Introduction
Corneal collagen crosslinking (CXL) is an effective treatment modality for progressive keratoconus and other corneal diseases [
1]. The corneal stroma is primarily comprised of regularly arranged collagen fibers and their interconnections. In patients with keratoconus, the interconnections are impaired, decreasing corneal mechanical strength. Wollensak et al. [
2] demonstrated for the first time that CXL increases corneal mechanical strength in keratoconus. This procedure involves photochemical reactions on the surface of the collagen fibers and in the protein networks surrounding the collagen [
3]. In addition, CXL increases collagen fiber diameters and improves resistance of the corneal stroma against multiple degrading enzymes.
These findings support the use of photochemical treatments in CXL to increase corneal strength and prevent or delay keratoconus progression. The most commonly used method in clinical practice is the standard Dresden protocol for corneal crosslinking (S-CXL). However, S-CXL is time-consuming, and although prior studies report comparable efficacies of accelerated CXL (A-CXL) and S-CXL [
4,
5], some clinical centers do not currently use conventional S-CXL to treat keratoconus. Moreover, removal of the corneal epithelium may cause pain, discomfort, temporary vision reduction, and decreased corneal clarity (haze), and can increase the risk of infectious keratitis [
6‐
8]. Many patients cannot undergo epithelium-off (epi-off) surgery due to insufficient corneal thickness (< 400 μm). Nevertheless, retaining the corneal epithelium (epi-on surgery) could decrease the efficacy of CXL as the epithelium can impede the penetration of ultraviolet (UV) radiation and riboflavin. Recently, multiple studies have reported means to improve CXL treatment procedures, such as immersing riboflavin into the stroma by various means, increasing UV irradiation energy and using an excimer laser rather than mechanics to remove the epithelium. Few randomized controlled trials have evaluated the relative efficacies of multiple CXL modalities. In available trials, the sample size or the treatment protocol types are limited, and inclusion criteria, such as ethnographics, age group, and baseline data, vary [
9], precluding definitive recommendations of CXL procedure types for patients with progressive keratoconus. For example, previous studies have demonstrated that pediatric keratoconus is more aggressive than adult keratoconus, and age is an important influence on the efficacy of CXL [
10,
11]. A study using a generalized estimating equation (GEE) could be used to compare the efficacy of different types of CXL procedures with varying combinations of riboflavin and irradiation power for treatment of progressive keratoconus in pediatric patients, as the generalized linear model correction could analyze outcomes to exclude the effects of sex, age, baseline data heterogeneity, and bilateral surgery [
12].
Currently, according to the Global Consensus on Keratoconus and Ectatic Diseases (2015) [
13], many investigators concur that a more comprehensive evaluation system should be used to evaluate keratoconus progression, rather than simply the maximum keratometry (K
max) of the corneal anterior surface. The ABCD Grading System [
14] uses the anterior (
A) and posterior (B) radiuses of curvature at the cone area, corneal thickness at the thinnest region of the cornea (
C), and best corrected distance vision (
D). Recent studies have identified that more than half of patients demonstrate quantifiable progress using the ABCD Grading System earlier than with measurement of K
maxchange alone [
15], pointing to the utility of a combined progression system to track progress following CXL [
16].
The present study compared the independent effects of five CXL procedures for progressive keratoconus at 1 postoperative year with follow-up based on keratometry and the ABCD Grading System.
Materials and methods
Data set and study design
A retrospective medical chart review was conducted on all consecutive patients with progressive keratoconus at the Hankou Aier Eye Hospital (Wuhan, Hubei province, China) who underwent CXL treatment between July 7, 2014 and August 22, 2021. Patients who returned for a follow-up visit after 1 year were included in the study. All patients provided written informed consent prior to surgery, and surgeries were performed by the same operator (Q. Y. Zeng).
An increase of at least 1 diopter (D) in maximum keratometry (Kmax) derived from computerized corneal topography during the preceding 12 months was required for inclusion. Patients with previous refractive surgeries or corneal history of ocular surface or other eye disorders were excluded. In addition, patients whose data could not be reviewed for any reason were classified as being lost to follow-up and excluded from the study.
