Purpose
To identify causes of reduced visual acuity and contrast sensitivity after big-bubble deep anterior lamellar keratoplasty (DALK) in keratoconus.
Design
Prospective interventional case series.
Methods
This study included 36 eyes in 36 patients with keratoconus who underwent DALK using the big-bubble technique. A bare Descemet membrane was achieved in all cases. Univariate analyses and multiple linear regression were used to investigate recipient-, donor-, and postoperative-related variables capable of influencing the postoperative quality of vision, including best spectacle-corrected visual acuity (BSCVA) and contrast sensitivity.
Results
The mean patient age was 27.7 ± 6.9 years, and the patients were followed for 24.6 ± 15.1 months postoperatively. The mean postoperative BSCVA was 0.17 ± 0.09 logMAR. Postoperative BSCVA ≥20/25 was achieved in 14 eyes (38.9%), whereas a BSCVA of 20/30, 20/40, or 20/50 was observed in 15 eyes (41.7%), 6 eyes (16.6%), and 1 eye (2.8%), respectively. Preoperative vitreous length was significantly associated with postoperative BSCVA (β = 0.02, P = .03). Donor-recipient interface reflectivity significantly influenced scotopic (β = −0.002, P = .04) and photopic (β = −0.003, P = .02) contrast sensitivity. The root mean square of tetrafoil was significantly negatively associated with scotopic (β = −0.25, P = .01) and photopic (β = −0.23, P = .04) contrast sensitivity. Recipient age, keratoconus severity, donor-related variables, recipient trephination size, and graft and recipient bed thickness were not significantly associated with postoperative visual acuity or contrast sensitivity.
Conclusion
Large vitreous length, higher-order aberrations, and surgical interface haze may contribute to poor visual outcomes after big-bubble DALK in keratoconus.
Different techniques for lamellar keratoplasty have evolved over time to achieve visual outcomes comparable to those of penetrating keratoplasty (PK). Among these techniques, deep anterior lamellar keratoplasty (DALK), in which a maximal depth of corneal stroma is removed, has gained popularity for its management of corneal stromal pathologies not involving the endothelium. Several studies have reported that visual function results are related to the type of donor-recipient interface that is accomplished with DALK. It has been demonstrated that when the recipient corneal stroma is removed down to the Descemet membrane, the optical quality of the interface is excellent and is comparable to that achieved through PK and that when layers of stroma are left adherent to the Descemet membrane, quality of vision is inferior to that achieved by PK. However, we have encountered patients whose best spectacle-corrected visual acuity (BSCVA) is not good, even when a bare Descemet membrane is achieved intraoperatively and the transparency of the donor cornea as well as the donor-recipient interface appears to be excellent postoperatively. In addition, some patients have reported limitations to their vision despite good corrected visual acuity.
Multiple factors can limit post-DALK visual performance, including lower- and higher-order aberrations and light scatter caused by the surgical interface or the use of low-quality grafts. The relative contributions of these factors in degraded visual performance after big-bubble DALK are poorly understood. Our study was designed to investigate, for the first time, the influence of donor-, recipient-, and postoperative-related variables on quality of vision (visual acuity and contrast sensitivity) after anatomically successful big-bubble DALK in keratoconus.
Methods
In this prospective interventional case series, patients who underwent big-bubble DALK between January 17, 2008 and March 5, 2014 for moderate (mean keratometry 48–55 diopters [D]) or advanced (mean keratometry >55 D or immeasurable keratometry) keratoconus were included. Indications for surgery included contact lens intolerance and poor corrected visual acuity. Ethics Committee approval was obtained from the Ophthalmic Research Center, which is affiliated with Shahid Beheshti University of Medical Sciences, Tehran, Iran, to conduct this prospective study, and the study adhered to the tenets of the Declaration of Helsinki. Informed consent was obtained from all participants after the purpose of the study was explained.
