For Keratoconus

CLINICAL SCIENCE

Morphology of the Corneal Limbus Following Standard and Accelerated Corneal Collagen Cross-Linking (9 mW/cm2) for Keratoconus

Ömür Ö. Uçakhan, MD, and Betül Bayraktutar, MD

keratoconus or other corneal ectatic diseases. The initial Purpose: To evaluate the morphological features of the corneal surgical method used by these authors is known as the Dresden limbus as measured by in vivo confocal microscopy (IVCM) protocol or standard protocol. According to this protocol, following standard and accelerated corneal collagen cross-linking epithelium is debrided, 0.1% riboflavin (in 20% dextran) (CXL) for keratoconus. solution is applied on the cornea every 2 minutes for 30 minutes, Methods: and this is followed by ultraviolet A (UVA) application (370 Patients with progressive keratoconus scheduled to 2 undergo standard CXL (group 1; 31 patients, 3 mW/cm2, 370 nm, nm, 3 mW/cm ) for another 30 minutes during which time 2 riboflavin is continued to be instilled every 2 minutes. Because 30 minutes), or accelerated CXL (group 2; 20 patients, 9 mW/cm , “ ” 370 nm, 10 minutes) in the worse eye were included in this this procedure takes 1 hour, accelerated CXL has been prospective study. Thirty eyes of 30 age-matched patients served introduced recently, in which the goal is to decrease the as controls (group 3). All patient eyes underwent IVCM scanning of irradiation time to shorten the procedure, by increasing the beam intensity. Thus, all accelerated CXL protocols adhere to the central cornea and the inferior limbal area at baseline and 1, 3, – fl and 6 months after CXL. the Bunsen Roscoe law of reciprocity, and the uence is maintained at 5.4 J/cm2. Several protocols of accelerated CXL Results: After CXL, epithelial regrowth was complete by day 4 in have been proposed (10 minutes at 9 mW/cm2, 5 minutes at 18 both groups 1 and 2. There were no statistically significant differ- mW/cm2, etc); however, little is known about the safety and ences between the baseline mean central corneal wing or basal cell efficacy of these alternate procedures. Some evidence from density, limbus-palisade middle or basal cell densities of groups 1, 2, recent studies suggest that UVA application at 9 mW/cm2 for or 3. At postoperative months 1, 3, and 6, there were no statistically 10 minutes or 10 mW/cm2 for 9 minutes might have the same significant differences in either central or limbus-palisade epithelial cross-linking efficacy as 3 mW/cm2 for 30 minutes.2–4 cell densities or diameters in keratoconic eyes that underwent Although the corneal limbus is not intentionally standard or accelerated CXL (P . 0.05). The morphology of the exposed to UVA during CXL, there is concern about limbal cells was preserved as well. accidental irradiation to the limbus, and limbal protection Conclusions: with metal or sponge rings have been proposed. The The morphology of limbus structures seems to be morphological effects of standard CXL on the central preserved following standard and accelerated CXL in short-term corneal structures have been studied using in vivo confocal follow-up, as measured using IVCM. microscopy (IVCM)5,6; however, its effects on the corneal Key Words: collagen cross-linking, ultraviolet A, riboflavin, kerato- limbus morphology have not been studied in humans. conus, limbus, confocal microscopy, corneal morphology, ectasia Similarly, there are no studies evaluating the limbus morphology after accelerated CXL. In this study, we sought (Cornea 2017;36:78–84) to evaluate the effects of standard and accelerated (9 mW/ cm2) corneal CXL on the corneal limbus morphology using IVCM. ince its introduction by Wollensak et al1 in 2003, corneal Scollagen cross-linking (CXL) has been shown to be the only modality that can stop the progression of ectasia in METHODS Fifty patients with the diagnosis of progressive grade I through grade IV keratoconus were enrolled in this prospective Received for publication June 23, 2016; revision received July 30, 2016; study. Progression was defined as at least 1 diopter (D) increase accepted August 7, 2016. Published online ahead of print October 05, in maximum keratometry (Kmax) value, manifest cylinder 2016. From the Department of Ophthalmology, Faculty of Medicine, Ankara error, or manifest refraction spherical equivalent, or loss of 2 or University, Ankara, Turkey. more lines of best-corrected visual acuity attributable to Supported in part by a grant from the Ankara University. keratoconus progression. Exclusion criteria were central cor- The authors have no conflicts of interest to disclose. neal thickness less than 400 mm as measured by the Pentacam Reprints: Ömür Ö. Uçakhan, MD, Department of Ophthalmology, Faculty of Medicine, Ankara University, Mamak Caddesi, Cebeci, Ankara, Turkey (Oculus GmbH, Weltzar, Germany), history of eye surgery, 06340 (e-mail: [email protected]). corneal scar, atopy, vernal keratoconjunctivitis, contact lens Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. use, or any other ocular surface or ocular pathology.