Surgical technique
Patients were included regardless of treatment protocols. In total, five different treatment combinations were included in the study: Accelerated Transepithelial CXL, Iontophoresis CXL for 10 min, CXL plus phototherapeutic keratectomy (CXL–plus-PTK), High-Fluence Accelerated CXL, and Accelerated CXL. When the thinnest corneal thickness of the eye was < 450 μm, patients were randomized to undergo either Accelerated Transepithelial CXL or Iontophoresis CXL. When the thinnest corneal thickness of the eye was ≥ 450 μm, the patients or their guardians elected whether to undergo the CXL-plus-PTK procedure to correct the irregularity of the epithelium. If preferred, the patients underwent the CXL-plus-PTK procedure. If the patients refused, they were randomly selected to receive the High-Fluence Accelerated CXL or Accelerated CXL procedure.
(1)
Accelerated Transepithelial CXL: In the first step, 0.25% riboflavin (Paracel Part I, Avedro Inc., USA) containing 0.02% benzalkonium chloride (BAC) and 0.85% hydroxypropyl methyl cellulose (HPMC) was applied onto the cornea every 90 s for 4 min. Thereafter, part I solution was rinsed with 0.22% riboflavin (Paracel Part II, Avedro), and part II solution was instilled every 90 s over the next 6 min. UV-A was applied using the Avedro KXL System (Avedro Inc., Waltham, USA) with 30 mW/cm2 UV power for 8 min with a 1 s on/off cycle (7.2 J/cm2).
(2)
Iontophoresis CXL: Patients underwent iontophresis for stromal imbibition, with 0.35 ml 0.1% Ricrolin + riboflavin solution (Sooft, Montegiorgio, Italy) applied to 0.8 mm of the cornea with a suction ring and delivered by an electric generator I-ON CXL (Sooft) set at 1 mA through inox electrodes for 10 min. The treated cornea was subsequently exposed to UV-A light (Vega, CSO, Firenze, Italy) for 9 min at an irradiance of 10 mW/cm2 (5.4 J/cm2).
(3)
CXL-plus-PTK: The corneal epithelium was ablated in a 7 mm zone with an intended depth of 50 μm using an excimer laser (Schwind eye-tech-solutions GmbH & Co. KG, Kleinostheim, Germany). CXL was then performed with 0.1% dextran-free riboflavin (VibeX Rapid, Avedro) instilled every 90 s for 10 min. Subsequntly, it was placed under UA irradiation for 4 min at 30 mW/cm2 (7.2 J/cm2, Avedro).
(4)
High-Fluence Accelerated CXL: The corneal epithelium was removed with a blunt knife in a 10 mm zone. Riboflavin and UV irradiation were used as described in the CXL-plus-PTK group (7.2 J/cm2).
(5)
Accelerated CXL: The epithelium was also removed by a blunt knife in a 10 mm zone. Riboflavin used for A-CXL was comprised of riboflavin 0.1% and dextran 20.0% (Ricrolin, Sooft). UV irradiation was used as described in the I-CXL group (5.4 J/cm2).
The operator verified irradiance prior to each treatment. The five CXL procedures are summarized in Table
1.
Table 1
Five Crosslinking Treatment Procedures
Fluence (total)(J/cm2) | 7.2 | 5.4 | 7.2 | 7.2 | 5.4 |
Soak time and interval (minutes) | 10 (1.5) | 10 | 10 (1.5) | 10 (1.5) | 30 (2) |
Intensity(mW/cm2) | 30 | 9 | 30 | 30 | 10 |
Treatment time(minutes) | 8 | 10 | 4 | 4 | 9 |
Irradiation mode (interval) | Pulsed(1 sec on-1 sec off) | Continuous | Continuous | Continuous | Continuous |
Epithelium status | On | On | Off | Off | Off |
De-epithelialization method | / | / | PTK | Mechanical | Mechanical |
Riboflavin | ParaCel | Ricrolin+ | Vibex Rapid | Vibex Rapid | Ricrolin |
Riboflavin osmolarity | Iso- | Iso- | Iso- | Iso- | Iso- |
Light source | KXL | VEGA | KXL | KXL | VEGA |
Pain medication and postoperative care
All patients received 0.5% levofloxacin drops four times daily for 3 days prior to surgery. Thirty minutes before surgery, patients received 2% pilocarpine (Sigma-Aldrich, St. Louis, MO, USA) and 0.4% oxybuprocaine hydrochloride (Bausch & Lomb Pty Ltd., NSW, Australia) drops three times, with 5 min between administrations.
