Successful prevention of bacterial endophthalmitis in eyes with the Boston keratoprosthesis.

Purpose: To determine the influence of topical vancomycin prophylaxis on the incidence of bacterial endophthalmitis in eyes with a Boston Keratoprosthesis (KPro).
Methods: A retrospective chart review was performed for 255 eyes of 231 patients who received a KPro between March 1990 and December 2006. Preoperative diagnoses were burn, ocular cicatricial pemphigoid (OCP), Stevens-Johnson Syndrome (SJS), and graft failure/other. Patients used topical antibiotic prophylaxis for the duration of the KPro: polymyxin-trimethoprim or a quinolone in the 1990s, or a quinolone with or without vancomycin beginning in late 1999. For each KPro eye, the follow-up interval was divided into months on or off vancomycin (vancomycin versus no-vancomycin group). The incidence of endophthalmitis was calculated with Kaplan-Meier survival curves.
Results: The 255 eyes were followed for 673.6 patient-years (mean, 2.64 years; range, 1 week to 13 years). There were 18 cases of bacterial endophthalmitis; 17 occurred at least 6 weeks postoperatively (range, 1.5 to 46 months). Gram-positive cocci caused over 80% of cases. Only 1 case, due to an atypical mycobacterium, occurred in a patient using vancomycin.The incidence of bacterial endophthalmitis was lower in the vancomycin group than in the no-vancomycin group: 0.35% versus 4.13% per patient-year (P = 0.001). It was also lower in SJS eyes using vancomycin versus no vancomycin: 1.76% versus 18.39% per patient-year (P = 0.009). In eyes with preoperative diagnoses of burn, OCP, or graft failure/other, the incidence in the vancomycin group was zero.
Conclusion: Topical vancomycin plus a quinolone is effective in preventing bacterial endophthalmitis in KPro eyes.