Keratoconus is a progressive corneal disorder that results in visual loss from irregular corneal astigmatism associated with high-order aberrations.1,2 It has a significant impact on quality of life.3–5 The prevalence of keratoconus varies by geography and ethnicity from 50 to 230 people per 100,000.6 The aim of treatment interventions is to optimize visual acuity and halt steepening and thinning of the cornea.7 The primary treatment for progressive keratoconus, or in young patients likely to progress, is corneal collagen crosslinking (CXL).8–10 CXL has been demonstrated to slow the progression through strengthening collagen fibers. For refractive empowerment, other surgical interventions include intrastromal corneal ring segments, penetrating keratoplasty, and deep anterior lamellar keratoplasty.11 The original Dresden CXL protocol has over time been optimized and adapted to improve patient outcomes.12 Most recently, there has been emerging evidence to suggest supplemental oxygen during CXL may be effective without adverse events in keratoconus.13,14 The first study investigating the effect of oxygen supplementation was published by Mazzotta et al,15 showing that using a transepithelial CXL approach with intraoperative supplemental oxygen resulted in clinically meaningful improvements in corneal curvature and corrected distance visual acuity (CDVA) without significant adverse events, suggesting a role for supplemental oxygen in deepening the CXL treatment efficacy.
In CXL, riboflavin acts as a photosensitizer for the ultraviolet (UV) irradiation to produce reactive oxygen species that create new covalent bonds between corneal collagen fibers.15,16 The amount of singlet oxygen released is dependent on the UV energy transfer from riboflavin and the stromal oxygen concentration.17 Richoz et al18 performed CXL in a hypoxic environment and demonstrated no stiffening effect and results similar to an untreated condition, which demonstrated the fundamental role of oxygen. Subsequently, Hill et al19 in an experiment with ex vivo porcine corneas showed supplemental oxygen significantly enhanced the effect of accelerated CXL. Moreover, oxygen supplementation has been demonstrated to increase the depth and strength of CXL in patients.20 Oxygen availability in the stroma during CXL can be increased by reducing the rate of consumption through pulsed illumination or using supplemental oxygen.15 Moreover, the use of pulsed illumination and supplemental oxygen during CXL have now entered clinical practice with a limited number of clinical studies reporting outcomes from the procedure.13,20,21 The pulsed light together with higher fluence compensating the presence of epithelium-on demonstrated enhancement of CXL transepithelial procedures with such as enhanced fluence pulsed light iontophoresis.22
This systematic review and meta-analysis aimed to evaluate the use of increased oxygen availability in CXL to treat keratoconus. In this study, we set out to report the efficacy of oxygen in CXL to treat patients with keratoconus based on visual acuity, curvature, and any adverse events.
This study constitutes a thorough literature search and meta-analysis of increased oxygen availability in CXL to treat keratoconus. We analyzed the data at a study-level (6 studies) and individual-level (203 patients, 233 eyes). The study was registered at the international database of prospectively registered systematic reviews (PROSPERO, CRD42021272745). The review method was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA) guidelines.23
We systematically searched the Medline, EMBASE, PubMed, Scopus, and Web of Science databases on November 3, 2021. Our search strategies were developed with the assistance of an academic librarian. The main search concepts were the population of interest (patients diagnosed with keratoconus), treatment involved (corneal collagen CXL), and element of interest (oxygen; Supplementary Digital Content Table 1, https://links.lww.com/APJO/A160). Retrieval was limited to articles published within the last 10 years. A search for grey literature was performed in OpenGrey, which yielded no results. Further, we hand-searched the reference lists of published articles to identify additional relevant studies and included an additional 8 studies for screening. A flowchart per PRISMA is presented in Figure 1.
We included studies that reported the use of methods to increase oxygen availability in the corneal stroma during CXL performed in patients with keratoconus published within the last 10 years. Studies in a language other than English and in nonhumans were excluded. Conditions other than keratoconus where CXL was used were also excluded.
All retrieved articles were exported to Covidence, a systematic review management software, to remove duplicates, screen titles and abstracts, and then performed full-text review. The data of interest included: (1) study ID (author, year of publication); (2) country; (3) demographic of each cohort; (4) study characteristics; primary outcomes including (5) CDVA, (6) maximum keratometry (Kmax) (7) adverse events; and (8) secondary outcomes including (9) demarcation line proportion (DLP); (10) demarcation line depth (DLD); (11) higher-order aberrations, trefoil and coma; and (12) corneal resistance factor (CRF).
