With the introduction of endothelial keratoplasty (EK) techniques such as Descemet stripping (automated) endothelial keratoplasty (DSEK/DSAEK) and Descemet membrane endothelial keratoplasty (DMEK), we may have entered a new era of corneal transplantation in the past decade, replacing penetrating keratoplasty as the technique of choice for treatment of endothelial disorders.1–5
EK may offer better and faster visual rehabilitation by avoiding refractive problems such as high or irregular astigmatism and diminishing ocular surface complications related to delayed wound healing and graft suturing. In addition, allograft rejection rates were reduced considerably.6–9 Especially, DMEK has been shown to result in very fast visual recovery with approximately 3 quarters of patients reaching a best-corrected visual acuity (BCVA) of ≥20/25 (0.8) at 6 months after DMEK.10,11
In a previous study, we had analyzed the clinical outcome and the effect of technique standardization at 6 months after DMEK of 500 consecutive DMEK cases at the Netherlands Institute for Innovative Ocular Surgery.12 This large study group has now been followed up to 2 years postoperatively to evaluate whether the clinical outcomes and postoperative complication rate remained stable after 6 months. In addition, we assessed whether parameters that may potentially influence these 2-year outcomes could be identified.
MATERIALS AND METHODS
Five hundred consecutive eyes of 393 patients who underwent DMEK at our institute for Fuchs endothelial corneal dystrophy (FECD), bullous keratopathy (BK), and other indications (Table 1) between October 2007 and September 2012 were analyzed at 12 and 24 months postoperatively. The preoperative and 6 months results of this study cohort have been described in a previous study.12 The first 25 DMEK cases, constituting the learning curve of the technique, were not part of the study group.12,13 Patient and donor demographics are reported in Table 1.
All patients signed an institutional review board–approved informed consent form before surgery, and the study was conducted according to the Declaration of Helsinki.
Donor Tissue and Surgical Protocol
All DMEK grafts were harvested at Amnitrans EyeBank Rotterdam as described previously14,15 and stored in organ culture medium (CorneaMax; Eurobio, France) until transplantation. Surgeries were performed by 2 surgeons (I.D. and G.R.J.M.) using the standardized “no-touch” surgical technique as described in detail previously.16,17 Postoperative medication included topical 0.5% chloramphenicol, 0.5% ketorolac, and 0.1% dexamethasone for 4 weeks followed by a routine 0.1% fluorometholone tapering regimen up to 12 months. Thereafter, patients were advised to continue using fluorometholone once a day or every other day indefinitely.
Measurements and Statistical Analysis
Patients were routinely examined at 12 and 24 months after surgery; preoperative and 6-month examinations had been analyzed within a previous study.12 Examinations included measurement of BCVA, central endothelial cell density (ECD), and central corneal thickness (CCT). Exclusion criteria and the number of eyes excluded from BCVA, ECD, and CCT analysis are shown in Table 2.
CCT analysis was based on Scheimpflug imaging (Pentacam HR; Oculus, Wetzlar, Germany). Donor ECD was measured preoperatively in vitro using an inverted light microscope (Axiovert 40; Zeiss) by multiple trained eye bank experts, and postoperatively in vivo ECD was evaluated using noncontact specular microscopy (Topcon SP3000; Topcon Medical Europe BV, Capelle a/d IJssel, the Netherlands). Commercial software of the specular microscope (ImageNet software; Topcon Medical Europe) provides an automatic ECD analysis program, which automatically recognizes cell borders. For all endothelial images of the central corneal window, the automatically delineated cell borders were checked, and when incorrectly assigned, the cell borders were manually reassigned by a trained technician. For each follow-up, the results of 3 measurements were averaged. All analyzed images were saved with the overlying cell border lines that were used for the analysis, and these images were checked by a second experienced reader for accuracy. For all images, the largest possible area of the image was used for the analysis.
