Hypothermia: the Bellevue Experience.

All adult patients (102 cases) presenting to Bellevue Hospital Medical Center over a calendar year (1978) with core temperatures less than 35 C were studied. Statistically significant correlations between hypothermia and mortality were identified according to mental status, hypoxia, hypotension, hyperamylasemia, duration and severity of hypothermia, and history of exposure and alcohol ingestion. Mortality could not be predicted on the basis of season, age (if greater than 40 years old), sex, presence of infection, or presenting temperature (if greater than 26 C). Thyroid and adrenal function were not significantly altered. Of only nine diabetic patients, four died in ketoacidosis or hyperosmolar states. There were no cases of meningitis, and the incidence of "occult" bacteremia was less than 1%. Prolonged hypothermia was uniformly associated with profound underlying medical disease. In patients presenting with temperatures less than 26 C, 50% of deaths resulted from temperature-induced ventricular arrhythmias. Alcoholics hypothermic from exposure had excellent prognoses; however, temperatures less than 26 C were associated with a marked and statistically significant incidence of death.