The Efficacy of Corneal Cross-Linking Shows a Sudden Decrease with Very High Intensity UV Light and Short Treatment Time

Abstract

Purpose.: Standard treatment in cases of progressive keratectasia is UV-triggered corneal cross-linking. For irradiances larger than 10 mW/cm² and treatment times below 10 minutes, the scientific proof of a biomechanical strengthening effect is insufficient. The authors investigated the biomechanical strengthening of ex vivo corneal tissue treated with irradiances between 3 mW/cm² and 90 mW/cm² and illumination times from 30 minutes to 1 minute, respectively.

Methods.: A total of 100 porcine eyes received riboflavin + UV treatment (constant irradiation dose of 5.4 J/cm²) with different intensities and illumination times and were randomly assigned into 10 groups. A control group (80 eyes) was not irradiated but underwent the same treatment otherwise. Young's modulus at 10% strain was determined for each strip after uniaxial stress-strain measurement. A Kruskal-Wallis test was used for statistical analysis.

Results.: A statistically significant difference (α = 0.01) was found between the median value of Young's modulus of the treatment groups up to 45 mW/cm² (illumination times from 30 minutes to 2 minutes) compared with the control group. There was no statistically significant difference between the treatment groups from 50 mW/cm² up to 90 mW/cm² (illumination times of less than 2 minutes) and the control group.

Conclusions.: The ex vivo results of corneal cross-linking performed in porcine corneas show that the Bunsen-Roscoe reciprocity law is only valid for illumination intensities up to 40 to 50 mW/cm² and illumination times of more than 2 minutes. Further experiments are necessary to validate these results for in vivo human corneal tissue. Additionally, safety aspects at high intensities must be investigated.

Introduction

Progressive keratectasia is either caused by progressive corneal disease ¹ or a sequela of refractive surgery. ^2,3 The standard treatment procedure before exploiting a fatal surgery like a keratoplasty is ultraviolet light–triggered corneal cross-linking using a riboflavin solution as a photosensitizer. ^4,5  

Corneal cross-linking was introduced by Seiler and Spoerl in 1997, ⁶ and a standardized protocol has been developed with time ⁷ : An abrasion of the corneal epithelium is performed, followed by riboflavin application (30 minutes, one drop every 2–5 minutes) as a photosensitizer. During the subsequent treatment phase, a recommended illumination intensity of 3 mW/cm² is applied to a 9-mm zone in the cornea for 30 minutes. This corresponds to a total energy dose of 3.4 J or a radiant exposure of 5.4 J/cm². The photochemical process that induces the additional cross-links between the corneal fibers is thereby dependent on the applied radiant exposure of UV light. A plurality of clinical studies following the standard protocol show that the progression of the ectasia can be halted. ^5,8–15  

One major disadvantage of the CXL procedure so far is the long total treatment time of 1 hour including a soaking time of 30 minutes for the riboflavin solution and an illumination time of 30 minutes for the UV light. Therefore, to address the patient's wish for increased comfort, and the surgeon's need for a higher throughput of patients, a shorter CXL procedure would be desirable. 

There are aims to shorten the soaking time by using a different protocol to apply the riboflavin and to shorten the illumination time by increasing the illumination intensity. 

By increasing the intensity, it is assumed that the Bunsen-Roscoe law of reciprocity is valid for the corneal cross-linking effect, having a constant energy dose of 3.4 J or a radiant exposure of 5.4 J/cm². It has been shown in ex vivo experiments that the biomechanical stiffening effect of the corneal tissue is equivalent with 10 mW/cm² (illumination time 9 minutes) to the standard protocol. ¹⁶  

However, it is known from photography that the Bunsen-Roscoe law is only valid for a certain range ¹⁷ and so far, it is not known how large this range is for corneal cross-linking. This means that the corneal cross-linking effect may be dependent on a threshold and the increase in illumination intensity (decrease in illumination time) is limited. So far, for illumination intensities larger than 10 mW/cm², the scientific proof of a biomechanical strengthening effect is insufficient and thus the aim of this work is to investigate the biomechanical strengthening of ex vivo corneal tissue, which is illuminated by intensities as large as 90 mW/cm² for a constant energy dose of 5.4 J/cm². 

