Comparison of DMEK and Ultrathin DSAEK
By Lynda Seminara
Selected By: Stephen D. McLeod, MD
Journal Highlights
Ophthalmology, September 2020
Download PDF
There have been many comparison studies of Descemet membrane endothelial keratoplasty (DMEK) and Descemet stripping automated endothelial keratoplasty (DSAEK) but only limited prospective research on DMEK versus ultrathin DSAEK (UT-DSAEK). Dunker et al. assessed DMEK and UT-DSAEK, looking primarily at high-contrast best spectacle-corrected visual acuity (BSCVA) in addition to endothelial cell density, refractive astigmatism, and perioperative complications. They found no difference in mean BSCVA at any post-op time point.
The study included 54 patients (54 eyes) from six centers in the Netherlands. All patients were pseudophakic adults with Fuchs endothelial corneal dystrophy. The authors defined ultrathin as central graft thickness of 100 μm. Participants were assigned to receive DMEK or UT-DSAEK by minimization randomization that included pre-op BSCVA and pre-op central corneal thickness.
All donor corneas except one were prepared at the same donor bank, and identical inclusion criteria were used for both procedures. Each of the six participating surgeons had performed hundreds of DSAEK and UT-DSAEK surgeries and at least 25 DMEK procedures. The surgeons were allowed to use their preferred DMEK and UT-DSAEK techniques during the operations. The primary outcome measure was BSCVA 12 months after surgery.
Findings at 12 months showed no significant between-group differences in BSCVA (p = .06), endothelial cell density (p = .12), hyperopic shift (p = .27), or spherical equivalent (p = .34). The percentage of eyes that attained BSCVA of 20/25 was greater in the DMEK group (66% vs. 33%; p = .02), but the difference in the percentage that achieved 20/20 was not significant (24% vs. 4%; p = .06). The most common complication with both procedures was the need for rebubbling due to graft detachment (one UT-DSAEK case, seven DMEK cases).
The authors acknowledged that a larger sample size would be valuable, and they noted that differing procedural techniques may have affected the outcomes. They also pointed out that lacking a standard for reporting outcomes of these techniques makes it difficult to properly compare them across different studies. “It would be helpful to set standards on reporting the most important outcome measure, that is, visual acuity,” they wrote. (Also see related commentary by Massimo Busin, MD, in the same issue.)
The original article can be found here.