代谢与胱抑素C对大动脉粥样硬化型缺血性 卒中风险的因果影响——一项两样本孟德尔随机化研究

Causal Effects of Metabolic and Renal Function-Related Biomarkers on the Risk of Large-Artery Atherosclerotic Stroke—A Two-Sample Mendelian Randomization Study

摘要: 目的：利用两样本孟德尔随机化(MR)方法，探讨胱抑素C、载脂蛋白A (ApoA)及谷氨酰转肽酶(GGT)与大动脉粥样硬化型缺血性卒中(LAA)之间的潜在因果关系。方法：本研究基于欧洲人群的全基因组关联研究(GWAS)数据构建两样本MR分析。分别选取与胱抑素C、ApoA及GGT相关的单核苷酸多态性(SNPs)作为工具变量。结局数据LAA的GWAS数据来源于欧洲人群的荟萃分析(4373例病例)。在全基因组显著性阈值(P < 5 × 10⁻⁸)下筛选，并进行连锁不平衡(r² < 0.001，距离 > 10,000 kb)剔除弱工具变量检验(F > 10)。采用逆方差加权(IVW)法作为主要因果估计方法，并结合加权中位数法、加权模式法、简单中位数法及MR-Egger回归进行敏感性分析。通过MR-Egger截距、MR-PRESSO及Cochran Q检验评估水平多效性与异质性。结果：IVW分析显示，遗传预测的ApoA水平升高与LAA风险显著降低相关(OR = 0.838, 95%CI 0.736~0.954, P = 0.0076)，多种MR方法方向一致，提示ApoA对LAA具有保护性因果作用。遗传预测的GGT水平升高显著增加LAA风险(OR = 1.233, 95%CI 1.092~1.393, P = 0.0007)，且在多种MR方法中结果一致，未发现显著多效性或异质性，遗传预测的胱抑素C水平与LAA风险呈正相关(OR = 1.145, 95%CI 1.015~1.291, P = 0.0278)。MR-Egger截距未发现明显方向性水平多效性。MR-PRESSO校正异常SNP后，因果效应方向与原始分析一致，表明结果稳健。结论：本研究的两样本孟德尔随机化分析表明，ApoA对LAA具有因果保护作用，GGT、胱抑素C升高可能增加LAA风险。上述结果为代谢及肾功能指标在LAA发生发展中的作用提供了遗传学因果证据。

Abstract: Objective: To investigate the potential causal relationships of cystatin C, apolipoprotein A (ApoA), and gamma-glutamyl transferase (GGT) with large-artery atherosclerotic ischemic stroke (LAA) using a two-sample Mendelian randomization (MR) approach. Method: This two-sample MR study was based on genome-wide association study (GWAS) data from European populations. Single-nucleotide polymorphisms (SNPs) associated with cystatin C, ApoA, and GGT were selected as instrumental variables. GWAS summary statistics for LAA were obtained from a meta-analysis of European ancestry populations (4373 cases). SNPs were selected at the genome-wide significance threshold (P < 5 × 10⁻⁸), and linkage disequilibrium was removed (r² < 0.001, distance > 10,000 kb). Weak instruments were excluded using an F-statistic threshold (F > 10). The inverse-variance weighted (IVW) method was used as the primary causal estimator, supplemented by weighted median, weighted mode, simple median, and MR-Egger regression for sensitivity analyses. Horizontal pleiotropy and heterogeneity were assessed using the MR-Egger intercept, MR-PRESSO, and Cochran’s Q test. Results: IVW analysis showed that genetically predicted higher ApoA levels were significantly associated with a reduced risk of LAA (OR = 0.838, 95% CI 0.736~0.954, P = 0.0076). The direction of effect was consistent across multiple MR methods, suggesting a protective causal role of ApoA against LAA. Genetically predicted higher GGT levels were significantly associated with an increased risk of LAA (OR = 1.233, 95% CI 1.092~1.393, P = 0.0007), with consistent results across different MR methods and no evidence of significant pleiotropy or heterogeneity. Genetically predicted higher cystatin C levels were also positively associated with LAA risk (OR = 1.145, 95% CI 1.015~1.291, P = 0.0278). The MR-Egger intercept indicated no evidence of directional horizontal pleiotropy. After correction for outlier SNPs using MR-PRESSO, the direction of causal effects remained consistent with the original analysis, indicating robust results. Conclusions: This two-sample MR analysis provides genetic evidence that ApoA has a causal protective effect against LAA, whereas elevated GGT and cystatin C levels may increase the risk of LAA. These findings offer causal support for the roles of metabolic and renal function biomarkers in the development and progression of large-artery atherosclerotic ischemic stroke.