Secondary Logo

Share this article on:

Mild Perioperative Hypothermia and Myocardial Injury

A Retrospective Cohort Analysis

Schacham, Yehoshua N., MD*,†; Cohen, Barak, MD, MHA*,‡; Bajracharya, Gausan R., MD*; Walters, Michael, BS*; Zimmerman, Nicole, MS*,§; Mao, Guangmei, PhD*,§; Tanios, Marianne A., MD, MPH*; Sessler, Daniel I., MD*

doi: 10.1213/ANE.0000000000003840
Perioperative Medicine: Original Clinical Research Report

BACKGROUND: We tested the primary hypothesis that final intraoperative esophageal temperature is associated with increased odds of a composite of in-hospital all-cause mortality and myocardial injury within 7 days after noncardiac surgery. Secondary exposures were time-weighted average intraoperative temperature and area <37°C threshold.

METHODS: Myocardial injury was defined by postoperative fourth-generation troponin T ≥0.03 ng/mL apparently due to cardiac ischemia. Data were extracted for inpatients who had noncardiac surgery with general anesthesia at the Cleveland Clinic between 2012 and 2015. All had esophageal temperature monitoring and routine postoperative troponin monitoring. We estimated the confounder-adjusted association between final intraoperative esophageal temperature and the collapsed composite with multivariable logistic regression. We similarly estimated associations with time-weighted average intraoperative temperature and area <37°C.

RESULTS: Two thousand two hundred ten patients were included. Nearly all final esophageal temperatures were 36°C–37°C. Ninety-seven patients (4.4%) had myocardial injury, and 7 (0.3%) died before discharge. Final intraoperative core temperature was not associated with the collapsed composite: odds ratio, 0.91 (95% confidence interval, 0.68–1.24) per 1°C decrease. Similarly, neither of the secondary exposures was associated with the composite outcome.

CONCLUSIONS: We did not observe an association between mild perioperative hypothermia and mortality or myocardial injury in adults having noncardiac surgery. However, the range of final intraoperative temperatures was small and largely restricted to the normothermic range (36°C–37°C). Trials are needed to further assess the effect of temperature on myocardial injury.

From the *Department of Outcomes Research, Anesthesiology Institute, Cleveland Clinic, Cleveland, Ohio

Internal Medicine C, Sheba Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

Division of Anesthesia, Critical Care and Pain Management, Tel Aviv Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel

§Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, Ohio.

Published ahead of print 8 August 2018.

Accepted for publication August 8, 2018.

Funding: This work was supported by a grant from 3M. The sponsor was not involved in data extraction, analysis, manuscript preparation, or the decision to publish.

Conflicts of Interest: See Disclosures at the end of the article.

Reprints will not be available from the authors.

Address correspondence to Daniel I. Sessler, MD, Department of Outcomes Research, Cleveland Clinic, 9500 Euclid Ave, P77, Cleveland, OH 44195. Address e-mail to DS@OR.org.

KEY POINTS

  • Question: Is intraoperative temperature associated with perioperative myocardial injury?
  • Findings: There was no association between intraoperative temperature and myocardial injury over a narrow range of final intraoperative core temperatures.
  • Meaning: Evidence for very mild hypothermia causing myocardial injury is currently weak.

Overall, 30-day postoperative mortality is 1%–2% among inpatients in the United States.1 , 2 About one-third of these deaths are cardiovascular or consequent to cardiovascular events—with myocardial injury being by far the most common.3 The term myocardial injury after noncardiac surgery (MINS) recognizes that troponin elevations without a nonischemic explanation (eg, sepsis, pulmonary embolus) are prognostic4—even in patients whose symptoms and signs do not meet the Third Universal Definition of Myocardial Infarction.5 Mortality increases markedly as a function of peak postoperative troponin concentration.6

