DMEK: The New Frontier in Endothelial Transplantation
DMEK: The New Frontier in Endothelial Transplantation
A short decade ago, penetrating keratoplasty (PK) was essentially the only procedure performed to repair endothelial dysfunction. Melles et al. had just begun their early clinical trials of the novel procedure he had developed, posterior lamellar keratoplasty (PLK), which would jump-start a new field in corneal transplantation called endothelial keratoplasty (EK). Terry and Ousley were working primarily in the laboratory to establish the ex vivo safety studies that would ultimately lay the groundwork for a large clinical study of 100 eyes that would establish the benefits of selective endothelial transplantation in their version of PLK, called deep lamellar endothelial keratoplasty (DLEK).
In 2012, EK is now well established as the treatment of choice for endothelial dysfunction and has been used for Fuchs' dystrophy, pseudophakic bullous keratopathy, irido–corneal–endothelial syndrome, and other endothelial diseases. Faster visual rehabilitation, a more physiologic topography and excellent visual outcomes are but a few of its benefits over PK. Innovations in EK have exploded as surgeons search for faster ways to provide better vision. Subsequent developments by Melles et al. simplified preparation of the donor by using a Descemetorrhexis rather than an intrastromal posterior resection. Gorovoy added the simplicity of the automated microkeratome for preparation of the donor tissue. These provided the basis for Descemet's stripping automated EK (DSAEK), the most commonly used method of endothelial transplantation today. Goals of transplantation have gone far beyond attempting to rectify frank opacity and corneal irregularity. Researchers have begun to look into models for reliably predicting refractive outcomes after EK.
Even now, current popular techniques do not restore the eye's native anatomy. DSAEK introduces a stroma-to-stroma interface which has been blamed for the lower-than-expected rates of visual acuity in the 20/20–20/25 range. A pure Descemet's membrane–endothelium complex transplant might address some of these issues, the first of which was performed by Melles et al. in 2006 and named Descemet's membrane EK (DMEK).
DMEK has become a new focus of study in EK (Figure 1). Although innovations in DSAEK remain strong as surgeons attempt to optimize the surgical course and minimize operative complications in a procedure that is already relatively easy and speedy compared with previous lamellar procedures, the benefits to patients have essentially been fully realized from the visual standpoint. DMEK may offer another option for endothelial replacement that can provide better visual outcomes. Of course, as with all new procedures, it is currently being performed by a handful of surgeons and many new challenges accompany the prospective benefits of this new iteration in endothelial transplantation. In this article we review the current published literature on this newest iteration of EK.
(Enlarge Image)
Figure 1.
Three weeks after Descemet's membrane endothelial keratoplasty. (A) Diffuse illumination. (B) Slit beam. Figures provided courtesy of Michael Straiko and Mark A Terry (Devers Eye Institute, OR, USA).
One of the primary benefits of DMEK is the lack of a stromal interface with the presumed advantage of improved vision by restoring the normal anatomic layering of a single layer of stroma, rather than a double stromal layer seen in other forms of EK. Early visual results in DMEK have confirmed this assumption with median best spectacle-corrected visual acuity of 20/30 as early as 1 month after DMEK. At 3 months, 60–63% of patients see 20/25 and 92–94% see 20/40 or better. This is an improvement of approximately one line of vision over DSAEK and is achieved 3 months earlier, although most authors have not attempted to report visual acuity this early in PLK/DLEK/DSAEK. Guerra et al. offered interesting insight in a study of 15 patients who had a DSAEK in one eye and a DMEK in the fellow eye. Average best spectacle-corrected visual acuity was statistically better in the DMEK eye by approximately one line and the majority of patients felt that this eye saw better.
The stability and normal physiologic refracting surface seen in other forms of EK are maintained in DMEK as well. There is a minimal change in refractive cylinder reported at 0–0.49 diopters (D) and a mild hyperopic refractive shift of 0.32–0.49 D. This hyperopic shift has come as a surprise to many, as the hyperopic refractive shift in DSAEK of approximately 1 D has commonly been attributed to the myopic shape of a microkeratome-prepared lenticule. It appears that stromal deturgescence may also play a role in the hyperopic shift seen after DMEK and may contribute to that seen after stromal-including EK procedures.
Preparation of a DMEK graft is more challenging than previous forms of transplantation because of the extremely thin and flimsy nature of the Descemet's membrane–endothelial cell complex, which tends to scroll into a cigar shape with the endothelium facing outward. This presents a challenge from both the tissue-preparation standpoint as well as the tissue-delivery standpoint.
