Comparative effects of glucagon-like peptide-1-based regimens on peripheral arterial and limb events in type 2 diabetes: a network meta-analysis of randomized trials

Abstract

Peripheral artery disease (PAD) is a disabling condition in patients with type 2 diabetes mellitus (T2DM), yet lower-extremity outcomes are rarely prespecified in trials of GLP-1-based therapies. We compared individual GLP-1-based regimens for PAD-related vascular and limb events. We searched PubMed, Embase, Scopus, Web of Science, the Cochrane Central Register of Controlled Trials (CENTRAL), ProQuest, medRxiv, ClinicalTrials.gov, and the WHO ICTRP from inception through June 3, 2025, and updated the search on July 16, 2026. Randomized controlled trials (RCTs) in adults with T2DM were eligible. Events from trial reports and registries were harmonized using MedDRA v26.0. Bayesian random-effects network meta-analyses were the primary analyses, frequentist models assessed robustness, and confidence was evaluated with CINeMA. Treatment effects were reported as risk ratios (RRs) with 95% credible intervals (CrIs). Forty-five RCTs involving 110,333 participants were included. For PAOD (33 RCTs; 90,742 participants; 572 events), subcutaneous semaglutide was associated with a lower risk than placebo (RR: 0.37; 95% CrI: 0.14–0.86), with low confidence; the finding was supported by frequentist and sensitivity analyses. In direct placebo-controlled trials, PAOD occurred in 18/4,088 participants receiving subcutaneous semaglutide (4.4 per 1,000) and 33/3,650 receiving placebo (9.0 per 1,000). For PVD (18 RCTs; 69,809 participants; 111 events), albiglutide (RR: 0.11; 95% CrI: 0.01–0.50) and efpeglenatide (RR: 0.09; 95% CrI: 0.01–0.57) were associated with lower risks than placebo, with high confidence for both comparisons. Insulin also showed a favorable Bayesian estimate versus placebo (RR: 0.06; 95% CrI: <0.01–0.49; moderate confidence). No credible Bayesian differences were identified for amputation, necrosis, or peripheral embolism and thrombosis; confidence was predominantly low or very low. The findings suggest agent- and outcome-specific signals rather than a class-wide limb benefit. Because events were rare and usually derived from adverse-event reporting, the results do not establish clinical superiority or support treatment selection based on limb protection alone. Dedicated PAD-focused trials with standardized adjudicated outcomes are needed. PROSPERO CRD420251244166.