Comparison of Posterior Lamellar Keratoplasty Techniques to Penetrating Keratoplasty

Comparison of Posterior Lamellar Keratoplasty Techniques to Penetrating Keratoplasty Irit Bahar, MD, Igor Kaiserman, MD, MSc, Penny McAllum, MBChB, FRANZCO, Allan Slomovic, MD, FRCSC, David Rootman, MD, FRCSC Purpose: To describe the visual outcomes and intraoperative and postoperative complications after penetrating keratoplasty (PK), deep lamellar endothelial keratoplasty (DLEK), Descemet stripping endothelial keratoplasty (DSEK), and Descemet stripping automated endothelial keratoplasty (DSAEK) and to compare the results with those of previously reported series. Design: Prospective, comparative, nonrandomized study. Participants: One hundred seventy-seven eyes of 161 consecutive patients who had corneal edema resulting from Fuchs endothelial dystrophy, pseudophakic bullous keratopathy, aphakic bullous keratopathy, failed graft or iridocorneal endothelial syndrome. Methods: All patients underwent either PK, DLEK, DSEK, or DSAEK at the Cornea Service of the Toronto Western Hospital. Main Outcome Measures: Uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), manifest refraction, corneal endothelial counts, and postoperative complications. Results: The average 12-month postoperative BSCVA was 20/53 in the PK group, 20/80 in the DLEK group, 20/56 in the DSEK group, and 20/44 in the DSAEK group. The mean spherical equivalent was similar between groups, but tended toward hyperopia in the DSEK and DSAEK groups. The average refractive astigmatism was 3.78⫾1.91 diopters (D) in the PK group and 1.61⫾1.26 D, 1.86⫾1.1 D, and 1.36⫾0.92 D in the DLEK, DSEK, and DSAEK groups, respectively (P⬍0.0001). Early postoperative donor disc dislocations occurred in 6 (8.8%) patients in the DLEK group, 2 (12.5%) in the DSEK group, and 7 (15.6%) in the DSAEK group (P ⫽ 0.0004). Detached grafts were reattached and repositioned by injecting an air bubble to press the donor against the recipient cornea. Primary graft failure occurred in 1 (2.1%) of the PK cases, 2 (2.9%) of the DLEK cases, and 1 (2.2%) of the DSAEK cases; all underwent the same procedure successfully. Average cell loss at 1 year after surgery was 40.11% and was similar in the 4 groups. Conclusions: The DSAEK surgery enabled rapid and better UCVA and BSCVA when compared with PK, DLEK, and DSEK, with significantly lower astigmatism. Endothelial cell loss was similar, but the dislocation rate was significantly higher in the DSAEK group. Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2008;115:1525–1533 © 2008 by the American Academy of Ophthalmology.

Penetrating keratoplasty (PK) has been the standard of care for treating endothelial failure for many years. However, the disadvantages of this procedure include prolonged visual rehabilitation, high astigmatism, and suture-related complications, such as broken sutures, infectious keratitis, wound dehiscence, and graft rejection, even years later.1–3 Endothelial transplantation was first described by Melles et al.4 They called this technique posterior lamellar keratoplasty (PLK). Terry and Ousley5,6 subsequently modified this technique, renaming it deep lamellar endothelial keratoplasty (DLEK). This surgery required manual lamellar dissections within the deep corneal stroma of both the recipient and the donor corneas. However, the lamellar dissections are difficult to perform and there is increased intraoperative tissue manipulation, a potential for early postoperative transplant dislocations because of the lack or sutures, and the presence of the lamellar interface, which may limit visual outcomes. © 2008 by the American Academy of Ophthalmology Published by Elsevier Inc.

The next modification of PLK was Descemet stripping endothelial keratoplasty (DSEK).7,8 In this procedure, instead of performing a lamellar dissection, the patient’s Descemet membrane is peeled off, using specially designed strippers. Compared with DLEK, DSEK is somewhat easier to perform, and stripping the Descemet membrane leaves a very smooth recipient interface onto which the donor can be applied. This may lead to better visual results, but has also been implicated as a cause of early postoperative donor dislocations. The current incarnation of PLK, described by Price and Price,9 is Descemet stripping automated endothelial keratoplasty (DSAEK). In this procedure, the donor dissection is carried out using a mechanical microkeratome, which makes a smooth dissection deep within the corneal stroma. The posterior donor button then is punched out, using standard corneal trephination techniques. DSAEK now is deISSN 0161-6420/08/$–see front matter doi:10.1016/j.ophtha.2008.02.010

1525

Ophthalmology Volume 115, Number 9, September 2008 Table 1. Comparison of Preoperative and Operative Parameters in the Various Groups: Penetrating Keratoplasty versus Deep Lamellar Penetrating Keratoplasty (n ⴝ 48)

Deep Lamellar Endothelial Keratoplasty (n ⴝ 68)

Descemet Stripping Endothelial Keratoplasty (n ⴝ 16)

Age (yrs) Gender (males) Follow-up (mos)

73.7⫾9.6 28 (58.3%) 15.8⫾11.1

76.4⫾8.3 22* (32.4%) 15.4⫾6.8

71.1⫾10 9 (56.3%) 12.9⫾3.0

Donor endothelial cell counts (cell/mm2) Donor size (mm)

2544⫾432 7.9⫾0.2

2762⫾476 7.7⫾0.5

2772⫾294 8.5⫾0.2

24 (50%) 16 (33.3%) 1 (2.1%) 5 (10.4%) 5 (10.4%) 1.27⫾0.75 (20/372) 1.39⫾0.64 (20/491)

