Too Thin for Cross-Linking? Five-Year Results of a Tailored Protocol

Some people with keratoconus are told their cornea is "too thin" for standard cross-linking. That can be a worrying thing to hear. Standard protocols generally need a minimum thickness of corneal tissue to protect the delicate inner layer of cells, the endothelium, from the ultraviolet light used in treatment. A 2026 study from Zurich reports five-year results of an approach designed for exactly these corneas.

If you're new to cross-linking, our guide to corneal cross-linking in New Zealand explains what the treatment involves, so we'll focus on the new evidence here.

What the researchers did

The protocol is called ELZA-sub400. Instead of using one standard dose of UV-A light, the treatment time is tailored to each person's stromal thickness (the main supporting layer of the cornea). The aim is for the treatment effect, seen afterwards as a "demarcation line", to sit about 70 µm in front of the endothelium, leaving a safety margin.

The team looked back at the records of 39 eyes of 33 people with progressive keratoconus or primary ectasia whose stroma was between 214 and 398 µm thick, that is, under 400 µm. Each eye had a single epithelium-off treatment: the surface layer was removed, riboflavin drops were applied, and UV-A light was shone at 3 mW/cm² for an individually calculated time. Eyes were checked yearly with Scheimpflug tomography (Pentacam HR) for up to five years.

This was a retrospective case series, which means the researchers reviewed past cases rather than setting up a trial in advance. There was no control group of untreated eyes or eyes treated another way.

What they found

Average follow-up was 3.6 ± 1.5 years (range 1 to 5 years). Importantly, only 18 of the 39 eyes (46.2%) completed the full five years.

• Stability: the authors report that 84.6% of eyes (33/39) remained stable or flattened, defined as Kmax (the steepest point of the cornea) changing by 1.00 dioptre or less. That implies 6 of 39 eyes did not meet this definition. The abstract doesn't explain how eyes with shorter follow-up were counted in this five-year figure.
• Corneal steepness: Kmax decreased from baseline by an average of −2.74 ± 2.65 D (P < .001), and the central 3 mm curvature measure also decreased (−1.29 D, P = .014).
• Vision: median corrected distance vision (CDVA) was unchanged at 0.40 logMAR (P = .572), and glasses prescription did not change significantly (sphere P = .211; cylinder P = .280).
• Thickness: the thinnest point of the cornea thinned slightly (median −10 µm, P = .049).
• Clarity: corneal haze readings (anterior densitometry) rose slightly (+2.40 greyscale units, P = .007), which the authors describe as within mild clinical ranges.
• Safety: the demarcation line averaged 275 ± 61 µm deep, and no endothelial decompensation (failure of the inner cell layer) occurred.

A note on P-values: P < .001 means a change this large would be very unlikely if nothing had really changed. P = .572 means the study found no evidence that vision changed, which is not quite the same as proving it stayed identical. Because there was no comparison group, these are before-and-after changes within the same eyes; there are no between-group results to report.

What this means, in plain terms

The results suggest that, in experienced hands, tailored cross-linking may help slow or stabilise keratoconus in some corneas once considered too thin to treat. It is not a way of reversing keratoconus or improving sight: average vision stayed where it was. The goal, as with all cross-linking, is to reduce the chance of further change.

Whether this kind of protocol is suitable or available for a particular person is a decision for the corneal ophthalmologist. Life after treatment matters too, including follow-up scans and returning to lenses; see what happens after cross-linking.

Where this fits in your care

Cross-linking is performed by an ophthalmologist, but the front door is a keratoconus-focused optometrist. At Rose Optometry in Hamilton, our optometrists, including Jagrut Lallu and Jacqueline Rowe, diagnose keratoconus, document progression with sequential corneal maps, refer to an ophthalmologist for cross-linking at the right moment, and fit scleral lenses before and after treatment. Thin corneas are one more reason to monitor closely and refer early. Read more about our keratoconus care in Hamilton.

What the studies can't tell us

• Only the abstract was available to us. We could not read the full paper, so funding, conflicts of interest, dropouts and side-effects beyond those listed could not be checked.
• Developers evaluating their own protocol. The study comes from the ELZA Institute in Zurich, whose name the protocol carries, so this is likely the protocol's own team assessing it. The abstract does not state funding or conflicts of interest.
• No control group and retrospective design. Without a comparison, we can't tell how these eyes would have done otherwise, or how this compares with other approaches.
• Incomplete five-year data. Fewer than half the eyes reached five years, so the 84.6% headline rests on incomplete follow-up.
• Too small for rare problems. With 39 eyes, uncommon but serious complications could easily be missed. Endothelial cell counts were not reported in the abstract.

If you've been told your cornea is thin, or you're unsure whether your keratoconus is changing, you can book a keratoconus assessment with our optometrists in Hamilton.

References

• Torres-Netto EA, Kling S, Akcan RE, et al. Five-Year Follow-up of the ELZA-sub400 Protocol: Individualized Corneal Cross-linking With Riboflavin and UV-A in Ultrathin Keratoconic Corneas. J Refract Surg. 2026;42(10):e416-e424. PMID: 42842445. https://doi.org/10.3928/1081597X-20260728-01