Journal of Refractive Surgery

Biomechanics Supplemental Data

Changes in Corneal Biomechanical Properties With Different Corneal Cross-linking Irradiances

Fangjun Bao, MD, PhD; YaRu Zheng, MD; Chang Liu, MD; XiaoBo Zheng, MSc; YiPing Zhao, MD; Yuan Wang, MD; LinNa Li, MD; Qinmei Wang, MD; Shihao Chen, MD, OD; Ahmed Elsheikh, PhD

Abstract

PURPOSE:

To evaluate whether different corneal cross-linking (CXL) irradiances, all with the same delivered total energy, achieve similar increases in corneal material stiffness.

METHODS:

One hundred twenty-six healthy white Japanese rabbits were randomly divided into seven groups (n = 18 each). After removing the epithelium of the left corneas, six groups were exposed to riboflavin (0.22% concentration by volume) and ultraviolet-A (370 nm) at different CXL irradiations, all with the same total dose (5.4 J/cm2), ranging from 3 mW/cm2 for 30 minutes to 90 mW/cm2 for 1 minute. The left corneas of the seventh group were exposed to riboflavin without irradiation. Twelve corneas of each group were prepared for inflation testing, where they were subjected to internal hydrostatic pressure simulating intraocular pressure, whereas the other six specimens were processed for electron microscopy measurements of fibril diameter and interfibrillar spacing. The inverse modeling process was used to estimate the tangent modulus of the tissue, which is considered an accurate measure of the material stiffness.

RESULTS:

The stiffening effect of CXL decreased when using high irradiation/short duration settings. Compared with the group with no irradiation (NUVA group), the tangent modulus increases reduced from 212.5% in the 3mW/30min group to 196.8% in the 90mW/1min group. These increases were significant (P < .05) in the 3mW/30min and 9mW/10min groups, but became insignificant in other CXL groups. The interfibrillar spacing in the anterior 50 μm of the corneal stroma also reduced with high irradiation/short duration settings, changing from 20.05 ± 1.89 nm in the NUVA group down to 13.06 ± 2.07 and 14.37 ± 1.90 nm in the 3mW/30min and 9mW/10min groups, respectively. These changes were significant (P < .05) between non-adjacent groups, but became non-significant otherwise. The corresponding changes in fibril diameter were small and nonsignificant in all cases (P > .05).

CONCLUSIONS:

Because the effect of CXL in stiffening the tissue and reducing the interfibrillar spacing consistently decreased with reducing the irradiance duration, the Bunsen–Roscoe law may not be readily applicable in the CXL of corneal tissue.

[J Refract Surg. 2018;34(1):51–58.]

Abstract

PURPOSE:

To evaluate whether different corneal cross-linking (CXL) irradiances, all with the same delivered total energy, achieve similar increases in corneal material stiffness.

METHODS:

One hundred twenty-six healthy white Japanese rabbits were randomly divided into seven groups (n = 18 each). After removing the epithelium of the left corneas, six groups were exposed to riboflavin (0.22% concentration by volume) and ultraviolet-A (370 nm) at different CXL irradiations, all with the same total dose (5.4 J/cm2), ranging from 3 mW/cm2 for 30 minutes to 90 mW/cm2 for 1 minute. The left corneas of the seventh group were exposed to riboflavin without irradiation. Twelve corneas of each group were prepared for inflation testing, where they were subjected to internal hydrostatic pressure simulating intraocular pressure, whereas the other six specimens were processed for electron microscopy measurements of fibril diameter and interfibrillar spacing. The inverse modeling process was used to estimate the tangent modulus of the tissue, which is considered an accurate measure of the material stiffness.

RESULTS:

The stiffening effect of CXL decreased when using high irradiation/short duration settings. Compared with the group with no irradiation (NUVA group), the tangent modulus increases reduced from 212.5% in the 3mW/30min group to 196.8% in the 90mW/1min group. These increases were significant (P < .05) in the 3mW/30min and 9mW/10min groups, but became insignificant in other CXL groups. The interfibrillar spacing in the anterior 50 μm of the corneal stroma also reduced with high irradiation/short duration settings, changing from 20.05 ± 1.89 nm in the NUVA group down to 13.06 ± 2.07 and 14.37 ± 1.90 nm in the 3mW/30min and 9mW/10min groups, respectively. These changes were significant (P < .05) between non-adjacent groups, but became non-significant otherwise. The corresponding changes in fibril diameter were small and nonsignificant in all cases (P > .05).

CONCLUSIONS:

Because the effect of CXL in stiffening the tissue and reducing the interfibrillar spacing consistently decreased with reducing the irradiance duration, the Bunsen–Roscoe law may not be readily applicable in the CXL of corneal tissue.

[J Refract Surg. 2018;34(1):51–58.]

Corneal cross-linking (CXL) with riboflavin and ultraviolet-A (UVA), first developed by Spoerl et al. in 1997,1 is a treatment modality designed to halt the progression of keratoconus and is considered a less invasive and lower cost option compared with keratoplasty. Unlike other treatments, the main role of CXL is to modify the intrinsic biomechanical properties of the collagen fibers.2 CXL increases the formation of intrafibrillar and interfibrillar covalent bonds by photosensitized oxidation,3 which can increase the mechanical stiffness4 and usually halts the progression of keratoconus and pellucid marginal degeneration ectasia after corneal refractive surgery.5–8

The effect of the standardized CXL has been proven in several studies that investigated the dosage parameters of riboflavin,9 safety of UVA radiation,10 penetration depth,11 and efficacy of the treatment in keratoconus.4 A significant long-term increase in corneal rigidity (Young's modulus) after CXL was achieved in porcine, human, and rabbit corneas.2,12 However, the application of the technique was judged to be too time-consuming,13 initiating efforts to speed up its implementation.

