Surgical pearls for CXL treatment of keratoconus, corneal ectasia

September 15, 2016

On the heels of the recent FDA approval of corneal collagen crosslinking, some surgical pearls are provided for the treatment of progressive keratoconus and corneal ectasia.

Take-home message: On the heels of the recent FDA approval of corneal collagen crosslinking, some surgical pearls are provided for the treatment of progressive keratoconus and corneal ectasia.

By Peter S. Hersh, MD, Special to Ophthalmology Times

Teaneck, NJ-Corneal collagen crosslinking (CXL) is a treatment designed to decrease the progression of keratoconus, a disease typified by distortion of the normal corneal optical dome secondary to biomechanical structural weakening. CXL aims to mitigate the progression of this distortion by strengthening the corneal stroma.

Related: Using a reverse side cut for better self-sealing efficacy

This year, the CXL procedure was approved by the FDA for the treatment of progressive keratoconus in patients > 14 years of age as well as for ectasia after refractive surgery. Approval was granted to Avedro, using its riboflavin products Photrexa Viscous and Photrexa. Photrexa comprises riboflavin 5’-phosphate 0.146% in solution containing 20 % dextran. It is used for riboflavin loading and during UV exposure. Photrexa is riboflavin 5’-phosphate 0.146% without dextran and is used for corneal swelling after the loading phase in corneas which have an intraoperative thickness <400 um. The riboflavin drops are used in conjunction with the Avedro KXL System, which operates at 365 nm UVA at a power of 3mW/cm2.

Recent: NASEM sets 2030 goals if a public health crisis is to be averted

The approved procedure is similar to the method used internationally for many years:

1.      The eye is prepped and draped in the usual fashion and a lid speculum is placed.

2.     A 9.0 mm removal of the central epithelium is performed per the surgeon’s preference. Usually, either a spatula or 20% alcohol is used to remove the epithelium.

3.     With the lid speculum removed, riboflavin with dextran (Photrexa Viscous) is applied topically every 2 minutes for 30 minutes. The patient is instructed to keep eyes closed between drops.

4.     After 30 minutes, confirm adequate riboflavin uptake by slit lamp examination. Uptake is confirmed by a homogeneous green fluorescence throughout the corneal stroma and by the presence of a yellow/green flare in the anterior chamber. If uptake is not adequate, additional riboflavin is administered until flare is observed. (Figure 1)

Procedure (cont.)


5.     UV administration should not be started in corneas thinner than 400 microns. If the corneal thickness is less than 400 microns, instill Photrexa drops (riboflavin without dextran) every 5 to 10 seconds until the corneal thickness increases to at least 400 microns. Photrexa is a relatively hypotonic solution which swells the corneal stroma, the goal of which is to protect the endothelium from damage by the riboflavin/UV interaction. Typically, we apply the drops for two minute intervals and then recheck pachymetry until this 400 micron threshold is met.

Related: Updated classification gives order to corneal dystrophy

6.     Place the patient under the KXL System. Place lid speculum and use the controller to center the treatment in the x,y, and z planes over the central cornea. Irradiate the eye for 30 minutes, assuring that the patient maintains gaze and the treatment is centered. During irradiation, continue topical instillation of Photrexa Viscous onto the eye every 2 minutes. (Figure 2)


Recent: How is making life easier for residents

FDA approvals of crosslinking for the two indications were each based on the results of 2 randomized, controlled clinical trials sponsored by Avedro. The primary efficacy criterion was a difference in the change in maximum keratometry on corneal topography analysis between treatment and sham control groups. Maximum keratometry was chosen as a quantitative descriptor of keratoconus severity. Change in maximum keratometry can be thought of as an indicator for progression or stability of keratoconus. For the multicenter studies of progressive keratoconus, average maximum keratometry improved by 1.6 diopters for patients enrolled with keratoconus, compared to average progression in control eyes. In corneal ectasia, average maximum keratometry improved by approximately 0.7 D.

