Discover the Advantages of Endothelial Keratoplasty

Transplanting just Descemet's membrane and the endothelium spares the anterior corneal structures. Here's how to get started.

By Sadeer B. Hannush, MD, Anthony Aldave, MD, and Henry D. Perry, MD

In recent years, a growing number of corneal surgeons have turned away from penetrating keratoplasty (PK) as their preferred means of corneal transplantation for the indication of corneal endothelial dysfunction in favor of a newer technique said to hold significant comparative advantages, Descemet's-stripping automated endothelial keratoplasty (DSAEK).

For more than a half century, full-thickness corneal transplantation, or penetrating keratoplasty (PK), was the un questioned transplant remedy of choice for those with corneal opacification secondary to multiple etiologies associated with endothelial dysfunction, most notably from Fuchs' endothelial dystrophy and pseudophakic or aphakic bullous keratopathy (Figure 1).

Figure 1. Clear graft after penetrating keratoplasty with eight interrupted sutures and a 24-bite running suture. ALL IMAGES COURTESY OF SADEER B. HANNUSH, MD

But penetrating keratoplasty has its drawbacks. Chief among them is compromised visual outcomes secondary to the ametropia that results from corneal toricity, regular and irregular, a product of many factors, including the use of 16 or more sutures to anchor the donor button in traditional full-thickness corneal transplantation. The sutures are also prone to loosen, invite infection or initiate neovascularization. Allograft rejection is one of the more troubling potential consequences. Also, compromised globe integrity secondary to the full-thickness incision at the graft-host junction carries a greater vulnerability to trauma. Visual rehabilitation can take from months to as long as two years or more following a full-thickness corneal transplantation procedure.

For the indication of corneal clouding secondary to endothelial dysfunction, pioneering surgeons thought they could avoid these adverse events and unsatisfactory outcomes by replacing just the affected back 2% of the cornea — Descemet's membrane and the endothelium — rather than the entire structure en masse.¹ The reasoning here is that the single-cell-thickness endothelial layer contains tens of thousands of what are essentially tiny “sump pumps,” whose job it is to suck fluid from the corneal stroma and pump it back to the anterior chamber to maintain corneal clarity. Failure of these endothelial sump pumps or compromise of the optically clear contour of the posterior corneal surface due to Fuchs' dystrophy or endothelial cell attrition leads to corneal swelling and visual loss.

It therefore stands to reason that a procedure that targets only Descemet's membrane and the endothelium addresses the chief source of visual loss in these conditions while sparing the rest of the cornea and preserving its normal contour. The following discussion offers a step-by-step guide to this exciting new procedure.

Advantages, Disadvantages

Even the most elegant and innovative new corneal surgical procedures have their pros and cons, and DSAEK is no exception.

• Advantages. Perhaps chief among DSAEK's advantages is the rapid visual rehabilitation — on the order of weeks vs. far longer following PK — facilitated by leaving the front-surface corneal contour relatively unscathed. (Some surgeons favor the use of one to three sutures, while others go sutureless.) You therefore end up with minimal anterior topographic change. The change in predicted post-operative refractive error likewise is minimal, usually a small hyperopic shift (1.0 to 1.5 diopters).

The complete or relative lack of sutures also eliminates the risk of suture-related complications, including vascularization that can lead to allograft rejection. A relatively intact globe resists trauma far more resiliently than in eyes following PK (Figure 2).

Figure 2. Upper left: Fuchs' dystrophy with corneal edema and micro-bullae.
Upper right: same patient one day post DSAEK with no sutures and a 30% residual air bubble.
Lower: Same patient one month post DSAEK with clear cornea.

• Disadvantages. Naturally, any new surgical procedure such as DSAEK entails a learning curve. It's also natural to expect some resistance to change, particularly among those who have performed PK for many years and are comfortable doing so.

One of the challenging aspects of DSAEK is the preparation of the donor tissue. While some surgeons prefer to prepare the tissue themselves — coauthor Dr. Hannush is one — most donor tissue is now prepared by eye banks. In DSAEK, the donor tissue includes Descemet's membrane and the endothelium together with 80-150 microns of carrier donor stromal issue that facilitates manipulation of the graft.

The risk of button dislocation, since the graft itself is not held by any sutures, is another potential complication. A successful procedure requires that the donor button adhere to the cornea's back surface. If not, the cornea will remain swollen and the surgeon must then institute measures to make the graft. Numerous events can cause donor button dislocation (Figures 3 & 4). See Table 1 for common causes of button dislocation.