Risk of Bias
The risk of bias was determined for each included study based on the National Health and Medical Research Council guidelines (Supplementary Digital Content Table 2, https://links.lww.com/APJO/A161).24 The risk of bias for the 2 randomized control trials was evaluated by Cochrane RoB 2.0 tool, and the remaining 4 studies that were nonrandomized prospective studies were assessed by ROBINS-I.25,26 These tools were used according to the Cochrane Handbook for Systematic Reviews of Intervention and had an overall low risk of bias in all studies (Supplementary Digital Content Fig. 1, https://links.lww.com/APJO/A162). This risk of bias assessment was conducted by 2 reviewers (G.B. and H.K.) independently, and disagreements were resolved through discussion with the third reviewer.27 To validate the quality assessment, the Critical Appraisal Skills Programme (CASP) tool was used and all studies met the minimum acceptable level (Supplementary Digital Content Table 3, https://links.lww.com/APJO/A163). CASP contains 11 questions for randomized control trials and 12 for cohort studies.28,29 Since there were a relatively small number of included studies (<10), we did not assess the publication bias.
The primary outcomes were changes in Kmax and CDVA from baseline and adverse events. The secondary outcomes included DLP, DLD, higher-order aberrations (trefoil and coma), and CRF. Data were reported across studies with different follow-up points. The most consistent point reported, 6 months of follow-up, was used in a meta-analysis. Data used in the meta-analysis were continuous data across 2 time points (baseline and 6 months).
A meta-analysis was performed with a random-effects model that determined the weighted effect-size estimation. Any possible heterogeneity between studies was considered with the I2 statistic. Statistical analysis and meta-analysis were carried out using Review Manager (open source desktop v5.4) developed by Cochrane Collaborations. A P value <0.05 was considered statistically significant.
Our search yielded 108 publication titles after duplicates were removed. The search results were screened and 21 full texts were assessed for eligibility. Then 6 studies were included in the systematic review and 5 included in the meta-analysis. The study selection process is illustrated in Figure 1.
The patient and intervention characteristics in each included study are outlined in Table 1. The reviewed studies included a total of 203 participants and 233 eyes. There was an overall average age of 24 (SD: 5.7) years and follow-up duration ranged from 1 to 12 months. About two-thirds (65.9%) of the participants were male and 34.1% were female. Each study was conducted in a different country (Italy, Iran, Turkey, China, France, Japan), which provided an international data set.
TABLE 1 -
Characteristics of Included Studies
||Age (31), y
||O2 rate (L/min)
|Mazzotta et al15
|Faramarzi et al32
|Aydin and Aslan20
||1, 6, 12
|Sun et al33
||1, 3, 6, 12
|Matthys et al21
||1, 3, 6, 12
|Kamiya et al13
||1, 3, 6, 12
NA indicates not applicable; P, prospective; RCT, randomized control trial.
There were 4 different methods used to increase the oxygen availability within the corneal stroma during CXL: goggles,13,15,21 facemask,30 eyelid speculum,20 and pulsed illumination.31 There were 5 of 6 studies that used epithelium-on (transepithelial) accelerated method of CXL included in the meta-analysis. One study had room air as a control20 and another conventional and accelerated epithelium-off CXL methods without oxygen supplementation as controls.30
All studies whether epithelium-off or epithelium-on demonstrated a reduction in Kmax from baseline to 6 months after increased oxygen (Fig. 2, Table 2, Supplementary Digital Content Fig. 2, https://links.lww.com/APJO/A164). In the only epithelium-on study included in the systematic review, Faramarzi et al30 compared oxygen delivered by a facemask during CXL to accelerated CXL and the conventional CXL as controls. At baseline, the Kmax was 54.3±3.6 diopter (D) in the oxygen CXL group, 54.6±5.0 D in accelerated CXL, and 56.0±5.2 D in the conventional CXL group, which was reduced to 53.5±3.2 D, 54.6±4.6 D, and 55.9±4.7 D at 6 months, respectively. The decrease in Kmax was significantly greater in the oxygen group (P=0.01). Aydin and Aslan20 demonstrated that the oxygen supplemented group had a significant decrease in Kmax group (55.1±4.0–54.4±3.9 D) compared to room air conditions as a control group (54.5±3.2–54.2±3.0 D) after 12 months of follow-up (P=0.019).
