The truth about DMEK

Proponents of DMEK say this technique could produce better results for patients than DSAEK. Three respected corneal surgeons explain why.

For years, Descemet stripping automated endothelial keratoplasty (DSAEK) has been the go-to technique for corneal transplant procedures. Now, proponents of Descemet’s membrane endothelial keratoplasty (DMEK) say this technique has the potential to produce better results for patients. So, why haven’t more corneal surgeons transitioned to this procedure?
In this roundtable discussion, three corneal experts discuss various misconceptions surrounding DMEK, the learning curve and what it will take for more surgeons to adopt this technique.
Francis Mah, MD, moderator: The focus of this roundtable is why more corneal surgeons have not transitioned to DMEK. Would you care to start, Dr. Price?

Francis Price, Jr., MD: Compared to DSAEK, there really has been no company that’s promoted DMEK. DMEK is a harder procedure than what preceded it, at least at the beginning, and I think people are reticent to change. And, like phacoemulsification, if you’re only doing two surgeries a month, you’re not going to keep your skill up. I think that’s been one of the issues, so the low-volume people have a hard time getting into it, and the high-volume people have been able to get by without it. But, I think the data are becoming overwhelming, and we’ve hit the tipping point where I think we’re going to see more people doing it.

Mark Terry, MD: I agree with Dr. Price; the perception is DMEK is harder than DSAEK. There’s a real issue in the learning curve in that you need more experience to get better at it, and it’s frustrating when you’re not better at it. So, if you’re not doing that many transplants a month, then it’s hard to improve.

About five years ago, I published an editorial1 examining why DMEK wasn’t more popular. It came down to the same reasons why surgeons continued to do PK and didn’t pick up DLEK (deep lamellar endothelial keratoplasty) 18 years ago. It’s a matter that the surgeons want to have a procedure that is easier, faster in the operating room and gives their patients better results than their current surgery.

Many of us have shown that DMEK definitely gives better vision, visual acuity and visual recovery, and it also gives you a lower rejection rate — all the reasons you should be doing it for the patient’s benefit. But surgeons still perceive that DMEK is not faster and not easier than DSAEK, so they are hesitant to switch.

I think that we’ve come a long way in the past five years since that editorial was published. We now have two or three standardized techniques by various high-level surgeons and high-volume surgeons that work with the standardized technique, and we now have better preparation of the tissue. When you think about the meteoric rise of DSAEK, it occurred when the eye banks started pre-cutting the tissue, and that’s when so many surgeons got on board because they didn’t risk losing tissue in the operating room. Now, we’ve got a similar situation with DMEK: you have the eye bank completely preparing the tissue for you: pre-stripping, pre-marking, pre-trephinating, pre-staining and now pre-loading the tissue in the injector for you. Now, all those factors that would normally risk losing the tissue or slowing down the operating room are taken on by the eye banks. Having the tissue prepared by the eye bank should make surgeons much more accepting of this procedure.

The only thing the surgeon has to do now is just learn the steps to un-scroll the tissue and put it up in place. There are now standardized techniques for doing this that have been published by Melles, by Price, by Gorovoy and by our group at Devers.

Dr. Mah: I think a major issue for DSAEK surgeons, who don’t wish to move to DMEK, is the dislocation rates associated with DMEK. What would be your response regarding the dislocation rates and then the extra time associated with that, compared to DSAEK?

Dr. Terry: I think that the first thing we need to do is to separate the words “dislocation” and “interface fluid separation.” When you have a partial tissue separation with DSAEK, you have interface fluid, and the tissue will reattach itself without intervention, even if you have a very large area of graft separation.

In DSAEK, if the tissue is completely dislocated and floating around in the anterior chamber, that’s when you need to re-bubble; the other cases are kind of questionable in terms of a re-bubble. So that’s one reason why the re-bubble rate has come down so much in DSAEK — surgeons have recognized that the tissue can have a lot of interface fluid and that it will attach spontaneously.

