Figure 1. Confocal images and image brightness intensity profiles before Descemet stripping with endothelial keratoplasty (DSEK) and through 2 years after the operation. The confocal images are from an individual cornea and are representative of the subepithelial, host stroma, lamellar wound interface, and donor stroma regions before and at 6 and 24 months after DSEK. Note the increased brightness of the subepithelial region and the surgical interface vs the normal control cornea (top right). The graph demonstrates the mean intensity of confocal brightness through the corneal stroma, with the subepithelial region referenced as 0% depth and the endothelium as 100% depth. The position of the surgical interface was scaled to 72% of depth, the mean interface depth, in each scan before calculating the mean. The brightness of both the subepithelial and interface regions improved partially after the operation. The stroma of the host and donor were similar to normal and did not change significantly during the study period. SU indicates scatter units.
Figure 2. Relationships between forward light scatter (straylight parameter) and corneal haze (confocal image intensity) at 12 months after Descemet stripping with endothelial keratoplasty. A, In the subepithelial layer, straylight was correlated with image brightness. B, In the surgical interface, no correlation was found. The solid line represents the linear regression fit across all subjects. The Pearson correlation coefficient and Bonferroni-adjusted P value are shown. SU indicates scatter units.
Figure 3. Improvement in confocal image brightness after Descemet stripping with endothelial keratoplasty (DSEK). A, Improvement in brightness of the subepithelial image at 12 months after DSEK diminished with increasing patient age. B, Improvement in the straylight parameter was directly related to improvement in brightness of the subepithelial image at 12 months after DSEK. The solid line represents the linear regression fit across all subjects. The Pearson correlation coefficient and Bonferroni-adjusted P value are shown. SU indicates scatter units.
Baratz KH, McLaren JW, Maguire LJ, Patel SV. Corneal Haze Determined by Confocal Microscopy 2 Years After Descemet Stripping With Endothelial Keratoplasty for Fuchs Corneal Dystrophy. Arch Ophthalmol. 2012;130(7):868-874. doi:10.1001/archophthalmol.2012.73
Author Affiliations: Department of Ophthalmology, Mayo Clinic, Rochester, Minnesota.
Objective To quantify corneal light scatter and its relationship to vision after Descemet stripping with endothelial keratoplasty (DSEK).
Methods Eyes with Fuchs corneal dystrophy were examined before and 1, 3, 6, 12, and 24 months after DSEK. Outcome measures were high- and low-contrast visual acuity, contrast sensitivity, and forward light scatter. Corneal reflectivity (backscatter), expressed in scatter units (SU), was quantified using in vivo confocal microscopy.
Results Comparing 49 study eyes with 35 normal eyes, the mean (SD) corneal subepithelial layer was more reflective than normal before (2325  vs 1208  SU, P < .001) and through 24 months after DSEK (1760  SU, P < .001). Interface reflectivity remained higher than in normal stroma through 24 months (1228  vs 827  SU, P < .001). At 1 year, forward light scatter correlated with subepithelial reflectivity (r = 0.28, P = .01) but not interface reflectivity. Recipient age was correlated with improvement in subepithelial reflectivity at 12 months (r = –0.41, P = .01, 34 eyes) and 24 months (r = – 0.36, P = .02, 26 eyes) after DSEK, and the improvement of subepithelial haze in eyes of participants aged 62 years or younger was correlated with improvement in forward light scatter at 12 months (r = 0.52, P = .008) and 24 months (r = 0.62, P = .004).
Conclusions In Fuchs corneal dystrophy, the corneal subepithelial region is a more important source of forward light scatter than the DSEK interface. Subepithelial haze improves more in younger patients and is associated with improvement in forward light scatter.
Clinical Relevance Visual function after DSEK is affected by residual haze in the anterior host cornea more so than the surgical interface. Haze, which likely is experienced as glare disability, improves after surgical intervention but improves more in younger patients.
