Severe accidental hypothermia treated in an ICU: prognosis and outcome

To assess the characteristics and outcomes of patients admitted to an ICU for severe accidental hypothermia, and to identify risk factors for mortality.

All consecutive patients admitted to an ICU between January 1, 1979, and July 31, 1998, with a temperature of < or = 32 degrees C were retrospectively analyzed. Rewarming was always conducted passively with survival blankets and conventional covers. Prognostic factors were studied by means of univariate analysis (Mann-Whitney U and chi(2) tests) and multivariate analysis (logistic regression).

Forty-seven patients were enrolled (mean +/- SD age, 61.7 +/- 16 years). Five patients had a cardiac arrest before ICU admission. Patient characteristics at ICU admission were as follows: temperature, 28.8 +/- 2.5 degrees C; systolic BP, 85 +/- 23 mm Hg; heart rate, 60 +/- 24 beats/min; Glasgow Coma Scale, 10.4 +/- 3.7; and simplified acute physiology score (SAPS) II, 50.9 +/- 27. Mechanical ventilation was necessary in 23 cases, and 22 patients in shock received vasoactive drugs. The mean length of stay in the ICU was 6.7 +/- 9 days. Eighteen patients (38%) died, but ventricular arrhythmia was never the cause. Univariate analysis identified several prognostic factors (p < 0.05): age (57 +/- 16 years vs 69 +/- 14 years), systolic arterial BP (93 +/- 20 mm Hg vs 71 +/- 21 mm Hg), blood bicarbonate level (23.5 +/- 5.2 mmol/L vs 16.6 +/- 6.2 mmol/L), SAPS II score (35.3 +/- 19.5 vs 72 +/- 21), mechanical ventilation (34% vs 81%), vasopressor agents (42% vs 82%), rewarming time (11.5 +/- 7.2 h vs 17.2 +/- 7 h), and discovery of the patient at home (2.3% vs 54.5%). The initial temperature did not influence vital outcome (28.9 +/- 2.6 degrees C vs 28.6 +/- 2.2 degrees C). Only the use of vasoactive drugs (odds ratio, 9; 95% confidence interval, 1.6 to 50.1) was identified as a prognostic factor in the multivariate analysis.

Severe accidental hypothermia is a rare cause of ICU admission in an urban area. Its mortality remains high, but there is no overmortality according to the SAPS II-derived prediction of death. Shock, requiring treatment with vasoactive drugs, is an independent risk factor for mortality, while initial core temperature is not. It remains to be determined whether aggressive rather than passive rewarming procedures are better.