Descemet Membrane Endothelial Keratoplasty: Safety and Outcomes: A Report by the American Academy of Ophthalmology

Purpose: To review the published literature on the safety and outcomes of Descemet membrane endothelial keratoplasty (DMEK) for the surgical treatment of corneal endothelial dysfunction. Methods: Literature searches were last conducted in the PubMed and the Cochrane Library databases most recently in May 2017. The searches, which were limited to English-language abstracts, yielded 1085 articles. The panel reviewed the abstracts, and 47 were determined to be relevant to this assessment. Results: After DMEK surgery, the mean best-corrected visual acuity (BCVA) ranged from 20/21 to 20/31, with follow-up ranging from 5.7 to 68 months. At 6 months, 37.6% to 85% of eyes achieved BCVA of 20/25 or better and 17% to 67% achieved BCVA of 20/20 or better. Mean endothelial cell (EC) loss was 33% (range, 25%–47%) at 6 months. Overall change in spherical equivalent was +0.43 diopters (D; range, –1.17 to +1.2 D), with minimal induced astigmatism of +0.03 D (range, –0.03 to +1.11 D). The most common complication was partial graft detachment requiring air injection (mean, 28.8%; range, 0.2%–76%). Intraocular pressure elevation was the second most common complication (range, 0%–22%) after DMEK, followed by primary graft failure (mean, 1.7%; range, 0%–12.5%), secondary graft failure (mean, 2.2%; range, 0%–6.3%), and immune rejection (mean, 1.9%; range, 0%–5.9%). Overall graft survival rates after DMEK ranged from 92% to 100% at last follow-up. Best-corrected visual acuity after Descemet's stripping endothelial keratoplasty (DSEK) ranged from 20/34 to 20/66 at 9 months. The most common complications after DSEK were graft detachment (mean, 14%; range, 0%–82%), endothelial rejection (mean, 10%; range, 0%–45%), and primary graft failure (mean, 5%; range, 0%–29%). Mean EC loss after DSEK was 37% at 6 months. Conclusions: The evidence reviewed supports DMEK as a safe and effective treatment for endothelial failure. With respect to visual recovery time, visual outcomes, and rejection rates, DMEK seems to be superior to DSEK and to induce less refractive error with similar surgical risks and EC loss compared with DSEK. The rate of air injection and repeat keratoplasty were similar in DMEK and DSEK after the learning curve for DMEK.