According to the Bunsen–Roscoe law, the same results can be achieved with a constant radiant energy (either a lower irradiance applied for a longer illumination time or a higher irradiance with a shorter time13). To our knowledge, all studies identified reductions in the stiffening effect of CXL with increases in irradiance (and shorter exposure time), even though the total energy delivered remained constant.14,15 However, these studies relied on uniaxial testing of strip specimens of corneal tissue, which does not represent the physiologic conditions affecting the cornea in vivo.16 This study aims to address this shortfall by relying instead on inflation tests, where the cornea is supported and loaded in close simulation of in vivo conditions. The study further relied on rabbit cornea specimens due to the similarity with human tissue17,18 and the shortage of human donor corneas available for experimental research.

Materials and Methods

Animals

One hundred twenty-six white Japanese rabbits (2 to 3 kg, aged 3 to 4 months) were obtained from the Animal Breeding Unit of Wenzhou Medical University and randomly divided into seven groups of 18 rabbits each. The rabbits were observed for 1 week before commencing the experimental study. All animals were treated in agreement with the Association for Research in Vision and Ophthalmology Statement for Use of Animals in Ophthalmic and Vision Research and with the approval of the Animal Care and Ethics Committee of the Eye Hospital, Wenzhou Medical University.

CXL Treatment

General anesthesia was administered by intramuscular injection of pentobarbital sodium (Merck KGaA, Darmstadt, Germany; 30 mg/kg) and SU-MIANXIN (Veterinary Institute at University of Munitions, Changchun, China) (0.2 mL/kg) and a wire eyelid speculum was positioned in the left eye of each rabbit. Prior to UVA irradiation, the epithelium was removed from the left eyes using a hockey knife and the corneas were saturated with 0.22% riboflavin drops (VibeX Xtra; Avedro, Inc., Waltham, MA) with 3-minute intervals over a total period of 30 minutes. The CXL procedure was then performed using a UVA irradiation system (CL-01; SiHaiTong Co., Suzhou, China) as per the protocols listed in Table 1. Irradiation was applied on the central 9-mm zone of the cornea and provided exposure to a total dose of 5.4 J/cm2.

Corneal Cross-linking Settings Adopted in Different Specimen Groups

Table 1:

Corneal Cross-linking Settings Adopted in Different Specimen Groups

The left eyes received tobramycin ophthalmic ointment (Tobrex; Alcon Laboratories, Inc., Fort Worth, TX) immediately after CXL, and continued to receive tobramycin ophthalmic drops and deproteinized calf-blood extract eye gel (Xingqi; Shenyang Xingqi Pharmaceutical Co., Ltd., Shenyang, China) three times a day for 1 week to help the new epithelium form completely. At this point, the rabbits were killed by intravenous injection of pentobarbital sodium overdose of 100 mg/kg and the left eyes were immediately enucleated. The corneas were separated along with a 3-mm–wide ring of scleral tissue and all other ocular components were removed. Twelve corneas from each group were prepared for inflation testing, whereas the other six were used in histological measurements. Corneal thicknesses in four directions (horizontal, vertical, 45°, and 135°) were measured using an ultrasonic pachymeter (SP-3000; Tomey, Inc., Nagoya, Japan). Central corneal thickness (CCT) in the initial stage before epithelium removal and exposure to UVA (CCT-initial), after epithelium removal and before CXL treatment (CCT-preCXL), at 5 minutes post-CXL (CCT-posCXL5min), and at 1 week post-CXL (CCT-posCXL1week) was measured.

Biomechanical Inflation Testing

As described previously,17,18 the corneas were mounted onto a custom-built pressure chamber used for corneal inflation testing. The pressure chamber was filled with phosphate buffered saline (Maixin, China) and connected to a syringe pump, which in turn was connected to a motor whose movement was controlled by bespoke LabView software. The pressure was controlled by the movement of the motor and continuously monitored using a pressure transducer (DMP-HS, Hangzhou, China) that connected with the pressure chamber. As indicated in a previous study,18 corneal diameters in four directions (horizontal, vertical, 45° and 135°) were measured using a Vernier caliper. Side images of the corneal profile were recorded with digital cameras (EOS 60D; Canon, Inc., Tokyo, Japan) positioned in the vertical and horizontal directions. The initial profiles and values of corneal diameters and thicknesses measured at 2 mm Hg pressure were used to construct a numerical model of each corneal specimen.

To ensure a fully inflated and wrinkle-free corneal surface, all specimens were subjected to an initial inflation pressure of 2 mm Hg. Connected to a personal computer to record the data automatically, a charge-coupled device laser displacement sensor (LK series; Keyence, Ithasca, IL) was used to monitor the displacement at the corneal apex continually. Each specimen was tested within 3 hours postmortem. To condition and stabilize the behavior, three cycles of loading and unloading up to a pressure of 30 mm Hg were applied at a rate of 0.41 mm Hg/sec. A recovery period of 90 seconds was allowed between each of the two loading cycles to ensure the behavior was not affected by the strain history of loading cycles.18 Finally, the specimens were subjected to a fourth loading cycle, the results of which were used in a subsequent inverse analysis.

