 1.23)
Smith 31 : 30 423 (562) 159 (268) 3.14 (1.82, 5.42)
Frank 142 : 158 390 (834) 520 (754) 0.56 (0.43, 0.73)
Mason 32 : 32 111 (40) 157 (73) 0.73 (0.60, 0.89)
Casati 25 : 25 470 (170) 442 (216) 1.11 (0.89, 1.40)
Murat 26 : 25 160 (61) 161 (100) 1.09 (0.84, 1.43)
Hohn 43 : 73 660(230, 1870) 956(340, 5480) 0.69 (0.36, 1.34)
Nathan 73 : 71 569 (356) 666 (405) 0.85 (0.70, 1.02)

Summary 0.84 (0.74, 0.96)

Treatment effect P = 0.009

0.4 1.0 2.0 3.0 4.0 5.0 6.0

Favors Favors
Normothermic Hypothermic

From Rajagopalan S, Mascha E, Na J, et al. (2008) The effects of mild perioperative hypothermia on blood loss and transfusion requirement. Anes-
thesiology 108: 71–77, with permission from Wolters Kluwer Health.
Treatment effect is expressed as ratio of geometric means of blood loss for normothermic (N) versus hypothermic (H) patients. Results indicate an
estimated 16% (95% confidence interval (CI) 4%, 26%) lower average blood loss in normothermic versus hypothermic patients, p < 0.009.
692 K. RUETZLER AND A. KURZ
Table 41.3
Transfusion meta-analysis and forest plot

Study Normothermic Hypothermic Outcome

n/N (%) n/N (%) RR (95%Cl)

Schmied 1/30 (3%) 7/30 (23%) 0.14 (0.02, 1.09)

Winkler 29/75 (39%) 40/75 (53%) 0.73 (0.51, 1.03)
Widman 9/22 (41%) 11/24 (46%) 0.89 (0.46, 1.73)
Hofer 5/29 (17%) 11/29 (38%) 0.45 (0.18, 1.14)
Johansson 15/25 (60%) 13/25 (52%) 1.15 (0.7, 1.89)
Kurz 23/104 (22%) 34/96 (35%) 0.62 (0.4, 0.98)
Bock 3/20 (15%) 9/20 (45%) 0.33 (0.11, 1.05)
Hohn 17/43 (40%) 18/43 (42%) 0.94 (0.57, 1.57)
Nathan 23/73 (32%) 24/71 (34%) 0.93 (0.58, 1.49)
Smith 2/31 (6%) 1/30 (3%) 1.94 (0.19, 20.24)

Summary 0.78 (0.63, 0.97)

Treatment effect P = 0.027

0.1 1.0 2.0 3.0 4.0

Favors Favors
Normothermic Hypothermic

Reproduced from Rajagopalan S, Mascha E, Na J, et al. (2008) The effects of mild perioperative hypothermia on blood loss and transfusion require-
ment. Anesthesiology 108: 71–77.
Treatment effect is expressed as the relative risk (RR) of transfusion in normothermic versus hypothermic patients. Normothermia is associated with
22% less risk of transfusion than hypothermia (95% confidence interval (CI) 3%, 37%), p ¼ 0.027. n, number transfused; N. number of patients.