With repeated-measures analysis of variance using linear mixed models, change in BCVA, ECD, and CCT over time (from preoperative to 24-month follow-up) and the time point of stabilization were assessed.18 In addition, relevant covariates (patient age, patient sex, lens status, surgery indication, graft attachment status, and donor storage time) were evaluated for their impact on these changes over time.
Multiple linear regression was used to assess whether ECD changes influenced the CCT outcome between the 6- and 24-month follow-up. Logistic regressions were performed to evaluate whether patients who were lost to follow-up after 6 months differed on baseline measurements (patient age, patient sex, lens status, surgery indication, graft attachment status, storage time, preoperative ECD, preoperative BCVA, and preoperative CCT) from those who were followed over 24 months.
In addition, for each linear mixed model, sensitivity analyses were performed to assess whether the results were influenced by patients being lost to follow-up. To prevent alpha inflation and power loss because of multiple testing, all P values were corrected using the Benjamini and Hochberg method.19 After applying the Benjamini and Hochberg correction, all P values can be compared with alpha = 0.05.
Second eyes of patients undergoing bilateral DMEK were excluded from statistical analysis, which included 102 FECD eyes, 2 BK eyes, and 2 eyes with failed DSEK transplants. For all statistical analyses, assumptions were checked, and if violations occurred, they are reported in the Results section. All analyses were performed with R (version 3.1.3) using the package “nlme” (version 3.1-120).
At 24 months after DMEK, 82% of eyes reached a BCVA of 20/25 (≥0.8), 52% ≥20/20 (1.0), and 16% ≥20/17 (1.2) (n = 360), which did not differ from the outcome after 12 months with 81% of eyes achieving a BCVA of 20/25 (≥0.8), 49% ≥20/20 (1.0), and 15% ≥20/17 (1.2) (n = 396) (P = 0.828) (Table 3). BCVA improved from 6 to 12 months after DMEK (P = 0.0059) and remained stable up to 24 months.
However, when excluding 82 eyes that only had 6-month follow-up but no 12 and 24 months data available, there was no significant change anymore between 6- and 12-month BCVA outcomes (P = 0.3975). Nonresponse analysis showed that these 82 eyes differed from the group that had BCVA data available at least one follow-up after 6 months, with respect to patient age [mean age, 71 (±16) years versus 67 (±12) years, P = 0.0282] and preoperative visual acuity [logMAR 0.81 (±0.55) versus 0.47 (±0.34), P < 0.0001], but not with respect to graft attachment status (P > 0.05).
Parameters related to changes in visual acuity (in logMAR) up to 24 months were patient age, preoperative pachymetry, surgery indication, and graft attachment status (P < 0.05) (Table 4). Eyes with FECD as surgery indication reached better visual acuity levels (on average 0.16 on the logMAR scale) than eyes with BK (P = 0.0005). Eyes with completely attached grafts reached better visual acuity outcomes (on average 0.13 points on the logMAR scale) than eyes with a partial graft detachment of ≤1/3 of the graft surface area (P = 0.0003). BCVA did not differ with respect to the extent of the graft detachment (>1/3 vs. ≤1/3, P = 0.4097). Patient sex, lens status, preoperative donor ECD, and graft storage time did not show a significant relation with postoperative visual outcomes (Table 4).
Endothelial Cell Density
Mean ECD decreased from preoperatively 2530 (±210) to 1600 (±490) cells/mm2 (n = 447), 1530 (±488) cells/mm2 (n = 427), and 1400 (±491) cells/mm2 (n = 392) at 6, 12, and 24 months, respectively, corresponding to a decrease of 37 (±18)%, 40 (±18)%, and 45 (±18)%, respectively (Table 3 and Fig. 1). Although the ECD decrease slowed down considerably after 6 months postoperatively, there was still a significant difference between ECD at 6, 12, and 24 months, respectively (P <0.05 for all time points). Sensitivity analysis showed that the pattern of the ECD decrease over time did not change after excluding eyes with missing values after 6-month follow-up.