TABLE 2 -
Primary Outcomes in Included Studies
||Kmax Baseline (D)
||Kmax at 6 mo (D)
||CDVA Baseline (logMAR)
||CDVA at 6 mo (logMAR)
|Mazzotta et al15
|Faramarzi et al32
|Aydin and Aslan20
|Sun et al33
||0.28 (−0.04 to 0.6)
||0.16 (−0.04 to 0.36)
|Matthys et al21
||0.19 (−0.05 to 0.43)
||0.12 (−0.02 to 0.26)
|Kamiya et al13
||0.19 (−0.17to 0.55)
||0.13 (−0.23 to 0.49)
CDVA indicates corrected distance visual acuity; D, diopter; Kmax, maximum keratometry.
The meta-analysis for Kmax included the 5 epithelium-on accelerated studies and found a significant decrease [mean decrease, 1.2 D; 95% confidence interval (CI): 0.17–2.28; P=0.02] after increased oxygen availability at 6-month follow-up. Heterogeneity between studies was low (I2=0%; P=0.9). This is shown in the forest plot in Figure 2 although there were not enough studies to evaluate publication bias with a funnel plot. The Kmax was followed up at 12 months which continue to show this similar trend (Supplementary Digital Content Fig. 3, https://links.lww.com/APJO/A165).
The visual outcome was assessed by the change in CDVA between the baseline and at 6 months follow-up after increased oxygen availability (Fig. 3; Table 2). Faramarzi and colleagues showed that in the supplemental oxygen group using the epithelium-off protocol, the CDVA improved significantly from 0.26±0.21 at baseline to 0.19±0.16 logMAR at 6 months (P=0.026). Aydin and Aslan20 reported that the CDVA significantly improved with increased oxygen availability compared to room air after 12 months (mean change: 0.04 D; 95% CI: −0.11 to 0.19; P=0.03). Further, CDVA improvement was significantly greater in the oxygen group compared to the control at 1, 6, and 12 months (all P<0.05).
The meta-analysis for CDVA included 5 studies with a 0.08 logMAR mean improvement with increased oxygen availability at 6 months compared to baseline (95% CI: 0.02–0.13; P=0.01). Heterogeneity between studies was low (I2=0%; P=0.9). Similarly, there was a significant difference in CDVA between baseline and at 12 months (mean change: 0.08 D; 95% CI: 0.03–0.12; P<0.05). This is illustrated in Figure 3.
There were no serious adverse events with increased oxygen availability in CXL. Matthys et al21 reported the most frequent adverse event was corneal haze in 22 patients (64.78%) although this decreased to 8 patients (23.5%) at 6-month follow-up. Mazzotta et al15 suggested that the haze was associated with keratocyte loss and edema related to the relative depth of the demarcation line.
The mean DLP and DLD were reported by Aydin and colleagues (Supplementary Digital Content Figs. 4A, B, https://links.lww.com/APJO/A166). DLP is determined by dividing the DLD value to central corneal thickness and was significantly higher in the oxygen group (0.68±0.05) compared to the room air group (0.58±0.05) (P<0.001).20 DLD was significantly different between the increased oxygen availability in CXL at 320±17 m and in the control at 269±20 m.
Corneal higher-order aberrations, particularly coma and trefoil, were significantly reduced after increased oxygen availability at 6-month follow-up in the study conducted by Mazzotta and colleagues. From baseline, trefoil decreased from 0.18±0.08 to 0.12±0.06 m, while coma reduced from 0.47±0.28 to 0.28±0.16 m (P<0.05).15 Overall, higher-order aberrations decreased from 0.53±0.28 to 0.44±0.27 m, although not significant it demonstrated a similar trend. This is illustrated in Supplementary Digital Content Figure 4C–E, https://links.lww.com/APJO/A166.
In the study by Faramarzi and colleagues, CRF increased significantly exclusively in the increased oxygen availability group from 6.32±2.12 mm Hg at baseline to 7.38±1.88 mm Hg at 6-month follow-up (P=0.009). This is shown in Supplementary Digital Content Figure 4F, https://links.lww.com/APJO/A166. Notably, this was only significant in the oxygen CXL group and not in either control groups. Sun et al31 showed that there was no significant change in CRF.
