For many years, penetrating keratoplasty (PK) has been the only reliable method for treating corneal edema from endothelial decompensation. Refractive outcomes after PK are difficult to predict, and patients typically experience significant refractive shifts as the full-thickness corneal wound heals and sutures are removed over several years.1"4 In some cases, visual recovery can be prolonged, and refractive changes may continue to occur for many more years. As a result, patients with Fuchs' corneal dystrophy often postponed having PK, despite experiencing significant glare and haze that interfered with daily activities, such as driving at night.
New techniques for replacing only the dysfunctional posterior portion of the cornea, rather than the full corneal thickness, offer significant advantages over PK. Most notably, posterior lamellar techniques maintain much of the structural integrity of the cornea and induce minimal refractive change. Melles' original technique, posterior lamellar keratoplasty (PLK), 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 scissors.5,6 The posterior donor stroma and endothelium were placed endothelial side down on a viscoelastic-covered spatula for insertion into the recipient eye through a 9-mm corneoscleral incision, and air was injected to press the donor button up into the recipient bed. Sutures were used only to close the corneoscleral incision. Melles subsequently showed that the donor tissue could be folded in half for insertion through a 5-mm scleral tunnel incision.7 Terry and Ousley8 reported outcomes with deep lamellar endothelial keratoplasty (DLEK), which is similar to PLK but uses modified instrumentation.9
Surgeons have been slow to adopt PLK/DLEK because it is technically challenging. The most difficult aspect of PLK/DLEK is lamellar dissection and excision of the posterior portion of the recipient cornea. Melles subsequently reported a technique for stripping Descemets membrane from the recipient cornea to remove the diseased or dysfunctional endothelial layer, thus eliminating the need to perform the intracorneal lamellar dissection and recipient button excision.10 Descemets stripping with endothelial keratoplasty (DSEK) has several advantages compared with the earlier PLK/ DLEK procedure: 1) it is less traumatic to the remaining recipient cornea and the underlying iris and anterior chamber; 2) it leaves the recipient cornea structurally intact because no stromal tissue is removed; and 3) the resulting recipient corneal interface is smooth, which may provide better visual acuity compared with a handdissected surface. This article reports early visual acuity and refractive outcomes from the initial 50 DSEK surgeries performed at our center.
PATIENTS AND METHODS
This was a retrospective analysis of the initial 50 consecutive DSEK cases performed at Price Vision Group, a tertiary referral center for corneal and refractive surgery in Indianapolis, Ind. A single surgeon (F.W.P.) performed the surgeries between December 2003 and July 2004. Surgery was performed on 50 eyes of 47 patients. All participants read and signed an informed consent document and the work was compliant with the Health Insurance Portability and Accountability Act of 1996.
Mean patient age was 70±12 years (range: 34 to 89 years), and 64% of patients were female. Two percent were Hispanic, 2% were African -American, and 96% were Caucasian. Five (10%) eyes were treated for corneal edema or bullous keratopathy and 45 (90%) eyes were treated for Fuchs' endothelial dystrophy. Seven eyes were phakic and 43 eyes were Pseudophakie at the time of surgery.
DSEK SURGICAL TECHNIQUE
The donor cornea was prepared first, followed by surgery on the recipient. A donor corneoscleral shell of at least 15 mm diameter was secured on an artificial anterior chamber (model discontinued, any suitable artificial anterior chamber could be used) and air was injected to fill the chamber. A set of three dissecting blades (DORC International, Zuidland, The Netherlands) was used to perform the dissection across the entire cornea of the donor using the technique described by Melles et al.5 The change in index of refraction between the cornea and the air behind it created a reflection that allowed the surgeon to estimate the depth of the lamellar dissection6 and an attempt was made to perform each dissection at approximately 90% stromal depth. Once the dissection was completed, the donor tissue was transferred to a punching system (Price Corneal Punch; ASICO, Westmont, 111) and cut from the posterior surface with an 8.0- or 8. 5 -mm trephine. This punching system was designed to provide vertical alignment and to center the trephination on the center of the lamellar dissection. Prior to removing the donor cornea from the artificial anterior chamber, the center of the dissection was marked on the epithelial surface with gentian violet. The tissue was covered with some of the tissue storage solution pending use later in the procedure.
