Secondary Logo

Journal Logo


Transepithelial photorefractive intrastromal corneal crosslinking versus photorefractive keratectomy in low myopia

El Hout, Safa MD1,*; Cassagne, Myriam MD, MSc1,2; de Gauzy, Thomas Sales FRCOphth1; Galiacy, Stéphane PhD1,2; Malecaze, François MD, PhD1,2; Fournié, Pierre MD, PhD1,2

Author Information
Journal of Cataract & Refractive Surgery: April 2019 - Volume 45 - Issue 4 - p 427-436
doi: 10.1016/j.jcrs.2018.11.008
  • Free


Corneal crosslinking (CXL), initially introduced by Wollensak et al. in 2003,1 has emerged as an efficient method for the treatment of keratoconus, a progressive disorder characterized by thinning and steepening of the cornea that leads to visual impairment. The aim of CXL is to create covalent bonds between amino groups within the collagen molecules or between proteoglycan core proteins and collagen to make the anterior corneal stroma more rigid.2 In the conventional CXL technique described in the Dresden protocol,3 the corneal stroma is soaked with a riboflavin 0.1% solution after epithelial removal; then the cornea is exposed to a uniform beam of 365 nm ultraviolet-A (UVA) radiation at 3 mW/cm2 for 30 minutes (5.4 J/cm2 dose). Studies of the efficiency of conventional CXL1,3–5 found that the procedure halted the progression of keratoconus in 90% of cases. They also reported mean flattening of the steepest corneal curvature of up to 2.0 diopters (D) and a significant improvement of visual acuity (>1 line).4,5 Over the past few years, there have been attempts to improve the conventional CXL technique. One new approach is transepithelial or epithelium-on (epi-on) CXL, which has gained attention as a potential means of improving patient comfort and the safety profile by leaving the barrier function of the epithelium intact, thus avoiding complications such as infectious keratitis or corneal healing disorders. Recent studies6,7 have shown that topography-guided CXL procedures, which are specifically aimed at stiffening the cone area, significantly decrease maximum keratometry (K) values and improve the corrected distance visual acuity (CDVA) compared with conventional CXL.

Today, the refractive surgery techniques used to correct low myopia are photorefractive keratectomy (PRK), laser in situ keratomileusis, and small-incision lenticule extraction. However, PRK requires corneal deepithelialization before excimer laser application. Similar to conventional CXL, PRK has a higher risk for complications as a result of epithelial debridement. In addition, the therapeutic effects of excimer laser refractive surgeries (PRK, laser in situ keratomileusis) and femtosecond laser refractive surgeries (small-incision lenticule extraction) are achieved through tissue ablation, resulting in weakening of the intrastromal corneal matrix.8 Thus, a procedure that preserves both the corneal epithelium and stromal tissue could improve the treatment of low myopia.

The principle of localized corneal flattening through custom treatment with UVA irradiation gave rise to a new refractive application of CXL called transepithelial photorefractive intrastromal CXL. It consists of planned intrastromal corneal remodeling by CXL to reduce the central mean K value and induce myopic correction. Hence, we performed a comparative study of patients who had transepithelial photorefractive intrastromal CXL in one eye and PRK in the other eye to assess the efficacy and safety of the new CXL technique in the refractive correction of low myopia in healthy nonectatic eyes.

Patients and methods

This prospective comparative nonrandomized single-center pilot cohort study with paired-eye control was performed at the Department of Ophthalmology, Purpan Hospital, Toulouse, France, from September 2016 through March 2017. Approval was obtained from the ethical committee (ID-RCB: 2015-A01526-43), and the study was performed in accordance with the tenets of the Declaration of Helsinki. All participants signed an informed consent form before enrolling in the study.

Patient Population

All patients presented for myopic correction and were eligible for corneal refractive surgery. Inclusion criteria were 18 years or older, stable myopia with a manifest refraction spherical equivalent (MRSE) of −1.00 to −2.50 D and cylinder of plano to 0.75 D, and minimum pachymetry greater than 480 μm. Patients with ocular disease, previous ocular surgery, trauma, or a corneal condition were excluded. Eyes with forme fruste keratoconus or keratoconus-suspect eyes were also excluded.

The nondominant eye of each patient had transepithelial photorefractive intrastromal CXL (CXL eye). The contralateral dominant eye had PRK during the same session in the same operating room (PRK eye). Patients were informed that after transepithelial photorefractive intrastromal CXL, there was a possibility of undercorrection that would result in anisometropia. They were told although most patients could tolerate the anisometropia, additional PRK surgery might be required.