Tissue preparation is currently performed by most surgeons in the operating room. The tissue is typically scored in some fashion and then separated from the overlying stroma. Several techniques have been developed and modified for preparation of the donor by a number of surgeons and is commonly performed while submerged in solution. In the 'submerged cornea using backgrounds away' (SCUBA) technique, the tissue is scored in the periphery to create a Descemetorhexis. It is then submerged in a viewing chamber filled with transport media to improve visualization and handling. The Descemet's membrane is stripped from the overlying stroma with nontoothed forceps, taking care to avoid tearing the delicate tissue. It may be trephinated partway through the stripping to minimize the amount of lifting of the tissue. The graft is then delivered into the anterior chamber, usually via an injection system, and unscrolled using gentle irrigation from a cannula. A modified 'big bubble' separation technique has also been described. Owing to the fragile nature of the tissue, tearing of the donor is not uncommon and can present a problem for the surgeon, the patient and the eye bank. If the tissue is destroyed in preparation, the surgeon may not have a back-up donor and surgery may need to be postponed to a later time. This can also place a tax on the eye-banking community as healthy tissue is essentially made useless for endothelial-replacement purposes and wastes an already limited, nonreplenishable resource. Early reports indicate a tissue complication rate of 0–16%, although in one study, half of the cases were successfully converted to DSAEK using the damaged tissue with no adverse effects. Eye banks currently offer preprepared tissue for DSAEK with comparable postoperative results to surgeon-prepared tissue. Early studies of DMEK-prepared tissue show good tissue quality with storage in organ culture medium for several weeks, and DMEK 'precut' tissue may alleviate some of the intraoperative donor preparation-related complications in DMEK as it has in DSAEK.
Delivery and positioning of such a thin layer of tissue can also be challenging. Once the Descemet's scroll has been inserted into the eye, it must be gently unscrolled while minimizing contact with and damage to the endothelium. Orientation and even successfully identifying orientation of the graft can be a challenge.
One of the new challenges that has accompanied the many benefits of EK has been the novel complication of graft detachment. Early experience in DMEK has demonstrated that air reinjection may be necessary for peripheral graft edge nonadherence in addition to more extensive graft detachments. In one study, only one out of 60 eyes had total graft detachment, but air reinjection was performed in 63% of eyes because of concern that the graft would not appropriately adhere. This is different than in other forms of EK (DSAEK, DLEK and PLK) in which peripheral edge lift can often be tolerated because these will seal on their own. The concept of 'graft detachment' or 'dislocation' may become less relevant and the concept of 're-bubbling' may become more relevant, as this is ultimately what the surgeon and patient experience as a true complication requiring intraocular intervention. Current air reinjection rates (re-bubbling) in these earlier forms of EK can be expected to be less than 10% in experienced hands.
Primary graft failure is the other major graft-related complication and has also been a challenge in endothelial transplantation. Early experience has shown graft failure in DMEK to be in the 8% range by multiple authors. This is lower than that experienced by many starting surgeons in other forms of DSAEK (experienced EK surgeon using a novel technique vs novice surgeon using novel technique), but higher than that achieved by similarly seasoned EK surgeons in DSAEK (experienced EK surgeon using a novel technique vs experienced surgeon using an established technique). Cases of primary graft failure will presumably decline as we become better at DMEK, just as cases have in other types of EK.
One of the primary concerns with fragile tissue such as a Descemet's scroll is the concern about significant handling damage. This is particularly a concern in endothelial transplantation where damage of the tissue directly affects the cells that we hope to transplant. Without a stromal carrier, the more delicate donor tissue of DMEK might be expected to fair less well than that used in DSAEK. Reports are limited at this time, but early results of endothelial cell loss following DMEK in experienced hands demonstrates results similar to early reports in DSAEK, with 32–34% cell loss at 6 months and then 9% loss per year for up to 4 years of post-operative follow-up. Endothelial cell loss after DSAEK has been further reduced by additional innovations in techniques, with recent series as low as 26% at 6 months with stability at 1 year, and further development in the DMEK technique may demonstrate the same.
Current data on graft rejection in DMEK are limited. One study found only one case of rejection over an average of 2 years follow-up in 120 eyes, yielding a 2-year graft-rejection rate of less than 1%, far lower than the approximate 7–12% found in other forms of EK over a similar time frame. This single case in DMEK presented as a typical post-graft rejection with corneal edema and Khodadoust line. It should be noted that the steroid regimen in this study included low-dose steroids beyond 1 year, whereas previous studies in DSAEK have included subgroups of patients in whom steroids were stopped before 1 year, as is representative of the varying steroid regimens used by different corneal surgeons.
