To evaluate the long-term endothelial cell density (ECD) changes and visual and refractive outcomes after deep anterior lamellar keratoplasty (DALK) using the big-bubble technique in eyes with keratoconus.
Retrospective interventional case series study.
setting: Single hospital. patients: Two hundred forty-one eyes of 214 patients who underwent DALK for moderate to advanced keratoconus. main outcome measures: Intraoperative and postoperative complications, postoperative uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), spherical equivalent refraction, topographic astigmatism, and ECD.
Two hundred thirty-four procedures (97%) were completed as DALK. Big bubble was successfully achieved in 193 eyes (82.4%). In 7 eyes (3%), the procedure was converted to penetrating keratoplasty because of Descemet membrane macroperforations. Microperforations occurred in 18 eyes (7.5%). The mean follow-up time was 50.5 ± 22.2 months (range 24 to 96 months). UCVA was lower than 20/100 in all eyes preoperatively and better than 20/100 in 191 eyes (81.6%) postoperatively. BSCVA was 20/40 or better in 187 eyes (79.9%) and 20/20 or better in 38 eyes (16.2%). The mean (± SD) preoperative ECD that was possible in 166 eyes was 2797 ± 561 cells/mm 2 . Mean (± SD) endothelial cell loss was 8.1% ± 4.6% at 1 year, 10.5% ± 5.7% at 2 years, 15.1% ± 14.8% at 6 years, and 22.5% ± 15.9% at 8 years. Stromal graft rejection episodes occurred in 4 eyes, which resolved with appropriate therapy.
Deep anterior lamellar keratoplasty that uses the big-bubble technique is effective in patients with keratoconus. Long-term endothelial cell loss was moderate and lower than penetrating keratoplasty grafts.
Although penetrating keratoplasty (PK) is still considered the gold standard for the treatment of advanced keratoconus (KCN), deep anterior lamellar keratoplasty (DALK) has emerged as an alternative treatment in the last decade. The main advantage of DALK is that the patient’s own endothelium is retained, which eliminates the risk of endothelial graft rejection. Because the integrity of the Descemet membrane is not violated, a tectonically stronger corneal wound is achieved, which allows earlier suture removal than PK. Also, since topical steroids can be discontinued earlier, there is less risk of secondary infection, delayed wound healing, cataract development, or glaucoma.
Despite the inherent advantages of DALK, the removal of the host stroma in DALK surgery is a technically difficult and time-consuming procedure. Apart from direct surgical excision of the stroma using layer-by-layer dissection, several other techniques, including injection of air, saline solution, or viscoelastic material into corneal stroma, have been proposed to ease the separation of the posterior stroma from the Descemet membrane. Big-bubble technique, proposed by Anwar and Teichmann, involves injection of air into the deep stroma following partial trephination to allow a safer dissection of posterior stroma from the Descemet membrane by creating a big bubble between Descemet membrane and posterior stroma. It was reported to be a safe and effective technique for DALK surgery in KCN patients.
Although DALK and PK have been found to provide similar visual and refractive outcomes, postoperative endothelial cell counts have been reported to be better after anterior lamellar keratoplasty in comparative studies. However, studies reporting endothelial cell density (ECD) after DALK surgery are limited and focused mainly on short-term data. In the current study, we present our long-term DALK results in a large series of eyes with KCN. To our knowledge, this is one of the largest series of DALK cases using big-bubble technique in KCN and first to present the long-term changes of ECD after DALK surgery in KCN patients.
In this noncomparative interventional case series study, clinical data of 241 eyes of 214 KCN patients undergoing DALK surgery at Kartal Training and Research Hospital between April 1, 2002 and May 31, 2008 were reviewed. Data were recorded prospectively and reviewed retrospectively. Patients who were lost to follow-up were excluded from the data analysis. DALK was performed using big-bubble technique and the eyes that were followed up at least 24 months were included.
