Endothelial Keratoplasty

DLEK & DSEK Sutureless Corneal Transplant Surgery Techniques

  - Videos of this procedure are found in the "Video" section of this site

Introduction

Endothelial dysfunction from disease or trauma is one of the leading indications for corneal transplantation. Over the past 100 years, the only solution for endothelial replacement was through full thickness corneal transplantation.1 While penetrating keratoplasty (PK) has been shown to yield healthy donor tissue with good endothelial function, this procedure has been plagued by the inherent problems of unpredictable surface topography, retained surface sutures, and poor wound strength.2-6

In 1993 Ko and Feldman presented an animal study at ARVO which described a new technique for endothelial replacement through a scleral limbal incision.7 In 1998, Melles et al described this technique in the first human patients and called it posterior lamellar keratoplasty.8 Terry and Ousley began laboratory work in 1999 on this technique, and after technical modifications and re-design of instrumentation, performed the first United States cases in 2000 and called the surgery Deep Lamellar Endothelial Keratoplasty (DLEK).9-21 All of this work represents a radical departure from the PK technique in that the DLEK surgery accomplished the goal of endothelial replacement without ever touching the surface of the recipient cornea. By eliminating surface corneal sutures and incisions, the advantages of normal corneal topography and faster wound healing were obtained, leading to faster visual rehabilitation and a more stable globe for the patient.10-14 We have investigated this technique in the largest prospective study of endothelial keratoplasty surgery in the world and have found it to be valid for endothelial replacement surgery.21

          While undoubtedly there will be further refinements of the technique and instrumentation in DLEK surgery, it is the purpose of this chapter to describe in detail the recent modification in preparation of the recipient bed by stripping of the recipient descemets membrane. This modification utilizing the stripping of Descemet’s membrane has been popularized as “Descemets Stripping Endothelial Keratoplasty”, or “DSEK”, and has the advantage of being easier for the surgeon to perform and of providing a smoother interface on the recipient side for the visual axis. Preparation of the donor tissue in endothelial keratoplasty has also been made easier with the utilization of an automated microkeratome, and the addition of this component to the surgical procedure has been popularized as “Descemets Stripping Automated Endothelial Keratoplasty”, or DSAEK. Currently, Moria has the only practical microkeratome for the preparation of donor material and other lamellar keratoplasty procedures. These advantages of DSEK/DSAEK over DLEK have been mitigated in the past, however, by the reported increase in donor dislocation the day after surgery, requiring a further surgical intervention for re-attachment. In addition, the abnormally increased thickness of the resulted DSEK/DSAEK cornea has unknown visual consequences for the patient and is still being investigated. We describe in this report a method for reducing the incidence of donor dislocation in DSEK/DSAEK surgery, by specific roughening of the peripheral recipient bed to promote donor adhesion. Other surgical maneuvers to prevent dislocation are also described.   

The reader is cautioned, however, that DLEK or DSEK/DSAEK surgery requires a separate skill set from standard full thickness PK surgery and therefore the experienced PK surgeon will find the maneuvers of DLEK or DSEK/DSAEK surgery initially unfamiliar. Therefore it is highly recommended that the DLEK/DSEK procedure be extensively practiced in the laboratory before embarking in clinical treatment of your patients.

 

 

 

Anesthesia

          DSAEK surgery is usually done under retrobulbar block anesthesia. General anesthesia (either endotracheal or laryngeal mask airway technique) is somewhat easier for DSAEK because it minimizes posterior pressure on the globe and this is important during the recipient resection and donor implantation phases of the surgery. In addition, general anesthesia minimizes the risk of patient movement during the more delicate portions of DSAEK surgery and may aid in the comfort level of the novice DSAEK surgeon.  Nonetheless, DSAEK surgery can be safely accomplished with good retrobulbar anesthesia combined with seventh nerve block (orbicularis block) local anesthesia as well. It is even possible to perform DSAEK surgery on selected high risk patients under topical anesthesia, but this is not recommended for the initial cases of the novice DSAEK surgeon.

 

Pre-operative Medications

In the usual case of patients with pseudophakia the pupil is constricted in order to stabilize the iris-lens diaphragm during the surgery. This is also done if the patient has a clear crystalline lens and concurrent cataract surgery is not planned. We have stopped using Pilocarpine for this, and instead use Miochol (intra-operatively) in order to only have a short acting constriction of the pupil. The Miochol is placed just after the Healon has been removed from the anterior chamber, after the recipient stripped tissue has been removed.

Pre-operatively, one set of aproclonidine 0.5% drops is also given just prior to surgery to reduce pressure and minimize conjunctival injection. Preoperative antibiotics may be given according to the surgeon’s preference. The eye is prepped in the usual sterile ophthalmic fashion with the use of povidone-iodine solution.

          In the case of patients with cataract and endothelial failure, cataract surgery is performed just prior to the DSAEK endothelial transplant and the pupil is dilated preoperatively with the surgeon’s standard dilating drops for cataract surgery. While pilocarpine is avoided, the rest of the pre-op medication regimen described above is utilized.

