This is the fourth in a series of articles that summarises an afternoon of presentations, plus a debate, on the topic of corneal ectasia that was held as part of the recent 29th Cologne Adventsymposium, an annual meeting organised by Laserforum e.V., in Cologne, Germany. Chaired by Drs Omid Kermani and Georg Gerten, the presentations each considered the condition from a different perspective and are available online.

Part 1, published in the June 2022 issue of Ophthalmology Times Europe®, considered the contribution of eye rubbing to the development of keratoconus. Part 2, published in the July/August 2022 issue, looked at developments in diagnostics and classification methods for ectatic corneal diseases. Part 3, published in the September 2022 issue, covered epithelial imaging. In this final instalment of the series, we look at recent developments in corneal cross-linking. This is based on the presentation that Prof. Farhad Hafezi gave during the event.

Reviewed by Dr Omid Kermani and Prof. Farhad Hafezi.

Figure 1: Comparison of corneal stiffening achieved with different epi-off CXL irradiation protocols: blue is the Dresden protocol (30 minutes at 3 mW/cm2); red is 10 minutes at 9 mW/cm²; green is 5 minutes at 18 mW/cm2; purple is control with no cross-linking. (Image courtesy of Prof. Hafezi, adapted from Hammer et al. 2014¹)

Prof. Hafezi began his presentation by explaining that there are three essential components for every corneal cross-linking (CXL) procedure: UV light, riboflavin and oxygen. When CXL was first developed in the early 2000s, it involved corneal epithelial cell debridement (making it an “epi-off” procedure) to enable riboflavin to be applied to the cornea and saturate the stroma.

This was followed by the delivery of UV, total dose (fluence) 5.4 J/cm², by irradiating the stroma with 3 mW/cm²-intensity light for 30 minutes. This is known as the Dresden protocol.

One of the basic laws of photochemistry is the Bunsen–Roscoe law of reciprocity. If all reagents are in excess, then the amount of reaction that occurs depends solely on fluence—not the speed at which the light energy is delivered. Under these conditions, doubling light intensity and halving the irradiation time should have the same effect.

But would this principle hold true in CXL? Prof. Hafezi’s group tested the biomechanics of several different accelerated CXL protocols (Figure 1) and found a 10-minute protocol that, although it induces significantly less corneal stiffening than the Dresden protocol, is enough to stabilise a keratoconus in most cases.¹

Figure 2: Comparison of stiffening results from different epi-off cross-linking protocols. (Image adapted from Abrishamchi et al. 2021³)

Prof. Hafezi explained that “the reason that increasing acceleration of the cross-linking reaction results in decreasing stiffening effects is that the speed of oxygen diffusion into the stroma is the rate-limiting factor in cross-linking. If the cross-linking reaction is too fast, the oxygen cannot diffuse in fast enough and efficacy is lost, unless intensity is increased to compensate.” If there is no oxygen in the cornea, no stiffening occurs at all.²

Until recently, Prof. Hafezi always reverted to the 30-minute Dresden protocol for particularly aggressive forms of keratoconus, such as in a young child or an ectasia after LASIK. However, he has now developed and published an accelerated epi-off protocol that compensates for the acceleration-related loss of stiffness by increasing fluence (Figure 2). This new protocol achieves the same stiffness as the Dresden protocol while saving the surgeon 20 minutes of time.³

Epi-on

“The holy grail for cross-linking would be a truly effective epi-on procedure,” Prof. Hafezi said. Leaving the epithelium intact would mean faster healing, less haze and an even lower infection risk. “The quest for good, efficient epi-on has been going on for more than a decade. What do we need to make it work?”

If the epithelium remains in place, then the UV irradiation needs to be increased because the epithelium blocks around 15 to 20% of the light. This is relatively easy: increase the fluence from 5.4 to 7.2 J/cm².

The next element is riboflavin. Riboflavin is a relatively large molecule, so transporting it through the tight junctions of the epithelium is not easy. One method is iontophoresis, which uses an electrical current to encourage transport.

