The surgical treatment of endothelial dysfunction has undergone rapid change in the past decade.1 Currently, Descemet stripping automated endothelial keratoplasty (DSAEK) has become the procedure of choice for repair of endothelial diseases such as Fuchs endothelial dystrophy, pseudophakic bullous keratopathy, and endothelial graft failure.
The advantages of DSAEK over penetrating keratoplasty (PKP) include faster visual recovery; relative refractive neutrality; increased intraoperative safety; and a smaller incision (5.0 mm), which leaves the eye stronger and less prone to traumatic rupture.2,3 In addition, most patients prefer DSAEK to PKP in the operative and postoperative experience.4 The superiority of DSAEK is evidenced by its remarkable adoption rate over the past few years. In 2007, DSAEK represented more than one third of the grafts performed in the United States and about 85% of the grafts performed to treat endothelial dysfunction (2007 Eye Banking Statistical Report, available from the Eye Bank Association of America at http://www.restoresight.org/pdfs/2007statreport.pdf. Accessed June 7, 2009).
A further refinement of endothelial transplantation, Descemet membrane endothelial keratoplasty (DMEK), has been developed.5 Melles et al.6 reported the first use of DMEK in 2006. This procedure involves stripping the recipient's Descemet membrane and replacing it with Descemet membrane only and the adherent endothelial layer of the donor. The main advantage of DMEK is the potential to provide superior vision.7,8 In the first 50 cases reported by Ham et al.,7 75% of the patients achieved 20/25 or better acuity within 6 months. The published literature on DMEK is limited. In a prospective study of 60 consecutive DMEK cases performed with corneas that had been stored in Optisol GS (Bausch & Lomb),8 we found that DMEK provides superior visual outcomes but also poses significant technical challenges compared with DSAEK. First, more donor tissue is lost with DMEK because donor tissue preparation is more difficult. Second, without attached stroma to provide stability, the endothelium–Descemet membrane tends to curl into a scroll and it is challenging to uncurl and correctly orient it in the recipient eye. Third, it is harder to achieve complete adherence with DMEK grafts. Finally, the regraft rate with DMEK is relatively high; Ham et al.7,9 reported a rate of 20% in the first 50 cases.
Studeny presented a variation of DMEK in which he left a narrow rim of stroma attached to the otherwise bare endothelium–Descemet membrane donor graft (P. Studeny, MD, “Update on Posterior Lamellar Keratoplasty/Descemet Membrane Transplantation (DSEK/DSAEK/FS-DSEK and DMEK,” course presented at the annual meeting of the American Academy of Ophthalmology, Atlanta, Georgia, USA, November 2008). He used hand dissection of the donor graft, as for deep lamellar endothelial keratoplasty and Descemet stripping endothelial keratoplasty.1,2 We have taken the logical next step and used a microkeratome instead of hand dissection. The hybrid graft potentially combines the stability of a DSAEK graft with the visual results of DMEK. Our term for this variation of Studeny's technique is Descemet membrane automated endothelial keratoplasty (DMAEK).
Donor Tissue Preparation
The donor tissue preparation consists of 4 steps: microkeratome lamellar dissection, creation of a big bubble to separate the central 6.0 to 7.0 mm of Descemet membrane and endothelium from the stroma, excision of the anterior stroma from the area of Descemet membrane detachment, and trephination to the desired graft diameter. By comparison, DSAEK donor preparation consists of 2 steps, lamellar dissection and trephination.
For the lamellar dissection, the donor cornea is mounted on an artificial anterior chamber (Moria ALTK, Moria). The epithelium is removed, and the central corneal thickness is measured by ultrasonic pachymetry. If the thickness is greater than 550 μm, a 350 μm head is chosen for the microkeratome (Moria); otherwise, a 300 μm head is chosen. The lamellar dissection is made with the microkeratome, and the free cap is replaced. The gutter is marked with a gentian violet marker to facilitate subsequent centering of the cornea on the cutting block.