78 | www.corneajrnl.com Cornea Volume 36, Number 1, January 2017

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea Volume 36, Number 1, January 2017 CXL and Limbus Morphology

Patients received either standard (group 1; 3 mW/cm2, bandage contact lens was removed, and patients were admin- 370 nm, 30 minutes) or accelerated CXL (group 2; 9 mW/cm2, istered topical loteprednol 0.5% eye drops (Lotemax; Bausch 370 nm, 10 minutes) to the worse eye. Thirty eyes of 30 age- & Lomb) 4 times daily. Topical steroids were tapered and matched patients served as the control (group 3). Written consent discontinued by month 1. was obtained from patients upon enrollment, and the study was approved by the Institutional Review Board. Patients underwent CXL no later than 1 month after baseline examinations. Preoperative and Postoperative Evaluation Patients were examined at baseline, everyday until epithelialization, at month 1, and month 3 after the procedure. Surgical Technique All patient eyes underwent confocal microscopy scanning of Standard CXL the central cornea and the inferior limbal area at baseline and 1 month, 3 months, and 6 months after CXL. Proparacaine hydrochloride 0.5% eye drops were instilled IVCM examination was performed with the HRT II 3 times at 5-minute intervals for topical anesthesia. After placing Rostock Cornea Module (Heidelberg Engineering GmbH, a lid speculum, the central 8- to 9-mm corneal epithelium was Heidelberg, Germany). During IVCM examination, patients fl manually removed. Ribo avin 0.1% drop in dextran 20% were asked first to look straight ahead for evaluation of the (MedioCross; Peschke Meditrade GmbH, Germany) was instilled central corneal epithelial morphology, and then to look up to onto the cornea every 2 minutes for half an hour. Then, the evaluate the limbal morphology at the 6 o’clock position. For patient was seated at the slit lamp to ensure the presence of each patient eye, cell density and diameter measurements fl yellow tint (ribo avin) in the anterior chamber. UVA light (3 were determined from the average of 3 different frames. One 2 mW/cm , 370 nm) (UV-X 1000, IROC, Switzerland) was examiner (B.N.F.) who was masked for the type of the CXL applied during the second half of the procedure for 30 minutes procedure, made the qualitative and quantitative analyses of 2 fl (total dose 5.4 J/cm ), and ribo avin continued to be instilled the image frames. One examiner (B.N.F.), who was masked every 2 minutes during this second half of the procedure. as to the type of CXL procedure, made the qualitative and quantitative analyses of the image frames. Accelerated CXL The central corneal wing cell density, central corneal After topical anesthesia and epithelial debridement as wing cell diameter, central corneal basal cell density, and described above, riboflavin 0.1% drops in hydroxypropyl methyl central corneal basal cell diameter were determined from the cellulose (VibeX Rapid; Avedro Inc, Waltham) were instilled central corneal IVCM scans; and limbus-palisade middle layer onto the cornea every 2 minutes for 30 minutes; then, after cell density, limbus-palisade middle layer cell diameter, limbus- observing riboflavin penetration to the anterior chamber, UVA (9 palisade basal cell density, limbus-palisade basal cell diameter, mW/cm2, 370 nm) (Avedro KXL; Avedro Inc) was applied for and palisade diameter were measured and averaged from 3 10 minutes (total dose 5.4 J/cm2), during which time riboflavin frames of each layer. Cell density was calculated using original continued to be instilled every 2 minutes. software of the confocal microscope, whereas cell diameters At the end of surgery, a bandage contact lens (balafilcon were measured using ImageJ software (ImageJ V.1.31; MD). A, PureVision; Bausch & Lomb) was placed on the cornea. Postoperatively, patients were prescribed topical moxifloxacin 0.5% eye drops (Vigamox, Alcon) 4 times daily (for 1 week) Statistical Analysis and topical preservative-free artificial tears as required. Patients Data were statistically analyzed through the SPSS were examined daily until epithelialization, at which time the program statistical package V.8.0 (SPSS Inc). The normality

