Corneal Cross-Linking for Keratoconus

Table of Contents

1. What the Treatment Does
2. How the Dresden Protocol Began
3. What the Two United States Labels Say
4. What the Randomized Trials Found
5. KERALINK and the Evidence in Children
6. Accelerated Protocols Compared With the Standard
7. Epithelium-On Compared With Epithelium-Off
8. Corneal Thickness Limits
9. Complications and Their Rates
10. Results at 5, 10 and More Than 10 Years
11. Corneal Transplant Numbers After Cross-Linking Arrived
12. What the Cochrane Reviews Concluded
13. Cost and Access as Reported
14. What Is Not Known
15. Key Research Papers
16. Connections
17. Featured Videos

What the Treatment Does

The cornea gets its strength from sheets of collagen. In keratoconus that framework weakens and the cornea bulges into a cone. Cross-linking adds chemical bonds between collagen molecules, much as extra stitches stiffen a fabric.

The drug is riboflavin 5′-phosphate, a form of vitamin B2. Both United States labels call it a photoenhancer: under ultraviolet A light it generates singlet oxygen, a reactive form of oxygen, and that is what forms the cross-links. Every study on this page applied riboflavin directly to the eye. None tested riboflavin taken by mouth.

The effect sits near the front of the cornea. Wollensak and colleagues measured stiffness in corneal strips from enucleated human eyes and found stress up by 328.9% after treatment, and they described cross-linking as maximal in the front 300 micrometers. Hafezi and colleagues give the treated depth as the front 250 to 350 micrometers of the corneal stroma, the thick middle layer.

Three terms recur below. A diopter is the unit of corneal curvature, and a falling number means a flatter cone. Maximum curvature (maximum keratometry) is the steepest reading anywhere on the cornea and is the usual trial yardstick. Epithelium-off means the thin surface cell layer is scraped away first, and epithelium-on means it stays. Vision changes are reported on a logarithmic chart scale; the 2015 Cochrane review equates 0.20 on that scale with about 2 lines, or 10 letters.

Cross-linking aims to stop the cornea from changing. Correcting the vision that remains is a separate matter, described under vision correction on the keratoconus page.

Back to Table of Contents

How the Dresden Protocol Began

The work started in pig eyes. In 1998 Spoerl and colleagues at Dresden treated enucleated pig corneas with chemical agents and with several kinds of light, and measured how the strips stretched. Riboflavin with ultraviolet light raised stiffness, as did weak glutaraldehyde. Those authors named the potential: a treatment for keratoconus that was not a transplant.

The first patients were reported in 2003. Wollensak, Spoerl and Seiler treated 23 eyes of 22 people with moderate or advanced progressive keratoconus, in a nonrandomized pilot with no control group. After the central surface layer was scraped off, riboflavin drops were applied and the eye was exposed to ultraviolet A light at 370 nanometers and 3 milliwatts per square centimeter, from 1 centimeter away, for 30 minutes. A later Dresden paper gives the resulting surface dose as 5.4 joules per square centimeter. That recipe became known as the Dresden protocol.

In that pilot, progression at least stopped in all treated eyes over 3 months to 4 years of follow-up. Sixteen eyes (70%) regressed, with maximum curvature down 2.01 diopters, and vision improved slightly in 15 eyes (65%). Corneal and lens clarity, inner-layer cell counts and eye pressure did not change. The authors wrote that the treatment's simplicity and "minimal costs" might suit developing countries, and that long-term results were still needed.

Back to Table of Contents

What the Two United States Labels Say

Two riboflavin 5′-phosphate products have United States Food and Drug Administration labels, each tied to its own ultraviolet device. The figures below are read from the labels themselves on the National Library of Medicine's DailyMed service: the epithelium-off label and the epithelium-on label. The agency's approval database dates them 15 April 2016 and 17 October 2025.

The Epithelium-Off Label

• Indication: progressive keratoconus, and corneal ectasia (similar bulging) after refractive surgery.
• Solutions: riboflavin 5′-phosphate at 1.56 milligrams per milliliter in 20% dextran, and at 1.46 milligrams per milliliter without dextran.
• Steps as written: under anesthetic drops, the surface layer is removed over about 9 millimeters. One drop of the dextran solution goes on every 2 minutes for 30 minutes. A slit-lamp check then looks for yellow riboflavin inside the eye, and corneal thickness is measured. The eye is irradiated for 30 continuous minutes at 3 milliwatts per square centimeter and 365 nanometers, with a drop every 2 minutes throughout.
• Age: safety and effectiveness have not been established below age 14. No trial patient was 65 or older.
• Contraindications: none are listed. The warning section states that ulcerative keratitis, a corneal ulcer, can occur.
• Trial results at 12 months, keratoconus: maximum curvature fell 1.4 and 1.7 diopters in treated eyes in two studies and rose 0.5 and 0.6 in sham eyes. The differences were 1.9 and 2.3 diopters. The label also reports that about 56% of sham eyes had crossed over to treatment by month 3 and 89% by month 6, so the sham comparison after that point carries forward earlier readings.
• Aftercare: the counseling section describes a bandage contact lens placed on the day of treatment, and tells patients not to rub their eyes for the first 5 days.

The Epithelium-On Label (2025)

• Indication: keratoconus in adults and in children aged 13 and older. The word "progressive" does not appear in this indication.
• Solutions: riboflavin 5′-phosphate 0.239%, then 0.177%, in single-dose syringes, used only with the matching light system and oxygen goggles.
• Steps as written: a sponge soaked in the 0.239% solution wipes the mucus layer off the eye without removing the surface cells. Two drops of 0.239% go on every 60 seconds for 4 minutes, then two drops of 0.177% every 30 seconds for 6 minutes, then a rinse and a thickness measurement. With oxygen flowing through the goggles, the system irradiates for 11 minutes 6 seconds at 30 milliwatts per square centimeter and 365 nanometers, cycling 1 second on and 1 second off. A bandage contact lens goes on afterward.
• Contraindications: known hypersensitivity to benzalkonium chloride or another ingredient; and people with no natural lens, or with an implanted lens that does not block ultraviolet light.
• Warning: caution in people with a history of herpes infection of the cornea, because the infection can reactivate.
• Side effects, 389 treated eyes: eye redness in 31%. Between 5% and 25% had haze, light sensitivity, surface spots of inflammation (punctate keratitis), pain, irritation, watering, a surface defect, eyelid swelling, stress lines (corneal striae), reduced vision, dry eye or inflammation inside the eye.
• Trial results, maximum curvature: in the first sham-controlled study (280 eyes), the change at 6 months was −0.3 diopters treated against +0.6 sham, a difference of 1.0. At 12 months it was −0.4 against +0.7, a difference of 1.1, though sham eyes could cross over after month 6. In the second (312 eyes), the 12-month change was −0.5 against +0.4, a difference of 1.0.
• Age subgroups: in the first study, people under 29 had a difference of 2.0 diopters at 6 months, while people 29 and older showed no improvement in either arm. In the second, people 30 and older improved about as much as younger people (−0.6 against −0.5), but their sham eyes barely changed.

