Yang Sui, Shinwan Kany, Shaan Khurshid, Kelvin Supriami, Joel T Rämö, Sean J Jurgens, Nobuyuki Enzan, Seung Hoan Choi, Zhi Yu, Linke Li, Buu Truong, Xingyu Chen, Min Seo Kim, Injeong Shim, Raphael Twerenbold, James P Pirruccello, Pradeep Natarajan, Patrick T Ellinor, Akl C Fahed
Across 2 large biobanks, genetically mediated differences in LDL-C were not consistently associated with VTE risk, suggesting that genetically mediated differences in LDL-C are unlikely to exert a large effect on venous thrombosis risk, although small effects cannot be excluded. Complementary common variant analyses further suggest that previously reported LDL-C associations with VTE may partly reflect broader metabolic and obesity-related pathways rather than LDL-C itself.
BACKGROUND: Venous thromboembolism (VTE) is a common and clinically significant cardiovascular condition. Although low-density lipoprotein (LDL) cholesterol is causally linked to atherosclerotic disease, its role in VTE remains unclear. Meta-analyses of randomized trials have suggested modest reductions in VTE with lipid-lowering therapy, but whether these effects are LDL mediated is uncertain.
OBJECTIVES: This study sought to evaluate whether lifelong genetically mediated differences in low-density lipoprotein cholesterol (LDL-C) are associated with VTE risk.
METHODS: Whole-genome sequencing data from 430,049 UK Biobank and 283,609 All of Us participants were analyzed. Rare protein-truncating and AlphaMissense-predicted damaging missense variants (minor allele frequency <0.1%) in LDLR, APOB, and PCSK9, as well as the common PCSK9 R46L variant, were evaluated. Associations with LDL-C were assessed using linear regression, and VTE associations using Firth logistic regression with fixed-effect meta-analysis. Incident VTE was evaluated using Cox models.
RESULTS: Rare functional variants in LDLR, APOB, and PCSK9 produced substantial gene-specific differences in LDL-C (up to ∼50% differences). Despite these marked lifelong differences, rare variant burdens were not consistently associated with VTE across cohorts, although smaller effect sizes could not be excluded. Meta-analyzed ORs ranged from 0.60 to 1.07. Sensitivity and prospective analyses yielded similar findings. Complementary common variant and multivariable Mendelian randomization analyses suggested that previously reported LDL-C associations with VTE may partly reflect broader metabolic and obesity-related pathways. Positive control analyses of established thrombophilia variants demonstrated expected associations.
CONCLUSIONS: Across 2 large biobanks, genetically mediated differences in LDL-C were not consistently associated with VTE risk, suggesting that genetically mediated differences in LDL-C are unlikely to exert a large effect on venous thrombosis risk, although small effects cannot be excluded. Complementary common variant analyses further suggest that previously reported LDL-C associations with VTE may partly reflect broader metabolic and obesity-related pathways rather than LDL-C itself.