Background

Type 2 diabetes (T2D) increases the risk of osteoarthritis, with both conditions sharing common risk factors including obesity. Glucagon-like peptide-1 receptor agonists (GLP1RA) might regulate osteoarthritis via their weight-reducing and anti-inflammatory effects. Nevertheless, whether the use of GLP1RA is associated with a lower risk of incident osteoarthritis (OA) among patients with T2D remains uncertain.

Objective

To assess the association between the use of GLP1RA and the risk of OA among patients with T2D

Methods

We conducted a nationwide, population-based cohort study applying a target trial emulation framework. From the National Health Insurance Service database (2010 - 2022) of Korea, eligible patients included adult T2D patients who initiated GLP1RA or dipeptidyl peptidase-4 inhibitors (DPP4I). The index date was defined as the date of study drug initiation. The primary outcome was a composite OA, comprising knee, hip, and hand OA, and secondary outcomes were the individual components of primary outcome, along with incident joint replacement (JR). Applying as-treated approach, patients were followed from the index date until the earliest of outcome occurrence, discontinuation, switching, all-cause death, or study end date (31 Dec 2022). Propensity score fine stratification weighting was applied to balance the covariate distributions between treatment groups. Incidence rates per 100 person-years (IRs) and hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using Poisson regressions and Cox models, respectively. We repeated our analysis by stratifying patients into obese and non-obese, with cutoff value of body mass index 25 kg/m², to assess possible effect modification.

Results

Among a total of 2 056 824 eligible patients, 9 866 patients started GLP1RA (mean age 45 yrs; 55% male; 67% dulaglutide) and 2 047 158 patients started DPP4I (mean age 56 yrs; 70% male). After weighting, during mean follow-up of 621 days, IR for OA were 3.99 (95% CI, 3.54 – 4.48) and 4.36 (4.34 – 4.38) among GLP1RA and DPP4I users, respectively. Compared to DPP4I, GLP1RA was not associated with a lower risk of OA (HR 0.93, 95% CI, 0.83 – 1.05). The results for the secondary outcomes were consistent with the primary outcome (HR 1.16 [0.52 – 2.59] for JR; 0.92 [0.78 – 1.09] for knee OA; 1.20 [0.62 – 2.34] for hip OA; 4.06 [0.98 – 16.75] for hand OA). We did not find significant effect modification by obesity (obese: HR 0.94 [0.79 – 1.12]; non-obese: 0.88 [0.63 – 1.22]; p for interaction 0.56).

Conclusions

In this nationwide cohort study, the use of GLP1RA was not associated with a lower risk of OA among patients with T2D. Although GLP1RA regulated OA symptoms in obese and osteoarthritic population, they did not prevent OA in patients with T2D.