 

Descemet’s Stripping Automated Endothelial Keratoplasty (DSAEK): Recipient Peripheral Scraping Technique

          The small incision DSAEK procedure is usually performed from the temporal side in order to provide the greatest manual access and visualization for the surgeon. Prior to forming the DSAEK scleral access incision, two clear corneal limbal stab incisions (about 1 mm diameter) are placed about 5 clock hours apart, to be used as access points to the anterior chamber later in the operation. We find that an inexpensive 1 mm diamond knife provides a better paracentesis wound than that made by a metal blade. The paracentesis wounds should be made relatively perpendicular to the corneal surface in order to allow quick access to the anterior chamber later in the surgery. Through one of the stab incisions, the cohesive viscoelastic Healon (Pfizer, New York, NY) is placed into the anterior chamber to replace the aqueous fully and to maintain normal pressure. We adamantly oppose the use of Viscoat (Alcon, Fort Worth, TX) or other dispersive viscoelastic materials during any portion of DLEK surgery as the dispersive materials can cause interface coating with subsequent non-adherence and dislocation of the donor tissue.

          Prior to the stripping of Descemets membrane, a template mark is placed on the corneal epithelial surface. A circular marker with a diameter of 8.0, 8.5, or 9.0 mm (depending upon recipient corneal diameter and surgeon preference) is used to make a circular impression on the central epithelial surface. If the position and centration of the mark is acceptable to the surgeon, then it is accentuated with gentian violet ink marks. This circle on the corneal surface will later be used as a template for stripping of the recipient descemets membrane. 

          A temporal limbal peritomy of the conjunctiva is performed with scissors allowing exposure of about 6mm arc length (about 3 clock hours) of limbal tissue. After cautery of the scleral bed, a trifaceted, guarded diamond knife is then set to a depth of 350 microns and a 5.0-mm length incision is made approximately 1 mm posterior to the corneal limbus and concentric with it. We have found that a deeper initial incision gives less of a beveled wound closure and also a greater chance of early perforation into the anterior chamber during DSAEK surgery. In lieu of a diamond knife, a sharp crescent blade or other steel scalpel can be used for the initial incision. A sharp crescent blade is then utilized to create a deep scleral-corneal lamellar pocket down to about 75% to 85% corneal depth along the entire length of the wound, extending centrally about 1 mm into clear cornea from the limbus.

A simple, blunt tip, reverse Sinskey hook (Bausch and Lomb, St. Louis, MO) is used for the descemets stripping portion of the procedure. The hook is entered into the Healon-filled anterior chamber through one of the limbal paracentesis sites. The tip is passed across the chamber and then lifted anteriorly until it contacts the recipient endothelium at the distal location of the epithelial circular template mark. The diseased endothelium and thickened descemets membrane of the recipient is easily punctured by the blunt tip of the reverse Sinskey hook and the hook then is used to score the recipient descemets membrane, following the path of the overlying circular template mark. This maneuver easily cuts a nearly perfect circle in the descemets membrane, outlining the edge limits of the recipient bed for endothelial transplantation. A blunt tip hook is far preferable to a sharp point or needle point in that it selectively punctures the descemets membrane, but does not penetrate the stronger overlying stromal fibers. This allows true descemets stripping (and not stromal dissection), yielding the easiest stripping and the smoothest possible surface for the visual axis

Fuchs’ dystrophy causes the recipient descemets membrane to be significantly thicker than what is normally encountered in other conditions of endothelial failure. This fact, in addition to posterior separation of posterior stromal lamellae by edema, works to the advantage of the surgeon in allowing very easy and smooth stripping of the diseased descemets to occur. In addition, the stripped descemets membrane is easily visualized, and does not require special staining to help with resection. Once the diameter of recipient bed has been scored, the same tip is used to gently peel descemets membrane away from the overlying tissue, pulling the edges into the anterior chamber for 360 degrees. The tissue then can be completely stripped off from the central posterior cornea using either the same reverse Sinskey hook tip or by using simple forceps or an irrigation/aspiration tip during the step of viscoelastic removal.

          Once descemets membrane has been completely stripped with the reverse hook, the diseased tissue can be removed from the chamber and sent to pathology. We utilize a standard cataract surgery diamond blade with a 2.8 mm width to enter the anterior chamber, but any blade is acceptable. Entry into the anterior chamber is done by passing the blade through the corneal-scleral tunnel, directing the blade posteriorly, and then entering the chamber. We prefer to enter the chamber at a site at least 1 mm peripheral to the most temporal portion of the circular template mark. Capsular or other gentle forceps are used to grasp the recipient descemets membrane, remove it from the eye and send it to pathology.

          It is at this point of the surgery that we feel it is critical to scrape the periphery of the recipient bed in order to aid in later donor adherence. When viewed by scanning electron microscopy we have found the recipient bed after stripping of descemets membrane to be exquisitely, glassy smooth.25 Compared to DLEK, there is complete absence of cut stromal fibrils. We believe that these cut fibrils of the recipient bed significantly contribute to donor adhesion, and their absence after stripping may partially explain why donor tissue after DSAEK has more of a propensity to dislocate than after DLEK. We therefore advocate a hybrid of the two techniques whereby the advantages of both are retained. In this hybrid modification, the periphery of the recipient bed is scraped to roughen the surface and expose stromal fibrils, while the central recipient bed is left untouched to yield the smoothest interface possible for the visual axis.