Crosslinking outcomes


In our own patient cohort treated within the U.S. multicenter trials we looked at a number of other aspects of crosslinking outcomes:

Corneal topography

 In a 71 eye study, we noted an improvement in corneal topography indices using Scheimpflug imaging, including the keratoconus index (KI), index of surface variance (ISV) and index of vertical asymmetry (IVA). The improvements observed in ISV indicate a decrease of the curvature variation compared to the mean curvature of the cornea, and IVA, a measurement of the difference between the superior and inferior curvature of the cornea, is analogous to an improvement in the more commonly used inferior-superior steepness.

More cornea: Frozen, fresh corneal donor carriers yield similar clinical outcomes

Furthermore, improvement in KI may indicate that there is an improvement normalization of the keratoconic topographic appearance postoperatively. The overall improvements in corneal topography indices suggest, in general, that the cone is flattening and that the post-CXL cornea is becoming more optically regular and symmetric. (Figure 3)

Higher-order aberrations

Corroborating our findings from corneal topography, we analysed higher order aberrations in 96 eyes undergoing crosslinking and found significant improvements in ocular and anterior corneal higher order aberrations 1 year after CXL. We found that the average total anterior corneal HOAs, total coma, 3rd-order coma, and vertical coma all improved after crosslinking. Looking at total ocular HOAs, total coma, 3rd-order coma, trefoil, and spherical aberration all improved. (Figure 4)

Patient satisfaction


Patient satisfaction

In an effort to expand on the objective postoperative assessment of the crosslinking procedure and to further elucidate the expected clinical response, self-reported patients’ optical symptoms and visual function were analyzed in 107 cases. In our study, we found that patients generally noted subjective improvement in visual symptoms. Specifically, night driving, difficulty reading, diplopia, glare, halo, starbursts, and foreign body sensation were all improved one year after CXL. (Figure 5)

Related: Why single-instrument screening for keratoconus remains a dream

Postoperative timecourse and corneal haze/demarcation line

It is important for surgeons and patients alike to be aware of the timecourse of healing and vision recovery after crosslinking. In general, there is a worsening of vision and steepening of the keratoconic cone at 1 month postoperatively with improvement thereafter. (Figure 6)

Corneal haze, thickness


On clinical examination, corneal haze is generally noted after the crosslinking procedure. It appears initially as a dust-like change in the corneal stroma and evolves over time to a mid-stromal demarcation line, delineating the posterior extent of the crosslinking reaction. (Figure 7)

More: Epi-on CXL: Safe, effective option for treating thin corneas

We quantified the natural history of corneal haze over time after crosslinking using Sheimpflug densitometry. Similar to the timecourse of clinical outcomes after crosslinking, there appears to be an increase in haze, which peaks at one month, and plateaus between one and three months. Between three and six months the cornea begins to clear, and continues to return towards baseline at one year. (Figure 8)

It remains unclear whether this postoperative haze is an unwanted side effect or rather a desired wound healing affect demonstrating the efficacy of the crosslinking procedure.

Related: Performing femtosecond laser-assisted keratoplasty for keratoconus

We also looked at corneal thickness after CXL and found that mild corneal thinning is a general concomitant of the early CXL postoperative course. Similar to the timecourse of crosslinking associated corneal haze and crosslinking clinical outcomes, the cornea appears to thin at 1 and 3 months, and to re-thicken between 3 and 12 months. In our study, at one year, corneas remained slightly thinner than preoperative measurements.

Predictors of outcomes


Predictors of outcomes

The essential goal of collagen crosslinking is to stabilize the progression of the ectatic cornea. With regard to this disease stabilization, crosslinking certainly appears efficacious. In our study, 98.1% of eyes showed < 2D and 91.6% showed <1.0D of topographic progression over one year postoperatively. Although there was, on average, 1.7 D flattening of the cone, looking at individual eyes, the maximum K value decreased by 2.00 D or more in 31% patients, remained unchanged within 1.0 D in 65%, and increased by 2.00 D or more in 4%.

Recent: Top 10 eye-healthy cities in America

Thus, to determine preoperative patient characteristics that may predict topography and visual acuity outcomes of corneal collagen crosslinking, we performed multiple regression and odds ratio analysis on a cohort of 104 eyes treated in our practice. The essential finding was that eyes with more preoperative steepening tended to have a greater improvement in topography; specifically, those with maximum K >55 D were 5.4x more likely to have topographic flattening >2 D after CXL compared with eyes with flatter corneas.

