Diversity in genetic risk of recurrent stroke: a genome-wide association study meta-analysis

Stroke is a leading cause of death and disability worldwide. Recurrent strokes are seven times more lethal than initial ones, with 54% leading to long-term disability. Substantial recurrent stroke risk disparities exist among ancestral groups. Notably, Africans face double the risk and higher fatality rates compared to Europeans. Although genetic studies, particularly GWAS, hold promise for uncovering biological insights into recurrent stroke, they remain underexplored. Our study addresses this gap through meta-analyses of recurrent stroke GWAS, considering specific ancestral groups and a combined approach.

We utilized four independent study cohorts for African, European, and Combined ancestry recurrent stroke GWAS with genotyping, imputation, and strict quality control. We harmonized recurrent stroke phenotype and effect allele estimates across cohorts. The logistic regression GWAS model was adjusted for age, sex, and principal components. We assessed how well genetic risk of stroke informs recurrent stroke risk using Receiver Operating Characteristic (ROC) curve analysis with the GIGASTROKE Consortium's polygenic risk scores (PRS).

Harmonization included 4,420 participants (818 African ancestry and 3,602 European ancestry) with a recurrent stroke rate of 16.8% [median age 66.9 (59.1, 73.6) years; 56.2% male]. We failed to find genome-wide significant variants (p < 5e-8). However, we found 18 distinct suggestive (p < 5e-6) genetic loci with high biological relevance consistent across African and European ancestries, including PPARGC1B, CCDC3, OPRL1, and MYH11 genes. These genes affect vascular stenosis through constriction and dilation. We also observed an association with SDK1 gene, which has been previous linked with hypertension in Nigerian and Japanese populations). ROC analysis showed poor performance of the ischemic stroke PRS in discriminating recurrent stroke status (area under the curve = 0.48).

Our study revealed genetic associations with recurrent stroke not previously associated with incident ischemic stroke. We found suggestive associations in genes previously linked with hypertension. We also determined that knowing the genetic risk of incident stroke does currently not inform recurrent stroke risk. We urgently need more studies to understand better the overlap or lack thereof between incident and recurrent stroke biology.