Prediction model for aortic dissection, aortic aneurysm, and peripheral artery disease

Abstract Aortic dissection (AD), aortic aneurysm (AA), and peripheral artery disease (PAD) are associated with higher mortality. However, validated prediction models for these conditions remain scarce. We developed risk prediction models for AD, AA, and PAD using routine health check-up data. We used data from the DeSC database, encompassing 1,082,369 participants aged 20–74 years without prior cardiovascular disease for model derivation. Study participants were randomly split into derivation (50%) and internal validation (50%) cohorts. The primary outcomes were the incidence of AD, AA, and PAD. Flexible parametric survival models were used to estimate 5-year risk using routine health check-up data, including age, sex, body mass index, blood pressure, lipid profile, glucose, smoking status, physical activity, and medication use. During follow-up, 756 AD, 2,230 AA, and 4,131 PAD events occurred. In the internal validation cohort, Harrell’s C-index was 0.781 (95% confidence interval: 0.761 to 0.801) for AD, 0.812 (0.799 to 0.825) for AA, and 0.701 (0.690 to 0.711) for PAD. The Royston D statistic was 1.695 (1.533 to 1.857) for AD, 1.975 (1.874 to 2.075) for AA, and 1.222 (1.152 to 1.292) for PAD. The models had good calibration for each outcome. Risk prediction equations were developed and validated to estimate risk for the development of AD, AA, and PAD based on routine health check-up data. These models may allow risk stratification to support earlier diagnosis and intervention for vascular disease. Further external validation using datasets from different target populations and clinical settings is warranted.