Surgical Technique

Transepithelial Photorefractive Intrastromal Corneal Crosslinking

After topical anesthesia of tetracaine and oxybuprocaine was administered, a custom oxygen delivery mask (Avedro, Inc.) was placed above the eye to be treated (Figure 1). The mask was connected to an oxygen wall outlet with a flow meter and bubble humidifier. A riboflavin 0.25% solution with benzalkonium chloride in hydroxypropyl methylcellulose (ParaCel Part 1, Avedro, Inc.) was applied at an interval of 1 drop every 90 seconds for 3 minutes. From the fourth minute on, another solution of riboflavin 0.22% (ParaCel Part 2, Avedro Inc.) was applied every 90 seconds for 6 minutes. The supply of oxygen was opened at a flow rate of 1.5 L/min 2 minutes before the end of riboflavin soaking. Excess riboflavin was flushed from the eye with a balanced salt solution. The concentration of oxygen in the mask was measured with a sensor (Fibox 4 Oxygen Meter, PreSens Precision Sensing GmbH). An oxygen concentration of at least 95% was required to initiate UVA irradiation.

Figure 1
Figure 1:
Oxygen delivery mask placed on the treated eye during the photorefractive intrastromal corneal crosslinking procedure.

A Conformité Européenne–marked (EU1504407) UVA delivery device with integrated active pupil-tracking technology (Mosaic, Avedro, Inc.) was used (Figure 2). Pupil centration was chosen because the angle κ is small in low myopia with low astigmatism. The stroma was irradiated with a 6.0 mm diameter spot of 365 nm UVA at 30 mW/cm2 for 16 minutes and 40 seconds, which was pulsed at 1-second intervals. The total dose was 15 J/cm2. The UVA beam was centered on the pupil via the pupil-tracking system. At the end of the procedure, the cornea was rinsed with a balanced salt solution.

Figure 2
Figure 2:
Ultraviolet-A delivery device with pupil-tracking system.

All eyes had the same protocol independent of their preoperative myopic error. This protocol was based on theoretic modeling studies by the device manufacturer.

Photorefractive Keratectomy

After topical anesthesia comprising tetracaine and oxybuprocaine was administered, the central 8.0 mm corneal epithelium was removed using a soft rotating surgical brush. Centration was confirmed, and excimer laser photoablation (Wavelight EX-500, Alcon Surgical, Inc.) was performed. The optical zone was 6.5 mm.

Postoperative Care

At the end of the procedures, topical ofloxacin (Quinofree) was administered. A bandage contact lens was applied to PRK eyes only. The postoperative treatment for both eyes included preservative-free ofloxacin 3 times daily for 7 days, preservative-free flurbiprofen (Ocufen) 3 times daily for 2 days, and single-dose lubricant eyedrops (Vismed, Horus Pharma) as needed.

Safety and Efficacy Outcomes

The main efficacy outcome measure was the mean change in the MRSE from baseline. The secondary efficacy outcomes measures were as follows:

  • The mean change in uncorrected distance visual acuity (UDVA), which was assessed using a standardized scotopic decimal projection chart at 5 m. The values were converted to logarithm of the minimum angle of resolution (logMAR) notation.
  • The mean change in the mean K value based on the axial front curvature map of a Pentacam HR Scheimpflug tomographer (Oculus Optikgeräte GmbH).
  • The demarcation line depth in the CXL eye, which was observed on anterior segment optic coherence tomography (AS-OCT) (Spectralis, Heidelberg Engineering GmbH) and measured centrally and at 4 peripheral points (3.0 mm temporally, nasally, in the upper limit, and in the lower limit).
  • The mean change in central epithelial thickness measurement on AS-OCT in both eyes.

The safety outcome measures included the CDVA in logMAR notation, the endothelial cell count (ECC) measured by specular microscopy (SP 2000P, Topcon Corp.), and the incidence of adverse events identified by ocular examination on slitlamp biomicroscopy. Subjective patient evaluation (eg, pain, dryness, itching, tingling, foreign-body sensation, watering, photophobia, blurred vision) was also performed using a visual analog scale. The score for each subjective symptom ranged from 0 to 10. Serious side effects were defined as infectious keratitis or a loss of CDVA of more than 2 lines.

Patient Examinations

Patients were examined preoperatively, immediately after the surgical procedures, 2 days 2 postoperatively for bandage contact lens removal in the PRK eye, and at 7 days to check for early complications. Patients were also examined 1 month, 3 months, and 6 months postoperatively.

