Development and Validation of a Predictive Model to Aid in the Management of Intact Abdominal Aortic Aneurysms

Abstract

Objective/Background: Predicting outcomes prior to elective abdominal aortic aneurysm repair (AAA) requires critical decision making, as the treatment offered is a prophylactic procedure to prevent death from a ruptured AAA. The aim of this work was to develop and validate a model that may predict outcomes for patients with an AAA and hence aid in clinical decision making. Methods: A discrete event simulation model was built to simulate the natural history of a patient with an AAA and to predict the 30 day and 2–5 year survival of patients undergoing treatment and surveillance. The input parameters of AAA behavior and impact of comorbidities on survival were derived from the published literature and the New Zealand national life tables. The model was externally validated using a cohort of patients that underwent AAA repair (n = 320) and a cohort of patients undergoing small AAA surveillance (n = 376). All patients had completed at least 5 years of follow up. Results: The model was run three times for each data set to test. This produced a SD < 1%, indicating excellent reproducibility. The observed 30 day mortality for the patients undergoing AAA repair was 9/320 (2.8%) and the expected (model predicted) mortality was 3.8% (c-statistic 0.87 [95 confidence interval 0.75–1.0]). The c-statistic for the predicted 2–5 year survival ranged from 0.68 to 0.71 for the repaired AAA cohort and 0.69 to 0.73 for patients with a small AAA on surveillance. Conclusion: The AAA clinical decision tool has the ability to accurately predict the 5 year survival of patients with an AAA. This tool can be used during clinical decision making to better inform clinicians and patients of long-term outcomes. Further validation studies in a wider AAA population are required to test the broader clinical utility of this AAA clinical decision tool.