33 (48.5%) 28 (41.2%) 0 4 (5.9%) 2 (2.9%) 1.12⫾0.74 (20/263) 1.17⫾0.74 (20/296)

11 (68.8%) 5 (31.3%) 0 0 0 0.52⫾0.11 (20/66) 1.11⫾0.71 (20/258)

Type of surgery Graft alone Graft ⫹ cataract extraction ⫹ IOL AC IOL Graft ⫹ IOL exchange or secondary IOL implantation Anterior Vitrectomy Preoperative best corrected visual acuity (logMAR) Preoperative uncorrected visual acuity (logMAR)

AC IOL ⫽ anterior chamber; DLEK ⫽ deep lamellar endothelial keratoplasty; DSAEK ⫽ Descemet stripping automated endothelial keratoplasty; DSEK ⫽ keratoplasty. *Statistical difference from the penetrating keratoplasty group (chi-square test for proportions): P⬍0.01.

scribed as the procedure of choice for corneal endothelial failure in many centers. There have been no reports, to the authors’ knowledge, comparing these 4 techniques using patients from 1 center. Ultimately, when adopting new techniques, one must compare the results with those of previously performed techniques. The purpose of the study was to determine the visual outcomes and intraoperative and postoperative complications after PK, DLEK, DSEK, and DSAEK and to compare these results with those of previously reported series.

Patients and Methods This prospective, comparative, nonrandomized study included a total of 161 consecutive patients with corneal edema secondary to Fuchs endothelial dystrophy, pseudophakic bullous keratopathy, aphakic bullous keratopathy, failed graft or iridocorneal endothelial syndrome who underwent either PK, DLEK, DSEK, or DSAEK at the Toronto Western Hospital, Toronto, Canada. The study began in February 2003 and is still ongoing. It was approved by the Institutional Research Ethics Committee at the University Health Network. Patients included in this study have reached at least 6 months of follow-up. All were operated on by 1 (D.R.) of 2 (D.R. and A.S.) surgeons. Some of the data collected for the DLEK group were published previously.10,11 These patients included the surgeon’s first DLEK, DSEK, and DSAEK cases, and thus include his learning curve. After surgery, all patients underwent a complete eye examination, including measurement of uncorrected Snellen visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), slit-lamp examination, intraocular pressure (IOP) measurement, and funduscopy. Details regarding preoperative data in all groups are given in Table 1. After surgery, UCVA was measured at 1 week and 1 month. Three-month examinations were performed until 1 year, and six-month examinations were performed thereafter. Patients underwent refraction, UCVA and BSCVA measurements, keratometry, corneal topography (NIDEK, OPD Scan II ARK 10000;

1526

Nidek Inc., Fremont, CA), and pachymetry (noncontact specular microscope; ROBO, Konan Storage System KSS 300; Konan Medical, Hyogo, Japan) at 3, 6, 12, 18, 24, 30, and 36 months after the operation.

Surgical Technique Surgery was performed using neuroleptic anesthesia with a retrobulbar and Van Lint block for akinesia and anesthesia.

Penetrating Keratoplasty In the PK group (n ⫽ 48), the donor button was cut with the Hanna punch trephine (average size, 7.9⫾0.2 mm). The donor buttons were punched from the endothelial side on curved blocks, to which they were held by suction. The recipient bed was prepared to conform closely to the shape of the donor button. A Hanna suction trephine was used to cut a partial depth, circular incision in the cornea, centered at the geometric center of the cornea. Excision of the recipient corneal button was completed with curved corneal scissors. Sixteen interrupted sutures (8 10-0 nylon sutures and 8 10-0 Vicryl sutures, Ethicon Inc., Somerville, NJ) and a single continuous 16-bite 11-0 nylon suture were placed in 25 cases, and 8 interrupted stitches and a single 16-bite continuous suture (all 10-0 nylon) were placed in 20 cases.

Deep Lamellar Endothelial Keratoplasty The DLEK technique was described previously.11 Briefly, in this group (n ⫽ 68), a 5-mm scleral incision was used. The incision was made using a diamond blade, set at 400-␮m depth. The Melles technique of dissection on an air bubble in the anterior chamber (AC) was used.12 Melles dissectors (Dutch Ophthalmic Research Corp., Rotterdam, The Netherlands) were used to separate the deep stromal plane. A trephine blade marked with gentian violet was used to outline the size of the recipient cut on the corneal epithelium. Then, the disc excision was started from the stroma with an inverted crescent knife, 180° from the wound. The disc excision was completed using Melles stromal scissors. The donor was prepared in a similar way on an

Bahar et al 䡠 Posterior Lamellar Keratoplasty versus Penetrating Keratoplasty Endothelial Keratoplasty versus Descemet’s Stripping Endothelial Keratoplasty versus Descemet Stripping Automated Endothelial Keratoplasty Descemet Stripping Automated Endothelial Keratoplasty (n ⴝ 45)

P Value (Analysis of Variance/ Chi-Square Test)

70.2⫾10.9 20 (44.4%) 9.8⫾4.3

0.02 0.03 ⬍0.001

2893⫾440 8.8⫾0.3

0.003 ⬍0.001

24 (53.3%) 20 (44.4%) 0 1 (2.2%) 0 0.9⫾0.5 (20/160) 1.01⫾0.63 (20/205)

0.53, chi-square test 0.63, chi-square test 0.44, chi-square test 0.26, chi-square test 0.08, chi-square test 0.12 0.15

