Purpose
To evaluate the long-term clinical outcome up to 5 years after Descemet membrane endothelial keratoplasty (DMEK).
Design
Retrospective, consecutive case series.
Methods
In this single-center study, 310 consecutive DMEK operations for endothelial decompensation were reviewed; 97 eyes of 84 patients met the inclusion criterion of a minimum 3-year follow-up. Retrospective evaluation of clinical examinations occurred at 1 and 3 months and annually up to 5 years after DMEK at the Department of Ophthalmology, Friedrich-Alexander-University Erlangen-Nürnberg, Erlangen (FAU), Germany. Main outcome measures were corrected distance visual acuity (CDVA), endothelial cell density (ECD), central corneal thickness (CCT), and graft survival (Kaplan-Meier analysis).
Results
Mean follow-up was 53 ± 13 months. CDVA improved from 0.62 ± 0.42 logMAR before DMEK to 0.13 ± 0.12 logMAR ( P < .001); 57% of eyes without ocular comorbidities reached ≥20/25 at 5 years after DMEK. ECD was stable after the initial postsurgical decrease (42% at 1 month, 44% at 5 years), from 2602 ± 243 cells/mm 2 before DMEK to 1460 ± 179 cells/mm 2 at 5 years. CCT decreased from 644 ± 67 μm before DMEK to 557 ± 49 μm at 5 years, with a minimum (530 ± 54 μm) at 3 months. Cumulative probability of 5-year graft survival was 95%.
Conclusions
The long-term sustainability of DMEK was confirmed. DMEK not only provides fast visual rehabilitation but maintains its clinical outcome within a follow-up of 5 years. Visual acuity and endothelial cell loss remain stable between 3 months and 5 years after DMEK.
Descemet membrane endothelial keratoplasty (DMEK), introduced by Melles in 2006, is becoming the treatment of choice in corneal endothelial dysfunction for an increasing number of corneal surgeons. In contrast to earlier endothelial keratoplasty (EK) techniques such as Descemet stripping (automated) endothelial keratoplasty (DSEK/DSAEK), DMEK allows the selective transplantation of the donor Descemet membrane and endothelium only, thereby preserving the corneal anatomy.
Compared with penetrating keratoplasty (PK) and DSEK/DSAEK, DMEK has shown faster visual rehabilitation and better visual outcome, as well as a reduced risk of transplant rejection in studies with 6 months or 1 year postsurgical follow-up. The technique was also successfully applied as a treatment for failed DMEK and DSEK/DSAEK. However, DMEK may be more challenging for surgeons than earlier EK techniques, as the 15-μm-thin graft is more vulnerable to tears during graft preparation.
Several modifications have been introduced to standardize the procedure, increasing the accessibility of the technique to corneal surgeons and reducing complications—for example, by use of sulfur hexafluoride gas or by modifying the size of the intracameral air bubble. The endothelial cell density (ECD) is highly associated with graft survival. Studies regarding the long-term course of ECD after DMEK have been published previously, showing a loss of ECD of 39%–55% after 5 years. However, long-term data concerning clinical outcome such as visual acuity, corneal thickness, and graft survival are not available in publication so far, to the best of our knowledge.
Recently, an Australian registry study was published by Coster and associates comparing the outcomes of lamellar, endothelial (DMEK, DSEK, DSAEK), and penetrating keratoplasty. Patients with Fuchs endothelial corneal dystrophy (FECD) reached 20/40 vision or better in 47% of eyes after PK and only in 35% after EK. For DMEK in particular, the rate of primary graft failure was 12.6% and a Kaplan-Meier analysis showed a graft survival of about 45% at 3 years, which was additionally significantly lower than survival for other EK techniques. This study was critically commented on in the literature. Evidence for the superiority of DMEK from a long-term perspective is missing.
The purpose of our study was to evaluate if DMEK provides long-term benefit up to 5 years.
Methods
Patients
We conducted a retrospective analysis of 97 DMEK operations on 84 patients that met the inclusion criterion of a minimum 3-year follow-up. The follow-up rate after 3 years was 83%, since some patients omitted the 3-year follow-up but attended the 4- or 5-year visit or experienced graft failure with secondary keratoplasty before the 3-year visit. The operations were performed between July 2009 and October 2011 at the Department of Ophthalmology, Friedrich-Alexander-University Erlangen-Nürnberg (FAU), Erlangen, Germany. Mean follow-up was 53 ± 13 months. Apart from the 42 eyes that completed the 5-year follow-up, 3 eyes were not available for the 5-year follow-up, because the 5-year interval was not completed at the time of the data acquisition. In summary, 55 eyes (57%) were lost to follow-up after 5 years.
