PDEK emerging as promising but challenging alternative to DSEK, DMEK

Endothelial keratoplasty has gained prominence over penetrating keratoplasty as the preferred corneal transplantation technique for numerous indications. Proponents say the newer technique improves vision with lower complication rates and faster visual recovery than full-thickness transplantation.

Currently, surgeons primarily perform two endothelial keratoplasty techniques: Descemet’s stripping endothelial keratoplasty and Descemet’s membrane endothelial keratoplasty.

DSEK involves removal of the Descemet’s membrane and endothelium from the recipient’s stroma and replacement with donor posterior stroma, Descemet’s membrane and endothelium.

DMEK also involves removal of the Descemet’s membrane and endothelium but with replacement of only these two layers, without posterior stroma. Between these two forms of endothelial keratoplasty, many surgeons believe that DMEK offers better visual and refractive outcomes, faster visual recovery and a lower graft rejection rate than DSEK.

An emerging technique, pre-Descemet’s endothelial keratoplasty, involves transplantation of the pre-Descemet’s, or Dua’s, layer, Descemet’s membrane and endothelium.

According to proponents of the technique, PDEK offers the ability to use corneal tissue from donors younger than 40 years. In addition, PDEK grafts are thicker than those used in DMEK and thinner than those used in DSEK, providing flexibility and ease of manipulation.

Image: PriceVisionGroup

“PDEK is going to be, in my opinion, a big boon for corneal surgery because of these several advantages,” Amar Agarwal, MS, FRCS, FRCOphth, OSN APAO Edition Board Member, said. Agarwal, Harminder Dua and colleagues first performed PDEK in September 2013.

“The most significant advantage of PDEK is that it is not dependent on the corneal donor age, whereas DMEK cannot be done unless the donor is above 40 years of age, given the increased tendency for graft scrolling,” Agarwal said. “Additionally, a PDEK graft is advantageous in that it is only 25 µm to 30 µm thick, minimizing graft-host interface postoperative haze. Notably, even though the PDEK graft is thin, the pre-Descemet’s layer preservation within the graft is the critical component that allows for ease of graft movement within the eye, minimizing the risk of graft tearing.”

For the time being, DSEK and DMEK will remain the leading corneal transplantation techniques, according to Mark A. Terry, MD.

“DSEK surgery is the standby procedure that can be done for all forms of disease that require endothelial keratoplasty. That’s really the go-to procedure for most surgeons,” Terry said. “The DMEK procedure offers pure anatomic replacement, and because of that, it offers better quality of vision and faster rehabilitation than you can get with DSEK. So, I think that in the future that DMEK is going to be the dominant procedure for routine cases of endothelial replacement. But DSEK will always have a place in corneal transplant surgery for those complex cases of endothelial problems.”

Donor age

Agarwal elaborated on the ability to use younger donors in PDEK.

“The problem with DMEK is that you need a donor above 40 years. It is critical to understand that the Descemet’s membrane and endothelium cannot be separated from the pre-Descemet’s layer until the donor is at least 40 years or older,” Agarwal said. “With PDEK, I am not limited by the age of the donor. In fact, the youngest graft I have used to date was with a 9-month-old donor.”

Agarwal and colleagues published a study in Cornea on PDEK with infant donor corneas. Infant corneas restored optical clarity and yielded good visual outcomes in three eyes of three patients with pseudophakic bullous keratopathy.

“Obviously, the number of hexagonal cells and the level of uniformity in young corneas would be expected to be higher, with a tendency to decrease with age. The corneal clarity directly correlates to the number and uniformity of these hexagonal cells. Thus, inherently, a younger cornea, say that of a 5-year-old child, would be preferred over one that is 45 years old,” Agarwal said.

Agarwal noted that full-thickness corneal transplants, with the use of infant donor tissue, have a high risk for rejection. However, with PDEK, the surgeon is not likely to encounter this potential risk because grafts do not contain the full stromal layer, he said.

“The reason for the high risk of rejection in full-thickness corneal transplants in these cases is that the graft contains the entire stromal layer. The stroma contains an abundance of antigens, the source and mechanism by which rejection of the graft can occur. With PDEK, the stromal layer is not utilized. I am strictly using the pre-Descemet’s layer, Descemet’s membrane and endothelium,” Agarwal said. “In opting for PDEK over other surgical transplantation techniques, the potential for corneal clarity is maximized, while at the same time minimizing the risk of rejection. In this way, postoperative outcomes can be optimized above and beyond the outcomes of PK, ultrathin DSEK, DSEK or DMEK.”

