Endothelial Keratoplasty:
History,
Mark A. Terry, MD
Director of Corneal Services
Devers Eye Institute,
Scientific Director
Lions Vision Research Laboratory of
Endothelial Keratoplasty: History,
The History of Endothelial Keratoplasty:
Selective endothelial keratoplasty (EK) has been the focus of intense investigation over the past decade, with an acceleration of technique modifications over the past 4 years. Initially, there were two separate approaches to posterior lamellar keratoplasty. Several surgeons revived a technique described long ago by Barraquer, Polack, 1 and others whereby an anterior flap was created either manually or with a microkeratome, the flap was retracted, and the posterior recipient stroma was trephined out. The tissue was replaced with a donor posterior lamellar button and the overlying flap was then sutured into place.2-4 While this approach was attractive to the corneal surgeon because it only required familiar surgical skills and promised a microkeratome smooth stromal interface, it ultimately fell into disfavor due to the problems associated with surface sutures, irregular astigmatism, flap problems and unpredictable corneal topography.2-7
The concept of a scleral-limbal approach to endothelial keratoplasty was first described by Ko and Feldman et al in 1993.8 They were the first to show proof of concept with their successful animal studies. This work was further developed by Gerrit Melles9 while he was a Cornea Fellow in
In 2002 Gerrit Melles modified his PLK procedure, reducing the incision size to 5 mm and advocated that the tissue be folded in half for insertion. His single case report demonstrated proof of concept for a folded graft to clear the cornea, but the resultant endothelial cell count was considerably less than his prior large incision series.17 After 38 cases of large incision DLEK, we adopted Melles’ small incision idea, moved the incision to the temporal side, and re-submitted our protocol to our IRB for prospective analysis of the effect of folding of the graft on endothelial cell survival.18 In addition, we were the first to advocate that the donor tissue be “over-folded” into a 60%/40% taco configuration in order to avoid the complication of the tissue unfolding “up-side-down”. Our subsequent prospective report on our first 100 DLEK cases with 100% follow-up at 6 months post-operatively demonstrated that folding and unfolding of the graft did not significantly increase the endothelial cell loss compared to the large incision technique where the graft is not folded.19 Yet this same series established that dislocated grafts suffered a significantly greater endothelial cell loss than grafts that did not dislocate.20 At this same time, other surgeons in the EKG were also advocating valuable technique changes to the DLEK procedure. Rob Shultze from
It became apparent from prospective data from several surgeons in the EKG that the interface after DLEK surgery was likely responsible for a limitation in the average Snellen visual acuity that could be attained with DLEK surgery. The average visual acuity in most series was about 20/40 to 20/50, with a dearth of 20/20 results.19,23-26 In an attempt to improve the smoothness of the recipient interface, Melles described a laboratory and three patient eye study whereby he stripped Descemet’s membrane from the recipient to provide a bed to stick the donor tissue directly onto the posterior surface, eliminating the recipient stromal dissection from the endothelial keratoplasty procedure.27 This accomplished two key objectives: it produced an easier procedure and it provided a possibly better optical interface. Our laboratory work has since confirmed by scanning electron microscopy that the recipient bed after Descemet’s stripping is significantly smoother than after DLEK.28 Frank Price was the first surgeon in the United States to publish clinical results of this Descemetorhexis technique, and he re-named it Descemet’s stripping endothelial keratoplasty (DSEK) .29 But the ease of the DSEK technique came with the price of a dramatic increase in dislocation rate compared with DLEK surgery, with Price and other surgeons reporting a 50% or higher dislocation rate in their initial cases.30 Much of the recent work in this field has been directed at reducing the rate of occurrence of this significant complication.