TABLE 1. Demographic Characteristics of the Patients in Groups 1, 2, and 3 Group 1 Group 2 Group 3 (Standard CXL) (Accelerated CXL) (Control) P No. patients/eyes 30/30 20/20 30/30 Age, yr 22.0 6 3.4 (18–34) 24.6 6 3.7 (20–34) 23.2 6 4.8 (18–35) 0.07 Sex Female 33.0% (10) 40.0% (8) 50.0% (15) 0.88 Male 66.6% (20) 60.0% (12) 50.0% (15) Eye lateralization (eyes) Right 43.3% (13) 60.0% (12) 50% (15) 0.57 Left 56.7% (17) 40.0% (8) 50% (15) Keratoconus grade (eyes) Grade 1 33.3% (10) 50.0% (10) Grade 2 43.3% (13) 30.0% (6) Grade 3 10.0% (3) 10.0% (2) Grade 4 13.3% (4) 10.0% (2)

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Uçakhan and Bayraktutar Cornea Volume 36, Number 1, January 2017

FIGURE 1. Preoperative IVCM images of one patient who underwent standard CXL: (A) central corneal epithelial wing cell layer, (B) central corneal basal epithelial layer, (C) rete peg middle epithelial layer, (D) rete peg basal epithelial layer. of the data was tested before analysis. Because data were not pegs. Cells of the rete pegs were morphologically similar to normally distributed, the Wilcoxon signed-rank test was used wing cells of the central cornea (Fig. 1C); the borders of the for comparison of preoperative versus postoperative measure- Palisades of Vogt were lined with basal epithelial cells that ments. Kruskal–Wallis analysis of variance was used for were smaller than other rete peg cells with similar morpho- intergroup comparison. P , 0.05 was considered logical features (Fig. 1D). statistically significant. Table 2 shows the mean baseline central corneal and limbus-palisade cell density and diameter and the mean baseline palisade width measurements of groups 1, 2, and RESULTS 3. At baseline, there were no statistically significant Table 1 shows the demographic characteristics of the differences between keratoconic or normal eyes in regard patients in groups 1, 2, and 3. All 3 groups were age-matched to the mean central corneal wing or basal cell density and (Kruskal–Wallis test, P = 0.07). Most patient eyes had grade the mean limbus-palisade middle or basal layer cell density 1 or 2 keratoconus in both groups 1 and 2.7 (P . 0.05). The epithelial wing cells of the central cornea of Tables 3 and 4 show the mean baseline, and post-CXL normal and keratoconic eyes were observed to have dark month 1, month 3, and month 6 central corneal and limbus- cytoplasm, well-defined cell borders, and no visible nuclei palisade cell densities and palisade width measurements of (Fig. 1A). Basal cells in the central cornea had similar groups 1 and 2. Accordingly, there were no statistically features with smaller diameters (Fig. 1B). In the limbal significant differences between the data of postoperative scans, vertically oriented, hyperreflective, double-contoured, months 1, 3, or 6 compared with the baseline (P . 0.05). parallel linear structures were observed, representing pali- Morphology of the central corneal and limbus-palisade region sade ridges. These structures alternated with islands of epithelial cells also appeared to be unchanged following epithelial cell columns corresponding to interpalisade rete standard (Figs. 1, 2) and accelerated (Figs. 3, 4) CXL.

