September 2016




Corneal collagen crosslinking

Epi-off vs. epi-on debate continues

by Liz Hillman EyeWorld Staff Writer


The Pentacam image at the far left, 3 months after epi-off crosslinking, shows inferior steepening with K readings reaching 59.4 D at the center of the cone. The middle image taken 1.5 months prior to crosslinking shows K readings reaching 60.9 D at the center of the cone. This is evidence of cone flattening of 1.5 D at 3 months after treatment as shown in the right image. Source: Uri Elbaz, MD

Deciding when to deviate from standard crosslinking protocol, if at all, comes down to careful patient selection

To scrape or not to scrape? That is the question. Removal of the epithelium prior to soaking the cornea in a riboflavin solution and exposing it to ultraviolet light is the standard protocol to create new collagen bonds and stabilize ectatic corneas, such as in patients with progressive keratoconus. But some say the epithelium does not need to be removed to achieve this stabilizing effect with the benefit of reducing patient discomfort, procedure time, and risk of infection. Traditional epi-off crosslinking on human subjects—the Dresden protocol—was first described in the American Journal of Ophthalmology in 2003 by Gregor Wollensak, MD, Eberhard Spoerl, MD, and Theo Seiler, MD.1 This study described the technique of scraping off keratoconus patients’ epithelium, soaking their eyes in a riboflavin solution, and exposing their eyes to UVA light at 1-cm distance for 30 minutes. It stopped progression in all 22 patients. Since then, several different iterations of crosslinking protocols to treat progressive keratoconus and ectasia have emerged in the hope of improving safety and efficacy.2

Perhaps one of the most discussed alternative protocols to the traditional Dresden protocol is that of transepithelial (epi-on) crosslinking, which leaves the epithelium intact, but requires the riboflavin to sit (or load) on the eye longer. Recent studies that have compared epi-on vs. epi-off crosslinking have found that both protocols improve corrected and uncorrected distance visual acuity.3 Other research found that both protocols might cause hypoesthesia, but all patients recovered completely regardless of the protocol taken.4 The researchers found hypoesthesia was more pronounced in epi-off cases and recovery was shorter for epi-on cases though. Uri Elbaz, MD, Department of Ophthalmology, Rabin Medical Center, Tel Aviv University, Tel Aviv, Israel, cited other research that suggests “evidence for epi-on [crosslinking] efficacy is weak,” he said.5–7 “Obviously, there are numerous advantages for the epi-on procedure, including patient comfort, throughput, fast recovery with less work. Also, the epi-on procedure can spare general anesthesia and save OR time,” Dr. Elbaz said. “On the other hand, the epi-off procedure may result in persistent epithelial defect, haze, and corneal ulcer, although the risk for these complications is quite low.” Dr. Elbaz said he only performs epi-off procedures on his patients, calling it “more efficient in arresting the progression of keratoconus” with “minimal” risk.

Picking the right protocol for the patient

Enrique Graue-Hernandez, MD, Institute of Ophthalmology Conde de Valenciana, Mexico City, agreed that the data for epi-on crosslinking is not as strong as that available for epi-off, but “that doesn’t mean that epi-on doesn’t work,” he added.

“I think there are good patients for both treatments. There are certain conditions that may cause me to do epi-on or epi-off,” Dr. Graue-Hernandez said, noting that most of the time he performs epi-off procedures. Young patients who might not be as tolerant of pain and who have limited progression of the disease make better epi-on candidates, Dr. Graue-Hernandez said. Those who have had significant progression within 6 months to a year, however, fare better with epi-off treatment. “We know the biomechanical effect for epi-on treatment is reduced, but the clinical results can be the same in the long run,” he said. “Epi-off has a greater amount of biomechanical effect. We know the demarcation line is deeper, and therefore we think that the biomechanical strengthening of the cornea is greater.” Should a patient receive epi-on treatment and experience progression, Dr. Graue-Hernandez said epi-off can still be performed. “The evidence in favor of doing epi-off is strong and it’s not refutable. The evidence of epi-on is debatable, and there is significantly more risk of progression with epi-on, but that doesn’t mean it doesn’t work in good patients. That means you have to choose the right patient,” Dr. Graue-Hernandez said, explaining that epi-on patients should be informed of the risk of progression and be carefully monitored postop.