Current Issues of Endothelial Keratoplasty
In 2003, we published and presented for consideration five goals for the ideal endothelial keratoplasty procedure.31 We described them as the following: 1. A smooth surface topography without significant change in astigmatism from pre-op to post-op, 2. A highly predictable and stable corneal power, 3. A healthy donor endothelium that resolves all edema, 4.A tectonically stable globe, safe from injury and infection, and 5. An optically pure cornea. In retrospect, a 6th goal should likely also have been added, and that is: 6.A surgical technique that is quickly and easily acquired. Over the past six years, the data that has been published independently by Melles,11 Terry,14,18-20,25,26, Price29,30 and Fogla24 has strongly supported the achievement of the first 4 goals in this list. The fifth goal of an optically pure cornea has been more elusive, and the new sixth goal is gradually being achieved through the evolution of the procedure. Endothelial keratoplasty is gradually becoming mainstream as the surgery has become easier to perform.
Ease of Surgery
The steps in DLEK which are most unfamiliar to PK surgeons include deep lamellar dissections for the donor and the recipient, excision of the recipient posterior stromal tissue, and atraumatic insertion and unfolding of the donor tissue. The utilization of air, rather than viscoelastic, to maintain space is also a lost art and yet is a necessary step in EK to secure the donor tissue. It has been my experience with surgeons that I have trained in DLEK, that those surgeons who are experienced and accomplished anterior lamellar surgeons with large surgical lamellar case volume have had the shortest learning curve for DLEK surgery. This scenario is more common outside the
When Melles introduced the technique modification of Descemet’s membrane stripping to EK surgery,27 this eliminated the more difficult step of recipient lamellar dissection and made EK surgery easier. In conjunction with the use of a microkeratome for preparation of the donor, the EK procedure now required no new lamellar dissection skills for the transplant surgeon, and these two technique modifications opened up the EK procedure to a wider arena of physicians. In addition, the possibility of complications with lamellar dissections that had been found in DLEK surgery, such as anterior perforation and pre-mature posterior perforation, were virtually eliminated with recipient Descemet’s stripping and use of the microkeratome for donor preparation. EK surgeons should be cautioned, however, that “button holes” and intra-operative destruction of the donor tissue have been reported with the microkeratome, and careful selection of the microkeratome head setting is important.32
Ease of EK surgery may soon be further enhanced by the widespread use of donor tissue which has been “pre-cut” with a microkeratome or femtosecond laser by the distributing Eye Bank prior to delivery to the surgeon. While preliminary unpublished data and my personal experience with pre-cut tissue is encouraging, it has not been established by a clinical scientific protocol, however, whether or not this pre-cut tissue will perform as well as tissue prepared at the time of surgery. If the dislocation rate is higher, or the endothelial survival is worse with pre-cut tissue, then any advantage in ease of use will be severely mitigated.
Although EK does not now require lamellar dissection skills, this should not be misconstrued to mean that EK is now an “easy” procedure. There is still a learning curve to the steps of atraumatic donor manipulation and insertion, air bubble manipulation, and prevention of donor dislocation. Surgeons contemplating adding EK to their surgical repertoire are strongly advised to read the literature, review the available teaching videos, and take an extensive hands-on course in EK before attempting their first case.
Optically Clear Cornea
There is ample evidence in the literature now on the advantages of EK over PK in the measured areas of corneal topography, spherical equivalent, and resultant spectacle corrected refractive error.11-26,29-31 It also appears that EK allows faster visual rehabilitation than standard PK.16,18,19,24,26,29 However, while the “average” visual acuity of DLEK and DSEK surgery appears to be as good as (or superior to) PK at six months post-op, the evidence to date shows that there are far less patients that achieve a level of 20/20 vision after EK surgery than what we have come to expect from our PK surgery.33-36 The obvious question in all EK reports is: “Why are not all of the patients with normal macular function and perfect topography after EK correctable to 20/20?” Similar to the studies in anterior lamellar surgery,37 the horizontal interface between the donor and recipient tissues has been indicted for this dearth of 20/20 results. However, other factors need to be explored as well.
