To investigate the outcome of Descemet membrane endothelial keratoplasty (DMEK) in patients with graft failure after Descemet stripping automated endothelial keratoplasty (DSAEK).
Retrospective cohort study.
setting : Institutional. study population : Fifteen eyes of 15 patients that underwent DMEK for graft failure with corneal decompensation following DSAEK were analyzed; 15 eyes with primary DMEK for Fuchs corneal dystrophy were included as control group. main outcome measures : Best-corrected visual acuity (BCVA), endothelial cell density (ECD), central corneal thickness (CCT), and rebubbling rate.
DMEK surgery was successful in all cases of both groups. Mean BCVA (logMAR) before DMEK was 1.27 ± 0.34 in the DMEK after DSAEK group and 1.0 ± 0.40 in the Primary DMEK group. After DMEK, mean BCVA increased significantly to 0.23 ± 0.21 ( P = .012, DMEK after DSAEK group) and 0.29 ± 0.23 ( P = .042, Primary DMEK group) after 3 months. There were no significant differences in mean BCVA between both groups at each visit. The rebubbling rate was 13% in the DMEK after DSAEK group and 40% in the Primary DMEK group ( P = .1). Mean CCT decreased significantly in both groups 1 month after DMEK ( P < .05). Mean ECD and change of ECD did not differ significantly between both groups at each visit ( P > .05).
The results after DMEK as a procedure to treat graft failure after DSAEK were as good as in patients that underwent DMEK as primary intervention to treat advanced Fuchs dystrophy. This indicates that the optical quality can be reestablished by DMEK in patients with failed DSAEK.
Posterior lamellar techniques for corneal transplantation allow the replacement of diseased corneal endothelium by healthy donor tissue while the recipient’s corneal stroma is left in place. The main difference among current techniques for posterior lamellar keratoplasty is the composition of the graft. In Descemet stripping automated endothelial keratoplasty (DSAEK) the graft consists of healthy donor endothelium, corresponding Descemet membrane, and parts of the donor’s posterior corneal stroma, whereas in Descemet membrane endothelial keratoplasty (DMEK) the graft consists of healthy donor endothelial cells and Descemet membrane only.
The functional outcome of DMEK surgery benefits from the use of a very thin graft: recent work has shown that transplantation of such thin grafts not only results in faster visual rehabilitation but also ends in better visual acuity. Better outcomes after DMEK might be attributable to reduced higher-order aberrations of the posterior surface, which might be explained by the re-formation of a physiologic interface between posterior stroma and Descemet membrane after DMEK.
The ease of graft preparation and transplantation as well as a fast visual rehabilitation have made DSAEK the preferred procedure for the treatment of diseased corneal endothelium; it has replaced penetrating keratoplasty in many parts of the world. In case of graft failure after DSAEK, repeat DSAEK has been shown to be a reasonable therapeutic option.
Since visual quality is better after DMEK, DSAEK surgeons are confronted with the question whether DMEK after failed DSAEK might be superior to repeat DSAEK because of better functional outcome. It has been shown by Dirisamer and associates and Ham and associates that DMEK is technically possible in eyes following DSAEK with unsatisfactory visual results. However, in eyes with corneal decompensation after graft failure, DMEK is more difficult to perform because of poor visibility of the anterior chamber in corneal edema. The question whether DMEK is feasible and successful in these eyes has, to our knowledge, not been addressed.
Recent findings provide evidence that adhesion of Descemet membrane to the posterior stroma is mediated by the interaction of a variety of adhesive glycoproteins such as fibronectin and vitronectin in the interfacial matrix with stromal proteins. Descemet stripping during DMEK exposes the posterior corneal stroma while the interfacial matrix remains at the cleavage plane of the transplanted endothelium–Descemet membrane complex. DSAEK might possibly modify the stromal proteins and thereby impair the adhesion of the graft in DMEK for failed DSAEK. Moreover, removal of the failed DSAEK graft might modify the posterior stromal surface and thus possibly induce optical aberrations. Thereby, the optical quality of the host-graft junction after DMEK for failed DSAEK might be reduced.
To examine the suitability of DMEK after failed DSAEK we analyzed the clinical and morphologic outcome of all cases with graft failure after DSAEK that were treated with secondary DMEK compared to patients with primary DMEK surgery.
Patients and Methods
The study group (DMEK after DSAEK group) of this retrospective cohort study consisted of 15 consecutive eyes from 15 patients (mean age: 67.3 ± 10.2 years, male: n = 5, female: n = 10) in which DMEK was performed for the treatment of graft failure following DSAEK using a standardized technique for graft preparation and transplantation. The indication for DSAEK was Fuchs corneal dystrophy in 14 patients (93%) and pseudophakic bullous keratopathy in 1 patient (7%). All corneas following failed DSAEK showed corneal decompensation with stromal edema. The reason for graft failure was not clear in all our patients. A presumed immunologic rejection was present in 2 eyes. All DSAEK surgeries were performed between October 2006 and December 2011, all secondary DMEK surgeries between October 6, 2009 and February 2, 2013. All surgeries (DMEK and DSAEK), which were performed in our department, were carried out by 3 experienced surgeons. The surgeons had similar learning curves and were experienced with the technique. DMEK surgeries included in this study were not among the first 75 DMEKs of each surgeon.
