Corresponding Author: Mee Kum Kim, MD, PhD. Department of Ophthalmology, Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul 03080, Korea. Keratoconus is a corneal ectatic disorder characterized by progressive protrusion and thinning of the corneal stroma, resulting in irregular astigmatism and visual disturbance [1,2]. Collagen cross-linking (CXL) with ultraviolet A (UV-A) light and riboflavin (vitamin B2) halts the progression of keratoconus and has the benefit of delaying corneal transplantation [1-4]. Overall, the 10-year data suggest that CXL is a valid treatment option for keratoconus [5-8].
Since the introduction of the conventional Dresden protocol for CXL [4], new treatment options have been widely investigated, such as energy delivery, riboflavin instillation, and riboflavin penetration-facilitating methods [2,9,10]. In addition, accelerated CXL (A-CXL) has been developed to shorten surgical time [11,12]; however, its efficacy compared with that of conventional protocols has long been debated [13-25]. Recently, a study reported that A-CXL using a retention ring for 10-minute continuous application of 0.1% isotonic riboflavin showed comparable efficacy to that of the conventional protocol in 1 year [9].
However, only a few studies have reported the long-term results of A-CXL in adult keratoconus patients [26,27]. Herein, we investigated the long-term effects of A-CXL with 10-minute riboflavin application on halting the progression of keratoconus. In addition, we investigated whether prognostic corneal parameters suggested a greater corneal flattening effect after treatment in 10 minutes of riboflavin applied A-CXL. Finally, we compared the effects of A-CXL according to the continuous riboflavin application time (10 minutes vs. 5 minutes).
Materials and Methods
Ethics statement
This study was approved by the Institutional Review Board of Seoul National University Hospital (No. 2012-180-1186), with a waiver of informed consent. The study adhered to the tenets of the Declaration of Helsinki.
Patients
The medical records of patients who underwent CXL procedures were retrospectively reviewed. A total of 37 eyes of 33 patients who were diagnosed with progressive keratoconus and underwent retention ring-assisted A-CXL with the application of riboflavin between 2016 and 2020 were included. Patients treated between 2016 and 2018 underwent A-CXL with continuous riboflavin application for 10 minutes (10-minute group), while those treated between 2018 and 2020 received A-CXL with continuous riboflavin application for 5 minutes (5-minute group). A total of 18 eyes from 16 patients were included in the 10-minute group, and 19 eyes from 17 patients were included in the 5-minute group. Progression was defined as an increase in maximum keratometry (Kmax) of more than 1.5 diopters (D) per year in serial topography [9]. Patients with a preoperative Kmax >60 D or thinnest corneal thickness <400 μm were excluded from the indication for A-CXL.
Surgical procedures
The patients underwent A-CXL following the same procedures as in the previous study [9], except for riboflavin application time (10 or 5 minutes). After removal of the corneal epithelium, 8.0-mm retention ring (Frimen Inc) was applied on the epi-off corneal surface, and 0.1% isotonic riboflavin (Vibex Rapid, Avedro Inc) with dextran-free hydroxypropyl methylcellulose (Vibex Rapid, Avedro Inc) was continuously applied for either 10 or 5 minutes. UV-A (Avedro Inc) was applied for 8 minutes at an intensity of 30 mW/cm2 with pulsation (1 second on/off), resulting in a cumulative dose of 7.2 J/cm2. Postoperatively, all patients in the 10- and 5-minute groups were administered topical 0.5% moxifloxacin and 1% prednisolone four times daily for 1 week. After 1 week, following complete epithelial healing, 0.02% ocumetholone was applied four times daily for 4 weeks.
Clinical outcomes and prognostic factor analysis
Preoperative and postoperative examinations included best-corrected visual acuity (BCVA) as a logarithm of the minimum angle of resolution (logMAR), refractive errors measured using an Auto Kerato-Refractometer (KR-8900, Topcon), keratometric values, including maximum (Kmax), minimum (Kmin), and average (Kavg) values, and 3- and 5-mm irregularity indices calculated using topography (ORBSCAN IIz, Technolas Perfect Vision GmbH). Corneal thickness was measured using anterior segment optical coherence tomography (Visante OCT, Carl Zeiss Meditec). Endothelial cell density (ECD) was measured using noncontact specular microscopy (SP-8800, Konan). Preoperative parameters were compared with postoperative measurements at 6, 12, 24, and 36 months in the 10- and 5-minute groups. Failure to halt the progression of keratoconus was defined as more than 1.5 D progression of Kmax postoperatively compared with baseline Kmax. Progression-free survival and final success rates were compared between the 10- and 5-minute groups.