In DLEK surgery, there is a stroma to stroma interface, and we have postulated, despite clarity seen with Slit Lamp biomicroscopy, that this stromal interface creates optical distortions which limit the final level of vision.19,26 However, even Wavefront testing has not yet been able to measure significant levels of higher order aberrations attributed to the interface after DLEK surgery. Yet in our prospective series of our initial 100 DLEK cases, we only had one case that achieved 20/20 vision at 6 months.19 It was hoped that the exquisitely smooth surface of the recipient bed in DSEK surgery would improve interface optical clarity, but in the initial 50 cases reported by Price, there was not a single case that achieved 20/20 vision after DSEK surgery.29 In our recent prospective study of our first 100 cases of DSEK (with and without microkeratome use), we feel that there is a tendency toward faster visual recovery and more cases of 20/20 vision after DSEK than our initial DLEK series, however statistical significance has not yet been achieved. Finally, in our analysis of our past 450 cases if EK, we have found that the only factors which significantly correlate with good post-operative vision are 1. Relatively good pre-operative vision and 2. Younger age of the patient. (unpublished data). Until these two confounding variables (age and pre-op vision) are controlled in a prospective study, pronouncements about 20/20 visual acuity advantages of one EK technique over another may be premature.
Other non-retinal factors which prevent EK eyes from achieving 20/20 vision may include posterior corneal curvature changes produced by the donor, persistent epithelial irregularity, and most recently, anterior recipient tissue light scatter. The anterior stroma of the chronically edematous cornea undergoes structural changes which may persist even after detergescence by the donor endothelium of EK. At Mayo Clinic, Bill Bourne and colleagues are able to measure increased light scatter from the anterior recipient tissue that has been left behind after EK.(Patel SV et al. “Comparison of corneal haze after DLEK and PK”, Federated Scientific Session, October 15th, 2005, Chicago) In light of these preliminary findings, it may be that EK, by its very nature of leaving behind slightly abnormal recipient tissue, will never achieve the consistency of 20/20 visual results that we desire, or that we can achieve with full thickness PK. As a practical matter, however, considering the multiple advantages of EK, consistently achieving levels of 20/25 or 20/30 may be all that is required to allow EK to replace PK as the procedure of choice for most patients and surgeons.
Dislocation rates and endothelial survival
At the current time, the primary complication resulting from EK surgery is separation and dislocation of the donor disc from the recipient bed. The rate of dislocation in DLEK surgery is documented at 4% in our first 100 cases of DLEK.19 The initial rate of dislocation in DSEK surgery, on the other hand, has been reported as high as 50% in the first cases of experienced EK surgeons.28,30 Technique modifications have been advocated to reduce the unacceptable dislocation rate from these initial DSEK cases. Mark Gorovoy reports a 25% rate of dislocation using the technique of a prolonged time of a full chamber air bubble.38 Price has utilized surface stroking to remove interface fluid combined with longer supine positioning postoperatively, and reduced his dislocation rate from 50% down to 13%. Upon adding to this regimen 3 to 4 stab incisions through the surface of the cornea into the interface to release even small amounts of interface fluid, his dislocation rate went down to 6%.30 We have not used surface stab incisions or large post-op air bubbles in our DSEK surgery. Instead, we have advocated gently scraping the stripped surface of the peripheral recipient bed in DSEK in order to expose the peripheral stromal fibrils, and promote donor edge adhesion. We have experienced 4 dislocations (4%) in our first 100 DSEK cases using this peripheral scraping technique.28
Obviously, many different techniques have evolved to reduce the incidence of donor dislocation after EK. Donor adherence is likely most dependent upon a healthy donor endothelium that begins pumping fluid from the overlying cornea as soon as possible. Therefore, more important than any precautious surgical maneuvers to prevent dislocation, the preservation of the health of the endothelium during donor preparation and insertion is paramount. How the tissue is mounted and dismounted on the artificial anterior chamber, how it is folded and inserted, and how it is unfolded and manipulated after insertion are all likely factors in endothelial function and donor adherence after transplantation. These are also the steps of the learning curve in DSEK that likely contribute to the increase in dislocation rate for the inexperienced EK surgeon. The independent data from large DSEK series by Price30 and also by Terry28 suggest that the current acceptable dislocation rate should be at least less than 10% in DSEK surgery. Aggressive modifications of technique such as “triple folding” of the endothelial tissue for injection and/or squeezing the donor tissue through a 3.0 mm recipient incision need to demonstrate an equal level of safety, with comparably low dislocation rates (<10%) and primary graft failure rates (<1%). Any perceived advantage for the surgeon using technique modifications must always be balanced against documented and measurable patient disadvantages and complications.