Our regular cohort of patients undergoing primary DMEK tends to have better values of best-corrected visual acuity (BCVA) before DMEK compared to our cohort undergoing DMEK for graft failure after DSAEK (DMEK after DSAEK group). In previous studies, patients undergoing DMEK had a preoperative BCVA between 0.6 and 0.7 logMAR (logarithm of the minimal angle of resolution); that is, BCVA was distinctly better than in the patients undergoing DMEK for failed DSAEK in the current study (1.27 ± 0.34 logMAR). Unpublished analysis of our general DMEK cohort showed a significant correlation between preoperative BCVA and BCVA 1, 3, 6, and 12 months after DMEK (n = 216, Spearman correlation coefficient 1 month: ρ 0.281, P < .001, 12 months: ρ 0.274, P = .005; Supplemental Figure , available at AJO.com ). Therefore, we compared the patients of the DMEK after DSAEK group retrospectively with 15 consecutive patients with a similarly low preoperative BCVA undergoing primary DMEK for Fuchs endothelial dystrophy.
The control group (Primary DMEK group) consisted of 15 eyes from 15 patients undergoing primary DMEK for Fuchs corneal dystrophy (mean age: 71.3 ± 10.6 years, P = .233, Mann-Whitney U ). The primary DMEK surgeries were performed between September 8, 2009 and February 27, 2013. Mean follow-up time after DMEK was 14.9 ± 10.6 months (range 1.0–31.2 months) in the DMEK after DSAEK group and 18.5 ± 9.1 months (range 1.0–30.8 months) in the Primary DMEK group ( P = .436, Mann-Whitney U ).
Donor corneas were stored in short-term culture (DMEK after DSAEK group: n = 3, primary DMEK group: n = 5) or long-term organ culture (DMEK after DSAEK group: n = 12, primary DMEK group: n = 10). DMEK grafts were prepared on the day of surgery directly prior to transplantation by the surgeon himself.
Informed consent was obtained from the patients. The study complied with the tenets of the Declaration of Helsinki and adhered to all state laws of the country. The Institutional Review Board of the University of Erlangen-Nuremberg, Germany waived the need for approval.
Graft preparation and graft transplantation for DMEK were performed as previously described in detail. The failed DSAEK grafts were removed from the stroma using an inverted hook (Price endothelial keratoplasty hook; Moria SA, Antony, France) and extracted with a forceps (DMEK after DSAEK group). In primary DMEK, the Descemet membrane of the recipient was removed in a central 9-mm area. After stripping, the graft (Descemet membrane with adherent corneal endothelial cells) spontaneously formed a roll, which was injected into the anterior chamber and positioned by repeated injections of balanced salt solution and air bubbles. At the end of the procedure, the anterior chamber was filled with an air bubble, which was reduced to 50% of the anterior chamber volume after 60 minutes.
The parameters analyzed in this study included BCVA, central corneal thickness (CCT) (Pentacam; Oculus, Wetzlar, Germany), and endothelial cell density (ECD) (SeaEagle; Rhine-Tec GmbH, Krefeld, Germany). The need for additional intracameral air injections after DMEK was assessed at slit-lamp examination and by the use of slit-lamp optical coherence tomography (SL-OCT; Heidelberg Engineering, Heidelberg, Germany). Examination results from the day before DMEK as well as 1, 3, 6, 12, and 18 months after grafting were analyzed.
Removed DSAEK grafts were fixed in buffered 10% formaldehyde solution (pH 7.2), dehydrated, and embedded in paraffin. Serial sections cut at 5 μm were stained with hematoxylin-eosin and periodic acid–Schiff. Main criteria of the histologic analyses included thickness of the DSAEK graft at the thickest and thinnest part, endothelial cell count per high-power field, loss of keratocytes, and occurrence of retained Descemet membrane, retrocorneal membranes, infectious infiltrates, inflammatory cells, or melanin granules.
Transmission Electron Microscopy
For electron microscopy, 1 explanted DSAEK graft was processed according to standard protocols, as described previously. The ultrathin sections were examined with a transmission electron microscope (EM 906E; Carl Zeiss AG, Oberkochen, Germany).
Statistical evaluation was performed using IBM SPSS software version 20.0 (SPSS, Armonk, New York, USA). Differences of samples between groups were assessed by Mann-Whitney U test. Categorical data were analyzed with χ 2 test and Fisher exact test, if the expected value of each group was less than 5. Differences of parameter values within groups were assessed using the Wilcoxon test. The significance level was set at P = .05.
