Association of the RC-HbA1c index with incident stroke in middle-aged and older Chinese adults: a nationwide prospective cohort study

Stroke is a major cause of disability and mortality among middle-aged and older adults. Although remnant cholesterol (RC) and glycated hemoglobin (HbA1c) are each associated with vascular risk, the association and incremental predictive value of their combination for incident stroke remain unclear. We analyzed 8,872 stroke-free adults from the China Health and Retirement Longitudinal Study (CHARLS). The RC-HbA1c index was calculated as RC (mg/dL) × HbA1c (%). Cox proportional hazards models, restricted cubic splines, and predictive performance analyses were performed. Cumulative exposure was evaluated in 5,412 participants with measurements at Wave 1 and Wave 3, using Wave 3 as the landmark. During the prespecified 7-year observation period, 468 participants developed stroke. Stroke incidence increased across RC-HbA1c quartiles (3.7%–7.3%, P < 0.001). Compared with the lowest quartile, the highest quartile was associated with increased stroke risk (hazard ratio [HR] = 1.65, 95% CI 1.18–2.31). Restricted cubic spline analysis showed no statistical evidence of departure from linearity. Adding RC-HbA1c to the basic model increased the C-statistic from 0.6584 to 0.6665 and yielded a significant continuous net reclassification improvement (NRI = 0.064, P = 0.016). In direct comparison with RC alone, RC-HbA1c provided a small improvement in discrimination (ΔC = 0.00217, 95% CI 0.00033–0.00454; P = 0.008). However, its discrimination did not differ significantly from a model incorporating RC and HbA1c separately (ΔC = − 0.00141, 95% CI − 0.00572–0.00125; P = 0.428). High cumulative RC-HbA1c exposure was also associated with increased stroke risk (HR = 1.99, 95% CI 1.35–2.93). The RC-HbA1c index was independently associated with incident stroke and provided a small incremental improvement in discrimination beyond conventional risk factors and RC alone. Its performance was not significantly different from that of separate RC and HbA1c predictors. The index may serve as a simple composite marker of lipid-glycemic metabolic status, although its clinical utility and generalizability require further evaluation.