Genomic modelling across more than 900,000 FinnGen and UK Biobank participants sorted seven hand and shoulder disorders into two heritable axes: tenosynovial (carpal tunnel, trigger finger, de Quervain; 61 loci) and fibro-capsular (frozen shoulder, Dupuytren's, plantar fascial disorder; 42 loci). Mendelian randomisation showed higher genetically predicted BMI raised carpal tunnel odds (OR 1.67 per SD) yet lowered Dupuytren's odds (OR 0.57). Pooling 210 placebo-controlled trials (134,138 participants), incretin receptor agonists increased serious fibro-capsular events (Peto OR 3.88), including frozen shoulder (4.16) and Dupuytren's (3.58), but not tenosynovial events (1.02).
The paradox is the interesting part: genetics suggests excess weight protects against Dupuytren's contracture, while the drugs that cut weight appear to raise its risk, hinting that weight loss itself, or a direct drug effect on fibroblasts, may matter. Frozen shoulder has long been linked to diabetes, so confounding by indication is a real concern. Serious adverse events are rare and poorly adjudicated in trials not designed to capture musculoskeletal outcomes, which explains the wide confidence intervals (1.17 to 10.93 for Dupuytren's). Absolute risks are likely small, and Mendelian randomisation tests lifelong BMI exposure, not drug-induced loss. This is a preprint awaiting peer review, so the estimates may shift. Treat it as a hypothesis-generating pharmacovigilance signal, not a reason to stop therapy; anyone on GLP-1 drugs developing new shoulder stiffness or palm nodules should mention it to their clinician. Incremental but worth validating in real-world data.