Statistical Analysis

Comparisons of parameters with the baseline values were performed using the Student t test. Data, expressed as the mean ± SD, were considered statistically significant when the P value was less than 0.05. The correlation between the mean K value and the demarcation line depth at 1 month was calculated using the Pearson correlation coefficient.


The mean age of the 19 patients in the cohort was 28.2 ± 4.8 years. There were no statistically significant differences in the baseline clinical characteristics between CXL eyes and PRK eyes (Table 1).

Table 1
Table 1:
Baseline clinical characteristics of CXL eyes and PRK eyes.


Table 2 and Figure 3 show the changes in the MRSE through 6 months postoperatively. There was a statistically significant reduction in the MRSE in CXL eyes (P < .001). The reduction in PRK eyes was less but still statistically significant (P < .001). Three CXL eyes had only a slight change in the MRSE (<0.5 D).

Figure 3
Figure 3:
Mean change in manifest refraction spherical equivalent over 6 postoperative months (CXL = transepithelial photorefractive intrastromal corneal crosslinking; PRK = photorefractive keratectomy).
Table 2
Table 2:
Clinical characteristics in both groups after 1, 3, and 6 months compared with baseline measurements (mean ± SD).

Visual Acuity

The improvement in UDVA from baseline was statistically significant 1 month, 3 months, and 6 months postoperatively in both groups (P < .001); however, there was less improvement in the PRK eyes than in CXL eyes (mean 0.75 ± 0.09 decimal versus 0.32 ± 0.28 decimal) (Table 2 and Figure 4). Thirteen patients (68.4%) were more satisfied with outcome in the PRK-treated eye. There was no regression in UDVA throughout the follow-up in either treatment group. Eight patients required retreatment in the CXL eye after the 6-month visit.

Figure 4
Figure 4:
A: Mean change in UDVA over 6 postoperative months. B: Change in lines of UDVA at 6 months. C: Comparison between postoperative UDVA and preoperative UDVA (CXL = transepithelial photorefractive intrastromal corneal crosslinking; LogMar = logarithm of the minimum angle of resolution; PRK = photorefractive keratectomy; UDVA = uncorrected distance visual acuity).

The CDVA remained stable through 6 months postoperatively in all eyes (mean logMAR, 0.0; mean decimal, 1.0).


Compared with baseline, there was a statistically significant reduction in the mean K value from the first month in CXL eyes and PRK eyes (P < .001 and P < .01, respectively) (Table 2 and Figure 5). The decrease was at least 1.0 D in 3 CXL eyes (16%) and 14 PRK eyes (73.6%) (Figure 6).

Figure 5
Figure 5:
Mean change in mean keratometry over 6 postoperative months (CXL = transepithelial photorefractive intrastromal corneal crosslinking; PRK = photorefractive keratectomy).
Figure 6
Figure 6:
Percentage of eyes with corneal flattening (CXL = transepithelial photorefractive intrastromal corneal crosslinking; Km = mean keratometry; PRK = photorefractive keratectomy).

Qualitative analysis of the differential axial front curvature maps between 6 months and baseline showed corneal central flattening of at least 1.0 D in 12 CXL eyes (63.2%) and in all PRK eyes (Figure 7).

Figure 7
Figure 7:
Difference axial front curvature maps: 6 months versus baseline.

Optical Coherence Tomography Demarcation Line

Table 3 shows the mean central and peripheral demarcation line depth at each visit. A stromal demarcation line was observed in all CXL eyes and was the most noticeable at 1 month (Figure 8). The depth decreased gradually starting from the peripheral stroma but remained visible at 6 months in many cases. There was a statistically significant correlation between the depth of demarcation line and the reduction in the mean K value at 1 month (r = .61, P < .01). There was also a correlation between the depth of demarcation line and the improvement in the MRSE (r = 0.51, P < .05) (Figure 9).

Figure 8
Figure 8:
Anterior segment optic coherence topography of a CXL eye at 1 month. Arrows show a demarcation line. A: Six millimeter whole-thickness corneal hyperreflectivity. B: Well-delimited anterior stromal hyperreflectivity.
Figure 9
Figure 9:
Correlation of demarcation line depth at 1 month. A: Mean change in mean keratometry (meanK). B: Mean change in manifest refraction spherical equivalent (MRSE).
Table 3
Table 3:
Depth of demarcation lines measures in CXL eyes over time.

Epithelial Thickness

There was no significant change from baseline in the mean central epithelial thickness at any follow-up visit in both treatment groups.

