Modified Harborview Risk Score accurately predicts mortality for patients with ruptured abdominal aortic aneurysm

Abstract

Objective: The modified Harborview Risk Score (HRS) is a simple measure initially derived from a single institutional dataset used to predict ruptured abdominal aortic aneurysm (rAAA) repair survival preoperatively using basic labs and vital signs collected upon presentation. However, validation of this widely applicable scoring system has not been performed. This study aims to validate this scoring system using a large multi-institutional database.

Methods: All patients who underwent repair of an rAAA from 2011 to 2018 in the National Surgical Quality Improvement Program (NSQIP) and at a single academic medical center were included. The modified HRS was calculated by assigning 1 point for each of the following: age >76 years, creatinine >2 mg/dL, international normalized ratio >1.8, and any systolic blood pressure less than 70 mmHg. Assessment of the prediction model was then completed. Using a primary outcome measure of 30-day mortality, the receiver operating characteristic area under the curve was calculated. The discrimination between datasets was compared using a Delong test. Mortality rates for each score were compared between datasets using the Pearson χ² test. Comparative analysis for patients with a score of 4 was limited due to a small sample size.

Results: A total of 1536 patients were identified using NSQIP, and 163 patients were assessed in the institutional dataset. There were 518 patients with a score of 0 (455 NSQIP, 63 institutional), 676 patients with a score of 1 (617 NSQIP, 59 institutional), 391 patients with a score of 2 (364 NSQIP, 27 institutional), 106 with a score of 3 (93 NSQIP, 13 institutional), and 8 patients with a score of 4 (7 NSQIP, 1 institutional). No difference was found in the receiver operating characteristic area under the curves between datasets (P = .78). Thirty-day mortality was 10% NSQIP vs 22% institutional for a score of 0; 28% NSQIP vs 36% institutional for a score of 1; 41% NSQIP vs 44% institutional for a score of 2; 45% NSQIP vs 69% institutional for a score of 3; and 57% NSQIP vs 100% institutional for a score of 4. Score 0 was the only score with a significant mortality rate difference between datasets (P = .01).

Conclusions: The modified HRS is confirmed to be broadly applicable as a clinical decision-making tool for patients presenting with rAAAs. Therefore, this easily applicable model should be applied for all patients presenting with rAAAs to assist with provider and patient decision-making prior to proceeding with repair.

Keywords: Risk score; Ruptured aortic aneurysm.