Inadvertent hypothermia and mortality in postoperative intensive care patients: retrospective audit of 5050 patients.

Abstract

We proposed that many Intensive Care Unit (ICU) patients would be hypothermic in the early postoperative period and that hypothermia would be associated with increased mortality. We retrospectively reviewed patients admitted to ICU after surgery. We recorded the lowest temperature in the first 24 h after surgery using tympanic membrane thermometers. We defined hypothermia as < 36 degrees C, and severe hypothermia as < 35 degrees C. We studied 5050 consecutive patients: 35% were hypothermic and 6% were severely hypothermic. In-hospital mortality was 5.6% for normothermic patients, 8.9% for all hypothermic patients (p < 0.001), and 14.7% for severely hypothermic patients (p < 0.001). Hypothermia was associated with in-hospital mortality: OR 1.83 for each degree Celsius ( degrees C) decrease (95% CI: 1.2-2.60, p < 0.001). Given the evidence for improved outcome associated with active patient warming during surgery we suggest conducting prospective studies of active warming of patients admitted to ICU after surgery.