S-22-2: PHENOME-WIDE ANALYSIS OF A BLOOD PRESSURE POLYGENIC RISK SCORE

Objective: Blood pressure (BP) is a polygenic trait with > 1400 known single nucleotide polymorphisms (SNPs), and BP polygenic risk scores (PRSs) strongly predict future hypertension. However, 90% of BP SNPs are also associated with other traits, such as anthropometric, biochemical, brain-related, hematological, and lifestyle traits. We conducted a phenome-wide survival analysis to explore the longitudinal associations between systolic BP (SBP) PRS and all major disease end points. Design and method: We used FinnGen study release 7 with 321,302 individuals (mean age at the end of follow-up 59.4 years, 56.2% women). Because Finnish permanent residents are linked to nationwide electronic health care registers, follow-up with virtually no loss to follow-up is possible for all major disease end points. First, we computed an SBP PRS based on 1,098,015 SNPs using PRS-CS and independent summary statistics from the UK Biobank. Then, using Cox models, we estimated the association between the SBP PRS and 2,350 incident disease end points across all ICD-10 categories and major medication classes. We considered a P value < 10 –5 significant. Results, and Conclusions: In total, we observed 157 phenome-wide significant associations between SBP PRS and incident disease end points (Figure). As expected, circulatory disease phenotypes had the highest number of significant associations (73), followed by endocrine and metabolic diseases (32), obstetric diseases (10), and respiratory diseases (8). Overall, 14 out of 17 ICD-10 categories had at least one phenome-wide significant association, with only neoplasms, ear-related, and perinatal diseases lacking any associations. Our results demonstrate that due to the high pleiotropy in BP genetics, SBP PRS is associated with a wide range of outcomes besides cardiovascular disease. The mechanisms underlying these links warrants further investigation. More broadly, our study also underlines the potential for PRSs from high-powered genome-wide association studies of pleiotropic traits to be utilized in the prediction of other incident diseases.