The donor endothelial cell loss after DLEK recorded at 6 and 12 months after surgery is about 25% and at 2 years is about 29%.19,39 In the initial series of DLEK surgery by Fogla, he reported an even lower rate of endothelial cell loss at 6 months of only 17% and no dislocations.24 The endothelial cell loss less than a year after DSEK is 40%, as reported by Gorovoy.38 This is the first report of endothelial cell counts after DSEK surgery in only 16 cases and so it is not possible to know if the higher cell loss from DSEK compared to DLEK is due to the difference in surgery or simply inter-surgeon variation. However, it is important to note that the dislocation rate in the Gorovoy DSEK series was 25%, 4% in our DLEK and our DSEK series, and 0% in the Fogla DLEK series.19,20,24,28,38 Dislocation is not an innocuous event, and the manipulations used to re-attach a dislocated donor disc result in a significant further reduction in endothelial cell counts.19,20 Long term, prospective data is needed to know whether the cell loss after DSEK is equivalent to what is seen after DLEK or PK.
The Future of Endothelial Keratoplasty
Dramatic changes in technique have taken place since Ko et al first described limbal based EK surgery in 1993.8 We now have a procedure in DSEK that can be safely executed by most transplant surgeons and which gives us at least one side of the interface which is exquisitely smooth. Further development in EK will be directed at perfecting the donor smoothness in order to achieve more cases that result in 20/20 vision. The microkeratome and the femtosecond laser are not currently capable of deep stromal dissections which are as smooth as the Descemet’s stripped recipient interface side.40-43 While Melles has shown that it is possible to strip donor Descemet’s membrane and transplant it to a smooth stripped recipient bed in the laboratory, 44 this work has not been successfully repeated in the clinical realm in any published reports. Pure Descemet’s membrane is quite fragile and manipulations of donor tissue which are well tolerated in the current EK techniques result in wrinkles, folds, tears, and unacceptable endothelial cell death when applied to pure Descemet’s transplantation. Hopefully, newer techniques and instrumentation will overcome these challenges.45,46
The current work with in-situ human corneal endothelial cell regeneration is exciting and complementary to the evolution of the surgical techniques of EK. Dimitri Azar’s group in Boston47 and Ray Tsai’s group in
These are exciting times for the corneal transplant surgeon. The 21st century heralds an era of custom keratoplasty, selectively treating only the diseased portion of the cornea and leaving the normal portions intact.50 The technical and laboratory innovations which have taken place this past decade and which are on the horizon promise a leap forward in our ability to provide quick and excellent visual rehabilitation for our patients suffering from endothelial dysfunction.
References
8. Ko, W, Freuh, B, Shield, C, Costello, M, Feldman, S.: Experimental posterior lamellar transplantation of the rabbit cornea. (ARVO abstract) Invest. Ophthalmol Vis Sci 34: 1102, 1993
23. Amayem AF, Terry MA, Helal MH, Turki WA, El-Sabagh H, El-Gazayerli E, Ousley PJ. Deep Lamellar Endothelial Keratoplasty (DLEK): surgery in complex cases with severe preoperative visual loss. Cornea 2005; 24: 587-592
24. Fogla R, Padmanabhan P. Initial results of small incision deep lamellar endothelial keratoplasty (DLEK). AJO 2006; 141:346-351
33. Mamalis N, Anderson CW, Kreisler KR, et al. Changing trends in the indications for penetrating keratoplasty. Arch Ophthalmol 1992;110:1409-11.
40. Terry MA, Ousley PJ, Wills B. A practical femtosecond laser procedure for DLEK endothelial transplantation: Cadaver eye histology and topography. Cornea 2005 : 24: 453-459