Francis W. Price Jr., MD, OSN Cornea/External Disease Board Member, described the difficulty of using tissue from younger donors in DMEK.

“The problem with DMEK is that the young donors curl up so tightly it’s difficult to use them because they are hard to uncurl and get into place. A lot of doctors won’t use a donor less than 50 years of age, and we typically don’t like them less than 40 years of age. We’ll occasionally go lower, but when you get below 40 years of age, it can be really difficult to unfold the tissue,” Price said.

Price said he has performed a few PDEK cases but had mixed results with younger donors.

“I did have a child’s cornea that I was able to do, and I also had a young adult, somebody who I think was 20 or 21, that I couldn’t get it to unfold. So, that’s why I think it’s just going to take more numbers to see how often you have problems with it and if PDEK actually gives us a better result or not. I’m open-minded about it, but I’m not really planning to do any PDEK surgeries myself because we get really good results with DMEK,” he said.

Graft preparation

DSEK grafts vary in thickness, from 100 µm to 200 µm.

“That means if a recipient cornea is 500 µm and I inject 100 µm or 200 µm tissue into that eye, the cornea then becomes 600 µm or 700 µm. That means the endothelium has to pump more fluid out from the excess stroma and inadvertently leads to graft-host interface haze. This is the inherent problem with DSEK,” Agarwal said.

Ultrathin DSEK grafts are generally between 60 µm and 100 µm, DMEK grafts are about 15 µm, and PDEK grafts are 25 µm to 30 µm.

Agarwal said that the pre-Descemet’s layer gives the PDEK graft additional rigidity, making it easier to manipulate.

“In DSEK, I would be using 100 µm corneal grafts, whereas in PDEK, I am using grafts that are only 25 µm,” Agarwal said. “In DMEK, the graft is very thin and flimsy. It can tear off. In PDEK, because the pre-Descemet’s layer remains attached to the Descemet’s, it gives us a scaffold that makes it easy to manipulate,” Agarwal said.

“That thickness is what its advantage is because it makes it where it doesn’t curl as tightly, and it’s a little bit easier to manipulate and get it to unfold,” Price said.

For DMEK graft preparations, eye banks peel Descemet’s membrane and endothelium from stroma, establish a small hinge at the periphery of the flap and then lay the flap back down, according to Terry Kim, MD, OSN Cornea/External Disease Board Member.

“PDEK is kind of a variation of DMEK where, instead of peeling the Descemet’s membrane off the back of the donor and starting in the peripheral part by the trabecular meshwork and peeling it off, they inject air and get a big bubble to form centrally. Then, as that extends out, it gives you the graft that you can have. That will usually be thicker than what the Descemet’s is that you peel off from the outside,” Price said.

For the PDEK graft, an air bubble is used to separate the pre-Descemet’s layer, Descemet’s membrane and endothelium from stroma, Kim said.

“This air dissection incorporating pre-Descemet’s layer can be done by the surgeon from a full-thickness corneal graft and, presumably, by eye banks in the near future, using similar steps. So, all you do is trephine out the size of the graft that you want and then implant it into the anterior chamber,” he said.

Different types of air bubbles can be used to prepare endothelial keratoplasty grafts. A type 1 bubble is a centralized bubble with a diameter of about 8 mm to 8.5 mm that does not go to the corneal periphery. It is this type of bubble that is used for PDEK.

“There is a space created. On one side you have pre-Descemet’s layer, Descemet’s membrane and endothelium, which is about 25 µm thick and comprises the PDEK graft. On the other side, you have the air and everything else, which can be preserved and used for a deep anterior lamellar keratoplasty. Let’s not waste precious tissue. One cornea, in this case, can be used for two distinct procedures,” Agarwal said.

A type 2 bubble reaches the corneal periphery, with a diameter of about 10 mm.

“That means that with a type 2 bubble, a separation has been created between the pre-Descemet’s layer and Descemet’s membrane. This will yield a DMEK graft, which is about 15 µm thick,” Agarwal said.

Price said that using a bubble to separate the PDEK graft from the stroma limits graft diameter.

“The most important thing about that layer is that it only goes out about 7 mm to 7.5 mm in diameter and then it becomes fused to the rest of the stromal tissue,” Price said. “Now, the reason that’s important here is because, when you inject the air and you form the central bubble, that means your graft size is limited by the size it expanded to and probably a little bit less if you cut it out with scissors or punch with a trephine. You’re not going to be all the way to the inside of that junction area.”