Surgery was performed using monitored anesthesia with retrobulbar block. The recipient corneal epithelium was lightly marked with the trephine used to cut the donor tissue, to outline the area for removal of Descemets membrane and placement of the donor tissue. A temporal 5-mm incision was made. Three-stepped clear corneal or limbal incisions were used for the first 22 cases, whereas cases 23 to 50 used scleral tunnel incisions to provide a more watertight wound and to minimize induced astigmatism from corneal flattening in the meridian of the incision. A diamond blade was used for the initial vertical incision to approximately 1/3 to 1/2 depth; a crescent blade was used to create a lamellar step, and finally the anterior chamber was entered with a 15° sharp metal blade. Descemets membrane was scored with a modified Price-Sinskey hook (ASICO) along the perimeter of the area to be excised, and either of two Descemets strippers, 45° or 90° models (DORC International), were used to strip off Descemets membrane with attached dysfunctional endothelium from within the circumscribed area. Once Descemets was fully stripped, it was removed from the anterior chamber with forceps. In cases with edematous or hazy epithelium, the central epithelium of the cornea was removed to provide a clearer view into the anterior chamber. No staining of Descemets membrane was used for this series of eyes. No other topical agents were used, such as glycerine, to clear the cornea. In some cases, Descemets membrane fragmented, and attention to detail was required to ensure that all portions in the central cornea had been removed from the eye. In all cases, Descemets membrane was laid on the anterior corneal surface of the patient as it was removed so the surgeon could determine how much was removed and if a portion of the central area was still present within the eye.
The donor cornea was brought into the operative field, a small amount of viscoelastic was placed on the endothelial surface, and the posterior portion of the donor button was folded over on itself in a "taco" configuration. The donor taco was grasped along its longer edge with long Kelman-McPherson forceps and placed into the eye in one smooth motion that carried the leading edge across the anterior chamber to the far side of the angle prior to releasing the hold of the forceps. When grasping the donor tissue, care was taken not to squeeze the tissue, as that could possibly damage the endothelial cells. The donor tissue was unfolded and pushed up against the recipient cornea with air. The air completely filled the anterior chamber and was left in the eye for 5 to 8 minutes. Five minutes was used in early cases, but eventually the time was increased to promote donor button adherence. The air was then removed and replaced with balanced salt solution using a standard 30-gauge cannula, leaving just a small bubble that was not large enough to cause pupillary block. The patient waited face-up in the recovery area for 1 hour prior to leaving the facility. Antibiotic and steroid ointment was placed in the eye, and the eye was patched and shielded.
Three cases of primary graft failure were regrafted 1 week after the original surgery. The original incision was reopened and the donor button was removed with forceps. A new donor button was prepared and placed in the eye as previously described. The results include the visual and refractive outcomes after donor button replacement in these three eyes.
Manifest refractions and visual acuities from the preoperative and 1-, 3-, and 6-month postoperative examinations were analyzed. Eyes with count fingers visual acuity prior to surgery were difficult to accurately refract, therefore the glasses prescription was used as the preoperative refraction, when available. In 2 of 50 cases, neither an accurate preoperative refraction nor glasses prescription was available. Snellen best spectacle-corrected visual acuity (BSCVA) measurements were converted to logarithms (logMAR) to allow averaging and statistical analysis.11 One way analysis of variance (ANOVA) was used to detect differences between examinations for each parameter. If a statistically significant difference was detected, Duncan's multiple range test was used to detect differences between pairs of means. Outcomes of the preoperative and 6-month examinations were also compared using the Student t test. Values were reported as mean± standard deviation. SAS statistical software (SAS Institute Ine, Cary, NC) was used for all analyses. A P value <.05 was considered statistically significant.
Manifest Refraction (D) Before and 6 Months After DSEK in 50 Eyes (Mean±Standard Deviation [range])
VISUAL AND REFRACTIVE OUTCOMES
Preoperative, 1-, 3-, and 6-month examinations were completed on all 50 eyes. The mean manifest cylinder and spherical equivalent refractions at the 6-month postoperative examination did not differ significantly from those at the preoperative examination (P=. 88 and P=. 10, respectively) (Table 1). The mean cylindrical component of the refraction transiently increased at the 1 -month examination (P=. 03 5) (Fig 1). However, after suture removal at approximately 2 months, mean cylinder returned to the preoperative level by the 3-month examination (Fig 1). Mean manifest spherical equivalent refraction did not change significantly between the pre- and postoperative examinations (P=. 10) (Fig 2). Comparative preoperative and 6-month postoperative cylinder and spherical equivalent refractive outcomes for the individual eyes in the series are shown in Figures 3 and 4.
