Descemet's Stripping Endothelial Keratoplasty Under Failed Penetrating Keratoplasty: Visual Rehabilitation and Graft Survival Rate

Purpose To evaluate graft survival, risk factors for failure, complications, and visual rehabilitation in patients who underwent Descemet's stripping endothelial keratoplasty (DSEK) under a failed penetrating keratoplasty (PK). Design Retrospective interventional case series. Participants Sixty eyes (60 patients) treated at Price Vision Group, Indianapolis, Indiana. Methods Graft diameters ranged from 8 to 9 mm and were ∼1 mm larger than the previous PK. The Descemet's membrane was not stripped in the majority (54, 84%). The graft was inserted using forceps or a Busin funnel glide (Moria, Anthony, France). The probability of graft survival was calculated by Kaplan–Meier survival analysis. Main Outcome Measures Graft survival, best-corrected visual acuity (BCVA), and complications. Results The mean recipient age was 68 years (range, 17–95 years). Forty eyes had 1 previous failed PK, 14 eyes had 2 previous failed PKs, and 6 eyes had 3 previous failed PKs. Thirty-one eyes (52%) had preexisting glaucoma, and 16 eyes (27%) had prior glaucoma surgery (trabeculectomy in 4, shunt procedure in 12). Fifty-five grafts were performed for visual rehabilitation, and 5 grafts were performed for pain relief. Median follow-up was 2.3 years (range, 2 months to 6 years). Median preoperative BCVA was 1.23 logarithm of the minimum angle of resolution (logMAR) (range, 0.2–3, Snellen 20/340), and median postoperative visual improvement was 0.6 logMAR (6 lines), range −0.3 to +2.7. Four eyes had graft detachment (6.6%), 7 eyes (10.5%) had endothelial rejection, and 10 eyes (16.6%) had graft failure (primary failure in 2, secondary failure in 8). The overall secondary graft survival rates were 98%, 90%, 81%, and 74% at 1, 2, 3, and 4 years, respectively. Prior glaucoma shunt was the principal risk factor for graft failure. The graft survival rates were 100%, 96%, 96%, and 96% in eyes without a prior shunt versus 93%, 74%, 44%, and 22% with a prior shunt at 1, 2, 3, and 4 years, respectively ( P =0.0005; relative risk = 20). Peripheral anterior synechiae ( P =0.14), neovascularization ( P =0.88), endothelial rejection ( P =0.59), and number of prior PKs ( P =0.13) were not independent risk factors for graft failure. Conclusions Endothelial keratoplasty under a previous failed PK is a useful alternative to a repeat standard PK, particularly in eyes with an acceptable topography and refractive outcome before failure. Financial Disclosure(s) Proprietary or commercial disclosure may be