Price also noted that using scissors to cut the graft tissue can be problematic.

“You have to cut the graft off with scissors. That can be a little bit more traumatic than using a trephine. I think it’s all going to come down to just long-term follow-up of a number of these eyes. It’s going to come down to how much donor endothelial cell loss you have,” Price said.

Terry also said endothelial damage can be caused by using scissors to cut grafts.

“The tissue tends to wrap around the scissors, and when it wraps around the scissors, it causes endothelial cell death. So, we need to have more data on larger numbers of patients to really know what the endothelial cell loss is acutely from the preparation of tissue before we can say this procedure is equivalent to DMEK or better than DSEK,” Terry said.

Handling and inserting tissue

A PDEK graft is easier to handle and inject than a DMEK graft, Agarwal said.

“It’s just 10 µm thicker, but because of the tissue characteristics of the pre-Descemet’s layer, it’s easy to unroll. PDEK grafts can be loaded into a normal IOL injector and inserted into the eye. PDEK behaves more like DMEK in this regard, differing from DSEK tissue. With a DSEK graft, you need a larger 5-mm opening. This is yet another huge advantage of PDEK,” he said.

“The goal for PDEK is to ensure better predictability of the graft. DMEK grafts, though beneficial from being so thin, lend themselves to an increased risk of tearing while handling the tissue. The goal with PDEK is to make the thinnest possible grafts that are more easily and consistently manageable during implantation and unfolding,” Kim said.

He said there are a few important points to address.

“No. 1, whether during preparation, loading, insertion or unfolding of a DMEK graft, bare Descemet’s membrane is fragile and easier to tear. No. 2, it’s difficult to predict what Descemet’s membrane is going to do and how it is going to behave,” Kim said.

A PDEK graft is inserted into the anterior chamber much like a DMEK graft.

“You put it in the eye and unfold it, using a similar technique to DMEK. But the difference is that PDEK tissue should behave more like thin DSEK tissue, sturdier and more manageable within the eye compared to DMEK tissue because it has pre-Descemet’s layer incorporated within the tissue,” he said.

Kim compared insertion of the different grafts.

“DSEK and ultrathin DSEK are quite predictable if you’re using a DSEK graft injector. I use an injector for all of my DSEK graft insertions, and the entire procedure is very predictable in terms of graft unfolding and graft manipulation. The injector helps me not only to insert the graft and get the orientation right side up, but it also lets me unfold it before I leave the chamber, so all I have left to do is to inject the air in under the graft and center it,” he said.

“With DMEK, on the other hand, I put the tissue in and I don’t know what’s going to happen. It may be folded. It may present with a single scroll or a double scroll. It can also be difficult to assess whether the tissue is inverted. And while I’m trying to manipulate the graft, I’m just delicately trying to tap the cornea without knowing how the DMEK graft is going to react,” he said.

Kim noted that surgeons may be able to fine-tune the positioning of PDEK grafts.

“Once you’ve put that air or gas bubble in under a DMEK graft, you’re fairly committed to the position of that DMEK graft (ie, you’re not going to detach the graft intentionally to recenter it and you cannot adjust the centration). With ultrathin DSEK grafts, once I put that air bubble in, I can move it. I can create venting incisions and move it, or I can take a LASIK flap roller and position the graft where I want it to be positioned. With PDEK, there lies a similar advantage in being able to mobilize the position of the graft,” Kim said.

Terry said that the bonding of Dua’s layer and recipient corneal stroma can be problematic in PDEK because the graft includes Dua’s layer, which is essentially stroma.

“The interface of stromal tissue to stromal tissue has been shown to not yield as pure an optical path as when you have a pure Descemet’s membrane transplant like you have with DMEK where you’re peeling the tissue,” Terry said. “When you’re peeling the tissue for DMEK, you get strictly Descemet’s membrane and endothelium as your replacement tissue. When you do PDEK, you get a thin layer of stromal tissue that we call Dua’s layer, and with DSEK, you get a thicker stromal layer. Even ultrathin DSEK cannot compare in visual quality or visual rehabilitation to DMEK. I think it’s likely that PDEK will be somewhere in between ultrathin DSEK and DMEK. It will be kind of that intermediate level of visual quality and visual recovery, but will never attain the anatomic and optical purity of DMEK.”