Randomized trials show that mild perioperative hypothermia causes surgical site infections7 and coagulopathy.8 Hypothermia also reduces drug metabolism,9 , 10 prolongs postoperative recovery11 and the duration of hospitalization,7 and provokes thermal discomfort.12 Among the major putative complications of hypothermia, myocardial outcomes are least established. A recent observational analysis suggests an association between Surgical Care Improvement Project compliance, an evidence-based guideline aimed to reduce perioperative complications (part Inf-10 [perioperative temperature management for reduction of surgical site infection]), and cardiovascular outcomes.13 But only a single, 1-center randomized trial of 300 patients specifically compared cardiac outcomes (35.4°C vs 36.7°C).14 The study was published in 1997, well before troponin screening was available; diagnosis of myocardial infarction was therefore primarily based on Holter electrocardiographic monitoring, which was then the best diagnostic approach. Because Holter monitoring is insensitive, the investigators may have missed ≥90% of the myocardial events that would now be expected in their high-risk population.6 In fact, only 1 patient was diagnosed with a myocardial infarction; the study is thus uninformative about the effects of hypothermia on myocardial injury.

Nonetheless, there are reasons to believe that mild hypothermia might provoke cardiovascular complications. For example, mild hypothermia provokes an adrenergic response that includes catecholamine release, vasoconstriction, tachycardia, and hypertension15 , 16—all of which increase myocardial stress. Hypothermia can also provoke shivering and consequently increase the metabolic rate.17 In the 1990s, patients commonly finished surgery with core temperatures <35°C. However, in most high-income countries, it is now routine to maintain “normothermia,” which is arbitrarily and nonphysiologically defined as core temperatures ≥36°C.18 Most patients warmed with forced air have final intraoperative temperatures between 36°C and 36.5°C, with the range extending from 35.5°C to 37°C.19 Whether temperature variations over the current clinical range provoke myocardial injury remains unknown.

We therefore tested the primary hypothesis that there is a confounder-adjusted association between intraoperative temperature, defined by final intraoperative esophageal temperature, and a composite of MINS defined as peak postoperative troponin ≥0.03 ng/mL within 7 days after surgery and in-hospital mortality. Secondarily, we tested the hypotheses that time-weighted average (TWA) intraoperative temperature and area under the threshold of 37°C are associated with the composite of MINS and mortality.

Back to Top | Article Outline

METHODS

With Cleveland Clinic Institutional Review Board approval and waived consent, we evaluated electronic medical records from the Cleveland Clinic Main Campus, Hillcrest Hospital, and Fairview Hospital. The a priori statistical plan was described in the study protocol. Analysis was restricted to inpatients who had esophageal temperature monitoring and routine postoperative troponin monitoring after noncardiac surgery 2015 with general or combined general and neuraxial anesthesia from January 2012 through December 2015.

The study population included patients who were enrolled in the Perioperative Ischemic Evaluation 2 (POISE-2),20 , 21 nitrous oxide and perioperative cardiac morbidity trial (ENIGMA-2),22 Vascular Events in Noncardiac Surgery Patients Cohort Evaluation Study (VISION),6 and Balanced Anesthesia trial (BALANCED)23 studies. The minimum enrollment criteria for each of these studies were inpatient surgery and ≥45 years of age. We also included moderate- to high-risk colorectal surgery patients who generally had routine postoperative troponin monitoring starting in 2015. We excluded patients with fourth-generation serum troponin T concentration ≥0.03 ng/mL within 3 days before surgery, nonischemic reasons for troponin elevation (Appendix Table 1), and American Society of Anesthesiologists physical status of V. Patients with incomplete baseline data or inadequate temperature monitoring were further excluded. We only included the first surgery for each patient in this study.

In most patients, troponin was evaluated preoperatively, 6–12 hours after surgery, and then daily for the initial 3 postoperative days as long as patients remained hospitalized. Routine screening was supplemented by troponin testing prompted by signs or symptoms of myocardial injury, but to minimize selection bias, only patients with scheduled troponin screening were included in our analysis. Core temperature was measured in the esophagus with a thermistor. Clinical thermistors are generally accurate to within about 0.1°C. All surgical patients at the Cleveland Clinic are warmed with forced air during surgery, while prewarming is hardly ever used.