Both labels state that the procedure "should not be performed on pregnant women", because it is not known whether it can cause fetal harm. Both list severe eye pain and any sudden decrease in vision as reasons for a patient to contact the physician immediately. Neither label's trials compared the two approaches with each other. Beckman and colleagues published the second epithelium-on study in full in 2026, reporting no serious eye side effects in the study eye and punctate keratitis in 6.5% against 1.8% with sham.

Back to Table of Contents

What the Randomized Trials Found

The early evidence for epithelium-off cross-linking rests on a small number of trials, most of them in adults.

• Wittig-Silva and colleagues, 3-year results. One hundred eyes were randomized, with results reported for 46 treated and 48 control eyes. Maximum curvature in control eyes rose by 1.20, 1.70 and 1.75 diopters at 12, 24 and 36 months. In treated eyes it fell by 0.72, 0.96 and 1.03. Unaided vision worsened by 0.10 on the chart scale in control eyes and improved by 0.15 in treated eyes. Two eyes had minor complications that did not affect final vision.
• Hersh and colleagues, United States multicenter trial. In 205 people, the sham group received riboflavin without removal of the surface layer. Treated eyes flattened by 1.6 diopters at 1 year while control eyes continued to progress. In the treated group, 31.5% of eyes flattened by 2.0 diopters or more and 5.6% steepened by 2.0 or more. Corrected vision improved on average, and the endothelial cell count, the inner-layer cells, did not change significantly. Haze was the most frequent side effect.
• Lang and colleagues, double-masked trial. With 29 eyes followed for about 3 years, refractive power fell 0.35 diopters a year with treatment and rose 0.11 a year with placebo, a significant difference. The authors added a caution: some treated patients still progressed and some untreated controls improved, so further work is needed to decide who needs treatment.

Pooled analyses point the same way. Sarma and colleagues in 2023 drew on 48 randomized trials, which also covered comparisons between protocols, and found 1.78 diopters at 1 year from 4 control-group trials and 0.10 in corrected vision from 7. Beyond 1 year, their comparison with control rested on a single trial.

Worse starting points did not predict a worse result in one analysis. Greenstein and Hersh found that eyes starting at 55 diopters or steeper were 5.4 times more likely to flatten by 2 diopters or more, and eyes with vision of 20/40 or worse were 5.9 times more likely to gain two lines. No starting feature predicted worsening. Other studies, listed under complications below, do report risk factors for failure.

Back to Table of Contents

KERALINK and the Evidence in Children

The corneal changes of keratoconus tend to progress until about age 30, so the youngest patients have the longest period of risk. KERALINK tested cross-linking in this group. It ran in four United Kingdom hospitals with 60 participants aged 10 to 16, masked assessors, and one study eye each. Half had epithelium-off cross-linking plus standard care, and half had standard care alone, meaning glasses or contact lenses as needed.

• Curvature at 18 months: mean steepest-axis curvature was 49.7 diopters with cross-linking and 53.4 with standard care, an adjusted difference of 3.0 diopters.
• Progression (a rise of 1.5 diopters): 2 of 30 patients (7%) against 12 of 28 (43%), with odds about 90% lower.
• Vision: adjusted differences favored cross-linking by 0.31 unaided and 0.30 corrected on the chart scale, about 3 lines each. These figures come from the National Institute for Health Research report. The journal abstract gives a different corrected figure (−0.51) with a confidence interval that crosses zero alongside a significant P value, so the two records do not agree on that one number.
• Quality of life: similar in both groups.
• Safety: no adverse events were attributed to cross-linking.

The authors listed limits. Device software flagged some curvature readings as suspect in the eyes with the most progression. Follow-up ended at 18 months. And they noted that keratoconus does not progress in everyone on standard care, which the 57% without progression in that group illustrates. Both of the questions they named as most important for later work concern the long term: whether the arrest lasts, and how many untreated children eventually progress.

Longer pediatric data come from uncontrolled follow-up, and they show more late progression than adult series do.

• 5 years (Godefrooij): 54 eyes, maximum curvature down 2.06 diopters, and 22% progressed by 1 diopter or more despite treatment. A more off-center cone was the identified cause.
• 10 years (Mazzotta): 62 eyes, an overall progression rate of 24%, including 2 eyes that went to a corneal graft.
• Up to 14 years (Vinciguerra): 38 eyes at a mean of 11.6 years. Vision improved. Thirteen eyes (34%) met a multi-measure definition of progression, and 3 (7.9%) changed significantly in front-surface curvature.

Maskill and colleagues found that people whose first treatment failed were significantly younger: 21.3 against 26.7 years. Bordais and colleagues found that all repeat treatments in their 10-year series were in people first treated under 18.

The label age limits are not the same as the trial ages. KERALINK enrolled from age 10, while the United States labels start at 14 (epithelium-off, "not established" below that) and 13 (epithelium-on).

Back to Table of Contents

Accelerated Protocols Compared With the Standard

The standard 30-minute exposure is long, and accelerated protocols shorten it by raising light intensity while keeping the total energy the same. Hashemi and colleagues, for example, gave the same 5.4 joules per square centimeter as 18 milliwatts for 5 minutes instead of 3 milliwatts for 30 minutes. A common alternative is 9 milliwatts for 10 minutes. The epithelium-off United States label describes only the 3-milliwatt, 30-minute schedule.