Agarwal and colleagues combine PDEK with glued IOL implantation in some cases.

“The glued IOL is fully fixed in the eye. Now, that acts like a trampoline, so when I inject the graft and I am putting in the air, the IOL acts like a trampoline, pushing the air against the cornea and the graft remains attached,” Agarwal said.

Donor loss rate

Price said that hybrid techniques such as Descemet’s membrane automated endothelial keratoplasty have resulted in a donor loss rate of 10% to 13% when an air bubble is used to separate Descemet’s membrane, “a pretty high rate of donor loss.” That donor loss rate has improved, but it is higher than DMEK or DSEK, he said.

Long-term data will give more insight into the donor loss rate of PDEK.

“We’ll just have to wait and see. Getting that needle right underneath Descemet’s in the center of the cornea can help in getting the separation of the big bubble, but it can still be difficult,” Price said. “We’ve done over 100 [cases] of the hybrid technique, and we trained our local eye bank to make them and I think they might still make some of them. But we didn’t really see a long-term advantage to it. They had more detachments than DMEK or DSEK. People got good vision, but then we had a higher donor loss rate, too, so that’s why we stopped. We also found that the rejection rate was kind of intermediate between DSEK and DMEK, probably because you have peripheral stroma. We’re just going to have to see long-term results for PDEK.”

PDEK trial with young donor corneas

Ashiyana Nariani, MD, MPH, of Duke University Eye Center, is collaborating with Kim and others on a prospective, double-masked, multicenter, international trial on PDEK.

“Logically, if the formula for optimizing postoperative success is to utilize the thinnest possible graft with the youngest endothelial cells, then PDEK is the solution,” Nariani said.

“The Young Pre-Descemet’s Endothelial Keratoplasty (Y-PDEK) trial will set out to evaluate PDEK outcomes with statistically significant numbers and long-term data,” Nariani said.

The Y-PDEK trial involves institutional collaboration among Duke University Eye Center, Dr. Agarwal’s Eye Hospital and Miracles in Sight, one of the largest and oldest eye banks in the United States. It was expected to be launched in November, with preliminary results scheduled for March 2016.

“One arm of the study will utilize donor grafts younger than 40 years, and the second arm with donors 40 years and older,” Nariani said. “Investigations will incorporate analysis and comparison of a number of factors, including visual outcomes, graft clarity and longevity, graft attachment rates and quantity of persistent corneal scarring between the arms.”

“So far, what we’re seeing from our preliminary data is promising,” Nariani said. “We hope to understand better through the Y-PDEK trial the effect of young donor corneas and pathophysiology behind the level of postoperative outcomes. Anecdotally, I have observed and been impressed by the clinical effect of using young donor grafts, while training in PDEK (surgical technique and graft preparation), and there has indeed been a remarkable level of difference in the corneal clarity that seems to be directly related to the age of the donor cornea.”

Nariani and colleagues hope that the study results will convince eye banks worldwide to prepare corneal tissue from young donors for PDEK.

“Now, if the trial indeed demonstrates statistically significant differences in postoperative outcomes when utilizing young donors with PDEK, it then becomes compelling for us to strive to make PDEK mainstream,” she said.

Nariani said that donors younger than 1 year are infrequently harvested because of a lack of surgeon demand for the tissue, which has an increased risk of PK rejection.

“I would think that’s a challenge globally because eye banks are not used to looking for infant donors for their surgeons,” she said.

“This [trial] will prove a number of things. It will prove whether or not infant donors need to be added to the pool. From a public health perspective, we’re dealing with corneal shortages in developing countries. The fact that we can add this additional pool is significant. Then you’re adding the fact that you have a lower level of rejection. You have less of a need for follow-up. You have less of a need for medications postoperatively when patients are not able to be compliant or are not able to afford their drops. It’s the best of all worlds,” Nariani said. “Now we just need to get the eye banks on board.”

PDEK and the results of the Y-PDEK trial may not only transform the way surgeons think about corneal transplantation surgery, but on an even larger scale, may lead to solutions to decrease corneal blindness globally and to overcome the corneal transplant tissue shortage in many parts of the world. Eye bank collaboration is critical to not only minimize the waste of tissue, but also to develop new strategies for safe preparation of PDEK grafts and to minimize endothelial cell loss in mass supply.

