RESUMO
It is commonly believed that hypothermia occurring during burn resuscitation is associated with poor outcome, but there is little direct supporting evidence. The authors conducted an analysis of a statewide trauma registry to determine whether hypothermia (T ≤36.5°C) was associated with mortality when controlling for clinical confounders. They included all patients treated at an accredited burn center from 2000 to 2011 where the trauma registrar recorded the primary injury type as a burn. They excluded records with missing data and nonphysiologic temperature (<26°C or >42°C). The primary exposure of interest was hypothermia. The authors constructed a hierarchical, multivariable logistic regression model to examine the effect of hypothermia on survival, controlling for potentially confounding variables. Predictors of mortality are presented as odds ratio (95% confidence interval). Primary burn injury was coded 17,098 times during the study period. Of these, 3809 were not treated at a burn center and 1192 were excluded for missing data. Admission hypothermia was independently associated with mortality (1.91 [1.58-2.29]) when adjusting for age, sex, total second- and third-degree burn surface area (TBSA), comorbid conditions, injury severity score, direct transport vs referral, method of temperature measurement, year, and the hospital providing care. Increasing age, female sex, TBSA >40%, presence of multiple comorbid conditions, and increasing injury severity score were associated with mortality. Other variables in the model were not independently associated with outcome. There was a weak correlation between TBSA and admission temperature (r = .18). Hypothermia at hospital admission is independently associated with mortality in burn patients when controlling for clinical confounders. Future studies should address potential causes underlying this observation.