Back to Top | Article Outline

Statistical Analysis

Our primary outcome was a composite of MINS defined as peak postoperative fourth-generation troponin ≥0.03 ng/mL within 7 days after surgery and in-hospital mortality. Our primary exposure was final intraoperative esophageal temperature. Secondary exposures were TWA intraoperative temperature and area under the threshold of 37°C.

A complete case analysis was done, as patients with missing data were excluded. The study population was summarized on potentially confounding baseline and procedural characteristics using appropriate summary, including means ± standard deviations, medians [Q1, Q3], or N (%). Types of surgery were categorized using Clinical Classifications Software, with categories with <5% incidence combined as an “other” classification.24

First, we estimated the association between final intraoperative esophageal temperature and a collapsed composite of in-hospital mortality and MINS using a multivariable logistic regression model. Potential confounding variables in Table 1 were added to logistic regression models if the significance level for entering the model was <.20; similarly, variables were retained in the model in the backward elimination step if the significance level for entering the model was <.30. Age, estimated blood loss, red blood cell transfusion, history of perivascular disease, hypertension, chronic kidney disease, and cardiac surgeries were finally selected in the model and adjusted for the association between intraoperative temperature and the outcome. Linearity of the relationship was evaluated by including restricted cubic spline terms at 10, 50, and 90 percentiles in the model. If the spline terms were significantly associated with the outcome (P < .05), then we would report the nonlinear association; otherwise, we would use the linear function of final temperature to assess the association.

Table 1

Table 1

Second, we estimated the associations of TWA intraoperative temperature and area under the curve (AUC) intraoperative temperature <37°C in separate multivariable logistic regression models using the same approach as in the primary analysis. TWA temperature was calculated as the AUC of the temperature measurements divided by total measurement time. AUC <37°C was calculated as the total area <37°C and above the temperature measurements. The associations were adjusted for surgery duration and other confounding variables.

We used an α of .05 for this analysis, thus using a significance criterion of .05 for the primary analysis and .025 for each secondary analysis (ie, .05/2; Bonferroni correction). We completed the analyses using SAS version 9.4 (SAS Institute, Carey, NC) and R version 3.3.2 (R Project for Statistical Computing, Vienna, Austria).

Back to Top | Article Outline

Sample Size and Power

A priori, we expected that ≥2500 patients would be eligible for analysis and that the incidence of myocardial injury would be 8% (as observed in the VISION cohort). The mean ± standard deviation final core temperature in a similar population of adults was 36.3 ± 0.5°C.19 We estimated that we would have 90% power at the 0.05 significance level to detect an odds ratio of ≥1.9 per 1°C for 8% incidence of MINS/mortality based on logistic regression.

Back to Top | Article Outline

RESULTS

A total of 3595 patients met the study inclusion criteria from 2012 to 2015, of whom 2210 were eligible for our analyses (Figure 1). The median [Q1, Q3] number of postoperative troponin measurements was 2 [2, 3]. Among eligible patients, 97 (4.4%) had MINS within 7 days after surgery, 7 (0.3%) died before discharge, and 1 had MINS before death. Patients are summarized on potentially confounding baseline and procedural characteristics listed in Table 1. The distribution of each hypothermia exposure is presented in Figure 2. The mean final temperature was 36.2°C (SD = 0.7), and 733 (33%) patients had a final core temperature <36°C.