The third factor is oxygen diffusion. The presence of the epithelium forms yet another barrier to oxygen diffusion into the cornea. It has been proposed that pulsing the light on and off might be effective: the idea is that oxygen is consumed when the light is on, so the oxygen can re-diffuse during the light’s off period.

The next question was: what happens with a slow procedure? Irradiating for 1 hour, instead of 30 minutes, with half the intensity, produces very efficient epi-on cross-linking.⁴ Prof. Hafezi said, “This was a proof of principle, and as a scientist I am happy—but as a clinician I don’t want to spend an hour in the operating room.”

One approach to increasing the amount of oxygen in the cornea during CXL, goggles that are fed a continuous stream of 100% oxygen, was proposed 2–3 years ago. However, Prof. Hafezi was not keen: “This makes sense. It should work. I don’t like it. Why? Because it is additional technology and I want to simplify the procedure, not make it more complicated and more expensive.”

“It is not necessary to have 100% oxygen surrounding the eye; what is needed is enough oxygen in the stroma for the cross-linking reaction to proceed effectively. What was needed was an intelligent combination of everything we have learned so far.”

There is now a published epi-on protocol that consists of normal air (no oxygen boost), iontophoresis to push riboflavin into the cornea, pulsed light, gentle acceleration, high fluence, and there are 3-year follow-up clinical data showing that it is effective.⁵

The next question was, could epi-on CXL be undertaken without iontophoresis? Prof. Hafezi’s group has been assessing the ability of penetration enhancers to get riboflavin to pass through the epithelial cell tight junctions, spending almost a year at the Swiss Federal Institute of Technology using multiphoton microscopy to assess the extent of riboflavin penetration, in order to find a penetration enhancer combination that achieves a stromal riboflavin concentration that is similar to that achieved with epi-off CXL.

This means there is now a gently accelerated high-fluence epi-on protocol, which takes approximately 30 minutes in total (including 15 minutes of irradiation), that can be performed at the slit lamp. This protocol is currently undergoing clinical evaluation.

Thin corneas

Prof. Hafezi reminded the audience that the original Dresden protocol required a minimum stromal thickness of 400 µm in order to protect the corneal endothelium from being exposed to potentially damaging levels of UV irradiation. Historically, surgeons have artificially increased the corneal thickness prior to treatment, either by swelling the cornea with hypo-osmolaric riboflavin⁶ or through the use of a riboflavin-soaked contact lens.⁷

However, these methods have disadvantages. The amount of swelling achieved by the hypoosmolar riboflavin is unpredictable, and, because contact lenses form an additional barrier to oxygen diffusion, their use reduces the stiffening efficacy by around 30%.

What Prof. Hafezi proposes now is what he describes as individualised cross-linking. Previously, surgeons adapted the cornea to the CXL protocol; now the technique can be adapted to each patient’s cornea. His research group modelled the interaction of oxygen, UV light, riboflavin and the stroma, and made an algorithm to predict the extent of cross-linking based on these parameters.⁸

Using this, they developed the “sub400” protocol, which uses measurements of the thinnest point of the cornea to calculate the optimal duration of a 3 mW/cm² intensity irradiation (via a simple look-up table) in an otherwise standard epi-off CXL protocol that leaves the same un-cross-linked stromal safety margin as the Dresden protocol to avoid harming the endothelium.

Clinical results have been very encouraging, with around 80% of corneas having successful prevention of ectasia progression after 1 year, even in corneas as thin as ~200 µm that would otherwise have gone on to require keratoplasty.⁹

This opens up a possibility for a new indication for CXL, keratoglobus. Prof. Hafezi has published a case report of sub400 CXL being used to successfully halt disease progression.¹⁰

Figure 3: The C-eye portable cross-linking device in use. (Images courtesy of Prof. Hafezi)

Describing his aims, Prof. Hafezi said, “We developed the sub400 protocol and epi-on for one reason: we want to democratise access to cross-linking, moving away from an expensive infrastructure toward equipment we all share, which is the slit lamp, and this is why we developed the C-eye device” (EMAGine AG, Zug, Switzerland; Figure 3).