Inverse Analysis

An inverse analysis process was used to evaluate the material's mechanical properties of corneal tissue based on the pressure-deformation experimental results. As described in a previous study,18 the finite element solver Abaqus (Dassault Systèmes Simulia Corporation, Forest Hill, MD) and optimization software package LS-OPT (Livermore Software Technology Corporation, Livermore, CA) were used to implement the iterative process of the inverse analysis procedure. Eighty-four finite element, specimen-specific models, each employing 1,728 fifteen-noded continuum elements (C3D15H), arranged in 12 rings and two layers, were developed from the specimens' initial geometry based on their thickness, corneal profile, and diameter measurements. An encastre connection was assumed along the limbus to simulate connection to the mechanical clamps. A first-order hyperelastic Ogden model18 was used to represent corneal material behavior using a strain energy density function in the form:

W=2μa2(λ1¯α+λ2¯α+λ3¯α−3)+1D(J−1)2
where λk is the deviatoric principal stretches = J−1/3 × λk (k = 1, 2, 3), λ1, λ2, λ3 the principal stretches, J = λ1λ2λ3. W represents the strain energy per unit volume, whereas material parameters μ and α are the strain hardening exponent and the shear modulus, respectively. D is a compressibility parameter
=3(1−2ν)μ(1+ν)
calculated assuming corneal tissue was nearly incompressible19 with a Poisson's ratio, ν, of 0.48.

Histological Analysis

Six corneas from each group were isolated from the ocular globe and earmarked for histological measurements, which were performed in a pressure-free state. They were fixed in glutaral, embedded in Epon (SPIPON 812 Kit; Structure Probe, Inc., West Chester, PA), sectioned on the sagittal plane, and stained with uranyl acetate and lead citrate to semi-quantify the collagen components and interfibrillar spacing. The histological assessment was conducted by a professional pathologist. Miron-thick (50 nm) sections were removed from each cornea and analyzed under a transmission electron microscope (TEM; H-7500; Hitachi, Ltd., Tokyo, Japan) at ×40,000 magnification. The mean diameter of fibrils (2*r) in a TEM image was calculated from the fibril's cross-sections, whereas the interfibrillar spacing was determined using imaging analysis software (Image J; National Institutes of Health, Bethesda, MD) (Figure A, available in the online version of this article). Because CXL has a depth-dependent stiffening effect and is most effective in the anterior cornea,20,21 the diameters of fibrils were measured in the anterior 50 μm of the corneal stroma just under the Bowman's membrane. Only fibrils with clearly defined, circular borders and high contrast were included. Fibril cross-sections with a slightly ellipsoidal form were discarded.22 On the other hand, the interfibrillar spacing (D) was calculated as:

D=Area/number of fibrils−2×r
The mean values of D and r obtained from five TEM images were used in the analysis.

The measurement of fibril diameter and interfibrillar spacing, where r is the radius of a collagen fibril and D is the interfibrillar spacing between two collagen fibrils.

Figure A.

The measurement of fibril diameter and interfibrillar spacing, where r is the radius of a collagen fibril and D is the interfibrillar spacing between two collagen fibrils.

Statistical Analysis

All analyses were performed using the PASW Statistics 20.0 software (SPSS, Inc., Chicago, IL). Comparisons of CCT values measured at different test stages and tangent modulus (Et) obtained for the seven specimen groups were performed using one-way analysis of variance test. Post hoc Tukey comparisons were used to isolate significant interactions. Comparison between CCT-initial and CCT-posCXL1week in each specimen group was tested using the paired t test. To determine the influence of irradiation on other parameters, including Et, CCT-posCXL5min, material parameters μ and α, root mean square (RMS), fibril diameter, and interfibrillar spacing, linear regression analyses were performed with irradiance as the independent variable and others as the dependent variables. P values less than .05 were considered statistically significant.

Results

CCT increased significantly after epithelium removal and before exposure to UVA (CCT-preCXL vs CCT-initial, P < .001), then decreased slightly at 5 minutes post-CXL (CCT-posCXL5min vs CCT-preCXL, P < .001) but remained much thicker than the initial thickness (CCT-posCXL1week vs CCT-initial, P = .002). This was followed by a gradual decrease toward the CCT-initial values, and the differences between CCT-initial and CCT-posCXL1week were not significant (P > .05) except in the 3mW/30min (P = .009) and 9mW/10min (P = .022) groups (Table A, available in the online version of this article). The fluctuation of thickness throughout the four stages (CCT-initial, CCT-preCXL, CCT-posCXL-5min, and CCT-posCXL1week) was similar among the seven specimen groups (P = .216, .641, .158, and .389, respectively). As shown in Figure 1, a clear difference in pressure-displacement behavior at the corneal apex was observed among the seven specimen groups.

Mean ± Standard Deviation of CCT in All Specimen Groups

Table A:

Mean ± Standard Deviation of CCT in All Specimen Groups

Mean pressure-displacement behavior at the corneal apex in the seven specimen groups.

Figure 1.

Mean pressure-displacement behavior at the corneal apex in the seven specimen groups.

Inverse Analysis of Inflation Test Results

The inverse modeling analysis was used to derive a constitutive model for each cornea that provides the best possible match (lowest RMS) with the experimentally obtained pressure-displacement results. With material parameters α and μ determined (Table 2), the stress-strain (σ − ε) relationship, and hence the tangent modulus (Et = dσ / dε) at any stress level, can be obtained. The mean stress-strain behavior for each specimen group is presented in Figure 2, whereas Et at a stress level of 0.01 MPa is plotted in Figure 3. The tangent modulus values showed a significant global difference between the seven groups (P < .05). Statistically significant differences were found in Et between the control group (NUVA) and the two groups with the longest irradiation duration; 3mW/30min (P = .012, increased by 212.5%) and 9mW/10min (P = .049, increased by 196.8%). With the other groups, the differences were not significant, although the mean ratios were considerable (Figure 3). Significant differences in Et were also observed between the CXL groups, but only between the groups with the longest and shortest irradiation times: 3mW/30min and 90mW/1min (P = .043). All other differences were not significant (Figure 3). Regression analysis of irradiance with Et (R2 = 0.141, P = .001), CCT-posCXL5min (R2 = 0.054, P = .049), α (R2 = 0.133, P = .002), and interfibrillar spacing (R2 = 0.480, P < .001) showed significant linear relationships (Figure B, available in the online version of this article). However, there was no significant correlation with μ (R2 = 0.051, P = .059), RMS (R2 = 0.002, P = .746), and fibril diameter (R2 = 0.011, P = .551).

