Uncontrolled exposure hypothermia is believed to be deleterious in the setting of major trauma. Prevention of hypothermia in the injured patient is currently practiced in both prehospital and in-hospital settings. However, this standard is based on studies of limited patient series that were not designed to identify the independent relationship between hypothermia and mortality. Recent studies suggest that therapeutically applied hypothermia may benefit selected patient subsets. The goal of this study was to evaluate the independent association between admission hypothermia and mortality after major trauma, with adjustment for clinical confounders.
Retrospective analysis of a statewide trauma registry. The primary outcome was death at hospital discharge. The key exposure was hypothermia, defined as body temperature ≤35°C at admission. Multivariate regression was used to risk-adjust for age, severity and mechanism of injury, and route of temperature measurement. Additional adjustment for prehospital exposure time and intravenous fluid therapy was also evaluated.
Trauma centers of the Commonwealth of Pennsylvania.
All trauma patients ≥16 yrs of age for the years 2000–2002. Transferred patients were excluded. Patients were excluded if temperature or route of temperature measurement was not known. Both the full cohort and a subset with isolated severe head injury were evaluated.
Of 38,520 patients, 1,921 (5.0%) were hypothermic at admission. Admission hypothermia was independently associated with increased odds of death in both the full cohort (odds ratio, 3.03; 95% confidence interval, 2.62–3.51) and the subset with isolated severe head injury (2.21; 1.62–3.03), with adjustment for age, severity and mechanism of injury, and route of temperature measurement.
Admission hypothermia is independently associated with increased adjusted odds of death after major trauma. The increase in mortality is not completely attributable to physiologic presentation or injury pattern or severity.
From the Department of Emergency Medicine (HEW, CWC), Department of Surgery (ABP, SAT), and Department of Critical Care Medicine (SAT), University of Pittsburgh School of Medicine, Pittsburgh, PA.
Supported, in part, by Clinical Scientist Development Award K08 HS013628 to Dr. Wang from the Agency for Healthcare Research and Quality. The authors of this article declare no financial or other conflicts of interest.
Presented at the Society for Academic Emergency Medicine Annual Meeting, May 2004, Orlando, FL.
Address requests for reprints to: Henry E. Wang, MD, MPH, Assistant Professor, Department of Emergency Medicine, University of Pittsburgh School of Medicine, 230 McKee Place, Suite 400, Pittsburgh, PA 15213. E-mail: firstname.lastname@example.org