believed. Data from the above-mentionedmeta-analysis

by Rajagopalan et al. (2008) reported that mild hypother- Table 41.4
mia significantly increased the relative risk for transfusion Hypothermia and transfusion requirement
by approximately 22% (CI 3–37%) (Table 41.3).
In a retrospective study performed at the Cleveland Adjusteda odds ratio Adjusteda ratio of
Clinic, about 4.6% of all patients received a blood trans- (pointwise 95% CI) geometric mean
fusion (Sun et al., 2015). Based on a multivariate logistic Area under for intraoperative duration of
regression model, the authors found a significant asso- 37°C erythrocyte hospitalization
ciation between core temperature <37oC and risk of hav- (degree transfusion (pointwise 95% CI)
hours) (n ¼ 45,866) (n ¼ 39,180)
ing a transfusion. The odds ratios relative to a reference
value of 1oC are presented in Table 41.4. 0.25 1.34 (1.04, 1.73) 0.96 (0.90, 1.03)
Generally, risk of transfusion significantly increases 0.50 1.10 (0.98, 1.24) 0.97 (0.93, 1.01)
as a function of duration and dimension of perioperative 1.00 (Reference) (Reference)
hypothermia. 2.00 1.00 (0.89, 1.13) 1.03 (1.00, 1.05)
Figure 41.5 displays the adjusted probability of trans- 4.00 1.12 (0.91, 1.39) 1.06 (1.03, 1.09)
fusion estimates versus integrated area above the core 8.00 1.41 (1.08, 1.84) 1.05 (1.01, 1.10)
temperature versus time curve and <37oC. The shaded 16.00 2.02 (1.30, 3.14) 0.96 (0.86, 1.08)
regions represent pointwise 95% CI.
Reproduced from Sun Z, Honar H, Sessler DI, et al. (2015) Intraoperative
core temperature patterns, transfusion requirement, and hospital dura-
SHIVERING tion in patients warmed with forced air. Anesthesiology 122: 276–285.
a
Estimates adjusted for year, type, and duration of surgery, body mass
Shivering is an autonomic thermoregulatory response,
index, age, preoperative platelet count, preoperative hemoglobin,
defined as rhythmic involuntary muscular activity. estimated blood loss, and individual anesthesiologist, as well as the
The ultimate goal of shivering is to produce heat in order Elixhauser comorbidities, which are listed in Table 41.2.
to restore normothermia. CI, confidence interval.
 CONSEQUENCES OF PERIOPERATIVE HYPOTHERMIA
35%
30%
Adjusted Pr(Transfusion)*
25%
20%
15%
10%
5%
0%
1/4 1/2 1 2 4 8
Area Under 37°C (degree⋅hours)
Fig. 41.5. Hypothermia and transfusion requirement. LOS,
length of stay. (Reproduced from Sun Z, Honar H, Sessler
DI, et al. (2015) Intraoperative core temperature patterns,
transfusion requirement, and hospital duration in patients
warmed with forced air. Anesthesiology 122: 276-285.)
Intraoperative shivering is rare, as many patients
receive muscle relaxants and the shivering threshold is
significantly reduced by anesthetic agents and only
patients having serious hypothermia hit this threshold.
In contrast, postoperative shivering in hypother-
mic patients is common. Shivering is largely thermo-
regulatory, in order to produce heat and compensate for
hypothermia, but is aggravated by volatile anesthetics.
However, thermoregulatory major muscular activity must
be distinguished from low-intensity shivering-like mus-
cular activity, which is common in patients in pain
(Horn et al., 1999). On average, postoperative shivering
increases oxygen consumption by approximately 40%
(Frank et al., 1995a). It appears to be logical that this
increase of oxygen consumption is associated with peri-
operative morbidity and mortality. Shivering is extremely
uncomfortable for patients and is a preventable complica-
tion of perioperative hypothermia. However, shivering
can be treated effectively by a variety of medications,
including pethidine, clonidine, dexmedetomidine, and
ketamine. Lowering the shivering threshold is the
assumed mechanism.
THERMAL DISCOMFORT AND ANXIETY
Thermal discomfort is a predictable and common conse-
quence of patients experiencing perioperative hypother-
mia. Even mild hypothermia produces marked thermal
discomfort and consequently affects patient satisfaction
(Reynolds et al., 2008). Patients often identify having
693
experienced cold during the postoperative period as the
worst part of their hospitalization, sometimes even worse
than surgical pain. Thermal discomfort is also stressful
for patients, and has been reported to lead to higher
blood pressure, heart rate, and plasma catecholamine
concentrations (Frank et al., 1995b). Active warming
improves thermal comfort in hypothermic patients, but