Patient age, surgery indication, lens status, and graft attachment status were related to postoperative ECD outcomes up to 24 months (P < 0.05) (Table 4), whereas patient sex, storage time, and preoperative pachymetry did not show a significant relation. Eyes with FECD had better ECD outcomes than eyes with BK (P = 0.0118), and phakic eyes had better outcomes than pseudophakic eyes (P = 0.007). Eyes with completely attached grafts had better ECD outcomes than eyes with a partial graft detachment of ≤1/3 of the graft surface area (P < 0.0001). Eyes with a detachment of >1/3 of the graft surface area had comparable ECD outcomes as eyes with a minor (≤1/3) detachment (P = 0.2314).
Postoperative CCT decreased from 667 (±92) μm preoperatively to 525 (±46) μm (n = 428), 527 (±40) μm (n = 423), and 534 (±43) μm (n = 378) at 6, 12, and 24 months, respectively, corresponding to a decrease of 20 (±11) %, 20 (±10) %, and 19 (±10) %, respectively (Table 3 and Fig. 1). No significant change was observed for CCT readings between 6 and 12 months or between 12 and 24 months (P > 0.05). However, average CCT increased significantly between 6 to 24 months postoperatively (P = 0.036), which was not related to the ECD decrease in the same period (P = 0.15). Sensitivity analysis showed that CCT outcomes were not affected after excluding eyes with missing values after 6-month follow-up.
Complications and Reinterventions
Within the first 6 months, graft detachment was the main complication with 34 eyes (6.8%) having a detachment of >1/3 of the graft surface area and 45 eyes (9.0%) having a detachment of ≤1/3 of the graft surface area as described previously.12 Further complications up to 2 years after DMEK were allograft rejection (1.4%, n = 7) and secondary graft failure (1.4%, n = 7).
Within the 24-month study period, a total of 32 eyes (6.4%) required retransplantation (Table 3), of which 11 (2.2%) were performed within the first 6 months, 15 (3.0%) between 6 and 12 months, and 6 (1.2%) between 12 and 24 months. Retransplantation was performed for primary graft failure (0.2%, n = 1), secondary graft failure (1.0%, n = 5), and significant graft detachment (5.2%, n = 26).
This study evaluated the clinical results of a group of 500 consecutive DMEK eyes up to 2 years postoperatively and assessed which parameters may influence the clinical outcome. Overall, our data show that the clinical outcome of this large DMEK group remained excellent within the study period, and the rate of retransplantation and major postoperative complications was very low.
With more than 80% of eyes reaching a BCVA of ≥20/25 (≥0.8) and more than 50% reaching ≥20/20 (≥1.0), DMEK provides an outstanding and stable level of visual acuity until 2 years postoperatively. In contrast to reports on DSEK/DSAEK and ultrathin DSAEK, which reported continuous improvement in BCVA up to 3 years, visual acuity after DMEK already stabilizes after 6 months, confirming that DMEK allows the fastest visual rehabilitation among all EK techniques.8,20–23 Visual acuity outcomes 2 years after DMEK were related to patient age, preoperative pachymetry, surgery indication, and graft attachment status. It has been shown in earlier studies that clinical outcomes may vary according to surgery indication with eyes operated on for FECD consistently achieving better visual outcomes than BK eyes. This may be attributed to the faster depletion of endothelial cells in BK eyes resulting in more severe corneal edema possibly causing visually relevant subepithelial fibrosis and disorganization of the stromal collagen structure.24 Earlier intervention for these eyes may be suggested and patients should be counseled appropriately on the expected clinical outcomes.
Interestingly, eyes with minor graft detachment, although most often considered visually insignificant, resulted in inferior BCVA outcomes compared with eyes with completely attached grafts. However, for each individual case with minor graft detachment, the risks associated with a rebubbling procedure may be carefully weighed against the theoretical visual improvement expected from air reinjection and the demands of the patient.