A final potential disadvantage has to do with the lamellar interface. Having two sections of adherent tissue (the donor and host) — as opposed to the one piece of a fullthicknesstransplant — invites some degree of visual compromise, on the order of one to two lines of Snellen acuity.2 But this typically prevails only for the short term, a matter of weeks to months. Final visual acuity of 20/20 may be attained one to two years postoperatively. Of critical importance is that most patients prefer a 20/25 result with minimal ametropia/refractive error to 20/20 with strong spectacle or contact lens correction.

Figure 3. One day post DSAEK with graft dislocation.

Figure 4. One day post DSAEK with graft separation in the setting of hypotony post filtering surgery for glaucoma.

Indications, Contraindications

Indications for DSAEK include conditions causing endothelial dysfunction, such as Fuchs' corneal endothelial dystrophy and posterior polymorphous corneal dystrophy. DSAEK is also indicated for pseudophakic or aphakic bullous keratopathy.

Among the contraindications are conditions affecting the corneal stroma, specifically stromal dystrophies. Examples of these include granular dystrophy and lattice dystrophy. Other contraindications would be corneal scars from infectious keratitis or trauma and corneal ectasias such as keratoconus. Corneal stromal opacification in the setting of a healthy Descemet's membrane and endothelium may be treated effectively using deep anterior lamellar keratoplasty, or DALK.

Patient Selection

For those new to the procedure, we recommend that you reserve your early DSAEK cases for patients with pseudophakic eyes with posterior chamber intraocular lens implants and preferably intact capsules.

At least early on, avoid performing DSAEK in patients with anterior chamber implants and those with direct communication between the anterior chamber and vitreous cavity (termed unicameral eyes). Also withhold this procedure in phakic patients, unless you plan to first remove the lens and make them pseudophakic. Finally, avoid at the outset performing DSAEK in patients who have undergone glaucoma filtering surgery and have large blebs or tube shunts with very low intraocular pressures.³

Preoperative/Early-Operative Patient Preparation

Most surgeons now performing DSAEK favor regional anesthesia, although topical anesthesia is certainly possible for this procedure. Consider removing the epithelium, especially if edematous, to better visualize the anterior chamber. Many also advocate using mannitol to shrink the vitreous and thereby lower the intraocular pressure and decrease posterior pressure. This makes it easier to operate on the eye.

Finally, it's helpful to dilate the eye a little using 2.5% phenylephrine. This allows you to obtain a red reflex for a better view when stripping Descemet's membrane.

Intraoperative Technique

The first decision in performing Descemet-stripping automated endothelial keratoplasty concerns how to construct the incision. Most surgeons agree that you can make either a scleral tunnel or a posterior limbal incision. Either way, it's critical that the incision's internal opening be peripheral to the proposed position of the DSAEK button. Also, any paracenteses should be made so that the internal lumen is peripheral to the graft button. This is an important pearl, because it makes it easier to inject balanced saline solution (BSS) or air without the risk of dislocating the button; it also makes it possible to burp air out of a paracentesis postoperatively, if necessary.

Most surgeons recommend that you make an incision of 4.0 mm to 5.5 mm. This minimizes the crushing effect on the donor tissue and resultant endothelial cell damage. Understand that an incision that large located at the posterior limbus instead of in the sclera will invariably lead to some astigmatism. Therefore, the more posterior your incision, the less likely it is to create astigmatism. Many surgeons favor the scleral tunnel incision for this reason.

The next step is important. You can maintain the chamber using either a cohesive viscoelastic (such as sodium hyaluronate) or a chamber maintainer connected to irrigation. The key pearl here is that, if you use a viscoelastic, avoid the dispersive type. A dispersive viscoelastic is difficult to remove completely at the end of the procedure. It may seep into the interface between the donor button and the host posterior stroma, which can lead to button dislocation.

The next step is to score Descemet's membrane with a reverse Sinskey hook. Here you want to be very gentle, making sure not to apply too much pressure to the posterior stroma and risk damaging it. You then strip Descemet's membrane and remove it from the eye. Some surgeons use a 90° stripper for this purpose, while others simply peel it off; it comes off pretty easily. (One coauthor, Dr. Hannush, prefers to peel it.)