A video demonstration of the following procedure is available online (Video 1). The donor is removed from the artificial anterior chamber and placed endothelial side up on the cutting block of the Hanna trephine system (Moria). A modification of the big-bubble technique10 is used to detach Descemet membrane from the posterior stroma using a 25-gauge needle bevel-up on an air-filled 5 cc syringe. Starting just outside the limbus within the scleral rim, the needle is tunneled at an oblique angle into the peripheral posterior corneal stroma. Air is injected until a bubble of Descemet membrane detaches (Figure 1). If no bubble is obtained, the rim is rotated and a new site is used. The target size of the big bubble is 6.0 to 7.0 mm in diameter.
An ophthalmic viscosurgical device (OVD), sodium hyaluronidase or methylcellulose, is used to coat the endothelium and the well of the artificial anterior chamber. The anterior cap is removed and the donor tissue placed endothelial side down on the artificial anterior chamber. The artificial anterior chamber is assembled without irrigation running through the system. A 15-degree sharp blade is used to make an incision through the stromal bed overlying the big bubble. The OVD protects the endothelium. Because there is no back pressure, Descemet membrane remains distended posteriorly, which allows a sufficiently large incision for safe access with a scissors. Trypan blue is used to fill the inside of the big bubble and stain Descemet membrane. Because the dye tracks out to the edge of the big bubble, the scissors can be placed near the edge of the separated stroma without inadvertently cutting or perforating Descemet membrane. Anwar corneoscleral scissors (Duckworth & Kent Ltd) is used to carefully cut away the disk of posterior stroma within the Descemet membrane detachment (Figure 2).
The anterior cap is replaced on the donor cornea, and the donor tissue is removed from the artificial anterior chamber and placed endothelial side up and centered on the Hanna trephine system. The tissue is punched with an 8.5 or 9.0 mm trephine, ensuring that the trephine is outside the Descemet detachment area (Figure 3). The tissue is then placed in tissue storage solution while the recipient is prepared.
Surgery is performed using topical anesthesia with monitored intravenous sedation. Silk 4-0 scleral traction sutures are placed superiorly and inferiorly to provide stability. The recipient epithelium is lightly marked with the trephine size used to punch the donor tissue. A 5.0 temporal scleral tunnel frown or 3.5 mm clear corneal incision is made. Two limbal paracentesis wounds are made, and the anterior chamber is filled with air. A modified Price-Sinskey hook (Moria) is used to score and strip Descemet membrane, while keeping the area of stripping inside the epithelial trephine reference mark. The eye is filled with balanced salt solution, and an inferior peripheral iridotomy is created using intraocular scissors (Duet, MicroSurgical Technology). The anterior chamber is then filled with trypan blue to identify residual tags of Descemet membrane. The tags as well as the posterior pigmented layer of the peripheral iridotomy are removed with bimanual irrigation and aspiration handpieces (American Surgical Instruments Corp.).
The donor cornea is carried into the operative field on a Patton spatula (Bausch & Lomb). A small amount of methylcellulose OVD (OcuCoat) is placed on the endothelial surface. The donor tissue is placed endothelial side up on a Busin funnel glide (Moria) and a micrograsping Tan forceps (American Surgical Instruments Corp.) is used to pull the tissue into the funnel. Using a bimanual technique, the forceps is inserted through the paracentesis, across the anterior chamber, and out the scleral tunnel wound. The donor tissue is pulled into the anterior chamber carefully to maintain the correct endothelial orientation (Figures 4 and 5). Air is injected through virgin limbus using a 30-gauge needle to push the donor tissue against the recipient cornea. The anterior chamber is completely filled with air, while making sure that the patient is able to perceive light. A Lindstrom laser in situ keratomileusis roller (BD Medical) is used to center the graft and remove residual fluid from the donor–recipient interface. A drop of homatropine 5% is administered. Patients are then moved to the recovery room and lie face up for 60 minutes. The intraocular pressure is checked in the clinic to be sure that air is not blocking the peripheral iridotomy. Postoperative instructions include tobramycin–dexamethasone ointment 4 times daily.