KCN diagnosis was based on the clinical findings of slit-lamp examination (corneal thinning, conical protrusion of the apical cornea, Fleisher ring, Vogt striae, epithelial and subepithelial scarring), keratometric measurements, and specific corneal topographic patterns. Only patients unsatisfied with spectacle- and contact lens–corrected vision were considered for DALK surgery. Patients with other concomitant diseases that could affect vision, with a history of previous intraocular surgery or acute hydrops, were not included in the study. Patients who required ocular surgery simultaneously with DALK surgery were excluded.
All surgical procedures were performed under retrobulbar or general anesthesia using big-bubble technique as described. The size of trephination was determined according to the horizontal corneal diameter and the location of the cone. A Barron suction trephine (range, 7.00 to 8.50 mm) was used for partial-thickness trephination of up to 60% to 80% corneal depth. A 30-gauge disposable needle bent at an angle of 60 degrees at 5 mm distance from its tip with the sharp side facing upwards was inserted 3 to 4 mm into the deep stroma, starting from the trephination groove and avoiding entry into the anterior chamber while maintaining the depth. The air was injected in a gradual manner into the deep stroma, aiming for an air-induced separation of Descemet membrane from the stroma. Complete separation of the Descemet membrane was noted as a large air bubble reaching the trephination line.
Once the big bubble was achieved, a peripheral paracentesis was performed to lower intraocular pressure. A crescent knife was used for a superficial corneal keratectomy, leaving a thin layer of corneal stroma in place anterior to the bubble. A 45-degree slit knife was inserted into the big bubble to collapse it. Blunt Vanna scissors were used to divide the stromal remnant tissue into 4 quadrants, which were excised for baring of Descemet membrane at the edge of the trephination.
In all procedures, the surgeon always aimed to achieve a big bubble to expose Descemet membrane (descemetic DALK; dDALK). In cases when a big bubble could not be generated with repeated attempts, a layer-by-layer manual stromal dissection was performed with a blunt-tipped wire spatula (predescemetic DALK; pdDALK). When Descemet membrane perforations occurred during surgery, air was injected into the anterior chamber to seal microperforations. If Descemet membrane perforation was large enough to preclude lamellar keratoplasty, the surgeon converted to PK.
Full-thickness corneoscleral donor buttons stored in Optisol GS (Bausch & Lomb, Rochester, New York, USA) were used for transplantation. The donor cornea was punched with Barron trephine punch blades with diameters ranging from 7.25 to 9.00 mm. The donor buttons were 0.25 mm (for vitreous length ≥16 mm) or 0.50 mm (for vitreous length <16 mm) larger than the recipients. The endothelium of donor corneal buttons was peeled using a dry sponge after staining with 0.06% trypan blue dye. The button was secured in place using 10/0 nylon sutures. The suturing technique comprised 16-bite interrupted or 1 single running with 16 to 18 bites or a combination of both, according to the surgeon’s preference. Interrupted sutures were chosen if there was a peripheral corneal vascularization or history of vernal keratoconjunctivitis, which could lead to suture complications such as vascularized or loose sutures. Keratoscopy was performed intraoperatively to adjust suture tension.
After the surgery, all patients received topical ciprofloxacin 0.3%, prednisolone acetate 1%, and preservative-free artifical tear eye drops 6 times per day for 1 month. Artificial tears and prednisolone acetate 1% eye drops were tapered off over 3 to 6 months. In case of high astigmatism, selective suture removal was performed from 4 weeks onward.
A complete ophthalmic examination was performed preoperatively and postoperatively, including slit-lamp examination, uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), refractive error, ECD, and corneal topographic keratometry. Endothelial cell counts and topography were performed using a noncontact specular microscope (Topcon SP2000p; Topcon Corp, Tokyo, Japan) and the Orbscan II (Bausch & Lomb, Rochester, New York, USA) respectively.