prevention of postoperative hypothermia is obviously
the preferable management strategy.
CARDIAC COMPLICATIONS
Among the major complications of hypothermia, myo-
cardial outcomes are least well established. An observa-
tional analysis suggests an association between Surgical
Care Improvement Project-10 (SCIP-10) compliance and
16
temperature and cardiovascular outcomes (Scott et al.,
2015). A single-center randomized trial of 300 patients
evaluated cardiac outcomes, although the study was seri-
ously underpowered (Frank et al., 1997). For example,
only two hypothermic patients had a cardiac arrest
and one had a myocardial infarction versus none in the
normothermic group (36.7°C). Ten hypothermic patients
experienced cardiovascular events versus two normother-
mic patients. These fragile results, based on fewer than
10 outcome events, are nearly as likely to be wrong as
to be right (Ioannidis, 2005), and are a poor basis for health
policy. The other relevant study had only 100 patients
and was thus even more fragile (Elmore et al., 1998).
An additional limitation of the study by Frank et al.
(1997) is that diagnosis was primarily based on Holter
electrocardiogram finding rather than troponin concentra-
tions, which are much more sensitive. Consequently, the
overall myocardial infarction rate was <1%, whereas the
true rate is no less than 10% in vascular surgery patients.
It is highly questionable whether any conclusion about
hypothermia and myocardial outcomes can be derived
from a study that missed 90% of the presumed myocardial
events. The only other relevant trial was restricted to 100
patients, leaving it even less powered to detect clinically
important outcomes (Elmore et al., 1998). There thus
remains considerable doubt as to the true effect of
moderate hypothermia on cardiovascular outcomes.
DELAYED DISCHARGE FROM THE PACU
It may not be not very surprising that perioperative hypo-
thermia is associated with delayed discharge from the
PACU. In a mayor randomized trial, Lenhardt et al.
(1997) assigned patients undergoing elective surgery to
normothermia or hypothermia during anesthesia. The
authors assessed length of stay in the PACU until a sus-
tained score of
13 was reached. Ability to discharge was
based on modification of the Aldrete and Kroulik (1970)
scoring system. The score is based on activity, ventilation,
694 K. RUETZLER AND A. KURZ
consciousness, and hemodynamic responses; 0, 1, or 2
points were assigned for each of eight responses. In this
study, hypothermic patients required about 40 minutes
longer than normothermic patients to reach a sustained
score
13, the defined criterion for discharge. Core
temperature in the hypothermic patients required 134
 60 minutes to reach 36°C. Consequently, hypothermic
patients required about 90 minutes longer than normother-
mic patients to reach both a sustained score
13 and a core
temperature >36°C (p < 0.001) (Figs 41.6 and 41.7)
(Lenhardt et al., 1997).
PROLONGED HOSPITALIZATION
Two hundred patients having elective colorectal resec-
tion for cancer or inflammatory bowel disease were
randomly assigned to normothermia (37  0.3°C) vs.
hypothermia (34.4  0.4°C) in a major landmark study
by Kurz et al. (1996). Hypothermia caused a threefold
Fig. 41.6. Kaplan–Meier “survival” analysis showing the
percentage of patients not sustaining a recovery score
13.
The probability value, using a Wilcoxon analysis, was < 0.0001.
(Reproduced from Lenhardt R, Marker E, Goll V, et al. (1997)
Mild intraoperative hypothermia prolongs postanesthetic
recovery. Anesthesiology 87: 1318–1323.)
Fig. 41.7. Kaplan–Meier “survival” analysis showing the
percentage of patients not sustaining a recovery score
13 and
a core temperature
36°C. The probability value, using a Wil-
coxon analysis, was 0.0001.34 (Reproduced from Lenhardt
R, Marker E, Goll V, et al. (1997) Mild intraoperative hypo-
thermia prolongs postanesthetic recovery. Anesthesiology 87:
1318–1323.)
higher risk of experiencing an SSI. It seems to be obvious
that corresponding hospitalization in patients having an
SSI is prolonged. Interestingly, the important effect of
prolongation of hospitalization was present in both, all
(12.1  4.4 vs. 14.7  6.5 days), and in infected (11.8
 4.1 vs. 13.5  4.5 days) patients. Therefore, perioper-
ative hypothermia per se (as well as the higher risk of
having an SSI) prolongs hospitalization (Table 41.1).
