A Randomized Multicenter Clinical Trial of Ultrathin Descemet Stripping Automated Endothelial Keratoplasty (DSAEK) versus DSAEK

Objective To compare visual acuity, refraction, endothelial cell density (ECD), and complications after Descemet stripping automated endothelial keratoplasty (DSAEK) and ultrathin DSAEK (UT-DSAEK). Design A multicenter, prospective, double-masked, randomized, controlled clinical trial. Participants From 66 patients with irreversible corneal endothelial dysfunction dues to Fuchs' dystrophy who enrolled from 4 tertiary medical centers in the Netherlands, 66 eyes were studied. Methods Participants were centrally randomized to undergo either UT-DSAEK or DSAEK, based on preoperative best spectacle-corrected visual acuity (BSCVA), recipient central corneal thickness, patient age, and recruitment center. Donor corneas were precut by a single cornea bank. Participants Participants underwent ophthalmic examinations preoperatively and 3, 6, and 12 months after the operation, including manifest refraction, BSCVA using an Early Treatment Diabetic Retinopathy Study chart, and endothelium imaging. Main Outcome Measures BSCVA 12 months postoperatively. Results Preoperative BSCVA did not differ between patients undergoing DSAEK (0.35 logarithm of the minimum angle of resolution [logMAR] [95% confidence interval {CI} 0.27–0.43]; n = 32) and UT-DSAEK (0.37 logMAR [95% CI 0.31–0.43]; n = 34; P = 0.8). BSCVA was significantly better after UT-DSAEK compared with that after DSAEK at 3 months (0.17 logMAR [95% CI 0.13–0.21], n = 31 vs. 0.28 logMAR [95% CI 0.23–0.33], n = 31; P = 0.001), 6 months (0.14 logMAR [95% CI 0.10–0.18], n = 30 vs. 0.24 logMAR [95% CI 0.20–0.28], n = 30; P = 0.002), and 12 months (0.13 logMAR [95% CI 0.09–0.17], n = 33 vs. 0.20 logMAR [95% CI 0.15–0.25], n = 29; P = 0.03). Refraction, ECD loss (40% at 3 months; P <0.001), donor loss (DSAEK n = 2 vs. UT-DSAEK n = 3 [relative risk {RR} 1.4 {95% CI 0.24–7.5}; P = 0.7]), and graft dislocation (DSAEK n = 5 vs. UT-DSAEK n = 5 [RR 1.0 {95% CI 0.34–3.33}; P = 0.9]) did not differ between UT-DSAEK and DSAEK. Donor thickness was significantly thinner for UT-DSAEK (101 μm [95% CI 93–110 μm]; range 50–145 μm) than for DSAEK (209 μm [95% CI 196–222 μm]; range 147–289 μm; P