The inclusion criteria required an uncomplicated postoperative course (absence of a double anterior chamber, graft rejection, graft opacity, interface haze or wrinkling, cataract development, or raised intraocular pressure) and a minimum follow-up of 1 year. Exclusion criteria were any ocular comorbidity (such as amblyopia and strabismus), neurologic problems, systemic diseases, or the taking of any medication that may affect visual acuity or contrast sensitivity. Eligible participants were enrolled on a consecutive attendance basis.
Preoperatively, complete ocular examinations, including tests for uncorrected visual acuity (UCVA) and BSCVA using the Snellen acuity chart (expressed in logMAR notations), slit-lamp examination, tonometry, dilated funduscopy, manifest refraction (when possible), corneal topography (TMS-1 Topographic Modeling System, version 1.61; Computed Anatomy Inc, New York, New York, USA), and vitreous length measurement using A-scan biometry (A/B scan; Sonomed Inc, Lake Success, New York, USA) were performed. All procedures were performed under general anesthesia using the big-bubble technique, as described in detail elsewhere. A bare Descemet membrane was achieved in all cases. For all transplants, we used fresh donor sclerocorneal buttons that were preserved using cold storage. The donor corneas, which were oversized by 0.25 mm, were punched from the endothelial side with a Barron punch (Katena, Denville, New Jersey, USA) after the donor Descemet membrane was gently stripped off with a dry cellulose sponge or forceps. The recipient stromal bed was completely washed out to remove viscoelastic material and debris before proceeding to graft suturing. A combined suturing technique that consisted of 16-bite single running and 8-bite interrupted nylon sutures (Sharpoint; Angiotech, Vancouver, Canada) was used. Three months after surgery, selective interrupted suture removal was initiated to reduce astigmatism. At the time of the study, all sutures had been removed.
At least 3 months after complete suture removal, postoperative examination was performed. This included analyses of UCVA, BSCVA, manifest refraction, contrast sensitivity, higher-order aberrations (HOAs), and central corneal thickness using confocal microscopy. Orbscan II topography maps (Orbscan II; Bausch & Lomb, Rochester, New York, USA) were used for topographic assessments. Data collected from these maps included postoperative keratometry readings and irregularity index, which was measured at central 3 mm and 5 mm. Data relevant to the donor corneas were retrieved from the Central Eye Bank of Iran, where the donor tissues had been procured.
Contrast Sensitivity Measurement
Monocular contrast sensitivity was measured using sine-wave gratings at 6 spatial frequencies (1, 2, 3, 6, 12, and 20 cycles per degree [cpd]) using the Metrovision Moniteur Ophtalmologique “STATphot” program (Metrovision, Pérenchies, France). During the determination of contrast sensitivity, the chart was viewed from a distance of 2 m with correcting glasses in place. After an initial demonstration of the procedure, the contrast threshold was measured for each spatial frequency. All patients were tested under both scotopic and photopic conditions, and the results were expressed as log units of contrast sensitivity. Scotopic and photopic contrast sensitivity were also expressed as the area under the log contrast sensitivity function (AULCSF). This converts contrast sensitivity measures at different spatial frequencies into a single digit, facilitating the evaluation of several explanatory factors that may influence contrast sensitivity.
Wavefront Aberration Measurement
After measuring contrast sensitivity, cyclopentolate (1%) eye drops were instilled, and when a pupil diameter greater than 6 mm was achieved the wavefront was measured using a Zywave II aberrometer with Zywave software version 5.2 (Bausch & Lomb, Rochester, New York, USA) in a dark room. This aberrometer was used to calculate HOAs for a 6 mm pupil in terms of Zernike polynomials up to the fifth order. Three measurements were taken for each eye, and the average of the 3 readings was used to calculate different root mean square (RMS) values, which were expressed in micrometers.
