This preprint reports an analysis of the contribution of inherited polygenic risk to residual vascular risk among stroke survivors. The investigators used large-scale polygenic risk scores to test whether genetic predisposition to coronary artery disease (CAD) and stroke predicts future major adverse cardiovascular events and relates to control of vascular risk factors after an index stroke. The authors note the work is a preprint and has not been certified by peer review.
The analysis included 2,701 stroke survivors from the UK Biobank with a mean age of 59.8 ± 7.1 years; 61.7% were male. Two meta-polygenic risk scores (metaGRS) were constructed from published genome-wide association study summary data using penalized regression methods: a stroke-metaGRS comprising approximately 3.2 million genetic variants and a CAD-metaGRS comprising approximately 1.7 million variants.
The primary outcome was incident major adverse cardiovascular events (MACE). Secondary outcomes were recurrent stroke and measures of baseline vascular risk-factor control, specifically glycemic control (HbA1c), LDL cholesterol, and blood pressure. Associations between metaGRS and incident outcomes (MACE, recurrent stroke) were tested with Cox proportional hazards models. Associations with baseline risk-factor control were evaluated using logistic regression. Mediation analyses were performed to quantify indirect effects of metaGRS on MACE via HbA1c, LDL cholesterol, and blood pressure.
The primary end point, incident MACE, and the secondary end point, recurrent stroke, were ascertained over up to 12 years of follow-up. The authors aimed to determine whether higher polygenic risk for CAD or stroke predicts these recurrent vascular events and whether genetic risk relates to suboptimal control of major vascular risk factors at baseline.
During the 12-year observation period, there were 731 MACE events (27.1% of the cohort) and 351 recurrent stroke events (13.0%). The CAD-metaGRS was independently associated with future MACE after adjustment for age and sex: hazard ratio per standard deviation increment 1.15 (95% CI 1.07–1.24; p < 0.001). Details on multivariable models beyond age and sex adjustment, such as additional clinical covariates, were not reported in the abstract.
Both higher stroke-metaGRS and higher CAD-metaGRS were associated with poorer glycemic control as measured by HbA1c. In addition, a higher CAD-metaGRS was associated with poorer lipid control (LDL cholesterol). Associations with blood pressure control were evaluated for mediation analysis, but specific effect sizes or odds ratios for baseline risk-factor associations were not reported in the abstract.
Mediation analyses identified glycemic regulation as a significant pathway linking polygenic risk scores to recurrent vascular events. In other words, part of the association between polygenic risk and MACE appears to operate indirectly through poorer glycemic control. The abstract reports this mediation finding but does not provide the proportion of effect mediated or detailed mediation statistics in the summary.
The authors conclude that polygenic risk scores for cardiovascular disease are associated with recurrent vascular events and with control of key vascular risk factors among stroke survivors. They suggest these findings may offer potentially actionable insights for secondary prevention, but emphasize that further investigation is warranted. As this report is a medRxiv preprint, the results should not be used to guide clinical practice until peer review and further validation are complete.
Ethical approval and participant consent were obtained through the UK Biobank; data for this work were accessed under UK Biobank application number 151281. The authors declare funding from multiple sources including the German Research Foundation, Munich Cluster for Systems Neurology, Collaborative Research Center 1744, Fritz-Thyssen Foundation, Hertie Foundation, and an LMU-TAU research cooperation program. One author reported consulting fees unrelated to this work; other authors reported no conflicts. The manuscript states that all data produced are contained in the manuscript.