To identify the prognostic factors that may be associated with greater postoperative corneal flattening, various corneal parameters were analyzed using univariate and multivariate regression methods in eyes with successfully halted progression for 3 years in the 10-minute group (n = 16). ΔKmax was used as the main outcome measure and was defined as follows: ΔKmax = (3-year postoperative Kmax) - (preoperative Kmax) [28]. A negative ΔKmax value indicated a decrease in corneal power. The more negative the ΔKmax, the more favorable the outcome after CXL.
Statistical analysis
Statistical analyses were performed using IBM SPSS ver. 27.0 (IBM Corp) and GraphPad Prism ver. 10.4.1 (GraphPad Software Inc). Chi-square test and independent t-test were used to compare baseline parameters in the 10- and 5-minute groups. Mixed-effects analysis was used to compare preoperative and postoperative BCVA, refractive errors, topographic indices, central corneal thickness (CCT), thinnest corneal thickness (TCT), irregularity (IR) at 3 and 5 mm, and ECD. Survival curves were generated using the Kaplan-Meier method, and the difference in survival after A-CXL between the 10- and 5-minute groups was compared using the log-rank test. In addition, the prognostic factors for a higher corneal flattening effect were evaluated. The effect of preoperative variables (age, BCVA, Kmax, Kavg, keratometric astigmatism [Kastig], spherical equivalent [SE], IR at 3 and 5 mm, CCT, and TCT) on ΔKmax was evaluated using univariate linear regression analysis and multivariate analysis with stepwise linear regression analysis. The data are presented as mean ± standard deviation, and statistical significance was set at p < 0.05.
Results
The demographic and baseline characteristics of the study groups are shown in Table 1. The mean age was 27.2 ± 7.6 years in the 10-minute group and 26.9 ± 5.0 years in the 5-minute group. The majority of participants in both groups were men, with 77.8% (14 out of 18) in the 10-minute group and 78.9% (15 out of 19) in the 5-minute group. There were no significant differences in the baseline measurements between the two groups.
In the 10-minute group, BCVA improved significantly at 3-year postoperative from preoperative 0.43 to 0.21 log-MAR (mixed-effects analysis, p = 0.021) (Fig. 1A), and SE showed a significant decrease in myopia from preoperative −10.84 to −9.06 D at 3 years after surgery (mixed-effects analysis, p = 0.036) (Fig. 1B). There was no significant difference between preoperative and postoperative astigmatism in refraction (Fig. 1C).
Long-term changes in topographic indices showed a significant decrease in Kmax and Kavg at 3-year postoperative (from 52.52 to 50.39 D and from 49.64 to 48.04 D, respectively) (Fig. 2A, 2B). Kastig showed a significant decrease at 12 and 24 months postoperatively (4.45 and 4.58 D, respectively) but did not show a significant difference at postoperative 36 months (4.69 D) compared with preoperative Kastig (5.76 D) (Fig. 2C). IR at 3 and 5 mm showed statistically significant decreases from 5.76 to 5.63 (mixed-effects analysis, p < 0.001) and 6.65 to 5.72 (mixed-effects analysis, p = 0.004), respectively, at 3 years after surgery (Fig. 2D, 2E). CCT, TCT, and ECD showed no statistically significant differences before and after surgery (Fig. 3A-3C).
Fig. 4A-4J shows the efficacy and safety of A-CXL with 5-minute application of riboflavin (5-minute group). Although the 3-year postoperative BCVA (0.30 logMAR) was significantly better than the preoperative BCVA (0.51 log- MAR; mixed-effects analysis, p = 0.027), no statistically significant differences were observed in SE, astigmatism in refraction, Kmax, Kavg, Kastig, IR at 3 and 5 mm, CCT, or TCT before and after surgery. ΔBCVA ([3-year postoperative BCVA] - [preoperative BCVA]) was compared between the 10-minute group and the 5-minute group and no significant difference was found between the two groups (independent t-test, p = 0.947).
Survival of the progression-free state in the 10- and 5-minute groups is shown in Fig. 5. Although there was no statistical difference in survival between the two groups, progression was observed in 2 of 18 eyes (11.1%) in the 10-minute group and in 5 of 19 eyes (26.3%) in the 5-minute group.