Methods

Experimental Study Design

To determine the efficacy of corneal cross-linking for higher intensities, the change in corneal stiffness that is evoked by the cross-linking treatment of ex vivo porcine corneal tissue was investigated. The elaborated study design is in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. As samples for the biomechanical measurements, one test strip was cut out of each cornea, which was randomly assigned to one of the groups with different treatment intensities shown in Table 1. The intensities varied from 3 to 90 mW/cm², with a constant energy dose of 5.4 J/cm² and corresponding illumination times ranging from 30 minutes to 1 minute. To each of these groups control eyes of the same date were tested, which were not irradiated with UV light, but otherwise underwent the same procedure including a soaking with Riboflavin solution. The experiment involved two main steps: Cross-linking procedure and stress-strain measurements. 

Table 1. 

 

Treatment Parameters of the Different Groups

Treatment Group	Irradiation Intensity, mW/cm2	Irradiation Time, min	Irradiation Dose, J/cm2	No. of Eyes Used for Statistical Analysis
1	3	30	5.4	10
2	6	15	5.4	10
3	9	10	5.4	10
4	20	4.50	5.4	9
5	34	2.65	5.4	10
6	45	2.00	5.4	10
7	50	1.80	5.4	9
8	60	1.50	5.4	8
9	67	1.35	5.4	8
10	90	1.00	5.4	7
Controls	–	–	–	65

Cross-Linking Procedure

A total of 180 porcine eyes were enucleated postmortem at a local abattoir and prepared for the experiments within 8 hours after the pigs were killed. All corneas were undamaged, undenaturated, clear, and possessed no visual autolytic changes, haze, or scratches. After epithelium abrasion, the eye bulb was placed upside down in a small contact lens container filled with a few drops of a 0.1% riboflavin solution (Innocross R^M Isotonic; IROC Innocross AG, Zurich, Switzerland) for 30 minutes. 

Ten eyes of a series were randomly assigned to the control group, which was not irradiated with UV light. Another 10 eyes underwent the cross-linking treatment, which was performed with a modified UV lamp (IROC Innocross AG), providing the intensities and the corresponding treatment times shown in Table 1. Treatment groups 1 to 3 were pooled in one series of measurement and therefore have one common control group of 10 eyes. 

The intensity was identical for all treated eyes of one series and only changed, according to Table 1, from one series to another. Therefore, radiant exposure was kept constant at 5.4 J/cm during all experiments. The UV lamp was calibrated with a power meter (LaserMate-Q; Coherent, Santa Clara, CA) to the respective intensities prior to the experiments. For the treatment, the UV lamp was placed over the eye bulb, ensuring full radiation of the cornea. 

Meanwhile, the eyes of the control group were kept moist under a glass dome containing a glass of water. 

Biomechanical Measurements

The time between the beginning of corneal cross-linking and biomechanical measurements was 30 minutes and resulted from the sample treatment time and a corresponding waiting time. Subsequently, the corneal tissue and surrounding parts of the limbus were extracted from the eye bulb by means of a circular cutter and its thickness was determined by a pachymeter measurement (Pach-Pen XL; Mentor, Norwell, MA). The extracted cornea was placed on a convex cutting underlay (Fig. 1a). With a two-blade knife, a strip was cut out of each cornea along the superior-inferior axis, measuring 5 mm in width (Fig. 1b). 

Figure 1. 

 

Preparation for biomechanical measurements. (a) Extracted cornea placed on a convex cutting underlay. (b) Preparation of the cut strip. (c) 5-mm wide corneal sample held by the clamps of the testing machine.

Corneal stiffness was determined by uniaxial stress-strain measurements using a material testing machine (MINIMAT; Polymer Laboratories, Stretton Shropshire, UK). After proper alignment in the testing machine with a clamp-to-clamp distance of 7 mm, the sample was fixed by tightening two screws (Fig. 1c). The use of a torque wrench ensured consistent clamping forces. 

