Background and purpose - There is no consensus on which type of shoulder prosthesis should be used in patients with rheumatoid arthritis (RA). We describe patients with RA who were treated with shoulder replacement, regarding patient-reported outcome, prosthesis survival, and causes of revision, and we compare outcome after resurfacing hemi-arthroplasty (RHA) and stemmed hemi-arthroplasty (SHA). Patients and methods - We used data from the national Danish Shoulder Arthroplasty Registry and included patients with RA who underwent shoulder arthroplasty in Denmark between 2006 and 2010. Patient-reported outcome was obtained 1-year postoperatively using the Western Ontario Osteoarthritis of the Shoulder index (WOOS), and rates of revision were calculated by checking revisions reported until December 2011. The patient-reported outcome of RHA was compared to that of SHA using regression analysis with adjustment for age, sex, and previous surgery. Results - During the study period, 167 patients underwent shoulder arthroplasty because of rheumatoid arthritis, 80 (48%) of whom received RHA and 34 (26%) of whom received SHA. 16 patients were treated with total stemmed shoulder arthroplasty (TSA), and 24 were treated with reverse shoulder arthroplasty (rTSA). 130 patients returned a completed questionnaire, and the total mean WOOS score was 63. The cumulative 5-year revision rate was 7%. Most revisions occurred after RHA, with a revision rate of 14%. Mean WOOS score was similar for RHA and for SHA. Interpretation - This study shows that shoulder arthroplasty, regardless of design, is a good option in terms of reducing pain and improving function in RA patients. The high revision rate in the RHA group suggests that other designs may offer better implant survival. However, this should be confirmed in larger studies.