• Trials pooled (Kobashi and Tsubota, 6 randomized trials): at 1 year, no differences in maximum curvature, thickness, unaided vision or inner-layer cell count. The standard protocol gave slightly better corrected vision, a difference the authors said may not be clinically significant. The authors concluded that both methods similarly stop progression.
• Trials and other studies pooled (Shajari, 22 studies, 1,158 eyes): the standard protocol was better for minimum curvature and depth of the treatment line. Accelerated had less corneal thinning. Vision, maximum curvature and cell density showed no difference.
• A paired-eye trial at 18 months and 4 years (Hashemi, 31 eyes per arm): the standard eye had more flattening. At 4 years the central curvature fell 1.35 diopters with the standard protocol and 0.36 with 18 milliwatts. The gap came from cones away from the center; central cones did about equally.
• Longer term pooled (Sarma): at 2 and 4 years the conventional protocol did better at improving maximum curvature.
• Real-world registry (Kandel, 684 eyes, 24 sites): 1-year changes in maximum curvature of −0.9 diopters standard and −1.2 accelerated (9 milliwatts), with adverse events in 8.3% and 5.5%, not statistically different.

The treatment line is the demarcation line, a visible boundary in the cornea roughly a month afterward. It was shallower after accelerated treatment by about 102 micrometers in the Kobashi analysis. Whether that matters is not settled: in the Hafezi thin-cornea series, line depth did not track the curvature outcome, and those authors suggested it reflects wound healing more than stiffening.

Ahmet and colleagues followed 175 pediatric eyes for 10 years after accelerated treatment and found re-progression in 9.14%, haze in 7.43% (4 eyes permanent), a transplant in 3 eyes and a repeat treatment in 3.

Back to Table of Contents

Epithelium-On Compared With Epithelium-Off

The 2021 Cochrane authors list the proposed advantages of keeping the surface layer: faster healing, less discomfort, quicker visual recovery and less haze. They also name the difficulty: getting enough riboflavin through intact cells and enough ultraviolet dose into the tissue. Results differ with the version tested.

• An early head-to-head trial (Soeters, 61 eyes): with a commercial transepithelial riboflavin solution. Eight transepithelial eyes (23%) steepened more than 1 diopter at a year against none with epithelium-off. Epithelium-off eyes flattened 1.2 to 1.5 diopters, and 4 of them (15%) had healing complications against none in the other group. The authors wrote that they did not recommend replacing the epithelium-off method at that time.
• Iontophoresis, an electric current that drives riboflavin in (Lombardo, 34 eyes): at 2 years, maximum curvature changed −1.05 diopters with the current (not statistically significant) and −1.51 with the Dresden protocol. Two eyes (10%) steepened more than 1 diopter with the current and none with Dresden. Vision improved slightly more with the current. No outcome difference between groups reached significance at 24 months.
• High oxygen (Elving, 32 people, one eye each way): customized epithelium-on treatment in high oxygen against customized epithelium-off in room air. At 24 months, maximum curvature fell 1.74 and 1.72 diopters, vision gains were similar, no adverse events occurred, and the epithelium-on eyes had less discomfort for the whole first week. This was a customized protocol, not the labeled one.
• Pooled 2024 (Borchert, 1,622 eyes): no significant difference at 12 months in corrected vision or maximum curvature. Unaided vision favored epithelium-off.
• Pooled 2025 (Mohammadpour, 28 studies, 1,764 eyes): epithelium-off gave more flattening, at 1 and 2 years, and more early thinning. Vision was equivalent.
• Pooled 2023 (Sarma): over the long term, chemically enhanced transepithelial treatment did worse than conventional treatment on maximum curvature, vision and cell density; iontophoresis was similar up to 2 years.

The trials behind the 2025 label compared the epithelium-on method with sham treatment, not with epithelium-off. Its 1.0-diopter difference from sham at 12 months is smaller than the 1.9 and 2.3 on the epithelium-off label, but those trials differed in design, severity and crossover, so that gap is not a direct comparison. A 2026 perspective essay by Jo and colleagues says it is uncertain whether the epithelium-on therapy has comparable short- and long-term effectiveness.

Back to Table of Contents

Corneal Thickness Limits

The corneal limit exists to protect the endothelium, the single inner layer of cells that keeps the cornea clear. The 400-micrometer figure traces to rabbit work: Wollensak and colleagues found in rabbit corneas that endothelial cells died from a light dose of 0.65 joules per square centimeter or more at the inner surface, a level reached at the standard surface dose when the cornea is thinner than 400 micrometers. In advanced keratoconus many corneas are thinner than that.

• Epithelium-off label: if the cornea measures under 400 micrometers after soaking, 2 drops of the dextran-free solution go on every 5 to 10 seconds until it reaches at least 400. The label states that irradiation is not to be done unless this threshold is met.
• Epithelium-on label: the label states that irradiation is not to be done below 325 micrometers. The label text read for this page gives no reason for the number.

Doctors have also tried treating thinner corneas outside these labels.

• Swelling with a low-salt (hypo-osmolar) riboflavin solution: Hafezi and colleagues reported 20 patients with no complications. Raiskup and Spoerl treated 32 eyes, which swelled from 337 to 451.8 micrometers, and found stable curvature and vision at 1 year, with all corneas clear.
• A limit to swelling: Hafezi reported a cornea of 268 micrometers that swelled to 406 and progressed anyway, by 2.3 diopters at 6 months. He suggested about 330 micrometers of stroma as a minimum.
• Standard treatment despite thinness: Kymionis and colleagues treated 14 eyes between 340 and 399 micrometers. Curvature improved, and inner-layer cell density fell from 2,733 to 2,441 cells per square millimeter. Those authors described this as a significant decrease.
• Individualized dose (the Sub400 protocol): light time is adjusted to the thickness. Hafezi and colleagues treated 39 eyes from 214 to 398 micrometers; 35 (90%) were stable at 12 months and none showed endothelial failure. In a smaller follow-up of 17 eyes over up to 24 months, Blaser and colleagues saw a mean drop of 56.4 cells per square millimeter that was not statistically significant.

All of these are small uncontrolled series with short follow-up.

Back to Table of Contents

Complications and Their Rates

Rates vary a great deal with definitions, protocol and follow-up. The Cochrane authors in 2015 described adverse effects as not uncommon but mostly brief and of low clinical significance. The rare ones can permanently damage vision.