At this point, some physicians favor scraping the peripheral aspect of the host posterior stroma to enhance the graft's adherence. (Dr. Hannush eschews this step — citing the opinions of biomedical engineers who question its utility — but prominent corneal surgeon Mark Terry, MD, one of the pioneers in DSAEK, strongly advocates it.) Proponents contend that scraping the peripheral posterior stroma creates a sort of Velcro effect that promotes allograft adherence.

After removing Descemet's, inspect it before sending it to the pathology laboratory to ensure that you have indeed stripped it off. If you used a cohesive viscoelastic, irrigate it out. Now you're ready to insert the donor button.

Methods of Insertion

There are numerous methods and devices for inserting the donor button:

1. Forceps (Charlie, Golosey, Ogawa).
2. Shovel (Rosenwasser).
3. Glide (Busin).
4. Disposable inserter (Neusidl Corneal Inserter, Tan Endoglide, EndoSerter).
5. Suture pull-through or needle push-through techniques.

Numerous online videos can help familiarize you with these devices and techniques.

Most of us who insert the donor tissue with a disposable inserter like to use a chamber maintainer to keep the anterior chamber fully formed. Inserters such as the Neusidl Corneal Inserter and the EndoSerter have the irrigation connected to the device itself, which maintains the anterior chamber. When using other devices or instruments, such as the Tan Endoglide or the Busin glide, you need a separate chamber maintainer.

A potentially useful step at this point is to mark the donor tissue with an asymmetric letter such as an S or a Z. Why? Obviously, you want to insert the donor tissue with the endothelial side down. If it flips over with the endothelial side up, clearly it's not going to work. Marking the tissue gives you a way to ensure that you've positioned it properly.

But be careful here. The ink in a surgical marking pen contains alcohol, which can seep through the stroma and damage the endothelium. Here's a useful pearl. The trick is to paint the ink onto your surgical glove and let the alcohol evaporate. You then take an imprint from the glove — using a metal S or Z somewhat like a cattle branding iron — and place the mark onto the stroma, without transferring any alcohol there.

Completing the Procedure

Close your incision before injecting balanced saline solution or air into the eye. Most of us advocate placing the air in the chamber for a full (100%) chamber fill, raising the intraocular pressure above 30 mm Hg but preferably not over 50. You don't want the eye rock-hard. Go for 30-40 mm Hg or so, and then leave it for 8-10 minutes.

Most surgeons then remove either all the air or just 50%. Fact is, after 10 minutes of tamponade, it makes little difference whether you leave any air behind. Everybody agrees, however, that you do not want a soft eye afterward. Hypotony is among the potential causes of button dislocation. After removing the air — whether all of it or just half — you now re-pressurize the eye with BSS and keep the eye at a normal or slightly-above-normal pressure, say 20-30 mm Hg.

Here's another pearl. When injecting the air for tamponade, make sure the air is entering posterior to the DSAEK button. If the air goes between the graft and host, clearly it's not going to work. Rather, it has to come in behind the graft to tamponade the graft against the host stroma. This may sound obvious, but keeping it in mind will help you avoid some of the difficulties surgeons encountered when first performing this procedure years ago.

You also want to center the graft on the posterior aspect of the host cornea. Most of us prefer not to use a reverse Sinskey hook to manipulate the graft, because when you manipulate it from the endothelial side, you risk damaging the endothelium. Instead, there are ways to position the graft anteriorly. This you can do either through massage, or else you can make fenestrations in the host stroma and then go in with a needle anteriorly and position the graft.

Creating ventilation incisions — four through the front of the host full-thickness to the back — offers numerous advantages. First, you can manipulate the graft through an anterior approach without touching the endothelium and potentially damaging endothelial cells. Second, you can burp any fluid that's in the interface and massage it out through the vents. Third, if the button dislocates and you have to re-bubble the anterior chamber in the office over the next several days, you can once again manipulate the graft anteriorly and release interface fluid through these vents.

But there are disadvantages as well. For instance, you can invite epithelial ingrowth through the vents, which can also serve as conduits for microorganisms to access the deep lamellar interface perhaps days, weeks or months following surgery and cause a deep stromal infection. Some surgeons, including Dr. Hannush, no longer use ventilation incisions for this reason.

It's recommended that you dilate the pupil before taking the patient to the recovery room. There, the patient lies on his or her back for an hour or so. This last step is not imperative, but it's useful. This way, you can turn to another case, and later come back and examine the initial patient at the slit lamp biomicroscope.