To assess the cases that qualified for inclusion in this study, all 310 DMEK operations in the given interval at our center were reviewed. Operations matching the time criterion were then systematically (earliest to most recent) evaluated. Eyes in which a secondary graft failure occurred during the follow-up period after DMEK were included until the repeat keratoplasty was performed. Data obtained after repeat keratoplasty were excluded from the analysis. We did not require participation in all intermediate follow-up examinations for patients to be included.
The Institutional Review Board/Ethics Committee approved the study. The study was in adherence to the tenets of the Declaration of Helsinki. Informed consent was obtained from all patients prior to operation. There were 41 (49%) male and 43 (51%) female patients. The mean age was 66 ± 9 years (range 42–89 years). Indications for DMEK operation included FECD (n = 88 eyes, 91%), failed DSAEK (n = 5 eyes, 5%), failed DMEK (n = 1 eye, 1%), failed PK (n = 1 eye, 1%), pseudophakic bullous keratopathy (n = 1 eye, 1%), and previous trabeculectomy with 5-fluorouracil (n = 1 eye, 1%).
Corneal Grafts
Corneoscleral buttons of donor tissue from eye banks in the United States were short-term cultured at 4 C in Optisol-GS (60%). The buttons of donor tissue from Europe were organ-cultured at 34 C in Dulbecco’s modified Eagle medium (40%).
Mean donor age was 67 ± 11 years (range 39–94 years), mean death-to-preservation time was 11 ± 8 hours (range 3–48 hours), and mean culture storage duration was 301 ± 164 hours (range 24–864 hours).
Surgical Technique
Using the technique presented in the following paragraphs, all DMEK operations were conducted under general anesthesia by 3 different surgeons (number of operations per surgeon: 45, 29, 23, respectively). To prevent complications, such as pupillary block, we performed 2 neodymium–yttrium-aluminum-garnet laser iridotomies at 6 and 12 o’clock prior to transplantation. The donor preparation was performed directly prior to DMEK operation.
Descemet membrane endothelial keratoplasty
The surgical technique was used as originally described by Melles with modifications described by Kruse. To prepare the graft, corneoscleral buttons were affixed onto a suction holder (Hanna trephination system; Moria SA, Antony, France). Using an 8-mm trephine, a mark was created on the endothelium and visualized with a 60-second 0.06% trypan blue (Vision blue; DORC, Zuidland, The Netherlands) staining. A narrow piece of the Descemet membrane was scratched off with a razor blade outside the mark. Then the central border was hoisted with a round blade. The graft was partly detached using 2 forceps in a parallel centripetal pulling motion and cut with an 8-mm trephine. For orientation, 3 trephine marks with a 1-mm diameter were set on the border, followed by an additional staining using trypan blue for 60 seconds to increase visibility. Further removal of the graft was completed bimanually with 2 forceps. The spontaneously formed graft roll was then moved into an injector cartridge, as used for implantation of an intraocular lens (Carl Zeiss Meditec, Jena, Germany). Using a 9-mm trephine, the patient’s corneal epithelium was marked. After filling the anterior chamber with air, the Descemet membrane and endothelium inside the selected area were stripped with an inverted hook (Price Hook; Moria SA). A phaco system (ARC Laser, Nuremberg, Germany) was used to stabilize the pressure in the anterior chamber, with the attached irrigation hand piece (0.8 mm; Geuder GmbH, Heidelberg, Germany) controlling the volume. The graft was then inserted into the anterior chamber with the cartridge. The graft roll was placed in a central location with directed streams of balanced salt solution and unrolled using air bubbles injected into the roll. Subsequently, the air bubble was removed, and air was inserted between the iris and the graft to attach the membrane to the corneal stroma. For the next 60 minutes the entire anterior chamber stayed filled with air; after that, the air bubble was reduced to 50% of the anterior chamber volume.
