Keratoconus Treatment: The Last 10 Years, From Cross-Linking to CAIRS

Keratoconus is a progressive disease in which the cornea — the clear front surface of the eye — thins and bulges forward into a cone. It usually begins in adolescence and, left untreated, can permanently degrade vision.

But where we stand today bears no comparison to ten years ago. Keratoconus is no longer a progressive and inevitable disease; it is one that can be halted and whose shape can be corrected.

1. Diagnosis: we now look for the tendency, not the cone

The greatest change happened not in the operating room but in the imaging suite.

Ten years ago, diagnosing keratoconus meant seeing clear thinning and clear steepening — the disease had to be established. Today, the triad of Scheimpflug tomography, epithelial thickness mapping and biomechanical assessment lets us catch the cornea before vision loss has begun.

The most striking recent finding came from screening: among the entirely asymptomatic children of keratoconus patients, 14.3% showed early signs of the disease.¹ If keratoconus runs in the family, the child should be screened without waiting for symptoms.

AI-assisted diagnosis is no longer a promise but a daily tool. A recent review of 107 studies found that models combining tomography, biomechanics and epithelial data meaningfully improve detection of early cases.²

2. Cross-linking (CXL): the treatment that halts the disease

Corneal cross-linking, which began with the Dresden protocol in 2003, is the single most important turning point in the history of keratoconus. For the first time we changed the course of the disease.

Riboflavin (vitamin B2) is applied to the cornea and activated with controlled ultraviolet-A light. The reaction creates new bonds between collagen fibres, strengthening the tissue and preventing further steepening of the cone.

Which protocol is better?

• A meta-analysis of 14 randomised controlled trials found conventional CXL superior for flattening and demarcation line depth, while accelerated protocols better preserve corneal thickness. No meaningful difference in visual acuity, refraction, biomechanics or endothelial cell counts.³
• A ten-year review is unambiguous: epithelium-off Dresden CXL remains the efficacy benchmark.⁴
• Six-year follow-up of high-energy pulsed accelerated CXL showed a 1.14 D reduction in maximum keratometry, sustained visual gain and a mean demarcation line depth of 261 µm, with no retreatment required.⁵

Epithelium-on cross-linking is now approved

The field's most recent milestone: on 20 October 2025 the FDA approved an oxygen-enriched topical cross-linking therapy that does not require epithelial removal, on the basis of two Phase 3 trials randomising more than 400 patients.

3. Ring segments: from ICRS to CAIRS

There is a critical distinction here, and it is the point patients most often confuse:

These two treatments are not alternatives but complements. In any patient with suspected progression, cross-linking should be planned alongside segment surgery.

Why the move from synthetic to biological?

Thin plastic (PMMA) ring segments introduced in the 2000s filled this gap. But synthetic material carried known problems: extrusion, migration, stromal thinning over the segment, and halo.

CAIRS (Corneal Allogenic Intrastromal Ring Segments), described in 2018, changed the equation: instead of a synthetic implant, segments of living tissue prepared from human donor cornea. Biocompatible, non-extruding, removable when needed, and shapeable to the individual patient.

What do the numbers say?

• A systematic review and meta-analysis pooling 14 studies and 442 eyes found CAIRS improved corrected distance visual acuity by 0.37 logMAR and uncorrected acuity by 0.43 logMAR, with 4.59 D improvement in spherical equivalent and 4.49 D reduction in maximum keratometry. Across the whole series there was a single severe adverse event (0.2%).⁷
• When CAIRS was compared directly against deep anterior lamellar keratoplasty (DALK) in advanced keratoconus, both produced significant improvement (all p < 0.001). Postoperative corrected acuity was better in the DALK group (p = 0.027), while uncorrected acuity, refractive and keratometric values were comparable.⁸ CAIRS is therefore not equivalent to transplantation at this stage — but it is a powerful, reversible option that defers it.
• In a Turkish series of 62 eyes in 49 patients followed for three years, visual outcomes improved markedly with no major complications.⁹

4. The age threshold is coming down

A decade ago, discussing segment surgery in a young patient was close to taboo — "let them grow up first," we would say. That taboo has fallen.

• In 107 eyes of 85 patients, all aged 18 or under, success rates after segment implantation were 84.3% at 5 years, 85.5% at 7 years and 88.9% at 10 years.¹²
• A seven-year analysis of segments combined with epithelium-off accelerated cross-linking in 73 paediatric eyes showed durable improvement.¹³

We can therefore now say that in a teenager with keratoconus, segment surgery is an option supported by ten-year data. We no longer have to make a child wait — and this is the age group in which the disease progresses fastest.

An honest boundary: these long-term data concern synthetic segments. No long-term paediatric CAIRS series has yet been published. That said, the fact that allogenic tissue does not extrude, can be removed and leaves no permanent foreign body in a growing cornea makes CAIRS conceptually better suited to this age group.

5. Keratoplasty: still here, but steadily less

One of the most telling indicators of the past decade: keratoconus is receding among indications for corneal transplantation, and the transplants performed are shifting from full-thickness to lamellar.

In a 20-year analysis of 3,281 corneal transplants from Türkiye, the share of keratoconus among indications fell while the odds of lamellar keratoplasty rose 1.59-fold per five-year interval.¹⁴

Meanwhile, techniques that add tissue to the cornea rather than removing it are maturing: Bowman layer transplantation, corneal tissue addition keratoplasty (CTAK) and stromal lenticule addition keratoplasty (SLAK). The shared conclusion of international reviews is that these techniques delay or entirely eliminate the need for full-thickness transplantation.¹⁵

6. The new framework: Stabilise – Reshape – Replace

A comprehensive review in one of the field's most respected journals gathers all of this into a single framework: Stabilise – Reshape – Replace.¹⁶

Transplantation is no longer the first step; it is the third.

And the hardest of these three steps is not the third — it is the first. The technology available to us has advanced faster than the moment at which patients reach us.

Frequently asked questions