As is consistent with a delay in clinical healing,
sutures were removed significantly later and the deposi-
tion of collagen (an index of scar formation and the
strength of the healing wound) was significantly less in
the hypothermia group than in the normothermia group.
That the patients assigned to hypothermia required
significantly more time before they could tolerate solid
food is also consistent with impaired healing.
In another study consisting of 39,180 patients, Sun
et al. (2015) reported a significant association between
area <37oC core temperature threshold and geometric
mean duration of hospitalization (Fig. 41.8).
Consequently, prolongation of hospitalization seems
to be essential, especially as healthcare-related costs tend
to rise sharply. Even extremely cautious assessments of
costs per in-hospital day of US$1000 represent increased
overall costs of about US$2000 in every single patient
experiencing perioperative hypothermia.
3.0
Adjusted Geometric Mean LOS*
2.8
2.6
2.4
2.2
2.0
1/4 1/2 1 2 4 8 16
Area Under 37°C (degree⋅hours)
Fig. 41.8. Estimates of geometric mean duration of hospital-
ization in days vs. integrated area above the core temperature
vs. time curve and <37oC core temperature. The shaded
regions represent pointwise 95% confidence interval. LOS,
length of stay. (Reproduced from Sun Z, Honar H, Sessler
DI, et al. (2015) Intraoperative core temperature patterns,
transfusion requirement, and hospital duration in patients
warmed with forced air. Anesthesiology 122: 276-285.)
 CONSEQUENCES OF PERIOPERATIVE HYPOTHERMIA
STRATEGIES FOR MAINTAINING
NORMOTHERMIA DURING
ANESTHESIA
Measurement of central core temperature
It is essential to focus on central core temperature, as the
core thermal component is usually relatively well and
homogeneously perfused. In contrast, the peripheral
thermal component is less perfused and consequently
relatively hypothermic compared to the central thermal
component. Therefore, measuring the central core tem-
perature is essential and rewarming should be guided
by central core temperature alone. The central core
temperature can be reliably measured at four sites:
1. pulmonary artery (using a Swan catheter)
2. distal esophagus
3. nasopharynx with the probe inserted 10–20 cm
4. tympanic membrane, if measured with a contact
thermistor or thermocouple.
The temperatures measured at these four sites do not
vary more than a couple of tenths of 1°C and therefore
even a single measurement gives a reliable estimation
of the actual core temperature.
Prewarming
Core temperature drops during the initial hour of anesthe-
sia based on the principle of redistribution (see above).
The redistribution effect might be partially attenuated
by prewarming patients. It is important to understand that
prewarming does not increase patients’ core temperature,
but increases the temperature of the peripheral tissues and
therefore decreases the amount of redistribution hypother-
mia. Consequently, prewarming alone does not prevent
perioperative hypothermia, but it does reduce the drop
in temperature during the initial period of anesthesia. Typ-
ically, the core temperature in prewarmed patients stays
about 0.4°C warmer compared to nonprewarmed patients.
Perioperative warming devices
Various perioperative warming devices are available and
can be divided into passive insulation and active warm-
ing devices of the skin surface, fluids, inspired and peri-
toneal gases, and endovascular heat exchangers.
1. Passive insulation: a single layer of passive insulation
reduces cutaneous heat loss by 30%. However, it is
essential to determine that the layer does not actively
transfer heat into the body, even by adding several
extra layers of insulation (Sessler and Schroeder,
1993). Consequently, passive insulation can decrease
heat loss but will not add any benefit in the mainte-
nance of perioperative normothermia (Sessler, 2016).
695
2. Active skin warming: Convective warming of the
skin with forced air is by far the most frequently used
and effective perioperative warming device. Forced-
air warming is especially attractive, as it is easy to
use, inexpensive, and for most patients and surgery
has a good cost/benefit ratio. In order to achieve the
highest effectiveness, the forced-air blankets should
be directly placed on the patient’s skin and should
cover as much of the body as possible. An alternative
approach is conductive warming using resistive