This systematic review and meta-analysis evaluates the effect of increased oxygen availability to the corneal stroma during CXL to treat keratoconus. We found 6 studies that utilized a variety of methods to increase local oxygen during CXL. These studies were of adequate quality with a low risk of bias. Overall, increased oxygen availability in CXL significantly improved Kmax and CDVA without serious adverse events at 6-month follow-up.13,15,20,30,31 Three separate studies reported a significant change in DLD and DLP, higher-order aberrations (coma and trefoil), and CRF, respectively.15,20,30
While CXL can occur in anaerobic or aerobic environments, the aerobic pathway has been found to be more effective in producing oxygen radicals for each molecule of riboflavin, through reoxidizing reduced riboflavin to the original state.19,32 Our clinical findings support previous theoretical modeling studies in the use of methods to increase oxygen during CXL performed for keratoconus.18,33 They are also consistent with studies reporting the increased oxygen availability with transepithelial, pulsed, high-dose, photorefractive CXL to treat low-grade myopia and hyperopia, and caused corneal flattening compared to earlier protocols.34,35
Increased oxygen availability to the corneal stroma by supplemental supply through goggles, eyelid speculum or mask, or pulsed illumination CXL to treat keratoconus was effective in demonstrating significant reductions in corneal curvature and improved vision without adverse events. Mazzotta et al15 showed improved corneal curvature at 6 months following accelerated transepithelial CXL with a mask method of enhancing oxygen delivery. Aydin and Aslan20 suggested that this significantly stopped progression after 12 months and Kmax was significantly decreased in the oxygen group compared to the control. Sun et al31 demonstrated that CDVA and uncorrected visual acuity were significantly improved at 12-month follow-up with pulsed illumination as an alternative way to enhance oxygen availability. Matthys et al21 demonstrated that at 12-month follow-up there was a significant decrease in Kmax and a significant improvement in CDVA. It has been hypothesized that the improvement in CDVA was a result of anterior corneal surface regularization and adjusted with corneal higher-order aberrations from posterior surface irregularity.
The availability of oxygen in the CXL microenvironment is thought to enhance efficacy for several reasons. The epithelium is a barrier and metabolically consumes oxygen so it has been suggested oxygen increases the rate of diffusion to the stroma.36 Oxygen is an important reagent in the lysyl oxidase pathway to produce free radicals to catalyze crosslinks between corneal fibers.19 Although CXL can occur in aerobic (type 1 pathway) and anaerobic (type II pathway) environments, the aerobic pathway results in more oxygen radicals through reoxidizing riboflavin to its original form.18 It has been demonstrated that, the biomechanical effect of epithelium-off accelerated CXL showed that Young modulus increased in corneas at a normal atmosphere level, although at low oxygen there was no increase compared to controls. Conversely, ex vivo findings with epithelium-on pulsed CXL in oxygen rich environments showed an increased efficacy.18,37
While the primary outcomes demonstrate clinical relevance, the secondary outcomes provided further insight on the efficacy of CXL with increased oxygen. Analysis of demarcation lines by optical coherence tomography suggested that oxygen has a role in deepening treatment efficacy with a deeper demarcation line in high-dose oxygen compared to low dose.15,38,39 The DLD has been previously reported between 150 and 350 m at 4–6 weeks after CXL. Transepithelial approach has been related to a more superficial DLD. The improvement in vision has been suggested to be attributed to a decrease in higher-order aberrations, measured by coma and trefoil.15 Finally, while the CRF significantly increased in the oxygen group after 6 months compared to the control groups, its clinical significance is not clear.30
Our results suggest that use of methods to increase oxygen supplementation during CXL results in changes expected during effective CXL, that is significant improvements in Kmax and CDVA without serious adverse events. High-quality randomized controlled trials are now needed to demonstrate that supplemental oxygen has a beneficial effect above that of CXL without supplemental CXL. This means that increased oxygen availability could improve the quality of life of keratoconus patients by halting the natural progression of keratoconus and improving the corneal curvature and visual acuity.3
While our systematic review and meta-analysis is comprehensive of the current published literature looking at oxygen in CXL for keratoconus, our study is not without limitations. Our study was limited to studies published in indexed journals. These varied in oxygen delivery method, varied in follow-up period and controls in the included published studies. We were limited by the number of studies and follow-up data currently published. The paucity of data evaluating the efficacy of increased oxygen availability highlights the need for further research in this area.