Descemet stripping endothelial keratoplasty (DSEK) is the most common surgical procedure for the treatment of corneal endothelial dysfunction. Compared with penetrating keratoplasty, DSEK results in a more stable refractive error, lower astigmatism,1,2 and fewer corneal high-order wavefront aberrations.3 The absence of an avascular penetrating corneal wound also avoids serious vision-threatening postoperative complications, such as wound dehiscence. Descemet stripping endothelial keratoplasty is associated with significant improvements in visual function, with good uncorrected visual acuity being a major advantage over penetrating keratoplasty. Nevertheless, vision can remain suboptimal for some patients. Vajaranant and colleagues4 reported a median acuity of 20/40 one year after DSEK in more than 300 eyes. In eyes with Fuchs corneal dystrophy, more than 90% may achieve 20/40 or better after DSEK, but even in eyes without visual comorbidities, fewer than 20% will attain 20/20 vision.5 Similarly, other objective measures of visual function, such as low-contrast visual acuity, contrast sensitivity, and disability glare, do not return to normal after DSEK, which can explain postoperative reports of glare and poor contrast.6,7
Degradation of vision after DSEK can be explained in terms of the retinal image point spread function, which describes the distribution of the image of a point source of light onto the retina and can be subdivided into small- and large-angle domains.8,9 Although degradation of the small-angle domain (<1° from center) is typically caused by high-order aberrations that affect visual acuity,8 degradation of the large-angle domain (>1° from center) is usually caused by forward light scatter, or straylight, and produces disability glare.7,9 Forward light scatter is caused by imperfections of the optical media and may arise from the cornea, lens, iris, sclera, or vitreous opacities. After DSEK, increased disability glare might be caused by scarring at the lamellar graft-host interface or by abnormalities in the residual anterior host cornea. Indeed, in a previous small study,6 backscattered light, which is an objective measure of corneal haze, remained increased through 6 months after DSEK for eyes with Fuchs corneal dystrophy compared with normal eyes, but we were unable to localize the sources of increased backscatter within the cornea.
The goal of this study was to better characterize and localize changes in corneal haze through 2 years after DSEK for Fuchs corneal dystrophy by analyzing the brightness of confocal microscopy images using a standardized technique.10 Although the confocal image originates from reflected and backscattered light, a correlation between image brightness and forward scatter measured by visual function would suggest a causative relationship between corneal haze and vision. In this article, we address a simple hypothesis that corneal features that are visible as persistent subepithelial haze in the host cornea contribute to forward light scatter, or disability glare, after DSEK for Fuchs corneal dystrophy. We examined relationships between corneal reflectivity and visual function after DSEK; a strong relationship would support this hypothesis. However, a weak or no relationship would suggest that structures that produce corneal haze are not contributors to vision loss in these patients. Espana and Huang11 found a correlation between poorer visual acuity and subepithelial haze, but not interface haze, in a retrospective study of 25 eyes. Other investigators12,13 have attempted similar evaluations of very small groups of patients, but their results have been inconclusive.
Eyes with corneal edema caused by Fuchs corneal dystrophy were prospectively examined before and at 1, 3, 6, and 12, and 24 months after DSEK. Eyes were excluded if they had any other corneal disease (eg, stromal scar or herpetic keratitis), an anterior chamber intraocular lens, aphakia, poorly controlled glaucoma, or any other disease that would preclude good vision (eg, optic neuropathy, maculopathy, or amblyopia) or had undergone previous surgery for glaucoma. Outcome measures were high- and low-contrast visual acuity, contrast sensitivity, forward light scatter (straylight), and corneal reflectivity measured using confocal microscopy. Corneal reflectivity measured in eyes treated with DSEK was compared with that measured in 35 normal, untreated control eyes. The control group was younger (mean, 41 years; range, 24-54 years) and was not matched to the study participants (corneal reflectivity varies minimally with age14). The study was approved by the Mayo Clinic Institutional Review Board and complied with the Health Insurance Portability and Accountability Act; informed consent was obtained from each participant after the possible risks of the study were explained.
