Risk factors for ischemic stroke in cold versus non-cold regions of China: a systematic review and meta-analysis

INTRODUCTION: Ischemic stroke (IS) exhibits pronounced geographic variation across China, with a consistently greater burden observed in colder regions. However, whether the strength of associations between major risk factors and IS differs between cold and non-cold regions remains unclear. METHODS: We systematically searched observational studies in PubMed, Embase, Web of Science, CNKI, Wanfang, VIP, and CBM from inception through September 6, 2025. Data were analyzed using Stata 16.0. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. Regional differences were evaluated using subgroup analysis. RESULTS: Ninety-six studies involving 800,741 participants were included. In the overall analysis, the largest pooled effect estimates were observed for atrial fibrillation (OR 4.48, 95% CI 2.41-8.36, I² = 62.1%), coronary heart disease (OR 3.20, 95% CI 2.29-4.46, I² = 86.2%), and hypertension (OR 3.05, 95% CI 2.55-3.64, I² = 92.9%), all of which were accompanied by moderate to substantial between-study heterogeneity. Subgroup analysis demonstrated significantly larger pooled associations in cold regions for hypertension (OR 4.31 vs. 2.69, I² = 94.8% vs. 92.0%), diabetes mellitus (OR 2.60 vs. 1.77, I² = 71.6% vs. 90.7%), triglycerides (TG; OR 3.41 vs. 1.48, I² = 96.7% vs. 92.4%), and low-density lipoprotein cholesterol (LDL-C; OR 1.91 vs. 1.43, I² = 0.0% vs. 67.2%) (all P < 0.05). In contrast, age was more strongly associated with IS in non-cold regions (OR 1.31 vs. 1.05, I² = 91.4% vs. 91.8%). CONCLUSION: Our analysis identified substantial geographic heterogeneity in the reported associations between selected cardiometabolic risk factors and IS in China. These exploratory, study-level findings generate hypotheses regarding contextual modification of risk factor associations, warranting confirmation in prospective, individual-level studies before informing region-specific prevention strategies.