SPSS statistics software package version 11.0 for Windows was used for the analysis. The Student t test, the Wilcoxon rank sum test, and the Mann-Whitney U test were used to compare the groups for the reviewed data. P < .05 was considered statistically significant. Snellen visual acuity measurements were converted to a logMAR score in order to simplify statistical analysis.
A total of 241 consecutive keratoconic eyes of 214 patients were evaluated. In 7 eyes (2.9%), the procedure was converted to PK intraoperatively because of Descemet membrane macroperforations. Therefore, data for 234 eyes of 207 patients were included for analysis ( Table 1 ).
|n (%) a|
|Recipient trephine size|
|7.00 mm||12 (5.1)|
|7.50 mm||69 (29.5)|
|7.75 mm||85 (36.3)|
|8.00 mm||49 (21.0)|
|8.50 mm||19 (8.1)|
|0.25 mm||175 (74.8)|
|0.50 mm||59 (25.2)|
|Suture removal (months)|
|Beginning, b mean ± SD (range)||5.33 ± 3.81 (1 to 20)|
|Completion, c mean ± SD (range)||14.05 ± 3.12 (9 to 20)|
At the time of the surgery the mean patient age was 25.4 ± 8.2 years (range 11 to 51 years). The mean follow-up period was 50.5 ± 22.2 months (range 24 to 96 months). The number of eyes that completed the follow-up periods of 8, 7, 6, 5, 4, and 3 years were 20, 44, 66, 106, 139, and 174, respectively.
Suture adjustment was performed in 21 of the 234 eyes (8.9%) at early postoperative period. There were no patients with sutures in at the last follow-up examination.
Table 2 shows preoperative and postoperative visual acuities. Postoperative improvement in UCVA and BSCVA was statistically significant ( P < .001). UCVA was lower than 20/100 in all 234 eyes preoperatively and better than 20/100 in 191 eyes (81.6%) postoperatively. Postoperative BSCVA was 20/40 or better in 187 eyes (79.9%) and 20/20 or better in 38 eyes (16.2%).
|logMAR, mean ± SD (range)||1.59 ± 0.38 (1.9 to 0.69)||0.48 ± 0.18 (1.3 to 0)|
|logMAR, mean ± SD (range)||1.14 ± 0.44 (1.9 to 0.4)||0.15 ± 0.13 (0.52 to 0)|
|≤0.3, n (%)||1 (0.4)||187 (79.9)|
|0.3-1.0, n (%)||63 (26.9)||45 (19.2)|
|>1.0, n (%)||170 (72.7)||2 (0.9)|
Preoperative keratometry was measurable in 186 eyes. Table 3 shows the mean preoperative and postoperative maximum keratometry and keratometric astigmatism. Postoperative decrease was statistically significant for both ( P values < .001). The keratometric astigmatism was lower than 4 diopters in 186 of 234 eyes (79.4%) postoperatively. Mean preoperative spherical equivalent value, which could be obtained in 179 eyes, also decreased significantly ( P < .001).
|Preoperative n = 186||Postoperative n = 234|
|Maximum keratometry, mean D ± SD (range)||60.53 ± 5.79 (49.0 to 71.3)||46.29 ± 4.14 (41.2 to 56.1)|
|Keratometric astigmatism, mean D ± SD (range)||8.7 ± 2.5 (2.6 to 12.1)||3.9 ± 3.1 (0.2 to 5.0)|
|Spherical equivalent, mean D ± SD (range)||−7.43 ± 3.20 (−2.0 to −3.25)||−0.90 ± 2.71 (−6.0 to 2.0)|
An exposed Descemet membrane (dDALK group) was achieved successfully in 193 of 234 eyes (82.4%) via big bubble. A layer-by-layer manual stromal dissection (pdDALK group) was performed in 41 eyes (17.6%) because of the lack of big bubble after several attempts. Interface haze was seen in 3 eyes (1.3%) in the pdDALK group, whereas no eye in the dDALK group developed interface irregularity. There was no statistically significant difference between dDALK and pdDALK groups in terms of UCVA, BSCVA, maximum keratometry, keratometric astigmatism, and spherical equivalent values at last visit ( P values: .670, .110, .830, .920, and .870, respectively) ( Table 4 ).