Considering the risk-benefit ratio

Epi-on might take less time and come with less pain, a quicker recovery, and less risk for infection, but continued progression might be looming. Epi-off, on the other hand, is a longer, more painful procedure with greater risk for infection and other side effects such as corneal haze, but with the stronger likelihood of stopping keratoconus progression. “Complications have been reported for both techniques,” said Sumit “Sam” Garg, MD, Gavin Herbert Eye Institute, University of California, Irvine, who performed epi-off crosslinking as part of an Avedro (Waltham, Massachusetts)- sponsored accelerated treatment protocol study. “It is important to follow these patients closely and to educate your patients about signs that may be harbingers to serious complications.” While research has shown mitomycin-C (MMC) can prevent corneal haze after photorefractive keratectomy (PRK),8 Dr. Graue-Hernandez doesn’t use MMC in crosslinking, calling the haze caused by this procedure different than that seen post-refractive surgery. Dr. Garg also does not use MMC. Standard postop care involves a course of antibiotics and topical steroids, although Dr. Graue-Hernandez said the steroid use might be longer for epi-off patients as there is more inflammation. Dr. Garg said he treats crosslinking patients as he would a PRK patient with topical steroid taper and aggressive preservative-free artificial tears.

Improving epi-on and enhancing epi-off

One way to possibly advance performance of the epi-on technique would be to improve riboflavin penetration into the corneal stroma. Several companies have developed riboflavin solutions that include “enhancers” for this purpose. “There are several reported additives to increase the permeability of riboflavin through the epithelium into the stroma: BAK, tetracaine, surfactants, tris hydroxymethyl aminomethane, and/or sodium EDTA,” Dr. Garg said. Dr. Elbaz said using a local anesthetic to disrupt epithelial integrity can also help improve penetrance of Ricolin (Carleton Optical, Buckinghamshire, U.K.), a 1% riboflavin solution with enhancers. Even still, Dr. Elbaz said better solutions for the efficacy of epi-on need to emerge before many ophthalmologists choose to convert from epi-off. Epi-off crosslinking continues to evolve outside of the Dresden protocol as well. Dr. Graue-Hernandez said his team has been using an accelerated crosslinking protocol for the last 4 years. This protocol increases UVA energy delivered to the eye but reduces the exposure time significantly. He said a “pulse” technique could be beneficial as well. “Theoretically, oxygen is needed for the reaction to occur. Whenever you use continuous illumination, oxygen is deprived and reaction is less,” Dr. Graue-Hernandez said. “There have been a few trials and a few in vivo experiments with pulse therapy. We don’t have it but you can buy a system that increases the oxygen concentration on the eye. There is a little patch that is plugged into an oxygen device that pumps 90% oxygen into an environment right in front of the eye. That’s used in combination with a pulse therapy to increase the reaction.” Overall, Dr. Garg said the current atmosphere of crosslinking is “very exciting.” “I do think that with time our treatments will evolve and perhaps we will have an epi-on treatment that will provide efficacy on par with epi-off treatments with improved safety and patient experience,” Dr. Garg said. “My take-home message is that we need to continue improving the epi-on technique so the efficacy will be at least equivalent to the epi-off technique. Until then, the epi-on technique should be reserved only for rare cases where, for example, a general anesthesia procedure is contraindicated in a child with progressive keratoconus, knowing that a retreatment may be needed in the future,” Dr. Elbaz said. “Otherwise, the vast majority of patient should receive the epi-off crosslinking.”


1. Wollensak G, et al. Riboflavin/ultraviolet -a-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135:620–7.

2. Cummings AB, et al. Newer protocols and future in collagen cross-linking. Indian J Ophthalmol. 2013;61:425–7.

3. Rossi S, et al. Standard versus trans-epithelial collagen cross-linking in keratoconus patients suitable for standard collagen cross-linking. Clin Ophthalmol. 2015;9:503–9.

4. Spadea L. Recovery of corneal sensitivity after collagen crosslinking with and without epithelial debridement in eyes with keratoconus. J Cataract Refract Surg. 2015;41:527–32.

5. Caporossi A, et al. Transepithelial corneal collagen crosslinking for progressive keratoconus: 24-month clinical results. J Cataract Refract Surg. 2013;39:1157–63.

6. Soeters N, et al. Transepithelial versus epithelium-off corneal cross-linking for the treatment of progressive keratoconus: a randomized controlled trial. Am J Ophthalmol. 2015;159:821–8.

7. Eraslan M, et al. Efficacy of epithelium -off and epithelium-on corneal collagen cross-linking in pediatric keratoconus. Eye Contact Lens. 2016 Feb 29. [Epub ahead of print].

8. Hashemi H, et al. Evaluation of the prophylactic use of mitomycin-C to inhibit haze formation after photorefractive keratectomy in high myopia: a prospective clinical study. BMC Ophthalmol. 2004;4:12.

Editors’ note: The physicians have no financial interests related to their comments.

Contact information


Epi-off vs. epi-on debate continues Epi-off vs. epi-on debate continues
Ophthalmology News - EyeWorld Magazine
283 110
220 144
True, 9