Development and validation of a risk prediction model for abdominal aortic aneurysm: a nationwide population-based cohort study

Abstract

Abdominal aortic aneurysm (AAA) is characterized by irreversible localized dilatation of the abdominal aorta. It poses a significant health risk. As AAA size tends to increase over time, there is a heightened risk of rupture, resulting in a substantially high mortality rate. Although AAA screening programs targeting specific demographics are available, there is room for improvement in terms of inclusivity and cost-effectiveness. This study aimed to develop a predictive model for AAA occurrence utilizing seven years of data from the Korean National Health Insurance Service database (NHIS). This study utilized NHIS data from 2009 to 2020. A total of 4,234,415 individuals who underwent health examinations in 2009 were identified. After applying exclusion criteria, a total of 3,937,535 individuals were selected. Of them, 70% were used for model development and 30% were used for validation. The mean follow-up duration was 10.11 ± 1.29 years, during which 2,836 cases of AAA were identified among 1,181,131 (2.4%) participants in the validation cohort. The model incorporated a set of 10 variables, encompassing age, sex, obesity, smoking, drinking, diabetes (DM), hypertension (HTN), dyslipidemia, chronic kidney disease (CKD), and cardiocerebrovascular disease (CVD). Evaluation of the model's predictive performance revealed an area under the curve (AUC) of 0.807 (95% CI: 0.80-0.81) when it was applied to the development cohort. The AUC remained high at 0.803 (95% CI: 0.79-0.81) when the model was applied to the validation cohort, indicating its effectiveness in forecasting AAA occurrence. A multivariable risk model for predicting the onset of AAA was successfully developed, showcasing an excellent performance with an AUC value of 0.807, surpassing traditional screening methods. This model has the potential to selectively identify high-risk patients from a slightly broader pool than current screening approaches. Priority should be given to proactive screening efforts targeting individuals at elevated risk for AAA, with the goal of reducing AAA-related mortality.

Keywords: Abdominal; Aortic aneurysm; Nomograms.

Fig. 1

Enrollment flow chart. AAA: Abdominal aortic aneurysm.

Fig. 2

A nomogram for predicting the 5-year probability of abdominal aortic aneurysm occurrence. Note: The 10 variables—age, sex, obesity, smoking status, drinking, presence of diabetes mellitus or hypertension, chronic kidney disease, cardiocerebrovascular disease, and total cholesterol level—were each assigned scores ranging from 0 to 100. The corresponding score for each variable can be determined by drawing a straight line to the scoring axis. The total score, calculated as the sum of the scores for all variables, ranges from 0 to 226 and is displayed at the bottom of the nomogram. DM: Diabetes mellitus, IFG: Impaired fasting glucose, HTN: Hypertension, CKD: Chronic kidney disease, CVD: Cardiocerebrovascular disease.

Fig. 3

The model’s receiver operating characteristic curve (ROC) when using development and validation cohorts. Note: A ROC curve is using development cohort, and B ROC curve is using validation cohort.

Fig. 4

Predicted incidence rate (per 1 000 person-years) using development and validation cohorts.