Expected Early Effects

The epithelium-off label's pooled table for keratoconus counts events up to month 3 in 102 treated people and 103 controls: haze in 64% against 9%, punctate keratitis 25% against 8%, a surface defect 24% against 1%, stress lines 24% against 12%, eye pain 17% against 3%, blurred vision 16% against 2%, foreign-body sensation 15% against 1%, light sensitivity 11% against none, and keratitis 1% against none. The label says most events resolved in the first month, some took up to 6 months, and haze took up to 12. In 1% to 2% of patients, a surface defect, corneal swelling, opacity or a scar was still present at 12 months.

Infection

• Frequency: a hospital series counted 6 cases in 4,863 standard treatments, or 0.12% (Farrokhpour). The same report describes 19 infected eyes in all, including referrals; in 5 patients, medicine did not stop the infection and keratoplasty followed. An accelerated series counted 7 in 532 eyes (Maharana).
• Severity: in that accelerated series the infected patients had a median age of 11, 4 of the 7 also had vernal keratoconjunctivitis, 3 were mixed infections, and 5 of 6 bacterial cases resisted fourth-generation fluoroquinolone antibiotics. One eye perforated, and 4 of 7 had vision worse than 6/60 at 6 months.
• Pattern: a systematic review of 26 publications (Murchison) found the overall frequency low, bacteria the commonest cause, and a mean onset of 4.8 days. Steroids, bandage contact lenses and a history of atopic or herpetic disease were associated with infection.
• Case reports: herpes keratitis with iritis 5 days after treatment in a woman with no such history (Kymionis), and a bacterial infection at 3 days that left a scar and a permanent drop in vision (Pollhammer).

Haze and Scarring

• Natural course: in the United States trial cohort, measured haze peaked at 1 month, fell between 3 and 12 months, and had not fully returned to baseline at 12 months in keratoconus. It did not correlate with final results (Greenstein).
• Clinically significant haze: 8.6% of 163 eyes at 1 year (Raiskup). Those eyes began thinner (420 against 478 micrometers) and steeper, and their vision fell, while vision improved in the other eyes.
• Other findings: sterile infiltrates in 7.6% and central scars in 2.8% of 117 eyes (Koller).

Endothelial Damage

Most series report no change in inner-layer cell counts, including at 10 and 13 years (Raiskup 2015; Enders). Serious cases still occur at thickness above the limit. Sharma and colleagues found lasting corneal swelling in 10 of 350 patients (2.9%), all above 400 micrometers. Corneal transplant was offered to 5 and 2 had it. Bagga and colleagues described an 18-year-old with generalized endothelial loss who needed a full-thickness transplant.

Failure and Vision Loss

• Koller, 117 eyes at 1 year: continued progression in 7.6%, loss of 2 or more lines in 2.9%. Age over 35 and good starting vision (better than 20/25) were risk factors for vision loss; high starting curvature was a risk factor for failure. The authors estimated that limiting treatment to under 58 diopters could bring failure below 3%.
• Range across studies (Shalchi review of epithelium-off studies): failure up to 33%, repeat treatment up to 8.6%, conversion to transplant up to 6.25%.
• Repeat treatment (Maskill): 20 of 21 eyes stabilized at a mean of about 30 months, with no complications noted.
• Too much flattening (Henriquez, 45 eyes, 10 to 13 years): flattening of 5 diopters or more in 15.5%, with vision loss in 2 eyes. One flattened 11.5 diopters, lost 7 lines and needed a transplant. These authors said the problem may be more common than is recognized.

Back to Table of Contents

Results at 5, 10 and More Than 10 Years

The control-group comparisons read for this page stop at about 3 years, and in the epithelium-off label trials most sham eyes had crossed over to treatment by month 6. The long-term record below is uncontrolled follow-up of those who returned, and it mostly comes from single centers.

About 5 to 7 Years

• Siena (Caporossi): 44 eyes at 4 to 5 years, all stable, with mean curvature down 2 diopters and corrected vision up 1.9 lines. Over 65% of untreated fellow eyes progressed by about 1.5 diopters within 24 months and were then treated.
• Hashemi, 40 eyes at 5 years: maximum curvature fell only 0.16 diopters, and corrected vision improved from 0.31 to 0.19.
• O’Brart, 36 eyes at 7 years: no treated eye progressed and maximum curvature fell 0.91 diopters. In 29 untreated fellow eyes it rose 0.87.

About 10 Years

• Dresden (Raiskup 2015): 34 eyes at a mean of 11 years. Maximum curvature went from 53.2 to 49.56 diopters and vision improved by 0.14; inner-layer cell counts did not change.
• France (Bordais), 89 eyes: maximum curvature down 2.31 diopters, with flattening still occurring between years 5 and 10. Repeat treatment in 4% of patients, and loss of more than 3 lines in 1%.
• United States pivotal trial, single center (Greenstein 2023): 19 returning eyes. Mean maximum curvature was unchanged (58.2 to 58.3 diopters), and 81.8% of keratoconus eyes stayed stable.
• Seifert, 131 eyes (44 at 10 years): the least favorable: people with a rise of more than 2 diopters went from 16% at 5 years to 33% at 10. Young age, high astigmatism, thin cornea, poor starting vision and atopic dermatitis were the risk factors, and the authors raised re-treatment for such patients.

Beyond 10 Years

• Vinciguerra, 27 eyes, mean 11 and maximum 13 years: failure in 2 eyes (7.4%), with stable corrected vision.
• Enders, 168 enrolled: data on 142 eyes at 1 year, 105 at 5, 61 at 10 and only 9 at 13. Keratometry remained lower. Three eyes were retreated, none had a transplant, and no infections occurred.
• Henriquez, 45 eyes, 10 to 13 years: progression in 2.22%.

A 2024 review of the long-term record by Cehelyk and Syed reports stabilization up to 13 years with the standard protocol and overall higher progression in children. A paper reporting results at 15 or 20 years did not surface in the searches for this page.

Back to Table of Contents

Corneal Transplant Numbers After Cross-Linking Arrived

The 2003 pilot's hope was fewer corneal transplants. Before-and-after counts from different places do not all agree, and a before-and-after count cannot prove cause.