Mean ± Standard Deviation of Constitutive Parameters α and μ in All Specimen Groups

Table 2:

Mean ± Standard Deviation of Constitutive Parameters α and μ in All Specimen Groups

Mean stress-strain behavior of corneas in each specimen group. Error bars represent standard deviation of stress values.

Figure 2.

Mean stress-strain behavior of corneas in each specimen group. Error bars represent standard deviation of stress values.

Mean and standard deviation of tangent modulus values for corneas included in each specimen group obtained at 0.01 MPa stress level. NUVA = no irradiation

Figure 3.

Mean and standard deviation of tangent modulus values for corneas included in each specimen group obtained at 0.01 MPa stress level. NUVA = no irradiation

Regression analysis of irradiance vs tangent modulus, central corneal thickness at 5 minutes after corneal cross-linking (CCT-posCXL5min), material parameter α, and interfibrillar spacing.

Figure B.

Regression analysis of irradiance vs tangent modulus, central corneal thickness at 5 minutes after corneal cross-linking (CCT-posCXL5min), material parameter α, and interfibrillar spacing.

Histological Analysis

Analysis of electron microscopy images (Figure C, available in the online version of this article and Table B, available in the online version of this article) showed small and non-significant differences in the fibril diameter among the seven specimen groups (P = .06), whereas interfibrillar spacing experienced large and significant differences (P = .00). Interfibrillar spacing was lower in the 3mW/30min group (indicating higher fibril density) compared with the 30mW/3min (P = .007), 45mW/2min (P = .002), 90mW/1min (P < .001), and NUVA (P < .001) groups, respectively. The differences between the 3mW/30min group and other groups at the higher end of irradiation times (9mW/10min and 18mW/5min) were not significant (P > .05). Additionally, a significant difference could also be found between the 9mW/10min and 45mW/2min (P = .033) groups, 9mW/10min and 90mW/1min (P = .001) groups, 9mW/10min and NUVA (P = .001) groups, and 18 mW/5min and NUVA (P = .033) groups.

Images of collagen fibrils in the anterior stroma obtained using transmission electron microscopy with 40,000 magnification. (A) 3mW/30min group, (B) 9mW/10min group, (C) 18mW/5min group, (D) 30mW/3min group, (E) 45mW/2min group, (F) 90mW/1min group, (G) no irradiation (NUVA) group.

Figure C.

Images of collagen fibrils in the anterior stroma obtained using transmission electron microscopy with 40,000 magnification. (A) 3mW/30min group, (B) 9mW/10min group, (C) 18mW/5min group, (D) 30mW/3min group, (E) 45mW/2min group, (F) 90mW/1min group, (G) no irradiation (NUVA) group.

Mean ± Standard Deviation of Interfibrillar Spacing and Fibril Diameter in Anterior 50 μm of Corneal Stroma in Different Specimen Groups

Table B:

Mean ± Standard Deviation of Interfibrillar Spacing and Fibril Diameter in Anterior 50 μm of Corneal Stroma in Different Specimen Groups

Discussion

This study examined the effectiveness of the CXL technique when changes are made in the irradiance and duration while keeping the magnitude of energy delivered constant. Our results indicate a consistent downward trend (significant negative correlation exists between irradiance and biomechanical parameters including tangent modulus Et and material parameter α), in which the effectiveness of CXL in stiffening the tissue (or increasing Et) decreased significantly as the irradiation time was reduced from 30 to 1 minute, even though the irradiance increased proportionally from 3 to 90 mW/cm2 to maintain the same level of energy delivery. This result is consistent with earlier reports, in which Et increases with 3mW/30min irradiation were found to decrease from 137%,15 210%,14 and 213% (current study) to 117%, 205%, and 197%, respectively, with 9mW/10min irradiation. In our study, the reduction in Et increases continued until they almost completely diminished in corneas with irradiation between 18mW/5min and 90mW/1min. However, some contrary evidence exists in the literature where similar biomechanical changes23 were achieved with standard 3mW/30min and faster 10mW/9min CXL, but in that study no further reductions in irradiation times were attempted.

These differences between current and earlier results could be related to various factors. First, the biomechanical measurements were performed at different times. Whereas Wernli et al.14 took their measurements at 30 minutes after starting irradiation, Hammer et al.15 recorded measurements at 30 minutes after the end of irradiation. In our study, measurements were taken 1 week after CXL. For this reason, wound healing and the different extents of epithelium regrowth at these time points may have played a role in the difference in results. Second, the three studies used different animal models: porcine corneas in Hammer et al.15 and Wernli et al.'s14 studies and rabbit corneas in our experiments. Third, the concentration of riboflavin was different: 0.1% in Hammer et al.15 and Wernli et al.'s14 studies and 0.22% in our study. The decisions adopted in our study followed the standard Dresden protocol, albeit with a different concentration of riboflavin, and our measurements were taken after the epithelium had regrown and the eye had largely recovered. The high riboflavin concentration (0.22%) adopted in our study (in all specimen groups) is becoming increasingly common in accelerated CXL treatment24 and is used to enhance the riboflavin penetration through corneal tissue and achieve higher UVA absorption,25 thus protecting the endothelium26 and improving the safety of CXL in rabbit corneas in which the initial CCT (approximately 360 μm) is less than 400 μm.