The ECD decrease after 2 years in our DMEK study group was 45%, with the main ECD decrease occurring within the first 6 months. Although ECD outcomes at 6 months postoperatively in a recent study could be related to graft storage times,25 no such relation was observed at 2 years postoperatively, which indicates that prolonged graft storage might influence the ECD decrease in the early but not in the longer term after surgery. Overall, the observed ECD decrease confirms results from previous DMEK studies26,27 and is comparable to the ECD decrease after DSEK/DSAEK and compares favorably to the midterm ECD decrease after PK.28,29 Although first studies reported the longer-term ECD outcome after DMEK,27,30,31 it will remain important to closely monitor the ECD decrease for long follow-up times to compare long-term outcomes of different techniques.
In addition, as already shown in earlier studies on ECD outcomes after different keratoplasty techniques, ECD results up to 2 years in the current DMEK cohort were better for eyes with FECD than for eyes operated on for BK, corroborating that FECD eyes have better clinical outcomes after DMEK than BK eyes.30,31
With respect to postoperative complication rates, these remained very low up to 2 years after DMEK. As reported earlier, partial graft detachment constitutes the major complication in the early postoperative course,12,31 whereas beyond 6 months, allograft rejection and secondary graft failure constitute the main complications, each occurring in less than 2% of cases. Approximately 80% of the repeat transplantation procedures in our study group were performed within the first postoperative year for persistent graft detachment. Because this complication is decreasing with surgical experience and technique standardization,12 early repeat transplantation may also decrease in the near future.
Overall, from the clinical results of this large DMEK study group up to 2 years postoperatively, a promising perspective for longer-term DMEK outcomes especially for eyes operated on for FECD and with completely attached grafts may be anticipated.
1. Melles GR. Posterior lamellar keratoplasty: DLEK to DSEK to DMEK
. Cornea. 2006;25:879–881.
2. Melles GR, Ong TS, Ververs B, et al. Descemet membrane endothelial keratoplasty (DMEK
). Cornea. 2006;25:987–990.
3. Gorovoy MS. Descemet-stripping automated endothelial keratoplasty. Cornea. 2006;25:886–889.
4. Price MO, Gorovoy M, Benetz BA, et al. Descemet's stripping automated endothelial keratoplasty outcomes compared with penetrating keratoplasty from the Cornea Donor Study. Ophthalmology. 2010;117:438–444.
5. Park CY, Lee JK, Gore PK, et al. Keratoplasty in the United States: a 10-year review from 2005 through 2014. Ophthalmology. 2015;122:2432–2442.
6. Heinzelmann S, Böhringer D, Eberwein P, et al. Outcomes of Descemet membrane endothelial keratoplasty, Descemet stripping automated endothelial keratoplasty and penetrating keratoplasty from a single centre study. Graefes Arch Clin Exp Ophthalmol. 2016;254:515–522.
7. Schoenberg ED, Price FW Jr, Miller J, et al. Refractive outcomes of descemet membrane endothelial keratoplasty triple procedures (combined with cataract surgery). J Cataract Refract Surg. 2015;41:1182–1189.
8. Goldich Y, Showail M, Avni-Zauberman N, et al. Contralateral eye comparison of Descemet membrane endothelial keratoplasty and Descemet stripping automated endothelial keratoplasty. Am J Ophthalmol. 2015;159:155–159.
9. Price MO, Scanameo A, Feng MT, et al. Descemet's membrane endothelial keratoplasty: risk of immunologic rejection episodes after discontinuing topical corticosteroids. Ophthalmology. 2016;123:1232–1236.
10. Dirisamer M, Ham L, Dapena I, et al. Efficacy of Descemet membrane endothelial keratoplasty: clinical outcome of 200 consecutive cases after a learning curve of 25 cases. Arch Ophthalmol. 2011;129:1435–1443.
11. Guerra FP, Anshu A, Price MO, et al. Descemet's membrane endothelial keratoplasty: prospective study of 1-year visual outcomes, graft survival, and endothelial cell loss. Ophthalmology. 2011;118:2368–2373.
12. Rodríguez-Calvo-de-Mora M, Quilendrino R, Ham L, et al. Clinical outcome of 500 consecutive cases undergoing Descemet's membrane endothelial keratoplasty. Ophthalmology. 2015;122:464–470.