Tamhane Test for All Pairwise Comparisons DLEK vs. DSAEK, P ⫽ 0.02 DSAEK vs. PK, P ⫽ 0.001; DSAEK vs. DLEK, P⬍0.001; DSAEK vs. DSEK, P ⫽ 0.001 PK vs. DSAEK, P ⫽ 0.002 PK vs. DSEK, P⬍0.001; PK vs. DSAEK, P⬍0.001; DLEK vs. DSEK, P⬍0.001; DLEK vs. DSAEK, P⬍0.001; DSEK vs. DSAEK, P ⫽ 0.001

DSEK vs. PK, P ⫽ 0.002; DSEK vs. DLEK, P ⫽ 0.001 None significant

Descemet stripping endothelial keratoplasty; IOL ⫽ intraocular lens; logMAR ⫽ logarithm of the minimum angle of resolution; PK ⫽ penetrating

artificial AC (Moria Instruments, Antony, France) and was cut on a Hanna punch block (same size for donor and recipient: average size, 7.7⫾0.5 mm). Then the donor was folded and placed on a plastic lens glide with minimal viscoelastic (Provisc; Alcon, Fort Worth, TX) and inserted into the AC. Positioning of the donor disc was performed using air and balanced saline solution (BSS), and a Sinskey hook was used to match the donor within the recipient dissection edges. After the AC was filled with air for 10 minutes, part of the air was removed and replaced with BSS. Tobradex (Alcon) drops were instilled into the eye.

Descemet Stripping Endothelial Keratoplasty In this group (n ⫽ 16), the donor disc preparation was performed as described above (DLEK procedure: average size 8.5⫾0.2 mm). The recipient epithelium was marked with the trephine used on the donor tissue to outline where to strip the Descemet membrane and to place the donor tissue. A 5-mm superior scleral tunnel incision was made. The Descemet membrane was scored in a circular pattern under the area of the epithelium marking using a reverse-Sinskey hook. The Descemet membrane and endothelium were stripped using the Sinskey hook. The Descemet membrane was spread on the anterior surface of the recipient cornea to make sure a sufficient area had been removed. Four full-thickness incisions in the midperipheral recipient were performed to help in drainage of fluids from the interface. The donor was folded with minimal viscoelastic (Provisc; Alcon) and was introduced into the AC. Positioning of the donor was performed using air and BSS as described before.

Descemet Stripping Automated Endothelial Keratoplasty In this group (n ⫽ 45), the recipient Descemet stripping was performed as described above (DSEK procedure), but using Descemet strippers (Moria, Antony, France). The donor disc dissection was performed with a microkeratome using the Moria ALTK microkeratome equipped with a 300- or 350-␮m head and associated artificial AC (Moria USA, Doylestown, PA). The 350-␮m head was used in cases where the precut donor pachymetry ex-

ceeded 550 ␮m after the epithelium was removed. After dissection, the donor tissue was transferred to a punching system and was cut with an 8-, 8.5-, or 9-mm diameter trephine (average size, 8.8⫾0.3 mm). A small amount of viscoelastic (Provisc) was placed on the endothelial surface, the posterior portion was folded over on itself, endothelial side inward, and inserted to the AC using the Price forceps (Moria; n ⫽ 12). In 22 cases, an anchoring 10-0 Prolene stitch (Ethicon Inc.) on a long straight needle was placed on the donor disc at the 6-o’clock position. After the suture was passed, it was tied to create a loop. The diameter of the suture loop was kept large enough (4 –5 mm) to facilitate its cutting after insertion. The straight needle was bent slightly and was passed through the 3- to 5-mm wide limbal incision exiting at the 6-o’clock limbus, then the folded disc was inserted to the chamber either by pulling the stitch and opening slightly the 5-mm scleral incision (n ⫽ 8) or with the Goosey forceps (Moria, Antony, France) (n ⫽ 22). In the 12 last cases, after the microkeratome pass, the anterior stromal cap was removed and the donor button mounted on the artificial AC was brought under view of the operating microscope. Using a curved Vannas scissors and a 0.12-colibri forceps (Katena Products Inc., Denville, NJ), a circumferential rim of peripheral stromal tissue that protruded outside the cut margin was excised. Then, the donor tissue was transferred to the punching system as described above (unpublished data).13 This technique is shown in the online Video (available at http://aaojournal.org). After surgery, all patients underwent pressure patching overnight. Beginning the next morning, 0.1% dexamethasone sodium phosphate and moxifloxacin or TobraDex (Alcon) eye drops were administered 4 times daily for 1 week. The antibiotic drops were discontinued 1 week after the surgery, and dexamethasone eyedrops were tapered for 12 months in all groups. Topical steroid drops were used 4 times daily in the first month after surgery, then tapered to 3 times daily in the second month, twice daily in the third month, and once daily from the fourth month onward.