Finally, we evaluated the possible prognostic factors associated with ΔKmax. Table 2 shows the results of univariate and multivariate analyses between preoperative variables and ΔKmax in eyes with successful progression-halting in the 10-minute group. In the univariate analysis, Kmax (β = −0.802), Kavg (β = −0.695), and Kastig (β = −0.814) were statistically significant factors, whereas in the multivariate analysis, only Kastig (β = −0.597) showed significant results. This suggests that the greater the preoperative Kastig is, the more flattened the postoperative Kmax at 3 years after A-CXL will be. As an example, Fig. 6A, 6B shows the preoperative and 3-year postoperative topographic images of the patient with the most flattened Kmax after A-CXL in the 10-minute group. Kmax decreased from 59.6 to 51.4 D (a decrease of 8.2 D).
Discussion
In this study, we demonstrated that pulsed-light A-CXL with retention ring-assisted continuous riboflavin application for 10 minutes is an effective and safe procedure for suppressing keratoconus progression, and that visual improvement is likely to be related to a decrease in irregular and regular astigmatism and flattening of the cornea. Also, 5-minute riboflavin application showed partial improvement in BCVA, but the corneal flattening effect and reduction of astigmatism seemed to be less evident than with 10-minute riboflavin application.
When comparing our results with those of the A-CXL effect in previous reports with a follow-up period of 24 months or more (Table 3) [23,26,29-33], the outcome of the 10-minute group of the present study shows comparable or more effective data demonstrating an average of 2.13 D of corneal flattening and BCVA improvement of 0.22 logMAR [23,26,29-33]. This may be related to deep CXL in the corneal stroma (demarcation line 279 μm), possibly due to the continuous application of riboflavin.
We also analyzed the factors associated with a higher Kmax flattening effect in successful patients in the 10-minute group at 3 years after surgery. Unsuccessful patients were defined as those in whom the cornea failed to halt the progression of keratoconus, and they were excluded from the analysis. Univariate analysis showed that preoperative Kmax, Kavg, and Kastig were statistically significant factors. Patients with higher preoperative Kmax, Kavg, or Kastig had a more flattened Kmax after surgery. In the multivariate analysis, Kastig was the only significant factor suggesting that a highly astigmatic cornea may show a maximal flattening effect after CXL. Wajnsztajn et al. [28] reported that eyes with higher preoperative Kmax, SE, and logMAR (BCVA) values had more successful results. However, that study included various types of CXL (epithelium-on/off, accelerated, and standard protocol) and consequently, the outcomes differed from those in our study. Higher Kastig may indicate greater biomechanical instability of the cornea, making it more responsive to the stiffening effects of A-CXL, thereby leading to a greater flattening effect. This hypothesis was also supported by two previous reports, presenting more effective in topographic flattening in advanced keratoconus but not as effective in less progressed keratoconus [34,35]. Additionally, the multivariate analysis did not show preoperative Kmax as a statistically significant factor for greater Kmax flattening. This may be due to the small sample size of the present study, which could have resulted in insufficient statistical power. Further studies with a larger number of patients and extended follow-up periods are needed.
In previous studies, conventional CXL showed long-term efficacy and safety in treating keratoconus. The present study showed comparable effects regarding Kmax flattening and improved BCVA compared with other conventional studies (Table 4) [5,36-43].
Regarding the progression of keratoconus, 11.1% (2 of 18) and 26.3% (5 of 19) showed progression in the 10- and 5-minute groups, respectively. Although the difference was not statistically significant, 10-minute riboflavin application in A-CXL showed more favorable outcomes in halting the progression of keratoconus; this is concordant with other results such as the flattening effect and improvement in visual acuity. A longer riboflavin application time allows for greater diffusion into the corneal stroma, resulting in a higher concentration and deeper penetration of riboflavin available for cross-linking. Furthermore, an increased riboflavin concentration enhances the absorption of UV-A light, thereby improving CXL efficacy. This may contribute to a more pronounced corneal flattening effect. Of the two eyes that showed progression of keratoconus in the 10-minute group, one showed progression at 1 year postoperatively, and the other eye at 3 years postoperatively. Therefore, long-term follow-up is required to monitor the progression of keratoconic progression after A-CXL.
Limitations of our study include a small sample size and the retrospective nature of the analysis. However, the efficacy of A-CXL with 10-minute riboflavin application is still noteworthy, along with the evidence that the application should be maintained for 10 minutes even if it is continuous.
In conclusion, retention ring-assisted continuous riboflavin application for 10 minutes in pulsed-light A-CXL is a comparably effective and safe treatment for halting keratoconus progression and showed comparable outcomes with those in the accelerated and conventional CXL protocols of previous studies. Higher preoperative Kastig may be related to a greater flattening effect of Kmax after 3 years of A-CXL.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