The DSEK was performed in a manner similar to the method described by Price and Price.15 In any phakic eyes, the lens was extracted by phacoemulsification and a posterior chamber intraocular lens was implanted at the time of DSEK. The central Descemet membrane was stripped from the host cornea. The donor tissue was prepared by the surgeon (K.H.B., L.J.M., or S.V.P.) or by a technician at the Minnesota Lions Eye Bank using an automated microkeratome (ALTK; Moria) with an intended anterior lamellar thickness of 300 μm or 350 μm. After placing a cohesive viscoelastic device on the endothelial surface, the donor tissue was folded and inserted into the eye with forceps through a 5- to 6-mm incision. Tamponade of the graft to the host cornea was facilitated by using an intracameral air bubble. Postoperatively, patients were treated with topical antibiotic solution and prednisolone acetate, 1% suspension. The latter was typically prescribed for administration 4 times daily, and the dosage was tapered to once daily at 3 to 4 months. Topical corticosteroid drops were continued at least through the 1-year postoperative visit.
Visual acuity was measured from all eyes with best spectacle correction after a manifest refraction at each visit. High-contrast visual acuity was measured by the electronic Early Treatment Diabetic Retinopathy Study protocol.16
Low-contrast visual acuity was measured in a darkened room at 3 m from a backlit 10% contrast chart (Sloan Translucent Low Contrast Chart; Precision Vision). Screen brightness was 139 candelas per square meter.
Contrast sensitivity was also measured (Functional Acuity Contrast Test; Vision Sciences Research Corporation).17 Participants were asked to discern the orientation (vertical, left, or right) of bars that had sinusoidal brightness at decreasing contrast. Contrast sensitivity was measured at spatial frequencies of 1.5, 3, 6, 12, and 18 cycles per degree (cpd).
Straylight, or intraocular forward-scattered light, was measured at 7° from fixation using a compensation comparison method (C-Quant Straylight Meter; Oculus).7,18
At every visit, the central cornea was examined by in vivo confocal microscopy (ConfoScan 4; Nidek) as described previously19 equipped with a z-ring adapter to stabilize the cornea and provide accurate depth information. The mean image intensity in the central 300 × 300 pixel area (167 μm × 169 μm) was calculated for each frame and expressed as scatter units (SU), the concentration of a standard turbidity solution (Amco Clear; GSF Chemicals) that produced the same mean brightness, as described by McLaren et al.10 Each scan was reviewed and the frames that corresponded to the epithelial surface, the anterior stroma, the surgical interface, and the endothelium were identified. To compare image brightness at the same depth relative to total thickness, the depth of each frame in the scan was scaled from 0% to 100% of stromal thickness, with the anterior stromal surface referenced as 0% and the endothelium as 100% of total thickness. In scans recorded after the operation, depth was scaled in 2 steps, first between the anterior stromal surface and the surgical interface, and second between the surgical interface and the endothelial surface. The surgical interface was scaled to 72%, with the mean interface depth a percentage of the total host and donor thickness (Figure 1). To compare the brightness of images before the procedure with those after the procedure, mean image brightness levels were calculated across a 7% stromal depth range for the subepithelial region, the host midstroma, the interface, and the donor midstroma.
High- and low-contrast visual acuity letter scores were expressed as logMAR. Variables before and after the operation were compared using generalized estimating equation models to account for possible correlation between fellow eyes of individual patients.20 Examination of changes in the brightness of confocal microscopy images was Bonferroni-adjusted for 5 comparisons, except for the analyses of interface and donor stromal image brightness, which were adjusted for 4 comparisons. Assessments of changes in visual function at 12 and 24 months vs visual function before surgical intervention were Bonferroni-adjusted for 3 comparisons. Correlations between variables were illustrated by calculating Pearson correlation coefficients. P values for these relationships were calculated using generalized estimating equation models. Minimum detectable differences were calculated for differences that were not significant (P > .05).