Statistical Analysis The 4 groups were compared using a 1-way analysis of variance for continuous variables (the Tamhane post hoc test was used

1527

Ophthalmology Volume 115, Number 9, September 2008 Table 2. Preoperative Clinical Data Penetrating Keratoplasty (n ⴝ 48) Indication for surgery PBK Fuchs dystrophy ABK ICE syndrome Failed graft Other preoperative disease Glaucoma Cataract AMD RD Macular hole Diabetic retinopathy Amblyopia

Deep Lamellar Endothelial Keratoplasty (n ⴝ 68)

16 (33.3%) 17 (35.4%) 3 (6.3%) 1 (2.1%) 11 (22.9%)

37 (54.4%)* 28 (41.2%) 1 (1.5%)

11 (22.9%) 16 (33.3%) 3 (6.3%) 2 (4.2%) 1 (2.1%) 1 (2.1%) 1 (2.1%)

13 (19.1%) 26 (38.2%) 3 (4.4%)

Descemet Stripping Endothelial Keratoplasty (n ⴝ 16) 8 (50%) 8 (50%)

Descemet Stripping Automated Endothelial Keratoplasty (n ⴝ 45) 12 (26.6%) 28 (62.2%)‡ 2 (4.4%) 3 (6.7%)

2 (2.9%)†

5 (31.3%) 1 (6.3%)

3 (6.7%)* 20 (44.4%) 7 (15.6%) 1 (2.2%) 1 (2.2%)

1 (1.5%)

P Value (Chi-Square Test) 0.08 0.003 0.16 0.31 0.003 0.04 0.67 0.17 0.35 0.43 0.60 0.76

ABK ⫽ aphakic bullous keratopathy; AMD ⫽ age-related macular degeneration; ICE ⫽ iridocorneoendothelial; PBK ⫽ pseudophakic bullous keratopathy; RD ⫽ retinal detachment. Symbols signify statistical difference from the penetrating keratoplasty group (chi-square test for proportions): *P⬍0.05; †P⬍0.01; ‡P⬍0.001.

thereafter for pairwise comparison) and the chi-square test for proportions. When the various PLK groups were compared individually with the PK group, the unpaired Student t test was used for continuous variables and the chi-square test was used for proportions. A P value ⬍0.05 was considered statistically significant. The computer statistics package SPSS (SPSS, Inc., Chicago, IL) was used for statistical analyses.

Results Patient Data This study included a total of 177 eyes of 161 patients aged 42 to 91 years. Mean follow-up for the entire group was 13.6⫾6.5 months. Table 1 and 2 summarize the preoperative characteristics and operative data of the patients in this study. Because of a significant difference in follow-up time among the 4 groups, data analysis of visual outcomes, endothelial cell counts, and complication rates were performed at 12 months of follow-up. In the DSAEK group, only visual outcome analysis was performed for patients with at least 6 months of follow-up (rather than 12

months) because of stability of vision and refraction in this subgroup of patients early on the postoperative course.

Visual and Refractive Results Table 3 demonstrates the postoperative BSCVA and UCVA in the 4 groups after 12 months of follow-up (except for the DSAEK group, in which patients with more than 6 months of follow-up were included because of early stability of refraction). The BSCVA was 20/53 in the PK group, with DSAEK still significantly better than PK (mean BCVA, 20/44; P ⫽ 0.001), DLEK (20/80; P ⫽ 0.001), and DSEK (20/56; not statistically significant [NS]; Fig 1). The worst UCVA was in the PK group (20/112). In the PLK techniques, there seems to be a progressive improvement in vision from DLEK (20/96; NS), to DSEK (20/89; NS), to DSAEK (20/71; P ⫽ 0.05; Table 3). Figure 2 and Table 3 show the mean spherical equivalent and manifest astigmatism in the groups studied after 12 months of follow-up (in PK patients after 12 months and removal of all interrupted sutures). The mean spherical equivalent did not differ significantly between groups, although it tended toward hyperopia Table 3. Comparison of 1-Year Postoperative

Endothelial cell counts Percent of endothelial loss Best-corrected visual acuity (logMAR) Uncorrected visual acuity (logMAR) Best-corrected visual acuity (logMAR), no retinal disease Uncorrected visual acuity (logMAR), no retinal disease Manifest spherical equivalent Manifest cylinder

Penetrating Keratoplasty (n ⴝ 48)

Deep Lamellar Endothelial Keratoplasty (n ⴝ 68)

1518⫾662 39.6⫾26.3% 0.42⫾0.14 (20/53) 0.75⫾0.35 (20/112) 0.41⫾0.14 (20/51; n ⫽ 43) 0.74⫾0.37 (20/110; n ⫽ 43) ⫺0.14⫾4.14 3.78⫾1.91

1601⫾687 43.4%⫾22.2 0.60⫾0.33 (20/80) 0.68⫾0.32 (20/96) 0.52⫾0.31 (20/60; n ⫽ 46) 0.59⫾0.28 (20/78; n ⫽ 46) ⫺0.07⫾1.40 1.61⫾1.26

DLEK ⫽ deep lamellar endothelial keratoplasty; DSAEK⫽ Descemet stripping automated endothelial keratoplasty; DSEK ⫽ Descemet stripping

1528

Bahar et al 䡠 Posterior Lamellar Keratoplasty versus Penetrating Keratoplasty

Figure 1. Graph showing best-corrected visual acuity (logarithm of the minimum angle of resolution [logMAR] units) in the penetrating keratoplasty (PK), deep lamellar endothelial keratoplasty (DLEK), Descemet stripping endothelial keratoplasty (DSEK), and Descemet stripping automated endothelial keratoplasty (DSAEK) groups over time. Asterisks signify statistical difference from the PK group (chi-square test for proportions): *P⬍0.05; **P⬍0.01.

in the DSEK and DSAEK groups. The refractive astigmatism was significantly higher in the PK group compared with all of the PLK groups: refractive astigmatism was 3.78⫾1.91 diopters (D) in the PK group and 1.61⫾1.26 D, 1.86⫾1.1 D, and 1.36⫾0.92 D in the DLEK, DSEK, and DSAEK groups, respectively (P⬍0.0001).