TABLE 2. Mean Baseline Palisade Width, Cell Density, and Diameter Measurements of the Central Cornea and the Inferior Limbus-Palisade Region of Groups 1, 2, and 3 Group 1 Group 2 Group 3 (Standard CXL) (Accelerated CXL) (Control) P* Central cornea Wing cell density, cells/mm2 4847 6 887 (2748–6341) 4999 6 636 (3922–6043) 4804 6 575 (3544–5664) 0.66 Wing cell diameter, mm 13.6 6 1.4 (11.0–16.8) 13.6 6 1.7 (10.4–15.2) 13.0 6 1.9 (9.5–17.7) 0.17 Basal cell density, cells/mm2 7593 + 856 (4734–8979) 7750 6 725 (6310–9024) 7908 6 656 (6348–9177) 0.83 Basal cell diameter, mm 8.0 6 1.2 (6.2–10.4) 7.8 6 1.1 (6.4–10.8) 7.7 6 1.3 (5.7–10.1) 0.21 Limbus-palisade Middle cell density, cells/mm2 4688 6 721 (3265–6376) 4819 6 595 (3811–6145) 4766 6 591 (3332–5834) 0.61 Middle cell diameter, mm 12.1 6 1.2 (9.3–14.8) 11.8 6 1.5 (10.0–15.2) 11.9 6 1.9 (7.9–15.3) 0.62 Basal cell density, cells/mm2 6900 6 789 (5446–8227) 7144 6 731 (5414–8880) 7113 6 660 (5641–8551) 0.49 Basal cell diameter, mm 8.2 6 1.1 (5.4–10.4) 8.9 6 1.1 (6.0–10.1) 7.4 6 1.2 (5.3–10.0) 0.06 Palisade diameter, mm 50.2 6 10.1 (26.0–77.0) 49.7 6 5.5 (38.9–62.0) 49.8 6 10.3 (22.4–69.3) 0.99

*Kruskal–Wallis analysis of variance.

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea Volume 36, Number 1, January 2017 CXL and Limbus Morphology

TABLE 3. Mean Baseline and Post-CXL Measurements at Months 1, 3, and 6 of Group 1 (Standard CXL) P* (Preoperative Pre-CXL Post-CXL Month 1 Post-CXL Month 3 Post-CXL Month 6 vs. Month-6) Central cornea Wing cell density, 4847 6 887 (2748–6341) 4584 6 700 (2622–5761) 4672 6 833 (2978–5944) 4751 6 812 (2876–6716) 0.07 cells/mm2 Wing cell 13.6 6 1.4 (11.0–16.8) 13.7 6 1.6 (10.0–17.0) 13.4 6 1.4 (11.0–16.5) 13.5 6 1.8 (9.1–16.3) 0.61 diameter, mm Basal cell density, 7593 + 856 (4734–8979) 7562 6 803 (5052–9200) 7585 6 783 (5018–8929) 7284 6 807 (5052–8766) 0.19 cells/mm2 Basal cell 8.0 6 1.2 (6.2–10.4) 7.9 6 1.1 (6.0–12.8) 8.2 6 1.3 (6.2–12.8) 7.8 6 1.0 (6.2–10.1) 0.9 diameter, mm Limbus-palisade Middle cell density, 4688 6 721 (3265–6376) 4714 6 689 (3570–6035) 4727 6 630 (3639–5976) 4800 6 671 (3639–6188) 0.19 cells/mm2 Middle cell 12.1 6 1.2 (9.3–14.8) 11.7 6 1.7 (8.7–15.4) 12.1 6 1.8 (8.0 6 15.3) 12.4 6 1.7 (8.0 6 15.5) 0.15 diameter, mm Basal cell density, 6900 6 789 (5446–8227) 7006 6 784 (5386–8315) 7058 6 761 (5441–8087) 6965 6 700 (5629–8156) 0.07 cells/mm2 Basal cell 8.2 6 1.1 (5.4–10.4) 7.9 6 1.2 (6.0–10.9) 8.0 6 1.0 (6.3–10.5) 8.0 6 1.1 (6.5–11.0) 0.68 diameter, mm Palisade 50.2 6 10.1 (26.0–77.0) 47.6 6 8.1 (28.2.0–62.5) 48.3 6 9.2 (26.4–70.8) 48.8 6 9.6 (24.1–70.8) 0.06 diameter, mm

*Wilcoxon signed-rank test.

DISCUSSION Confocal microscopy studies investigating the mor- CXL with neither the standard, nor the accelerated phology of keratoconic eyes in humans have concentrated on protocol (9 mW/cm2 for 10 minutes) seemed to cause any the morphology of the central cornea. Because of easier significant limbal stem cell damage at 6-month follow-up visualization, basal epithelial cells have been studied mostly, using IVCM. Additionally, there were no significant differ- and in 2 studies, the mean basal epithelial cell density was ences between the limbal morphology of keratoconic eyes at reported to be significantly lower in keratoconic eyes baseline compared with normal eyes. compared with normal eyes.8,9 In our study, the mean central

