Evaluation of risk prediction models, V-POSSUM and GAS, in patients with acute abdominal aortic rupture treated with EVAR or an open procedure

Abstract

Background: In this time of outcome measurements predictive models are needed for correction of comorbidity, in order to perform audits on the outcome of patients. The aim of this study was to evaluate the use of risk-predicting scoring methods in the treatment of RAAA with emergency endovascular aneurysm repair (eEVAR) and open repair in order to analyze outcomes.

Methods: Two hundred and six consecutive patients underwent open repair or eEVAR for RAAA. Patients data were compared with preoperative risk assessments according to the V-POSSUM and the Glasgow Aneurysm Score (GAS).

Results: The 30-day mortality rate was 45/206 (22%) in the whole cohort, 41/179 (23%) (95% Confidence Interval [CI] 17-30) after open repair and 4/27 (15%, 95% CI 6-32) after eEVAR. The patient population did not significantly change over time, according to age, gender, V-POSSUM and GAS, neither did the 30-day mortality. The GAS appeared to be a reasonable predictor of postoperative outcome for both open and endovascular RAAA repair, with an area under the curve (AUC) of 0.73 (SD 0.04; 95% CI 0.65-0.80; P<0.001). The V-POSSUM predicted mortality was somewhat less accurate: AUC 0.69 (SD 0.05; 95% CI 0.60-0.77; P<0.001). The only pre-operative variables independently predicting 30-day mortality were age (P=0.006) and lowest pre-operative systolic blood pressure (P=0.032).

Conclusions: Preoperative risk assessment with either V-POSSUM or GAS does not identify patients with fatal outcome at the individual level. However, these risk models may be useful as a relative estimate, in surgical audit or for inter- and intra-hospital comparisons.