Recent: Academy brings innovation theme to Chicago in October

However, with regard to eyes in which corneal topography continued to steepen, that is, those in which the crosslinking procedure failed to completely stabilize the disease, there were no independent predictors of continued disease progression even at the more refined >1 D level; all eyes were equivalently likely to be stabilized by the CXL procedure. Specifically, in patients with an initial maximum K >55 D, 40/44 (90%) eyes showed less than 1.0 D of progression, one year after CXL; similarly, in patients with initial maximum K <55.0 D, 55/60 (92%) eyes were stable. Finally, preoperative cone location may play an important role in the efficacy of the crosslinking procedure. There appears to be more topographic flattening in those eyes with centrally located cones. We found that maximum keratometry flattened by 2.6 D in eyes with centrally located cones, and by only 1.0 D and 0.05 D, in those eyes with paracentral and peripheral cones, respectively.

More: 10 least eye-healthy cities in the United States

With regard to corrected visual acuity, we found, on average, an approximately one line improvement in the cohort analyzed. When stratified by individual eyes, however, vision improved by 2 or more Snellen lines in 21% and remained stable within one line in 78%; 1 eye in our patient cohort lost 2 Snellen lines. In our analysis, the only independent predictor of a change in postoperative best corrected vision after CXL was preoperative best corrected visual acuity. Those eyes with worse preoperative best corrected visual acuity were more likely to experience an improvement of >2 Snellen lines. Specifically, eyes with a preoperative Snellen visual acuity of 20/40 or worse were 5.9 times more likely to improve by two lines or more; 43% of eyes with best corrected vision 20/40 or worse had on improvement of >2 lines compared with only 11% of eyes who were better than 20/40 preoperatively. With regard to eyes which lost vision from the procedure, the most salient indicator of an unwanted outcome, there was no independent preoperative indicator; however, it is very important to advise those patients with very good correctable vision before crosslinking of the changes in vision that may be experienced during the healing process.




1.     Hersh PS, Greenstein SA, Fry KL. Corneal Collagen Crosslinking for Keratoconus and Corneal Ectasia: One Year Results of a Randomized Prospective Study. J Cat Refract Surg. 2011:37:149-160.

2.     Greenstein SA, Fry KL, Bhatt J, Hersh PS. Natural History of Corneal Haze after Collagen Crosslinking for Keratoconus and Corneal Ectasia: A Scheimpflug and Biomicroscopic Analysis. J Cat Refract Surg. 2010:35:2105-2114.

3.     Greenstein SA, Fry KL, Hersh PS. Corneal topography indices after corneal collagen crosslinking for keratoconus and corneal ectasia: one year results. J Cat Refract Surg. 2011:37:1282-1290

4.     Greenstein SA, Fry KL, Hersh, MJ, Hersh, PS. Higher-order aberrations after corneal collagen crosslinking for keratoconus and corneal ectasia. J Cat Refract Surg. 2012;38:292-302.

5.     Greenstein SA, Shah VP, Fry KL, Hersh PS. Corneal thickness changes after corneal collagen crosslinking for keratoconus and corneal ectasia: one year results. J Cat Refract Surg 2011:37:691-700.

6.     Brooks NO, Greenstein SA, Fry KL, Hersh, PS. Patient subjective visual function after corneal collagen crosslinking for keratoconus and corneal ectasia. J Cat Refract Surg 2012;38:615-619.

7.     Greenstein SA, Hersh PS. Characteristics influencing outcomes of corneal collagen crosslinking for keratoconus and ectasia: Implications for patient selection. J Cat Refract Surg. 2013;39:1133-1140.

8.     Greenstein SA, Fry KL, Hersh, PS. Effect of Topographic Cone Location on Outcomes of Corneal Collagen Cross-linking for Keratoconus and Corneal Ectasia. J Cat Refract Surg 2012; 28: 397-405.


Peter S. Hersh, MD, is director, CLEI Center for Keratoconus–Hersh Vision Group, Teaneck, NJ. He did not indicate any financial interest in the subject matter.