To report endothelial cell counts (ECC) in patients 1 year after Descemet membrane endothelial keratoplasty (DMEK) and suggest surgical techniques that may minimize intraoperative endothelial cell loss.
Retrospective, observational case series.
This is a retrospective case series of 125 consecutive cases with 1 year of postoperative follow-up. Eight-millimeter DMEK grafts were prepared at a single institution via the submerged cornea, using backgrounds away technique. Grafts were introduced through a 2.4-mm incision and a “no-touch” technique was used to position them.
The mean preoperative donor ECC was 2740 ± 210 (cells/mm 2 ). Mean postoperative year 1 ECC was 2210 ± 550 for an ECC loss of 19% ± 10%. Rebubbling was performed in 5% of cases.
Postoperative year 1 endothelial cell counts in DMEK can equal and even surpass those of Descemet stripping automated endothelial cell keratoplasty (DSAEK) or even penetrating keratoplasty. Given reported decreased rejection rates in DMEK, these patients may have notably superior ECC 5 and 10 years out from surgery compared to if they had undergone DSAEK.
Unique in transplant surgery, Descemet membrane endothelial keratoplasty (DMEK) offers a one-to-one transplant of pathologic tissue for healthy donor tissue. The advantages of DMEK compared to Descemet stripping automated endothelial keratoplasty (DSAEK) include faster postoperative visual recovery, decreased incidence of graft rejection, and more predictable postoperative refractive outcomes. However, the driving force for adoption of DMEK is improved visual outcomes secondary in large part to the lack of a host-graft stromal interface found in DSAEK and other lamellar keratoplasties. The full extent of this improvement in visual quality includes sharpened contrast sensitivity and less higher-order optical aberration in DMEK, which are not fully captured by Snellen acuity.
Despite these advantages, there are few reports of postoperative endothelial cell counts (ECC) after DMEK, and because of technical variations these reports are not necessarily comparable. Postoperative year 1 ECC presumably serve as a surrogate for the long-term survival of DMEK grafts, especially in light of the decreased rate of long-term endothelial loss of these grafts.
Patients and Outcome Analysis
This is an observational, retrospective analysis of 125 consecutive patients at a single institution with at least 1 year of follow-up. This study adheres to the tenets of the Declaration of Helsinki and Institutional Review Board approval at Gorovoy Eye Specialists was obtained. Baseline donor ECC was measured by the eye bank. Postoperative ECC was measured with specular microscopy (Konan Medical, Inc, Greensboro, North Carolina, USA) using the manufacturer’s software and calibration. Manual counts were averaged from 3 measurements. All patients had posterior chamber intraocular lens. Patients with very difficult anatomy, including large iris defects or a history of pars plana vitrectomy, were excluded from DMEK and this study.
The DMEK technique has been described in detail previously. These steps are also outlined on a video online (available at www.youtube.com/watch?v=VhgwqoiPEQU ). The technique will be reviewed briefly here. However, aspects of the surgical technique that the authors believe are critical in helping preserve endothelial cells are reviewed in the discussion.
The SCUBA (“submerged cornea, using backgrounds away”) technique was performed under Optisol media (Bausch & Lomb, Aliso Viejo, California, USA). Donor tissue was placed endothelial side up and scored just inside the Schwalbe line with a y-hook nucleus rotator (Walcott, Ocean View, New Jersey, USA) for 360 degrees. Descemet membrane (DM) was stained with Vision Blue (D.O.R.C. International, Zuidland, Netherlands) and stripping was completed to nearly 50% to the center in all quadrants with a single tying forceps. The stripped edges were repositioned and a superficial trephination was performed. The peripheral ring of trephined DM was separated, which allowed the donor to spontaneously scroll (peeling technique available at www.youtube.com/watch?v=yrMsL0-0PFc ).
Descemet Membrane Endothelial Keratoplasty Technique
An 8-mm surface trephination mark was centered on the patient’s cornea. Host tissue was delineated with Vision Blue and descmetorrhexis was performed first by scoring with the Irrigating Gorovoy Stripper (Ambler, Exton, Pennsylvania, USA) and then stripping with the irrigation and aspiration (I&A) of the available phacoemulsification unit in cortical mode. Care was taken to closely align the area stripped with the surface marking of 8 mm. A Viscojet intraocular lens (IOL) injector system (Bausch & Lomb) was used to forcefully inject the graft through the main 2.4-mm incision, which was then sutured. The graft was then oriented with the rolls of the scroll facing ventrally with short bursts of balanced salt solution.
The graft was then uncurled and centered with a combination of corneal dome compression and sweeping of the corneal surface with cannulas. Finer centration was accomplished similarly with the same 2 cannulas on the surface in a sweeping motion. No instrumentation ever touched the donor after it had been injected in the eye. Once the graft was centered, an air bubble was then injected under the center of the graft and a 90% air fill was completed to a normal intraocular pressure, as measured by corneal ballotment (roughly 15–20 mm Hg). The patient was taken to the recovery room, where he or she lay supine for 90 minutes. The patient was then examined at the slit lamp to ensure that the air bubble cleared the inferior peripheral iridotomy, which was either created preoperatively with a yttrium-aluminum-garnet laser or intraoperatively with microscissors.
Postoperative Steroid Regimen
For the vast majority of cases, patients were placed on prednisolone acetate 1% 4 times daily for 2 months and then tapered 1 drop a month to once daily indefinitely. Steroid responders were switched to loteprednol etabonate 0.5% with the same tapering regimen. Refractory steroid responders were tapered earlier and more aggressively.
Demographic information included a mean patient age of 70. The majority of patients were transplanted secondary to primary Fuchs endothelial dystrophy (76%, 95/125). A history of failed endothelial keratoplasty (16%, 20/125) or pseudophakic bullous keratopathy (7%, 9/125) rounded out virtually all of the other patients.
The mean preoperative donor ECC was 2740 ± 210 (cells/mm 2 ). Mean postoperative year 1 ECC was 2210 ± 550 for an ECC loss of 19% ± 10%. Rebubbling was performed in 5% (6/125) of the cases. Asymptomatic peripheral donor scrolls were mostly located inferiorly and simply observed. Primary graft failure occurred in 3% (4/125) of cases. These patients were regrafted and excluded from the study unless they had 1 year of follow-up with the new graft.