Descemet's stripping automated endothelial keratoplasty (DSAEK) is indicated for patients with endothelial failure. Even though dense edema is not a contraindication for this procedure, it may be appropriate to select pseudophakic patients with a posterior chamber lens. Collectively, we have performed more than 500 DSAEK procedures. Specifically, Dr. Busin has performed approximately 100 and Dr. Gorovoy has performed more than 400.
The improvement in refractive results as well as the increased safety of a closed system procedure, compared with an open-sky penetrating keratoplasty, make DSAEK an important advance in refractive surgery that has revolutionized corneal transplantation for endothelial dysfunction. Before, with penetrating keratoplasty, refractive results were unpredictable and could take 1 year or more to stabilize. By contrast, the majority of DSAEK patients see 20/40 within 6 weeks. The results are so superior that penetrating keratoplasty should no longer be recommended for patients with endothelial disease.
HOW DO YOU PERFORM DSAEK?
There are several DSAEK techniques, and in fact, our techniques differ significantly.
Dr. Gorovoy's technique. DSAEK has been evolving, particularly in terms of the wound. I perform DSAEK under topical anesthesia. First, I measure the patient's corneal diameter with a marking trephine (Figure 1) to determine what size tissue I want to insert. In approximately 75% of cases, I use a 9-mm button. On rare occasions, I use a button that is 8.75-mm or smaller. The donor cornea is mounted on the artificial anterior chamber of the Moria DSAEK system (Antony, France) (Figure 2). I use the Moria CB microkeratome with a 300-µm head to cut an anterior lamellar cap, and then I punch the appropriately sized button from the posterior portion of the cornea (Figure 3). I do not mark the tissue.
Once satisfied with the donor button, I make a 3.2-mm clear corneal incision and several 1-mm paracentesis wounds in the patient's limbus. I introduce an irrigating stripper to circumferentially strip Descemet's membrane (Figure 4) just inside the trephine mark, through the paracentesis wounds. Next, I use a phacoemulsification I/A tip to aspirate the membrane and make three midperipheral full-thickness venting stab incisions, as described by Francis W. Price, Jr., MD, of Indianapolis, Indiana (personal communication). I make a 60/40 fold in the donor cornea with a small amount of Healon (Advanced Medical Optics, Inc., Santa Ana, California) on the center of the endothelial side. Using a modified Goosey forceps (Moria), I insert the folded tissue through the 3.2-mm keratome incision. To unfold it, I go back in with the I/A unit and aspirate the posterior leaflet. Once fully unfolded, I shallow the chamber slightly—making sure the tissue is well centered—and then insert a 100% air bubble into the anterior chamber. Finally, I drain the venting incisions. The eye is left high tensive to normotensive with the air bubble and no sutures.
The patient is dilated and spends 1 hour lying supine in the recovery room. Before discharge, I examine the patient at the slit lamp and burp some air out of the anterior chamber. This is done until the air bubble clears the bottom of the pupil, eliminating the risk of pupillary block. We start the patient on a fluoroquinolone antibiotic and prednisone q.i.d., and we place a shield on the eye until the next day.
Dislocation rates have been dropping as the procedure improves. In my last 100 cases, 9% dislocated, however, the rate is only 4% in eyes with normal anterior segments (ie, eyes with no communication between the anterior and posterior chambers). Dislocation is more likely to occur in eyes with shunts or previous vitrectomy. It is realistic to aim for a dislocation rate that is below 5%.
Dr. Busin's technique. I use a blunt cannula with an attached air syringe to break, strip, and remove Descemet's membrane—usually in a single piece. Donor preparation is as Dr. Gorovoy described, although I feel it is particularly important to mark the tissue, for two reasons. First, the outer limit of the area from which the anterior stroma has been excised is marked on the posterior lamella to avoid it punching through to the outer edge. Second, a central mark (I usually use the letter "F") on the stromal side of the donor lamella allows the surgeon to distinguish it from the endothelial side after the graft is inserted into the anterior chamber.
I also have a different method for insertion (Figure 5), which I find to be simpler and gentler to the endothelium. Instead of making a taco fold, I place the tissue on a specially designed glide with the endothelium up. I drag the tissue into the anterior chamber with a crocodile vitreous forceps that is inserted through a paracentesis performed opposite to the surgical wound. While sliding into the glide, the tissue folds onto itself on both sides, but the endothelium remains untouched. As the two sidefolds are approximately 20% of the entire width, unfolding is also much less problematic, even in a shallow anterior chamber. To obtain maximum adhesion, it is important to ensure that no viscoelastic remains between the donor and recipient tissue. I avoid pupillary block by performing a surgical peripheral iridectomy prior to graft insertion.
To date, I have only had four cases of dislocation in which it was necessary to postoperatively place a second air bubble. I had two primary graft failures in my first 10 cases, but I have experienced none since that time.