Most importantly, the experimental platforms employed in the studies were not the same. The strip extensiometry used in Hammer et al.15 and Wernli et al.'s studies14 involves cutting the cornea into strips (causing the termination of fibrils along the specimen sides) and subjecting the strips to uniaxial tension. Both of these factors violate the cornea's in vivo, physiologic conditions and are expected to cause remodeling of the tissue's microstructure, aligning collagen fibers more in the direction of the load, and affecting the accuracy of the results' representation of real behavior. The technique has further limitations caused by the initially curved form of the specimen, whereby the straightening of the specimens from their curved form results in initial strains that affect the behavior under subsequent loading.16 In contrast, inflation testing maintains the cornea's integrity and applies a load that closely simulates the eye's intraocular pressure.

Earlier studies identified two distinctive phases in corneal stress-strain behavior: an initial matrix-regulated phase with low stiffness followed by a collagen-regulated phase with much higher stiffness.27 In the first phase, the corneal biomechanical properties are dominated by the extracellular matrix while the collagen fibril remain loose and unable to contribute to the overall performance. However, as the tissue stretches under load, the collagen fibrils become increasingly taut and take over the overall behavior, making the cornea much stiffer. In our study, the initial, low-stiffness phase was observed to shorten in specimens cross-linked with long-duration/low-irradiance (3mW/30min, 9mW/10min, 18mW/5min). This feature is important because it indicates the success of CXL, with some of the irradiance-duration combinations, in reducing the tissue's deformation, especially in the behavior stage where most of the distortion takes place.

Stromal ultrastructural abnormalities found in keratoconus comprise altered spatial distribution of proteoglycans, changes in collagen organization, and uneven distribution of collagen mass, as well as changes in fibril diameter and interfibrillar spacing.28 The CXL stiffens the cornea through the creation of covalent bonds between collagen and proteoglycan molecules,29 making the tissue more efficient in carrying the load and resisting deformation. Other effects may materialize in altering the fibril diameter and spacing, which would induce changes in the biomechanical contribution of fibrils. Earlier studies have demonstrated that the CXL treatment increases the diameter of corneal collagen fibrils,22 which is regulated by the protein core on the keratan sulfate proteoglycan,30 but this apparent increase could be an artefact caused by tissue swelling.31 Although our study showed consistent increases in fibril diameter with CXL, these increases were not significant.

On the other hand, the effect of riboflavin and UVA on interfibrillar spacing remains controversial, with reports showing increases,32 decreases,33 or no significant changes.34 In the cases where increased spacing was observed, this may be related to postoperative edema of the corneas31,32 in the early stage. Because of the different hydration states of the cornea over the different stages of preparation and treatment, the CCT increased after epithelium removal, reduced slightly at 5 minutes after CXL while remaining thicker than the initial state, and then reduced toward the initial values 1 week after treatment (except in the 3mW/30min and 9mW/10min groups). These findings are similar to earlier trends.35 The effect of CXL is not homogeneous among different layers of the stroma, whereby lower collagen interfibrillar spacing was observed in only the anterior layer.34 In the current study, adjusting the interfibrillar spacing in the anterior 50 μm of the stroma for the changes in CCT before and after CXL showed a significant positive correlation between irradiance and interfibrillar spacing (R2 = 0.480, P < .001). With lower interfibrillar spacing, there would be an increase in the density of corneal collagen, which would in turn be expected to make the cornea stiffer.33

Keratoconus is commonly characterized by corneal thinning and biomechanical instability, both of which are related to reduced activity of protease inhibitors and increased activity of proteinase enzymes.36 An increased resistance of corneal stroma to enzymatic digestion after standard CXL was reported in Spoerl et al.'s study37 and confirmed by others.34 Increased resistance to proteinase digestion is considered an important factor for CXL to prevent keratoconus progression.37 Furthermore, the dry weights of the corneal specimen were significantly different between standard and accelerated CXL: 3 mW/30min group > 9 mW/10min group > 18 mW/5min group. The phenomenon described above38 can be used to explain the study finding that the stiffening effect of CXL and the interfibrillar spacing in the anterior cornea both decreased significantly with high irradiation/short duration settings.

Oxygen diffusion flux and local oxygen uptake influence the oxygen levels in the cornea, and the transformation of oxygen into reactive oxygen species during CXL makes corneal oxygen levels decrease.39 The CXL process with riboflavin and UVA is oxygen dependent,40 and therefore lower or higher oxygen availability would be expected to decrease or increase the overall efficacy of CXL, respectively.41 Further, oxygen conversion to free radicals may outpace oxygen replenishment by diffusion in high UVA irradiances, which may explain the reduced effectiveness of CXL in cases with high irradiances and short durations.15,39 Our findings support this hypothesis and agree with previously reported data.15

Some study limitations should be noted, including restricting the ultrastructural analysis to the anterior 50 μm of the stroma and the follow-up period to 1 week after CXL. Overall, the study demonstrated a gradual and consistent reduction in CXL effectiveness in increasing tissue stiffness in cases with shorter irradiation duration, even though the total amount of energy delivered to the cornea remained constant. The increases in tissue stiffness (as measured by the tangent modulus) were significant only with 3mW/30min and 9mW/10min combinations, and shorter durations resulted in non-significant changes in stiffness. These trends indicate that the Bunsen–Roscoe law may not be applicable in the CXL of corneal tissue, possibly due to the replenishment of oxygen levels experienced in cases with high irradiance.