According to Prof. Hafezi, the device is “small but high-tech. It can treat at the slit lamp with more than 11 different protocols—for keratoconus, for refractive, for keratitis—so if you ask me how often I go back to the laying position, I do, but only for special cases such as patients that are not compliant or who require general anaesthesia.” Figure 4 shows the view through the slit lamp.

Treatment at the slit lamp

Figure 4: Surgeon’s view of the cornea through C-eye during slit lamp cross-linking. (Images courtesy of Prof. Hafezi)

What about riboflavin distribution when treating at the slit lamp: is it affected by gravity? Prof. Hafezi said, “The short answer is yes but it takes an hour before the first changes become apparent, which is more than enough time to perform cross-linking”.¹¹

With regards to fixation, it is much better with an upright patient who is comfortable and has a fixation light on the other eye so they know where to look. However, is the Dresden protocol, which takes half an hour, achievable?

“We don’t need to [perform this],” explained Prof. Hafezi. “Most cross-linking procedures take around 10–12 minutes; PACK [photo-activated chromophore for keratitis] cross-linking for keratitis takes around 6 minutes, and that’s easily feasible, especially if you give the patient the surgeon’s chair so they are really comfortable and spend 30 seconds at the beginning to make sure that the height is well adjusted.”

He added: “The last question is the most important: can we leave the operating room for the procedure room or even the examination room? Yes, we can. With epi-on there is no question at all: the epithelium is on. With epi-off, you do not only stiffen the cornea but also sterilise it, so the answer is yes.”

The trend of the past 4 or 5 years has been for office-based procedures. Intravitreal injections¹² and cataract surgery¹³ have been shown to be safe performed in the procedure room, or even the examination room, with no increased risk for endophthalmitis, so cross-linking in the procedure room should be possible.

Prof. Hafezi concluded his presentation by reiterating that cross-linking at the slit lamp broadens access to the procedure.¹⁴ It is suitable for all indications—epi-on, epi-off, keratoconus, PACK cross-linking—and is available to patients with a variety of special needs.

The debate

Following Prof. Hafezi’s presentation, its content was debated with the chairs and other presenters—Dr Omid Kermani (Germany), Dr Georg Gerten (Germany), Prof. Damien Gatinel (France), Prof. Renato Ambrósio Jr. (Brazil), Prof. Dan Reinstein (UK) and Dr Tim Archer (UK)—and with members of the online audience.

Prof. Kermani: It seems that we are coming to a time when cross-linking can be performed in the practice room. I am sure our audience wants to learn a little bit more about this technology. How do you soak the cornea when the patient is sitting upright? And how do you make sure that, due to the position of the patient, the lower cornea is not soaked with more riboflavin than the upper cornea?

My second question is just to clarify epi-on or epi-off: do you do epi-on cross-linking with the C-eye device? How do you avoid the necessity of doing iontophoresis?

Prof. Hafezi: This procedure is like any type of surgery and comprises multiple steps that come one after the other. We looked at each step and tried to predict potential obstacles.

We instil riboflavin when the patient is in a simple reclining chair at an angle of around 20°. The patient then comes to the slit lamp and is positioned vertically: we showed about 4 years ago¹⁵ that we have 60 minutes from the moment when the cornea becomes vertical until you see the first gradient between upper and lower cornea. You are right that a gradient develops and more riboflavin accumulates in the lower cornea but the difference after 1 hour is 2%, so you have plenty of time.

The third thing is that we should not change too many paradigms at the same time, so the C-eye device is created to make the procedures simple and efficient and to democratise access; any type of epi-off procedure can be performed.

