Corneal cross-linking: the treatment that can slow keratoconus down

If you have just been told your keratoconus is progressing, the next word was probably cross-linking. Corneal collagen cross-linking (CXL) is the only treatment with randomised-trial evidence that it can slow keratoconus down. Here is what the trials measured, what the numbers mean for you, and how to reach the right person at the right time.

What keratoconus is doing

In keratoconus the cornea, the clear front window of the eye, thins and bulges into an irregular cone, which is why glasses stop working and your prescription keeps shifting. It affects roughly 1 in 2,000 people and changes fastest in the teens and twenties. Around a quarter (26.8%) of people with keratoconus eventually need a corneal transplant, which is why catching progression early matters.

What cross-linking does

All the trials below used the standard “Dresden” method: the surface layer of the cornea is removed, 0.1% riboflavin (vitamin B2) drops are applied, and ultraviolet-A light is shone on the cornea for about 30 minutes. Together they create new bonds (cross-links) between collagen fibres, making the cornea stiffer and more resistant to bulging. It is an outpatient procedure of about an hour.

Cross-linking is designed to reduce the chance of further progression. It does not reverse keratoconus, rebuild shape already lost, or by itself give sharp vision. A modest flattening afterwards is a bonus, not the goal.

What the trials found

A 2015 Cochrane review found three randomised trials (Australia, the UK and the USA) comparing cross-linking with no treatment: 219 eyes analysed, 119 treated and 100 untreated, all adults with documented progression and a cornea at least 400 microns thick in two of the trials (Sykakis et al., 2015). The trials measured outcomes differently and could not be pooled, so each result comes from a single study:

• Progression at 3 years. In the Australian trial, none of 46 treated eyes steepened by 2 dioptres or more, compared with 19 of 48 untreated eyes (relative risk 0.03, 95% CI 0.00 to 0.43).
• Corneal steepness at 12 months. Treated corneas were on average 1.92 D flatter than untreated ones (95% CI 1.30 to 2.54; 94 eyes).
• Unaided vision at 12 months. About two lines (10 letters) better in treated eyes (0.20 logMAR; 95% CI 0.09 to 0.31).
• Corneal thickness. Results were inconsistent between trials.

A 2017 review by Kobashi and Rong widened the net to five randomised trials (289 eyes) with one-year follow-up. All five showed the steepest part of the cornea (Kmax) was less steep after cross-linking, though the results were too varied to combine into one figure. Best glasses-corrected vision was 0.09 logMAR better in treated eyes (95% CI 0.04 to 0.14; P = 0.0005), which the authors point out is less than one chart line and within normal test-to-test variation. Corneal thickness (mean difference 1.46 microns, 95% CI −2.27 to 5.68) and astigmatism (−0.25 D, 95% CI −0.76 to 0.26) did not differ significantly from untreated eyes.

What a confidence interval means. The 95% CI is the range the true effect probably sits in. For the 3-year progression result the interval runs from 0.00 to 0.43, so the true risk reduction could be anywhere from very large to about 57%, but the whole range favours cross-linking. When an interval crosses zero (as for thickness and astigmatism above), the study cannot tell whether there was any effect.

Side effects

Cochrane reported side effects in roughly 1 in 10 treated participants: 3 of 12 in the UK trial and 3 of 50 over three years in the Australian trial. They included corneal swelling, inflammation inside the front of the eye, recurrent surface erosions, small infiltrates and, in one person, blood vessels growing into the corneal edge. Most were temporary and of low clinical significance; none occurred in untreated eyes.

What the studies can’t tell us

Cochrane rated the certainty of every outcome as very low. In plain language, the true effect could turn out quite different from the numbers above, even though the potential benefit looked large. Why: the trials were small with wide confidence intervals; patients and most assessors knew who had been treated; 21 control eyes left the Australian trial early (12 crossed over to cross-linking, 5 had transplants, 4 withdrew); and three unpublished trials were found, raising concern about publication bias. Kobashi and Rong add that three of their five trials used the patient’s other eye as the control, questionable in an asymmetric disease. No trial included children, measured quality of life or cost, followed people beyond three years, or tested newer epithelium-on or accelerated protocols.

Timing: why the corneal map matters

Every trial only enrolled eyes with documented progression. Kobashi and Rong’s definition is a useful benchmark: steepening of 0.75 D or more, an increase in astigmatism of 0.75 D or more, or a drop in vision over 12 months. Progression can only be shown by comparing corneal maps over time, so a baseline map as soon as keratoconus is suspected, and repeat maps at sensible intervals, are the foundation of the whole pathway.

Start with your optometrist

The procedure is performed by an ophthalmologist (eye surgeon), but the journey does not start there. The front door is an optometrist who focuses on keratoconus:

• Diagnosis and mapping. Corneal tomography confirms keratoconus and sets a baseline.
• Documenting progression. Sequential maps are the evidence a surgeon needs, and they catch change early, while the cornea still has thickness to spare.
• Referral at the right moment. A keratoconus-focused optometrist refers you to a corneal specialist with imaging and history prepared.
• Vision before and after. Cross-linking aims to protect shape; it is scleral lenses, fitted by your optometrist, that restore clear vision around it.

You do not need a GP referral to see an optometrist.

Being assessed in New Zealand

Rose Optometry in Hamilton runs a dedicated keratoconus clinic with Pentacam and ANTERION corneal mapping, monitors progression over time, refers to corneal surgeons when cross-linking is indicated, and fits the lenses that restore vision. It sees patients and accepts referrals from across New Zealand and, with the New Zealand Eye Research Centre, is the birthplace of the Rose K lens.

If you or someone in your whānau has keratoconus, or a prescription that will not settle, the first step is a corneal map with an optometrist who knows the condition. You can book a keratoconus assessment directly, or ask your own optometrist to refer you.

This article is general information, not a substitute for a clinical assessment. Whether cross-linking is right for you depends on your corneal measurements over time and an examination by a qualified clinician.

References

Sourced and verified via PubMed.

• Sykakis E, Karim R, Evans JR, et al. Corneal collagen cross-linking for treating keratoconus. Cochrane Database of Systematic Reviews. 2015;(3):CD010621. PMID 25803325. doi:10.1002/14651858.CD010621.pub2
• Kobashi H, Rong SS. Corneal Collagen Cross-Linking for Keratoconus: Systematic Review. BioMed Research International. 2017;2017:8145651. PMID 28691035. doi:10.1155/2017/8145651