Validation of three models predicting in-hospital death in patients with an abdominal aortic aneurysm eligible for both endovascular and open repair

Abstract

The Medicare, the Vascular Governance North West (VGNW), and the British Aneurysm Repair (BAR) models can be used to predict in-hospital death after an intervention for an asymptomatic abdominal aortic aneurysm (AAA). Validation of these models in patients with suitable aortic anatomy for endovascular repair and a general condition fit for open repair is lacking. We validated the Medicare, VGNW, and BAR models in patients from a randomized controlled trial comparing open and endovascular AAA repair. A per-protocol analysis was done of 345 Dutch and Belgian patients with in-hospital death as the primary end point. The prediction models were validated taking into account discrimination (the ability to distinguish between death and survival) and calibration (the agreement between predicted and observed death rates). Discrimination was assessed using the area under the receiver-operating characteristics curve (AUC). An AUC >0.70 was considered to be sufficiently accurate. Calibration was assessed using the Hosmer-Lemeshow (HL) test, and P > .05 was considered to be sufficiently accurate. The AUC was 0.77 (95% confidence interval [CI], 0.64-0.90; HL test, P= .52) for the Medicare model, 0.88 (95% CI, 0.81-0.95; HL test, P= .31) for the VGNW model, and 0.79 (95% CI, 0.67-0.91; HL test, P= .15) for the BAR model. In AAA patients eligible for endovascular and open repair, the predictions of in-hospital death by the Medicare, VGNW, and BAR models were sufficiently accurate. Therefore, these models can be used to support deciding between endovascular and open repair.