TABLE 4. Mean Baseline and Post-CXL Measurements at Months 1, 3, and 6 of Group 2 (Accelerated CXL) P* (Preoperative Pre-CXL Post-CXL Month 1 Post-CXL Month 3 Post-CXL Month 6 vs. Month 6) Central cornea Wing cell density, 4999 6 636 (3922–6043) 4730 6 432 (3896–5591) 4715 6 414 (3812–5535) 4917 6 786 (3879–6251) 0.28 cells/mm2 Wing cell 13.6 6 1.7 (10.4–15.2) 13.1 6 1.9 (10.2–17.0) 13.3 6 2.4 (12.8–18.2) 15.0 6 1.7 (12.6–18.2) 0.1 diameter, mm Basal cell density, 7750 6 725 (6310–9024) 7558 6 750 (6478–8973) 7663 6 715 (6536–8892) 7849 6 588 (7112–8815) 0.06 cells/mm2 Basal cell 7.8 6 1.1 (6.4–10.8) 7.5 6 1.0 (6.0–10.0) 7.7 6 1.1 (6.3–9.9) 8.3 6 0.6 (7.4–9.3) 0.08 diameter, mm Limbus-palisade Middle cell density, 4819 6 595 (3811–6145) 4752 6 521 (4107–5942) 4800 6 448 (3947–5732) 5031 6 490 (4425–5948) 0.31 cells/mm2 Middle cell 11.8 6 1.5 (10.0–15.2) 11.2 6 1.4 (8.7–15.3) 11.8 6 1.8 (9.6–15.7) 12.2 6 1.7 (9.3 6 15.9) 0.4 diameter, mm Basal cell density, 7144 6 731 (5414–8880) 7157 6 773 (5413–8744) 7210 6 753 (5425–8722) 7574 6 654 (6550–8677) 0.42 cells/mm2 Basal cell 8.9 6 1.1 (6.0–10.1) 8.1 6 1.1 (6.1–10.9) 8.1 6 1.1 (6.7–11.1) 8.6 6 1.5 (6.7–11.1) 0.99 diameter, mm Palisade 49.7 6 5.5 (38.9–62.0) 47.5 6 9.0 (33.0–62.5) 46.6 6 7.0 (32.3–62.6) 46.0 6 9.0 (34.5–58.2) 0.07 diameter, mm

*Wilcoxon signed-rank test.

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Uçakhan and Bayraktutar Cornea Volume 36, Number 1, January 2017

FIGURE 2. Postoperative month 6 IVCM images of the same patient eye as in Figure 1, with no difference in cellular morphology detected at any layer. corneal epithelial wing and basal cell densities of keratoconic accompanies keratoconus induces the synthesis of proinflam- eyes were not statistically significantly different from those of matory cytokines, which in turn lead to keratocyte apopto- age-matched healthy controls (Table 2). The difference sis.10,11 Such a mechanism may also be the underlying between our study and the previous 2 studies in regard to etiology for the lower limbal epithelial cell counts encoun- the mean basal epithelial cell density probably reflects the tered in keratoconic eyes, in the absence of contact lens wear, differences in the study populations. Differences in age range, or in vernal keratoconjunctivitis. presence of atopy and/or vernal conjunctivitis, keratoconus The efficacy of corneal CXL in halting the progression severity, and previous/current use of contact lenses may lead of keratoconus has been shown in several studies. The to such disparity between results. In our study, most patient procedure also has a fairly good safety profile, with eyes had grade I to II keratoconus, no patients had been using a complication rate of about 1% in 2-year follow-up.12 During contact lenses, no patient had a history of atopy, and the age CXL, although the central cornea is irradiated, partial range was narrower and younger compared with those of irradiation of the limbus seems to be unavoidable. However, previous studies. until now, limbal stem cell deficiency due to the procedure To our knowledge, our study is the first to concentrate has not been reported even in cases of pellucid marginal on the limbal microstructure of keratoconic eyes. Limbal stem degeneration (PMD), in which the procedure is performed cells are known to reside in the basal epithelial layer of the eccentrically.13,14 This has been proposed to be due to limbal palisades; therefore, changes in the limbal palisade protection offered by the epithelium left around the 8-mm epithelial structure may predate and lead to the central keratectomy, which absorbs 95% of UVA energy in a 400-mm epithelial changes encountered in keratoconus. Furthermore, thick cornea.15 Lateral diffusion of UVA has been reported to it has been suggested that the epithelium is affected early in be less than 20 mm.16 the disease process in keratoconus and epithelial injury that Previously, Wollensak et al17 applied sectorial UVA may arise due to eye rubbing, or atopic disease that frequently irradiation (5.4 J/cm2, 370 nm, 3 mW/cm2 for 30 minutes) to

FIGURE 3. Preoperative IVCM images of one patient who underwent accelerated CXL: (A) central corneal epithelial wing cell layer, (B) central basal epithelial layer, (C) rete peg middle epithelial layer, (D) rete peg basal epithelial layer.

