Risk Factors for Repeat Keratoplasty after Endothelial Keratoplasty in the Medicare Population

Purpose: To evaluate long-term rates of repeated keratoplasty after endothelial keratoplasty (EK), and to explore risk factors associated with graft failure. Design: Retrospective, cohort study. Methods: Population-based using 2010-2019 Medicare carrier claims. Study population: Medicare beneficiaries aged ≥65 years who underwent endothelial keratoplasty (EK) procedures. Main outcome measures were 1) occurrence of repeated keratoplasty for the overall cohort as well as stratified by clinical indication, and 2), risk factors associated with repeated keratoplasty. The probability of undergoing repeated keratoplasty for different indications was estimated using the Kaplan-Meier method and compared using the log-rank test. A Cox proportional hazards regression model was constructed to determine patient and physician factors associated with the risk of repeated keratoplasty. Results: A total of 94,829 Medicare EK procedures (N = 71,040 unique patients) were included in the analysis. Over the study period, 11,013 of 94,829 EK procedures were followed by repeated keratoplasty. The overall probability of receiving a repeated keratoplasty was 6.1% at 6 months, 7.6% at 1 year, 14.3% at 5 years and 16.9% at 8 years. The probability of repeated keratoplasty was highest for procedures performed for prior failed grafts: 10.2% at 1 year, 29.9% at 8 years. Whereas 6.4% and 11.4% of grafts performed for Fuchs' endothelial dystrophy (FED) required repeated keratoplasty at 1 and 8 years of follow-up, respectively. For bullous keratoplasty and/or corneal edema, 8.6% and 22% of grafts underwent repeated keratoplasty at 1 and 8 years, respectively. In a multivariate analysis, patient factors associated with increased risk of repeated keratoplasty were black (hazard ratio [HR]: 1.29; 95% confidence interval [CI]: 1.20-1.38) or Asians (HR: 1.26; 95% CI: 1.10-1.45) having a glaucoma diagnosis (HR: 1.53; 95% CI: 1.46-1.60), prior glaucoma surgery (HR: 1.26; 95% CI: 1.18-1.35), and concurrent glaucoma surgery (HR: 1.31; 95% CI: 1.20-1.44). Surgeons with higher EK volumes had a decreased risk of repeated keratoplasty. Physicians who graduated medical school >30 years previously had increased risk of repeated keratoplasty compared to those who graduated within 10 years (HR: 1.16; 95% CI: 1.10-1.23). Conclusions: Blacks or Asians, comorbid glaucoma, concurrent or prior glaucoma surgery, and lower volumes of surgery are associated with increased risk of repeated keratoplasty. Racial disparities identified in this study warrant further investigation.