Mean BSCVA steadily improved from 1 to 6 months after surgery (Fig 5). Mean BSCVA was 0.47±0.34 (Snellen equivalent 20/60) at the 3-month examination, and 0.39±0.34 (Snellen equivalent 20/50) at the 6-month examination, a significant improvement over the preoperative BSCVA of 0.66±0.54 (Snellen equivalent 20/100) (P=.007). By the 6-month examination, 31 (62%) of 50 eyes refracted to ^20/40, 38 (76%) of 50 eyes saw ≥20/50, and 47 (94%) eyes had ^20/200 visual acuity. Ten of 12 eyes that did not achieve ^20/50 vision and all 3 eyes with visual acuity < 2 0/2 00 had significant retinal problems, which limited useful vision. The incidence of retinal problems, including age-related macular degeneration (n=8), diabetic retinopathy (n=1), macular hole (n=1), macular changes from endophthalmitis prior to DSEK (n=1), and retinal branch vein occlusion (n=1), was typical of the patient demographics. Excluding the 12 eyes with known retinal problems, mean BSCVA improved from 0.52 ±0.92 (Snellen equivalent 20/60) before DSEK to 0.29±0.45 (Snellen equivalent 20/40) at the 6-month examination (Fig 6).
Figure 1. Manifest refractive cylinder (diopters; mean ± S D) before and after DSEK. The 1-month examination was the only postoperative time point when mean cylinder differed significantly from that of the preoperative examination (P=. 039).
Figure 2. Manifest spherical equivalent refraction (diopters; mean±SD) before and after DSEK. Differences between examinations were not statistically significant (P=. 11).
Figure 3. Scatterplot of pre- and postoperative manifest refractive cylinder for the individual eyes (n=48; accurate preoperative refractions were not available for 2 eyes). The diagonal line is a reference line showing where points would lie with no difference between pre- and postoperative cylinder.
Figure 4. Scatterplot of pre- and postoperative spherical equivalent refraction for the individual eyes (n=48; accurate preoperative refractions were not available for 2 eyes).
NORMAL POSTOPERATIVE COURSE
The donor tissue typically adhered to the posterior surface of the recipient cornea without any fluid or space between the two. Variable amounts of corneal edema would persist for a few days to weeks as the new tissue eliminated the excess corneal hydration. Thus for the first few days after surgery, the central cornea with the combined two layers of tissue was typically not as clear as seen in PK grafted tissue. The peripheral grafted tissue would also initially appear edematous along the edge where bare stroma was exposed to the anterior chamber fluid; the edge edema typically took a few weeks to clear, presumably until endothelial cells migrated over the edge to provide a barrier to fluid entry. Whenever 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 would appear edematous; these areas could take a few months to clear, depending on their size. Therefore, after the first few cases in the series, the standard practice was to remove a smaller diameter of Descemet's membrane and endothelium on the recipient than was to be covered with the donor tissue. Figure 7 shows a representative eye 6 months after DSEK. The entire central cornea is crystal clear and the edge of the donor shows no residual edema.
Figure 5. BSCVA (logMAR; mean±SD) before and after DSEK for all 50 eyes, including those with documented retinal problems. Mean BSCVA was significantly improved at the 3- and 6-month examinations compared with the preoperative value (P=. 00 70).
Figure 6. BSCVA (logMAR; mean±SD) before and after DSEK for the 38 eyes without documented retinal problems. Mean visual acuity was significantly improved at the 6-month examination compared with the preoperative value (P=. 0034).
Seven eyes were phakic at the time of surgery; one eye developed a cataract, which was successfully removed using a 2.8-mm clear corneal temporal incision for phacoemulsification, and BSCVA was 20/30 at the 6-month examination. Three cases of primary graft failure occurred; the donor buttons were replaced within 1 week of the original surgery and BSCVA was 20/30 to 20/40 at the most recent examination.
Figure 7. Slit-lamp photograph of a typical eye in this case series 6 months after DSEK. The white arrowhead points to the edge of the donor tissue illuminated by the slit-beam. The edge of the donor graft appears as a light circular line on the cornea.
