Although glenoid component loosening has been recognized as a common reason for failure after total shoulder arthroplasty, there are few studies on the outcome of revision surgery for this problem. The purpose of this study is to determine the outcome of patients who underwent revision for glenoid component loosening. Between 1976 and 2002, 68 shoulders in 66 patients underwent revision for glenoid loosening at our institution. Group I consisted of 33 shoulders that underwent placement of a new glenoid component, and group II consisted of 35 shoulders that had removal and bone grafting without glenoid reimplantation. Follow-up averaged 3.8 years for group I and 6.2 years for group II. There was significant overall improvement in pain from preoperatively to postoperatively in both groups (P = .0001). Pain relief occurred in 23 of 33 shoulders in group I and in 24 of 35 in group II (P = .9203). Regarding range of motion, there was no significant change from preoperatively to postoperatively (P > .05), except for active elevation in group I (P = .0387). Patient satisfaction occurred in 24 in group I and in 19 in group II (P = .1150). The rate of survival free of reoperation at 5 years was 91% (95% confidence interval, 81% to 100%) in group I and 78% (95% confidence interval, 63% to 96%) in group II (P = .3019). When the Neer result rating was applied, 9 shoulders in group I and 3 in group II had an excellent or satisfactory result (P = .0432). Twenty shoulders had late positive cultures, most commonly, Propionibacterium acnes. Glenoid revision surgery will often lead to pain relief and patient satisfaction. There is a slight clinical benefit to reimplanting a glenoid component whenever structurally possible. Positive cultures in revision surgery are common, with uncertain clinical significance.