At the end of the surgery, the corneal surface was dressed with a therapeutic soft contact lens (Bausch & Lomb Pty Ltd.) for at least 24 h until the epithelium was completely healed.
Assessments
Contact lens wearers were instructed to discontinue use for a minimum of 3 weeks prior to the preoperative eye examination. During the baseline visit and postoperative visits at 1, 3, 6, and 12 months, the following assessments were performed: uncorrected distance visual acuity (UDVA; in logarithm of the minimum angle of resolution [LogMAR] units), corrected distance visual acuity (CDVA; in LogMAR units), and spherical equivalence (SE); mean keratometry (Kmean), maximum keratometry (Kmax), minimum corneal thickness (MCT), A (staging index for ARC; ARC = anterior radius of curvature), B (staging index for PRC, PRC = posterior radius of curvature), and C (staging index for MCT, measured with an Oculus Pentacam; Oculus, Wetzlar, Germany). Confocal microscopy was performed using an HRT3 microscope (Heidelberg Engineering, Heidelberg, Germany) to measure endothelial cell density (ECD). Adverse events were defined as any documented medical complication, including keratitis or corneal ulceration, or retreatment at any time point within 1 year after CXL.
Statistical analysis
A multivariable linear regression model with GEE correction was used to correct for cases in which a patient underwent bilateral CXL. The Accelerated CXL group was the reference group, and the other four procedures were compared against this group. Further, to eliminate the effect of age on the determined variable, patients were divided into two groups depending on age (< 18 years or ≥ 18 years). All secondary outcomes were analyzed using GEE correction.
Baseline measurement normality was assessed using Q-Q plots. Categorical variables were presented as numbers and percentages. A one-way analysis of variance was used to analyze differences in baseline characteristics among the five groups. To account for multiple comparisons, a Gabriel post hoc test was performed. A paired Student’s t-test statistical analysis was performed to compare the change of clinical parameters between the baseline and 1 year follow-up time points in each group. Variables were assessed for multicollinearity. All statistical analyses were performed with SPSS version 25.0 (IBM, Armonk, NY, USA). A P-value < 0.05 was considered statistically significant.
Discussion
To our knowledge, this is the first retrospective longitudinal study to compare the independent effects of different CXL procedures in treatment of keratoconus based on keratometry and the ABCD Grading System. All five procedure types stabilized disease, and improved visual acuity and keratometry values at 1 year postoperatively, according to findings without GEE corrections (Fig.
1 and Supplemental Digital Content
1). Secondary analysis revealed that no treatment procedure resulted in significant differences in UDVA, CDVA, and SE compared with the
Accelerated CXLgroup. These findings were consistent with some previous reports [
18‐
20]. Contrastingly, some studies have reported that different CXL procedures result in significantly improved visual acuity [
21,
22]. All of these results are influenced by interaction effects of various factors. Therefore, we used the GEE to correct for the effects of age, baseline data heterogeneity, and bilateral surgery on treatment outcomes.
The UV-A irradiation energy used for the
Accelerated Transepithelial CXL procedure was 7.2 J/cm
2, which was higher than that of the
Accelerated CXL procedure (5.4 J/cm
2). It was less effective, as defined by K
mean, K
max, and Avalue at 1 year postoperatively. This could be because the corneal epithelium block the penetration of riboflavin and ultraviolet irradiation. Increasing UV-A irradiation energy could potentially improve the efficacy of CXL treatment, and benzalkonium chloride could disrupt the tight junctions of the corneal epithelium [
23]. A prior study used the conventional UV-A energy (5.4 J/cm
2) for
Accelerated Transepithelial CXLand reported the same trend of kerametory changes not only at 1 year but also at 6, 18, and 24 months postoperatively [
24]. The
Accelerated Transepithelial CXL procedure had the lowest efficacy among the five CXL procedures 1 year postoperatively, as measured by K
mean and K
max values.