The biomechanical measurement of each sample was carried out as follows: First, the two clamps of the testing machine were moved closer together to ensure complete relaxation of the sample tissue. Then, the clamps were moved apart until a preload of 20 mN was reached. At this stage, initial length L₀ of the sample was recorded as reference for the stress-strain curve. Afterwards, the actual measurement was performed, thereby the sample was stretched with a velocity of 2 mm/min up to a maximum force of 10 N. During the measurement, the load curve was automatically recorded up to 18% strain and loads were converted to stress by dividing it by the cross-sectional area (i.e., the sample's width times its thickness). Figure 2 shows a recorded stress-strain curve. 

Figure 2. 

 

Example of an obtained stress-strain curve, which shows the typical behavior of biological tissue: A toe region of nonlinear elasticity at low strains and a linear elastic behaviour at medium strains. A precipitous fall at high strains is missing as the maximum force of 10 N was not exceeded and therefore the tissue was not torn apart.

The stiffness (Young's modulus) as a derivative of the stress-strain curve was determined. For the following statistical analysis, Young's modulus was consistently evaluated at 10% strain. 

Statistical Analysis

The eyes within the control group were found not to be normally distributed. Therefore, a Kruskal-Wallis test was performed on all eyes in the control group on different days (charges of eyes), which showed no statistically significant difference (P = 0.2417). Thus, the control eyes of all groups could be pooled in one group as they behaved equally. Each treatment group was then compared with the control group for statistical significance by means of a nonparametric Kruskal-Wallis test (with Bonferroni–Dunn post hoc test [α < 0.01]). Additionally, boxplots of Young's moduli of all groups were generated for visual comparison as well as the relative stiffness increase as a function of illumination intensity. Statistical calculations (Table 2) were performed with online statistic software (www.stattools.net, in the public domain) and visualization (Figs. 3, 4) was created with data analysis and graphing software (Origin 6; OriginLab Corporation, Northampton, MA). 

Figure 3. 

 

Young's moduli at 10% strain for the control and different treatment groups. Box plot whiskers indicate the fifth and the 95th percentiles, crosses (x) indicate the first and the 99th percentiles and dashes (–) indicate the minimum and maximum values within the groups.

Figure 4. 

 

Stiffness increase of all treatment groups compared with the control group. The second x-axis at the top indicates the corresponding irradiation times to maintain a constant energy dose of 5.4 J/cm².

Table 2. 

 

Post Hoc Analysis of the Kruskal-Wallis Test

Group A	Group B, mW/cm2	Q	Q(0.01)	Significant Difference
Control	3	4.812	3.743	Yes
Control	6	4.916	3.743	Yes
Control	9	4.832	3.743	Yes
Control	20	4.899	3.743	Yes
Control	34	4.372	3.743	Yes
Control	45	3.867	3.743	Yes
Control	50	1.345	3.743	No
Control	60	1.972	3.743	No
Control	67	0.240	3.743	No
Control	90	0.378	3.743	No

 All treatment groups (group B) were compared against the control group (group A). Q represents the calculated values that are compared with Q(0.01) from the Kruskal-Wallis test.

Results

Excluded from data analysis were 24 samples where a procedural error systematically occurred (e.g., slip of the probe in the probe holder). Thus, only 156 of the 180 samples were used for data analysis. The numbers of eyes in each group used for data analysis are listed in Table 1. 

Box plots of Young's moduli at 10% strain for the control and treatment groups are displayed in Figure 3. Young's modulus of the control group varied between 1300 kPa and 6550 kPa, with a median of 3350 kPa. Within the treatment groups, the median of the Young's modulus varied between 3500 kPa and 7800 kPa, with associated ranges displayed in Figure 3. The performed Kruskal-Wallis test (Dunn criteria) showed a significant difference between groups (P < 0.0001). Post hoc analysis (shown in Table 2) showed a statistically significant difference (α = 0.01) between the treatment groups up to 45 mW/cm² and the control group and no statistically significant difference between the treatment groups from 50 mW/cm² up to 90 mW/cm² and the control group. 

Evaluating the average increase in stiffness of all groups, which is displayed in Figure 4, one observes that the data follows a typical threshold function (Boltzmann function). The 50% limit is associated at approximately 47 mW/cm² ± 1.5 mW/cm². 