References

  1. Spoerl E, Huhle M, Kasper M, Seiler T. Increased rigidity of the cornea caused by intrastromal cross-linking [article in German]. Ophthalmologe. 1997;94:902–906.
  2. Wollensak G, Spoerl E, Seiler T. Stress-strain measurements of human and porcine corneas after riboflavin-ultraviolet-A-induced cross-linking. J Cataract Refract Surg. 2003;29:1780–1785. doi:10.1016/S0886-3350(03)00407-3 [CrossRef]
  3. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-a-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135:620–627. doi:10.1016/S0002-9394(02)02220-1 [CrossRef]
  4. Wollensak G. Crosslinking treatment of progressive keratoconus: new hope. Curr Opin Ophthalmol. 2006;17:356–360. doi:10.1097/01.icu.0000233954.86723.25 [CrossRef]
  5. Caporossi A, Mazzotta C, Baiocchi S, Caporossi T. Long-term results of riboflavin ultraviolet a corneal collagen cross-linking for keratoconus in Italy: the Siena eye cross study. Am J Ophthalmol. 2010;149:585–593. doi:10.1016/j.ajo.2009.10.021 [CrossRef]
  6. Gomes JA, Tan D, Rapuano CJ, et al. Group of Panelists for the Global Delphi Panel of Keratoconus and Ectatic Diseases. Global consensus on keratoconus and ectatic diseases. Cornea. 2015;34:359–369. doi:10.1097/ICO.0000000000000408 [CrossRef]
  7. Spadea L. Corneal collagen cross-linking with riboflavin and UVA irradiation in pellucid marginal degeneration. J Refract Surg. 2010;26:375–377. doi:10.3928/1081597X-20100114-03 [CrossRef]
  8. Richoz O, Mavrakanas N, Pajic B, Hafezi F. Corneal collagen cross-linking for ectasia after LASIK and photorefractive keratectomy: long-term results. Ophthalmology. 2013;120:1354–1359. doi:10.1016/j.ophtha.2012.12.027 [CrossRef]
  9. Gore DM, O'Brart DP, French P, Dunsby C, Allan BD. A comparison of different corneal iontophoresis protocols for promoting transepithelial riboflavin penetration. Invest Ophthalmol Vis Sci. 2015;56:7908–7914. doi:10.1167/iovs.15-17569 [CrossRef]
  10. Spoerl E, Mrochen M, Sliney D, Trokel S, Seiler T. Safety of UVA-riboflavin cross-linking of the cornea. Cornea. 2007;26:385–389. doi:10.1097/ICO.0b013e3180334f78 [CrossRef]
  11. Wollensak G, Aurich H, Pham DT, Wirbelauer C. Hydration behavior of porcine cornea crosslinked with riboflavin and ultraviolet A. J Cataract Refract Surg. 2007;33:516–521. doi:10.1016/j.jcrs.2006.11.015 [CrossRef]
  12. Wollensak G, Iomdina E. Long-term biomechanical properties of rabbit cornea after photodynamic collagen crosslinking. Acta Ophthalmol. 2009;87:48–51. doi:10.1111/j.1755-3768.2008.01190.x [CrossRef]
  13. Medeiros CS, Giacomin NT, Bueno RL, Ghanem RC, Moraes HV Jr, Santhiago MR. Accelerated corneal collagen crosslinking: technique, efficacy, safety, and applications. J Cataract Refract Surg. 2016;42:1826–1835. doi:10.1016/j.jcrs.2016.11.028 [CrossRef]
  14. Wernli J, Schumacher S, Spoerl E, Mrochen M. The efficacy of corneal cross-linking shows a sudden decrease with very high intensity UV light and short treatment time. Invest Ophthalmol Vis Sci. 2013;54:1176–1180. doi:10.1167/iovs.12-11409 [CrossRef]
  15. Hammer A, Richoz O, Arba Mosquera S, Tabibian D, Hoogewoud F, Hafezi F. Corneal biomechanical properties at different corneal cross-linking (CXL) irradiances. Invest Ophthalmol Vis Sci. 2014;55:2881–2884. doi:10.1167/iovs.13-13748 [CrossRef]
  16. Elsheikh A, Anderson K. Comparative study of corneal strip extensometry and inflation tests. J R Soc Interface. 2005;2:177–185. doi:10.1098/rsif.2005.0034 [CrossRef]
  17. Bao F, Deng M, Wang Q, et al. Evaluation of the relationship of corneal biomechanical metrics with physical intraocular pressure and central corneal thickness in ex vivo rabbit eye globes. Exp Eye Res. 2015;137:11–17. doi:10.1016/j.exer.2015.05.018 [CrossRef]
  18. Bao F, Deng M, Zheng X, et al. Effects of diabetes mellitus on biomechanical properties of the rabbit cornea. Exp Eye Res. 2017;161:82–88. doi:10.1016/j.exer.2017.05.015 [CrossRef]
  19. Dhaliwal DK, Romanowski EG, Yates KA, et al. Valacyclovir inhibition of recovery of ocular herpes simplex virus type 1 after experimental reactivation by laser in situ keratomileusis. J Cataract Refract Surg. 2001;27:1288–1293. doi:10.1016/S0886-3350(01)00804-5 [CrossRef]
  20. Seifert J, Hammer CM, Rheinlaender J, et al. Distribution of Young's modulus in porcine corneas after riboflavin/UVA-induced collagen cross-linking as measured by atomic force microscopy. PLoS One. 2014;9:e88186. doi:10.1371/journal.pone.0088186 [CrossRef]
  21. Kymionis GD, Tsoulnaras KI, Liakopoulos DA, et al. Corneal stromal demarcation line determined with anterior segment optical coherence tomography following a very high intensity corneal collagen cross-linking protocol. Cornea. 2015;34:664–667. doi:10.1097/ICO.0000000000000427 [CrossRef]