We thought about how to remove the epithelium at the slit lamp for a long time because, even if a surgeon has been doing this with the hockey knife for years with the patient horizontal, the movement of the hand needed to do it when the patient is upright is very different due to the curvature of the cornea. It can be quite hard to remove the epithelium without cutting into Bowman’s layer.

We eventually developed a process in which we tap the cornea with freshly prepared 40% ethanol on the tip of a cotton pad. After 70 seconds there is a beautiful erosion of 8 mm, with no harm to the cornea, and this can be done by any beginner in the field.

The way we avoid iontophoresis in epi-on was to go back to earlier studies¹⁶ on mixing penetration enhancers and modify the osmolarity of the solution to increase the paracellular transport within the epithelium, so the riboflavin can pass through the epithelial cell tight junctions directly into the stroma.

Dr Gerten: Regarding the therapy of keratoconic patients: when they walk into our office they do not say, “Doctor, please cross-link my cornea.” They say, “Doctor, how can I see better?”

You showed us these sub400 protocol patients and I fully agree that it makes a lot of sense to cross-link these keratoglobus or almost pellucid marginal degeneration cases you showed us. After that, these patients often have huge astigmatisms. Would it be possible, after cross-linking, to use intra-corneal rings and so combine cross-linking and other procedures?

Prof. Hafezi: You are absolutely right; cross-linking might make sense to stabilise the tissue but then we have to improve on vision. Especially in thin corneas, this is a question we get a lot: what is the use of cross-linking a cornea if the patient cannot see afterwards?

We are in the lucky situation in Switzerland to have three manufacturers of scleral lenses within 50 kilometres and I have seen cases that came from transplantation waiting lists with a cornea of 200 µm of residual stroma, huge epithelial compensation, K readings of 160 D. After cross-linking, they obtain a visual acuity of 0.7 with a scleral lens. So even extreme cases can now profit, and the best cornea is your own: the longer you can preserve your own cornea the better.

For the other aspects, I have been performing wavefront-guided PRKs on much less extreme cases for a number of years now, to regularise the corneas to improve corrected distance visual acuity, and epithelial mapping makes my Trans-PRK much more precise.

Dr Gerten: Damien, how do you treat thin corneas?

Prof. Gatinel: I think for optical restoration there is nothing better than a contact lens. So, when I have a patient with keratoconus, I fight on two fronts: the cessation of eye rubbing—I give them a patch at night, send them to the allergy specialist, ask them to make sure that they do not rub any more, etc.—and for visual rehabilitation, I am disappointed sometimes with surgical procedures. They may be satisfying for the surgeon, but the patient is eventually dissatisfied because he does not really see the improvement.

I would go for a hard contact lens. I do not really use cross-linking because, again, when patients stop rubbing, they do not progress.

That is the time to do PRK on those who have just a minor astigmatism to correct. As long as I am sure that they do not rub any more I am happy to do PRK or fit contact lenses. I am very conservative in terms of surgery with those keratoconus patients.

Prof. Ambrósio: I agree that eye rubbing is very bad for anyone. For rehabilitation of vision we have to go through the sequence of glasses and contact lenses. Imaging can help tremendously with the scleral lens, but some cases need surgery, which means rings, and patients are disappointed because they have very high expectations.

The first thing in treating keratoconus is patient education, which includes not rubbing the eye but goes very far beyond that. You have to tell them about the disease, what they may expect and the possible need for transplant. I agree that even with a very thin cornea, if you cross-link it a transplant has a higher chance to be more stable in the future.

This has not been proven by a scientific study, but we see a lot of post-op patients late after transplants even though they may have some residual ectasia in the periphery. I think cross-linking is a very good thing even for a pre-transplant case. A lot of people would debate this, because it may make the cornea harder, but I think ectasia should be treated as a specialty and we have to respect the disease and not to oversimplify.

Even though eye rubbing is bad for everybody I think many patients still progress; I cannot explain how but no explanation does not mean it does not happen. I have seen many patients progress who promise they do not sleep on the eye and do not rub the eye. These cases have to be prognostically detected early because they need cross-linking very early; as soon as possible in many cases.

