Purpose
To compare outcomes of big-bubble deep anterior lamellar keratoplasty (DALK) and penetrating keratoplasty (PK) for macular corneal dystrophy.
Design
Prospective, randomized, interventional case series.
Methods
setting : Single hospital. patients : Eighty-two eyes of 54 patients requiring keratoplasty for the treatment of macular corneal dystrophy without endothelial involvement were included. main outcome measures : Operative complications, uncorrected visual acuity, best-corrected visual acuity, contrast sensitivity function, higher-order aberrations, and endothelial cell density were evaluated.
Results
The DALK and PK group consisted of 35 and 41 eyes, respectively. Best-corrected visual acuity after surgery was 20/40 or better 68.5% and 70.7% of the eyes in the DALK and PK groups, respectively ( P > .05). No statistically significant differences between groups were found in contrast sensitivity function with and without glare for any spatial frequency ( P > .05). Significantly higher levels of higher-order aberrations were found in the DALK group ( P < .01). In both groups, a progressive and statistically significant reduction in endothelial cell density was found ( P < .01). At the last follow-up, the mean endothelial cell loss was 18.1% and 26.9% in DALK and PK groups, respectively ( P = .03). Graft rejection episodes were seen in 5 eyes (12.1%) in the PK group, and regrafting was necessary in 3 eyes (7.3%). Recurrence of the disease was documented in 5.7% and 4.8% of the eyes in the DALK and PK groups, respectively.
Conclusions
Deep anterior lamellar keratoplasty with the big-bubble technique provided comparable visual and optical results as PK and resulted in less endothelial damage, as well as eliminating endothelial rejection in macular corneal dystrophy. Deep anterior lamellar keratoplasty surgery is a viable option for macular corneal dystrophy without endothelial involvement.
Macular corneal dystrophy is a bilateral autosomal recessive disorder that may result in significant visual discomfort. It is characterized by multiple grayish-white stromal opacities with indistinct and hazy borders that extend from limbus to limbus. As corneal opacity slowly becomes more dense and involves the visual axis, loss of functional visual acuity occurs. Therefore, keratoplasty eventually becomes necessary for the restoration of vision and the recovery of corneal transparency.
Traditionally, penetrating keratoplasty (PK) has been considered as the definitive treatment option for a variety of corneal pathologic features, including corneal stromal dystrophies. However, deep anterior lamellar keratoplasty (DALK) currently is considered to be the first-choice surgical procedure in patients with corneal disease not involving the endothelium, such as keratoconus, stromal scars, and stromal dystrophies. The main advantage of DALK is that the patient’s own endothelium is retained, which eliminates the risk of endothelial graft rejection and preserves endothelial cell density. Although DALK has several advantages over PK, there have been concerns about its role in macular dystrophy because of the involvement of deeper layers of stroma and possibly the Descemet membrane. Some investigators believe that DALK is not suitable for the treatment of macular corneal dystrophy, claiming that the stromal and endothelial involvement as well as the fragility of Descemet membrane in macular corneal dystrophy would lead to interface opacities and higher rates of endothelial cell attrition after DALK.
In the peer-reviewed literature to date, there are a few reports comparing therapeutic outcomes of PK and DALK surgery in the context of macular corneal dystrophy. In the current study, we prospectively compared big-bubble DALK and PK in terms of optical and visual outcomes, complications, and their effect on the endothelial cell density for the management of macular corneal dystrophy.
Methods
All patients were informed about the study as well as the advantages and disadvantages of the procedure. Informed consent was obtained from all patients to participate in this research study in accordance with the Declaration of Helsinki. The Institutional Review Board of Kartal Training and Research Hospital, Istanbul, Turkey, approved the study before it commenced. This study is registered at http://www.controlled-trials.com with identification number ISRCTN41152132. This study was a prospective, randomized clinical trial. Patients requiring keratoplasty for the treatment of macular corneal dystrophy without endothelial involvement were enrolled between January 1, 2006, and June 1, 2010. Macular corneal dystrophy diagnosis was made based on patient history, family history, and slit-lamp findings, including multiple gray-white stromal opacities extending to the deep stroma. Eighty-two eyes of 54 patients were included. All eyes were assigned a number randomly based on a surgical chart (even number, DALK; odd number, PK). Patients who were lost to follow-up, who had undergone previous eye surgery, or who underwent additional surgery combined with keratoplasty were excluded from the data analysis.
All eyes were operated on by a single experienced surgeon (A.K.). Deep anterior lamellar keratoplasty was performed by using the big-bubble technique previously described by Anwar and Teichman. When a big bubble could not be obtained after repeated attempts, a layer-by-layer manual dissection was performed. If during manual dissection the bare Descemet membrane could not be reached and diseased unclear stromal tissue remained, we converted to PK.