  22. Wollensak G, Wilsch M, Spoerl E, Seiler T. Collagen fiber diameter in the rabbit cornea after collagen crosslinking by riboflavin/UVA. Cornea. 2004;23:503–507. doi:10.1097/01.ico.0000105827.85025.7f [CrossRef]
  23. Schumacher S, Oeftiger L, Mrochen M. Equivalence of biomechanical changes induced by rapid and standard corneal cross-linking, using riboflavin and ultraviolet radiation. Invest Ophthalmol Vis Sci. 2011;52:9048–9052. doi:10.1167/iovs.11-7818 [CrossRef]
  24. Ng AL, Kwok PS, Wu RT, Jhanji V, Woo VC, Chan TC. Comparison of the demarcation line on ASOCT after simultaneous LASIK and different protocols of accelerated collagen crosslinking: a bilateral eye randomized study. Cornea. 2017;36:74–77. doi:10.1097/ICO.0000000000001012 [CrossRef]
  25. Schumacher S, Mrochen M, Spoerl E. Absorption of UV-light by riboflavin solutions with different concentration. J Refract Surg. 2012;28:91–92.
  26. Caruso C, Epstein RL, Ostacolo C, Pacente L, Troisi S, Barbaro G. Customized corneal cross-linking: a mathematical model. Cornea. 2017;36:600–604. doi:10.1097/ICO.0000000000001160 [CrossRef]
  27. Anderson K, El-Sheikh A, Newson T. Application of structural analysis to the mechanical behaviour of the cornea. J R Soc Interface. 2004;1:3–15. doi:10.1098/rsif.2004.0002 [CrossRef]
  28. Akhtar S, Bron AJ, Salvi SM, Hawksworth NR, Tuft SJ, Meek KM. Ultrastructural analysis of collagen fibrils and proteoglycans in keratoconus. Acta Ophthalmol. 2008;86:764–772. doi:10.1111/j.1755-3768.2007.01142.x [CrossRef]
  29. Turkcu UO, Yuksel N, Novruzlu S, Yalinbas D, Bilgihan A, Bilgihan K. Protein oxidation levels after different corneal collagen cross-linking methods. Cornea. 2016;35:388–391. doi:10.1097/ICO.0000000000000731 [CrossRef]
  30. Musselmann K, Kane B, Alexandrou B, Hassell JR. Stimulation of collagen synthesis by insulin and proteoglycan accumulation by ascorbate in bovine keratocytes in vitro. Invest Ophthalmol Vis Sci. 2006;47:5260–5266. doi:10.1167/iovs.06-0612 [CrossRef]
  31. Hayes S, Boote C, Kamma-Lorger CS, et al. Riboflavin/UVA collagen cross-linking-induced changes in normal and keratoconus corneal stroma. PLoS One. 2011;6:e22405. doi:10.1371/journal.pone.0022405 [CrossRef]
  32. Akhtar S, Almubrad T, Paladini I, Mencucci R. Keratoconus corneal architecture after riboflavin/ultraviolet A cross-linking: ultrastructural studies. Mol Vis. 2013;19:1526–1537.
  33. Choi S, Shin JH, Cheong Y, Jin KH, Park HK. Structural and biomechanical effects of photooxidative collagen cross-linking with photosensitizer riboflavin and 370 nm UVA light on human corneoscleral tissues. Microsc Microanal. 2013;19:1334–1340. doi:10.1017/S1431927613001669 [CrossRef]
  34. Hayes S, Kamma-Lorger CS, Boote C, et al. The effect of riboflavin/UVA collagen cross-linking therapy on the structure and hydrodynamic behaviour of the ungulate and rabbit corneal stroma. PLoS One. 2013;8:e52860. doi:10.1371/journal.pone.0052860 [CrossRef]
  35. Hovakimyan M, Guthoff R, Knappe S, et al. Short-term corneal response to cross-linking in rabbit eyes assessed by in vivo confocal laser scanning microscopy and histology. Cornea. 2011;30:196–203. doi:10.1097/ICO.0b013e3181e16d93 [CrossRef]
  36. Zhou L, Sawaguchi S, Twining SS, Sugar J, Feder RS, Yue BY. Expression of degradative enzymes and protease inhibitors in corneas with keratoconus. Invest Ophthalmol Vis Sci. 1998;39:1117–1124.
  37. Spoerl E, Wollensak G, Seiler T. Increased resistance of crosslinked cornea against enzymatic digestion. Curr Eye Res. 2004;29:35–40. doi:10.1080/02713680490513182 [CrossRef]
  38. Aldahlawi NH, Hayes S, O'Brart DP, Meek KM. Standard versus accelerated riboflavin-ultraviolet corneal collagen crosslinking: resistance against enzymatic digestion. J Cataract Refract Surg. 2015;41:1989–1996. doi:10.1016/j.jcrs.2015.10.004 [CrossRef]
  39. Kamaev P, Friedman MD, Sherr E, Muller D. Photochemical kinetics of corneal cross-linking with riboflavin. Invest Ophthalmol Vis Sci. 2012;53:2360–2367. doi:10.1167/iovs.11-9385 [CrossRef]
  40. Richoz O, Hammer A, Tabibian D, Gatzioufas Z, Hafezi F. The biomechanical effect of corneal collagen cross-linking (CXL) with riboflavin and UV-A is oxygen dependent. Trans Vis Sci Technol. 2013;2:6. doi:10.1167/tvst.2.7.6 [CrossRef]
  41. Kling S, Richoz O, Hammer A, et al. Increased biomechanical efficacy of corneal cross-linking in thin corneas due to higher oxygen availability. J Refract Surg. 2015;31:840–846. doi:10.3928/1081597X-20151111-08 [CrossRef]