Intraoperative and Postoperative Complications Table 4 summarizes the intraoperative and postoperative complications in the 4 groups. There were few major complications in all groups. Early postoperative donor disc dislocations occurred in 6 (8.8%) patients in the DLEK group, 2 (12.5%) in the DSEK group, and 7 (15.6%) in the DSAEK group, all of which were managed successfully with rebubbling, except for 1 DLEK patient, who was diagnosed as a primary failure in this group (1 of the 2 [2.9%] primary failures in this group; one had successful repeat DLEK and the other underwent PK). Primary graft failure also occurred in 1 (2.1%) of the PK cases, 2 (2.9%) of the DLEK cases, and 1 (2.2%) of the DSAEK cases; all underwent the same procedure successfully.

Acute rejection was uncommon in all groups: 2 (4.2%) in the PK, 3 (4.4%) in the DLEK group, and 1 (2.2%) in the DSAEK group. Graft dehiscence or postoperative leakage occurred in 3 (6.3%) cases in the PK group only. Persistent epithelial defect occurred in 4 (8.3%) patients in the PK group only. Cystoid macular edema was diagnosed in 1 patient in each group. Interface opacity occurred in the PLK groups only: 3 (4.4%) in the DLEK group, 1 (6.3%) in the DSEK group, and 2 (4.4%) in the DSAEK group. This opacity affected the visual outcomes in these eyes (BSCVA range, 20/80 –20/50) and was located in the deep stroma. Steroid-induced glaucoma appeared in 3 patients (6.3%) in the PK group, 9 (13.2% ) in the DLEK group, 2 (12.5%) in the DSEK group, and 3 (6.7%) in the DSAEK group. Figure 3 and Table 3 show the endothelial cell loss in various groups. There were no significant differences between the 4 groups.

Discussion The DLEK, DSEK, and DSAEK procedures represent posterior lamellar corneal transplantation that allow for the

Parameters in the Various Groups Descemet Stripping Endothelial Keratoplasty (n ⴝ 16)

Descemet Stripping Automated Endothelial Keratoplasty (n ⴝ 45)

P Value (Analysis of Variance)

Tamhane Test for All Pairwise Comparisons

1726⫾671 38.2⫾22.0% 0.45⫾0.22 (20/56) 0.65⫾0.43 (20/89) 0.38⫾0.2 (20/48; n ⫽ 13) 0.61⫾0.50 (20/81; n ⫽ 13) 0.41⫾1.33 1.86⫾1.1

1735⫾416 36.4⫾15.2% 0.34⫾0.17 (20/44) 0.55⫾0.21 (20/71) 0.30⫾0.36 (20/40; n ⫽ 41) 0.54⫾0.35 (20/69; n ⫽ 41) 0.96⫾1.86 1.36⫾0.92

0.70 0.70 0.001 0.05 0.2 0.45 0.11 ⬍0.0001

None significant None significant DLEK vs. DSAEK, P ⫽ 0.001; DLEK vs. PK, P ⫽ 0.001 PK vs. DSAEK, P ⫽ 0.05 None significant None significant DLEK vs. DSAEK, P ⫽ 0.03 DLEK, DSEK, DSAEK vs. PK, P⬍0.0001

endothelial keratoplasty; logMAR ⫽ logarithm of the minimum angle of resolution; PK ⫽ penetrating keratoplasty.

1529

Ophthalmology Volume 115, Number 9, September 2008

Figure 2. Bar graph showing mean manifest astigmatism in the penetrating keratoplasty (PK), deep lamellar endothelial keratoplasty (DLEK), Descemet stripping endothelial keratoplasty (DSEK), and Descemet stripping automated endothelial keratoplasty (DSAEK) groups after 6, 12, 18, and 24 months of follow-up. Asterisks signify statistical difference from the PK group (chi-square test for proportions): **P⬍0.01. SE ⫽ spherical equivalent.

selective replacement of diseased recipient endothelium. Descemet stripping automated endothelial keratoplasty differs from DLEK and DSEK in avoiding manual lamellar dissection of both donor and host corneas, and thus intuitively should lead to improved visual outcomes.6 Indeed, a trend to progressive improvement in visual outcomes from DLEK to DSEK to DSAEK and a significant improvement from DLEK to DSAEK are demonstrated herein.

Visual and Refractive Outcomes Refraction. The mean spherical equivalent did not differ significantly between groups, although it tended toward

hyperopia in the DSEK and DSAEK groups. The manifest cylinder was significantly lower in the posterior lamellar corneal transplantation groups versus the PK group, another significant advantage that enables better and faster rehabilitation of vision after corneal transplantation. Stability of the refraction is a major advantage of all endothelial keratoplasty techniques as compared with PK. Penetrating keratoplasty usually results in unstable refraction as long as sutures are present and even after their removal.3,14 –16 In many patients in the posterior lamellar corneal transplantation group, and especially in the DSAEK group (n ⫽ 8; 17.7%), the postoperative refraction was stable by as early as 1 month. However, the PK patients’

Table 4. Intraoperative and Postoperative Complications in the 4 Groups

Disc dislocation Graft leak/dehiscence Persistent epithelial defect (⬎1 mo) Induced glaucoma Acute rejection Primary failure Interface opacity CME

Penetrating Keratoplasty (n ⴝ 48)

Deep Lamellar Endothelial Keratoplasty (n ⴝ 68)

Descemet Stripping Endothelial Keratoplasty (n ⴝ 16)

