Objective scoring systems of medical risk: A clinical tool for selecting patients for open or endovascular abdominal aortic aneurysm repair

Abstract

Objective scoring systems have been developed for risk stratification of open infrarenal aneurysm repair. To date, none have been applied for the selection of patients who would most benefit from either an open or an endovascular approach. This study assessed the utility of comorbidity-based objective scoring systems for defining subgroups of patients who might most benefit from open or endovascular aneurysm repair. A retrospective database review was performed for the period January 1999 to December 2004 to identify patients who had undergone elective open aneurysm repair (open repair) or elective endovascular aneurysm repair (EVAR). Validation of the Glasgow Aneurysm Score (GAS), the Modified Leiden Score (M-LS), and the Modified Comorbidity Severity Score (M-CSS) was performed for perioperative mortality risk in the open repair group. GAS, M-LS, and M-CSS were then calculated for the EVAR group. Differences in open repair vs EVAR mortalities were evaluated. During the time period, 558 patients underwent open repair and 304 underwent EVAR. Overall mortality was 4.7% for open repair patients and 2.0% for EVAR. All three scoring systems were validated to our open repair data set (C statistic: GAS, 0.72; M-LS, 0.71; M-CSS, 0.74). A score was calculated for each system that separated patients into groups of either low or high risk of death for open repair. This score (cut point) was 76.5 for the GAS, 5.2 for the M-LS, and 8 for the M-CSS. Analysis of the EVAR population revealed that patients at low medical risk for open repair did not derive statistically significant mortality benefit with EVAR; however, patients at high medical risk for open repair derived significant benefit from EVAR (GAS>76.5 mortality: open repair, 7.8%; EVAR, 1.9% [P<.01]; M-LS mortality: open repair, 8.1%; EVAR, 2.5% [P<.01]; and M-CSS mortality: open repair, 10.3%; EVAR, 3.4% [P<.025]). Despite a very small number of deaths (n=6), receiver operator curve analysis identified M-LS and M-CSS as having some predictive ability for mortality risk with EVAR (C statistic: M-LS, 0.70; M-CSS, 0.69). Three validated objective scoring systems can be used to categorize patients into two groups of medical risk: one that has excellent outcome with open repair and derives no early mortality benefit from EVAR, and another that has significant mortality with open repair and derives important benefit with EVAR.