Identification of novel prognostic indicators in burned patients

The size and depth of burn and patient age are useful early prognostic indicators in burned patients, but have limited value in predicting which patients in a given cohort are likely to die. The objective of this study was to identify additional variables in the first 10 days of burn injury which could better predict patient outcome. Variables consisting of demographic information, routine laboratory data, and clinical observations on 89 burned patients (63 survivors and 26 nonsurvivors) were analyzed. Compared to survivors, nonsurvivors had significantly larger burns (58 +/- 23% vs. 37 +/- 17%; p less than 0.0002) and a higher incidence of Gram-negative septicemia (20 of 26 [77%] vs. 24 of 63 [38%]; p less than 0.001). Potential prognostic variables were subjected to multivariate logistic regression analysis for each of the first 10 days following burn injury in order to identify a combination of parameters which predicted patient outcome. The regression analyses revealed that, as previously demonstrated, patient age and burn size were significant predictors of mortality on admission and throughout the first 10 days postburn. In addition, absolute monocyte count (AMC), absolute lymphocyte count (ALC), maximum daily temperature (Tmax), and BUN were also significant predictors (p less than 0.05). These data indicate that logistic regression models can identify simple prognostic variables in burned patients which may improve clinicians' ability to identify high-risk patients early in the course of their burn injuries.