Confocal Scan Examination
A confocal scan (Confoscan 3.4; Nidek Technologies, Padova, Italy) was used to measure central graft and recipient bed thickness and to quantitatively evaluate the donor-recipient interface. Using 3 Z-scan graphs in each cornea, central graft thickness (the distance between the epithelial and interface reflectivity peaks) and recipient bed thickness (the distance between the interface and the endothelial reflectivity peaks) were calculated and averaged. The donor-recipient interface was defined as the corneal sublayer located in the posterior stroma that displayed a discontinuity in stromal keratocytes and extracellular matrix architecture. Features of the interface that were evaluated for the purposes of the study included folds, deposits, and reflectivity. Interface reflectivity was calculated as the average of 3 maximum light reflectance unit values (expressed in arbitrary numerical units), which were obtained using Z-scan graphs.
Statistical Analysis
The data were analyzed using SPSS statistical software version 21 (IBM Corp, Armonk, New York, USA). Values indicating means and standard deviations, ranges, frequencies, and percentages were used to express data. The normal distribution of continuous variables was verified using a Kolmogorov-Smirnov test and a Q-Q plot. The Spearman correlation coefficient was used to analyze the influence of donor features (including age, death-to-preservation time, storage time, stromal status, endothelial cell density, and graft rating), recipient parameters (including age, preoperative mean keratometry and keratometric astigmatism, keratoconus severity, vitreous length, and trephination size), and postoperative outcomes (including follow-up period, spherical equivalent refraction, mean keratometry, keratometric astigmatism, graft irregularity indices measured at central 3 mm and 5 mm, RMS of each HOA, RMS of total HOAs, central graft thickness, recipient bed thickness, and interface reflectivity) on postoperative BSCVA, and scotopic and photopic AULCSF. Variables selected by the Spearman correlation coefficient based on a .05 significance threshold were introduced into a multiple linear regression model to evaluate the simultaneous effect of the variables. A P value <.05 was considered to be statistically significant. All reported P values are 2-sided.
Results
Recipient and Donor Characteristics
Forty-one eyes of 41 patients with keratoconus who underwent big-bubble DALK were initially enrolled. Five eyes were excluded from the study owing to subepithelial graft haziness (n = 1), interface haziness (n = 1), and interface wrinkling (n = 3). Therefore, data of 36 eyes were included for analysis. The mean age of the participants was 27.7 ± 6.9 years (range, 15–41 years). The mean vitreous length was 16.91 ± 1.28 mm and ranged from 15.22 to 20.59 mm. Moderate keratoconus was observed in 6 eyes (16.7%), whereas 30 eyes (83.3%) had severe keratoconus. The recipient trephination size was 7.75 mm in 8 eyes (22.2%) and 8.0 mm in 28 eyes (77.8%). The mean follow-up duration after corneal transplantation was 24.6 ± 15.1 months (range, 13–82 months). All grafts were clear at the final follow-up, and slit-lamp examination showed a clear interface with no visible opacities or wrinkling.
A total of 36 corneoscleral buttons from 36 cadavers, including 31 male and 5 female donors with a mean age of 35.1 ± 15.3 years (range, 10–70 years), were procured. The donor data are presented in Table 1 .