• Netherlands, national registry (Godefrooij): 269 transplants for keratoconus in people under 50 in 2005 to 2007, and 201 in 2012 to 2014, about 25% fewer.
• One corneal department (Sandvik): 55 keratoconus transplants in 2005 to 2006 and 26 in 2013 to 2014, while the eyes operated on were more advanced.
• Canada, two provinces (Sklar): the keratoconus share of all grafts fell, but the absolute number did not (179 against 198 a year). The authors attributed the falling share to a rise in other kinds of transplant.
• United States, a large records network (Gad El Sayed): among 18,467 matched patients, full-thickness transplant was recorded in 9.76% before the April 2016 approval and 3.89% after. Partial-thickness grafts went from 3.88% to 1.16%.

Back to Table of Contents

What the Cochrane Reviews Concluded

Cochrane reviews apply strict rules to how trials are judged. Two have addressed cross-linking in keratoconus, and both are more guarded than most of the papers above.

2015: Cross-Linking Compared With No Treatment

Sykakis and colleagues found 3 randomized trials, 225 eyes, adults only, all at high risk of bias, with data that could not be pooled. Treated eyes were about 2 diopters less steep and about 2 chart lines better at 12 months, from one trial and very low-quality evidence. Data on quality of life and costs were absent. In one trial 3 of 12 people had side effects, and in another 3 of 50. The conclusion was that the evidence "is limited due the lack of properly conducted RCTs". KERALINK and the sham-controlled trials on the epithelium-on label came later. Searches for this page found no later Cochrane review of this comparison.

2021: Transepithelial Compared With Epithelium-Off

Ng and colleagues included 13 trials with 723 eyes. The difference in maximum curvature at 12 months or later was 0.99 diopters, with a confidence interval from −0.11 to 2.09, and very low certainty. Moderate-certainty evidence showed slightly more haze or scarring with epithelium-off. The conclusion was that it "remains unknown" whether transepithelial cross-linking or any other approach has an advantage over epithelium-off for progression, vision or patient-reported outcomes. The authors asked for arrest of progression as the main outcome of future trials, standard definitions of progression, and longer follow-up. A journal summary of the same review repeated these findings.

Back to Table of Contents

Cost and Access as Reported

Cost-Effectiveness Models

These studies are computer models, not observed spending. They measure cost per quality-adjusted life year, one year of life in full health. Each result depends heavily on how long the effect is assumed to last.

• United Kingdom National Health Service (Salmon): £3,174 per quality-adjusted year in the base case, rising above £33,263 if the effect lasts only 5 years.
• A European payer model (Godefrooij): €54,384 assuming a 10-year effect, and €10,149 assuming a lifelong effect.
• Canada (Leung): Can$9,090, compared with conventional care ending in a transplant.
• Brazil's public system (Hansen): US$58.26, with a mean of 968.8 fewer transplants per 10,000 eyes treated.

The models predate the 2025 product, so none uses its price.

United States Price and Access

A 2026 perspective essay by 19 authors (Jo and colleagues) is the main source here, and it is opinion backed by a literature review, not a measured study. It reports that one approved epithelium-off protocol exists in the United States and that access has been limited by high costs, insurance coverage and proprietary protocols. It gives a recently reported wholesale acquisition cost of $78,500 for the epithelium-on therapy, without saying in the abstract what unit that price covers. It also reports that the epithelium-off product is to be phased out in 2026, which would leave no approved epithelium-off option in the United States, and it asks for a pause on that. The DailyMed listing for the epithelium-off label read for this page carried a July 2026 date; this page did not confirm the product's current market status.

Claims data give a measured picture for 2016 to 2020. Gupta and colleagues compared 552 treated people with 2,723 matched controls in a national insurance database. People 30 or older had lower odds of treatment (0.34), as did people in health maintenance organization plans (0.64) and on the east coast (0.37) and in the Lower Midwest (0.31). Sex, race, education and income were not associated with treatment in this insured group.

Access Elsewhere

At a public service in Auckland, New Zealand, Angelo and colleagues studied 454 patients with keratoconus and found 69% attendance. Pacific Peoples attended least (58.9%), and Pacific Peoples and Māori came in younger with worse disease. Unemployment was associated with worse vision.

Back to Table of Contents

What Is Not Known

• How long the effect lasts. Few eyes have been followed past 13 years, and follow-up numbers shrink along the way. The 2003 Dresden authors named this as an open question, and the cost models turn on it.
• Who needs it. In KERALINK, 57% on standard care did not progress in 18 months. Natural-history and trial data do not yet identify which eyes would stay stable without treatment, and Lang and colleagues reached the same point.
• How to define progression. Studies use different cut-offs and measures. The Vinciguerra pediatric series shows how much this matters: 34% by one definition, 7.9% by a narrower one.
• How the epithelium-on method compares directly. The label trials were against sham, follow-up there was 12 months, and head-to-head trials used other protocols.
• How accelerated protocols hold up past a few years. Most comparisons stop by 4 years.
• The safest lower thickness limit. Thin-cornea series are small, and endothelial follow-up is short.
• Long-term effects in younger children and after treatment under the labeled ages. Pediatric series show more late progression, and they are uncontrolled.
• How much cross-linking itself reduced transplants. National counts disagree and cannot separate it from other changes in care.
• Effects in pregnancy and in older adults. Both labels carry the pregnancy statement above, and the epithelium-off trials enrolled no one 65 or older.