          Peripheral scraping of the recipient bed is accomplished with the use of a specialized instrument, the Terry Scraper. (Bausch and Lomb, St. Louis, MO). This instrument is similar to a reverse Sinskey hook or Descemets stripper, but instead has a broad (1mm) roughened tip like an ice scraper. The tip is entered through the scleral-tunnel incision, with Healon still filling the anterior chamber. The tip of the Terry Scraper is then used to scrape the peripheral 1.5 mm of the recipient bed, and the creation of stromal fibrils is visually verified. The periphery of the recipient bed is scraped for 360 degrees, taking care to leave the central bed of 5 to 6 mm diameter untouched and smooth.    

          After scraping of the recipient bed, the temporal scleral wound is temporarily closed with 1 interrupted 10-0 nylon suture. An irrigation/aspiration tip is then entered into the anterior chamber and extensive effort is expended to remove all of the viscoelastic from the eye. Absolutely no Healon should remain in the anterior chamber prior to insertion of the donor disc or the donor tissue will not stick in place. Therefore, care is taken to irrigate and aspirate the anterior chamber, pupillary area, angle, and even the peripheral recipient bed as necessary. Once the surgeon is confident that all Healon has been removed, then Miochol is injected to constrict the pupil for the safety of the donor during insertion and unfolding. The intra-ocular pressure of the eye is left slightly soft and attention is turned to preparation of the donor.

 

         

Automated Donor Tissue Preparation (DSAEK)

(This step of donor preparation can be done just prior to the surgery on the patient’s eye, depending on surgeon preference) (“Pre-cut” tissue is also available now through a few EBAA certified Eye Banks.)

          The operating microscope is brought over to the separate donor table for preparation of the donor tissue. Because whole globes are rarely available here in the United States, an artificial anterior chamber is necessary for preparation of the donor posterior disc. The only automated microkeratome system available for DSAEK surgery is from Moria. (Moria LLP, Pennsylvania) The Optisol-GS preservation fluid (Bausch and Lomb, Rochester, NY) from the donor tissue container is pulled up in a syringe (which has a three way stop cock) and then the syringe is then used to fill the I/A port of the artificial anterior chamber. The syringe is also attached to the port to be used to vary the pressure inside the chamber for the duration of the resection. The standard donor cornea-scleral cap tissue is first coated with a thin layer of Healon on the endothelium. It is then placed endothelial side down onto the post/piston of the artificial anterior chamber (taking care not to include an air bubble in the chamber) and the tissue is oriented with the largest diameter of the cornea in the horizontal meridian noted. The donor tissue cap in proper noted orientation, the piston/post of the unit is then raised until the tissue is firmly locked into place. The Optisol filled syringe is then used to raise the pressure in the artificial chamber to over 65 mm Hg and the stop-cock closed to stabilize the high pressure. The epithelial cells are then wiped from the surface of the cornea with a Miracel sponge. The horizontal meridian is marked at the peripheral cornea with a marking pen so that the horizontal meridian of the donor tissue can be identified and for proper anterior cap orientation later in the procedure.

For anterior automated lamellar keratoplasty, the guide ring for the microkeratome can be adjusted in height to yield the desired diameter for the tissue resection. For DSAEK surgery, the greatest diameter possible of at least 10 mm is required, and therefore the guide ring should be placed at the lowest possible position without the use of a spacer.

The microkeratome head is usually set at 300 microns and the diameter of resection is at least 9.5 mm or more. For corneas which are thicker than 580 microns, a 350 micron microkeratome head can be utilized to obtain an even thinner posterior donor tissue. Caution should be utilized, however, in using a 350 micron head in order to avoid the occurrence of posterior perforation or “button holes” of the tissue. With the pressure in the artificial chamber elevated to at least 65 mm Hg and verified with a gravity tonometer or by finger touch, the microkeratome head is mounted on the guide ring, positioned for resection, and then passed over the donor cornea at a rate of about 3 to 4 seconds for the pass.

 A free cap of anterior tissue is resected and remains above the blade on the microkeratome head. After drying the residual stromal bed with Miracel sponges, checking the stromal bed for smoothness of cut and diameter of cut, several marks are placed at the peripheral edge of the resected bed to help with the centering of the subsequent trephination of the donor tissue. The anterior cap is placed back in position, using the previously placed peripheral reference surface marks. One additional mark is then placed at the exact central point of the anterior cap of tissue to further facilitate centration of the posterior trephination. A moment is given for the anterior cap to adhere to the bed.

          The donor tissue now must be carefully dismounted from the artificial anterior chamber without damaging the endothelial cells from chamber collapse. We have found that the easiest way to avoid chamber collapse is to leave the tissue attached to the post/piston. To achieve this, the stop cock on the syringe is turned to the position to allow Optisol flow from the syringe to the chamber. A tying forceps is then used to sweep along the inside ring of the metal cap which locks the tissue onto the post, pushing slightly posterior on the scleral rim, and breaking the seal that binds the donor scleral tissue to the metal cap. Very gently the post/piston is lowered, and Optisol is gently infused, as needed, to keep the donor chamber from collapsing. The locking cap is then removed, with the donor tissue left on the post with a formed chamber. The scleral edges of the donor are gently lifted in each quadrant to release the seal of the tissue to the post, and the tissue is removed from the post very slowly to allow the Healon to simply flow off in one cohesive “glob” from the endothelial layer onto the post. Once again care is taken not to collapse the chamber and damage the endothelium. We believe that minimal irrigation of the endothelium with BSS or Optisol during tissue removal from the Moria unit aids in the health of the endothelium and in the subsequent donor adherence to the recipient bed.

          Once the donor corneal-scleral tissue has been removed from the post, we have been gently irrigating the sclera above and below the endothelial surface of the donor tissue with Optisol solution (taken from the same transport container for the donor tissue) in order to remove excess residual Healon from the donor. We then use a Miracel sponge to wick the excess fluid from the endothelium, placing the sponge tip at the scleral edge, away from the endothelium.

          The donor tissue is then placed endothelial side up onto a standard punch trephine block. The previously placed ink mark on the central point of the anterior resected cap is used as a guide to position the tissue for trephination, in order to make sure that the posterior punch trephination is centered on the bed of the previous keratome pass. We utilize a Barron donor punch (Katena, Denville, NJ). The same size diameter punch is used as the diameter of the descemet’s membrane disc that was removed from the recipient. The tissue is punched out with the trephine.

          Because the 5 mm wound of small incision DLEK surgery is smaller than the 8.0 mm (or larger) diameter of the donor disc, the donor tissue must be folded prior to insertion. To accomplish this, a very thin strip of Healon is placed onto the endothelial surface along the previously identified and marked horizontal meridian of the donor button. Stabilizing the anterior edge of the donor button with a 0.12 mm forceps, the posterior stromal tissue edge is gently grasped with non-toothed delicate forceps. The posterior tissue is then gently folded with the endothelium on the inside protected by the layer of Healon, and it is folded into an asymmetric “taco” shape, in a 60%/40% ratio, the most anterior side of the taco being 60% and the posterior side 40%. We were the first to initiate the idea of an “overfolded, 60%/40% ratio” over 5 years ago (in 2001) in order to avoid having the tissue unfold upside-down in the patients anterior chamber. The donor tissue is then brought over to the operative field still on the trephine block.

 

Donor Tissue Preparation Without a Microkeratome

          The operating microscope is brought over to the separate donor table for preparation of the donor tissue. Because whole globes are rarely available here in the United States, an artificial anterior chamber is necessary for preparation of the donor posterior disc. We utilized a Bausch and Lomb (St. Louis, MO) artificial anterior chamber that is all stainless steel and has dual irrigation/aspiration ports. The Optisol-GS preservation fluid (Bausch and Lomb, Rochester, NY) from the donor tissue container is pulled up in a syringe and is then used to fill the I/A ports of the artificial anterior chamber. The syringe is also attached to the port to be used to vary the pressure inside the chamber for the duration of the resection. The standard donor cornea-scleral cap tissue is first coated with a layer of Healon on the endothelium. It is then placed endothelial side down onto the post of the artificial anterior chamber and oriented with the largest diameter of the cornea in the horizontal meridian. This meridian is marked with a marking pen so that the horizontal meridian of the donor tissue can be identified later in the procedure. The donor tissue is capped into place and the chamber is filled with Optisol-GS and the pressure normalized. There are several ways to create a deep dissection plane in the donor tissue in the absence of a microkeratome. Each way has advantages and surgeon preference is the over-riding determinant of which method is utilized. The peripheral incision and central manual trephine approachs are described below.

In the peripheral incision technique, the preparation of the donor is similar to the preparation of the recipient deep lamellar plane. A diamond knife set to a depth of 350 microns is used to make a 3 or 4 clock hour-length incision in the peripheral donor limbal area, right next to the edge of the metal cap of the artificial anterior chamber. The crescent blade is then used to cut to the deeper stromal tissue and then once the desired plane has been reached, the Devers Dissectors are used to continue the dissection plane all the way to the limbus of the donor tissue for 360 degrees. We have found it helpful to place multiple ink marks onto the surface of the donor tissue. These marks accentuate the visualization of the surface of the donor tissue and therefore allow better depth perception (and estimation of the level of the dissection) when viewing the Devers Dissectors plane compared to the surface plane. It is important to make sure that the dissection plane is carried out all the way to the limbus in every quadrant in order to avoid problems that may result from an eccentric trephination of the donor lenticule later on.

In the manual trephine approach, an 8.5 mm diameter Barron suction recipient trephine (Katena Products, Denville, NJ) is placed onto the surface of the donor tissue and suction is applied. Trephination is carried out to about 60% depth with the trephine. It is noteworthy that after the blade touches the epithelial surface of the donor, it only takes about 4 or 5 quarter turns of the Barron trephine to reach this depth. This is much sooner than when the same trephine is used on the recipient in standard PK surgery. The trephine is then removed and the cut inspected for depth. Ideally, an 80% depth should be attained for the plane of the pocket of the donor tissue. Any deeper than this, and the donor tissue is so thin that it spontaneously rolls up like a rug causing confusion as to which side is the endothelial side and undoubtedly causing endothelial damage. If the dissection depth of the donor is less than 60% depth, then the stromal surface may not be as smooth and the tissue may be much thicker than the recipient bed is deep. However, whether disparity between donor and recipient disc thicknesses causes a later visual problem is unknown at this time.22

 

Transplantation of the Donor Tissue

          With the microscope in place, the temporary scleral suture of the superior wound is cut. The anterior chamber of the patient is then filled completely with BSS. The donor tissue is then brought into the field and grasped along the horizontal meridian of the stromal surface of the donor tissue with specialized insertion forceps (“Charlie Forceps”, Bausch and Lomb, St. Louis, MO). The Charlie forceps are non-toothed fine forceps that coapt only at the distal tips. The forceps have a block that allows a significant spacing along the blades length to prevent crushing of the donor tissue. This specially designed stop in the Charlie forceps enables the surgeon to transfer and hold the folded donor disc tissue without crushing it. The folded donor tissue is placed into the anterior chamber in one deft movement, by inserting the donor tissue with the anterior 60% stromal side facing the recipient bed and the posterior 40% stromal side facing the iris. Again, the endothelial layer remains protected on the inside by Healon. The tissue can be gently prodded with the forceps along the stromal sides if centration of the tissue within the recipient bed needs to be improved. Additionally, the surface of the cornea can be massaged, and this too often allows the donor taco tissue to be gently prodded into proper centration. The opening of the taco is placed to the surgeon’s left. The 60% stromal side gently adheres to the overlying recipient bed with the 40% stromal edge lying nearly perpendicular to the iris plane.

Three sutures of 10-0 nylon are then used to close the scleral wound to secure the chamber. A cannula is then placed through the right hand stab incision and the tip placed onto the iris surface. BSS is then gently injected into the anterior chamber to fill the chamber and deepen it. With deepening of the anterior chamber, the “taco” will usually begin to open. The ideal is to open the taco just enough so that the posterior 40% edge lies at an 80 degree angle, just short of perpendicular to the iris plane. If the tissue does not open at all, then the irrigating cannula can be moved to the left paracentesis site, and irrigation with BSS can be used to loosen the Healon from the endothelial surface and gently unfold the tissue to the 80 degree angle noted above. Because the donor tissue was folded into an asymmetric shape, the tissue invariably will spontaneously unfold in the correct orientation (i.e., endothelium down), as long as the chamber is deep enough and there is no impediment. Once the tissue has unfolded to the 80 degree angle, then an air bubble is gently injected into the anterior chamber from the LEFT paracentesis site, very gently into the interior of the 80 degree folded taco. This gentle injection will then fully open the taco and push it up into position onto the recipient bed. Once the tissue is unfolded fully, then air is forcibly and quickly injected to fill the chamber with air and stabilize the tissue, locking it into position. If at any time the surgeon has trouble opening or positioning the donor taco, then the first thing to do is to close the wound with three sutures to stabilize the anterior chamber and then repeat the unfolding and chamber deepening maneuvers mentioned above.

          The donor disc may not have perfect centration after insertion. If not, it can be positioned from either the endothelial side or the stromal side. A reverse Sinskey hook (Bausch and Lomb, St. Louis, MO) is used for endothelial-side positioning. The hook is placed through the stab incision, the peripheral endothelium is engaged, and the tissue moved over to whatever position is desired. This should only be done with a partial air bubble in the chamber. If the chamber is completely filled with air, then such a maneuver will cause striae in the donor tissue. Although this maneuver undoubtedly causes endothelial damage at that point of peripheral contact, we have not found that the central endothelial cell counts 6 months after surgery are any worse than after standard PK.10-11,14,21 Care is taken, however, to minimize this maneuver and also to avoid the central posterior striae that can occur and can compromise vision. An alternative technique for positioning can be done from the stromal interface side using a 30-gauge needle tip. A slight “barb” is placed on a standard short 30-gauge needle, and the tip is placed through the temporal wound directly into the interface. The barb is rotated posteriorly to engage a few stromal fibers of the donor disc, and this grasp is used to move the tissue over into the proper centration. During both the endothelial and stromal positioning maneuvers, the anterior chamber is completely filled with air.

          Once the tissue is in proper centration, it is critical to make sure that there is no residual fluid in the central or peripheral interface which can interfere or delay adhesion. One technique described by Price et al involves the “squeezing” of interface fluid from the central interface to the periphery and into the anterior chamber. In this technique, the anterior chamber must be completely filled with air, past the edges of the donor tissue. The epithelial surface of the cornea is kept wet with BSS, and then the smooth surface of the shaft of an irrigating cannula is used to deeply compress the central surface of the cornea. The cannula is then swept across the surface from central to peripheral, stroking the corneal surface, attempting to move central interface fluid to the periphery and then into the anterior chamber. Significant compression of the surface must be employed. This maneuver is done from center to periphery in every quadrant, until the surgeon is satisfied that there is no residual fluid in the interface. There is now in 2006 a specific instrument called the “Cindy Sweeper” for this maneuver. (Bausch and Lomb).

          Another maneuver for interface fluid removal involves the placement of stab incisions from the surface to the interface, which are placed in the mid-peripheral cornea in each quadrant. Price has advocated these stab incisions as a means of further releasing interface fluid and reducing donor dislocations. We do not advocate this technique at this time as it violates the primary philosophy of endothelial keratoplasty, which is to avoid surface corneal incisions or sutures of any kind. Furthermore, we do not find that this additional manipulation offers any further protection against dislocation than the recipient preparation steps we have outlined above.

          With the tissue in good position, the interface fluid removed, and the anterior chamber completely filled with air, the microscope light is turned off and the tissue is left undisturbed for a full ten minutes intra-operatively. We believe that this time allows the donor tissue to fully warm up to body temperature and endothelial pump function to begin. This also allows full cohesive interaction of the peripheral recipient and donor interface stromal fibrils. Admittedly, the decision for 10 minutes of donor respite time is wholly arbitrary, and less or more time may be recommended in the future as multiple surgeon experience dictates.

It is also at the beginning of this 10 minute waiting time that we place several drops of dilating solution (Cyclogyle 1% and Phenylephrine 2.5%) onto the corneal surface to allow for the pupil to dilate post-operatively. Because we leave about an 8 or 9 mm air bubble in place at the end of surgery, the dilated pupil will prevent any pupillary block. In nearly 500 cases of endothelial keratoplasty, we have not had a single case of pupillary block.    

     Once satisfied that the donor disc is in final position with no interface fluid, the surgeon then removes the air in the anterior chamber and replaces it with BSS. Care is taken to avoid pupillary block by the air bubble in the anterior chamber, but if it occurs, simple suctioning of the air from the pupillary area resolves the problem. Occasionally air can get trapped behind the iris, giving the impression of posterior pressure with the iris coming forward to the donor edges. Again, suctioning with a cannula from the pupillary area will resolve this issue. The BSS placed into the anterior chamber creates a normal IOP and the chamber deepens. An air bubble of approximately 8 to 9 mm is usually left in place to help further stabilize the donor disc position over the first 24 hours postoperatively.

The suture knots of the scleral incision are cut short, and buried on the scleral side. The wound is checked to be watertight. The conjunctival peritomy is closed with either sutures or cautery. We routinely place on the corneal surface a 24-hour collagen shield soaked in antibiotics (moxifloxacin) and steroids (dexamethazone) at the close of surgery in order to deliver medication until the patch is removed the next day; however, each surgeon’s usual routine for antibiotics (subconjunctival or otherwise) is certainly acceptable.

          An occlusive patch and shield are routinely placed, and the patient is brought to the recovery room. We usually instruct the nurses to have the patient lie in a supine position, flat, facing the ceiling, for the first hour after surgery and then as much as reasonably possible to allow the retained air bubble to further stabilize the graft position, but this is not critical. The patient is discharged from this outpatient procedure when fully recovered from anesthesia.

 

Combined Surgery with DSAEK

          The most common procedure done in combination with DSAEK surgery is cataract extraction with intra-ocular lens placement. This is usually accomplished with phacoemulsification, but other nucleus manipulative techniques can be employed. The cataract extraction phase of the combined procedure is usually done first and from the same surgical scleral-corneal tunnel incision site. If there is a problem with visualization of the procedure due to surface bullae, nodular scarring or epithelial edema, we simply scrape the surface of the cornea over a 7 mm diameter area or larger and remove any irregularities. Even with poor visualization, the nucleus can be easily removed by bringing the lens up into the anterior chamber and performing the emulsification there. Damage to the central endothelium is not an issue, because the endothelium will be removed during the DSEK procedure which follows. The viscoelastic material used must be the most cohesive type available, therefore we prefer Healon or Healon V (Pfizer, New York, NY) for use during any surgery that is combined with DSEK. By no means should a dispersive viscoelastic like Viscoat (Alcon, Fort Worth, TX) be used at any time of a combined procedure, and a concentrated effort must be made to ensure that absolutely no viscoelastic of any kind remains in the eye prior to insertion of the donor tissue lenticule. After the cataract extraction with IOL placement is completed, the surgeon then strips the descemets membrane as described above using either the paracentesis sites or the scleral tunnel incision for access .

          Other surgeries such as IOL lens exchange, vitrectomy, iridoplasty, or suturing of an IOL to the ciliary sulcus can also be accomplished at the same time as DSEK surgery. Because many of these surgeries require a larger incision than a 3 mm cataract wound, they can be done through the DSAEK wound, and the DSAEK wound fashioned for whatever length is necessary to accomplish the task. All surgical maneuvers must be completed however, BEFORE the donor tissue is inserted, to insure the secure adhesion and safety of the donor endothelium. Interestingly, many intra-ocular surgeries can safely be accomplished only 3 or 4 months after endothelial keratoplasty surgery without concern of dislodging the grafted corneal endothelial tissue.23

 

Postoperative course

          The patient is seen the next morning and the patch is removed. Most patients will remark that the eye was as comfortable as after standard cataract surgery and that they did not require narcotic pain relief at all. Once the patch is removed, the vision is usually about 20/400. The vision is unimportant on post-op day one, and the only reason for the visit is to insure that the donor disc is attached and in good position. In our prospective series of our first 100 DLEK patients, we experienced only 4 cases where the donor disc was dislocated on the first postoperative day.21, 24 All four cases were easily treated by taking the patient back to surgery, and usually under topical anesthesia, with a 15 minute operation, another air bubble is placed in the anterior chamber and the disc repositioned as before. All repositioned grafts resulted in clear corneas, but the endothelial cell counts at 6 months post-op are significantly lower than grafts that have not had to undergo re-positioning.24 In our first 4 cases of DSEK without the benefit of scraping and roughening of the peripheral recipient bed, we experienced a 50% rate of dislocation. Utilizing the modification of peripheral scraping, however, we have experienced only 4 cases (4%) of dislocation in our initial 100 DSEK cases using peripheral scraping.25

          If the graft is in good position on day one, it generally will heal in good position. Although we had no late dislocations in DLEK surgery, of the 4 cases of dislocation in DSEK surgery, 3 of those cases were attached at day one and dislocated on days 2, 3, and 4 after the surgery date. This has led us to recommend to patients that they use the residual air bubble for one more day of supine position after DSEK surgery and to absolutely not rub their eye for a full 2 weeks after DSEK surgery. The edges of the graft seal down with solid healing sometime within the first 3 months. The overlying cornea has a variable rate of clearing, but some patients are able to see as well as 20/25 only one week after DLEK/DSEK/DSAEK surgery with a crystal clear central cornea.

Only prospective data with control of the variables of patient age and pre-operative vision level will sort out the differences in DLEK and DSEK/DSAEK surgical outcome. As a guide for DSAEK, however, we have found that the usual visual progression postoperatively of patients with minimal or no macular disease, has been the following:

 

One day: 20/400

One week: 20/70

One month: 20/40

Three months: 20/30

Six months: 20/25

One year: 20/25 to 20/20

Two years: 20/25 to 20/20

 

There is, of course, high variability of vision in any series of elderly patients undergoing ocular surgery, but especially DLEK or DSEK/DSAEK. The interface may clinically appear exceptionally clear, but it remains an interface with at least the donor tissue with a stromal resection. It is this stromal interface of the donor that likely contributes about one line of visual loss to the macular potential.13-14 Extensive work continues to be done to improve the interface in DSAEK surgery. Investigators are working in the areas of femtosecond laser preparation of the donor tissue, but currently, the interface after femtosecond resections in the deep stroma are inferior to that created by a microkeratome. While all of these high and low technology approaches have individual appeal, the worth of any modification will have to be determined by comparing the safety level of the modified technique to the safety of the current DLEK and DSEK/DSAEK techniques. The ultimate visual results must also be compared in like fashion. If any of these modifications cause a higher dislocation rate or other complication for the grafted tissue, then any apparent advantage for the ease of surgery is fatally mitigated. Patient safety and improved patient results should always take precedence over the comfort level in surgery for the surgeon.

          The endothelial survival after small incision DLEK/DSEK/DSAEK surgery is quite remarkable. Even with folding the tissue and other donor manipulations described above, the average endothelial cell count after small incision DLEK surgery at 6 months is comparable to PK surgery and is not significantly different from large incision DLEK where the tissue is not folded.14,21 However, our most recent 2 year analysis indicates that folded tissue has a more accelerated loss of endothelial cells from one to two years post-operatively than that found in eyes with large incision DLEK surgery where the tissue is inserted without folding it.26

          The postoperative medical therapy after DSAEK surgery is identical at this time to what is done with PK surgery patients and in our DLEK surgery patients. Topical prednisolone acetate 1% is used four times a day for 3 months, then three times a day until 6 months, then twice a day until 9 months, and then once a day until one year postoperatively. The steroids are then tapered down further until discontinued entirely. In our first 100 cases of DLEK with follow-up between 6 months and 5 years, we have experienced only 4 episodes of graft rejection inflammation and only one case where the graft function was lost.24 This is a lower rate of rejection and graft loss due to rejection than is seen in a similar cohort of PK patients. Steroid therapy after DLEK surgery, therefore, may not be as critical as after PK, but this remains speculation at this point.

          Fluoroquinolone antibiotics are used on a four times a day dosage for the first two weeks after DSAEK surgery and then discontinued.

          Outside of a scientific protocol, DLEK/DSEK/DSAEK patients do not require the same degree of monitoring as standard PK patients and therefore require less post-op clinic time. With no sutures or corneal incisions to worry about, wound healing or ulcerations are not an issue. Astigmatism management is also not an issue after DLEK/DSEK/DSAEK surgery, much to the joy of patient and surgeon alike! The only critical monitoring is for steroid-induced glaucoma as long as the patient is on topical steroids, and this is done according to the clinician’s standard routine.

          The DLEK/DSEK/DSAEK surgical procedure is different than PK and requires a commitment to exacting detail and thorough practice prior to incorporation of this procedure into the surgeon’s operative repertoire. However, with its superior topography, rapid wound healing and long term safety, the endothelial keratoplasty of DSAEK is well worth the effort.

