Organ-specific biological age gaps show heterogeneous and direction-dependent genetic relationships with age-related diseases

Organ-specific biological age gap (BAG) traits are model-derived measures of ageing-related variation, but their genetic relationships with age-related diseases and the genomic scale of these relationships remain unclear. We analysed European-ancestry genome-wide association study (GWAS) summary statistics for nine UK Biobank-derived organ-specific BAG traits and 19 age-related disease or disease-related phenotypes using bidirectional two-sample Mendelian randomization (MR). Associations were prioritised using a false-discovery-rate-based evidence framework and further evaluated using complementary MR analyses, statistical power assessment, sample-overlap auditing, colocalisation, multivariable MR, and genome-wide and local genetic-correlation analyses. Among 342 planned comparisons, 306 were estimable and 31 met Tier 1 or Tier 2 prioritisation criteria. Prominent forward associations included renal BAG with cystatin C-based estimated glomerular filtration rate, cardiovascular BAG with coronary artery disease and large artery stroke, and cardiovascular, metabolic, and pulmonary BAGs with type 2 diabetes. In reverse analyses, type 2 diabetes liability was associated with metabolic, cardiovascular, brain, and hepatic BAGs. A pleiotropy-robust MR analysis preserved the direction of all 31 prioritised associations, although between-variant heterogeneity was widespread. Colocalisation and genetic-correlation analyses showed heterogeneous patterns of local and genome-wide sharing, while multivariable MR retained only a subset of associations. Reverse non-lobar intracerebral haemorrhage associations were dependent on a single locus. Organ-specific BAG traits show heterogeneous, direction-dependent, and multi-scale genetic relationships with age-related phenotypes. Cardiometabolic and renal relationships were among the most consistently supported, but differential power, widespread heterogeneity, conditional instrument weakness, and locus dependence argue against a single dominant ageing axis or simple causal pathways.