One patient with long-standing bullous keratopathy had a dense anterior stromal scar stripped off at the time of surgery and subsequently had residual anterior stromal haze limiting vision to 20/200. His surgery had primarily been performed to relieve pain from the bullae. Three patients with Fuchs' dystrophy had superficial keratectomies performed postoperatively to treat anterior corneal dystrophy that was apparent once the dysfunctional endothelium had been replaced allowing the epithelial edema to clear. These three patients had BSCVA of 20/50 to 20/60 at their most recent examinations.
Comparative Visual and Refractive Outcomes With Different Keratoplasty Procedures
The following 2 cases, which were among the 50 reported, highlight the most significant advantages of DSEK compared with PK.
A 34-year-old female, who was the youngest patient we have seen with Fuchs' dystrophy, presented with edema and bullae of the central 3.5 mm of the cornea. Although she could read the 20/30 line on the Snellen acuity chart in a darkened room, BSCVA decreased to 20/200 with glare. She was unable to drive at night and had difficulty with driving during the day and performing her job as an accountant. Three months after DSEK, BSCVA was 20/25 and she reported less visual disturbance with glare. It had taken 18 months for her to recover BSCVA of 20/25 in her other eye after PK, so she was pleased with the more rapid visual recovery in the DSEK eye. Her BSCVA was 20/20 in the DSEK eye at 1-year follow-up, suggesting that visual acuity continues to improve between 6 months and 1 year after DSEK.
A 48-year-old policeman with preoperative BSCVA of 20/40 in the right eye and 20/30 in the left eye had bilateral central corneal edema and bullae. With glare his vision deteriorated to 20/200 in the right eye and 20/70 in the left eye. Within 1 month after DSEK on the right eye, BSCVA was 20/30 and he requested surgery on the left eye. In his line of work, the relatively small 5-mm DSEK scleral tunnel incision provided greater protection against traumatic rupture compared with standard PK.
Results from our initial 50 consecutive cases using the DSEK technique show that treated eyes experienced minimal refractive change (Figs 1 and 2). This is a major advantage shared by the endothelial keratoplasty techniques as compared with PK, which can have unpredictable refractive outcomes with prolonged refractive instability. We found that DSEK did not cause a statistically significant change in mean spherical equivalent refraction. The mean cylindrical component transiently increased at 1-month follow-up but returned to the preoperative level following suture removal approximately 2 months after surgery. Similarly, the mean astigmatic error after PLK is low, 1.54±0.81 D,12 and DLEK does not induce a significant change in mean corneal curvature or spherical equivalent refraction.8 In contrast, high astigmatism after PK can prevent functional success in 10% to 20% of treated eyes.2 Mean refractive cylinder of 4.0 to 5.0 D is common after PK (Table 2),1314 and the mean spherical equivalent refraction can shift by several diopters following suture removal.13 The comparatively stable and predictable refractive outcomes obtained with endothelial keratoplasty are beneficial to patients.
Visual recovery after DSEK compared favorably with that after PK or PLK/DLEK (Table 2). By 6 months postoperatively, 62% of eyes had ≥20/40 visual acuity, and 76% saw ≥20/50. In fact, the DSEK visual outcomes at 6 months were comparable to the best outcomes obtained after PK even when compared with some longer PK postoperative time points of 2 to 4 years (Table 2).13-16 Furthermore, none of the eyes required rigid contact lenses for best corrected visual acuity after DSEK, in contrast to PK, where it is common for 10% to 15% of eyes to require rigid gas-permeable contact lenses for best vision.14,15 Another notable difference is that in our DSEK series, suture-related issues were completely resolved within 2 to 3 months after surgery, whereas most eyes still have sutures in place 6 months after PK, and visual acuity can change, for better or worse, after suture removal.15
An additional significant advantage of endothelial keratoplasty, compared with PK, is that the eye remains structurally stronger. Penetrating keratoplasty wounds can rupture with minimal trauma many years after the original surgery, leading in some cases to loss of the eye from suprachoroidal hemorrhage.17 In fact, the loss of several of our patients' eyes from trauma after PK was what provided the initial impetus for us to learn the PLK technique. After performing 101 PLK procedures, we switched to the DSEK technique because it is a simpler procedure and is less traumatic to the remaining recipient cornea and the underlying iris and anterior chamber. Furthermore, PLK/DLEK requires placing together two hand-dissected surfaces, whereas with DSEK, the recipient interface is smooth, and this may help minimize potential visual limitations from interface irregularities.
