Abstract

Background: The non-high-density lipoprotein cholesterol (non-HDL-C) / high-density lipoprotein cholesterol (HDL-C) ratio has emerged as a marker of residual vascular risk; however, its role in patent foramen ovale (PFO)-associated cryptogenic stroke (CS) remains unclear. We investigated the association between the non-HDL-C/HDL-C ratio and CS in patients with PFO and evaluated its incremental predictive value beyond the Risk of Paradoxical Embolism (ROPE) score. Methods: This retrospective study included 316 patients with confirmed PFO, including 56 patients with CS. Multivariable logistic regression, restricted cubic spline analysis, ROC analysis, net reclassification improvement (NRI), integrated discrimination improvement (IDI), decision curve analysis, and bootstrap internal validation were performed. Results: Patients with CS had significantly higher non-HDL-C/HDL-C ratio levels than those without CS (p < 0.001). In multivariable analysis, the non-HDL-C/HDL-C ratio remained independently associated with CS (OR: 1.881, 95% CI: 1.310–2.700, p < 0.001). Restricted cubic spline analysis demonstrated a significant nonlinear association between the non-HDL-C/HDL-C ratio and CS risk (overall p = 0.001; nonlinear p = 0.015). Addition of the non-HDL-C/HDL-C ratio to the ROPE score improved discrimination, increasing the AUC from 0.781 to 0.819 (DeLong p = 0.010), and significantly improved risk reclassification (continuous NRI: 0.555, p = 0.002; IDI: 0.057, p = 0.002). Internal validation demonstrated stable model performance with minimal optimism. Conclusions: The non-HDL-C/HDL-C ratio was independently associated with CS and demonstrated potential incremental predictive value beyond the ROPE score in patients with PFO. These findings suggest that metabolic lipid burden may contribute to thromboembolic susceptibility and may improve individualized risk stratification in PFO-related stroke.

IPC Classification

Keywords