Ten-year outcomes of endothelial keratoplasty versus penetrating keratoplasty in Hong Kong: a retrospective cohort study

Abstract Background This study compares the long-term survival of primary endothelial (DSEK/DSAEK) versus penetrating keratoplasty in Hong Kong patients with Fuchs endothelial dystrophy (FED) or pseudophakic bullous keratopathy (PBK). Methods A ten-year retrospective cohort of patients who received primary Descemet stripping (automated) endothelial keratoplasty (DSEK/DSAEK) or penetrating keratoplasty (PKP) for either FED or PBK at an academic ophthalmology department in Hong Kong between 2006 and 2015 was conducted. Results Seventy-four eyes of 65 patients received primary keratoplasties with 95 months of mean follow-up, comprising 41 DSEK/DSAEK and 33 PKP. Overall graft survival rates were 76.1% at 5 years and 55.0% at 10 years. DSEK/DSAEK demonstrated higher survival rates than PKP, with 82.1% versus 68.8% survival at 5 years ( P = 0.31) and 63.8% versus 43.5% at 10 years ( P = 0.15). Older age was associated with graft failure risk. DSEK/DSAEK eyes had better observed best-corrected visual acuity (BCVA) at 5 years ( P = 0.006), although this difference was attenuated after adjustment for preoperative BCVA, and the visual difference was no longer apparent at 10 years ( P = 0.42). Graft ulcers occurred in fewer eyes receiving DSEK/DSAEK (4.9%) compared to PKP (21.2%) at 10 years ( P = 0.033). The cumulative rejection rates were 12.2% for DSEK/DSAEK and 9.1% for PKP at 10 years ( P = 0.67) while 14.6% of DSEK/DSAEK and 12.1% of PKP required a re-graft ( P = 0.75). Conclusions Both primary DSEK/DSAEK and PKP provided meaningful long-term graft survival in this Hong Kong cohort of patients with FED or PBK. DSEK/DSAEK eyes had better observed BCVA at 5 years, although this association was attenuated after adjustment for preoperative BCVA, and the visual difference was no longer apparent at 10 years. DSEK/DSAEK was associated with fewer graft ulcers, while rejection and re-graft rates were similar between groups. These findings should be interpreted in the context of the retrospective design, baseline differences between treatment groups, small subgroup sample sizes, and the historical DSEK/DSAEK techniques represented in this cohort.