13. Ham L, Dapena I, van Luijk CM, et al. Descemet membrane endothelial keratoplasty (DMEK
) for Fuchs endothelial dystrophy: review of the first 50 consecutive cases. Eye. 2009;23:1990–1998.
14. Groeneveld-van Beek EA, Lie JT, van der Wees J, et al. Standardized 'no-touch' donor tissue preparation for DALK and DMEK
: harvesting undamaged anterior and posterior transplants from the same donor cornea. Acta Ophthalmol. 2013;91:145–150.
15. Lie JT, Birbal R, Ham L, et al. Donor tissue preparation for Descemet membrane endothelial keratoplasty. J Cataract Refract Surg. 2008;34:1578–1583.
16. Dapena I, Moutsouris K, Droutsas K, et al. Standardized “no-touch” technique for Descemet membrane endothelial keratoplasty. Arch Ophthalmol. 2011;129:88–94.
17. Liarakos VS, Dapena I, Ham L, et al. Intraocular graft unfolding techniques in Descemet membrane endothelial keratoplasty. JAMA Ophthalmol. 2013;131:29–35.
18. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38:963–974.
19. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B Stat Methodol. 1995;57:289–300.
20. Li JY, Terry MA, Goshe J, et al. Three-year visual acuity
outcomes after Descemet's stripping automated endothelial keratoplasty. Ophthalmology. 2012;119:1126–1129.
21. Wacker K, Baratz KH, Maguire LJ, et al. Descemet stripping endothelial keratoplasty for Fuchs' endothelial corneal dystrophy: five-year results of a prospective study. Ophthalmology. 2016;123:154–160.
22. Busin M, Madi S, Santorum P, et al. Ultrathin Descemet's stripping automated endothelial keratoplasty with the microkeratome double-pass technique: two-year outcomes. Ophthalmology. 2013;120:1186–1194.
23. Tourtas T, Laaser K, Bachmann BO, et al. Descemet membrane endothelial keratoplasty versus Descemet stripping automated endothelial keratoplasty. Am J Ophthalmol. 2012;153:1082–1090.
24. Shimizu T, Yamaguchi T, Satake Y, et al. Topographic hot spot before Descemet stripping automated endothelial keratoplasty is associated with postoperative hyperopic shift. Cornea. 2015;34:257–263.
25. Rodríguez-Calvo de Mora M, Groeneveld-van Beek EA, Frank LE, et al. Association between graft storage time and donor age with endothelial cell density
and graft adherence after Descemet membrane endothelial keratoplasty. JAMA Ophthalmol. 2016;134:91–94.
26. Feng MT, Price MO, Miller JM, et al. Air reinjection and endothelial cell density
in Descemet membrane endothelial keratoplasty: five-year follow-up. J Cataract Refract Surg. 2014;40:1116–1121.
27. Schlögl A, Tourtas T, Kruse FE, et al. Long-term clinical outcome after Descemet membrane endothelial keratoplasty. Am J Ophthalmol. 2016;169:218–226.
28. Price MO, Fairchild KM, Price DA, et al. Descemet's stripping endothelial keratoplasty five-year graft survival and endothelial cell loss. Ophthalmology. 2011;118:725–729.
29. Lass JH, Gal RL, Dontchev M, et al. Donor age and corneal endothelial cell loss 5 years after successful corneal transplantation
. Specular microscopy ancillary study results. Ophthalmology. 2008;115:627–632.
30. Baydoun L, Ham L, Borderie V, et al. Endothelial survival after Descemet membrane endothelial keratoplasty: effect of surgical indication and graft adherence status. JAMA Ophthalmol. 2015;133:1277–1278.
31. Ham L, Dapena I, Liarakos VS, et al. Mid-term results of Descemet membrane endothelial keratoplasty (DMEK
): 4 to 7 years clinical outcome. Am J Ophthalmol. 2016;171:113–121.