A 75-year-old woman with Fuchs endothelial dystrophy had DMAEK surgery in the left eye. She previously had cataract surgery in both eyes, as well as DMEK in the right eye. Preoperatively, the corrected distance visual acuity was 20/30 in the right eye and 20/40 in the left eye. Ultrasonic pachymetry measurements were 513 μm and 594 μm, respectively. Confocal microscopy of the left eye revealed undefined endothelial cell borders and indistinct cell margins, indicating endothelial dysfunction. Slitlamp examination was significant for 3+ diffuse guttata with endothelial pigment plaques and central corneal thickening.
The DMAEK surgery was uneventful, with a 9.0 mm donor graft inserted via a pull-through technique. On postoperative day 1, the uncorrected distance visual acuity (UDVA) was counting fingers at 3 feet with a near total air fill of the anterior chamber. At 1 week, the UDVA had improved to 20/25 with an attached posterior graft and a clear cornea (Figure 5). Anterior segment optical coherence tomography (Visante OCT, Carl Zeiss Meditec) showed the graft was adherent to the recipient cornea (Figure 6). The endothelial cell density, assessed at 1 month by an experienced technician using the manual centers method, was 3875 cells/mm2 (Nidek confocal microscope, Nidek Inc.). The cell count of the donor cornea from a 4-year-old child was 3993 cells/mm2.
We believe that DMAEK may offer a further advancement of endothelial keratoplasty. Overall, the visual results of transplanting Descemet membrane and endothelium without stroma via the DMEK technique are superior to those of DSAEK, allowing more patients to achieve 20/20 or 20/25 acuity.7,8 However, donor preparation, positioning, and attachment are more challenging with DMEK, resulting in more donor tissue loss and a higher regraft rate than typically experienced with DSAEK.7,8 Currently, the disadvantages of DMEK compared with DSAEK include (1) pre-cut tissue not available; (2) additional donor preparation time; (3) greater risk for losing the donor during preparation; (4) difficulty detecting which side is the endothelial side once the tissue is in the recipient eye; (5) longer operative time due to challenges with tissue unfolding and positioning; (6) higher rate of re-bubbling to obtain complete attachment; and (7) higher rate of graft failure.
The DMAEK technique addresses some of the difficulties seen with DMEK while still providing enhanced potential for 20/20 vision. The challenge with manipulation of only Descemet membrane is virtually eliminated by keeping a rim of donor stroma. Intraoperatively, the tissue behaves differently from the tissue in DMEK and similar to a DSAEK graft with regard to tissue insertion, unfolding, and positioning. One beneficial difference from DSAEK is that the aeration of the stromal rim from preparation in DMAEK helps to unfold and buoy the graft against the recipient cornea. This aids injection of air behind the graft and protection of the endothelium from hitting the iris, glaucoma tubes, or intraocular lenses and provides an early “life preserver” from posterior dislocation in aphakia. In addition, no new instrumentation is required beyond that used for DSAEK. These attributes would ease the transition for surgeons currently performing endothelial transplantation. Finally, because of the stability provided by the stromal ring, eye banks could prepare, store, and ship tissue as well as perform post-processing quality control assessments, as currently done with DSAEK (ie, slitlamp evaluation and specular microscopy).
The principal challenge with DMAEK is the donor preparation. In our experience, as in DMEK, the risk for losing the graft is higher than with DSAEK. Generating a big bubble can cause tears in Descemet membrane if the air is injected too vigorously. In addition, the bubble can be troublesome to obtain and its size can be difficult to predict reliably. If the bubble is too small, it could result in reduced visual acuity or visual disturbances from the edge of the stromal ring. It can be difficult to expand a localized small detachment without causing a tear or popping Descemet membrane. While a decentered bubble can be managed by decentering the trephination, this results in asymmetric thickness of the stromal ring. Occasionally, the big bubble can extend to the limbus, completely detaching Descemet membrane. This requires conversion to a DMEK technique for endothelial transplantation.
