External Validation of Risk Stratification Models Predicting the Immediate Mortality After Open Repair of Ruptured AAA.

Abstract

Preoperative risk stratification of the immediate postoperative death after surgery may be helpful for the decision-making for patients with ruptured AAAs (rAAs). The purpose of this study was to apply and validate the Glasgow aneurysm score (GAS) and the Hardman Index (HI) in predicting 30-day/in-hospital mortality in open surgical repair (OSR) and endovascular repair (rEVAR) of rAAAs. We conducted a retrospective review of a prospectively created database including all patients with a rAAA referred to our Institution between August 1998 and December 2014. Primary outcome was 30-day/in-hospital mortality. Multivariate logistic regression analysis was used to identify independent risk factors. The receiver-operator characteristic curve was used to determine the value of the HI and GAS in predicting 30-day/in-hospital death. A total of 150 patients (130 patients received OSR, 20 patients rEVAR) were included in our analysis. The 30-day/in-hospital mortality was 34.0% for the entire cohort: 36.15% for OSR group and 20.0% for rEVAR group (p 0.210). A multivariate analysis in the OSR group evidenced that unconsciousness was a statistically significant [adjusted odds ratio (OR) 8.00] predictor of 30-day/in-hospital mortality. The mean GAS was 86.9±16.1 for the OSR group and 88.1±11.2 for the rEVAR group (p 0.773). The AUC for GAS was 0.805 among OSR patients and 0.975 among rEVAR patients. The mean HI in the OSR group was 1.11±1.0 and the AUC for HI was 0.82. Surgical repair of rAAAs is still associated with a considerable mortality rate. We confirmed the great discriminative ability of GAS in patients with rAAAs treated with OSR. With regard to HI, this scoring system could accurately predict early mortality after OSR in our cohort but failed to identify patients at highest risk for postoperative mortality.