On the other hand, DMEK tissue doesn’t completely dislocate unless the tissue is dead or upside down. So, if the tissue has interface fluid, which a large proportion of the DMEK tissue have at the edges, then those tissues that have less than 30% interface fluid area don’t need a re-bubble most of the time. There are some nuances to that.
Unlike re-bubbling a dislocated DSAEK, with DMEK tissue, the tissue is still attached and centered. With the DMEK tissue, all you have to do is put air inside the anterior chamber. You don’t have to center the tissue or remove interface fluid, like you need to do when re-bubbling DSAEK. So with DMEK, re-bubbling can be done in the clinic, and you can re-bubble easily at the slit lamp. We published a paper with Chris Sales as first author on that technique , which takes less than five minutes to accomplish, so it’s not an ordeal to re-bubble a DMEK tissue. Equating the time and effort involved with DSAEK re-bubbling to DMEK re-bubbling is kind of a fallacy .
So, I explain to surgeons about the ease of re-bubbling a DMEK graft and I say, “It doesn’t interrupt the clinic flow doing it at the slit lamp any more than taking a stitch out on a PK.” And, you are very successful with a single re-bubble. The vast majority of these grafts will attach, and the re-bubble rate now is down with standardized techniques anywhere from 2.5%, which Mark Gorovoy reported, to 15% or 20% on most of the series. And with standardized techniques, they’re down anywhere from 6% to 10%.

Dr. Price: Like Mark, I think it’s important to distinguish between dislocations and partial detachments. In DSAEK, the graft is going to attach on its own if it’s less than 30%. In DMEK, one of the reasons we do reinjections of air is because you don’t want to get a total detachment. So, if you have a detachment that’s increasing in size or affecting the central vision with edema, then you just go in. It’s fairly easy to put more air in. A number of people are using long-acting gasses, which you can do. In our practice, a lot of our patients are from out of state. We like to use air, because it goes away faster and they can fly sooner.

But, we’re also more likely to do an early re-injection of air so they don’t have to stick around. With local patients, you can watch it and see how a partial detachment fares, whether it’s going to get better or worse. But if they’re from far away, you just re-inject air, get it to stick down and usually they’re alright.

With that said, that is extra work because you have to look at the patients more than you did with a PK and a little bit more than you did with DSAEK to make sure these things don’t happen. The flip side is once you get it to adhere, it’s just tremendously better from a rejection rate, topical steroid rate, and vision. So, it’s a tradeoff. It’s a little bit more work, but it can be a lot better for our patients.

Dr. Terry: I’d like to add a note about patients flying. Like Frank, we have a lot of patients flying in to Portland to get the DMEK done, and I was always so scared about letting them go home with any air or gas in the eye at all. I had an experience about two years ago with an attorney from Los Angeles who had to fly home for a trial, and he had a 25% fill of the anterior chamber of the gas bubble (20% concentration of SF6) still in his eye. I said, “Just stay around because I don’t know what this gas is going to do.” But, he flew home anyway against medical advice, and I had him take a picture of his eye at the gate before he got on the plane, another picture of his eye when he was at 35,000 feet in flight, and another picture when he got back to the gate. Airlines pressurize the cabins at 7,000 feet, or about that equivalent. And what I found was that his gas bubble expanded about 15% or 20% in mid-flight, and was back down to 25% when he landed. It never got lower than the pupil edge and certainly never occluded the peripheral iridectomy that we put inferiorly. He reported no symptoms.

Now I tell patients that fly in, even from the East Coast, that it’s OK to fly if their gas bubble is 25% or less. And I do give them some Cosopt (dorzolamide hydrochloride and timolol maleate) to use if they’re having discomfort on the airplane, but none of them have used the Cosopt. There was some small expansion in other patients that sent me photos mid flight, but it wasn’t enough to cause discomfort or problems for the graft.

Dr. Price:  I would personally caution people about flying with a 10 to 25% air/gas bubble.  If there is loss of cabin pressure, the risk to the eye is substantial.