Trephination can sometimes pose a challenge. It can be difficult to center the stromal rim on the block because the tissue is opaque from injecting air. The trephination must be outside the edge of the Descemet detachment or an edge of the donor graft will not have a stromal rim. This would create an area more prone to fluid entering the interface, which would promote detachment. Centering the graft is facilitated by a trephination system such as the Hanna in which the surgeon can look down the barrel of the trephine with its mounting apparatus to check centering on the edges of the donor Descemet membrane. In addition, we have seen that DMAEK tissue, similar to other forms of endothelial keratoplasty tissue, can stick to the trephine after it is punched because it is so thin centrally and easily folds up in air. A slit attachment on the operating microscope can be helpful when orientation is lost during preparation or storage.
We feel that preparation of DMAEK tissue could be performed by a skilled eye-bank technician or in close association with a surgeon's local eye bank. Donors who develop complete Descemet detachments to the limbus could be converted to a DMEK procedure.
In conclusion, our experience has shown that donor endothelium and Descemet membrane with a rim of stroma can be successfully transplanted. We call this technique DMAEK when a microkeratome is used for the initial lamellar dissection. The technique offers an alternative form of endothelial keratoplasty, which may allow DSAEK surgeons to achieve the superior visual results of DMEK with easier unfolding and positioning of the graft. However, there is a steep learning curve in tissue preparation, with a risk for tissue loss. We believe eye banks may be instrumental in the preparation of donor tissue and therefore in the overall success of this technique. We are currently conducting a prospective study to evaluate the rate of success with donor tissue preparation, graft survival, visual acuity outcomes, and endothelial cell loss in an initial consecutive series of DMAEK cases.
1. Melles GRJ., 2006. Posterior lamellar keratoplasty: DLEK to DSEK to DMEK [editorial], Cornea, 25, 879-881.
2. Price MO, Price FW. Descemet's stripping endothelial keratoplasty. Curr Opin Ophthalmol. 2007;18:290-294.
3. Price FW Jr., 2005. Corneal transplantation as a refractive surgical procedure [guest editorial], J Refract Surg, 21, 216-217.
4. Kelley K, Voci E., 2009. From the patient's perspective. In: Price FW, Price MO, editors., DSEK: What You Need to Know About Endothelial Keratoplasty. Slack, Inc, Thorofare, NJ, pp. 17-23.
5. Melles GRJ, Lander F, Rietveld FJR. Transplantation of Descemet's membrane carrying viable endothelium through a small scleral incision. Cornea. 2002;21:415-418.
6. Melles GRJ, Ong TS, Ververs B, van der Wees J. Descemet membrane endothelial keratoplasty (DMEK). Cornea. 2006;25:987-990.
7. Ham L, Dapena I, van Luijk C, van der Wees J, Melles GRJ. Descemet membrane endothelial keratoplasty (DMEK) for Fuchs endothelial dystrophy: review of the first 50 consecutive cases. In press, Eye 2009.
8. Price MO, Giebel AW, Fairchild KM, Price FW. Descemet membrane endothelial keratoplasty: prospective multicenter study of visual and refractive outcomes and endothelial survival. In press, Ophthalmology 2009.
9. Ham L, van der Wees J, Melles GRJ. Causes of primary donor failure in Descemet membrane endothelial keratoplasty. Am J Ophthalmol. 2008;145:639-644.
10. Anwar M, Teichmann K. Big-bubble technique to bare Descemet's membrane in anterior lamellar keratoplasty. J Cataract Refract Surg. 2002;28:398-403.