Figure 1

Figure 1

Figure 2

Figure 2

Table 2

Table 2

Figure 3

Figure 3

Final intraoperative core temperature was not associated with the collapsed composite, with an estimated odds ratio (95% confidence interval) of 0.91 (0.68–1.24) per 1°C decrease (P = .53; Table 2 and Figure 2), adjusted for selected confounding variables. Results were similar in a sensitivity analysis restricted to MINS. Similarly, none of the secondary exposures was associated with the composite, with estimated odds ratios (97.5% confidence interval) of 0.90 (0.59–1.38) per 1°C decrease in TWA intraoperative temperature (P = .57) and 1.00 (0.92–1.09) per 1°C × hour increase in AUC temperature <37°C (P = .95; Figure 3).

Back to Top | Article Outline

DISCUSSION

There was no association between final intraoperative esophageal temperature and a collapsed composite of in-hospital mortality and MINS. Furthermore, there was no association between intraoperative TWA temperature or AUC temperature and our composite outcome. These results are somewhat surprising given that even mild hypothermia provokes a substantial adrenergic response15 , 16 and shivering,17 both of which seem likely to promote myocardial stress and injury.

An important distinction between our current analysis and previous randomized trials is that the control group in most was colder than is now common, for example, 34.5°C vs 36.5°C. (The major trials of hypothermia7–10 were conducted before active warming was routine, and patients were randomized to routine care [hypothermia] versus extra warming.) The current range of final intraoperative temperatures is smaller and higher, at least in high-income countries’ operating rooms. For example, the average final intraoperative temperature in our patients was 36.4°C. Most final temperatures were between 36.0°C and 36.8°C and only rarely <36°C or >37°C. Because core temperatures in our patients were so tightly clustered, our statistical models lack accuracy at lower and higher temperatures and, thus, do not add sufficient information regarding the hypothermia and hyperthermia ranges.

Final intraoperative temperatures near 37°C are unusual, even in actively warmed surgical patients. About 65% of included patients had digestive surgery, nearly all colorectal. About half of the patients undergoing colorectal surgery at the Cleveland Clinic carry a diagnosis of inflammatory bowel disease. A consequence of their chronic inflammation is that many of these patients have a mild fever. Relatively high temperatures in these patients may reflect an underlying pathology rather than aggressive warming, and those patients with the highest temperatures may have the worst underlying inflammation. One consequence of inflammation is platelet activation, which may in turn aggravate the risk of myocardial injury. Confounding by mild fever in patients with inflammatory bowel disease may thus distort the expected protective relationship between high core temperature and myocardial injury.

How best to characterize intraoperative core temperature patterns remains unknown and may well depend on the outcome of interest. For example, hypothermia-induced coagulopathy is presumably an instantaneous function of tissue temperature because coagulation enzymes and platelets have no memory for previous thermal perturbations. TWA temperature or area under a threshold, such as 37°C or 36°C, may thus best characterize intraoperative temperature with respect to bleeding. Myocardial injury differs from bleeding in that it appears to be rare during surgery, with 94% occurring later on the day of surgery and over the initial 2 postoperative days.6 Final intraoperative temperature may be most relevant for myocardial injury because awakening temperature provokes the potentially harmful autonomic response to hypothermia. Thus, we considered final intraoperative temperature to be our primary exposure while considering other temperature characterizations as secondary exposures.

We minimized selection bias by restricting analysis to patients who had routine troponin monitoring either because they were enrolled in trials or because they were monitored per clinical routine. Our study is strengthened by the relatively large sample size, our ability to adjust for many potentially confounding variables, and the inclusion of a variety of surgical procedures. External validity is strengthened by the diverse study population from 3 Cleveland Clinic hospitals.

Given the retrospective nature of this analysis, there is surely a degree of unobserved confounding, although we adjusted for many patient and procedural characteristics thought to be related to myocardial injury. Given a 0.3% incidence of in-hospital mortality, we used a collapsed composite (ie, any MINS or mortality versus none) approach to this analysis rather than assessing the average hypothermia association across both components of the composite. The collapsed composite approach overweighs components with higher incidences, so our analysis is strongly driven by MINS, potentially missing an important difference in mortality.20 It was predictable that the incidence of MINS would far exceed mortality, but mortality was nonetheless included to reduce attrition bias.

