Being told your cornea is too thin to treat is one of the harder moments in a keratoconus journey. For years, the standard cross-linking recipe simply stopped being an option once the cornea dropped below a set thickness – and the people affected were often the ones who needed help most. A narrative review in Die Ophthalmologie now sets out how ELZA-sub400 cross-linking changes that calculation by fitting the treatment to each cornea instead of the other way around (Hafezi et al., 2026, doi:10.1007/s00347-026-02467-z).
Key takeaways
- Conventional cross-linking needs a cornea at least 400 µm thick, which rules out many patients with advanced keratoconus.
- ELZA-sub400 individualises the UV dose to the measured corneal thickness, so cross-linking (CXL) becomes possible in corneas as thin as 214 µm.
- In a published prospective study of 39 eyes, 90 percent stayed stable at one year with no sign of endothelial damage.
- The honest caveat: this is early evidence – 39 eyes, single centre, 12-month follow-up – and larger, longer, comparative trials are still needed.
What is keratoconus, and why does corneal thickness matter?
Keratoconus is a condition in which the cornea – the clear front window of the eye – gradually thins and bulges into a cone shape, blurring and distorting vision. The main treatment to stop it progressing is corneal cross-linking (CXL), which stiffens the cornea using riboflavin (vitamin B2) drops activated by ultraviolet (UV) light. The original recipe, the Dresden protocol, requires the cornea to be at least 400 µm thick. That threshold exists to protect the endothelium, the delicate innermost cell layer that cannot regenerate, from UV damage. In practice, the treatment cross-links roughly the top 330 µm and leaves about 70 µm untreated as a safety buffer above the endothelium – think of it as a margin of protection at the base of the cornea.
Why standard cross-linking excludes thin corneas
The problem is that keratoconus itself thins the cornea, so the patients with the most advanced disease are often below the 400 µm line and are turned away – the exact group that most needs stabilising. Surgeons have tried several workarounds. One swells the cornea with hypo-osmolar riboflavin to reach 400 µm, but the swelling is variable and hard to predict. Another places a riboflavin-soaked contact lens on the eye, which the authors note reduces the available oxygen and weakens the strengthening effect. A third leaves an island of surface cells over the thinnest spot, at the cost of an uneven result. Each adapts the cornea to fit the technique.
How ELZA-sub400 cross-linking is different
ELZA-sub400 cross-linking reverses that logic: it adapts the technique to the cornea. Rather than forcing every eye to 400 µm, the protocol measures the stromal thickness during surgery and individualises how long the UV light is applied. The approach rests on an algorithm – developed with corneal biomechanics researcher Sabine Kling – that combines two physical laws: Fick’s law, which describes how riboflavin and oxygen diffuse through the cornea, and the Lambert-Beer law, which describes how the tissue absorbs UV energy. From the measured thickness, the algorithm sets the irradiation time needed for effective cross-linking while still preserving the 70 µm safety margin over the endothelium. The protocol keeps the gentle, non-accelerated 3 mW/cm² epi-off setting, which also means it runs on most cross-linking devices already in clinics rather than requiring new hardware. Prof. Farhad Hafezi and colleagues at the ELZA Institute in Zurich led the work.
What the ELZA-sub400 cross-linking study showed
It is worth being clear about what kind of paper this is: a review that gathers the published clinical evidence rather than a new trial. The central data come from a prospective study of 39 eyes with corneas measuring 214–398 µm. At one year, 90 percent (35 of 39 eyes) were tomographically stable, defined as a change in maximum corneal steepness (Kmax) of less than 1.0 D. On average, Kmax flattened by −2.06 ± 3.66 D, from 58.5 ± 7.6 D at baseline to 56.4 ± 7.8 D at 12 months (p = 0.001), and there were no signs of endothelial decompensation. The protocol has also been used once in keratoglobus – a rarer condition in which the whole cornea thins – where a single case remained stable over 32 months (Kmax 81.90 → 80.60 D) with no endothelial damage. We describe this experience on our cross-linking for keratoglobus page.
What this could mean for you
- If you have been told your cornea is too thin for cross-linking, that door may not be fully closed – individualised protocols are designed for exactly this situation.
- Because the method uses a standard 3 mW/cm² light source, it does not depend on a specialised machine.
- Cross-linking aims to halt progression and can modestly flatten the cornea; it is not a way to reverse keratoconus or replace glasses and contact lenses.
The honest limits of the evidence
The authors are candid about what is not yet settled, and it matters. The prospective study included 39 eyes – a meaningful but small number. There is, as yet, no large multicentre randomised controlled trial comparing ELZA-sub400 against the older swelling method. Follow-up beyond five years is still needed to confirm the effect lasts in these very thin corneas. The protocol also depends on accurate ultrasound pachymetry measured at the right moment during surgery; a mismeasurement could mean over- or under-treatment, so the surgeon must be experienced with the technique. Whether the 70 µm safety margin is strictly necessary is increasingly debated, and reducing it might extend the range further – but the authors stress that would require robust safety data first. They also note that in very advanced disease with poor vision and contact-lens intolerance, a corneal transplant or CAIRS (corneal allogenic intrastromal ring segments) may be the more appropriate first step. Newer variants – a second-generation protocol and an epi-on version that leaves the surface cells intact – are in development and are not yet standard care.
Questions worth asking your ophthalmologist
- How thin is my cornea at its thinnest point, and how was that measured?
- Is my keratoconus actually progressing, which is what cross-linking is meant to stop?
- Am I a candidate for an individualised thin-cornea protocol, or would another option suit me better?
- What are the realistic goals here – stabilising my vision, and what it will not do?
- How much experience does the centre have with cross-linking very thin corneas?
- What follow-up will I need, and over what period?
Frequently asked questions
Can cross-linking be done if my cornea is under 400 µm?
Sometimes. Standard cross-linking needs 400 µm, but individualised protocols such as ELZA-sub400 are specifically designed to treat thinner corneas by tailoring the UV dose. Whether it suits you depends on your eyes and is a decision for your ophthalmologist.
Does ELZA-sub400 cross-linking reverse keratoconus?
No. The goal is to halt progression. Some flattening of the cornea can occur, but cross-linking is not a vision-correction procedure and does not remove the need for glasses or contact lenses.
Is it safe for the inside of the eye?
The protocol keeps a 70 µm untreated safety margin above the endothelium, and in the published study of 39 eyes there were no signs of endothelial damage. Long-term data in very thin corneas are still being gathered.
How thin a cornea can be treated?
In the prospective study, corneas as thin as 214 µm were treated. The thinnest safe limit for any individual depends on their specific anatomy.
Do I need a special machine for this?
No. ELZA-sub400 uses a standard 3 mW/cm² UV source and ultrasound pachymetry, so it works on most cross-linking devices already in use.
The ELZA Institute in Zurich uses individualised cross-linking protocols for thin corneas as part of its keratoconus care. You can read more on our cross-linking for thin corneas page or arrange a consultation to discuss whether it applies to your eyes.
References
- Hafezi F, Hillen M, Kling S, Hafezi NL, Aydemir ME, Torres-Netto EA. Cross-linking treatment of ultrathin corneas: the ELZA-sub400 protocol. Die Ophthalmologie. 2026 (online ahead of print). doi:10.1007/s00347-026-02467-z
- Hafezi F, Kling S, Gilardoni F, et al. Individualized corneal cross-linking with riboflavin and UV-A in ultrathin corneas: the Sub400 protocol. Am J Ophthalmol. 2021;224:133–142. doi:10.1016/j.ajo.2020.12.011
- American Academy of Ophthalmology. Corneal cross-linking (patient information). aao.org