| Death-to-preservation time, N (%) | |
| <24 h | 10 (27.8) |
| 24–48 h | 26 (72.2) |
| Storage time (d), mean ± SD (range) | 3.2 ± 2.9 (0–11) |
| Stroma status, N (%) | |
| Clear | 26 (72.2) |
| Cloudiness | 10 (27.8) |
| Endothelial cell density (cells/mm 2 ), mean ± SD (range) | 2803.1 ± 629.6 (1128–3890) |
| Graft rating | |
| Excellent | 9 (25.0) |
| Very good | 12 (33.3) |
| Good | 8 (22.2) |
| Fair | 7 (19.5) |
Visual, Refractive, and Contrast Sensitivity Outcomes
The mean preoperative UCVA and BSCVA were 1.31 ± 0.30 logMAR (range, 0.60–2.10 logMAR) and 1.02 ± 0.49 logMAR (range, 0.48–2.10 logMAR), respectively. These figures were 0.66 ± 0.43 logMAR (range, 0.0–1.50 logMAR) and 0.17 ± 0.09 logMAR (range, 0.0–0.38 logMAR), respectively, at the final follow-up. There was a significant increase in postoperative UCVA ( P < .001) and BSCVA ( P < .001). Postoperatively, a BSCVA ≥20/25 was achieved in 14 eyes (38.9%), whereas a BSCVA of 20/30, 20/40, or 20/50 was observed in 15 eyes (41.7%), 6 eyes (16.6%), and 1 eye (2.8%), respectively. Preoperative spherical equivalent refractive error, mean keratometry, and keratometric astigmatism were −9.30 ± 5.0 D (range, −17.50 to −5.50 D), 53.53 ± 5.91 D (range, 48.50–64.0 D), and 5.88 ± 3.76 D (range, 1.50–15.75 D), respectively. Postoperative spherical equivalent refractive error, mean keratometry, and keratometric astigmatism were −3.84 ± 3.56 D (range, −12.88 to +1.38 D), 46.01 ± 3.22 D (range, 40.0–53.75 D), and 3.31 ± 2.24 D (range, 0.50–9.25 D), respectively ( P < .05 for all comparisons with corresponding preoperative values).
Postoperatively, graft irregularity indices measured at the 3 mm and 5 mm zones were 3.09 ± 0.85 D (range, 2.0–5.40 D) and 5.46 ± 1.44 D (range, 3.30–10.60 D), respectively. The scotopic and photopic contrast sensitivity measured at each spatial frequency is presented in Table 2 . The scotopic AULCSF was 1.40 ± 0.18 (range, 1.05–1.74) and the photopic AULCSF was 1.41 ± 0.21 (range, 0.99–1.74). The wavefront analysis of HOAs is summarized in Table 3 .
| Spatial Frequency | Scotopic Contrast Sensitivity (Decibels) | Photopic Contrast Sensitivity (Decibels) |
|---|---|---|
| 1 cycle/degree | 16.50 ± 2.45 (11.0–19.0) | 16.44 ± 2.37 (10.0–19.0) |
| 2 cycles/degree | 17.31 ± 3.96 (8.0–23.0) | 17.19 ± 4.15 (7.0–22.0) |
| 3 cycles/degree | 17.19 ± 4.72 (11.0–26.0) | 17.28 ± 5.24 (8.0–26.0) |
| 6 cycles/degree | 14.53 ± 5.0 (6.0–24.0) | 14.16 ± 5.70 (5.0–24.0) |
| 12 cycles/degree | 8.38 ± 5.12 (2.0–21.0) | 9.28 ± 5.46 (2.0–21) |
| 20 cycles/degree | 4.66 ± 3.31 (2.0–15.0) | 5.78 ± 3.59 (2.0–14.0) |
| Higher-Order Aberration | Mean ± Standard Deviation (μm) | Range (μm) |
|---|---|---|
| Trefoil | 1.44 ± 0.76 | 0.58–3.03 |
| Coma | 1.17 ± 0.64 | 0.08–2.70 |
| Spherical | 0.84 ± 0.47 | 0.20–2.07 |
| Tetrafoil | 0.54 ± 0.37 | 0.07–1.44 |
| Secondary astigmatism | 0.20 ± 0.18 | 0.01–0.77 |
| Pentafoil | 0.13 ± 0.14 | 0.0–0.49 |
| Third order | 1.99 ± 0.76 | 0.95–3.37 |
| Fourth order | 1.11 ± 0.46 | 0.31–2.19 |
| Fifth order | 0.18 ± 0.17 | 0.0–0.58 |
| Total | 2.20 ± 0.75 | 0.76–3.82 |
Confocal Scan Findings
Mean central corneal thickness was 524.3 ± 62.2 μm (range, 398.0–668.0 μm). Mean central graft thickness was 503.0 ± 61.4 μm (range, 370–637.4 μm), and mean recipient bed thickness was 21.8 ± 6.1 μm (range, 10.4–31.4 μm). A deep lamellar interface was easily identified in the examined eyes. There were hyporeflective striae in the rear stroma, which represented microfolds, and sheets of moderately to highly reflective amorphous deposits together with scattered high-contrast microdots at the interface area. The mean interface reflectivity value was 118.9 ± 35.2 light reflectance units (range, 60.0–196.0 light reflectance units).
