Novel Predictors of Peri-Operative Mortality in a Series Of 931 Consecutive Patients with Intact Aortoiliac Aneurysms Managed at a Single Centre

Abstract

Introduction - The aim of our study was to identify preoperative and intraoperative factors associated with in-hospital mortality after repair of intact abdominal aortic or iliac aneurysms. Methods - This was a retrospective analysis of prospectively collected information on patients undergoing repair of intact aortoiliac aneurysms during a 28-year period. Recorded information included preoperative demographics, risk factors and co-morbidities, aneurysm characteristics, investigations and operative variables including special aneurysm presentation (inflammatory, mycotic/infected, aortocaval fistula) and procedure type (open vs. endovascular repair). Univariate and multivariate logistic regression analyses (with odds ratios, OR) were performed, the latter performed to develop a predictive model using the generated output. The area under the curve of the receiver operating characteristics (ROC) curve of our score, the Glasgow aneurysm score (GAS),1 and the Vascular Study Group of New England (VSGNE) score,2 was calculated. Results - During the study period we operated on 931 patients with intact aortoiliac aneurysms, on an elective (n=882) or urgent (n=49) basis, and an in-hospital mortality of 1.7% and 8.2%, respectively (p=0.015). Endovascular aneurysm repair (EVAR, n=414) was a predictor of lower mortality (0.5% vs. 3.3% for open repair, n=517, p=0.003). Other significant predictors included the presence of abdominal or back pain (7.5% vs 1.3% for no pain, p<0.001), preoperative angiography (6.9% vs 1.6% for no angiography, p=0.01), special aneurysm presentation (10.7% vs 1.5% for 875 ordinary atherosclerotic aneurysms, p<0.001), concomitant major procedures (19% vs 1.6% for the rest, p<0.001), serious intraoperative complications (9.1% vs 1.5% for no such complications, p<0.001), median number of transfused units of blood intraoperatively (2 and 0 for cases with and without mortality, respectively, p<0.001) and duration of the procedure (300min and 150min for cases with and without mortality, respectively, p<0.001). On multivariate analysis, EVAR was an independent predictor of lower mortality (OR=0.22, p=0.05) and special aneurysm presentation (OR=6.45, p=0.001) and concomitant major procedures (OR=14.3, p<0.001) were independent predictors of higher mortality. The area under the ROC curve, and p values for the GAS, VSGNE score and our score was 0.46, 0.62 and 0.81, and 0.59, 0.084 and 0.00003, respectively (Figure). The difference between the VSGNE and our score was significant (p=0.029). Repeat analysis for electively operated cases demonstrated an area under the ROC curve of 0.66 (p=0.029) and 0.80 (p=0.000069), respectively, for the VSGNE and our score, with the difference becoming non-significant. Conclusion - Our study in patients undergoing aortoiliac aneurysm repair has demonstrated novel risk factors for mortality and externally validated VSGNE in different healthcare settings.