June 1997

Perioperative Maintenance of Normothermia Reduces the Incidence of Morbid Cardiac Events: A Randomized Clinical Trial.

Frank SM, Fleisher LA, Breslow MJ, Higgins MS, Olson KF, Kelly S, Beattie C;

JAMA 1997; 277:1127-1134.

[ see abstract below ]

In this article Steve Frank and colleagues serve up some chilling information about the risk of mild hypothermia. The effect of hypothermia on the incidence of perioperative cardiac morbidity is examined. Kurz et al¹ recently published a much discussed article demonstrating that postoperative infections and length of stay were affected by mild hypothermia in patients at high risk for postoperative infections.

Frank demonstrates that this mild hypothermia to 35-35.5 degrees--a temperature most clinicians would rarely treat and about which they would almost never be concerned--apparently has additional cardiac implications.

Frank et al hypothesized that hypothermia would impact perioperative cardiac morbidity. A randomized controlled clinical trial of 300 patients was done. Inclusion criteria were a highly stressful procedure (undergoing thoracic, vascular, or abdominal surgery) along with a high risk of, or documented history of coronary artery disease (CAD) and age > 60. They also were to be admitted to the ICU.

These patients were studied to determine if cardiac morbidity, still the leading cause of perioperative death, was impacted by the mild hypothermia that normally accompanies the administration of anesthesia and the conduct of surgery. The mean temperature for the hypothermic group was 35.4 degrees C, a temperature that is routinely seen in patients undergoing surgery.

This group received routine care - blood transfusions ands fluids being warmed, in line passive humidifier, and a normal room temperature of 21 degrees C. They did not receive forced air convection warming blankets intra or postoperatively. Treatment of the normothermia group only differed in that they were warmed with a forced air warming device intra and post operatively.

A generic (pent, sux, tube, isoflurane, nitrous, fentanyl) general anesthetic was administered most often for this study. Patients undergoing peripheral vascular reconstructions usually received epidurals. Postoperative pain management was state of the art, employing PCA, epidural infusions, and intrathecal morphine injections as appropriate.

Determinants of myocardial supply and demand--e.g. heart rate, blood pressure, hematocrit--were both measured and controlled using standard approaches. No manipulations were performed or treatments withheld that would exaggerate the effect of what was being measured. In other words, the study was conducted so that it would be applicable to virtually any anesthesiologist's practice.

Morbid cardiac events were assessed in a blinded fashion--cardiac arrest, myocardial infarction, and unstable angina/ischemia. Holter monitors were used for the first 24 hours to define ventricular tachycardia (5 or more beats) and ECG ischemia.

Intraoperatively no differences were found between groups. Postoperatively the hypothermic group had significantly more morbid cardiac events (p = ..02) 10 hypothermic vs. 2 normothermic patients suffered morbid events. Contradicting Mangano's recent publication about the usefulness of perioperative beta blockers², Frank discovered that preoperative beta blocker use was a predictor of ECG events and morbid cardiac events. No information was noted about continuation of beta blockers.

Two theories are advanced to explain hypothermia's effect on perioperative cardiac events. The first is that hypothermia directly elicits an adrenergic response. The second cause suggests that these adverse effects might be the result of more cardiac demand due to shivering brought on by the hypothermia.

What does all of this mean to the practicing clinician? Keeping the patient warm is likely to reduce morbid cardiac events by 55%. The cost of using a warming device is not zero, however, and the study relied on intermediate variables (e.g. ischemia not progressing to an MI, clinically insignificant non-sustained VT) in order to attain statistical significance.

There is no statistically significant evidence that the ultimate hard outcomes of the patients were affected by hypothermia, and this conclusion is reinforced by the fact that patients experiencing these morbid cardiac events did not have a longer length of stay. Given these facts, we cannot absolutely conclude that the cost of the intervention is warranted. However, all of the hard outcomes--cardiac arrests² and MI's¹ did occur in the hypothermic group.

Combining this data with that from Kurz argues for a much more concerted effort on our part to keep the patients at high risk of postoperative infection (e.g. colon surgery) or at high risk for cardiac events (vascular surgery, documented significant CAD) normothermic throughout the procedure.

     References:
     ¹ Kurz A, Sessler DI, Lenhardt R, The Study of Wound Infection and Temperature Group: Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization. N Engl J Med 1996; 334:1209-15

     ² Mangano DT, Layug EL, Wallace A, Tateo I, The Multicenter Study of Perioperative Ischemia Research Group: Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 1996; 335:1713- 20. Click here to read Dr. Lubarsky's commentary on this article.

Return to the Current Literature Review Front Page, or read the abstract: