Descemet s stripping with endothelial keratoplasty in 200 eyes

J CATARACT REFRACT SURG - VOL 32, MARCH 2006 Descemet s stripping with endothelial keratoplasty in 200 eyes Early challenges and techniques to enhance donor adherence Francis W. Price Jr, MD, Marianne O. Price, PhD PURPOSE: To describe early challenges and techniques to promote donor tissue adherence in Descemet s stripping with endothelial keratoplasty (DSEK). SETTING: Price Vision Group, Indianapolis, Indiana, USA. METHODS: The first 200 consecutive cases of DSEK performed by a single surgeon were analyzed retrospectively. Follow-up was 7 to 20 months for 124 eyes and 2 to 6 months for 76 eyes. The surgical technique consisted of stripping Descemet s membrane and endothelium from the recipient s central cornea and transplanting an 8.0 to 9.0 mm disc of donor endothelium and posterior stroma through a 5.0 mm incision, with sutures used only to close the incision. RESULTS: The most frequent challenge was inadequate donor attachment. Using techniques to remove fluid from the donor recipient graft interface, the donor detachment rate in the last 64 cases was 6%, with half attributable to patient eye rubbing. Detached grafts were reattached by injecting an air bubble to press the donor against the recipient cornea. There were 7 primary graft failures, with only 1 occurring in the second 100 cases, which primarily used microkeratome-dissected donor tissue. Other complications were infrequent and included pupillary block glaucoma (1), aqueous misdirection syndrome (1), and cataract development in 2 of 27 phakic eyes. The DSEK procedure was performed safely before and after laser in situ keratomileusis (1 each). CONCLUSIONS: Early outcomes in the initial 200 consecutive DSEK procedures suggest the technique provides significant advantages over penetrating keratoplasty, including more rapid healing, more predictable refractive outcomes, and better retention of corneal strength and integrity. Although donor adherence was more challenging, DSEK was technically easier and should be less traumatic to anterior chamber structures than earlier posterior grafting techniques. J Cataract Refract Surg 2006; 32:411 418 Q 2006 ASCRS and ESCRS Transplantation of the posterior layers of the cornea to treat endothelial dysfunction has been gaining interest as an alternative to penetrating keratoplasty (PKP) because it provides faster healing and more predictable refractive outcomes. 1 The original posterior lamellar keratoplasty (PLK) technique, described by Melles et al., 2 consisted of manually dissecting the recipient and donor corneas at 80% to 90% stromal depth using curved blades, then excising the posterior recipient stroma and endothelium with an intrastromal trephine and small, curved scissors. A donor button consisting of posterior stroma and endothelium was inserted into the recipient eye through a 9.0 mm incision, and air was injected into the anterior chamber to press the donor button up against the recipient bed. Melles et al. 3 later showed that the donor tissue could be folded in half for insertion through a 5.0 mm scleral tunnel incision as part of a completely sutureless procedure. Deep lamellar endothelial keratoplasty (DLEK) is the same basic technique as Melles et al. s PLK but with instrumentation and procedural refinements. 4,5 Melles et al. 6 subsequently reported a simplified recipient preparation procedure that consisted of scraping Descemet s membrane and endothelium from the recipient cornea instead of performing the more difficult lamellar dissection and excision procedures. Descemet s stripping with endothelial keratoplasty (DSEK) offers a number of Q 2006 ASCRS and ESCRS Published by Elsevier Inc. 0886-3350/06/$-see front matter doi:10.1016/j.jcrs.2005.12.078 411

advantages over the earlier posterior keratoplasty or DLEK technique because no complex recipient trephination is required, there is less potential for trauma to the anterior chamber and lens, and there is minimal concern about inducing ectasia in eyes that have had previous refractive surgery that thinned the cornea. 1 After performing 101 posterior lamellar keratoplasty procedures using Melles et al. s earlier technique, DSEK was then used. As previously reported, visual outcomes of the first 50 DSEK cases showed that this corneal transplantation technique caused minimal refractive change and provided rapid visual recovery. 1 This promising new procedure for treating endothelial dysfunction has also presented challenges. In this article, early difficulties in the first 200 consecutive DSEK procedures are reported and techniques to enhance donor tissue adherence are described. PATIENTS AND METHODS Results are reported for the first 200 consecutive cases of Descemet s stripping with DSEK performed by a single surgeon (F.W.P.) at Price Vision Group, Indianapolis, Indiana, a tertiary referral center for cornea and refractive surgery. Surgery was performed in 200 eyes of 179 patients between December 2003 and May 2005. All patients read and signed an informed consent document. The work was compliant with the Health Insurance Portability and Accountability Act of 1996 and consistent with Good Clinical Practices and the Declaration of Helsinki, 1996. Surgical Technique For DSEK, the donor cornea was prepared first, followed by surgery on the recipient. In 112 of the 200 cases, primarily the earlier cases in the series, the donor tissue was dissected manually, whereas in the remaining 88 cases, primarily the later cases, the donor tissue was dissected with a microkeratome (model CB, Moria USA). For manual dissection, a donor corneoscleral shell of at least 15.0 mm diameter was mounted on an artificial anterior chamber and the chamber was inflated with air to create a light reflection that enabled the surgeon to estimate the depth of the Accepted for publication August 1, 2005. From the Price Vision Group (F.W. Price) and the Cornea Research Foundation of America (M.O. Price), Indianapolis, Indiana, USA. Presented in part at the ASCRS Symposium on Cataract, IOL and Refractive Surgery, Washington, DC, USA, April 15 19, 2005. F.W. Price received travel grants from Moria. The other author has no financial or proprietary interest in any material or method mentioned. Clorissa Quillin, Stacey Henderson, Susan Simon, and the staff of Price Vision Group for provided outstanding technical assistance. Reprint requests to Marianne O. Price, PhD, Cornea Research Foundation of America 9002 North Meridian Street, Suite 212 Indianapolis, Indiana 46260, USA. E-mail: fprice@pricevisiongroup. net. lamellar dissection. The tissue was dissected at approximately 80% to 90% stromal depth using a series of 3 curved blades of increasing length (DORC International). 7 Microkeratome dissections were performed using a CB microkeratome equipped with a 300 mm head (83 cases) or 350 mm head (5 cases) and associated artificial anterior chamber (Moria). The 350 mm head was received late in the series, and it was used in cases in which the precut donor pachymetry exceeded 575 mm after the epithelium was removed. After dissection with either technique, the donor tissue was transferred to a punching system and cut with an 8.0 mm, 8.5 mm, or 9.0 mm diameter trephine. The donor tissue was then covered with tissue storage solution pending use later in the procedure. Surgery was performed on the recipient using monitored anesthesia with retrobulbar block. The recipient epithelium was marked lightly with the trephine used on the donor tissue to delineate where to strip Descemet s membrane and place the donor tissue. A 5.0 mm temporal incision was made. In most cases, a scleral tunnel incision was used to make the wound as watertight as possible and to minimize induced astigmatism from flattening the cornea in the meridian of the incision. However, clear corneal incisions were used early in the series and when cataract surgery was performed as a combined procedure with DSEK. A modified Price-Sinskey hook (Asico) was used to score Descemet s membrane in a circular pattern under the area of the epithelial reference mark. A 45-degree or 90-degree Descemet s stripping instrument (DORC International) was used to strip off Descemet s membrane and endothelium within the scored area. Once fully stripped, Descemet s membrane and the endothelium were removed from the anterior chamber with a forceps. The donor cornea was brought into the operative field, a small amount of viscoelastic material comprising sodium hyaluronate 1% (Healon or Provisc) was placed on the endothelial surface, and the posterior portion was folded over itself with the endothelial side inward. The folded donor tissue was grasped gently with a long Kelman-McPherson forceps and inserted into the eye. Air was injected into the anterior chamber to help unfold the donor tissue and press it up against the recipient cornea. In 1 case, a vitreous tap was performed to relieve back pressure. An 18-gauge needle was introduced through the sclera 3.5 mm posterior to the limbus through the pars plana of the interior of the eye into the midportion of the vitreous cavity. Vitreous was aspirated, allowing the iris to drop posteriorly, which deepened the anterior chamber. This allowed the donor tissue to be inserted into the anterior chamber and the case to proceed as normal. Techniques to Promote Donor Tissue Adherence In the first 10 cases, the air was left in the eye for 5 minutes in the operating room. It was then removed and replaced with a balanced salt solution because this was the procedure that had been used in 101 previous cases of PLK. However, the donor tissue had to be repositioned the day after surgery in 5 of the first 10 DSEK cases, suggesting that it did not adhere as readily to the smooth recipient surface left behind after stripping Descemet s membrane as it did to a hand-dissected stromal surface. In all subsequent cases, the anterior chamber was left filled with air for 8 minutes in the operating room, after which most of the air was removed and 1 drop of homatropine 5% was applied to prevent pupillary block. Patients were then sent to the recovery room to lie with their face up for 30 to 60 minutes to allow the remaining air to push the donor tissue up against the recipient cornea. The corneal 412 J CATARACT REFRACT SURG - VOL 32, MARCH 2006

surface was massaged with a Lindstrom laser in situ keratomileusis (LASIK) flap roller (BD Medical) while the anterior chamber was filled with air to help remove residual fluid trapped between the donor and recipient corneas. In the last 64 cases, to further reduce the donor detachment rate, a 15-degree blade was used to make 3, or usually 4, equally spaced small incisions in the midperipheral recipient cornea down to the graft interface. This was done while the anterior chamber was filled with air, immediately after the corneal surface was massaged with the flap roller. Twisting the blade slightly often helped drain a significant amount of fluid (Figure 1) even though the view through the operating microscope suggested that the donor was already in perfect apposition to the recipient. Repositioning Donor Tissue In 28 of the 200 cases, the donor tissue was not well adherent postoperatively. Two detachments were noted several weeks after surgery, and the rest were noted within the first week. To reattach the donor tissue, the patient was taken to a minor surgery room and laid under an operating microscope. Using standard sterile technique, a small amount of anterior chamber fluid was removed with a 30-gauge needle and air was injected to fill the anterior chamber and push the donor tissue up against the recipient cornea. The air injection needle was placed through virgin posterior limbal tissue into the anterior chamber. In cases in which the donor had completely detached, it was positioned carefully between the donor cornea and underlying iris before air was injected. In all cases, the patient was asked repeatedly if the light still appeared bright to ensure there was no vascular occlusion of the retina. The air was left in place for 1 to 2 hours; then most was removed and replaced with a balanced salt solution to prevent pupillary block. Reoperations for Primary Donor Failures Any graft that failed to clear in the first 2 weeks after surgery was termed a primary graft failure, consistent with the definition used in earlier reports of a series of 3992 consecutive PKPs. 8,9 Seven cases of primary graft failure were regrafted 1 week after the original surgery. In 6 cases, the prior incision was reopened and the donor button was removed with a forceps. A new donor button was prepared and implanted in the eye as described above. The remaining case was an Alzheimer s patient who kept rubbing her eye; the eye was regrafted with a PKP. In each of these cases, the donor tissue had first detached and been repositioned with a repeat air injection procedure; it detached again and was regrafted. RESULTS Demographics Tables 1 and 2 summarize demographic, surgical, and postoperative procedure information for the 200 DSEK cases in 179 patients. Length of follow-up was 13 to 20 months in 50 eyes, 7 to 12 months in 74 eyes, and 2 to 6 months in 76 eyes. Normal Postoperative Course As described, the donor tissue usually adhered to the posterior surface of the recipient cornea without fluid or space between the 2. Variable amounts of corneal edema persisted for several days to weeks as the new donor tissue eliminated excess corneal hydration. For the first few days after surgery, the central cornea was typically not as clear as seen in PKP-grafted tissue. The peripheral portion of grafted tissue also initially appeared edematous along the edge where bare stroma was exposed to the anterior chamber fluid, and this usually took a few weeks to fully clear. When donor tissue did not completely cover any area of the peripheral recipient cornea from which Descemet s membrane and endothelium had been removed, the overlying stroma and epithelium appeared edematous, often with bullae, and these areas took up to a few months to clear, depending on size. Therefore, after the first few cases, it became standard practice to remove the recipient Descemet s membrane and endothelium from an area smaller than the diameter of the donor button. Figure 2 shows a representative eye 6 months after DSEK; the entire central cornea was crystal clear, and the donor edge showed no residual edema. Table 1. Patient demographics. Demographic Value (%) Figure 1. Intraoperative photograph shows the fluid draining from a midperipheral corneal incision penetrating to the donor recipient graft interface (arrow). No. of patients 179 d No. of women 118 (66) No. of whites 175 (97) No. of African Americans 4 (2) No. of Hispanics 1 (1) Mean age (y) G SD 69 G 12 J CATARACT REFRACT SURG - VOL 32, MARCH 2006 413

Table 2. Preoperative, operative, and postoperative data for initial 200 consecutive DSEK cases. Finding Number (%) Diagnosis Fuchs dystrophy 181 (90) Corneal edema 19 (10) Recipient preoperative lens status Pseudophakic 147 (73) Phakic 51 (26) Aphakic 2 (1) Procedures combined with DSEK Phacoemulsification/IOL implantation 24 (12) IOL exchange 1 (0.5) Secondary IOL placement 1 (0.5) Epithelial removal 62 (31) Pars plana vitrectomy 4 (2) Vitreous tap 1 (0.5) Postoperative surgical procedures Donor tissue reposition 28 (14) DSEK for primary donor failure 6 (3) PKP for primary donor failure 1 (0.5) PKP for secondary graft failure 2 (1) Phacoemulsification/IOL implantation 2 (1) Secondary IOL implantation 1 (0.5) Superficial keratectomy 4 (2) Pars plana vitrectomy 1 (0.5) LASIK 1 (0.5) DSEK Z Descemet s stripping with endothelial keratoplasty; IOL Z intraocular lens; LASIK Z laser in situ keratomileusis; PKP Z penetrating keratoplasty Donor Dissection Technique Donor tissue was dissected manually in 112 cases and by microkeratome in 88 cases (Table 3). The microkeratome cases were primarily in the latter half of the series, so a higher percentage of them were done in combination with the fluid drainage technique to promote donor adherence. With microkeratome dissection, there were no donor tissue perforations, and the incidence of primary donor failure was lower compared with hand dissection. The microkeratome-dissected tissue was never extremely thin, as occasionally happens with hand-dissected tissue, so it was easier to handle and unfold, and this may have been less traumatic to the endothelium (Table 3). Recipient Lens Status Before DSEK, 147 eyes were pseudophakic, 2 were aphakic, and 51 were phakic. An anterior chamber intraocular lens (IOL) was exchanged for a posterior chamber IOL as a combined procedure with DSEK in 1 pseudophakic eye. A secondary IOL was implanted as a combined procedure with DSEK in 1 of 2 aphakic eyes; the other aphakic eye had a secondary IOL implanted 4 months after DSEK. Figure 2. Slitlamp photograph of a typical DSEK graft 6 months after surgery. The edge of the donor tissue is illuminated by the slitbeam (arrow). Phacoemulsification and IOL implantation were performed as a combined procedure with DSEK in 24 phakic eyes. Two of 27 eyes left phakic subsequently had cataracts and were treated successfully with phacoemulsification and IOL implantation through a 2.8 mm clear corneal incision. In both phakic eyes that had cataracts, the donor tissue had dislocated after DSEK, requiring extra manipulation and a postoperative air injection procedure. Removal of Recipient Epithelium The epithelium was removed as a combined procedure with DSEK in 62 cases to treat anterior corneal dystrophy or to remove cloudy epithelium and provide a clearer view into the eye. In 4 cases, superficial keratectomies were performed postoperatively to treat anterior corneal dystrophy that became apparent once the dysfunctional endothelium had been replaced, allowing the epithelial edema to clear. Table 3. Incidence of intraoperative and early postoperative difficulties with manual and microkeratome dissection techniques. Manual Donor Dissection Microkeratome Dissection Difficulty n (%) n (%) Total cases 112 d 88 d Donor tissue perforation 5 (4) 0 (0) Donor detachment 18 (16) 8 (9) (Excluding first 10 cases)* 13 (12) d d Primary graft failure 6 (5) 1 (1) *The microkeratome dissections were clustered toward the end of the series. As described in Patients and Methods, a partial-air-bubble technique was used to promote donor adherence after the first 10 cases. A fluid drainage technique was added in the last 64 cases to further promote donor adherence. 414 J CATARACT REFRACT SURG - VOL 32, MARCH 2006

Other Combined Procedures In 1 case, a vitreous tap was performed because vitreous back pressure was noted with rapid and firm shallowing of the anterior chamber when an attempt was made to insert the donor tissue. No postoperative complications of the posterior segment occurred in this eye. Donor Detachments and Graft Failures Donor tissue dislocation was the most common postoperative challenge. Figure 3 illustrates a total donor detachment at the 1-day postoperative examination. The incidence of donor detachment decreased as the series progressed (Figure 4), reflecting the learning curve described in the Patients and Methods section. After a 50% donor dislocation rate in the first 10 cases, patients were then asked Figure 3. Slitlamp photograph of donor tissue detachment 1 day after DSEK. 60% 50% 40% 30% 20% 10% 0% Donor detachments Primary Graft Failures cases 1-10 cases 11-135 cases 136-200 Figure 4. Incidence of donor tissue detachment and primary graft failure. Beginning with case 11, the cornea was massaged to remove fluid from the donor recipient graft interface and the patient remained face up in the recovery area for 30 to 60 minutes with a partial air bubble in the eye to promote donor tissue attachment. In cases 136 through 200, fluid was also drained from the graft interface through 4 midperipheral corneal incisions. to lie face up for 30 to 60 minutes in the recovery area with a partial air bubble in the eye to help press the donor tissue up against the recipient cornea. Also, the cornea was massaged with a Lindstrom LASIK roller to help remove fluid from the graft interface. These additional procedures reduced the donor detachment rate to 13% (17 in 126 cases). In the last 64 cases, fluid was drained from the donor recipient graft interface through midperipheral corneal incisions to further promote donor adherence (Figure 1). The incidence of spontaneous donor detachment dropped to 2 of 64 (3%) after fluid was drained. Two additional donor detachments occurred in the last 64 cases as a result of patients rubbing their eyes despite instructions not to, this resulted in a total detachment rate of 6% in the last 64 cases. Twelve cases of donor dislocation were detected and treated within the first 2 days after surgery; 12 were treated 3 to 8 days after surgery, and 2 were treated between 2 and 3 weeks after surgery. Occasionally at the 1-day examination, some detachment was noted, but it was difficult to determine the extent, so the eye was rechecked 1 day to several days later. In 1 case, an eye had aphakic bullous keratopathy for at least 12 years before surgery and the early postoperative view through the cornea was extremely poor; at the 2-week examination, it was determined that the graft was fully detached. It was repositioned with air without any problem. In another case, detachment was noted at a 3-week examination after the graft had been clear and attached at 1 day and 1 week. This detachment was thought to be secondary to eye rubbing. At least 2 of the dislocations detected and treated at the 1-week examination also seemed to be caused by eye rubbing after the graft was attached on the first postoperative day. In many cases, it was not obvious why donor adherence was suboptimal, but in some cases it may have been related to the dissection technique. In 1 case, a manual donor tissue dissection was extremely thin, comprising little more than Descemet s membrane and endothelium. This tissue was difficult to manipulate and position in the anterior chamber. After an attempt was made to improve attachment with a postoperative air injection procedure, the donor button again detached and was replaced with new donor tissue. In another case involving a manual dissection, a partial donor detachment appeared to be caused by residual tags of stromal tissue along the edge of the donor button, which allowed fluid into the host graft interface and prevented tight donor recipient apposition. This initially caused edema in approximately one third of the graft, but over the next couple of months the interface sealed over and the edema resolved. In another case, the trephination of the corneal donor tissue was not centered on the microkeratome dissection, causing 1 edge of the donor to be full thickness. The thick edge was on the underside as the donor tissue was folded into a taco configuration for implantation and was not noted. The thick J CATARACT REFRACT SURG - VOL 32, MARCH 2006 415

edge along about one third of the donor button prevented it from applanating against the recipient, so fluid continued to enter the interface. A postoperative air injection procedure was performed before the graft was replaced. This was the only primary graft failure in 88 cases using microkeratome-dissected donor tissue, and it was one of the few detachments in the last 64 cases. The incidence of primary donor failures decreased with the reduction in donor detachments as the series progressed (Figure 4). Donor detachment and repeat air injection preceded each primary donor failure. In 6 of 7 cases, the donor tissue was removed and replaced with a freshly dissected donor button. The remaining case was an Alzheimer s patient who could not be restrained from rubbing her eyes, and she had regrafting by PKP. During this series, no cases of planned primary DSEK were converted to PKP. Two grafts that failed within the first year of follow-up were regrafted with PKP. In 1 case, a central scar in the lamellar graft interface limited best spectaclecorrected visual acuity (BSCVA) to 20/80. The other patient experienced acute toxic anterior chamber syndrome after surgery. Other Postoperative Procedures There was 1 complication associated with a residual air bubble. A patient had pupillary block and intraocular pressure (IOP) was 54 mm Hg at the 1-day postoperative examination; a paracentesis site was opened, and the air was removed. The pupil remained fixed and dilated. In another case, a patient had aqueous misdirection syndrome. The IOP was 72 mm Hg on the first postoperative day, so a pars plana vitrectomy was performed, which resolved the problem. Laser in situ keratomileusis was performed in 1 eye 8 months after DSEK to correct a refractive error. DSEK After LASIK One patient in the series had Fuchs corneal dystrophy diagnosed after she had bilateral LASIK and bilateral central corneal edema developed with dense guttata. The LASIK was performed 12 months earlier at another center, and it is not known whether guttata were present preoperatively. With glare, her visual acuity was 20/200. One month after combined DSEK, phacoemulsification, and IOL implantation was performed, visual acuity was 20/40. DISCUSSION Descemet s stripping with endothelial keratoplasty offers significant advantages over penetrating keratoplasty and earlier posterior keratoplasty techniques. 10 The most significant advantages of DSEK over PKP are that the recipient cornea remains structurally intact and is more resistant to injury. 11 Recovery of useful vision occurs within weeks. 1 The preoperative refractive status of the eye is maintained, and visual fluctuations are minimal during the healing process. 1,12 No intricate suture placement methods are required, questions about when to remove sutures are eliminated, and concerns about late wound dehiscence after suture removal are alleviated. Furthermore, the risk for expulsive intraoperative suprachoroidal hemorrhage is minimized because a 5.0 mm scleral tunnel incision can be closed quickly. These advantages are especially important when trying to restore corneal clarity to patients who have visual potential in 1 eye only; 9 patients in our series had useful vision in the surgical eye only. The principal advantages of DSEK over PLK or DLEK are that no complex recipient dissection and trephination techniques are required and there is less potential for trauma to the anterior chamber and lens. In other reported series, 2 of 32 DLEK cases and 1 of 7 PLK cases were converted to PKP intraoperatively because of difficulties with recipient dissection and trephination. 4,13 We had no attempted DSEK cases that had to be converted to PKP. The DSEK technique also presents new challenges, primarily associated with trying to fixate the donor tissue to the smooth posterior surface of the recipient stroma. In Melles et al. s 2 original PLK technique, the donor button diameter was cut slightly smaller than the excised recipient bed and the donor tissue was placed into the recessed cavity created by the recipient excision. This helped prevent the donor tissue from sliding out of position. Terry recommended extending the recipient lamellar dissection beyond the area to be excised and attempting to match the recipient excision diameter to the diameter of the donor disc. To accommodate any size mismatch between the free-hand-cut recipient disc and the trephine-punched donor disc and to help minimize the chance of donor tissue dislocation, he further recommended tucking the edges of the donor disc into the peripheral recipient lamellar dissection with a reverse Sinskey hook. Despite these modifications, Terry experienced a 6% donor detachment rate in his first 90 DLEK cases (Ophthalmology Times, May 15, 2005). Sano 14 reported 1 donor detachment in his first 3 DLEK cases. With DSEK, the donor tissue is not held in place by a recessed edge, nor is it tucked into position. Furthermore, the recipient stromal interface left behind after stripping Descemet s membrane clinically appears smoother than a PLK/DLEK hand-dissected stromal surface and probably provides less texture to prevent the donor from sliding out of position. Initially, we did not fully appreciate how smooth and slick the posterior corneal surface was after stripping Descemet s membrane. The surgeon in our series (F.W.P.) had performed 101 PLK surgeries with a 5% detachment rate before converting to DSEK and experienced 5 detachments in the first 10 DSEK cases. Techniques 416 J CATARACT REFRACT SURG - VOL 32, MARCH 2006

developed to eliminate fluid from the donor recipient graft interface significantly reduced the donor detachment rate. We caution patients not to rub their eyes after surgery because we believe that pushing or rubbing the eye in the early postoperative period, and possibly even hard squeezing, can indent the cornea and separate the edges of the donor and recipient, allowing fluid into the interface. Once enough fluid separates the donor and recipient, the donor can detach. We have reproduced donor tissue dislocation in the operating room. In eyes with the anterior chamber filled with a balanced salt solution and only a small central air bubble (after 8 minutes of air in the anterior chamber and drainage of any interface fluid), pushing on the cornea or rubbing across it with an instrument will cause dislocation of the donor tissue into the anterior chamber. This dislocation occurs very quickly and easily in the operating room, and we suspect it can in the early postoperative period as well. In our experience, the advantages of DSEK far outweigh the challenges. As previously reported, the visual outcomes in the first 50 eyes in this series were excellent. 1 Six months after DSEK, mean manifest cylinder was 1.5 G 0.94 diopters (D), unchanged from the preoperative value of 1.5 G 1.0 D, and mean manifest spherical equivalent refraction was 0.15 G 1.5 D, statistically comparable to the preoperative value. Preoperative mean BSCVA was 20/100. Statistically significant improvement in BSCVA was noted at the 3-month and 6-month examinations (P Z.007). Six months after DSEK, 62% of the eyes refracted to R20/40 and 76% saw R20/50. These visual outcomes equaled or exceeded those in the largest reported PKP and DLEK series of Fuchs and endothelial dystrophy cases. 4,15 18 In addition to the advantages previously cited, the risk for postoperative ocular surface complications is minimal with DSEK because it is performed through a small incision, which may require no sutures or, at most, a few that are removed within a couple of months. Our previous analysis of long-term outcomes of 3992 PKP procedures showed ocular surface complications were the most common cause of graft failure in the first year after surgery and were responsible for 18% of the graft failures overall. 8,19 Ocular surface complications are of greater concern after PKP because there are multiple corneal sutures and the corneal surface is anesthetic. In contrast, endothelial keratoplasty can be performed through a 5.0 mm scleral tunnel incision that is covered with conjunctiva, so suture-related and neurotrophic concerns are virtually eliminated (Baratz KH, et al. IOVS 2005; 46:ARVO E-Abstract 2703). Confocal microscopy shows subepithelial central corneal innervation 3 months and 6 months after DSEK (Figure 5). The major advantage of DSEK over PLK or DLEK techniques is that it is easier to perform because it does not Figure 5. Confocal microscope images of subepithelial central corneal innervation in 2 different patients 3 months (A) and 6 months (B) after DSEK. require recipient lamellar dissection and excision of the recipient button using scissors inside the eye. This is particularly advantageous in phakic eyes, which have a shallower anterior chamber. Five of our first 50 PLK eyes were phakic, and 4 (80%) developed cataracts, 20 whereas 24 of our first 200 DSEK eyes were phakic after surgery, and 2 (8%) developed a cataract. Another important advantage of DSEK over PLK or DLEK is that there is no reduction in stromal thickness, alleviating concerns about postoperative ectasia. Currently, there is no information available on what dissection depth or degree of corneal thinning is safe in PLK/DLEK procedures. Experience from millions of LASIK procedures suggests that a residual corneal bed thickness of at least 250 mm or 300 mm should be maintained to prevent corneal ectasia J CATARACT REFRACT SURG - VOL 32, MARCH 2006 417

and refractive instability. As the current population of LASIK and PRK patients grow older, some will eventually experience endothelial dysfunction requiring surgical treatment. Surgical techniques such as PLK or DLEK, which involve lamellar dissection with removal of posterior stromal tissue, could lead to ectasia. In fact, we believe it would be inadvisable to perform PLK or DLEK in eyes that previously had LASIK or to perform LASIK in eyes with previous PLK or DLEK because both the anterior and posterior stroma would have been compromised, leaving only a thin section of uncut central stroma. In contrast, DSEK does not thin the stroma, so it should not cause corneal instability or significant refractive change. We have performed LASIK without problem on a postoperative DSEK patient to correct residual refractive error. Also, 1 patient in this series who previously had LASIK subsequently had uneventful DSEK surgery. The alternative in this case would have been to perform a standard PKP; however, the refractive effect would have been unpredictable and most likely left the patient myopic again. The DSEK technique was the logical choice to treat the corneal decompensation and guttata because there was no evidence of ectasia or weakening of the cornea from the LASIK. CONCLUSION Descemet s stripping with endothelial keratoplasty is a promising new corneal transplant procedure that provides rapid visual recovery and predictable refractive outcomes. We hope this report of early experiences with DSEK will accelerate the learning curve of other surgeons and help prevent problems in their initial series of patients. REFERENCES 1. Price FW Jr, Price MO. Descemet s stripping with endothelial keratoplasty in 50 eyes; a refractive neutral cornea transplant. J Refract Surg 2005; 21:339 345 2. Melles GRJ, Eggink FAGJ, Lander F, et al. A surgical technique for posterior lamellar keratoplasty. Cornea 1998; 17:618 626 3. Melles GRJ, Lander F, Nieuwendaal C. Sutureless, posterior lamellar keratoplasty: a case report of a modified technique. Cornea 2002; 21:325 327 4. Terry MA, Ousley PJ. Rapid visual rehabilitation after endothelial transplants with deep lamellar endothelial keratoplasty (DLEK). Cornea 2004; 23:143 153 5. Terry MA, Ousley PJ. Small-incision deep lamellar endothelial keratoplasty (DLEK); six-month results in the first prospective clinical study. Cornea 2005; 24:59 65 6. Melles GRJ, Wijdh RHJ, Nieuwendaal CP. A technique to excise the Descemet membrane from a recipient cornea (descemetorhexis). Cornea 2004; 23:286 288 7. Melles GRJ, Rietveld FJR, Beekhuis WH, Binder PS. A technique to visualize corneal incision and lamellar dissection depth during surgery. Cornea 1999; 18:80 86 8. Thompson RW Jr, Price MO, Bowers PJ, Price FW Jr. Long-term graft survival after penetrating keratoplasty. Ophthalmology 2003; 110: 1396 1402 9. Price FW Jr, Whitson WE, Collins KS, Marks RG. Five-year corneal graft survival; a large, single-center patient cohort. Arch Ophthalmol 1993; 111:799 805 10. Price FW Jr. Corneal transplantation as a refractive surgical procedure [guest editorial]. J Refract Surg 2005; 21:216 217 11. Elder MJ, Stack RR. Globe rupture following penetrating keratoplasty; how often, why, and what can we do to prevent it? Cornea 2004; 23:776 780 12. Terry MA, Ousley PJ. In pursuit of emmetropia: spherical equivalent refraction results with deep lamellar endothelial keratoplasty (DLEK). Cornea 2003; 22:619 626 13. Melles GRJ, Lander F, van Dooren BTH, et al. Preliminary clinical results of posterior lamellar keratoplasty through a sclerocorneal pocket incision. Ophthalmology 2000; 107:1850 1856; discussion by HE Kaufman, 1857 14. Sano Y. Corneal endothelial transplantation: results of a clinical series using deep lamellar endothelial keratoplasty (DLEK). Cornea 2004; 23(suppl):S55 S58 15. Price FW Jr, Whitson WE, Marks RG. Progression of visual acuity after penetrating keratoplasty. Ophthalmology 1991; 98:1177 1185 16. Williams KA, Hornsby NB, Bartlett CM, et al, eds. The Australian Corneal Graft Registry 2004 Report. Bedford Park, Australia, AGCR Publication, 2004: 154 17. Pineros O, Cohen EJ, Rapuano CJ, Laibson PR. Long-term results after penetrating keratoplasty for Fuchs endothelial dystrophy. Arch Ophthalmol 1996; 114:15 18 18. Claesson M, Armitage WJ, Fagerholm P, Stenevi U. Visual outcome in corneal grafts: a preliminary analysis of the Swedish Corneal Transplant Register. Br J Ophthalmol 2002; 86:174 180 19. Price MO, Thompson RW Jr, Price FW Jr. Risk factors for various causes of failure in initial corneal grafts. Arch Ophthalmol 2003; 121: 1087 1092 20. Price MO, Price FW Jr. Cataract progression and treatment following posterior lamellar keratoplasty. J Cataract Refract Surg 2004; 30: 1310 1315 418 J CATARACT REFRACT SURG - VOL 32, MARCH 2006