The outcome incidence turned out to be about half what we expected (4.4% vs 8%) presumably because we included many patients undergoing colorectal surgery who were relatively young and healthy and few patients undergoing vascular surgery who have especially high cardiovascular risk. We have much data between 35.5°C and 37.5°C but relatively little at lower and higher temperatures. It thus remains distinctly possible that more serious hypothermia is associated with MINS. However, it is now uncommon to allow core temperatures to be <35.5°C.

In summary, we did not observe a significant association between intraoperative temperature, defined various ways, and a collapsed composite of in-hospital all-cause mortality and myocardial injury in adults having noncardiac surgery. However, the range of final intraoperative temperatures was small and largely restricted to the normothermic range (36°C–37°C), our analysis provides little information about temperatures <36°C. Even highly significant retrospective associations should be considered hypothesis generating rather than evidence of causality. Similarly, a lack of association should not be considered proof of no effect. Thus, while our results do not suggest that there is a large effect of temperature on myocardial injury, only a well-powered randomized trial will provide a definitive answer. Fortunately, 1 is in progress (NCT03111875).

Back to Top | Article Outline

DISCLOSURES

Name: Yehoshua N. Schacham, MD.

Contribution: This author helped design the study, draft the initial manuscript, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Barak Cohen, MD, MHA.

Contribution: This author helped review and draft the manuscript and approve the final manuscript.

Conflicts of Interest: None.

Name: Gausan R. Bajracharya, MD.

Contribution: This author helped review and draft the manuscript and approve the final manuscript.

Conflicts of Interest: None.

Name: Michael Walters, BS.

Contribution: This author helped review and draft the manuscript and approve the final manuscript.

Conflicts of Interest: None.

Name: Nicole Zimmerman, MS.

Contribution: This author helped design the study, extract and analyze the data, and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Guangmei Mao, PhD.

Contribution: This author helped extract and analyze the data and review and approve the final manuscript.

Conflicts of Interest: None.

Name: Marianne A. Tanios, MD, MPH.

Contribution: This author helped review and draft the manuscript and approve the final manuscript.

Conflicts of Interest: None.

Name: Daniel I. Sessler, MD.

Contribution: This author helped design the study, draft the initial manuscript, and review and approve the final manuscript.

Conflicts of Interest: D. I. Sessler is a consultant for 3M, with all fees donated to charity or used to support research.

This manuscript was handled by: Tong J. Gan, MD.