The average time between DSAEK and DMEK was 26 ± 17 months (range: 4–70 months). DMEK surgery was successful in all procedures. No primary graft failure occurred. Corneal edema, thickening, and opacification of the cornea improved in all patients after DMEK.
Four eyes in the DMEK after DSAEK group were excluded from the analysis of the visual acuity owing to amblyopia, ischemic optic neuropathy, macular pucker, and age-related macular degeneration. In the Primary DMEK group 4 eyes were excluded because of epithelial basement membrane dystrophy (Map-dot-fingerprint), subepithelial haze/scarring (n = 2), and a macular hole. There was no difference in the number of eyes with ocular comorbidities limiting the visual acuity in both groups (Fisher exact test, P = .659).
Mean BCVA in the DMEK after DSAEK group was 1.27 ± 0.34 before DMEK and increased significantly, to 0.43 ± 0.33 (Wilcoxon, P = .003) at 1 month and to 0.23 ± 0.21 (Wilcoxon, P = .012) at 3 months after DMEK ( Table 1 ). At the following visits 6, 12, and 18 months after DMEK, mean BCVA did not increase significantly compared to the previous visit ( P > .05). In the Primary DMEK group, mean BCVA increased significantly from 1.0 ± 0.40 before surgery to 0.32 ± 0.09 (Wilcoxon, P = .027) 1 month and to 0.29 ± 0.23 (Wilcoxon, P = .042) 3 months postoperatively ( Table 1 ). Thereafter, mean BCVA increased slightly but not significantly up to 0.10 ± 0.09 at the follow-up visit 18 months after surgery ( P > .05). Comparative analysis of mean BCVA values at baseline and at each follow-up visit showed no significant differences between both groups ( Table 1 , Figure 1 ).
|Follow-up (Months After DMEK)||DMEK After DSAEK Group (Mean ± SD, logMAR)||Primary DMEK Group (Mean ± SD, logMAR)||P Value (Mann-Whitney U )|
|Preoperative||1.27 ± 0.34||1.0 ± 0.40||.065|
|1||0.43 ± 0.33||0.32 ± 0.09||.733|
|3||0.23 ± 0.21||0.29 ± 0.23||.247|
|6||0.19 ± 0.05||0.19 ± 0.07||1.0|
|12||0.19 ± 0.08||0.13 ± 0.09||.429|
|18||0.14 ± 0.14||0.10 ± 0.09||.731|
Central Corneal Thickness
In the DMEK after DSAEK group, mean CCT decreased significantly, from 917 ± 184 μm before DMEK to 480 ± 68 μm 1 month after DMEK (Wilcoxon, P = .005), and in the Primary DMEK group from 727 ± 79 μm to 506 ± 61 μm (Wilcoxon, P = .018, Table 2 ). There was no significant difference in CCT at each follow-up visit after DMEK surgery between both groups. During the following months no significant change in mean CCT could be observed within both groups ( P > .05 between each follow-up visit, Table 2 , Figure 2 ).
|Follow-up (Months After DMEK)||DMEK After DSAEK Group (Thickness in μm ± SD)||Primary DMEK Group (Thickness in μm ± SD)||P Value|
|Preoperative||917 ± 184||727 ± 79||N/A a|
|1||480 ± 68||506 ± 61||.247|
|3||479 ± 51||522 ± 88||.18|
|6||480 ± 60||505 ± 43||.299|
|12||520 ± 63||520 ± 51||1.0|
|18||506 ± 66||512 ± 35||.529|
Endothelial Cell Density
Mean ECD decreased by 39%, from 2470 ± 120 cells/mm² to 1496 ± 311 cells/mm², in the DMEK after DSAEK group ( P = .005, Wilcoxon) within the first month after surgery. In the Primary DMEK group, mean ECD decreased by 45%, from 2455 ± 181 cells/mm² to 1341 ± 201 cells/mm² ( P = .043, Wilcoxon) within the first month after surgery. During the following months mean ECD did not change significantly. There was no significant difference in the endothelial cell counts between both groups at baseline and at each follow-up examination ( Table 3 , Figure 3 ).
|Follow-up (Months After DMEK)||DMEK After DSAEK Group (Cell Count/mm² ± SD)||Primary DMEK Group (Cell Count/mm² ± SD)||P Value (Mann-Whitney U )|
|Donor||2470 ± 120||2455 ± 181||.683|
|1||1496 ± 311||1341 ± 201||.428|
|3||1545 ± 294||1591 ± 193||.573|
|6||1611 ± 265||1519 ± 272||.679|
|12||1393 ± 223||1499 ± 306||.536|
|18||1390 ± 239||1391 ± 193||1.00|