Extracorporeal life support rewarming rate is associated with survival with good neurological outcome in accidental hypothermia

Abstract

Abstract OBJECTIVES Favourable outcomes have been demonstrated after extracorporeal life support (ECLS) facilitated rewarming for severe accidental hypothermia. The clinical impact of varying rewarming rates however is unclear. We sought to quantify the change in the probability of good neurological outcome with ECLS rewarming rate and identify the optimal rewarming rate threshold. METHODS We performed a secondary analysis of the International ACcidental Hypothermia Extracorporeal Life Support Collaborators, an individual patient data data set (n = 658) for ECLS-assisted rewarming for accidental hypothermia. The independent variable of interest was rewarming rate. The primary outcome was survival with good neurological status. We applied an adjusted marginal effects model to quantify the probability of good neurological outcome over clinically observed rewarming rates. We examined strata defined by sex, initial potassium level and history of asphyxiation. RESULTS Of 658 cases, the median age and initial core temperature were 36 years (22–55) and 24.5°C (22.1–26.2) respectively; 190 (29%) were female, and 547 (83%) had a non-perfusing initial cardiac rhythm. The mean rewarming rate was 7.0°C/h. The median ECLS duration was 5.8 h (range: 0.5–158 h). The overall survival was 46% (n = 303/658), and good neurological outcome was 40% (n = 265/658). The median intensive care unit and hospital length of stay was 5 days (range: 1–35 days) and 18 days (range: 1–106 days), respectively. Marginal effects analysis demonstrated a 1.9% decrease in the probability of survival with good neurological outcome for each 1°C/h increase in rewarming. Across the reported range of rewarming rates (0.05–30.8°C/h), the probability of good neurological outcome declined from 49.6% to 4.1% for an average patient. The relationship was similar within various subgroups. The optimal cut-off threshold for the rate of rewarming to distinguish between a good and poor neurological outcome was ≤5.0°C/h. CONCLUSIONS Among cases with severe accidental hypothermia treated with ECLS, slower rewarming rates are associated with improved survival with good neurological outcomes. Slow rewarming, at rates ≤5.0°C/h, may improve clinical outcomes.