Dr. Mah: One argument from prominent corneal surgeons is that results from ultra-thin DSAEK are comparable to DMEK. Do you think that’s just a rationalization, or do you think there’s really something to that?

Dr. Price: I think there are two aspects to it. First of all, the thinner they get, the closer to DMEK they get. And so, you would expect that the rejection rates will get better, the visions will get better, but you still have stroma. With everything we’re seeing with our DSAEK and DMEK cases, you’re going to get some increased rejection rate with the stroma and it probably decreases as it’s getting lower.

With that said, what they’re not talking about is the tissue-loss rates with these ultra-thins, and the eye banks are pretty good at cutting the tissue. But in our experience, when we were cutting the ultra-thins on our own, there was a significant tissue loss rate, and we really don’t have a significant tissue loss rate with the DMEK preps. And, as they’re doing these ultra-thins, I would wonder what their tissue loss rate is, and that’s something to consider.
Dr. Mah: Do you think it’s more difficult to prepare the ultra-thin DSAEK grafts than the DMEKs grafts?

Dr. Price: There’s some variability with microkeratomes, but you may feel differently, Mark. Also the supply and equipment costs are much greater for DSAEK preps than DMEK — important implications in the third world or those who have global contracts that include the tissue costs.

Dr. Terry: No, Frank, I completely agree with you. I think that surgeons tend to order tissue like they’re ordering an IOL and that scares me. This is a precious gift, and it is so true that the ultra-thin DSAEK-prepared grafts have a higher loss of tissue in the preparation than the DMEK. We look at our own processing of tissue for DSAEK, ultra-thin DSAEK and DMEK, and there’s a higher rate of tissue loss with the ultra-thin DSAEK than with the DMEK. So, I think that surgeons need to realize you can’t just order tissue at a certain thickness and expect that it’s not going to have any impact on the tissue availability.

The other assertion that “ultra-thin is just as good as DMEK” just does not jive with the data. If you look at the best paper ever written on ultra-thin DSAEK, Massimo Busin’s Ophthalmology paper, he utilized DSAEK tissue that was 70 μm or less. And if you look at the visual acuity result that he reports at six months, the number of cases that are 20/20 or better was only 26%, which is far below that of DMEK, where 45% to 60% are getting 20/20 or better vision at 6 months. There’s a paper by Paul Phillips that was published just this year looking at his very first 100 DMEK cases, and the percentage of 20/20 or better at six months was 55%. And although a few more of the ultra-thin Busin cases got to 20/20 or better in a couple of years, similar to what we published about the progressive improvement in vision with standard thickness DSAEK, there is remodeling, and the eye will help you to get better vision over time. But with DMEK, it’s so much faster..
The other thing that hasn’t been talked about is the quality of vision. We gave several presentations at the Academy and at ASCRS and other venues showing it’s not just the Snellen visual acuity, but it’s the quality of vision that patients rave about when they’re talking about their DMEK eye as opposed to their DSAEK eye. So, when you look at the data, the ultra-thin DSAEK eyes don’t even come close to the DMEK eyes in achieving that high level of vision.

Dr. Mah: What will need to happen for DSAEK surgeons to adopt or change over to DMEK?

Dr. Terry: The quality always comes to the fore. With DSAEK, it was the same thing. Patients asked, “Are you doing DSAEK? Are you doing PK?” That’s happening with DMEK, and the patients are choosing it.

I think a big factor will be surgeons saying, ‘Well, if I don’t start doing DMEK, I’m going to lose my corneal transplant practice.” But moreover, I think that with the eye banks preparing the tissue for you, cutting the tissue for you, trephining it for you, stripping it for you, marking it for you, even putting it into the injector for you, there’s not much else that you have to worry about in terms of the tissue. You just have to learn the steps of opening it up and putting it in place, and I think that more and more surgeons are open to that. They say, “If I can do a DSAEK, I sure can learn how to do this DMEK,” and more and more surgeons I think are going to come to the fore.

