Abstract 4367664: Glycated Hemoglobin Predicts the Presence of High-Risk Coronary Plaque on Coronary Computed Tomographic Angiography (CCTA) in Patients without Diabetes

Abstract

Introduction: A substantial risk gradient for incident cardiovascular disease has been seen across glycated hemoglobin (HbA 1c ) levels below diabetic and pre-diabetic thresholds, highlighting the need for better risk prediction across the glycemic spectrum. Glycemia is associated with greater non-calcified plaque, limiting the predictiveness of calcium scores in this population. Generalizable data on the association between HbA 1c , semiquantitative CAD burden, and high-risk plaque (HRP) features in pre and nondiabetics is limited. We hypothesized that higher HbA 1c is associated with a higher comprehensive CCTA risk (Leiden) score and HRP. Methods: Patients undergoing clinically indicated CCTA were prospectively enrolled (2021-2023, N=419) in a study of blood biomarkers predictive of CAD, of whom 347 were nondiabetic. CCTAs were independently analyzed using a 17-segment model and comprehensively phenotyped by two experts in consensus to derive the Leiden score, determine HRP. HbA 1c , measures of insulin resistance and comorbidities were compared across those with or without ≥1 coronary segment with HRP and CAD severity groups [Leiden scores 0-5 (low), 6-20 (moderate) and >20 (high)]. Multivariable logistic regression assessed the association between HbA 1C groups and presence of HRP. Receiver operating characteristic (ROC) analysis and Youden index were used to define an optimal HbA 1c cutoff predictive of HRP. Results: Patients with HRP (93, 26.8%), were older (64±10 years), more hypertensive (HTN, 69%) and dyslipidemic (79%). HRP was associated with more severe CAD (mean Leiden score 17.62 vs. 7.35, P <0.001). Median HbA 1c was significantly higher in those with HRP [5.55, IQR 5.3-5.7 vs. 5.6, IQR 5.5-5.8, P <0.001] (Table 1), but did not vary significantly by CAD severity (Table 2). HbA 1c was associated with a greater risk of HRP in age, sex adjusted models [OR 2.54; 95% CI 1.24-5.35, per unit increase]. An optimal HbA 1c threshold of 5.55 had modest predictive performance for HRP (c-statistic 0.62). HbA 1c ≥ 5.55 was independently predictive of HRP in models adjusted for age, sex, race, smoking, HTN, statin use, and Leiden score [OR 1.79; 95% CI 1.19–2.71]. Conclusion: Higher HbA 1c is independently associated with high-risk plaque in pre- and nondiabetics undergoing clinically indicated CCTA. Further prospective evaluation of prediabetic HbA 1c thresholds predictive of subclinical and incident CAD in larger populations, and their preventive implications is warranted.