Authors: Cristina Bostan, Taylor Nayman, Andrei Alexandru-Szigiato, Mona H. Dagher
Author Disclosure Block: C. Bostan: None. T. Nayman: None. A. Alexandru-Szigiato: None. M. H. Dagher: None.

Abstract Body:

Purpose: To determine the epidemiology, visual outcomes, and microbiological profile of Boston keratoprosthesis type I (KPro)-associated endophthalmitis.
Study Design: Retrospective cohort study
Methods: Retrospective medical chart review of 140 consecutive KPro eyes with a follow-up of 61±32 months. Clinically-diagnosed endophthalmitis cases were identified. Outcome parameters included: endophthalmitis incidence rate, surgery-to-infection time, recurrence incidence rate, vision loss compared to baseline, loss of eye (evisceration required to control infection), risk factors and culture results (positivity, isolated organism, antibiotic susceptibility).
Results: Endophthalmitis occurred in 12 eyes (9%) at 20±17 months postoperatively (incidence 0.0015 cases/eye-year). Four cases of recurrence were observed in 2 eyes (incidence 0.0086 cases/eye-year). Loss of vision >2 Snellen lines, progression to “no light perception” and eye loss occurred in 75, 33, and 17%, respectively. All patients were on prophylactic topical fluoroquinolones and steroids before endophthalmitis episode; 75% used extended-wear contact lenses. Ocular surface disease and prior glaucoma tube shunt or trabeculectomy did not affect endophthalmitis risk. Positive cultures (58%) isolated gram-positive bacteria sensitive to fluoroquinolones in all but 1 eye.
Conclusions: KPro-associated endophthalmitis occurred with an incidence of 0.0015 cases/eye-year in our cohort, which is lower than the incidence reported in other North-American centers. Unlike these, where gram-negative and fungal infections are increasing, we report only gram-positive cases with use of fluoroquinolone-alone prophylactic topical therapy.