Dr. Mah: Do you think DMEK will ever take over for DSAEK? Or are there newer techniques that might make DSAEK obsolete?

Dr. Price: I think there’s always going to be room for DSAEK in complicated cases, just like there’s always room for PKs. But for uncomplicated cases, it gets back to what the surgeons say. Some say, “Well, the visions I get with my DSAEK are good enough for my patients. They don’t mind it.” It’s typically patients who haven’t been told about the difference. And if you tell people, “Well, I can do this procedure where you’re going to get your vision back pretty quickly, or this other one, it’s good, but will take two years but you’ll get 20/20.’ Which will they choose? They really need to present it that way.

Dr. Terry: I think that what we’re coming to is a point where if you’re going to be a surgeon that does corneal transplantation, you need to know how to do PK, DSAEK and DMEK. And, we also have to bring up the topic: what do you think is going to happen to DMEK, with the introduction of Descemet’s membrane stripping and no transplant placed at all?

Dr. Price: Actually, earlier this fall, we started a randomized study of DMEK patients where we’re giving them a ROCK inhibitor or placebo, and it will be interesting to see how that goes.

But there are two problems I see with doing just the stripping alone. The first problem is they are only a 4-mm diameter area, and we know from lasers, we know from IOLs that 4 mm just doesn’t cut it — especially on the back of the cornea. I think you’re going to get too much scatter with the peripheral part.

The other problem is that some of these patients take a few months to clear, and we’ve shown that if we replace a primary failure in a month there’s no problem with the visual result. But, both Nellis and Cruz have reported that  when they wait six months to do that, the visions are not as good as primary cases that don’t fail. Because you get severe edema when you strip Descemet’s off, even in Fuchs’ when you have that severe edema for a number of months there may be some permanent changes that are occurring in the corneas where they can lose some best corrected vision.

Dr. Terry: There’s a nice paper by Peter Veldman and Kathryn Colby looking at 13 eyes where they did the Descemet stripping alone. Ten of those 13 cleared without a problem within six weeks to 12 weeks, and the other 3 did not clear. However, in another paper they reported that they rescued those by putting a DMEK graft in, and the vision recovery was just as good as their series of DMEK eyes in Fuchs’ dystrophy eyes ( So, we now have a concept that you can rescue with a DMEK those eyes that do not respond to pure stripping. You can rescue them out to three months. Once you get past three months, I think all of us start getting nervous, as Frank said, that there’s going to be permanent changes from the chronic edema of the cornea.

We’re really in this kind of a transition zone where I think that it’s not entirely clear which patients should have DMEK versus just stripping without DMEK. There may be a hybrid where we decide to do something like pure stripping with ripasudil and see what happens and wait a few weeks. Then, if the patient does not clear and gets to the point where we get nervous, such as six weeks, then we do a rescue with DMEK.

And I know that Kathryn Colby in Chicago is doing a randomized study looking at DMEK versus the stripping with ROCK inhibitor, and I think they’re going out a little further out than 4 mm. We’re going to have a lot of information in the next two years. This is a very exciting time for corneal transplantation.

Dr. Price: It really is, and there can be some huge changes in the whole dynamics of tissue and transplants.

Dr. Mah: That’s why I asked what do you think is going to happen moving forward, what’s going to push people to make a change from DSAEK. It might be one of these variations on the theme instead of DMEK itself.
Endothelial Keratoplasty: Why Aren't We All Doing Descemet Membrane Endothelial Keratoplasty?
Terry, Mark A. MD
Cornea: May 2012 - Volume 31 - Issue 5 - p 469–471
doi: 10.1097/ICO.0b013e31823f8ee2
About the participants:
Francis S. Mah, MD, is director of Cornea Service at Scripps Clinic in La Jolla, CA. 
Francis W. Price, Jr., MD, Price Vision Group, Indianapolis, and founder and president of the board, Corneal Research Foundation of America, Indianapolis.
Mark A. Terry, MD, is Director of Cornea Services, Devers Eye Institute, Portland, Ore.