We demonstrate that oxygen has an important role in optimizing CXL in the treatment of patients with keratoconus by improving visual and keratometry outcomes without serious adverse events. By using transepithelial approaches and higher fluencies such as 30 mW/UV-A power with shorter exposure times, intraoperative oxygen supplementation improves the functional outcomes of transepithelial CXL. Considering the novelty of enhanced oxygen availability in CXL, it is important to examine long-term data to understand the safety profile and stability of changes.
1. Ferdi AC, Nguyen V, Gore DM, et al. Keratoconus natural progression: a systematic review and meta-analysis of 11 529 eyes. Ophthalmology. 2019;126:935–945.
2. Kandel S, Chaudhary M, Mishra SK, et al. Evaluation of corneal topography, pachymetry and higher order aberrations for detecting subclinical keratoconus. Ophthalmic Physiol Opt. 2022;42:594–608.
3. Kandel H, Pesudovs K, Watson SL. Measurement of quality of life in keratoconus. Cornea. 2020;39:386–393.
4. Kandel H, Pesudovs K, Ferdi A, et al. Psychometric properties of the keratoconus outcomes research questionnaire: a Save Sight Keratoconus Registry Study. Cornea. 2020;39:303–310.
5. Tan JCK, Nguyen V, Fenwick E, et al. Vision-related quality of life in keratoconus: a save Sight Keratoconus Registry Study. Cornea. 2019;38:600–604.
6. Hashemi H, Heydarian S, Hooshmand E, et al. The prevalence and risk factors for keratoconus: a systematic review and meta-analysis. Cornea. 2020;39:263–270.
7. Kobashi H, Rong SS. Corneal collagen cross-linking for keratoconus: systematic review. Biomed Res Int. 2017;2017:8145651.
8. Raciti M, Epstein R, Majmudar P, et al. Corneal endothelial cell density following transepithelial collagen cross-linking. Invest Ophthalmol Vis Sci. 2013;54:5263.
9. Hersh PS, Greenstein SA, Fry KL. Corneal collagen crosslinking for keratoconus and corneal ectasia: one-year results. J Cataract Refract Surg. 2011;37:149–160.
10. Mazzotta C, Traversi C, Baiocchi S, et al. Corneal collagen cross-linking with riboflavin and ultraviolet a light for pediatric keratoconus: ten-year results. Cornea. 2018;37:560–566.
11. Espandar L, Meyer J. Keratoconus: overview and update on treatment. Middle East Afr J Ophthalmol. 2010;17:15–20.
12. Kandel H, Nguyen V, Ferdi AC, et al. Comparative efficacy and safety of standard versus accelerated corneal crosslinking for keratoconus: 1-year outcomes from the Save Sight Keratoconus Registry Study. Cornea. 2021;40:1581–1589.
13. Kamiya K, Kanayama S, Takahashi M, et al. Visual and topographic improvement with epithelium-on, oxygen-supplemented, customized corneal cross-linking for progressive keratoconus. J Clin Med. 2020;9:1–10.
14. Subasinghe SK, Ogbuehi KC, Dias GJ. Current perspectives on corneal collagen crosslinking (CXL). Graefes Arch Clin Exp Ophthalmol. 2018;256:1363–1384.
15. Mazzotta C, Sgheri A, Bagaglia SA, et al. Customized corneal crosslinking for treatment of progressive keratoconus: clinical and OCT outcomes using a transepithelial approach with supplemental oxygen. J Cataract Refract Surg. 2020;46:1582–1587.
16. Raiskup F, Spoerl E. Corneal crosslinking with riboflavin and ultraviolet A. Part II. clinical indications and results. Ocul Surf. 2013;11:93–108.
17. Kamaev P, Smirnov M, Friedman MD, et al. Photochemical kinetics model of corneal cross-linking with riboflavin. Investig Ophthalmol Vis Sci. 2015;56:3004.
18. Richoz O, Hammer A, Tabibian D, et al. The biomechanical effect of corneal collagen cross-linking (cxl) with riboflavin and UV-A is oxygen dependent. Transl Vis Sci Technol. 2013;2:6.
