Coronary artery calcium score and measures of prediabetes for cardiovascular risk prediction in patients referred for coronary CT angiography

Background Coronary artery calcium score (CACS) is a key marker of cardiovascular (CV) risk, reflecting subclinical atherosclerosis and predicting future CV outcomes. However, metabolic factors such as prediabetes also contribute to CV risk but are not captured by CACS alone. Incorporating prediabetes markers, including glycated hemoglobinA1c (HbA1c) and impaired glucose tolerance (IGT), may enhance risk stratification and improve CV outcome prediction. Purpose This study evaluates the prognostic value of combining models of CACS with prediabetes markers and type 2 diabetes mellitus (T2DM) for prediction of CV outcomes in patients with suspected coronary artery disease (CAD). Additionally, we compare the predictive performances of the models to determine whether this approach improves risk classification. Methods This prospective study included 723 patients with suspected CAD and no known CV disease, who were referred for coronary computed tomography angiography. CACS was measured and categorized using a cut-off of 100. Prediabetes was defined using two markers: HbA1c (39-47mmol/L) and IGT assessed by an oral glucose tolerance test. Cox regression analysis evaluated the association between CACS ≥100, alone and in combination with measures of prediabetes, for prediction of incident CV outcomes. Risk prediction models based on CACS were compared to models incorporating HbA1c, IGT, and T2DM, using Harrell’s C-index and area under the curve (AUC) to evaluate predictive performance. The CV outcome comprised cardiovascular death, acute myocardial infarction, stroke, coronary revascularization procedures, and heart failure diagnosis. Results Over a median follow-up of 57 months, 84 patients (11.6%) experienced a CV event. In adjusted models, CACS ≥100 was associated with a 1.88-fold increased risk of CV events (adjusted HR (aHR) = 1.88, 95% CI: 1.01–3.49). Adding prediabetes markers to CACS ≥100 further increased risk, with aHRs of 2.74 (95% CI: 1.12–6.72) for HbA1c and 3.12 (95% CI: 1.28–5.59) for IGT. Adding T2DM to the CACS model increased CV risk fourfold, while T2DM patients with CACS <100 had a risk similar to those with normal glucose tolerance. The CACS ≥100 + IGT model had the highest predictive performance (C-index = 0.76, p < 0.0001), while the CACS ≥100 + HbA1c model had a C-index of 0.73, both outperforming CACS ≥100 alone(C-index 0.61). Similarly, AUC improved from 0.62 (CACS alone) to 0.71 (CACS + HbA1c) and 0.75 (CACS + OGTT), indicating better risk discrimination. Conclusions Incorporating prediabetes markers to the traditional CACS-based risk stratification improved CV risk prediction, demonstrating superior predictive performance compared with CACS alone. Suggesting that measurement of prediabetes should be performed when CACS ≥100, to further help identify those at a higher risk of adverse CV outcomes.