Are interventional radiology techniques ideal for nonpenetrating splenic injury management: Robust statistical analysis of the Trauma Quality Program database.

Abstract

Splenic artery embolization (SAE) is increasingly favored for adult blunt splenic injury management. We compared SAE to other splenic injury management strategies using robust statistical techniques. Univariate analyses of demographics and outcomes were performed for four patient groups: observation, SAE, splenic surgery, splenic surgery + SAE in the American College of Surgeons Trauma Quality Program (TQIP) database. To address nonlinear associations of ED vital signs with mortality, multivariable spline-based logistic regression models with interaction terms between hemodynamic status and management strategy and either splenic Abbreviated Injury Score (AIS) or Injury Severity Score (ISS), were generated. In 44,187 splenic injury patients meeting study inclusion criteria, the most common management strategy was observation alone (77.9%). The observation group had median spleen AIS of 2, ISS 20, with 6.3% mortality; SAE (2.6%) had median spleen AIS3, ISS 24, with 6.6% mortality; splenic surgery (22.4%) AIS4, ISS 29, with 15.4% mortality; and splenic surgery + SAE (0.04%) AIS4, ISS 29, with 15.2% mortality. In multivariable models, SAE had lower predicted probability of mortality than surgery over most initial ED systolic blood pressures (SBPs). At all spleen AIS, SAE had lower predicted mortality than surgery. SAE had lower mortality than surgery except at very high ISS, where it was comparable. SAE had lower predicted mortality than observation management at spleen AIS≥3. In subgroup analysis of patients without severe multi-system injuries, predicted mortality did not differ by management strategy. SAE is associated with decreased mortality at spleen AIS 3-5. The benefits of SAE appear to be largely for spleen AIS 3-5 in the setting of severe (AIS≥3) multi-system injuries.