GENOME-WIDE META-ANALYSIS AND POLYGENIC RISK SCORE IN A RISK ASSESSMENT TO INTRACEREBRAL HEMORRHAGE

Intracerebral hemorrhage (ICH), defined as bleeding within the brain parenchyma, is the second most common cause of stroke. We conducted a meta-analysis combining race-specific genome-wide association studies (GWASs) to identify genetic variants associated with ICH risk. The analysis included 141 ICH cases and 8,990 controls from Korean cohorts, along with 2,700 ICH cases and 396,624 controls from the UK Biobank (UKB). Meta-analysis was performed using fixed- and random-effect models with inverse-variance weighting. Candidate loci were further evaluated using approximately 450,000 whole-exome sequencing (WES) data. We additionally developedm weighted polygenic risk scores (wPRS) based on feature-selected variants and assessed performance using the area under the receiver operating characteristic curve (AUROC). Among 2,028,834 shared variants, six variants in the 19q13.32 region reached genome-wide significance (p<5×10⁻⁸) in the multi-ethnic meta-analysis. The strongest signal was identified at the APOE p.C130R mutation (rs429358), showing a pooled odds ratio (pOR) of 1.21 (p = 4.57×10⁻¹⁰). This association was replicated in European-ancestry WES data (OR=1.24, p=1.61×10⁻⁸) and showed a nominal effect in Asian-ancestry samples. Another variant, rs5117 in APOC1, also demonstrated significant association (pOR=1.20, p=2.60×10⁻⁹). Fifteen additional variants across CTNNA2, ARFGEF1-CPA6, GRIN2B, TOMM40, PVRL2, and APOE showed suggestive significance (p<1×10⁻⁵). A wPRS model using multiple markers achieved an AUROC of 0.82 across populations. While the combined approach increases statistical power and highlights the relevance of the 19q13.32 locus, further improvements in analytical models are needed to clarify the polygenic architecture of ICH