Triple Descemet Membrane Endothelial Keratoplasty
The triple procedure consists of phacoemulsification and intraocular lens implantation before EK (DMEK). Hyaluronic acid was applied as viscoelastic for cataract operation with intraocular lens implantation and subsequently removed entirely before the graft roll was inserted. The pupil was constricted before DMEK operation using acetylcholine chloride (Miochol-E; Novartis Pharma GmbH; Nuremberg, Germany). We used a multipiece acrylic AcrySof (MA 60 AC; Alcon Pharma GmbH, Freiburg I. Breisgau, Germany) or a spherical single-piece acrylic intraocular lens (46 S AcriSmart; Carl Zeiss Meditec, Jena, Germany) for implantation. Triple DMEK was applied in 49 of 97 operations (51%).
Postoperative Medication
The postoperative medication regimen consisted of topical pilocarpine 4 times a day as long as the air bubble was in the anterior chamber to avoid a secondary angle block. Topical ofloxacin (0.3%) was applied twice a day for 10 days. During the first 2 weeks, hyperosmolar eye drops (Omni-Sorb; OmniVision GmbH, Puchheim, Germany) were used up to 5 times a day, if necessary, to clear the edema in the initial postoperative phase. Topical prednisolone 1% was used 5 times a day and reduced monthly to once a day, which was continued during the first year and then stopped.
Statistics and Measurements
We assessed the outcomes of patients undergoing DMEK operation using corrected distance visual acuity (CDVA) in logarithm of the minimum angle of resolution (logMAR) units, ECD, and central corneal thickness (CCT), as mean ± standard deviation (SD) or proportion. CDVA was measured by supplying the patients with the optimal spectacle correction. None of the patients needed contact lenses after DMEK to achieve a better visual acuity than with spectacles.
Follow-up examinations were performed at 1 and 3 months, as well as annually up to 5 years after DMEK. We used Scheimpflug imaging (Pentacam; Oculus, Wetzlar, Germany) to gather CCT and specular microscopy (SeaEagle; HAI Laboratories, Lexingtion, Massachusetts, USA) for ECD. The endothelial cell count analysis was performed at the Department of Ophthalmology, University of Erlangen-Nürnberg by an independent examiner, who was not involved in this study.
Statistics were analyzed using SPSS for Windows (version 20.0; SPSS Inc, Chicago, Illinois, USA). We compared measurements using Wilcoxon signed rank test with a significance level of 5% ( P = .05). The Mann-Whitney U test was used for statistical analysis of differences in the outcomes of subgroups (organ-cultured vs short-term cultured, triple DMEK vs DMEK, outcomes of individual surgeons).
Results
Visual Outcome
Analyzing all eyes included in the study, preoperative mean CDVA ± SD of 0.64 ± 0.41 logMAR improved to 0.18 ± 0.16 logMAR at 5 years ( P < .001). A visual acuity of 20/40 or better was reached in 88% of eyes and of 20/25 or better in 48% 5 years after DMEK ( Table 1 ).
| Baseline | 1 Month | 3 Months | 1 Year | 2 Years | 3 Years | 4 Years | 5 Years | |
|---|---|---|---|---|---|---|---|---|
| CDVA excluding eyes with comor-bidities | ||||||||
| Mean ± SD (logMAR) | 0.62 ± 0.42 | 0.27 ± 0.24 | 0.17 ± 0.13 | 0.15 ± 0.10 | 0.13 ± 0.10 | 0.14 ± 0.13 | 0.15 ± 0.10 | 0.13 ± 0.12 |
| N | 70 | 65 | 60 | 68 | 58 | 64 | 48 | 35 |
| ≥20/40 (0.5) | 19% | 75% | 92% | 99% | 98% | 92% | 96% | 97% |
| ≥20/25 (0.8) | 0% | 29% | 42% | 47% | 57% | 61% | 60% | 57% |
| CDVA including all eyes | ||||||||
| Mean ± SD (logMAR) | 0.64 ± 0.41 | 0.35 ± 0.36 | 0.25 ± 0.24 | 0.22 ± 0.27 | 0.20 ± 0.22 | 0.20 ± 0.20 | 0.21 ± 0.21 | 0.18 ± 0.16 |
| N | 97 | 87 | 79 | 88 | 76 | 80 | 60 | 42 |
| ≥20/40 (0.5) | 15% | 66% | 80% | 90% | 92% | 85% | 87% | 88% |
| ≥20/25 (0.8) | 0% | 22% | 33% | 40% | 43% | 49% | 48% | 48% |
For further analysis of CDVA we excluded eyes with ocular comorbidities limiting visual outcome after DMEK. These ocular comorbidities included glaucoma with optic nerve atrophy (n = 5), diabetic or cystoid macular edema (n = 3), age-related macular degeneration (n = 3), amblyopia (n = 4), proliferative diabetic retinopathy (n = 1), macular pucker (n = 1), discrete central corneal scar (n = 1), and previous retinal detachment with scleral buckling and cryocoagulation (n = 1). The diagnosis of ocular comorbidity was made before DMEK in all eyes excluded from the analysis of CDVA except for the 3 eyes that developed macular edema after DMEK. Graft failures with secondary keratoplasty (n = 12) were also excluded from this analysis. In some cases we found multiple reasons for exclusion. A total of 27 DMEK operations met these exclusion criteria.