|pdDALK n = 41||dDALK n = 193|
|Mean ± SD||95% CI||Mean ± SD||95% CI||P Value|
|UCVA (logMAR)||0.56 ± 0.20||0.49 to 0.63||0.62 ± 0.27||0.58 to 0.66||.670|
|BSCVA (logMAR)||0.21 ± 0.07||0.18 to 0.24||0.16 ± 0.22||0.14 to 0.19||.110|
|Maximum keratometry (D)||45.53 ± 2.75||44.69 to 46.37||45.77 ± 3.52||45.41 to 46.14||.830|
|Keratometric astigmatism (D)||3.73 ± 1.42||3.27 to 4.19||3.52 ± 1.53||3.30 to 3.75||.920|
|Spherical equivalent (D)||−1.29 ± 2.27||−2.13 to 0.45||−1.03 ± 2.75||−1.43 to −0.62||.870|
Preoperative ECD could not be determined in 68 of 234 eyes (29.1%) because of the advanced KCN and apical scarring. The mean preoperative and postoperative ECD of the remaining 166 eyes were shown in Figure 1 , demonstrating endothelial cell loss with respect to preoperative ECD. Preoperative mean ECD was 2797 ± 561 cells/mm 2 . Endothelial cell loss was 8.1% ± 4.6% at 1 year, 10.5% ± 5.7% at 2 years, 15.1% ± 14.8% at 6 years, and 22.5% ± 15.9% at 8 years. Overall, mean endothelial cell loss at the last follow-up visit was 10.9% ± 6.3% in 166 eyes (decreased from 2797 ± 561 cells/mm 2 to 2490 ± 607 cells/mm 2 ). Figure 2 shows endothelial cell loss of 234 DALK eyes in respect to postoperative 1-year ECD. Endothelial cell loss was 3.6% ± 4.7% at 2 years, 12.6% ± 8.1% at 6 years, and 15.9% ± 7.3% at 8 years.
ECD change of 166 eyes was also evaluated in different subgroups. Although the mean endothelial cell loss was higher in eyes with Descemet membrane microperforation (n = 22, from 2811 ± 470 cells/mm 2 to 2473 ± 501 cells/mm 2 , 12.3% ± 7.9% loss) compared to the eyes without Descemet membrane microperforation (n = 144, 2757 ± 671 cells/mm 2 to 2498 ± 714 cells/mm 2 , 9.5% ± 4.7% loss), the difference did not reach a statistically significant level ( P = .531). When endothelial cell loss was compared between dDALK and pdDALK groups, it was more prominent in the dDALK group than in the pdDALK group, but the difference was not statistically significant ( P = .970). ECD decreased from 2774 ± 522 cells/mm 2 to 2496 ± 603 cells/mm 2 (n = 38, 10.4% ± 4.8% loss) in the pdDALK group and from 2805 ± 714 cells/mm 2 to 2484 ± 493 cells/mm 2 (n = 128, 11.3% ± 5.7% loss) in the dDALK group.
Descemet membrane perforation occurred in 25 of 241 eyes (10.4%) during the surgery. The surgery was continued at a pre-descemetic plane with the help of air injection into the anterior chamber in 18 cases with microperforations (7.5%). Preoperative mean central corneal thickness (CCT) was 361 ± 47.6 μm (range 289-411 μm) and there was no statistically significant difference in CCT between the eyes with or without Descemet membrane perforation ( P = .720, paired t test). Double anterior chamber developed in 2 eyes (0.8%) immediately after operation, which resolved after air injection into the anterior chamber.
Table 5 compares clinical data of the present study with other DALK series using the big-bubble technique for the treatment of KCN.