“We’re on the verge of finding a corneal transplantation technique that combines the best of all worlds,” Nariani said. “With the innovation of PDEK and the expanded utility of young donor corneas, we may not only improve patient outcomes but also contribute a solution to the global challenge of corneal transplant tissue shortage. We need to ensure our eye banks are then ready to meet this potential demand for PDEK tissue.” – by Matt Hasson

• References:
• Agarwal A, et al. Br J Ophthalmol. 2014;doi:10.1136/bjophthalmol-2013-304639.
• Agarwal A, et al. Cornea. 2015;doi:10.1097/ICO.0000000000000486.
• Kumar DA, et al. J Cataract Refract Surg. 2015;doi:10.1016/j.jcrs.2015.05.015.
• Pereira C, et al. Br J Ophthalmol. 2011;doi:10.1136/bjo.2010.191494.

• For more information:
• Amar Agarwal, MS, FRCS, FRCOphth, can be reached at Dr. Agarwal’s Eye Hospital, 10 Cathedral Road, Chennai-600 086, India; email: dragarwal@vsnl.com.
• Terry Kim, MD, can be reached at Refractive Surgery Service, Duke University Eye Center, 2351 Erwin Road, Box 3802, Durham, NC 27710; email: terry.kim@duke.edu.
• Ashiyana Nariani, MD, MPH, can be reached at Duke University Eye Center, 2351 Erwin Road, Durham, NC 27710; email: ashiyana.nariani@duke.edu.
• Francis W. Price Jr., MD, can be reached at Price Vision Group, 9002 N. Meridian St., Suite 100, Indianapolis, IN 46260; email: fprice@pricevisiongroup.net.
• Mark A. Terry, MD, can be reached at Devers Eye Institute, 1040 NW 22nd Ave., Suite 200, Portland, OR 97210; email: mterry@deverseye.org.

Disclosures: Agarwal, Kim, Nariani, Price and Terry report no relevant financial disclosures.

Which corneal transplantation technique do you prefer: DSEK, DMEK or PDEK?

DMEK yields best, fastest visual recovery

Endothelial keratoplasty is a major landmark in the corneal transplant arena and represents one of the greatest advances in corneal surgery in the present century. EK represents one of the selective corneal transplant procedures. Such a procedure significantly improves the postoperative quality of vision and augments the patient’s quality of life.

The major players in this EK arena are DSEK, DMEK and the newly introduced PDEK. Most surgeons are performing DSEK while the interest in DMEK continues to grow globally. PDEK, on the other hand, takes advantage of the newly described Dua’s layer. The presence of stroma in DSEK and PDEK makes donor tissue handling easier as compared with DMEK, which does not utilize any stroma. Older donor corneas are preferred for DMEK because it is easier to harvest the donor Descemet’s membrane. In contrast, PDEK can use very young donor corneas as well as older corneas.

Ultimately, the quality of vision and the speed of visual recovery after surgery are of paramount importance to the patient. In both these areas, DMEK best fits the bill as compared with DSEK and PDEK. This is because the deeper the donor-recipient interface, the better the visual quality. DMEK has the most posteriorly located interface among all EK procedures. Additionally, this tissue integration in DMEK restores the patient’s cornea to near-normal corneal anatomy.

Thomas “TJ” John, MD, is an OSN Cornea/External Disease Board Member. Disclosure: John reports no relevant financial disclosures.

DSEK more versatile than DMEK, PDEK

Endothelial transplantation has advanced significantly over the last several years, with multiple options now available, including DMEK, DSEK and PDEK. One reason that I still do some DSEK procedures is the versatility of the procedure. DSEK can be used with excellent success even in aphakic eyes, patients with an existing anterior chamber lens or patients with large iris defects. We are participating in trials of the artificial iris from HumanOptics, and in many of these patients, the combination procedure of a vitrectomy, scleral-fixated IOL, scleral-fixated artificial iris and DSEK is performed.

Because the DSEK tissue is thicker, it can be manipulated easily in this type of eye. Additionally, a retention suture can keep it against the cornea until the endothelial pump function allows permanent adherence. If PDEK or DMEK were being performed in this type of an eye, it would be very possible for the thin tissue to go around the iris prosthesis into the vitreous cavity and thus require more invasive surgery. Because of the versatility, I think that corneal surgeons should be well versed in the DSEK procedure, even if they also commonly perform DMEK or PDEK.

David R. Hardten, MD, is OSN Cornea/External Disease Section Editor. Disclosure: Hardten reports no relevant financial disclosures.