Immediate Postoperative Steps

Following the procedure, confirm at the slit lamp biomicroscope that the button is properly positioned. If in fact you had left a 50% air bubble — and remember, it's not imperative that you do so — you now want to make sure that the inferior meniscus of the bubble is superior to the inferior edge of the pupil. If not, you'll get pupillary block with elevated intraocular pressure. So if the bubble is too big, you can burp some of it at the slit lamp before sending the patient home.

If the eye gets soft, you need to replace the air with BSS — because your biggest enemy now is hypotony, the major cause of button dislocation. Finally, apply an antibiotic ointment and patch the eye, and send the patient home.

Postoperative Day One

At day one, expect a 30% air bubble (unless you had removed it all at the end of the procedure). Anticipate a visual acuity between 20/200 and counting fingers at several feet (assuming a competent macula and optic nerve, as well as a visual axis unobstructed by the air bubble). If the patient is counting fingers only at a few inches, you may have a problem, usually marked corneal edema secondary to button dislocation; they should be able to see a little farther out.

Inspect the button to make sure: (a) it's centered, (b) it's apposed to the host cornea and (c) there's no fluid in the inter face. Do not manipulate the button at this stage, unless it's completely dislocated and floating in the anterior chamber.

If that's the case, you can reposition and re-bubble/tamponade it now. Some surgeons — including coauthor Dr. Hannush — prefer to wait a few days before doing so. The reason here is that one of the theories behind button dislocation is that the donor endothelial “sump pumps” have not yet recovered and are not pumping. If so, maybe it's useful to let the button swim in aqueous humor for a few days before attempting to re-bubble. This you can do at the slit lamp biomicroscope in your office or in a minor operating room.

We recommend that you have on hand a re-bubbling instrument kit (Figure 5). This includes: a 3-cc syringe, a 30-gauge anterior chamber cannula, a lid speculum, a sharp blade, and two reverse Sinskey hooks. Also have on hand a topical anesthetic, cycloplegic and antibiotic.

Start the patient on a topical steroid and antibiotic, and consider an NSAID if you performed a combined procedure with cataract surgery. See the patient one week later.

Figure 5. A rebubbling kit is important to have on hand.

A New Era in Corneal Transplantation

Endothelial keratoplasty no doubt will continue to evolve, improve and likely yield increasingly better outcomes. Consider its advantages over traditional penetrating eratoplasty. Even though DSAEK has its disadvantages, we believe these are far outweighed by its benefits.

Also be on the alert for information and training opportunities on a new corneal transplantation procedure, Descemet's membrane endothelial keratoplasty (DMEK), in which only Descemet's and endothelium — without carrier donor corneal stroma — is transplanted. This procedure remains in its infancy and is not yet ready for prime time.

Both procedures figure to take their place among the state-of-the-art surgical repertoire of corneal surgeons. Their relative advantages over PK will redound ultimately to the patient in the form of quicker visual rehabilitation, fewer potential complications and better outcomes. OM

References

1. Hannush SB. Preoperative considerations and decision-making in keratoplasty. In: Krachmer, Mannis & Holland's: Cornea, 3rd Ed. Elsevier, London, U.K., 2010:1327-1334.
2. Shulman J, Kropinak M, Ritterband DC, Perry HD, Seedor JA, McCormick SA, Milman T. Failed Descemet-stripping automated endothelial keratoplasty grafts: a clinicopathologic analysis. Am J Ophthalmol 2009;148(5):752-759.
3. Wiaux C, Baghdasaryan E, Lee OL, Bourges JL, Deng SX, Yu F, Aladave AL. Outcomes after Descemet-stripping endothelial keratoplasty in glaucoma patients with previous trabeculectomy and tube shunt implantation. Cornea 2011;30(12):1304-1311. PubMed PMID: 21963858.
4. Hannush SB, Chew HF, Eagle RC Jr. Deep stromal infectious keratitis after DSAEK with fenestrations. Cornea 2011;30(2):229-232.

Sadeer B. Hannush, MD (pictured), is Attending Surgeon on the Cornea Service at the Wills Eye Institute in Philadelphia. Anthony Aldave, MD, is Director of the Cornea Service at the Jules Stein Eye Institute in Los Angeles. Henry D. Perry, MD, is Director of the Cornea Service at the Nassua University Medical Center in East Meadow, NY.