ADVANTAGES, DISADVANTAGES
The major disadvantage of precut tissue is that it is associated with a higher rate of graft dislocations versus tissue cut by the surgeon. Jayne Weiss, MD, of the Kresge Eye Institute in North Carolina, analyzed mid-2006 survey data on the outcome of its precut tissue grafts from the North Carolina Eye Bank.1 She found dislocation rates and graft failure rates of 34% and 23%, respectively. These rates may be partly due to the learning curve of inexperienced surgeons, however, it is unacceptably high; the primary graft failure rate should be closer to 1%.
We typically think of graft failures as being surgeon-induced, but it would be interesting to see data from the same surgeons comparing precut tissue versus their own dissections. This would show whether the tissue itself is impacting failure rates. Precut tissue is typically thicker than self cut. In some cases, it may be up to 20% thicker, and thus, difficult to handle. If you do not cut your own tissue, you do not have the opportunity to mark it, and it may be quite difficult to correctly orient in the eye. The only advantage of precut tissue is that it provides surgeons the opportunity to perform DSAEK rather than penetrating keratoplasty if Moria equipment is not available.
SWOLLEN DONOR TISSUE
Overly swollen tissue is undesirable, and swelling of precut donor tissue may contribute to dislocation. One mystery of DSAEK is that the cause of sticking tissue is unknown, although we do know that it includes the stromal-fibril interaction and intracellular adhesions. If cells are swollen and distorted, they may become less adherent, but we do not, however, fully understand this mechanism.
Stacey Gorovoy, MD, (Dr. Gorovoy's daughter) recently presented a paper at The Cornea Society and Eye Bank Association of America's Federated Scientific Session in which she analyzed a series of 148 DSAEK cases.2 In all cases, donor tissue—prepared by Dr. Mark Gorovoy—that was preserved in Optisol (Oxonica Ltd., Oxford, UK) for more than 3 days prior to dissection was twice as likely to detach than tissue stored for less than 3 days. Moreover, tissue preserved for more than 5 days was three times as likely to detach. This may be related to increased swelling associated with the time the tissue was in the media. A donor cornea that has already been dissected prior to storage is going to swell even more, because the stroma is directly bathed in the media.
A thick and swollen donor button is more difficult to handle, insert, and manipulate once inside the eye. Overmanipulation of the tissue may increase endothelial cell loss and could even lead to graft failure. If there is a lot of edema, it will also take longer for the cornea to clear, and visual rehabilitation is delayed.
COST-EFFECTIVENESS
Some surgeons are concerned with the (1) recent shortage of donor corneas and (2) significant cost for precut tissue, which may constitute 50% to 100% of the surgical fee for the entire procedure. In Europe, precut tissue costs more than a standard donor cornea. It may not be cost-effective in the long run, because the European reimbursement system pays a fixed amount per procedure, no matter what the tissue costs.
With the Moria DSAEK system, you can do more than just DSAEK. We both have found that it may be advantageous to combine a DSAEK procedure with a lamellar graft for keratoconus, therefore using two pieces of a single donor cornea. It may even be possible to use one donor cornea for three procedures if you use the posterior portion for DSAEK, the scleral rim for scleral grafts associated with shunts, and the anterior portion for an anterior graft. If the anterior portion is not needed, do not discard it. Instead, keep it as emergency reserve tissue for 1 week, and use it in case of a corneal perforation.
For surgeons new to DSAEK, a hands-on course and mentoring from an experienced surgeon are critical. Better technique and more experience with DSAEK, particularly with insertion of the donor button, will help protect the endothelium. Although the procedure is not technically any more challenging than penetrating keratoplasty, there is a learning curve to reduce rates of dislocation, graft failure, and endothelial damage. Dr. Busin saw his average endothelial cell loss go from approximately 50% to approximately 20% at 6 months, a level comparable with one observed after conventional penetrating keratoplasty.
Studies on endothelial cell loss after DSAEK are ongoing, and it appears that results will be similar to penetrating keratoplasty and should still be the procedure of choice. Even in the midst of a learning curve, the results should be dramatically superior to penetrating keratoplasty.
Mark S. Gorovoy, MD, is in private practice at Gorovoy Eye Specialists, in Fort Myers, Florida. Dr. Gorovoy states that he has no financial interest in the products or companies mentioned. He may be reached at +1 239 939 1444; mgorovoy@gorovoyeye.com.
Massimo Busin, MD, is in private practice at Villa Serena Hospital, in Forlí, Italy. Dr. Busin states that he has no financial interest in the products or companies mentioned. He may be reached at +39 347 24 49 343; mbusin@yahoo.com.
Up Front | Jun 2007
DSAEK: Current Procedures and Controversies Over Precut Tissue
Two surgeons discuss their DSAEK technique, share thoughts on donor and precut tissue.
Mark S. Gorovoy, MD, and Massimo Busin, MD
