Histidine-rich glycoprotein (HRG) is a plasma protein implicated in coagulation regulation and experimental thrombosis models. Its causal role in human thrombotic disease has remained uncertain. The investigators aimed to determine whether genetically determined variation in HRG influences risk of thrombosis by performing Mendelian randomisation (MR) analyses in the UK Biobank.
The study used a two-sample MR framework with non-overlapping samples drawn from UK Biobank. Separate genome-wide association studies (GWAS) were performed: one GWAS to identify single nucleotide polymorphisms (SNPs) associated with measured circulating HRG protein levels and a second GWAS to estimate SNP associations with thrombosis outcomes.
Genetic instruments were derived from the HRG-level GWAS and then applied to assess associations with three primary outcomes: overall thrombosis, venous thrombosis, and arterial thrombosis. Multiple MR methods were used for sensitivity analyses. In addition, observational analyses were performed in participants with measured HRG levels using logistic regression adjusted for age, sex, and C-reactive protein (CRP), and stratified analyses by genotype at the lead SNP were reported.
Among 30,680 participants with measured HRG, the GWAS identified multiple loci associated with HRG concentration. The strongest genetic signal was at SNP rs9898, with effect size β = 0.52 and very strong statistical significance (P = 1.1 × 10⁻³⁰⁶). These SNPs provided the basis for genetic instruments used in MR analyses.
Using rs9898 as a single genetic instrument, MR analyses showed no evidence that genetically predicted HRG levels were associated with risk of thrombosis:
Analyses using multiple instruments and alternative MR methods produced similar null findings, indicating consistent absence of association across approaches.
Complementary observational analyses in participants with measured HRG demonstrated that measured HRG levels were not associated with thrombosis after adjustment for age, sex, and C-reactive protein. Stratification by rs9898 genotype did not materially alter these results. The consistency between observational and genetic MR results supports the conclusion that HRG variation does not translate into measurable thrombosis risk in this dataset.
In this UK Biobank-based MR study, genetically determined variation in HRG was not associated with risk of overall, venous, or arterial thrombosis. The authors interpret these findings to mean that HRG-related coagulation phenotypes shown in experimental systems do not appear to confer clinically meaningful thrombotic risk in humans, at least within the populations and outcome definitions evaluated.
Clinicians and researchers should note that, per these analyses, HRG does not emerge as a likely causal target for thrombosis prevention based on genetic evidence presented here. However, the report is a preprint and has not been peer reviewed.
The authors reported that UK Biobank approved the use of data under application number 194362. Ethical approval details for UK Biobank were reported as North West Multi-centre Research Ethics Committee reference 11/NW/0274, and participant informed consent was obtained for use of data and samples.
Competing interest statements were provided by the authors; several authors disclosed research support or personal fees from multiple industry sources, while others reported no competing interests. Funding was declared from the Thrombosis Research Institute.
Specific methodological details beyond those reported in the abstract (for example, exact sample sizes for each outcome GWAS, precise MR methods used beyond statement of multi-instrument approaches, and full sets of instrument SNPs and their F-statistics) were not reported in the source abstract and are therefore not summarized here.
Overall, the preprint concludes that genetically predicted HRG levels are not associated with thrombosis risk in UK Biobank, but the findings should be interpreted in the context that the manuscript is a preprint pending peer review.