Full-thickness corneoscleral donor buttons stored in Optisol GS (Bausch and Lomb, Rochester, New York, USA) were used for transplantation. Penetrating keratoplasty procedures were performed using the standard technique. The donor buttons in both the DALK and PK groups were sutured with a single continuous suture or interrupted 10-0 nylon sutures.
A complete ophthalmologic examination was performed before the operation and postoperative visit in both groups. The examination included logarithm of the minimal angle of resolution uncorrected visual acuity (UCVA), logarithm of the minimal angle of resolution best-corrected visual acuity (BCVA), manifest refraction, slit-lamp biomicroscopy, and corneal topographic analysis with the CSO topography system (Costruzione Strumenti Oftalmici, Firenze, Italy). Contrast sensitivity measurements and corneal aberrometric analysis also were performed after surgery after all sutures were finally removed. The CSV-1OOOE chart (VectorVision, Greenville, Ohio, USA) was used for the assessment of contrast sensitivity. This test consists of 4 rows of sine-wave gratings (3, 6, 12, 18 cycles/degree) that had to be observed by the patient with full correction in place at a distance of 2.5 m. After an initial demonstration, the contrast threshold was measured for each spatial frequency. All patients were tested under both mesopic and photopic conditions, and the results were expressed in log units of contrast sensitivity. Corneal aberrometry was recorded and analyzed with the CSO topography system, whose software automatically converts the corneal elevation profile into corneal wavefront data using Zernike polynomials with an expansion up to the seventh order. The corneal aberration coefficients and root mean square (RMS) values were calculated for a 6.0-mm pupil.
Endothelial cell density of donor corneas were assessed by a specular microscope before storage in Optisol medium. The endothelium was photographed and evaluated using a Topcon SP 2000p noncontact specular microscope (Topcon Corp, Tokyo, Japan). Images of the central corneal window were reviewed by the same observer (E S.), and manual correction of the cell borders was performed before final analysis of the endothelium. Twenty endothelial cells were marked for each analysis. For each examination, 3 measurements of endothelial cell density were averaged.
Data were described as mean ± standard deviation (range). SPSS statistics software package version 15.0 for Windows (SPSS, Chicago, Illinois, USA) was used for statistical analysis. Normality of all data samples was checked by means of Kolmogorov-Smirrnov test. When parametric analysis was possible, the Student t test for unpaired data was used for comparisons between PK and DALK groups. When parametric analysis was not possible, the Mann–Whitney U test was applied for between-group comparisons. A P value less than .05 was considered statistically significant.
Results
A total of 82 eyes of 54 patients with macular corneal dystrophy were included. An equal number of eyes (41 eyes) underwent DALK and PK surgery. No intraoperative complication occurred in the PK group. Conversion to PK was needed in 6 eyes (14.6%) in the DALK group because of macroperforation. Therefore, the data analysis included 41 eyes in the PK group and 35 eyes in the DALK group. Complete Descemet membrane exposure was achieved in 27 eyes (77.1%; desmetic DALK [dDALK]) via the big-bubble technique; however, layer-by-layer manual stromal dissection was needed in 8 eyes (22.8%; predesmetic DALK [pdDALK]). Table 1 summarizes the preoperative conditions in the 2 groups of eyes analyzed. As shown, no statistically significant differences were found between the 2 groups in terms of age, sex, or preoperative visual acuity.
| DALK Group | PK Group | P Value | |
|---|---|---|---|
| Age (y) | |||
| Mean (SD) | 29.7 (11.3) | 33.0 (13.0) | .40 |
| Median (range) | 28.0 (16 to 61) | 31.0 (17 to 67) | |
| Sex (male/female) | 9/12 | 16/13 | .61 |
| UCVA (logMAR) | |||
| Mean (SD) | 1.34 (0.44) | 1.40 (0.46) | .51 |
| Median (range) | 1.30 (0.69 to 3.00) | 1.30 (0.69 to 3.00) | |
| BCVA (logMAR) | |||
| Mean (SD) | 1.30 (0.46) | 1.36 (0.48) | .53 |
| Median (range) | 1.30 (0.69 to 3.00) | 1.30 (0.69 to 3.00) |
The mean period between surgery and complete suture removal was 14.0 ± 3.4 months and 16.4 ± 3.8 months in DALK and PK groups, respectively ( P = .07). Mean complete follow-up time was 30.5 ± 8.75 months and 31.2 ± 9.78 months in the DALK and PK groups, respectively ( P = .53).