Fifty-two eyes of 45 participants with Fuchs corneal dystrophy were enrolled in the study and examined preoperatively. Three eyes of 3 participants failed to clear after the operation and were excluded from further analysis. Forty-eight eyes were examined at 1 month, 48 eyes at 3 months, 49 eyes at 6 months, 39 eyes at 12 months, and 29 eyes at 24 months. The mean patient age was 67 years (range, 41-87 years.) All eyes were pseudophakic after the operation with posterior chamber intraocular lenses.
Before the operation, 47 eyes had confocal scans with sufficient quality for assessment. Usable scans were available for 43 eyes at 1 month, 44 eyes at 3 months, 46 eyes at 6 months, 36 eyes at 12 months, and 27 eyes at 24 months after the operation.
Before DSEK was performed, confocal images of the study eyes were brightest in a region immediately deep to the epithelium, and this region was brighter than that in healthy control eyes (mean [SD], 2325  vs 1208  SU, P < .001). The brightness of the host midstroma before the operation was similar to that of the control group (1067  vs 915  SU, P = .09; minimum detectable difference, 171 SU, power = 80%, α = .05).
By 3 months after DSEK, the image brightness of the subepithelial region had decreased (mean [SD], 1953  SU, P < .005). Image brightness did not change significantly between 6 months (1803  SU) and 24 months (1760  SU, P = .99). At 24 months after DSEK, subepithelial reflectivity remained higher than that in controls (P < .001).
At 1 month after DSEK, interface reflectivity (mean [SD], 1613  SU) was higher than reflectivity of the deep stroma of normal controls (827 , P < .001). Interface brightness decreased between 1 and 12 months (1335  SU, P = .03) but did not change significantly between 12 and 24 months (1228  SU, P = .72; minimal detectable difference, 225 SU, power = 80%, α = .01). At 24 months, reflectivity from the interface remained higher than reflectivity of the deep stroma of normal controls (P < .001).
Host stromal reflectivity did not change significantly through 24 months after DSEK (1034  SU) compared with before DSEK (P > .99). Donor stromal reflectivity transiently increased between 1 month (923  SU) and 3 months (1029 , P = .02) but otherwise remained stable through 24 months (977  SU, P > .99 vs 1 month) (Figure 1).
Best-corrected high-contrast visual acuity was 0.43 (0.21) logMAR (Snellen equivalent, 20/54) before DSEK compared with 0.17 (0.15) logMAR (Snellen equivalent, 20/30) at 12 months (P < .001) and 0.14 (0.15) logMAR (Snellen equivalent, 20/28) at 24 months (P < .001) after the procedure. At 12 and 24 months after the operation, all other visual acuity variables also were improved compared with before the operation. There were no significant differences in variables from 12 months to 24 months, except a small decrease in contrast sensitivity at 1.5 cpd (Table). Only contrast sensitivities at 1.5 and 3 cpd were evaluated because many eyes could not discern any pattern orientation at higher spatial frequencies before and after DSEK.
At 12 months after DSEK, subepithelial reflectivity correlated with forward light scatter (r = 0.28, P = .01, 36 eyes) (Figure 2A) but not with high-contrast (r = 0.24, P = .76) or low-contrast (r = 0.11, P = .50) visual acuity or with contrast sensitivity (r = 0.09, P = .80, for 3 cpd). There were no other significant correlations between the reflectivity of any layer and either high-contrast visual acuity or forward light scatter at 12 months after the operation; specifically, there was no relationship between interface brightness and forward light scatter (r = –0.07, P = .81, 36 eyes) (Figure 2B). At 24 months after DSEK, although observing fewer eyes, we found no correlation between subepithelial reflectivity and forward light scatter (r = 0.38, P = .52, 27 eyes). However, at 2 years, the graft stromal brightness was correlated with high-contrast visual acuity (r = 0.38, P < .002, 27 eyes), but there was no correlation between image brightness from any other region and any visual function variable.
