HYPOTHESIS This study investigated the hypothesis that functional outcome remains significantly improved over the preoperative state beyond 15 years of reverse total shoulder arthroplasty (RTSA) for irreparable rotator cuff dysfunction. METHODS Operations were performed on 22 shoulders at a mean age of 68 (range, 54-77) years. The patients could personally be reviewed clinically and radiographically in intervals of 2 to 5 years and with a final follow-up examination at no less than 15 years (mean, 16.1; range, 15-19 years). The RTSA was a primary procedure in 7 procedures, and 15 patients had undergone at least 1 previous nonarthroplasty procedure. RESULTS The mean absolute Constant score (CS) had improved from 23 ± 11 to 58 ± 19 points (P < .001) and the relative CS (rCS) from 30% ± 13% to 73% ± 23% (P < .001) at final follow-up. Significant improvements were seen in mean pain scores (4 to 12 points; P < .001), active anterior elevation (53° to 101°; P = .001), abduction (55° to 86°; P = .005), and Subjective Shoulder Value (27% to 78%; P = .001). The mean rCS and Subjective Shoulder Value did not significantly deteriorate over 15 years, but mean active abduction was significantly reduced over time (P = .018). One or more complications were recorded in 13 patients (59%), with 6 failed RTSAs (27%).The mean rCS of patients with complications (explantations excluded) was not significantly inferior compared with that of patients without complications (62% vs. 81%; P = .090). CONCLUSIONS This early series of RTSA shows a substantial complication and failure rate. If, however, complications can be treated without removal of the implants, outcome is not compromised. Overall shoulder function and subjective outcome remained significantly improved and highly satisfactory during the entire study period.