HbA1c-to-HDL-C ratio and risk of incident cardiovascular disease among older adults: evidence from a 7-year prospective cohort study in southwestern China

Abstract

Cardiovascular disease (CVD) remains the leading cause of death worldwide, particularly among older adults. The glycated hemoglobin A1c-to-high-density lipoprotein cholesterol (HbA1c/HDL-C) ratio is an emerging biomarker reflecting glucolipid metabolic derangement. However, prospective evidence regarding its association with incident CVD among community-dwelling older adults in southwestern China remains limited. This study aimed to investigate the association between the HbA1c/HDL-C ratio and incident CVD, characterize its dose–response relationship, and evaluate its incremental predictive value beyond conventional cardiovascular risk factors in an older adult population. This prospective cohort study included 1,875 community-dwelling adults aged ≥60 years without baseline CVD from the Early Screening and Intervention Program for High-Risk Populations of Cardiovascular and Cerebrovascular Disease Risk Factors in Wanzhou District. Participants were followed during 2024. The HbA1c/HDL-C ratio was calculated from baseline measurements and analyzed both continuously and as quartiles. Incident CVD events were ascertained using a dual-source validation approach that combined standardized questionnaire data with linkage to the regional surveillance system. Multivariable Cox proportional hazards models, restricted cubic spline (RCS), time-dependent area under the receiver operating characteristic curve (AUC), C-statistic and decision curve analysis (DCA) were fitted to estimate the association between HbA1c/HDL-C and CVD risk. Subgroup, sensitivity analyses were performed to verify robustness. During a median follow-up of 5.08 years, 207 participants developed incident CVD, with a cumulative incidence of 11.04%. After fully adjusting for confounding factors, participants in the highest quartile (Q4) of the HbA1c/HDL-C ratio exhibited a significantly elevated risk of incident CVD compared with the lowest quartile (Q1) (HR = 2.111, 95% CI : 1.276–3.170, P = 0.003), whereas the linear trend across quartiles did not reach statistical significance ( P for trend = 0.068). RCS analysis revealed a nonlinear dose-response relationship ( P for overall < 0.001, P for nonlinear = 0.017), with a data-driven inflection point at an HbA1c/HDL-C ratio of 4.83 (95% CI : 3.55–6.06). Below this threshold, each unit increment was associated with a 44.10% increase in CVD risk (HR = 1.441, 95% CI : 1.138–1.825, P = 0.002). Incorporation of the HbA1c/HDL-C ratio into a conventional risk factor model enhanced the time-dependent AUC (test-set C-statistic: 0.795 vs. 0.777). DCA demonstrated enhanced net benefit, particularly during mid-to-long-term follow-up. Subgroup and sensitivity analyses verified the robustness. After multivariable Cox regression adjustment, participants in the highest quartile (Q4) of the HbA1c/HDL-C ratio showed a significant association with incident CVD, while the linear trend across Q1 to Q4 was borderline significant. RCS analysis further identified a nonlinear dose–response relationship between HbA1c/HDL-C ratio and incident CVD. Given its accessibility, low cost, and ease of calculation, the HbA1c/HDL-C ratio may serve as a complementary indicator for community-based CVD risk stratification, particularly in resource-constrained primary care settings. The apparently stronger association observed should be considered exploratory and requires confirmation in larger prospective cohort studies.