Fig. 1: Workflow of the study.
From: Genetic predictors of lifelong medication-use patterns in cardiometabolic diseases
GWAS were performed for 12 phenotypes of medication-use patterns in treating hyperlipidemia, hypertension and T2D (three continuous analyses of the total number of medication purchases and nine binary analyses of medication changing and fast discontinuation) in FinnGen, with data capturing all medication purchases since 1 January 1995. Meta-analyses of up to 567,671 participants combined data in FinnGen (n = 29,990–193,933), EstBB (n = 5,110–184,892) and UKBB (n = 188,846). Fine-mapping was performed in FinnGen for all associated (P < 5 × 10−8) regions. Genetic architectures between medication-use traits and the underlying cardiometabolic risk were juxtaposed by comparing genome-wide significant associations, calculating LDSC regression genetic correlations and testing associations between PRS for LDL, SBP and T2D and the medication-use phenotypes. A medication-use enhanced multitrait PRS for CAD was built using MTAG method and its performance was compared to a traditional CAD PRS by testing associations with CAD in an independent sample (UKBB, n = 343,676).
