Operative vs Nonoperative Management of Pediatric Blunt Pancreatic Trauma: Evaluation of the National Trauma Data Bank

Abstract

Traumatic pancreatic injury is associated with significant morbidity and mortality. We evaluated the differences in outcomes among children with blunt pancreatic injuries managed operatively and nonoperatively. The National Trauma Data Bank was evaluated from 2002 to 2011. Patients less than18 years of age with blunt pancreatic injuries and Abbreviated Injury Scale (AIS) scores ≥ 3 were identified. Patients were divided into nonoperative (NO), operative (O), and delayed operative (DO; operation performed 48 hours or more after admission) groups. Outcomes evaluated were total length of stay (LOS), ICU use/LOS, complications, and death. Univariate comparisons were performed using Fisher's exact and Kruskal-Wallis rank tests. Multivariable analyses were performed using robust regression and logistic regression. There were 424 cases analyzed. Mean (± SD) age was 10.6 ± 5.3 years, and mean Injury Severity Score (ISS) was 23.4 ± 13.4. Operative groups differed by age (p= 0.002), AIS severity (p= 0.04), and concomitant head injury (p= 0.01), but were similar with regard to sex, race, and ISS. Length of stay was significantly higher in the DO group compared with the NO or O groups; the NO group had the lowest LOS (covariate-adjusted: 18.7 days vs 11.8 days, p < 0.001 and 12.6 days, p < 0.001, respectively) and infection rates (10.2% vs 1.6% and 6.2%, respectively, p= 0.04). The ICU LOS was greatest in the DO group (vs NO, p= 0.03; O, p= 0.29), as was the likelihood of ICU use (vs NO, p= 0.02; O, p= 0.75). Groups did not differ with respect to outcomes including death (p= 0.94) and overall complication rate (p= 0.63). Overall, children managed nonoperatively have equivalent or better outcomes when compared with operative and delayed operative management in regard to death, overall complications, LOS, ICU LOS, and ICU use.