How effective is penetrating corneal transplantation? Factors influencing long-term outcome in multivariate analysis

Background. In a large patient cohort, we investigated long-term corneal graft outcome, risk factors for graft failure, and whether corneal graft survival had improved over time. Methods. Records of 10,952 full-thickness corneal grafts with associated archival follow-up were examined within a prospectively-maintained, national database of 13,831 records, with follow-up extending for up to 18 years. Kaplan-Meier survival analysis was used to indicate variables of interest for Cox proportional hazards regression analysis. A model clustered by individual patient to control for inter-eye or inter-graft dependence was constructed to identify variables best predicting penetrating corneal graft failure. Results. Probability of corneal graft survival was 0.86 at I year, 0.73 at 5 years, 0.62 at 10 years, and 0.55 at 15 years. Graft survival did not improve over a 15-year timeframe. Variables predicting graft failure in multivariate analysis included transplant centre, donor age, preoperative diagnosis, number of previous ipsilateral grafts, lens status, history of corneal neovascularisation, ocular inflammation or raised intraocular pressure in the grafted eye, requirement for anterior vitrectomy, graft size, early suture removal, postoperative events including graft neovascularisation, rise in intraocular pressure, and rejection episodes, type of treatment for raised intraocular pressure, and arrangements for recipient follow-up. A further 11 variables showing a significant influence on graft survival in univariate analysis were not included in the final Cox model. Conclusion. The long-term results of corneal transplantation are no better than for other forms of transplantation and have shown no measurable improvement over the past 15 years.