19. Hill J, Liu C, Deardorff P, et al. Optimization of oxygen dynamics, UV-A delivery, and drug formulation for accelerated epi-on corneal crosslinking. Curr Eye Res. 2020;45:450–458.
20. Aydin E, Aslan MG. The efficiency and safety of oxygen-supplemented accelerated transepithelial corneal cross-linking. Int Ophthalmol. 2021;41:2993–3005.
21. Matthys A, Cassagne M, Galiacy SD, et al. Transepithelial corneal cross-linking with supplemental oxygen in keratoconus: 1-year clinical results. J Refract Surg. 2021;37:42–48.
22. Mazzotta C, Bagaglia SA, Sgheri A, et al. Iontophoresis corneal cross-linking with enhanced fluence and pulsed UV-A light: 3-year clinical results. J Refract Surg. 2020;36:286–292.
23. Page MJ, Bossuyt PM. The PRISMA 2020 statement: an updated guideline for reporting systemic reviews. BMJ. 2021;372:n71.
24. National Health and Medical Resesarch Council (NHMRC). Guidelines for Guidelines: Assessing risk of bias. Australian Government. 2019. Available at: https://nhmrc.gov.au/guidelinesforguidelines/develop/assessing-risk-bias
. Accessed March 3, 2022.
25. Higgins J, Altman D. Cochrane Handbook for Systematic Reviews of Interventions. Chichester (UK): The Cochrane Collaboration; 2011.
26. Higgins JP, Sterne JAC, Savović J, et al. RoB 2: a revised Cochrane risk-of-bias tool for randomised trials. 2016. Available at: http://www.riskofbias.info
. Accessed March 3, 2022.
27. Olson J, Sharp P, Goatman K, et al. Improving the economic value of photographic screening for optical coherence tomography-detectable macular oedema: a prospective, multicentre, UK study. Health Technol Assess. 2013;17:1–142.
28. Programme CAS. Randomised Control Trial Standard Checklist 2018. Available at: https://casp-uk.b-cdn.net/wp-content/uploads/2020/10/CASP_RCT_Checklist_PDF_Fillable_Form.pdf
. Accessed May 3, 2022.
30. Faramarzi A, Hassanpour K, Rahmani B, et al. Systemic supplemental oxygen therapy during accelerated corneal crosslinking for progressive keratoconus: randomized clinical trial. J Cataract Refract Surg. 2021;47:773–779.
31. Sun L, Li M, Zhang X, et al. Transepithelial accelerated corneal collagen cross-linking with higher oxygen availability for keratoconus: 1-year results. Int Ophthalmol. 2018;38:2509–2517.
32. Kamaev P, Friedman MD, Sherr E, et al. Photochemical kinetics of corneal cross-linking with riboflavin. Invest Ophthalmol Vis Sci. 2012;53:2360–2367.
33. Schumacher S, Mrochen M, Wernli J, et al. Optimization model for UV-riboflavin corneal cross-linking. Invest Ophthalmol Vis Sci. 2012;53:762–769.
34. Sachdev GS, Ramamurthy S, Dandapani R. Photorefractive intrastromal corneal crosslinking for treatment of low myopia: clinical outcomes using the transepithelial approach with supplemental oxygen. J Cataract Refract Surg. 2020;46:428–433.
35. Stodulka P, Halasova Z, Slovak M, et al. Photorefractive intrastromal crosslinking for correction of hyperopia: 12-month results. J Cataract Refract Surg. 2020;46:434–440.
36. Atalay E, Ozalp O, Yildirim N. Advances in the diagnosis and treatment of keratoconus. Ther Adv Ophthalmol. 2021;13:25158414211012796.
37. Wang J, Wang L, Li Z, et al. Corneal biomechanical evaluation after conventional corneal crosslinking with oxygen enrichment. Eye Contact Lens. 2020;46:306–309.
38. Malhotra C, Jain AK, Gupta A, et al. Demarcation line depth after contact lens-assisted corneal crosslinking for progressive keratoconus: comparison of dextran-based and hydroxypropyl methylcellulose-based riboflavin solutions. J Cataract Refract Surg. 2017;43:1263–1270.
39. Seiler T, Hafezi F. Corneal cross-linking-induced stromal demarcation line. Cornea. 2006;25:1057–1059.