Purpose: Conjunctival complications may occur after glaucoma drainage device surgery. We analyzed the frequency, risk factors, management and outcomes of Ahmed glaucoma valve (AGV)-related conjunctival complications.
Methods: Retrospective review of postoperative conjunctival complications in patients undergoing AGV insertion. Only subjects with ≥1-year follow-up were included.
Results: The charts of 158 subjects with a median age of 64±16.2 years were reviewed. Median follow-up was 43.5 months (range, 12 to 103 mo). Fifty-three (33.5%) wound dehiscences and 14 (8.9%) device exposures were diagnosed 31.6±35.7 and 996±735 days after procedure, respectively (P<0.001). Ninety-one subjects (57.6%) had no conjunctival complications. This uncomplicated group used 3.3 (±1.1) [95% confidence interval (CI): 3.07, 3.51] hypotensive medications before surgery as compared with 3.8 (±1.1) (95% CI: 3.48, 4.10) and 3.9 (±0.9) (95% CI: 3.36, 4.36) for dehiscence and exposure groups, respectively (P=0.01). The inferonasal quadrant was associated with the highest rate of dehiscences (4/7, 57.1%) (95% CI: 18.4, 90.1), followed by the inferotemporal quadrant (30/65, 46.2%) (95% CI: 33.1, 58.2), the superotemporal (15/61, 24.6%) (95% CI: 12.9, 33.8), and the superonasal (4/25, 16%) (95% CI: 10.9-52.0; P<0.0073). There were no differences in dehiscence and exposure rates between limbal versus fornix-based approaches (P=0.54; 95% CI: 24.8-44.9, 24.4-45.7, 5.9-19.6, 3-45.1, respectively). Forty-eight (90.6%) dehiscent wounds resolved with conservative management and 5 were resutured, whereas all exposed devices were managed surgically.
Conclusions: Conjunctival dehiscence is usually a benign, common complication after AGV insertion. It does not need repair as long as the tube is well covered. AGV tube or plate exposures are less common, occur later and were promptly repaired as per current practice. Important factors predisposing to these problems include a greater number of preoperative hypotensive medications and the implantation quadrant.