A new algorithm to allow early prediction of mortality in elderly burn patients

Abstract

IntroductionThe elderly are the fastest growing population segment, and particularly susceptible to burns. Predicting outcomes for these patients remains difficult. Our objective was to identify early predictors of mortality in elderly burn patients. MethodsOur Burn Center's prospective database was reviewed for burn patients 60+ treated in the past 10 years. Predictor variables were identified by correlative analysis and subsequently entered into a multivariate logistic regression analysis examining survival to discharge. Results203 patients of 1343 (15%) were eligible for analysis. The average age was 72±10 (range 60–102) and the average total body surface area (TBSA) burned was 23±18% (range 1–95). Age, TBSA, base deficit, pO2, respiratory rate, Glasgow Coma Score (GCS), and Revised Trauma Score (RTS, based on systolic blood pressure, respiratory rate, and GCS) all correlated with mortality (p≤0.05). Using multiple logistic regression analysis, a model with age, TBSA and RTS was calculated, demonstrating:increased risk of mortality=β0+1.12 (age)+1.094 (TBSA)+0.718 (RTS)In this model, β0 is a constant that equals −8.32. ConclusionsPredicting outcomes in elderly burn patients is difficult. A model using age, TBSA, and RTS can, immediately upon patient arrival, help identify patients with decreased chances of survival, further guiding end-of-life decisions.