APPENDIX Table 1

APPENDIX Table 1

Back to Top | Article Outline

REFERENCES

1. Henderson WG, Khuri SF, Mosca C, Fink AS, Hutter MM, Neumayer LA. Comparison of risk-adjusted 30-day postoperative mortality and morbidity in Department of Veterans Affairs hospitals and selected university medical centers: general surgical operations in men. J Am Coll Surg. 2007;204:1103–1114.
2. Semel ME, Lipsitz SR, Funk LM, Bader AM, Weiser TG, Gawande AA. Rates and patterns of death after surgery in the United States, 1996 and 2006. Surgery. 2012;151:171–182.
3. Devereaux PJ, Sessler DI. Cardiac complications in patients undergoing major noncardiac surgery. N Engl J Med. 2015;373:2258–2269.
4. The Vascular Events in Noncardiac Surgery Patients Cohort Evaluation (VISION) Investigators. Myocardial injury after noncardiac surgery: a large, international, prospective cohort study establishing diagnostic criteria, characteristics, predictors, and 30-day outcomes. Anesthesiology. 2014;120:564–578.
5. Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial infarction. Circulation. 2012;126:2020–2035.
6. Devereaux PJ, Biccard BM, Sigamani A, et al; Writing Committee for the Vision Study Investigators. Association of postoperative high-sensitivity troponin levels with myocardial injury and 30-day mortality among patients undergoing noncardiac surgery. JAMA. 2017;317:1642–1651.
7. Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization: study of Wound Infection and Temperature Group. N Engl J Med. 1996;334:1209–1215.
8. Rajagopalan S, Mascha E, Na J, Sessler DI. The effects of mild perioperative hypothermia on blood loss and transfusion requirement: a meta-analysis. Anesthesiology. 2008;108:71–77.
9. Heier T, Clough D, Wright PM, Sharma ML, Sessler DI, Caldwell JE. The influence of mild hypothermia on the pharmacokinetics and time course of action of neostigmine in anesthetized volunteers. Anesthesiology. 2002;97:90–95.
10. Leslie K, Sessler DI, Bjorksten AR, Moayeri A. Mild hypothermia alters propofol pharmacokinetics and increases the duration of action of atracurium. Anesth Analg. 1995;80:1007–1014.
11. Lenhardt R, Marker E, Goll V, et al. Mild intraoperative hypothermia prolongs postanesthetic recovery. Anesthesiology. 1997;87:1318–1323.
12. Kurz A, Sessler DI, Narzt E, et al. Postoperative hemodynamic and thermoregulatory consequences of intraoperative core hypothermia. J Clin Anesth. 1995;7:359–366.
13. Scott AV, Stonemetz JL, Wasey JO, et al. Compliance with surgical care improvement project for body temperature management (SCIP Inf-10) is associated with improved clinical outcomes. Anesthesiology. 2015;123:116–125.
14. Frank SM, Fleisher LA, Breslow MJ, et al. Perioperative maintenance of normothermia reduces the incidence of morbid cardiac events: a randomized clinical trial. JAMA. 1997;277:1127–1134.
15. Frank SM, el-Gamal N, Raja SN, Wu PK. alpha-Adrenoceptor mechanisms of thermoregulation during cold challenge in humans. Clin Sci (Lond). 1996;91:627–631.
16. Frank SM, Higgins MS, Fleisher LA, Sitzmann JV, Raff H, Breslow MJ. Adrenergic, respiratory, and cardiovascular effects of core cooling in humans. Am J Physiol. 1997;272:R557–R562.
17. Frank SM, Fleisher LA, Olson KF, et al. Multivariate determinants of early postoperative oxygen consumption in elderly patients: effects of shivering, body temperature, and gender. Anesthesiology. 1995;83:241–249.
18. Sessler DI. Perioperative thermoregulation and heat balance. Lancet. 2016;387:2655–2664.
19. Sun Z, Honar H, Sessler DI, et al. Intraoperative core temperature patterns, transfusion requirement, and hospital duration in patients warmed with forced air. Anesthesiology. 2015;122:276–285.
20. Devereaux PJ, Mrkobrada M, Sessler DI, et al; POISE-2 Investigators. Aspirin in patients undergoing noncardiac surgery. N Engl J Med. 2014;370:1494–1503.
21. Devereaux PJ, Sessler DI, Leslie K, et al; POISE-2 Investigators. Clonidine in patients undergoing noncardiac surgery. N Engl J Med. 2014;370:1504–1513.
22. Myles PS, Leslie K, Chan MT, et al; ANZCA Trials Group for the ENIGMA-II investigators. The safety of addition of nitrous oxide to general anaesthesia in at-risk patients having major non-cardiac surgery (ENIGMA-II): a randomised, single-blind trial. Lancet. 2014;384:1446–1454.
23. Short TG, Leslie K, Chan MT, Campbell D, Frampton C, Myles P. Rationale and design of the balanced anesthesia study: a prospective randomized clinical trial of two levels of anesthetic depth on patient outcome after major surgery. Anesth Analg. 2015;121:357–365.
24. Elixhauser A, Steiner C, Palmer L; US Agency for Healthcare Research and Quality. Clinical Classifications Software (CCS), 2015. Available at: http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp.Accessed September 30, 2018.
© 2018 International Anesthesia Research Society