heating or circulating water (Ruetzler et al., 2011;
Hasegawa et al., 2012).
(a) Fluid warming is not efficient, as the tempera-
ture of the infusion (usually not warmer than
38–39°C) can only slightly exceed core temper-
ature. Consequently, active warming of the
patient is basically impossible unless very large
amounts of fluid are given over a short period of
time. The most important effect of fluid warm-
ing is decreasing the amount of heat loss. For
example, 1 liter of crystalloid given at room
temperature decreases core temperature by
0.25°C. Equally 1 unit of blood from the refrig-
erator decreases core temperature by 0.25°C.
(b) Warming inspired and peritoneal gases: warm-
ing patients by warming inhaled air is insuffi-
cient, as the heat capacity of air is low. Only a
marginal amount of heat is lost via the respira-
tory system. Thus warmed air is unable to trans-
fer a considerable amount of heat to patients
(Sammour et al., 2010).
(c) Vascular heat exchange catheters transfer much
more heat compared to skin-warming devices.
On the other hand, heat exchange catheters
are quite expensive, and placement is invasive.
Thus, their use is mostly restricted to patients
requiring a rapid onset of hypothermia, for
example after distinct accidental hypothermia
(Sessler, 2016).
SUMMARY
Perioperative hypothermia is common, and significantly
affects perioperative morbidity and mortality. In particu-
lar, perioperative hypothermia causes impaired pharmaco-
dynamics, SSIs, blood loss and coagulopathy, transfusion
requirements, thermal discomfort, prolonged recovery,
and prolonged duration of hospitalization.
Maintenance or restoration of normothermia is essen-
tial in all surgical patients and can effectively be achieved
by consequent (re-)warming during surgery. Core temper-
ature should be measured in all patients having general or
regional anesthesia lasting more than 30 minutes.
696 K. RUETZLER AND A. KURZ
Important points
● Perioperative hypothermia is common.
● Perioperative hypothermia is a significant contri-
butor to morbidity and mortality.
● Maintenance and restoration of normothermia are
essential.
● Monitoring the core temperature during surgery and
consequent warming of the patients are imperative.
REFERENCES
Aldrete JA, Kroulik D (1970). A postanesthetic recovery
score. Anesth Analg 49: 924–934.
Allen DB, Maguire JJ, Mahdavian M et al. (1997). Wound
hypoxia and acidosis limit neutrophil bacterial killing
mechanisms. Arch Surg 132: 991–996.
Carli F, Emery PW, Freemantle CA (1989). Effect of peropera-
tive normothermia on postoperative protein metabolism in
elderly patients undergoing hip arthroplasty. Br J Anaesth
63: 276–282.
Castren M, Silfvast T, Rubertsson S et al. (2009). Scandinavian
clinical practice guidelines for therapeutic hypothermia
and post-resuscitation care after cardiac arrest. Acta
Anaesthesiol Scand 53: 280–288.
Edwards SW, Hallett MB, Campbell AK (1984). Oxygen-
radical production during inflammation may be limited
by oxygen concentration. Biochem J 217: 851–854.
Eger 2nd EI, Johnson BH (1987). MAC of I-653 in rats, includ-
ing a test of the effect of body temperature and anesthetic
duration. Anesth Analg 66: 974–976.
Elmore JR, Franklin DP, Youkey JR et al. (1998).
Normothermia is protective during infrarenal aortic surgery.
J Vasc Surg 28: 984–992.
Esnaola NF, Cole DJ (2011). Perioperative normothermia
during major surgery: is it important? Adv Surg 45:
249–263.
Forstot RM (1995). The etiology and management of inadver-
tent perioperative hypothermia. J Clin Anesth 7: 657–674.
Frank SM, Fleisher LA, Olson KF et al. (1995a). Multivariate
determinants of early postoperative oxygen consumption in
elderly patients. Effects of shivering, body temperature,
and gender. Anesthesiology 83: 241–249.
Frank SM, Higgins MS, Breslow MJ et al. (1995b). The cate-
cholamine, cortisol, and hemodynamic responses to mild
perioperative hypothermia. A randomized clinical trial.
Anesthesiology 82: 83–93.
Frank SM, Fleisher LA, Breslow MJ et al. (1997).
Perioperative maintenance of normothermia reduces the
incidence of morbid cardiac events: a randomized clinical
trial. JAMA 277: 1127–1134.
Fritz HG, Holzmayr M, Walter B et al. (2005). The effect of
mild hypothermia on plasma fentanyl concentration and
biotransformation in juvenile pigs. Anesth Analg 100:
996–1002.
Hart SR, Bordes B, Hart J et al. (2011). Unintended perioper-
ative hypothermia. Ochsner J 11: 259–270.
