To analyze the longer-term surgical outcomes of Ahmed and Baerveldt shunt implants in the treatment of refractory glaucoma. Retrospective, nonrandomized, comparative trial. Consecutive patients who underwent glaucoma shunt implantation (70 Baerveldt, 48 Ahmed) by a single surgeon. Follow-up retrospective analysis of 118 consecutive cases of refractory glaucoma that required placement of either Ahmed or Baerveldt shunt implants for intraocular pressure (IOP) control. The longer term clinical course (survival curve analysis up to 48 months) was reviewed. The primary outcome measure was surgical success (6 mmHg < or = IOP < or = 21 mmHg without additional glaucoma surgery or devastating complication). Secondary outcome measures included mean IOP and number of medications used at the following postoperative visits: day 1, 1 week, 1 month, 3 months, 6 months, and every 6 months thereafter (up to 48 months). In unsuccessful cases, the median time to failure was determined. No significant differences in success rates were noted at 1, 2, 3, or 4 years after surgery, with 4-year success rates of 62% for the Ahmed group and 64% for the Baerveldt group (P = 0.843). Mean IOPs were comparable from 1 to 48 months (P>0.05). However, patients in the Ahmed group required more glaucoma medications at 18, 24, 30, and 36 months (P<0.05). The median times to failure were 3.2 months in the Baerveldt group, compared with 15.0 months in the Ahmed group (P = 0.009). No differences were observed in longer term survival rates and IOP profiles between Ahmed and Baerveldt shunt implants up to 48 months following surgery. Patients in the Baerveldt group were more likely to develop early postoperative hypotony-related complications and failure, whereas patients undergoing Ahmed implants were more likely to be on additional glaucoma medications (starting at 18 months postsurgery) and develop later onset failure.