New Millennium, New Transplant
A short decade ago, penetrating keratoplasty (PK) was essentially the only procedure performed to repair endothelial dysfunction. Melles et al. had just begun their early clinical trials of the novel procedure he had developed, posterior lamellar keratoplasty (PLK), which would jump-start a new field in corneal transplantation called endothelial keratoplasty (EK). Terry and Ousley were working primarily in the laboratory to establish the ex vivo safety studies that would ultimately lay the groundwork for a large clinical study of 100 eyes that would establish the benefits of selective endothelial transplantation in their version of PLK, called deep lamellar endothelial keratoplasty (DLEK).
In 2012, EK is now well established as the treatment of choice for endothelial dysfunction and has been used for Fuchs' dystrophy, pseudophakic bullous keratopathy, irido–corneal–endothelial syndrome, and other endothelial diseases. Faster visual rehabilitation, a more physiologic topography and excellent visual outcomes are but a few of its benefits over PK. Innovations in EK have exploded as surgeons search for faster ways to provide better vision. Subsequent developments by Melles et al. simplified preparation of the donor by using a Descemetorrhexis rather than an intrastromal posterior resection. Gorovoy added the simplicity of the automated microkeratome for preparation of the donor tissue. These provided the basis for Descemet's stripping automated EK (DSAEK), the most commonly used method of endothelial transplantation today. Goals of transplantation have gone far beyond attempting to rectify frank opacity and corneal irregularity. Researchers have begun to look into models for reliably predicting refractive outcomes after EK.
Even now, current popular techniques do not restore the eye's native anatomy. DSAEK introduces a stroma-to-stroma interface which has been blamed for the lower-than-expected rates of visual acuity in the 20/20–20/25 range. A pure Descemet's membrane–endothelium complex transplant might address some of these issues, the first of which was performed by Melles et al. in 2006 and named Descemet's membrane EK (DMEK).
DMEK has become a new focus of study in EK (Figure 1). Although innovations in DSAEK remain strong as surgeons attempt to optimize the surgical course and minimize operative complications in a procedure that is already relatively easy and speedy compared with previous lamellar procedures, the benefits to patients have essentially been fully realized from the visual standpoint. DMEK may offer another option for endothelial replacement that can provide better visual outcomes. Of course, as with all new procedures, it is currently being performed by a handful of surgeons and many new challenges accompany the prospective benefits of this new iteration in endothelial transplantation. In this article we review the current published literature on this newest iteration of EK.
(Enlarge Image)
Figure 1.
Three weeks after Descemet's membrane endothelial keratoplasty. (A) Diffuse illumination. (B) Slit beam. Figures provided courtesy of Michael Straiko and Mark A Terry (Devers Eye Institute, OR, USA).
Postoperative Vision
One of the primary benefits of DMEK is the lack of a stromal interface with the presumed advantage of improved vision by restoring the normal anatomic layering of a single layer of stroma, rather than a double stromal layer seen in other forms of EK. Early visual results in DMEK have confirmed this assumption with median best spectacle-corrected visual acuity of 20/30 as early as 1 month after DMEK. At 3 months, 60–63% of patients see 20/25 and 92–94% see 20/40 or better. This is an improvement of approximately one line of vision over DSAEK and is achieved 3 months earlier, although most authors have not attempted to report visual acuity this early in PLK/DLEK/DSAEK. Guerra et al. offered interesting insight in a study of 15 patients who had a DSAEK in one eye and a DMEK in the fellow eye. Average best spectacle-corrected visual acuity was statistically better in the DMEK eye by approximately one line and the majority of patients felt that this eye saw better.
Postoperative Refractive Outcomes
The stability and normal physiologic refracting surface seen in other forms of EK are maintained in DMEK as well. There is a minimal change in refractive cylinder reported at 0–0.49 diopters (D) and a mild hyperopic refractive shift of 0.32–0.49 D. This hyperopic shift has come as a surprise to many, as the hyperopic refractive shift in DSAEK of approximately 1 D has commonly been attributed to the myopic shape of a microkeratome-prepared lenticule. It appears that stromal deturgescence may also play a role in the hyperopic shift seen after DMEK and may contribute to that seen after stromal-including EK procedures.
Intraoperative Graft Complications
Preparation of a DMEK graft is more challenging than previous forms of transplantation because of the extremely thin and flimsy nature of the Descemet's membrane–endothelial cell complex, which tends to scroll into a cigar shape with the endothelium facing outward. This presents a challenge from both the tissue-preparation standpoint as well as the tissue-delivery standpoint.
