Design: Genetic association and Mendelian randomisation across two biobanks
Population: 430,049 UK Biobank and 283,609 All of Us participants
Primary outcome: Association of genetically mediated LDL cholesterol with venous thromboembolism
Effect: No consistent association; meta-analysed odds ratios 0.60 to 1.07 despite up to ~50% LDL differences

Whole-genome data from two biobanks (430,049 UK Biobank and 283,609 All of Us participants) were used to test whether genetically determined LDL cholesterol influences venous thromboembolism, a cleaner test of causation than observational lipid measurements.

Rare functional variants in LDLR, APOB and PCSK9 produced lifelong LDL cholesterol differences of up to about 50%, yet venous thromboembolism risk was not consistently associated (meta-analysed odds ratios 0.60 to 1.07). Complementary Mendelian randomisation suggested earlier LDL-venous thromboembolism links may reflect broader metabolic and obesity pathways rather than LDL itself.

For practice this is a boundary marker: LDL lowering remains firmly about arterial atherosclerotic risk, and these data argue against expecting a meaningful effect on venous clots. It does not change lipid targets or who should be on a statin.

Genetically determined LDL cholesterol differences of up to about 50% did not consistently alter venous thromboembolism risk.
Meta-analysed odds ratios spanned 0.60 to 1.07, straddling no effect.
Earlier LDL-clot associations may reflect metabolic and obesity pathways, not LDL itself.
Keep LDL lowering framed as arterial, not venous, risk reduction.
This changes neither lipid targets nor statin indications.

Why it matters

It separates a genuine arterial benefit of lipid lowering from an assumed venous one.

The statistics, in plain English

Odds ratios whose confidence crosses 1.0 mean no reliable effect; a large genetic study finding no association is stronger evidence against causation than observational data suggesting one.