DMEK has become the gold standard for the surgical treatment of endothelial diseases of the cornea. DMEK not only appears to be useful as a primary surgical procedure, but is also increasingly being considered for the surgical treatment of endothelial decompensation following prior corneal transplantation.
This study therefore investigated the outcome of repeat DMEK in case of endothelial graft failure following not just previous DMEK but also previous PK.
In our study, in terms of BCVA patients with initial DMEK benefit more from repeat DMEK than PK patients. It has been reported that an intact corneal surface provides better visual results in DMEK compared to PK 1. According to our data (see Fig. 1) and the findings of another comparative study 13, this seems also to be true for repeat DMEK and is consistent with the optical conditions before the graft failure.
However, as concerns ECD, PK patients show higher numbers during the first three years (see Fig. 2). This may be due to the surgical trauma to the peripheral endothelium of the host in initial DMEK, as well as to the additional air/gas contact during the first procedure. There are also more indications for PK where there is a healthy endothelial reservoir in the periphery (as in keratoconus for example). This may lead to longer graft survival rates 15,16. In addition, ECD loss may become similar between DMEK and PK patients in the long-term follow-up of more than three years. Even though there are no comparative studies yet, long-term data from various large keratoplasty centers allow a valid overview. For DMEK, after initial ECD loss of about 34% after six months (reflecting the surgical trauma), the annual loss rate decreases to 9% 17–19. For PK, an initial annual decrease of ECD of 20% is reported 20,21, consisting of a rapid and a slow component, the latter leading to slower decrease in the long-term period 22. These data suggest that after three years, ECD numbers should balance between DMEK and PK. In addition, this difference in ECD does not seem to be relevant for the rate of a new graft failure, which is comparable between both groups (Fig. 4).
Looking at immune reactions, one could speculate that the higher alloantigen load grafted in PK-patients could lead to higher rejection rates, which is not supported by our data (Fig. 5). In contrary, repeat PK after failed PK is associated with a significantly higher graft rejection rate 23 and therefore considered as high risk keratoplasty 24 which seems not the case in DMEK after failed PK. First steps to elucidate the rejection/failure mechanisms in DMEK have been made recently, highlighting non-cellular components of the innate immune system 25. Furthermore, rebubbling rates are comparable between PK and DMEK patients (Fig. 3), despite the fact that PK eyes show a more complex altered anatomy in comparison to DMEK eyes. There are often defects of iris, aphakia or glaucoma drainage devices are present, all of them complicating graft unfolding and adherence as well as hampering the presence of the air/gas tamponade. Additional factors such as incapability of the patient to stay supine for example, or the amount of the tamponade seem to be relevant in this context. Compared to rebubbling rates in primary and uncomplicated DMEK, about 20% is consistent with our initial rebubbling rate during the first three years 1, which decreased during the last few years to 11% (data not shown).
According to these data, repeat DMEK is a considerable option in treating endothelial graft failure following PK and DMEK. Taking into consideration the current literature further helps to validate our findings and to answer the question if this success of DMEK is independent from the initial keratoplasty technique:
-
Most studies have been published including patients with DMEK for DMEK failure 4–6, 25–29. Even though numbers of included patients differ (6–55) and reported clinical parameters are heterogeneous, all studies confirm a fast visual rehabilitation and similar courses of ECD, rebubbling and immune reactions.
-
Some studies address the results from DMEK for PK failure 7–10, 14,30,31. Despite varying numbers of patients (5–93) and outcome parameters, all these papers highlight the fast rehabilitation following DMEK in PK eyes. The longest follow-up is three years which corresponds with the present study 7. Even in complicated, vascularized corneas with limbal insufficiency, comparable clinical courses of DMEK are reported 32.
-
There are two studies reporting successful treatment of DSAEK-failure by DMEK 11,12.
These different studies are quite heterogeneous concerning cohorts and data analysis, but some of them conclude in common:
-
Possible risk factors for graft failure following DMEK are glaucoma, complex preoperative anterior segment situations, and difficult primary DMEK surgery 13,26,27. One study states that a loss of endothelial cells is the main cause for graft failure following DMEK 25.
-
Visual acuity improves in all mentioned studies (amount dependent on rate of extracorneal eye diseases limiting BCVA).
-
Central corneal thickness decreases in successful repeat DMEK.
-
There is an endothelial cell loss of about 30% during the first 3–6 months postoperatively.
-
Rebubbling rates vary between 15% and 60% in DMEK as repeat keratoplasty procedure 4,7,27,28.
As far as we know, the only comparative study that currently exists (PK-DSAEK versus PK-DMEK) is also retrospective and reports a favorable outcome in DMEK as treatment for graft failure; it comprises 52 cases in total 13.
Despite these results in favour of DMEK as treatment of graft failure, our study has some weaknesses: heterogeneity and limited size of both groups, and the retrospective character. In addition, we did not check for the increase in quality of life of the patients systematically as outcome measure of repeat DMEK. There are patients with low visual acuity after repeat DMEK who nevertheless are extremely enthusiastic about the relief of bullous keratopathy-complaints and the fast rehabilitation.
However, in spite of limited data availability and worldwide still widely spread DSAEK and PK 2, DMEK seems to be appropriate in treating graft failure even in complicated eyes. Especially when rapid visual rehabilitation and/or anatomical reconstruction and minimally invasive surgery are required (e.g. in the elderly or in eyes with comorbidities that limit visual prognosis and impending pain due to bullous keratopathy), DMEK is a reasonable alternative to repeat DSAEK or PK. In particular, patients with endothelial failure following successful Limbo-PK (restored limbal function and absence of graft rejection; 4/35 in our cohort) benefit from repeat DMEK as they retain their limbal function and stromal clarity 32. Given increasing surgical experience, more and more complex situations will be manageable.