Descemet Stripping Automated Endothelial Keratoplasty (n ⴝ 45)

P Value (Chi-Square Test)

0 3 (6.3%) 4 (8.3%) 3 (6.3%) 2 (4.2%) 1 (2.1%) 0 1 (2.1%)

6 (8.8%) 0 0 9 (13.2%) 3 (4.4%) 2 (2.9%) 3 (4.4%) 1 (1.5%)

2 (12.5%) 0 0 2 (12.5%) 0 0 1 (6.3%) 1 (6.3%)

7 (15.6%)* 0 0 3 (6.7%) 1 (2.2%) 1 (2.2%) 2 (4.4%) 1 (2.2%)

0.0004 0.04 0.01 0.52 0.78 0.91 0.48 0.71

CME ⫽ cystoid macular edema. Asterisks signify statistical difference from the penetrating keratoplasty group (chi-square test for proportions): *P⬍0.01.

1530

Bahar et al 䡠 Posterior Lamellar Keratoplasty versus Penetrating Keratoplasty

Figure 3. Endothelial cell survival curves of patients undergoing penetrating keratoplasty (PK), deep lamellar endothelial keratoplasty (DLEK), Descemet stripping endothelial keratoplasty (DSEK), and Descemet stripping automated endothelial keratoplasty (DSAEK). Values are represented as a proportion of the preoperative graft cell counts. SE ⫽ spherical equivalent.

vision usually stabilizes after 1 year (after suture removal and stabilization of refraction). Price and Price17 showed recently that visual recovery is accelerated by use of microkeratome-dissected, rather than hand-dissected, donor tissue. They attributed this finding to the smoother surface created by the microkeratome versus the hand dissection, which may help to reduce interface irregularities. The current results support these findings. Best Spectacle-Corrected Visual Acuity. In this study, the average BSCVA 3 months after DSAEK was 20/40, similar to 3-month data reported by others after DSAEK and DLEK (in which both donor and host lamellar corneal dissections were performed manually).6,18 –22 The current DLEK group, however, demonstrated worse visual outcomes (20/80). This could be attributed to a relatively high proportion of other vision-limiting diseases in this subgroup of patients (n ⫽ 22). It is possible that with the beginning of change of surgical technique from PK to DLEK, the surgeon chose the first cases to be those with limited visual potential, which could contribute to a selection bias in this study. In this series, other factors contributing to postoperative BSCVA worse than 20/40 included AMD (n ⫽ 14), epiretinal membrane (n ⫽ 1), end-stage glaucoma (n ⫽ 13), retinal detachment (n ⫽ 2), diabetic retinopathy (n ⫽ 2), macular hole (n ⫽ 1), and central retinal vein occlusion (n ⫽ 1). When eliminating those cases from this series, an improvement was found in average UCVA and BSCVA, especially in the DLEK group, to 20/78 and 20/60, respectively. The greatest remaining reservations of PK over PLK concern interface opacity. Scarring and irregularity of the interface between graft and host have not been reported as

major complications with the new techniques of PLK (i.e., DSAEK). Interface opacity occurred in 4.4% to 6.3% of cases in this series. This may be the result of the smooth donor surface obtained by dissecting it with the microkeratome (DSAEK) and the smooth recipient surface, created by stripping the Descemet membrane. The influence of this interface on contrast sensitivity and high-order ocular aberration is yet to be determined. Nevertheless, the haze seems to dissipate with time, even when present early on. Another important observation in the current study is the significantly better visual outcomes (UCVA and BSCVA) of all the PLK techniques (DLEK, DSEK, and DSAEK) in comparison with PK. Table 1 shows that preoperative UCVA and BSCVA were similar between the 4 groups (except for better preoperative BSCVA in the DSEK group only). Thus, it can be assumed that the better vision in the PLK group is not a result of any preoperative bias, but rather reflects a real difference that can be attributed to the avoidance of irregular or high levels of postoperative astigmatism. Another possible explanation may be that PK and its complications more likely compromise the vision (i.e., persistent epithelial defects, induced glaucoma as a consequence of long-term treatment with steroids).

Complication Rate Endothelial Cell Loss. One year after surgery, endothelial cell loss was similar in the 4 groups (an average 40.1% cell loss from preoperative donor counts). These counts could be affected by the donor preoperative counts and size, the incision size, and the insertion technique. Terry and Ousley6 reported an average cell loss of 25% at 6 months after

