Abstract 4141929: Genetic Evidence for GIP as a Therapeutic Target for Peripheral Artery Disease

Background: Peripheral arterial disease (PAD), when comorbid with type 2 diabetes (T2D), is associated with increased risk of adverse limb events, and therapeutic options for these patients are limited. The STARDUST trial recently demonstrated that Liraglutide significantly improves lower limb perfusion in PAD patients with concomitant T2D in an open label randomized setting, potentially expanding the use of Glucagon-like Peptide-1 Receptor (GLP1R) agonists to this population. Using data from the Million Veteran Program (MVP), we aimed to estimate the effects of GLP1R and gastric inhibitory peptide (GIP) agonists on PAD using genetic proxies. Methods: We obtained previously published GWAS summary statistics for T2D from the DIAMANTE consortium, and performed an association analysis for PAD in MVP at the GLP1R and GIP loci. For our primary analysis in MVP, association testing was performed with clinical PAD using logistic mixed models controlling for age, sex and population structure. We performed a colocalization analysis of PAD and T2D association signals at these loci using hyprcoloc, and then tested the association with PAD stratified by T2D status. Genome-wide significance (5x10e-8) was used to determine statistical significance for our primary analysis. Results: We identified 51330 participants with and 256807 participants without PAD in MVP. Among the European population, we observed evidence of association for PAD at the GIP locus (lead variant rs35895680, OR = 1.048, P = 1.723e-10 for the C risk allele), but not at the GLP1R locus (lead variant rs113678890, OR = 1.6434, P = 1.150e-3). The lead PAD variant at the GIP locus was observed to be the lead T2D variant in the recent DIAMANTE consortium GWAS analysis, and we observed evidence of colocalization between T2D and PAD at the GIP locus, (Posterior Probability of the same causal variant = 0.99). Logistic regression for rs35895680 in the MVP PAD data revealed significant associations in both diabetics (OR=1.04, P = 9.02e-5) and non-diabetics (OR 1.04, P = 1.35e-3), without evidence of heterogeneity (P = 0.60281). Conclusions: These findings highlight GIP as a promising target for PAD intervention, potentially benefiting a broader patient population beyond those with concomittent diabetes. Consequently, future trials incorporating therapies inhibiting both GLP1R and GIP, such as tirzepatide, could offer superior therapeutic benefits for PAD compared to GLP1R agonists alone.