Targeting High-Risk Populations for Abdominal Aortic Aneurysm Screening: A Sex-Stratified Analysis in an Ethnicity Diverse Cohort Incorporating Machine Learning Tools

Background Abdominal aortic aneurysm (AAA) is usually asymptomatic, but rupture carries up to 90% mortality. Ultrasound screening reduces rupture-related mortality in ever-smoking men older than 65. Women have 4-fold lower prevalence but worse outcomes and were markedly underrepresented in major AAA trials. We analyzed a nationwide database to identify high-risk subgroups. Methods We retrospectively analyzed individuals aged ≥50 years who underwent ultrasound screening for AAA from the nationwide Clalit-Health-Services database (2013–2023). Sex-stratified logistic regression and machine learning models were applied to identify high-risk groups. Results A total of 70,149 individuals underwent screening, including 32,269 (46%) of Middle Eastern or North African descent and 24,776 (35.3%) women. AAA prevalence was 9.5% in men and 3.2% in women. In women, AAA was associated with age >75 (odds ratio [OR]: 2.22, 95% confidence interval: 1.74–2.84, P < 0.001), ever smoking (OR: 3.08, 2.63–3.60, P < 0.001), and high CHA₂DS₂-VASc >4 (OR: 3.27, 2.25–4.74, P < 0.001). In men, age >75 (OR: 1.93, 1.71–2.18, P < 0.001), smoking (OR: 1.52, 1.38–1.67, P < 0.001), and high CHA₂DS₂-VASc (OR: 1.84, 1.65–2.06, P < 0.001) were significant. Ethnicity modestly increased AAA odds in men. Consistent with logistic regression, machine learning-based probability curves demonstrated higher predicted AAA risk with advancing age, higher CHA₂DS₂-VASc scores, smoking, and peripheral arterial disease, particularly among women >75 years. Conclusion Our findings indicate that AAA screening should consider age, smoking status and cardiovascular risk factors, including in women, where higher ORs were observed.