1531

Ophthalmology Volume 115, Number 9, September 2008 surgery in small-incision DLEK surgery. Although their numbers are better, those data were obtained at 6 months after surgery, compared with 12 months in this study. The technique of folding and insertion are similar in all the current groups, except the PK group, although 22 patients in the DSAEK group had undergone the surgery using a stitch, with or without the forceps assistance. Donor size in the current DLEK group is smaller than Terry’s 8.5-mm donors and the 9-mm donors in the present DSAEK group. The bigger donor increases the number of transplanted endothelial cells, which may improve long-term endothelial cell density. The current PK cell loss results are comparable with those reported by Ing et al23 12 months after surgery. Dislocation Rate. Dislocation of the donor corneal disc was the most common perioperative complication and occurred in 6 (8.8%) patients in the DLEK group, 2 (12.5%) in the DSEK group, and 7 (15.6%) in the DSAEK group. Separation or displacement of the donor lenticula was treated by repositioning the graft with a 30-gauge bent needle and air bubble tamponade. It was performed successfully as late as 1 week after surgery. The current DLEK, DSEK, and DSAEK donor dislocation rate is higher than that reported by Terry and Ousley24 (4%) for their DLEK procedure, but is consistent with the reported dislocation rates in other studies (varying from 14% to 25%).9 This rate has been shown to improve as the surgeon gains more experience with this procedure. The current DLEK dislocation rate was lower than the that of the DSAEK group. Several differences in the surgical technique may explain this difference. Whereas DLEK has 2 manually dissected rough surfaces in opposition, DSEK has only 1, and DSAEK has 2 very smooth surfaces, without a recipient pocket. This may interfere with the donor adhesion. Endothelial pump function has an important role in graft adhesion. Thus, all donor corneas in this study were warmed before the implantation to overcome the endothelial shock induced by the cold refrigerated environment. Nevertheless, endothelial damage may be induced by the folding itself. Unknown stromal factors also may have an important role in graft sticking. This needs to be investigated further. Terry et al25 recently published their laboratory study on cadaver eyes and showed that the high rate of dislocation of the donor disc in DSEK may be caused by the absence of recipient stromal fibrils to bind to the donor stromal fibrils initially. They suggested that selectively scraping the peripheral recipient bed to promote donor edge adhesion (while leaving the central bed untouched for vision) may aid in prevention of donor dislocation after DSEK. Another factor, more relevant in the triple procedures, is retained viscoelastic substance in the graft– host interface. The surgeon should be aware of this factor during the operation and should be meticulous in aspirating the viscoelastic before inserting the lenticula into the AC. In the last 12 DSAEK cases, BSS was used rather than Optisol in the artificial AC, and the thick peripheral stroma of the donor disc was trimmed before insertion. The authors believe that this thick stromal rim on the donor disc is induced during microkeratome preparation of a 9-mm donor disc. This thick edge can lead to incomplete adherence of

1532

the donor lenticula to the host stromal bed, thus allowing fluid to seep into the interface, leading to decentration and disc dislocation. Some surgeons advocate the use of BSS rather then Optisol in the preparation of the DSAEK lenticula, claiming that the Optisol solution contains dextran, which causes imbibitions of fluids into the cornea and may be associated with disc detachment and dislocation. Rejection Rate. A total of 6 patients in this series experienced an acute endothelial graft rejection within 12 months of surgery: 2 (4.4%) in the PK group, 3 (4.4%) in the DLEK group, and 1 (2.2%) in the DSAEK group. In all cases, the rejection episode was treated successfully. The rejection rate reported here is similar to and even lower than that reported in the literature for PK (9%–29%; n ⫽ 3, 27), DLEK (4%; n ⫽ 22), and DSAEK (7.5%–9% at 6 months).21,26 In the PLK group, large-diameter, excessively thick, or eccentric donor corneal discs may increase the risk of graft rejection because of the proximity between the donor endothelium and host uveal tissue and limbus. Longer-term studies are needed to determine whether the increased donor size in the DSAEK group increases the rate of rejection. However, the risk for epithelial and stromal rejection should be lower in the PLK group. The absence of graft sutures in the PLK patients also may reduce the risk of rejection and other complications like wound dehiscence and infectious keratitis.26 –29 Recently, Allan et al30 examined the corneal transplant rejection rate and the rate of graft failure subsequent to rejection in the first 2 years after endothelial keratoplasty (EK) and compared this with similar cases of PK. They reported a rejection rate of 7.5 % in the EK group and 13% in the PK group over 2 years of follow-up. A lower rejection rate is reported herein, but at a follow-up time of 1 year. They concluded that rejection is less frequent and may be less severe after EK than after PK. However, they stated that it is not yet clear whether these differences are simply a product of relatively prolonged postoperative topical steroid cover in EK patients. Other Complications. No patient in this series experienced postoperative angle-closure glaucoma because of a pupillary block from the residual intraoperative air bubble. Steroid-induced glaucoma occurred in 6.3% to 13.2% of the patients. This compares well to the known incidence of up to 30% of the patients receiving topical steroid therapy.31,32 To the best of the authors’ knowledge, and after reviewing the literature, this is the first prospective study that compares the outcomes of the 4 techniques for endothelial keratoplasty. The limitations of this study include relatively small sample size, relatively short follow-up, the nonrandomized nature of this study, and possible selection bias. In conclusion, DSAEK surgery allows rapid and better UCVA and BCVA recovery when compared with that of PK, DLEK, and DSEK, with significantly lower astigmatism. Endothelial cell loss was similar, but the dislocation rate was significantly higher in the DSAEK group. This should be investigated further to minimize the chances of a second penetrating procedure to the eye.