References

1. Sugar A, Sugar J. Techniques in penetrating keratoplasty: a quarter century of development. Cornea 2000;19:603-10.

2. Abou-Jaoude ES, Brooks M, Katz DG, Van Meter WS. Spontaneous wound dehiscence after removal of single continuous penetrating keratoplasty suture. Ophthalmology 2002;109:1291-6.

3. Tseng SH, Lin SC, Chen FK. Traumatic wound dehiscence after penetrating keratoplasty: clinical features and outcome in 21 cases. Cornea 1999;18:553-8.

4. Stechschulte SU, Azar DT. Complications after penetrating keratoplasty. Int Ophthalmol Clin 2000;40:27-43.

5. Akova YA, Onat M, Koc F, Nurozler A, Duman S. Microbial keratitis following penetrating keratoplasty. Ophthalmic Surg Lasers 1999;449-55.

6. Confino J, Brown SI. Bacterial endophthalmitis associated with exposed monofilament sutures following corneal transplantation. Am J Ophthalmol 1985;99:111-13.

7. Ko WW, Frueh BE, Shields CK, Costello ML, Feldman ST. Experimental posterior lamellar transplantation of the rabbit cornea Invest Ophthalmol Vis Sci 1993;34(4):1102.

8. Melles GR, Eggink FA, Lander F, Pels E, Rietveld FJ, Beekhuis WH, Binder PS. A surgical technique for posterior lamellar keratoplasty. Cornea 1998;17:618-26.

9. Terry MA, Ousley PJ. Endothelial replacement without surface corneal incisions or sutures: topography of the deep lamellar endothelial keratoplasty procedure. Cornea 2001;20:14-8.

10. Terry MA, Ousley PJ. Deep lamellar endothelial keratoplasty in the first United States patients: early clinical results. Cornea 2001;20:239-43.

11. Terry MA, Ousley PJ. Replacing the endothelium without corneal surface incisions or sutures: the first United States clinical series using the deep lamellar endothelial keratoplasty procedure. Ophthalmology 2003;110:755-64.

12. Terry MA, Ousley PJ. In pursuit of emmetropia: spherical equivalent refraction results with deep lamellar endothelial keratoplasty (DLEK). Cornea 2003;22:619-26.

13. Terry MA, Ousley PJ. Rapid visual rehabilitation after endothelial transplants with deep lamellar endothelial keratoplasty (DLEK). Cornea 2004;23:143-53.

14. Terry, MA, Ousley, PJ: Small Incision Deep Lamellar Endothelial Keratoplasty (DLEK): 6 months results in the first prospective clinical study. Cornea 2005;24:59-65

15. Terry MA. Endothelial replacement: the limbal pocket approach. Ophthalmol Clin North Am 2003;16:103-12.

16. Terry MA. Deep lamellar endothelial keratoplasty (DLEK): pursuing the ideal goals of endothelial replacement. Eye 2003;17:982-8.

17. Terry MA. A new approach for endothelial transplantation: deep lamellar endothelial keratoplasty. Int Ophthalmol Clin 2003;43:183-93.

18. Terry MA, Ousley PJ. Corneal endothelial transplantation: advances in the surgical management of endothelial dysfunction. Contemporary Ophthalmology 2002;1(26):1-8.

19. Terry, MA: “Endothelial Replacement. In: Krachmer J, Mannis M, Holland E, eds. Cornea: surgery of the cornea and conjunctiva. St. Louis: Elsevier Mosby; 2005: Chapter 140: 1707-1718.

20. Terry MA. The evolution of lamellar grafting techniques over twenty-five years. Cornea 2000;19:611-6.

21. Terry MA, Ousley PJ: Deep Lamellar Endothelial Keratoplasty (DLEK): Visual acuity, astigmatism, and endothelial survival in a large prospective series. Ophthalmology 2005; 112: 1541-154922.

Armour RL, Wilson DJ, Ousley PJ, Terry MA. Invest Ophthalmol Vis Sci. 2004;45:ARVO E-Abstract 2898.

23. Amayem AF, Terry MA, Helal MH, Turki WA, El-Sabagh H, El-Gazayerli E, Ousley PJ. Deep Lamellar Endothelial Keratoplasty (DLEK): surgery in complex cases with severe preoperative visual loss. Cornea 2005; 24: 587-592

24. Terry MA, Ousley PJ: Deep Lamellar Endothelial Keratoplasty (DLEK): Early complications and their management. Cornea  2006; 25: 37-4325.

25. Terry MA, Hoar KL, Wall J, Ousley, PJ. The Histology of Dislocations in Endothelial Keratoplasty (DSEK and DLEK): Prevention of dislocation with a laboratory-based surgical solution in 100 consecutive DSEK cases. Cornea 2006 (in press)

26. Terry MA, Wall J, Hoar KL, Ousley PJ. Endothelial Keratoplasty: A prospective study of endothelial cell loss during the 2 years after deep lamellar endothelial keratoplasty. Ophthalmology 2006 (in press)