Although outcomes of shoulder, hip, and knee arthroplasties have been well-described, there have been no studies directly comparing the outcomes of these procedures as treatments for osteoarthritis. We compared the inpatient mortality, complications, length of stay, and total charges of patients who had shoulder arthroplasty for osteoarthritis with those of patients who had hip and knee arthroplasties for osteoarthritis. A review of the Maryland Health Services Cost Review Commission discharge database identified 994 shoulder arthroplasties, 15,414 hip arthroplasties, and 34,471 knee arthroplasties performed for osteoarthritis from 1994 to 2001. There were no in-hospital deaths after shoulder arthroplasty, whereas 27 (0.18%) and 54 (0.16%) deaths occurred after hip and knee arthroplasties, respectively. Compared with patients who had hip or knee arthroplasties, patients who had shoulder arthroplasties had, on average, a lower complication rate, a shorter length of stay, and fewer total charges. The latter had 1/2 as many in-hospital complications, were 1/6 as likely to have a length of stay 6 days or greater, and were 1/10 as likely to be charged more than $15,000. We believe shoulder arthroplasty is as safe as the more commonly performed major joint arthroplasties.