J. P. Meizoso1, C. J. Allen1, J. J. Ray1, C. M. Thorson1, L. R. Pizano1, N. Namias1, K. G. Proctor1, J. E. Sola2, C. I. Schulman1 1University Of Miami,Trauma, Surgical Critical Care, And Burns,Miami, FL, USA 2University Of Miami,Pediatric And Adolescent Surgery,Miami, FL, USA
Introduction: Although trauma is the leading cause of death and morbidity in children in the US, pediatric burns continue to represent a large source of morbidity with an estimated 30,000 children requiring inpatient admission every year for treatment. In addition, children account for approximately 25% of burn deaths each year. The objective of this study is to identify major predictors of mortality in the pediatric burn population at a large American Burn Association verified burn center.
Methods: A retrospective review of all pediatric burn patients (≤ 17y) from January 1993 to December 2013 were surveyed. Demographics, laboratory studies, total body surface area (TBSA) burn, need for emergent procedures, length of stay (LOS), and survival were obtained. Univariate analysis was performed to identify factors significantly associated with mortality. A multiple logistic regression model was used to identify independent predictors of mortality. Data are expressed as M±SD if normally distributed or median (interquartile range) if not.
Results: 943 patients evaluated at our center were average age 4.9±5 years, 63% male, 44% black, 33% white, and 21% Hispanic, TBSA 8%(28), base deficit -2 mEq/L(8), Glasgow Coma Scale (GCS) 15(0), scene GCS 15(0), and hospital LOS 3(30) days. The vast majority of burns occurred at home (95%). Only 14% of patients had a TBSA burn >20%. Intubation was required in 6.7% of patients. Overall mortality was 2.3%. Initial base deficit [-8(11) vs -2(8), p<0.001], bicarbonate level (18±5 vs 23±3, p<0.001), hematocrit (46±11 vs 34±12, p=0.012), pCO2 (49±21 vs 41±10, p=0.013), pH (7.24±0.15 vs 7.36±0.09, p<0.001), scene GCS [3(12) vs 15(0), p<0.001], hospital GCS [3(12) vs 15(0), p<0.001], TBSA [50(55) vs 7(25), p<0.001], and the need for intubation (67% vs 5.4%, p<0.001) were significantly associated with mortality on univariate analysis. The logistic regression model identified TBSA burn [odds ratio (OR): 1.09, confidence interval (CI): 1.03-1.15] and scene GCS (OR: 0.83, CI: 0.68-0.99) as significant independent predictors for mortality (area under receiver operator characteristic curve: 0.979).
Conclusion: Pediatric burn patients are typically young and male with predominantly small burns (<20% TBSA) that occur in the home. Overall mortality over a 20-year period in our burn center was 2.3%. Independent risk factors for mortality included TBSA burn and Glasgow Coma Scale at the scene of the incident. This suggests pre-hospital determinants such as GCS might serve as an indicator for poor outcome in the pediatric burn patient.