Discussion

The presented data show the dependence of increase in corneal stiffness on illumination intensity while keeping a constant irradiation dose of 5.4 J/cm². According to Figure 4, an equivalent stiffness increase can be achieved up to an illumination intensity of approximately 40 to 45 mW/cm², corresponding to illumination times of approximately 2 minutes. For higher intensities ranging from 50 mW/cm² up to 90 mW/cm², no statistically significant stiffness increase could be achieved. Of all the treatment groups that were significantly different from the control group, no significant difference among each other was found in the post hoc analysis. Therefore, we don't claim the highest value occurring at 34 mW/cm² to be an optimum. 

The reported stiffness increase for the lower intensities (3–45 mW/cm²) of a factor of 2.2 is similar to previously published data by Wollensak and colleagues, who measured a 1.8-fold increase. ⁵ Other studies investigating the stiffness increase according to the standard protocol with 3 mW/cm² used different measurement techniques like supersonic shear imaging, ¹⁸ ultrasound, ¹⁹ or comparing corneal geometry. ²⁰ They found stiffness increases of 4.6, 1.04, and 1.6, respectively. The only available study that investigated higher intensity while the irradiation dose is kept constant found an equivalence in stiffness increase due to CXL between the standard protocol (3 mW/cm² for 30 minutes) and a 10 mW/cm² for 9 minutes, ¹⁶ supporting the finding of this study for the low intensities. The absolute stiffness increase in the mentioned study was only by a factor of 1.3. The difference compared with this study might be caused by a different protocol used for the biomechanical measurements. 

The most interesting finding of this study is the failure of the Bunsen-Roscoe reciprocity law for short illumination time and high intensities. The Bunsen-Roscoe law describes the photo-response of a material to a certain energy dose. It concludes that all photochemical reaction mechanisms depend only on the total absorbed energy and are statistically independent of the two factors that determine total absorbed energy—that is, radiant intensity or irradiance, and exposure time. A review of the validity of the Bunsen-Roscoe law in biology and medicine shows that approximately 95% and over 80%, respectively, of the evaluated reactions follow the law of reciprocity. ¹⁷ The failure of the law observed in this study is probably due to the relative complex photochemistry that is not fully understood at this time. As a consequence, the corneal CXL-treatment has an upper limit for the applied illumination intensity or lower limit for the illumination time. From the current data, this limit seems to be at approximately 40 to 45 mW/cm² corresponding to an illumination time of approximately 2 minutes. In order to decrease the illumination time further and still have the same effect, maybe not only the intensity has to be increased, but also the total energy dose. For higher energy doses in corneal cross-linking, the outcome is reported controversially. In a study by Lanchares et al.—who doubled the energy dose to 10.6 J/cm² by applying 3 mW/cm² for 60 minutes—no stiffening effect at all was achieved. ²¹ Thus, it seems that the induced cross-links within the first 30 minutes are somehow destroyed again by the UV light in the additional 30 minutes of illumination. However, Xiu et al. increased the dose by applying 4.2 mW/cm² for 30 minutes and measured an increase in stiffness of 1.45. ²²  

In conclusion, the performed investigations of corneal cross-linking in ex vivo tissue show that the Bunsen-Roscoe reciprocity law is only valid for illumination intensities up to approximately 40 to 45 mW/cm² with illumination times of more than 2 minutes. At higher intensities, the achieved stiffness increase is not significant anymore. In order to clarify the validity of these results for in vivo human corneal tissue, further experiments are necessary. Additionally, safety aspects at high intensities must be investigated. Particularly, the susceptibility of substructures of the cornea—such as the endothelium—must be considered. If transferability of our results and safety could be assured in the end, a new rapid corneal cross-linking procedure could be introduced in clinical practice. The shorter treatment time could increase both the patient's comfort and the doctor's patient throughput. 

Other aspects that should be investigated in future studies are the influences of different concentrations of riboflavin and/or different irradiation energy doses. Potential effects of higher riboflavin concentrations could be higher protection of corneal endothelium and lens epithelium from UVA damage, and greater ability to allow tissue strengthening at irradiation intensities above 50 mW/cm² and irradiation times of only 1 to 2 minutes. 

Footnotes

 Disclosure: J. Wernli, IROC Innocross AG (F, C); S. Schumacher, IROC Innocross AG (F, C); E. Spoerl, None; M. Mrochen, IROC Innocross AG (I)