Our findings suggested that the
CXL-plus-PTK procedure was the most effective based on reduction in K
maxvalue. Excimer laser ablation combined with CXL treatment has been proposed as an ideal technique due to its optimal refractive outcome that can provide both stability and functional vision, by both stabilizing and reshaping the cornea [
25]. Previous studies also suggested that the
CXL-plus-PTKprocedure is more effective than the S-CXL procedure [
26,
27]. Further, the improved efficacy could last up to 2 years after surgery [
28]. Ozge et al. reported no obvious differences between the
CXL-plus-PTK and S-CXL groups until the third postoperative year, although the corneal epithelium was removed in both procedures. One possible reason for the improved efficacy of the
CXL-plus-PTK procedure is that PTK could ablate part of the Bowman layer and the irregular stroma on top of the cone area. Ablation of these corneal tissues could alleviate corneal irregularity and enhance riboflavin penetration [
29,
30].
The Bowman layer and partial stroma were ablated with the excimer laser in the
CXL-plus-PTK group, and the regeneration of these structures was limited [
31,
32]. The MCT yield change in the
CXL-plus-PTK group was less effective than in the
Accelerated CXLgroup. A previous study found that the cornea gradually thickens up to 3 years postoperatively [
33]. Longer term follow-up and larger sample sizes are needed for more definitive evaluation of the long-term efficacies of these modalities.
Contrastingly, the
Iontophoresis CXL group yielded similar efficacy in the change of keratometry but performed better on the improvement of the
C value of the ABCD Grading System. The Global Consensus on Keratoconus and Ectatic Diseases (2015) states that no consistent or clear definition of keratoconus progression is available, and acknowledges the lack of specific quantitative data. The ABCD Grading System is increasingly recognized as a more appropriate evaluation system for keratoconus and ectactic corneal disease as the change of the keratometry at a ‘single’ point on the anterior surface, as measured by K
max, cannot represent the morphological changes of the entire cornea [
13,
34,
35]. Presently, clinical studies of the efficacies of different crosslinking procedures according to the more accurate ABCD Grading System are lacking. Our findings suggested that the
Iontophoresis CXL procedure performed better in reduction of the C value compared with the
Accelerated CXL group, while the
CXL-plus-PTK group had decreased performance. Therefore,
Iontophoresis CXLfor 10 min could be the most effective treatment among the five different CXL procedures at 1 year postoperatively. Iontophoresis can effectively deliver appropriate amounts of riboflavin to the stroma through the intact epithelium [
36]. Although
Iontophoresis CXL for 10 min group did not improve keratometry of the anterior surface as much as the
CXL-plus-PTK group when compared to the reference group,
Iontophoresis CXL more effectively protected cornea thickness.
Age is an important influence on the effectiveness of CXL [
10,
11]. To compensate for the effect of age on CXL efficacy, patients were divided into two age groups (age < 18 years and age ≥ 18 years). After GEE correction, the MCT reduction in pediatric patients was more significant than in adults at 1 year postoperatively following CXL. A prior study reported that pediatric central and paracentral corneal thicknesses increase slowly but did not elaborate on the reason for this phenomenon [
37]. However, MCT changes did not significantly differ between adult and pediatric patients at 2 and 4 years after CXL [
38,
39]. A longer term follow-up is therefore needed to validate this finding.
Moreover, we found that pediatric patients who underwent mechanical epi-off CXL procedure with a higher UV-A energy (7.2 J/cm
2) were more likely to develop corneal opacity. Another study that used the same procedure to treat pediatric keratoconus patients did not report complications through 2 years postoperatively [
40]. Corneal haze following CXL has been reported in previous studies [
41], but the reasons remain unclear at present. Potential reasons for this phenomenon are as follows: (1) more severe corneal ectasis caused by the fibroblast proliferation, which is more common in pediatric patients than in adults due to a more active proliferation response than adults [
42,
43]; (2) the haze could be related to slow spontaneous crosslinking reactions triggered by residual riboflavin in the corneal stroma and UV-A rays in natural light [
44] as two patients had a history of sunlight exposure early in the postoperative period; or (3) endothelial toxicity caused by reduced corneal thickness [
45]. To improve the efficacy of CXL and ensure treatment safety, the
High-Fluence Accelerated CXL procedure should be applied to pediatric patients with additional caution.
The study has several limitations that should be considered in its interpretation: the lack of a group treated by S-CXL, the accuracy of pre-existing data, and the inherent biases introduced by retrospective analysis. Further, the cornea biomechanics are not completely stable at 1 year following CXL. Hence, these findings should be further confirmed by prospective trials with a longer follow-up period, larger sample size, and better variable selection.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.