Analyzing 378,772 UK Biobank participants alongside multi-ancestry GWAS data, researchers identified 149 genetic loci shared between type 2 diabetes (T2D) and coronary artery disease (CAD). Strikingly, 35 of 42 novel loci showed discordant effects — raising risk for one condition while lowering the other. Seven mechanistic clusters emerged, three centered on liver fat with opposing T2D-CAD directionality. Mendelian randomization confirmed that genetically higher liver fat raises T2D risk but paradoxically lowers CAD risk — an effect explained by whether hepatic lipid is retained locally or exported as ApoB-containing lipoproteins into circulation.
This finding reframes a long-standing clinical puzzle: why some metabolic interventions simultaneously worsen glycemia while improving cardiovascular outcomes, or vice versa. The ApoB partitioning mechanism offers a compelling explanation — liver fat itself isn't uniformly harmful; its cardiovascular consequences depend on downstream lipid trafficking. This has immediate implications for interpreting lipid-lowering therapies like ANGPTL3 inhibitors or PCSK9 inhibitors, which reduce ApoB-containing lipoproteins and may carry distinct metabolic tradeoffs. For precision prevention, clinicians may eventually need to characterize patients' hepatic lipid phenotype before selecting cardiometabolic therapies. Limitations include the predominantly European UK Biobank ancestry despite multi-ancestry GWAS inputs, and observational constraints inherent to biobank designs. As a preprint not yet peer-reviewed, these mechanistic conclusions warrant independent replication before clinical translation. Still, the liver-centric framework represents a potentially paradigm-shifting advance in understanding discordant cardiometabolic risk.