Hasegawa K, Negishi C, Nakagawa F et al. (2012). Core
temperatures during major abdominal surgery in patients
warmed with new circulating-water garment, forced-air
warming, or carbon-fiber resistive-heating system. Journal
of Anesthesia 26: 168–173.
Heier T, Caldwell JE, Sessler DI et al. (1991). Mild intraopera-
tive hypothermia increases duration of action and sponta-
neous recovery of vecuronium blockade during nitrous
oxide-isoflurane anesthesia in humans. Anesthesiology
74: 815–819.
Heier T, Clough D, Wright PM et al. (2002). The influence of
mild hypothermia on the pharmacokinetics and time course
of action of neostigmine in anesthetized volunteers.
Anesthesiology 97: 90–95.
Horn EP, Schroeder F, Wilhelm S et al. (1999).
Postoperative pain facilitates nonthermoregulatory tremor.
Anesthesiology 91: 979–984.
Ioannidis JP (2005). Contradicted and initially stronger effects
in highly cited clinical research. JAMA 294: 218–228.
Jonsson K, Hunt TK, Mathes SJ (1988). Oxygen as an isolated
variable influences resistance to infection. Ann Surg 208:
783–787.
Kurz A, Sessler DI, Lenhardt R (1996). Perioperative normo-
thermia to reduce the incidence of surgical-wound infection
and shorten hospitalization. Study of Wound Infection and
Temperature Group. N Engl J Med 334: 1209–1215.
Lenhardt R, Marker E, Goll V et al. (1997). Mild intra-
operative hypothermia prolongs postanesthetic recovery.
Anesthesiology 87: 1318–1323.
Leslie K, Sessler DI, Bjorksten AR et al. (1995). Mild hypo-
thermia alters propofol pharmacokinetics and increases
the duration of action of atracurium. Anesth Analg 80:
1007–1014.
Leslie K, Bjorksten AR, Ugoni A et al. (2002). Mild core hypo-
thermia and anesthetic requirement for loss of responsive-
ness during propofol anesthesia for craniotomy. Anesth
Analg 94: 1298–1303. table of contents.
Liu M, Hu X, Liu J (2001). The effect of hypothermia
on isoflurane MAC in children. Anesthesiology 94:
429–432.
Madrid E, Urrutia G, Roque i Figuls M et al. (2016). Active
body surface warming systems for preventing complica-
tions caused by inadvertent perioperative hypothermia in
adults. Cochrane Database Syst Rev 4CD009016.
Rajagopalan S, Mascha E, Na J et al. (2008). The effects of
mild perioperative hypothermia on blood loss and transfu-
sion requirement. Anesthesiology 108: 71–77.
Reynolds L, Beckmann J, Kurz A (2008). Perioperative
complications of hypothermia. Best Pract Res Clin
Anaesthesiol 22: 645–657.
Rohrer MJ, Natale AM (1992). Effect of hypothermia on the
coagulation cascade. Crit Care Med 20: 1402–1405.
Ruetzler K, Kovaci B, Guloglu E et al. (2011). Forced-air and a
novel patient-warming system (vitalHEAT vH2) compara-
bly maintain normothermia during open abdominal sur-
gery. Anesthesia and Analgesia 112: 608–614.
Sammour T, Kahokehr A, Hayes J et al. (2010). Warming
and humidification of insufflation carbon dioxide in
 CONSEQUENCES OF PERIOPERATIVE HYPOTHERMIA
laparoscopic colonic surgery: a double-blinded random-
ized controlled trial. Ann Surg 251: 1024–1033.
Scott AV, Stonemetz JL, Wasey JO et al. (2015). Compliance
with Surgical Care Improvement Project for body tempera-
ture management (SCIP Inf-10) is associated with improved
clinical outcomes. Anesthesiology 123: 116–125.
Sessler DI (2016). Perioperative thermoregulation and heat
balance. Lancet 387: 2655–2664.
Sessler DI, Schroeder M (1993). Heat loss in humans
covered with cotton hospital blankets. Anesth Analg 77:
73–77.
Sessler DI, Lee KA, McGuire J (1991). Isoflurane anesthesia
and circadian temperature cycles. Anesthesiology 75:
985–989.
Smeulers NJ, Wierda JM, van den Broek L et al. (1995).
Effects of hypothermic cardiopulmonary bypass on the
697
pharmacodynamics and pharmacokinetics of rocuronium.
J Cardiothorac Vasc Anesth 9: 700–705.
Sun Z, Honar H, Sessler DI et al. (2015). Intraoperative
core temperature patterns, transfusion requirement, and
hospital duration in patients warmed with forced air.
Anesthesiology 122: 276–285.
Tayefeh F, Plattner O, Sessler DI et al. (1998). Circadian
changes in the sweating-to-vasoconstriction interthreshold
range. Pflugers Arch 435: 402–406.
Valeri CR, Khabbaz K, Khuri SF et al. (1992). Effect of skin
temperature on platelet function in patients undergoing
extracorporeal bypass. J Thorac Cardiovasc Surg 104:
108–116.
Winkler M, Akca O, Birkenberg B et al. (2000). Aggressive
warming reduces blood loss during hip arthroplasty.
Anesth Analg 91: 978–984.