Back to Table of Contents

Key Research Papers

1. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-a-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135(5):620-7. PubMed PMID: 12719068
2. Spoerl E, Huhle M, Seiler T. Induction of cross-links in corneal tissue. Exp Eye Res. 1998;66(1):97-103. PubMed PMID: 9533835
3. 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(9):1780-5. PubMed PMID: 14522301
4. Hafezi F, Mrochen M, Iseli HP, Seiler T. Collagen crosslinking with ultraviolet-A and hypoosmolar riboflavin solution in thin corneas. J Cataract Refract Surg. 2009;35(4):621-4. PubMed PMID: 19304080
5. Wollensak G, Spoerl E, Wilsch M, Seiler T. Endothelial cell damage after riboflavin-ultraviolet-A treatment in the rabbit. J Cataract Refract Surg. 2003;29(9):1786-90. PubMed PMID: 14522302
6. Beckman KA, Parkhurst GD, Lee JH, Al-Mohtaseb ZN, Greenwood MD, Armijo M, et al. Randomized, Controlled Study to Evaluate the Safety and Efficacy of Oxygen-Enriched Epithelium-On Corneal Cross-Linking for the Treatment of Keratoconus. Ophthalmol Ther. 2026;15(4):1463-1484. PubMed PMID: 41824263
7. Wittig-Silva C, Chan E, Islam FM, Wu T, Whiting M, Snibson GR. A randomized, controlled trial of corneal collagen cross-linking in progressive keratoconus: three-year results. Ophthalmology. 2014;121(4):812-21. PubMed PMID: 24393351
8. Hersh PS, Stulting RD, Muller D, Durrie DS, Rajpal RK, United States Crosslinking Study Group. United States Multicenter Clinical Trial of Corneal Collagen Crosslinking for Keratoconus Treatment. Ophthalmology. 2017;124(9):1259-1270. PubMed PMID: 28495149
9. Lang SJ, Messmer EM, Geerling G, Mackert MJ, Brunner T, Dollak S, et al. Prospective, randomized, double-blind trial to investigate the efficacy and safety of corneal cross-linking to halt the progression of keratoconus. BMC Ophthalmol. 2015;15:78. PubMed PMID: 26194634
10. Sarma P, Kaur H, Hafezi F, Bhattacharyya J, Kirubakaran R, Prajapat M, et al. Short- and long-term safety and efficacy of corneal collagen cross-linking in progressive keratoconus: A systematic review and meta-analysis of randomized controlled trials. Taiwan J Ophthalmol. 2023;13(2):191-202. PubMed PMID: 37484615
11. Greenstein SA, Hersh PS. Characteristics influencing outcomes of corneal collagen crosslinking for keratoconus and ectasia: implications for patient selection. J Cataract Refract Surg. 2013;39(8):1133-40. PubMed PMID: 23889865
12. Larkin DFP, Chowdhury K, Doré CJ, Bunce C, Burr JM, Caverly E, et al. Epithelium-off corneal cross-linking surgery compared with standard care in 10- to 16-year-olds with progressive keratoconus: the KERALINK RCT. Efficacy Mech Eval. Southampton (UK): NIHR Journals Library; 2021. PubMed PMID: 34644043
13. Larkin DFP, Chowdhury K, Burr JM, Raynor M, Edwards M, Tuft SJ, et al. Effect of Corneal Cross-linking versus Standard Care on Keratoconus Progression in Young Patients: The KERALINK Randomized Controlled Trial. Ophthalmology. 2021;128(11):1516-1526. PubMed PMID: 33892046
14. Godefrooij DA, Soeters N, Imhof SM, Wisse RP. Corneal Cross-Linking for Pediatric Keratoconus: Long-Term Results. Cornea. 2016;35(7):954-8. PubMed PMID: 27027921
15. Mazzotta C, Traversi C, Baiocchi S, Bagaglia S, Caporossi O, Villano A, et al. Corneal Collagen Cross-Linking With Riboflavin and Ultraviolet A Light for Pediatric Keratoconus: Ten-Year Results. Cornea. 2018;37(5):560-566. PubMed PMID: 29319598
16. Vinciguerra R, Bordignon N, Ferraro V, Mazzotta C, Rosetta P, Vinciguerra P. Corneal Collagen Cross-Linking for Progressive Keratoconus in Pediatric Patients: Up to 14 Years of Follow-up. Am J Ophthalmol. 2023;255:170-177. PubMed PMID: 37478962
17. Maskill D, Okonkwo A, Onsiong C, Hristova S, Dodd A, Anand S. Repeat corneal collagen cross-linking after failure of primary cross-linking in keratoconus. Br J Ophthalmol. 2024;108(5):662-666. PubMed PMID: 37344124
18. Bordais J, Cassagne M, Touboul D, Saunier V, Butterworth J, Malecaze F, et al. Conventional Epithelial-Off Corneal Crosslinking in Patients With Progressive Keratoconus: 10-Year Outcomes. Cornea. 2024;43(7):874-882. PubMed PMID: 38537067
19. Hashemi H, Miraftab M, Seyedian MA, Hafezi F, Bahrmandy H, Heidarian S, et al. Long-term Results of an Accelerated Corneal Cross-linking Protocol (18 mW/cm2) for the Treatment of Progressive Keratoconus. Am J Ophthalmol. 2015;160(6):1164-1170.e1. PubMed PMID: 26314662
20. Hashemi H, Mohebbi M, Asgari S. Standard and accelerated corneal cross-linking long-term results: A randomized clinical trial. Eur J Ophthalmol. 2020;30(4):650-657. PubMed PMID: 30943777
21. Kobashi H, Tsubota K. Accelerated Versus Standard Corneal Cross-Linking for Progressive Keratoconus: A Meta-Analysis of Randomized Controlled Trials. Cornea. 2020;39(2):172-180. PubMed PMID: 31369463
22. Shajari M, Kolb CM, Agha B, Steinwender G, Müller M, Herrmann E, et al. Comparison of standard and accelerated corneal cross-linking for the treatment of keratoconus: a meta-analysis. Acta Ophthalmol. 2019;97(1):e22-e35. PubMed PMID: 29855152
23. Kandel H, Nguyen V, Ferdi AC, Gupta A, Abbondanza M, Sullivan L, 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(12):1581-1589. PubMed PMID: 33935236
24. Hafezi F, Kling S, Gilardoni F, Hafezi N, Hillen M, Abrishamchi R, et al. Individualized Corneal Cross-linking With Riboflavin and UV-A in Ultrathin Corneas: The Sub400 Protocol. Am J Ophthalmol. 2021;224:133-142. PubMed PMID: 33340508