Recipient age was correlated with improvement in subepithelial reflectivity from before DSEK to 12 months (r = –0.41, P = .01, 34 eyes) (Figure 3A) and 24 months (r = –0.36, P = .02, 26 eyes) after the procedure. To evaluate the relationship between improvement of confocal image brightness and improvement of visual function from before to after DSEK, we separately analyzed the participants aged 62 years or younger at the time of the operation in whom lens opacity was minimal and presumably not contributing to decreased vision. We found that improvement in subepithelial reflectivity was correlated with improvement in forward light scatter (through 12 months: r = 0.52, P = .008, 17 eyes; through 24 months: r = 0.62, P = .004, 13 eyes) (Figure 3B). Similarly, improvement in subepithelial reflectivity was correlated with improvement in photopic low-contrast visual acuity (through 12 months: r = 0.60, P < .001, 12 eyes; through 24 months: r = 0.63, P = .001, 13 eyes). Fewer individuals were included in the latter comparisons because some could not read any letters in low-contrast conditions before DSEK.
Subepithelial reflectivity in corneas with Fuchs dystrophy is higher than normal; although it improves after DSEK, it does not return to normal by 2 years. Subepithelial reflectivity after DSEK is associated with forward light scatter but not with visual acuity, and the improvement in subepithelial reflectivity is higher in younger recipients. The surgical lamellar interface remains brighter than normal corneal stroma through 2 years after DSEK, but interface reflectivity is not associated with visual function.
Abnormalities have been recognized in all layers of corneas with Fuchs dystrophy by using confocal microscopy in vivo,21 but these were of little importance when penetrating keratoplasty was the treatment of choice for the disease. However, in the current era of endothelial keratoplasty, understanding the changes in the residual host cornea and their effect on visual outcomes is becoming more important. In this study, we found a striking increase in subepithelial reflectivity of corneas before DSEK, confirming previous observations12,13 in corneas with Fuchs dystrophy and other causes of corneal edema. The bright subepithelial zone most likely corresponds to subepithelial fibrosis, ultrastructural matrix changes, and corneal edema that accompany chronic endothelial dysfunction.22- 25 Increased subepithelial reflectivity in Fuchs corneal dystrophy is also accompanied by cellular abnormalities of the anterior stroma, including depletion of keratocytes26 and disorganization of corneal nerves.27 The roles of these cells, their interactions with the extracellular matrix, and their effects on surgical outcomes are not clearly understood.26
A previous small study28 found that corneal backscatter remained higher than normal through 6 months after DSEK for Fuchs corneal dystrophy. We confirmed and extended these observations in the present study by using confocal microscopy to assess corneal reflectivity through 2 years after DSEK. In addition, by using confocal microscopy, we were able to better localize the main sources of increased corneal backscatter to the subepithelial and interface regions of the cornea. Although subepithelial reflectivity improved during the first year after DSEK, it remained higher than normal, even at 2 years after the procedure. We suspect that the early improvement in subepithelial reflectivity after DSEK corresponded to resolution of corneal edema,29 whereas the late persistently increased subepithelial reflectivity was associated with the chronic ultrastructural changes of Fuchs corneal dystrophy. Similar changes in the profiles of confocal image brightness have been reported12,13 in small studies with limited follow-up. Of particular interest in the present study was a greater improvement in subepithelial reflectivity in younger recipients, suggesting that in younger host corneas, changes are more reversible or repair faster than in older recipients. Whether the increased subepithelial reflectivity and other anterior stromal abnormalities will normalize beyond 2 years after DSEK is unknown.