The pathophysiology and pathomorphology of corneal ectasia: Part 1 | Part 2 | Part 3

Farhad Hafezi, MD, PhD, FARVO

E: farhad@hafezi.ch

Prof. Hafezi is a refractive and cataract surgeon based at the ELZA Institute in Zurich, Switzerland. He is the chief scientific officer of EMAGine AG, producer of C-eye, and holds a patent on a UV light source (PCT/CH 2012/000090).

Omid Kermani, MD

E: o.kermani@augenportal.de

Dr Kermani specialises in cataract and refractive surgery and has worked in private practice since 1993. He is based at the Augenklinik am Neumarkt in Cologne. He has no financial interest regarding the topic of the article.

References

1. Hammer A, Richoz O, Arba Mosquera S, et al. Corneal biomechanical properties at different corneal cross-linking (CXL) irradiances. Invest Ophthalmol Vis Sci. 2014;55:2881-2884.

2. Richoz O, Hammer A, Tabibian D, et al. The biomechanical effect of corneal collagen cross-linking (CXL) with riboflavin and UV-A is oxygen dependent. Transl Vis Sci Technol. 2013;2:6.

3. Abrishamchi R, Abdshahzadeh H, Hillen M, et al. High-fluence accelerated epithelium-off corneal cross-linking protocol provides dresden protocol-like corneal strengthening. Transl Vis Sci Technol. 2021;10:10.

4. Torres-Netto EA, Kling S, Hafezi N, et al. Oxygen diffusion may limit the biomechanical effectiveness of iontophoresis-assisted transepithelial corneal cross-linking. J Refract Surg. 2018;34:768-774.

5. Mazzotta C, Bagaglia SA, Sgheri A, et al. Iontophoresis corneal cross-linking with enhanced fluence and pulsed UV-A light: 3-year clinical results. J Refract Surg. 2020;36:286-292.

6. Hafezi F, Mrochen M, Iseli HP, Seiler T. Collagen cross-linking with ultraviolet-A and hypoosmolar riboflavin solution in thin corneas. J Cataract Refract Surg. 2009;35:621-624.

7. Jacob S, Kumar DA, Agarwal A, et al. Contact lens-assisted collagen cross-linking (CACXL): A new technique for cross-linking thin corneas. J Refract Surg. 2014;30:366-372.

8. Kling S, Hafezi F. An algorithm to predict the biomechanical stiffening effect in corneal cross-linking. J Refract Surg. 2017;33:128-136.

9. 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.

10. Hafezi F, Torres-Netto EA, Randleman JB, et al. Corneal cross-linking for keratoglobus using individualized fluence. J Refract Surg Case Rep. 2021;1:e10-e14.

11. Salmon B, Richoz O, Tabibian D, et al. CXL at the slit lamp: no clinically relevant changes in corneal riboflavin distribution during upright UV irradiation. J Refract Surg. 2017;33:281.

12. Tabandeh H, Boscia F, Sborgia A, et al. Endophthalmitis associated with intravitreal injections: office-based setting and operating room setting. Retina. 2014;34:18-23.

13. Ianchulev T, Litoff D, Ellinger D, et al. Office-based cataract surgery: population health outcomes study of more than 21 000 cases in the United States. Ophthalmology. 2016;123:723-728.

14. Hafezi F, Richoz O, Torres-Netto EA, et al. Corneal cross-linking at the slit lamp. J Refract Surg. 2021;37:78-82.

15. Salmon B, Richoz O, Tabibian D, et al. CXL at the slit lamp: no clinically relevant changes in corneal riboflavin distribution during upright UV irradiation. J Refract Surg. 2017;33:281.

16. Gatzioufas Z, Raiskup F, O’Brart D, et al. Transepithelial corneal cross-linking using an enhanced riboflavin solution. J Refract Surg. 2016;32:372-377.