Background and purpose — 32-mm heads are widely used in total hip arthroplasty (THA) in Scandinavia, while the proportion of 36-mm heads is increasing as they are expected to increase THA stability. We investigated whether the use of 36-mm heads in THA after proximal femur fracture (PFF) is associated with a lower risk of revision compared with 32-mm heads. Patients and methods — We included 5,030 patients operated with THA due to PFF with 32- or 36-mm heads from the Nordic Arthroplasty Register Association database. Each patient with a 36-mm head was matched with a patient with a 32-mm head, using propensity score. The patients were operated between 2006 and 2016, with a metal or ceramic head on a polyethylene bearing. Cox proportional hazards models were fitted to estimate the unadjusted and adjusted hazard ratio (HR) with 95% confidence intervals (CI) for revision for any reason and revision due to dislocation for 36-mm heads compared with 32-mm heads. Results — 36-mm heads had an HR of 0.9 (CI 0.7–1.2) for revision for any reason and 0.8 (CI 0.5–1.3) for revision due to dislocation compared with 32-mm heads at a median follow-up of 2.5 years (interquartile range 1–4.4). Interpretation — We were not able to demonstrate any clinically relevant reduction of the risk of THA revision for any reason or due to dislocation when 36-mm heads were used versus 32-mm. Residual confounding due to lack of data on patient comorbidities and body mass index could bias our results.