December 05, 2019
4 min read

Evidence seems to favor epithelium-on for future of cross-linking

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

The debate over epithelium-on vs. epithelium-off collagen cross-linking for keratoconus and post-corneal refractive surgery ectasia is, in my opinion, similar to the much older LASIK/SMILE vs. PRK competition of the past three decades. Looking at global adoption rates, LASIK/SMILE is the clear winner over time vs. PRK in refractive corneal surgery. With equal efficacy and similar total complication rates, although different complications, patients and their surgeons have voted loud and clear that they prefer the reduced pain and more rapid visual recovery of LASIK/SMILE vs. PRK.

Richard L. Lindstrom
Richard L. Lindstrom

The challenge in CXL is to develop a system that delivers similar efficacy for epithelium-on vs. epithelium-off CXL. Once this is developed and achieves regulatory approval, nearly all patients and surgeons will adopt epithelium-on CXL. Epithelium-on CXL, besides being more surgeon- and patient-friendly, will also nearly eliminate the not uncommon side effects and complications associated with a large epithelial defect.

Over the past 15 years, I have gained experience with both epithelium-on and epithelium-off CXL in several clinical trials (Peschke, Topcon, Avedro, CXLO/CXLUSA) and in daily practice. I would like to share a few thoughts gained from this experience. As a disclosure, I have consulted widely in the field of CXL and own or have owned equity in CXLO, Avedro, Minnesota Eye Consultants and Unifeye Vision Partners.

First, my patients and I much prefer epithelium-on CXL. I nearly always do bilateral same-day treatments with epithelium-on CXL whereas I usually do one eye at a time, starting with the worst eye, when using epithelium-off CXL. While there is some pain and blurring of vision associated even with epithelium-on CXL, it is milder and short-lived, usually lasting only 24 to 48 hours at most vs. 5 days or more with epithelium-off CXL. I still use a bandage contact lens with epithelium-on CXL, prescribe a short course of NSAIDs and usually also prescribe a mild narcotic for the first 24 hours. Antibiotic coverage in the face of an intact epithelium is less necessary, and steroid treatment can be tapered earlier as there is less to no surface or stromal corneal haze. Most patients are functioning well enough with their old glasses to drive themselves back to the office the next day for their first postoperative visit, just like LASIK. Those patients who can only function with a contact lens in place can resume wear within days, rather than weeks. So, while epithelium-on CXL is not a “painless 5-minute lunchtime procedure” with immediate full visual recovery and no pain, it is much easier on the patient than epithelium-off CXL.


As an aside, for refractive CXL in the non-keratoconus patient to be successful, I believe it will have to be epithelium on.

For the surgeon, the procedure takes as long or longer for epithelium-on CXL as epithelium-off CXL. Surface preparation for riboflavin delivery can be more demanding than simple epithelial removal. In addition, another challenge in epithelium-on CXL is to ensure the stroma is well loaded with riboflavin before treatment with light delivery, and this can require longer riboflavin delivery time and a skilled observer using a slit lamp to evaluate adequate stromal riboflavin loading before light delivery.

Several techniques to improve riboflavin loading of the corneal stroma in the face of an intact epithelium have evolved, including chemical and mechanical treatments to disrupt the epithelial cell tight junctions, iontophoresis, the use of specially formulated higher concentration riboflavin drops and specialty riboflavin solution delivery devices. In addition, pulsed treatment with the light source and supplemental oxygen are logical and showing promise.

The data are still evolving, and different investigators are generating dissimilar outcomes with some far more enthusiastic than others. I find myself one of the believers that epithelium-on CXL with the proper riboflavin solution concentration and stromal loading methods, pulsed light delivery and oxygen supplementation will combine to give us an effective epithelium-on CXL treatment. My experience is that the outright failure rate with a well-designed well-performed epithelium-on CXL is no higher than that experienced with epithelium-off CXL. The amount of corneal flattening I have observed as measured by topography and/or refraction with epithelium-on CXL is somewhat less than what I achieve with epithelium-off CXL, but flattening is still present. Progression is halted and some flattening achieved, enough in my experience to allow epithelium-on protocols to pass regulatory approval barriers when compared with no treatment or even to be confirmed noninferior to current epithelium-off treatment methods. In regard to patient experience and morbidity, epithelium-on CXL will surely be proven not only noninferior but superior to epithelium-off CXL.

Time will tell if I am right, but I am optimistic we can achieve an FDA-approved epithelium-on CXL in the United States. I am also personally OK with the somewhat lower magnitude and variability of corneal flattening I see with epithelium-on vs. epithelium-off CXL. I think this will make the use of other refractive adjuncts such as Intacs (Addition Technology), conductive keratoplasty, PTK/PRK, ICLs (STAAR Surgical) and refractive cataract surgery easier to constructively combine with CXL. It may also allow a more predictable custom application of light as used in so-called “selective CXL.”


Finally, my experience and that of others suggest Snellen visual acuity and especially patient-reported post-treatment quality of vision are superior after epithelium-on CXL when compared with epithelium-off CXL. This may be related to the reduced surface and stromal haze seen with epithelium-on CXL. Once epithelium-on CXL achieves FDA approval in the U.S., I believe procedure volumes will increase as surgeons, referring doctors and especially patients become more comfortable with the patient journey required to undergo the CXL procedure. I anticipate patients will be followed more carefully and offered treatment earlier. This will be a big win for patients but also benefit treating and diagnosing eye care providers as well as the industry that supports us. Epithelium-on CXL will also greatly improve the likelihood that CXL can evolve into a safe, effective and predictable corneal refractive surgery. The future is bright for CXL and, in my opinion, will evolve toward epithelium-on CXL, treatment at an early stage of disease and potential applications to modify refractive error.

Disclosure: Lindstrom reports relevant financial disclosures with Avedro, Glaukos, CXLO/CXLUSA, Minnesota Eye Consultants and Unifeye Vision Partners.