25. Ahmet S, Yayla Akincilar G, Kirgiz A, Kandemir Besek N, Kemer Atik B, Topcu H, et al. Long-term results of accelerated corneal collagen crosslinking in paediatric patients with progressive keratoconus: 10-year follow-up. Eye (Lond). 2024;38(13):2522-2529. PubMed PMID: 38609652
26. Soeters N, Wisse RP, Godefrooij DA, Imhof SM, Tahzib NG. Transepithelial versus epithelium-off corneal cross-linking for the treatment of progressive keratoconus: a randomized controlled trial. Am J Ophthalmol. 2015;159(5):821-8.e3. PubMed PMID: 25703475
27. Lombardo M, Serrao S, Lombardo G, Schiano-Lomoriello D. Two-year outcomes of a randomized controlled trial of transepithelial corneal crosslinking with iontophoresis for keratoconus. J Cataract Refract Surg. 2019;45(7):992-1000. PubMed PMID: 31003798
28. Lombardo M, Giannini D, Lombardo G, Serrao S. Randomized Controlled Trial Comparing Transepithelial Corneal Cross-linking Using Iontophoresis with the Dresden Protocol in Progressive Keratoconus. Ophthalmology. 2017;124(6):804-812. PubMed PMID: 28283279
29. Elving S, Fredriksson A, Beckman Rehnman J, Behndig A. Randomized clinical trial comparing customized corneal crosslinking: epi-on in high oxygen and epi-off in room air for keratoconus. J Cataract Refract Surg. 2024;50(7):746-753. PubMed PMID: 38465837
30. Borchert GA, Kandel H, Watson SL. Epithelium-on versus epithelium-off corneal collagen crosslinking for keratoconus: a systematic review and meta-analysis. Graefes Arch Clin Exp Ophthalmol. 2024;262(6):1683-1692. PubMed PMID: 37938377
31. Mohammadpour M, Esfandiarifard S, Pourmirbabaei S, Bolourinejad P. A Systematic Review and Meta-Analysis Comparing Epithelium-On and Epithelium-Off Corneal Collagen Cross-Linking in Keratoconus Management: Integrating New Evidence. Cornea. 2025;44(12):1564-1579. PubMed PMID: 40856685
32. Shalchi Z, Wang X, Nanavaty MA. Safety and efficacy of epithelium removal and transepithelial corneal collagen crosslinking for keratoconus. Eye (Lond). 2015;29(1):15-29. PubMed PMID: 25277300
33. Jo JJ, Bian Y, Chamberlain WD, Chen M, Colby K, Dmitriev P, et al. Challenges to the Access and Utilization of Corneal Cross-Linking in Patients With Keratoconus in the United States. Am J Ophthalmol. 2026;292:58-62. PubMed PMID: 42526732
34. Raiskup F, Spoerl E. Corneal cross-linking with hypo-osmolar riboflavin solution in thin keratoconic corneas. Am J Ophthalmol. 2011;152(1):28-32.e1. PubMed PMID: 21529763
35. Hafezi F. Limitation of collagen cross-linking with hypoosmolar riboflavin solution: failure in an extremely thin cornea. Cornea. 2011;30(8):917-9. PubMed PMID: 21389853
36. Kymionis GD, Portaliou DM, Diakonis VF, Kounis GA, Panagopoulou SI, Grentzelos MA. Corneal collagen cross-linking with riboflavin and ultraviolet-A irradiation in patients with thin corneas. Am J Ophthalmol. 2012;153(1):24-8. PubMed PMID: 21861976
37. Blaser F, Torres-Netto EA, Gatzioufas Z, Perschak P, Hafezi F, Said S. Assessing Endothelial Integrity in Patients With Progressive Keratoconus and Thin Corneas Treated With the Sub400 Corneal Cross-linking Protocol. J Refract Surg. 2025;41(7):e682-e689. PubMed PMID: 40626429
38. Farrokhpour H, Soleimani M, Cheraqpour K, Masoumi A, Tabatabaei SA, Shahriari M, et al. A Case Series of Infectious Keratitis After Corneal Cross-linking. J Refract Surg. 2023;39(8):564-572. PubMed PMID: 37578174
39. Maharana PK, Sahay P, Sujeeth M, Singhal D, Rathi A, Titiyal JS, et al. Microbial Keratitis After Accelerated Corneal Collagen Cross-Linking in Keratoconus. Cornea. 2018;37(2):162-167. PubMed PMID: 29111996
40. Murchison CE, Petroll WM, Robertson DM. Infectious keratitis after corneal crosslinking: systematic review. J Cataract Refract Surg. 2021;47(8):1075-1080. PubMed PMID: 33769765
41. Kymionis GD, Portaliou DM, Bouzoukis DI, Suh LH, Pallikaris AI, Markomanolakis M, et al. Herpetic keratitis with iritis after corneal crosslinking with riboflavin and ultraviolet A for keratoconus. J Cataract Refract Surg. 2007;33(11):1982-4. PubMed PMID: 17964410
42. Pollhammer M, Cursiefen C. Bacterial keratitis early after corneal crosslinking with riboflavin and ultraviolet-A. J Cataract Refract Surg. 2009;35(3):588-9. PubMed PMID: 19251154
43. Greenstein SA, Fry KL, Bhatt J, Hersh PS. Natural history of corneal haze after collagen crosslinking for keratoconus and corneal ectasia: Scheimpflug and biomicroscopic analysis. J Cataract Refract Surg. 2010;36(12):2105-14. PubMed PMID: 21111314
44. Raiskup F, Hoyer A, Spoerl E. Permanent corneal haze after riboflavin-UVA-induced cross-linking in keratoconus. J Refract Surg. 2009;25(9):S824-8. PubMed PMID: 19772259
45. Koller T, Mrochen M, Seiler T. Complication and failure rates after corneal crosslinking. J Cataract Refract Surg. 2009;35(8):1358-62. PubMed PMID: 19631120
46. Raiskup F, Theuring A, Pillunat LE, Spoerl E. Corneal collagen crosslinking with riboflavin and ultraviolet-A light in progressive keratoconus: ten-year results. J Cataract Refract Surg. 2015;41(1):41-6. PubMed PMID: 25532633
47. Enders C, Vogel D, Dreyhaupt J, Wolf W, Garip-Kuebler A, Hall J, et al. Corneal cross-linking in patients with keratoconus: up to 13 years of follow-up. Graefes Arch Clin Exp Ophthalmol. 2023;261(4):1037-1043. PubMed PMID: 36197498
