Association between remnant cholesterol inflammatory index and new-onset stroke in different glycemic metabolic states: insights from the China health and retirement longitudinal study

Abstract

The remnant cholesterol inflammatory index (RCII) has been proposed as a novel biomarker for lipid dysregulation and inflammation, and has also been suggested as a potential risk factor for diabetes. However, the association between RCII and the incidence of new-onset stroke across different clinical populations, particularly in individuals with varying glycemic metabolic states (GMS), remains unclear. This study included 8,310 participants aged ≥ 45 years without a history of stroke from the China Health and Retirement Longitudinal Study (CHARLS). RCII was calculated as [RC (mg/dL) × hs-CRP (mg/L)] / 10. Participants were categorized into four groups according to baseline RCII levels: Q1 (≤ 0.779), Q2 (0.779–1.950), Q3 (1.950–5.250), and Q4 (> 5.250). The primary endpoint was the occurrence of incident stroke. Kaplan-Meier curves, multivariable Cox proportional hazards models, receiver operating characteristic (ROC) analysis, and restricted cubic splines (RCS) were applied to evaluate the association between RCII and stroke risk across GMS. During a median follow-up of 9 years, 771 (9.28%) participants experienced a first stroke event. Stroke risk increased progressively across RCII quartiles. Kaplan-Meier analysis showed significant differences in the occurrence of stroke in the overall population and in all glycemic subgroups, including normal glucose regulation (NGR), prediabetes mellitus (Pre-DM) and diabetes mellitus (DM) (all log-rank P < 0.0001). Compared to Q1, participants in Q4 showed a 66.9% higher risk of stroke (HR 1.669, 95% CI 1.343, 2.075). Formal interaction testing revealed that glycemic status significantly modified the association ( P = 0.047). Specifically, the hazard ratios for stroke in Q4 were 1.755 (95% CI 1.249, 2.486) in the NGR group and 1.708 (95% CI 1.220, 2.427) in the Pre-DM group, whereas the association was not statistically significant in the DM group (HR 1.158, 95% CI 0.684, 1.961). RCS demonstrated a non-linear association between RCII and stroke risk in the overall population, NGR and Pre-DM groups. ROC analysis revealed that RCII had the highest discriminatory ability (AUC 0.594) in the NGR group, followed by the pre-DM (AUC 0.578) and DM (AUC 0.547) groups. The addition of RCII to the baseline clinical model yielded a marginal improvement in the C-index (0.700 to 0.705). Sensitivity analyses confirmed the robustness of the findings. Elevated RCII was independently associated with an increased stroke risk, with significant heterogeneity across glycemic states. The association was stronger and exhibited a non-linear pattern in individuals with NGR and Pre-DM, but not in those with DM. These findings suggest that RCII may serve as an integrated early metabolic marker for stroke risk stratification and early prevention.