The minimal induced refractive change, rapid visual recovery, and retention of corneal integrity achieved with DSEK have made it our procedure of choice for treating endothelial dysfunction, such as Fuchs' endothelial dystrophy or bullous keratopathy. Furthermore, we believe that the favorable DSEK refractive and visual outcomes, along with the advantages for patients illustrated in the two case reports, have shifted the risk-benefit ratio of performing surgery, such that we now perform DSEK earlier than we used to perform PK on patients with Fuchs' dystrophy. Transplant surgery is now undergoing a similar paradigm shift as occurred in cataract surgery when phacoemulsification replaced intracapsular and extracapsular techniques. With improved transplant procedures, surgeons and patients may no longer wait for endothelial dysfunction to become as incapacitating before proceeding to surgery.
Several reports have indicated that mean visual acuity continues to improve in DLEK eyes for up to 2 years after surgery.9,18 The gradual visual improvement in some PLK/DLEK eyes may be associated with stromal remodeling at the lamellar interface. We do not yet know whether visual acuity will likewise continue to improve over time after DSEK. As additional follow-up is obtained, complications, endothelial viability, and longer-term visual results will be reported.
1. Binder PS, Waring GO. Keratotomy for astigmatism. In: Waring GO III, ed. Refractive Keratotomy for Myopia and Astigmatism. St Louis, Mo: Mosby Year Book; 1992:1157-1186.
2. Riddle HK Jr, Parker DA, Price FW Jr. Management of postkerato plasty astigmatism. Curr Opin Ophthalmol. 1998;9:15-28.
3. Touzeau O, Borderie VM, Allouch C, Scheer S, Laroche L. Effects of penetrating keratoplasty suture removal on corneal topography and refraction. Cornea. 1999;18:638-644.
4. 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.
5. Melles GR, Eggink FA, Lander F, Pels E, Rietveld FJ, Beekhuis WH, Binder PS. A surgical technique for posterior lamellar keratoplasty. Cornea. 1998;17:618-626.
6. Melles GR, Rietveld FJ, Beekhuis WH, Binder PS. A technique to visualize corneal incision and lamellar dissection depth during surgery. Cornea. 1999;18:80-86.
7. Melles GR, Lander F, Nieuwendaal C. Sutureless, posterior lamellar keratoplasty: a case report of a modified technique. Cornea. 2002;21:325-327.
8. Terry MA, Ousley PJ. In pursuit of emmetropia: spherical equivalent refraction results with deep lamellar endothelial keratoplasty (DLEK). Cornea. 2003;22:619-626.
9. Terry MA, Ousley PJ. Rapid visual rehabilitation after endothelial transplants with deep lamellar endothelial keratoplasty (DLEK). Cornea. 2004;23:143-153.
10. Melles GR, Wijdh RH, Nieuwendaal CP. A technique to excise the descemet membrane from a recipient cornea (descemetorhexis). Cornea. 2004;23:286-288.
11. Holladay JT, Prager TC Mean visual acuity. Am J Ophthalmol. 1991;111:372-374.
12. Melles GR, Lander F, van Dooren BT, Pels E, Beekhuis WH. Preliminary clinical results of posterior lamellar keratoplasty through a sclerocorneal pocket incision. Ophthalmology. 2000;107:1850-1857.
13. Davis EA, Azar DT, Jakobs FM, Stark WJ. Refractive and keratometric results after the triple procedure: experience with early and late suture removal. Ophthalmology. 1998;105:624-630.
14. Pineros O, Cohen EJ, Rapuano CJ, Laib son PR. Long-term results after penetrating keratoplasty for Fuchs' endothelial dystrophy. Arch Ophthalmol. 1996;114:15-18.
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, Holland HK, Esterman A, Coster DJ, eds. The Australian Corneal Graft Registry 2004 Report. Bedford Park, Australia: AGCR Publications; 2004:154.
17. 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.
18. Price MO, Price FW Jr. Cataract progression and treatment following posterior lamellar keratoplasty. J Cataract Refract Surg. 2004;30:1310-1315.
Manifest Refraction (D) Before and 6 Months After DSEK in 50 Eyes (Mean±Standard Deviation [range])
Comparative Visual and Refractive Outcomes With Different Keratoplasty Procedures