Disease-wide Association Results

Disease phenotypes were pre-defined by FinnGen study1 using ICD-10 code and included diseases screened to ensure they had a minimum number of 300 cases after merging the participants with metabolome data available.

Following rigorous quality control, 527 prevalent and 859 incident diseases were identified based on the date of initial diagnosis. Incident diseases were tracked over a median follow-up period of 14.9 years, concluding in November 2023. A comprehensive list of these diseases and their summaries is available for download.

For cross-sectional analysis, odds ratios derived from logistic regressions were calculated to explore the associations between metabolites and prevalent diseases. Hazard ratios from Cox proportional hazard regressions were also derived for incident diseases.

Both sets of associations were adjusted for participants‘ baseline characteristics, including age, sex, Townsend deprivation index, Body Mass Index (BMI), smoking status, statin medication use, and fasting duration prior to blood collection. Subgroup analyses were conducted based on sex and age (mid-aged: <60 years and elderly: ≥60 years). Furthermore, we performed additional sensitivity analysis by adjusting additional covariate of eGFR. Replication analysis in White and non-White ancestries were performed to validate the findings. Results can be accessed through downloaded data files. Detailed data partition

Analysis were performed through statsmodels (v0.13.1) and lifelines (v0.27.4) packages under Python (v3.9.6).