To evaluate the association between metformin use and incident osteoarthritis (OA) in people with diabetes and the impact of dosing.

Nested case-control study within a cohort of >1.4 million individuals from Sweden.

Participants were aged 35–80 years in 2005, without diabetes or OA. We identified persons with incident diabetes between 2006 and 2016 and excluded those with OA before the diabetes diagnosis and those with an incident OA diagnosis within 3 years of the diabetes diagnosis. Cases were defined as individuals with incident OA before 2020 and were matched with up to four controls without OA in the same period, by sex, diabetes duration, birth year (±1 year) and date of diabetes diagnosis (±180 days) using incidence density sampling. Metformin use before the index date (OA diagnosis) was the main exposure. Secondary exposures were quartiles of total metformin use (defined daily doses (DDD)) and duration-adjusted use (DDD/day), reflecting average daily use. We estimated risk ratios with 95% CIs using conditional logistic regressions, adjusted for age at diabetes diagnosis, education, immigration status and comorbidities.

Incident OA diagnosis in primary or specialist care (International Classification of Diseases codes M15–M19).

We identified 4007 cases and 14 111 controls. Any metformin use was not associated with OA risk (risk ratio (RR) 1.02, 95% CI 0.93 to 1.12). Results for higher total use (0.98 (95% CI 0.86 to 1.11)) and duration-adjusted use (0.92 (95% CI 0.79 to 1.07)) showed no or inconclusive associations.

In individuals with incident diabetes and no prior OA, metformin was not linked to a lower risk of developing OA.