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Cornea Volume 36, Number 1, January 2017 CXL and Limbus Morphology

FIGURE 4. Postoperative month 6 IVCM images of the same patient eye as in Figure 3, with no difference in the cellular morphology of the investigated corneal layers. the corneal limbus of rabbit eyes between the 8 and 10 double-fluence CXL application to the corneal limbus did not o’clock position with or without prior riboflavin (with alter the regenerative capacity of limbal epithelial cells or p63 dextran), and examined the corneas with immunohistochem- expression. The authors concluded that CXL did not induce istry and TUNEL staining 24 hours after irradiation. The potential harm to limbal stem cells in the immediate post- authors demonstrated no differences between irradiated and operative period or in midterm follow-up; however, long-term normal eyes with or without riboflavin use and concluded that potential effects including UVA-mediated mutagenicity might accidental irradiation of the limbal epithelium during CXL take years to appear. would not pose a threat for the induction of degenerative or In an IVCM study, Mazzotta et al5 evaluated morpho- neoplastic changes. logical changes induced by CXL in the central irradiated In a human ex vivo study, after CXL (5.4 J/cm2, 370 portion of human corneas. Although the study concentrated nm, 3 mW/cm2 for 30 minutes), counts of limbal viable cells on central corneal changes, the authors mentioned that there and reverse transcriptase polymerase chain reaction for was no loss of limbal germinal structures. In the aforemen- corneal differentiated epithelial cells and stem cells were tioned study, there was no mention of the part of the limbus performed.18 The investigators found a reduced number of that was examined, and no quantitative analysis had been viable limbal cells and a reduction of the expression of stem performed. cell marker p63 in 3 of 10 corneas. To our knowledge, this is the first IVCM study Shetty et al19 compared the effects of accelerated CXL investigating the effects of standard and accelerated (9 mW/ using different protocols (9 mW/cm2 for 10 minutes, 18 mW/ cm2 for 10 minutes) CXL on corneal limbal microstructure of cm2 for 5 minutes and 30 mW/cm2 for 3 minutes) and standard human eyes. We did not observe any changes in any structural CXL (3 mW/cm2 for 30 minutes) on ex vivo-cultured limbal component of the limbal–palisade zone of the treated eyes. epithelial cells using the quantitative real-time polymerase The cells appeared to be morphologically normal, and the cell chain reaction, vital staining, immunofluorescence staining, density and diameters were preserved throughout the 6-month and fluorescence-activated cell-sorting staining to evaluate the follow-up. One limitation of our study is the short-term apoptotic status. The authors reported a lower number of viable follow-up. A longer follow-up is required to definitively cells and more damage to limbal epithelial cells after standard exclude any potential mutagenic effect of CXL. Nevertheless, CXL compared with accelerated CXL with 30 mW/cm2 according to the results of this preliminary study, the for 3 minutes. morphology of the corneal limbus seems to be preserved after In an experimental animal study, Richoz et al20 applied standard (3 mW/cm2 for 30 minutes) and accelerated CXL with the standard protocol and fluence (5.4 J/cm2)tothe (9 mW/cm2 for 10 minutes) corneal CXL for keratoconus central 7 mm of 4 rabbit eyes and the cornea and limbus during 6 months of follow-up. (13 mm) of 4 rabbit eyes. In another set of 8 eyes, they used double-standard fluence of 10.8 J/cm2 to irradiate the central cornea (4 eyes) and the cornea and limbus (4 eyes). After REFERENCES epithelialization, the animals were killed and the corneal 1. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-a-induced buttons were examined with light microscopy and immuno- collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. histochemistry. Time to reepithelialization was similar to that 2003;135:620–627. of untreated controls in all eyes, and no endothelial cell damage 2. Hammer A, Richoz O, Arba Mosquera S, et al. Corneal biomechanical properties at different corneal corss-linking irradiances. Invest Ophthal- was seen in any rabbit eye. For both irradiation diameters and mol Vis Sci. 2014;55:2881–2884. fluences tested, no differences were observed in the p63 3. Kymionis GD, Tsoulnaras KI, Grentzelos MA, et al. Corneal stroma putative stem cell marker expression pattern. Therefore, even demarcation line after standard and high-intensity collagen crosslinking

