Outcomes of New Partial-Thickness Corneal Grafts: A Paradigm Shift in Practice, but at What Cost?

Abstract

Introduction: Over the past 7-8 years, there has been a paradigm shift in the pattern of practice of corneal transplantation. Partial thickness and corneal endothelial cell allografts are now being performed for indications that previously were treated by full-thickness corneal grafts. We examined the evidence for the success of these new lamellar procedures, with a focus on graft survival and visual outcome. Methods: In a prospectively-maintained, national register of >23,000 corneal grafts with up to 25 years of annual follow-up, 2983 lamellar grafts were identified, of which 42% were endokeratoplasties (posterior corneal endothelial cell grafts), 39% were traditional, peripheral lamellar keratoplasties, and 19% were deep anterior lamellar keratoplasties (DALKs). Kaplan-Meier plots were used to determine graft survival times, Cox proportional hazards regression was used for multivariate graft survival analysis, and visual outcomes were investigated using best-corrected Snellen acuity, that is, with any prescribed spectacle lens or contact lens. Results: Kaplan-Meier graft survival at one year was 74% for endokeratoplasties, 80% for traditional lamellar procedures, and 93% for DALKs. The major indications for endokeratoplasty were Fuchs’ dystrophy (47%) and bullous keratopathy (33%). Over the time frame 2004-2012, penetrating corneal grafts performed for either of these indications exhibited significantly better graft survival than did endokeratoplasties for the same indications (p< 0.001). The major indication for DALK was keratoconus (76%). Over the time frame 1995-2012, penetrating corneal grafts performed for keratoconus exhibited significantly better graft survival than did DALKs for keratoconus (p< 0.001). Approximately 80% of lamellar procedures were performed to improve visual acuity, but at the time of most recent follow-up, more patients with penetrating grafts achieved good post-operative Snellen acuity than did those with lamellar grafts. A best-corrected Snellen acuity of 6/12 or better (=20/40 or better) at most recent follow-up was achieved in 18% of endokeratoplasties, 34% of traditional lamellar grafts, and in 37% of DALKs. Surgeons who performed more than 15 lamellar procedures per year achieved significantly better graft survival than those who performed fewer grafts (p=0.02). Conclusion: There appears to be a surgeon learning curve for new lamellar corneal graft procedures. Currently, outcomes in terms of graft survival and visual acuity are better for penetrating grafts than for the newer lamellar procedures, even when matched for era and indication for graft.