Correlations
In the univariate analysis, postoperative BSCVA was significantly associated with preoperative vitreous length (r = 0.41, P = .04), postoperative spherical equivalent refraction (r = −0.39, P = .03), and the RMS of coma (r = 0.52, P = .02). Scotopic AULCSF was significantly associated with postoperative keratometric astigmatism (r = −0.56, P = .001), the RMS of coma (r = −0.50, P = .04), the RMS of tetrafoil (r = −0.62, P = .006), graft thickness (r = −0.37, P = .04), and interface reflectivity (r = −0.37, P = .04). Photopic AULCSF was correlated with postoperative keratometric astigmatism (r = −0.51, P = .003), the RMS of coma (r = −0.51, P = .03), the RMS of tetrafoil (r = −0.56, P = .02), graft thickness (r = −0.37, P = .04), and interface reflectivity (r = −0.44, P = .01). The RMS of total HOAs had no correlation with BSCVA (r = 0.26, P = .18) but demonstrated borderline association with scotopic AULCSF (r = −0.37, P = .048) and photopic AULCSF (r = −0.35, P = .06). The follow-up period was not correlated with postoperative outcomes including spherical equivalent refraction (r = −0.22, P = .13), keratometric astigmatism (r = 0.19, P = .58), the RMS of coma (r = 0.22, P = .37), the RMS of tetrafoil (r = 0.12, P = .21), graft thickness (r = 0.48, P = .79), and interface reflectivity (r = 0.36, P = .23). Additionally, the follow-up period had no significant influence on postoperative BSCVA (r = 0.03, P = .82), scotopic AULCSF (r = −0.16, P = .29), or photopic AULCSF (r = −0.17, P = .23). Recipient age, preoperative mean keratometry and keratometric astigmatism, keratoconus severity, donor-related variables, recipient trephination size, postoperative mean keratometry, and recipient bed thickness were not significantly associated with visual acuity or contrast sensitivity.
Multiple regression analyses revealed that postoperative BSCVA was significantly associated with preoperative vitreous length (β = 0.02, 95% confidence interval [CI]: 0.002–0.04, P = .03) ( Figure 1 ). However, postoperative spherical equivalent refraction (β = 0.001, P = .91) and the RMS of coma (β = 0.04, P = .18) lost their significance in the multiple regression model. Interface reflectivity significantly influenced the scotopic (β = −0.002, 95% CI: −0.005 to 0.0, P = .04) and photopic (β = −0.003, 95% CI: −0.006 to −0.001, P = .02) AULCSF ( Figure 2 ). The same analysis showed that the RMS of tetrafoil was significantly negatively associated with the scotopic (β = −0.25, 95% CI: −0.44 to −0.07, P = .01) and photopic (β = −0.23, 95% CI: −0.44 to −0.01, P = .04) AULCSF ( Figure 3 ). Scotopic AULCSF was not significantly correlated with postoperative keratometric astigmatism (β = 0.003, P = .91), the RMS of coma (β = −0.06, P = .38), graft thickness (β = 0.001, P = .60), or the RMS of total HOAs (β = 0.005, P = .14) in multiple linear regression. Similarly, photopic AULCSF was not significantly influenced by postoperative keratometric astigmatism (β = −0.006, P = .84), the RMS of coma (β = −0.01, P = .71), or graft thickness (β = −0.09, P = .27).