Copyright Ó 2016 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. Uçakhan and Bayraktutar Cornea Volume 36, Number 1, January 2017

determined with anterior segment optical coherence tomography. 12. Koller T, Mrochen M, Seiler T. Complication and failure rates after corneal J Cataract Refract Surg. 2014;40:736–740. crosslinking. J Cataract Refract Surg. 2009;35:1358–1362. 4. Elbaz U, Shen C, Lichtinger A, et al. Accelerated (9-mW/cm2) corneal 13. Spaeda L. Corneal collagen cross-linking with riboflavin and UVA irradiation collagen crosslinking for keratoconus-A 1-year follow-up. Cornea. 2014; in pellucid marginal degeneration. J Refract Surg. 2010;26:375–377. 33:769–773. 14. Hassan Z, Nemeth G, Modis L, et al. Collagen cross-linking in the 5. Mazzotta C, Traversi C, Baiocchi S, et al. Corneal healing after riboflavin treatment of pellucid marginal degeneration. Indian J Ophthalmol. 2014; ultraviolet-A collagen crosslinking determined by confocal laser scan- 62:367–370. ning microscopy in vivo: early and late modifications. Am J Ophthalmol. 15. Mazzotta C, Traversi C, Baiocchi S, et al. Conservative treatment of 2008;146:527–533. keratoconus by riboflavin-UVA-induced cross-linking of corneal collagen: 6. Mastropasqua L, Nubile M, Lanzini M, et al. Morphological modification qualitative investigation. Eur J Ophthalmol. 2006;16:530–535. of the cornea after standard and transepithelial corneal cross-linking as 16. Mazzotta C, Balestrazzi A, Traversi C, et al. Treatment of progressive imaged by anterior segment optical coherence tomography and laser keratoconus by riboflavin-UVA-induced cross-linking of corneal colla- scanning in vivo confocal microscopy. Cornea. 2013;32:855–861. gen: ultrastructural analysis by Heidelberg Retinal Tomograph II in vivo 7. Krumeich JH, Kezirian GM. Circular keratotomy to reduce astigmatism confocal microscopy in humans. Cornea. 2007;26:390–397. and improve vision in stage I and II keratoconus. J Refract Surg. 2009; 17. Wollensak G, Mazzotta C, Kalinski T, et al. Limbal and conjunctival 25:357–365. epithelium after corneal cross-linking using riboflavin and UVA. Cornea. 8. Mocan MC, Yilmaz PT, Irkec M, et al. In vivo confocal microscopy for 2011;30:1448–1454. the evaluation of corneal microstructure in keratoconus. Curr Eye Res. 18. Vimalin J, Gupta N, Jambulingam M, et al. The effect of riboflavin-UV-A 2008;33:933–939. treatment on corneal limbal epithelial cells—a study on human cadaver 9. Bitirgen G, Ozkagnici A, Malik RA, et al. Evaluation of contact lens- eyes. Cornea. 2012;31:1052–1059. induced changes in keratoconic corneas using in vivo confocal microscopy. 19. Shetty R, Matalia H, Nuijts R, et al. Safety profile of accelerated corneal Invest Ophthalmol Vis Sci. 2013;54:5385–5391. cross-linking versus conventional cross-linking: a comparative study on ex 10. Kim WJ, Rabinowitz YS, Meisler DM, et al. Keratocyte apoptosis vivo-cultured limbal epithelial cells. Br J Ophthalmol. 2015;99:272–280. associated with keratoconus. Exp Eye Res. 1999;69:475–481. 20. Richoz O, Tabibian D, Hammer A, et al. The effect of standard and high- 11. Lema I, Duran JA. Inflammatory molecules in the tears of patients with fluence corneal cross-linking (CXL) on cornea and limbus. Invest keratoconus. Ophthalmology. 2005;112:654–659. Ophthalmol Vis Sci. 2014;55:5783–5787.