48. Sharma A, Nottage JM, Mirchia K, Sharma R, Mohan K, Nirankari VS. Persistent corneal edema after collagen cross-linking for keratoconus. Am J Ophthalmol. 2012;154(6):922-926.e1. PubMed PMID: 22959362
49. Bagga B, Pahuja S, Murthy S, Sangwan VS. Endothelial failure after collagen cross-linking with riboflavin and UV-A: case report with literature review. Cornea. 2012;31(10):1197-200. PubMed PMID: 22960648
50. Henriquez MA, Perez L, Hernandez-Sahagun G, Rojas RP, Stulting RD, Izquierdo L. Long Term Corneal Flattening After Corneal Crosslinking in Patients with Progressive Keratoconus. Clin Ophthalmol. 2023;17:1865-1875. PubMed PMID: 37425031
51. 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(4):585-93. PubMed PMID: 20138607
52. Hashemi H, Seyedian MA, Miraftab M, Fotouhi A, Asgari S. Corneal collagen cross-linking with riboflavin and ultraviolet a irradiation for keratoconus: long-term results. Ophthalmology. 2013;120(8):1515-20. PubMed PMID: 23583165
53. O'Brart DP, Patel P, Lascaratos G, Wagh VK, Tam C, Lee J, et al. Corneal Cross-linking to Halt the Progression of Keratoconus and Corneal Ectasia: Seven-Year Follow-up. Am J Ophthalmol. 2015;160(6):1154-63. PubMed PMID: 26307513
54. Greenstein SA, Yu AS, Gelles JD, Huang S, Hersh PS. Long-Term Outcomes After Corneal Cross-linking for Progressive Keratoconus and Corneal Ectasia: A 10-Year Follow-Up of the Pivotal Study. Eye Contact Lens. 2023;49(10):411-416. PubMed PMID: 37565471
55. Seifert FK, Theuersbacher J, Schwabe D, Lamm O, Hillenkamp J, Kampik D. Long-Term Outcome of Corneal Collagen Crosslinking with Riboflavin and UV-A Irradiation for Keratoconus. Curr Eye Res. 2022;47(11):1472-1478. PubMed PMID: 36173395
56. Vinciguerra R, Pagano L, Borgia A, Montericcio A, Legrottaglie EF, Piscopo R, et al. Corneal Cross-linking for Progressive Keratoconus: Up to 13 Years of Follow-up. J Refract Surg. 2020;36(12):838-843. PubMed PMID: 33295997
57. Cehelyk EK, Syed ZA. Long-term outcomes of corneal crosslinking. Curr Opin Ophthalmol. 2024;35(4):315-321. PubMed PMID: 38700950
58. Godefrooij DA, Gans R, Imhof SM, Wisse RP. Nationwide reduction in the number of corneal transplantations for keratoconus following the implementation of cross-linking. Acta Ophthalmol. 2016;94(7):675-678. PubMed PMID: 27213687
59. Sandvik GF, Thorsrud A, Råen M, Østern AE, Sæthre M, Drolsum L. Does Corneal Collagen Cross-linking Reduce the Need for Keratoplasties in Patients With Keratoconus?. Cornea. 2015;34(9):991-5. PubMed PMID: 25970434
60. Sklar JC, Wendel C, Zhang A, Chan CC, Yeung SN, Iovieno A. Did Collagen Cross-Linking Reduce the Requirement for Corneal Transplantation in Keratoconus? The Canadian Experience. Cornea. 2019;38(11):1390-1394. PubMed PMID: 31335524
61. Gad El Sayed M, Patel R, Hwang FS. TriNetX Analysis of Surgical Treatment Patterns of Keratoconus Before and After the Approval of Corneal Crosslinking in the United States. Cornea. 2026. PubMed PMID: 42228419
62. Sykakis E, Karim R, Evans JR, Bunce C, Amissah-Arthur KN, Patwary S, et al. Corneal collagen cross-linking for treating keratoconus. Cochrane Database Syst Rev. 2015;2015(3):CD010621. PubMed PMID: 25803325
63. Ng SM, Ren M, Lindsley KB, Hawkins BS, Kuo IC. Transepithelial versus epithelium-off corneal crosslinking for progressive keratoconus. Cochrane Database Syst Rev. 2021;3(3):CD013512. PubMed PMID: 33765359
64. Ng SM, Hawkins BS, Kuo IC. Transepithelial Versus Epithelium-Off Corneal Crosslinking for Progressive Keratoconus: Findings From a Cochrane Systematic Review. Am J Ophthalmol. 2021;229:274-287. PubMed PMID: 34048801
65. Salmon HA, Chalk D, Stein K, Frost NA. Cost effectiveness of collagen crosslinking for progressive keratoconus in the UK NHS. Eye (Lond). 2015;29(11):1504-11. PubMed PMID: 26315704
66. Godefrooij DA, Mangen MJ, Chan E, O'Brart DPS, Imhof SM, de Wit GA, et al. Cost-Effectiveness Analysis of Corneal Collagen Crosslinking for Progressive Keratoconus. Ophthalmology. 2017;124(10):1485-1495. PubMed PMID: 28532974
67. Leung VC, Pechlivanoglou P, Chew HF, Hatch W. Corneal Collagen Cross-Linking in the Management of Keratoconus in Canada: A Cost-Effectiveness Analysis. Ophthalmology. 2017;124(8):1108-1119. PubMed PMID: 28457614
68. Hansen LO, Garcia R, Torricelli AAM, Bechara SJ. Cost-Effectiveness of Corneal Collagen Crosslinking for Progressive Keratoconus: A Brazilian Unified Health System Perspective. Int J Environ Res Public Health. 2024;21(12). PubMed PMID: 39767411
69. Gupta AS, Yu Y, Orlin SE, VanderBeek BL. Real-world socioeconomic determinants of corneal crosslinking in a national cohort. J Cataract Refract Surg. 2024;50(3):217-223. PubMed PMID: 37847112
70. Angelo L, Gokul A, Wadhwa H, McGhee CNJ, Ziaei M. Assessment of Health Disparities and Barriers to Accessing Keratoconus and Crosslinking at a Tertiary Referral Center. Cornea. 2024;43(3):277-284. PubMed PMID: 37098115

PubMed Topic Searches

1. Corneal cross-linking for keratoconus: randomized trials
2. Epithelium-on versus epithelium-off corneal cross-linking
3. Accelerated corneal cross-linking for keratoconus
4. Complications of corneal cross-linking
5. Cross-linking in thin corneas

Back to Table of Contents

Connections

• Keratoconus
• Ophthalmology
• Vitamin B2 (Riboflavin)
• Corneal Ulcer
• Corneal Dystrophy
• Ultraviolet Radiation
• Eczema
• Blurred Vision

Back to Table of Contents

Featured Videos