Bahar et al 䡠 Posterior Lamellar Keratoplasty versus Penetrating Keratoplasty

References 17. 1. Maeno A, Naor J, Lee HM, et al. Three decades of corneal transplantation: indications and patient characteristics. Cornea 2000;19:7–11. 2. Mamalis N, Anderson CW, Kreisler KR, et al. Changing trends in the indications for penetrating keratoplasty. Arch Ophthalmol 1992;110:1409 –11. 3. Pineros O, Cohen EJ, Rapuano CJ, Laibson PR. Long-term results after penetrating keratoplasty for Fuchs’ endothelial dystrophy. Arch Ophthalmol 1996;114:15– 8. 4. Melles GR, Eggink FA, Lander F, et al. A surgical technique for posterior lamellar keratoplasty. Cornea 1998;17:618 –26. 5. 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. 6. Terry MA, Ousley PJ. Deep lamellar endothelial keratoplasty: visual acuity astigmatism, and endothelial survival in a large prospective series. Ophthalmology 2005;112:1541– 8. 7. Melles GR, Wijdh RH, Nieuwendaal CP. A technique to excise the Descemet membrane from a recipient cornea (descemetorhexis). Cornea 2004;23:286 – 8. 8. Price FW Jr, Price MO. Descemet’s stripping with endothelial keratoplasty in 50 eyes: a refractive neutral corneal transplant. J Refract Surg 2005;21:339 – 45. 9. Price FW Jr, Price MO. Descemet’s stripping with endothelial keratoplasty in 200 eyes: early challenges and techniques to enhance donor adherence. J Cataract Refract Surg 2006;32: 411– 8. 10. Hyams M, Segev F, Yepes N, et al. Early postoperative complications of deep lamellar endothelial keratoplasty. Cornea 2007;26:650 –3. 11. Yepes N, Segev F, Hyams M, et al. Five-millimeter-incision deep lamellar endothelial keratoplasty: one-year results. Cornea 2007;26:530 –3. 12. Melles GR, Lander F, Beekhuis WH, et al. Posterior lamellar keratoplasty for a case of pseudophakic bullous keratopathy. Am J Ophthalmol 1999;127:340 –1. 13. Srinivasan S, McAllum P, Bahar I, et al. Stromal trimming of the donor disc: a technique to promote graft adherence and centration following Descemet’s stripping automated endothelial keratoplasty. Br J Ophthalmol. In press. 14. Riddle HK Jr, Parker DA, Price FW Jr. Management of postkeratoplasty astigmatism. Curr Opin Ophthalmol 1998;9: 15–28. 15. Akova YA, Onat M, Koc F, et al. Microbial keratitis following penetrating keratoplasty. Ophthalmic Surg Lasers 1999;30: 449 –55. 16. Davis EA, Azar DT, Jakobs FM, Stark WJ. Refractive and

18. 19. 20. 21.

22. 23. 24. 25.

26. 27. 28.

29. 30. 31. 32.

keratometric results after the triple procedure: experience with early and late suture removal. Ophthalmology 1998;105:624 –30. Price MO, Price FW Jr. Descemet’s stripping with endothelial keratoplasty: comparative outcomes with microkeratomedissected and manually dissected donor tissue. Ophthalmology 2006;113:1936 – 42. Fogla R, Padmanabhan P. Initial results of small incision deep lamellar endothelial keratoplasty (DLEK). Am J Ophthalmol 2006;141:346 –51. Terry MA, Ousley PJ. Rapid visual rehabilitation after endothelial transplants with deep lamellar endothelial keratoplasty (DLEK). Cornea 2004;23:143–53. Koenig SB, Covert DJ. Early results of small-incision Descemet’s stripping and automated endothelial keratoplasty. Ophthalmology 2007;114:221– 6. Koenig SB, Covert DJ, Dupps WJ Jr, Meisler DM. Visual acuity, refractive error, and endothelial cell density six months after Descemet stripping and automated endothelial keratoplasty (DSAEK). Cornea 2007;26:670 – 4. Gorovoy MS. Descemet-stripping automated endothelial keratoplasty. Cornea 2006;25:886 –9. Ing JJ, Ing HH, Nelson LR, et al. Ten-year postoperative results of penetrating keratoplasty. Ophthalmology 1998;105: 1855– 65. Terry MA, Ousley PJ. Deep lamellar endothelial keratoplasty: early complications and their management. Cornea 2006;25: 37– 43. Terry MA, Hoar KL, Wall J, Ousley P. Histology of dislocations in endothelial keratoplasty (DSEK and DLEK): a laboratorybased, surgical solution to dislocation in 100 consecutive DSEK cases. Cornea 2006;25:926 –32. Jonas JB, Rank RM, Budde WM. Immunologic graft reactions after allogenic penetrating keratoplasty. Am J Ophthalmol 2002;133:437– 43. Koenig SB, Schultz RO. Penetrating keratoplasty for pseudophakic bullous keratopathy after extracapsular cataract extraction. Am J Ophthalmol 1988;105:348 –53. van der Schaft TL, van Rij G, Renardel de Lavalette JG, Beekhuis WH. Results of penetrating keratoplasty for pseudophakic bullous keratopathy with the exchange of an intraocular lens. Br J Ophthalmol 1989;73:704 – 8. Thompson RW Jr, Price MO, Bowers PJ, Price FW Jr. Longterm graft survival after penetrating keratoplasty. Ophthalmology 2003;110:1396 – 402. Allan BD, Terry MA, Price FW Jr, et al. Corneal transplant rejection rate and severity after endothelial keratoplasty. Cornea 2007;26:1039 – 42. Jones R III, Rhee DJ. Corticosteroid-induced ocular hypertension and glaucoma: a brief review and update of the literature. Curr Opin Ophthalmol 2006;17:163–7. Kersey JP, Broadway DC. Corticosteroid-induced glaucoma: a review of the literature. Eye 2006;20:407–16.

Footnotes and Financial Disclosures Originally received: October 24, 2007. Final revision: February 9, 2008. Accepted: February 12, 2008. Available online: April 28, 2008.

Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Manuscript no. 2007-1386.

From the Department of Ophthalmology, Toronto Western Hospital, University of Toronto, Toronto, Canada.

Correspondence: Irit Bahar, MD, Department of Ophthalmology, Toronto Western Hospital, 399 Bathurst Street, Toronto, Ontario, Canada M5T 2S8. E-mail: [email protected].

1533