Corneal Cross-linking Settings Adopted in Different Specimen Groups

GroupNo.Irradiation
3mW/30min183 mW/cm2 for 30 minutes
9mW/10min189 mW/cm2 for 10 minutes
18mW/5min1818 mW/cm2 for 5 minutes
30mW/3min1830 mW/cm2 for 3 minutes
45mW/2min1845 mW/cm2 for 2 minutes
90mW/1min1890 mW/cm2 for 1 minute
NUVA group18Unirradiated but underwent epithelium removal and riboflavin instillation

Mean ± Standard Deviation of Constitutive Parameters α and μ in All Specimen Groups

GroupαμRMS (mm)
3mW/30min group0.0328 ± 0.0461119.2957 ± 59.41270.0102 ± 0.02
9mW/10min group0.0156 ± 0.021104.3486 ± 42.41150.0034 ± 0.0066
18mW/5min group0.0164 ± 0.021105.7002 ± 60.41810.0035 ± 0.0056
30mW/3min group0.0104 ± 0.007881.2774 ± 41.95640.0074 ± 0.0105
45mW/2min group0.0093 ± 0.010474.0538 ± 17.66330.0039 ± 0.0038
90mW/1min group0.0097 ± 0.006167.3235 ± 16.86130.0057 ± 0.0102
NUVA group0.0107 ± 0.004159.0354 ± 21.47180.0213 ± 0.0614

Mean ± Standard Deviation of CCT in All Specimen Groups

GroupCCT-initial (μm)CCT-preCXL (μm)CCT-posCXL5min (μm)CCT-posCXL1week (μm)
3mW/30min group366.3 ± 16.6486.0 ± 68.1447.2 ± 49.0411.7 ± 57.8
9mW/10min group369.8 ± 15.8504.6 ± 79.3468.2 ± 70.6416.1 ± 59.6
18mW/5min group364.3 ± 29.4523.1 ± 57.8477.9 ± 53.5380.2 ± 53.0
30mW/3min group354.2 ± 22.7494.7 ± 61.6446.7 ± 55.3385.7 ± 70.1
45mW/2min group371.8 ± 25.5501.3 ± 49.6461.2 ± 51.2380.6 ± 45.9
90mW/1min group362.6 ± 29.3521.3 ± 39.3499.1 ± 37.6396.1 ± 56.7
NUVA group349.3 ± 23.2491.1 ± 48.7365.2 ± 40.3

Mean ± Standard Deviation of Interfibrillar Spacing and Fibril Diameter in Anterior 50 μm of Corneal Stroma in Different Specimen Groups

GroupNo.Fibril Diameter (nm)Interfibrillar Spacing (nm)
3mW/30min626.56 ± 1.3913.06 ± 2.07
9mW/10min626.33 ± 1.1314.37 ± 1.9
18mW/5min626.49 ± 1.2216.15 ± 1.78
30mW/3min626.92 ± 1.2117.68 ± 2.14
45mW/2min626.22 ± 0.6818.26 ± 2.61
90mW/1min626.85 ± 0.4519.68 ± 1.8
NUVA626.05 ± 0.6920.05 ± 1.89
Authors

From the Eye Hospital (FB, YZheng, CL, XZ, YZhao, YW, LL, QW, SC) and the Institution of Ocular Biomechanics (FB, YZheng, XZ, QW, SC), Wenzhou Medical University, Wenzhou, Zhejiang Province, China; the School of Engineering, University of Liverpool, Liverpool, United Kingdom (AE); and the National Institute for Health Research (NIHR) Biomedical Research Centre at Moorfields Eye Hospital, NHS Foundation Trust, UCL Institute of Ophthalmology, London, United Kingdom (AE).

The authors have no financial or proprietary interest in the materials presented herein.

Drs. Bao and YaRu Zheng contributed equally to this work and should be considered as equal first authors.

Supported by the Natural Science Foundation of Zhejiang Province (LY16H120005, LY18A020008), the National Natural Science Foundation of China (81600712, 31771020), the Science and Technology Plan Project of Wenzhou Science and Technology Bureau (Y20170198), the Scientific Research Project of Zhejiang Provincial Department of Education (Y201534199), and the Projects of Medical and Health Technology Development Program in ZheJiang Province (2016ZHB012, 2018RC057).

AUTHOR CONTRIBUTIONS

Study concept and design (FB, QW, SC, AE); data collection (YZheng, CL, XZ, YZhao, YW, LL); analysis and interpretation of data (FB, YZheng); writing the manuscript (FB, YZheng, CL, XZ, YZhao, YW, LL); critical revision of the manuscript (FB, QW, SC, AE); statistical expertise (FB, YZheng, CL, XZ, YZhao, LL); obtaining funding (FB); administrative, technical, or material support (FB, SC); supervision (QW, SC, AE)

Correspondence: ShiHao Chen, MD, OD, 270 Xueyuan West Road, Wenzhou City, Zhejiang Providence, 325027, China. E-mail: chenle@rocketmail.com

Received: May 01, 2017
Accepted: October 19, 2017

10.3928/1081597X-20171025-01

Sign up to receive

Journal E-contents