Reflectivity at the surgical interface was initially high after DSEK; similar to subepithelial reflectivity, however, it partially improved after the procedure. Increased light scatter from the surgical interface has often been assumed to cause poor visual outcomes after lamellar operations.30,31 Although the interface was a source of scatter after DSEK, subepithelial reflectivity was higher than interface reflectivity at all examinations in this study, suggesting that the anterior cornea was a more important source of scatter. Furthermore, interface reflectivity was not directly associated with poor visual outcomes, and this also has been noted by Espana and Huang.11 In contrast, subepithelial reflectivity (backscatter) was correlated with forward light scatter at 1 year after DSEK, although at 2 years, the correlation was not significant, possibly because of a smaller sample size. Nevertheless, in recipients aged 62 years or younger, in whom lenticular changes were a minimal source of forward scatter preoperatively, improvement in forward scatter was significantly correlated with improvement in subepithelial reflectivity. Thus, our data establish a relationship between subepithelial, but not interface, reflectivity and forward light scatter, and the greater improvement in subepithelial reflectivity in younger recipients might explain the greater improvement in disability glare in that population after DSEK.7
We did not find a relationship between subepithelial or interface reflectivity (haze) and high-contrast visual acuity, a notion that is sometimes used to explain decreased visual acuity after lamellar operations.11,30- 32 Corneal haze represents light that is scattered back to the observer, whereas it is forward light scatter that degrades the retinal image and affects vision; both probably originate from the same sources within the cornea.6 Nevertheless, forward light scatter has minimal effect on the center of the retinal image point spread function (small-angle domain), which affects visual acuity9 and thus would not be expected to decrease visual acuity, except under extreme conditions.8 In contrast, forward light scatter affects the peripheral flange of the point spread function and is experienced as disability glare.7 Decreased visual acuity after lamellar keratoplasty procedures is more likely the result of increased high-order aberrations that degrade the central point spread function33 more than corneal haze.34
The strength of this study is the rigorous prospective data collection, with standardized vision measurements and a standardized imaging technique10 in which we compensated for fluctuations in sensitivity of the confocal microscope. Use of the confocal microscope enabled us to better localize the sources of haze within the cornea, in contrast to previous measurements of corneal haze by using a custom slitlamp scatterometer.6 Measurements from these 2 instruments might not be directly comparable because of the nature of their measurement; the slitlamp scatterometer measures backscattered light,35 whereas the confocal microscope measures backscattered and reflected light. This study was limited by measurement of visual function in a standardized clinical setting and not under suboptimal conditions, such as with a glare source or poor ambient light. In patients interacting in the real world, visual function may be worse than what we measured if they perform under certain conditions that increase scatter from the cornea after DSEK.
In summary, subepithelial haze in corneas with Fuchs dystrophy, as measured by using confocal microscopy, improves after endothelial keratoplasty but does not return to normal by 2 years. Subepithelial reflectivity is associated with forward light scatter, which contributes to disability glare before and after DSEK for Fuchs corneal dystrophy. Subepithelial reflectivity improves more in younger recipients after DSEK for Fuchs corneal dystrophy, which corresponds to the greater improvement in disability glare in younger recipients.7 Haze at the surgical interface is not associated with forward light scatter, and haze from the subepithelial or interface regions is not associated with postoperative visual acuity. We conclude that persistent abnormalities in the host cornea affect the quality of vision after endothelial keratoplasty. Persistent haze of the host tissue, which improves more in younger graft recipients, suggests that younger patients will recover faster and have fewer optical limitations than will older patients. This finding has possible implications on the age at which surgical intervention is recommended, although this should be balanced against the risks of operating at a younger age. In addition, changes in subepithelial haze and the subsequent effect on vision are directly applicable to future endothelial replacement procedures, including DSEK.
Correspondence: Keith H. Baratz, MD, Department of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Submitted for Publication: August 14, 2011; final revision received November 19, 2011; accepted November 22, 2011.
Published Online: March 12, 2012. doi:10.1001/archophthalmol.2012.73
Author Contributions: Dr Baratz had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: None reported.
Funding/Support: This study was supported by Research to Prevent Blindness (Dr Patel was an Olga Keith Wiess Scholar and received an unrestricted departmental grant) and the Mayo Foundation.
Additional Contributions: David O. Hodge, MS, provided expertise in the statistical analysis of the data. Mr Hodge did not receive any compensation.