Infectious Keratitis and Endophthalmitis Following Boston Type I Keratoprosthesis: Incidence, Risk Factors, and...

To evaluate the incidence, clinical characteristics, and risk factors for infectious keratitis and endophthalmitis following Boston Type I keratoprosthesis (KPro) implantation.

We retrospectively reviewed 153 eyes (141 patients) that underwent KPro implantation at the University of Illinois Chicago between 2005 and 2023. Demographic, clinical, microbiologic, and surgical data were analyzed. Incidence rates were calculated per eye-year at risk, and Cox proportional hazards models were used to assess risk factors.

Among 153 eyes from 141 patients, 30 developed at least one infection: 12 developed endophthalmitis alone, 5 developed infectious keratitis alone, and 13 developed both. The incidence rates of endophthalmitis and infectious keratitis were 2.77 and 1.97 cases per 100 eye-years at risk, respectively. In univariable Cox analysis, immune-mediated ocular disease (HR, 3.05; 95% CI, 1.04-8.95; P = .043) was associated with endophthalmitis, and preexisting eyelid/adnexal abnormality was associated with infectious keratitis (HR, 5.67; 95% CI, 2.02-15.95; P = .001). In the model adjusted for age, immune-mediated ocular disease, and prior herpes simplex virus (HSV) keratitis, immune-mediated ocular disease was associated with endophthalmitis (adjusted HR, 3.62; 95% CI, 1.19-11.04; P = .024), whereas prior HSV keratitis showed a borderline association (adjusted HR, 3.05; 95% CI, 1.00-9.26; P = .050). The proportional hazards assumption was not met for immune-mediated ocular disease or preexisting eyelid/adnexal abnormality; the corresponding HRs represent average associations over follow-up. No multivariable model was fitted for infectious keratitis.

Infectious keratitis and endophthalmitis remain serious complications after Boston Type I KPro. Endophthalmitis occurred most often in the first postoperative year but continued during long-term follow-up. Immune-mediated ocular disease was associated with endophthalmitis in the adjusted model, whereas prior HSV keratitis showed a borderline adjusted association. Preexisting eyelid/adnexal abnormality was associated with infectious keratitis in univariable analysis; this finding is exploratory. Long-term surveillance and careful management of the ocular surface and eyelids remain central to postoperative care.