Eyes without ocular comorbidity or graft failure had a preoperative mean CDVA ± SD of 0.62 ± 0.42 logMAR, improving to 0.13 ± 0.12 logMAR at 5 years ( Figure 1 ). A visual acuity of 20/40 or better was reached in 97% of eyes and of 20/25 or better in 57% 5 years after DMEK ( Table 1 ).
Statistically significant improvement of mean CDVA was achieved in the first month after DMEK ( P < .001) and between 1 and 3 months ( P < .001). No further statistically significant changes occurred between 3 months and 5 years after DMEK.
Endothelial Cell Density
Mean donor ECD ± SD was 2602 ± 243 cells/mm 2 , and decreased to 1460 ± 179 cells/mm 2 at 5 years after DMEK ( Table 2 , Figure 2 ). Within the first month after DMEK a statistically significant ( P < .001) decrease in mean ECD was observed. After that ECD was stable; no further statistically significant changes occurred in the remaining follow-up. Total ECD loss reached 44% at 5 years after DMEK (with 42% in the first month).
| Baseline | 1 Month | 3 Months | 1 Year | 2 Years | 3 Years | 4 Years | 5 Years | |
|---|---|---|---|---|---|---|---|---|
| ECD | ||||||||
| Mean ± SD (cells/mm 2 ) | 2602 ± 243 | 1502 ± 260 | 1488 ± 274 | 1485 ± 252 | 1428 ± 251 | 1404 ± 295 | 1408 ± 269 | 1460 ± 179 |
| N | 97 | 76 | 74 | 80 | 70 | 73 | 51 | 33 |
| CCT | ||||||||
| Mean ± SD (μm) | 644 ± 67 | 548 ± 75 | 530 ± 54 | 538 ± 61 | 535 ± 51 | 550 ± 55 | 560 ± 54 | 557 ± 49 |
| N | 97 | 82 | 77 | 84 | 71 | 77 | 56 | 41 |
Central Corneal Thickness
Before DMEK mean CCT ± SD was 644 ± 66 μm, with a decrease to 530 ± 54 μm at 3 months, followed by an increase up to 557 ± 49 μm at 5 years after DMEK ( Table 2 , Figure 3 ). Mean CCT reached a minimum at 3 months with an 18% decrease compared to preoperative data. At 5 years the total decrease was 13%.
Within the first month after DMEK ( P < .001) and between 1 and 3 months ( P < .001), a statistically significant decrease in mean CCT was observed, followed by a statistically significant increase between 1 and 2 years ( P = .01) and 2 and 3 years ( P = .02). Between 3 months and 1 year CCT remained stable ( P = .06). Over the total 5 years of follow-up the decrease was statistically significant ( P < .001).
Graft Survival
The Kaplan-Meier estimator was used to assess graft survival. For survival analysis we excluded the initial learning curve of the surgeons involved (operations conducted within the first 6 months after introduction of DMEK at our center). Within the first 6 months after the introduction of DMEK at our center 8 operations led to graft failure.
The survival analysis using the Kaplan-Meier estimator showed a 95% graft survival 5 years after DMEK ( Figure 4 ). Grafts that failed to clear after operation were categorized as primary graft failures. Primary graft failure occurred in 2 cases (2%) and made up 50% of all graft failures. We observed 2 additional graft failures (2%) at 23 and 39 months after DMEK. Operations used for secondary keratoplasty were PK and repeat DMEK. One episode of immunologic rejection occurred (1%), but did not lead to graft failure.