Tissue preparation is currently performed by most surgeons in the operating room. The tissue is typically scored in some fashion and then separated from the overlying stroma. Several techniques have been developed and modified for preparation of the donor by a number of surgeons and is commonly performed while submerged in solution. In the 'submerged cornea using backgrounds away' (SCUBA) technique, the tissue is scored in the periphery to create a Descemetorhexis. It is then submerged in a viewing chamber filled with transport media to improve visualization and handling. The Descemet's membrane is stripped from the overlying stroma with nontoothed forceps, taking care to avoid tearing the delicate tissue. It may be trephinated partway through the stripping to minimize the amount of lifting of the tissue. The graft is then delivered into the anterior chamber, usually via an injection system, and unscrolled using gentle irrigation from a cannula. A modified 'big bubble' separation technique has also been described. Owing to the fragile nature of the tissue, tearing of the donor is not uncommon and can present a problem for the surgeon, the patient and the eye bank. If the tissue is destroyed in preparation, the surgeon may not have a back-up donor and surgery may need to be postponed to a later time. This can also place a tax on the eye-banking community as healthy tissue is essentially made useless for endothelial-replacement purposes and wastes an already limited, nonreplenishable resource. Early reports indicate a tissue complication rate of 0–16%, although in one study, half of the cases were successfully converted to DSAEK using the damaged tissue with no adverse effects. Eye banks currently offer preprepared tissue for DSAEK with comparable postoperative results to surgeon-prepared tissue. Early studies of DMEK-prepared tissue show good tissue quality with storage in organ culture medium for several weeks, and DMEK 'precut' tissue may alleviate some of the intraoperative donor preparation-related complications in DMEK as it has in DSAEK.
Delivery and positioning of such a thin layer of tissue can also be challenging. Once the Descemet's scroll has been inserted into the eye, it must be gently unscrolled while minimizing contact with and damage to the endothelium. Orientation and even successfully identifying orientation of the graft can be a challenge.
Postoperative Graft Complications
One of the new challenges that has accompanied the many benefits of EK has been the novel complication of graft detachment. Early experience in DMEK has demonstrated that air reinjection may be necessary for peripheral graft edge nonadherence in addition to more extensive graft detachments. In one study, only one out of 60 eyes had total graft detachment, but air reinjection was performed in 63% of eyes because of concern that the graft would not appropriately adhere. This is different than in other forms of EK (DSAEK, DLEK and PLK) in which peripheral edge lift can often be tolerated because these will seal on their own. The concept of 'graft detachment' or 'dislocation' may become less relevant and the concept of 're-bubbling' may become more relevant, as this is ultimately what the surgeon and patient experience as a true complication requiring intraocular intervention. Current air reinjection rates (re-bubbling) in these earlier forms of EK can be expected to be less than 10% in experienced hands.
Primary graft failure is the other major graft-related complication and has also been a challenge in endothelial transplantation. Early experience has shown graft failure in DMEK to be in the 8% range by multiple authors. This is lower than that experienced by many starting surgeons in other forms of DSAEK (experienced EK surgeon using a novel technique vs novice surgeon using novel technique), but higher than that achieved by similarly seasoned EK surgeons in DSAEK (experienced EK surgeon using a novel technique vs experienced surgeon using an established technique). Cases of primary graft failure will presumably decline as we become better at DMEK, just as cases have in other types of EK.
Postoperative Endothelial Cell Loss
One of the primary concerns with fragile tissue such as a Descemet's scroll is the concern about significant handling damage. This is particularly a concern in endothelial transplantation where damage of the tissue directly affects the cells that we hope to transplant. Without a stromal carrier, the more delicate donor tissue of DMEK might be expected to fair less well than that used in DSAEK. Reports are limited at this time, but early results of endothelial cell loss following DMEK in experienced hands demonstrates results similar to early reports in DSAEK, with 32–34% cell loss at 6 months and then 9% loss per year for up to 4 years of post-operative follow-up. Endothelial cell loss after DSAEK has been further reduced by additional innovations in techniques, with recent series as low as 26% at 6 months with stability at 1 year, and further development in the DMEK technique may demonstrate the same.
Rejection
Current data on graft rejection in DMEK are limited. One study found only one case of rejection over an average of 2 years follow-up in 120 eyes, yielding a 2-year graft-rejection rate of less than 1%, far lower than the approximate 7–12% found in other forms of EK over a similar time frame. This single case in DMEK presented as a typical post-graft rejection with corneal edema and Khodadoust line. It should be noted that the steroid regimen in this study included low-dose steroids beyond 1 year, whereas previous studies in DSAEK have included subgroups of patients in whom steroids were stopped before 1 year, as is representative of the varying steroid regimens used by different corneal surgeons.