Visual Outcomes
After surgery, UCVA and BCVA improved significantly in both groups ( P < .01). Table 2 summarizes the comparative outcomes at the last visit. Logarithm of the minimal angle of resolution UCVA was significantly better in the PK group ( P = .02).
| DALK Group | PK Group | P Value | |
|---|---|---|---|
| UCVA (logMAR) | |||
| Mean (SD) | 0.62 (0.27) | 0.47 (0.21) | .02 |
| Median (range) | 0.69 (0.15 to 1.60) | 0.52 (0.00 to 1.00) | |
| BCVA (logMAR) | |||
| Mean (SD) | 0.35 (0.30) | 0.23 (0.11) | .13 |
| Median (range) | 0.30 (0.09 to 1.60) | 0.22 (0.09 to 0.40) |
At the last follow-up, UCVA was 20/40 or better in 11 eyes (31.4%) in the DALK group and in 14 eyes (34.1%) in the PK group ( P > .05). Best-corrected visual acuity was 20/40 or better in 24 eyes (68.5%) in the DALK group and in 29 eyes (70.7%) in the PK group ( P > .05). Uncorrected visual acuity was significantly better in the dDALK group than in the pdDALK group ( P = .04). Although not significant, BCVA also was better in dDALK group ( Table 3 ).
| pdDALK (n = 8) | dDALK (n = 27) | P Value (Mann–Whitney U Test) | |
|---|---|---|---|
| UCVA (logMAR) | |||
| Mean (SD) | 0.72 (0.39) | 0.58 (0.21) | .04 |
| Median (range) | (0.52 to 1.60) | 0.52 (0.15 to 1.00) | |
| BCVA (logMAR) | |||
| Mean (SD) | 0.44 (0.12) | 0.32 (0.30) | .05 |
| Median (range) | (0.15 to 1.60) | 0.22 (0.09 to 0.40) |
Contrast Sensitivity Outcomes
Figure 1 shows the mean contrast sensitivity function under mesopic conditions measured with and without a glare source. As shown, no statistically significant differences between groups were found in contrast sensitivity measured without glare for any spatial frequency (3 cycles/degree, P = .39; 6 cycles/degree, P = .77; 12 cycles/degree, P = .72; and 18 cycles/degree, P = .94). The same trend was observed for contrast sensitivity measured with glare (3 cycles/degree, P = .48; 6 cycles/degree, P = .60; 12 cycles/degree, P = .88; and 18 cycles/degree, P = .80). Comparing the subgroups, the mean contrast sensitivity function under photopic and mesopic conditions were not different between the dDALK and pdDALK groups for each of the spatial frequencies ( P > .05).
Corneal Aberrometric Outcomes
Aberrometric analyses were obtained in 27 eyes (77.1%) in the DALK group and in 30 eyes (73.2%) in the PK group ( Figure 2 ). Significantly higher levels of higher-order aberrations (HOAs), primary coma aberrations, spherical aberrations, and residual aberrations were found in the DALK group (HOA RMS, P < .01; coma aberration RMS, P < .01; spherical aberration RMS, P < .01; residual aberration RMS, P = .03). No significant differences between groups were detected in the total RMS ( P = .17). Significant correlations between contrast sensitivity and HOA RMS were observed only in the DALK group (no glare: 3 cycles/degree, r = −0.48, P = .02; 6 cycles/degree: r = −0.43, P = .04; 12 cycles/degree, r = 0.46 and P = .03; 18 cycles/degree, r = −0.40, P = .05; glare: 3 cycles/degree, r = −0.41, P = .05; 6 cycles/degree, r = −0.47, P = .02; 12 cycles/degree, r = −0.44, P = .04; 18 cycles/degree, r = −0.44, P = .04). We could not make a comparison between the dDALK and pdDALK groups because only 1 pdDALK patient had corneal aberrometric analysis.
Corneal Endothelial Density Changes
Figure 3 shows the endothelial cell density during the postoperative follow-up. Before surgery, endothelial cell density was measured in all donor corneas in the PK group; however, endothelial cell density was obtained in 27 eyes (65.8%) in the DALK group. The mean preoperative endothelial cell density was 2881 ± 449 cells/mm 2 and 2734 ± 549 cells/mm 2 in the DALK and PK groups, respectively ( P = .27). In both groups, a progressive and statistically significant reduction in endothelial cell density was found during the follow-up ( P < .01). Endothelial cell loss was 13.25 ± 11.21% at 6 months, 15.44 ± 12.31% at 12 months, 16.32 ± 14.76% at 24 months, and 18.12 ± 16.43% at the last follow-up in the DALK group. In the PK group, the mean endothelial cell loss was 12.17 ± 10.92%, 15.04 ± 13.21%, 22.12 ± 18.93%, and 26.98 ± 21.76% at 6 months, 12 months, 24 months, and the last visit, respectively. Significantly higher endothelial cell density loss was observed in the PK group compared with the DALK group at 24 months after surgery ( P = .03) as well as at the last postoperative follow-up visit ( P < .01). Comparing the subgroups, the mean endothelial cell losses were not different between the dDALK and pdDALK groups ( P > .05).