Endothelial Cell Count

There were no statistically significant differences between the baseline ECC and the postoperative ECC (Table 2). The images of the endothelial cells at 1 month and 3 months were blurry in CXL eyes. At 6 months, the image quality was improved (Figure 10).

Figure 10
Figure 10:
Specular microscopy photograph at 1 month showing endothelial cell count of 2652 cells/mm2 in a CXL eye. Note the blurry image.

Adverse Events

More patients reported no pain postoperatively in the CXL eye compared with the PRK eye; all patients reported pain in the PRK eye at 2 days. Patients reported a foreign-body sensation in 17 CXL eyes (89.5%) within the 2 first days postoperatively and moderate photophobia in both eyes at 2 days (19 patients [100%]) and during the first month (17 patients [89.5%]; both phenomena tended to decrease over time. There were no serious side effects. Punctate epithelial staining was noticed immediately after CXL in all 19 eyes; the epithelium was fully intact by 2 days. Ulcers were still present 7 PRK eyes (36.8%) at 2 days.

Transient delineated corneal haze was seen in 16 CXL eyes (84.2%) and 1 PRK eye (5.3%) (Figure 11). The haze was most pronounced at 1 month, after which it gradually improved. It remained slightly visible in some eyes at 6 months.

Figure 11
Figure 11:
Slitlamp photograph of a CXL eye shows circular stromal haze at 1 month.


Today, corneal refractive surgery techniques are often used to correct low myopia because they are relatively easy to perform and efficient and they yield predictable outcomes. These techniques use custom stromal photoablation performed using an excimer laser. However, tissue removal affects the biomechanical properties of the cornea, resulting in weakening of and a reduction in stromal rigidity.8 Thus, these procedures are not used in eyes with subclinical keratoconus or thin corneas because of the risk for corneal ectasia.

Studies9,10 suggest that screening tests for refractive surgery should be thorough and should include corneal topography to exclude eyes that are suspect for keratoconus. However, others report refractive surgery in eyes with keratoconus or keratoconus-suspect eyes gives acceptable results. For example, a study11 found that PRK was safe and effective for keratoconus-suspect corneas over a 5-year follow-up. Another study12 found that topography-guided laser ablation treatment improved the UDVA and CDVA in patients with mild to moderate degrees of keratoconus. However, in clinical practice, excimer laser keratectomy is still considered to be contraindicated in eyes with an unstable thin cornea with keratectasia. Therefore, to avoid further weakening and ectasia of the keratoconic cornea after excimer laser corneal ablation, many refractive surgeons perform simultaneous topography-guided PRK and riboflavin–UVA CXL.13

Unlike corneal refractive surgery techniques, there is no stromal ablation in CXL. Standard CXL induces corneal flattening of approximately 2.0 D in patients with keratoconus.3–5 This flattening is thought to be the result of stromal stiffening induced by increased formation of covalent bonds within or between collagen molecules and proteoglycan core proteins.2

Many variations on the conventional CXL Dresden protocol have been developed. They not only modify the protocol of riboflavin soaking through the epithelium (transepithelial CXL14 or iontophoresis15) but also the UVA irradiation protocol, reducing the time of the procedure (accelerated CXL16) or customizing the irradiation pattern (topography-guided CXL). Greater corneal flattening can be achieved with topography-guided CXL than with conventional CXL because the former applies cone-centered patterns of irradiation delivered by devices that allow custom treatment plans (Mosaic KXL II device, Avedro, Inc.).6,7

Given the multitude of CXL protocols, finite element modeling of the cornea and the kinetics of the CXL have been developed.17 These numeric models are used to simulate the topographic changes after CXL, predict the theoretic maximum depth of the treatment, and optimize several parameters (eg, UVA dose, irradiance, pulse interval, riboflavin concentration, addition of supplemental oxygen).18 Seven et al.19 developed and applied a biomechanical model to study the potential refractive applications of CXL, notably the correction of astigmatism using custom irradiation patterns guided by topography.

Kanellopoulos20 published the first reported case of photorefractive CXL using a high dose of UVA (14 J/cm2) for astigmatism correction, resulting in a toric reduction of 0.8 D. They then used a peripheral irradiation profile for hyperopia correction with encouraging results. Finally, they showed a statistically significant reduction in the MRSE in a feasibility study that evaluated photorefractive intrastromal CXL in the treatment of low myopia.

Elling et al.21 recently published a case series evaluating epithelium-off (epi-off) photorefractive intrastromal CXL for the treatment of low myopic refractive error in 24 eyes of 14 patients. Although similar irradiation parameters were applied in our study (30 mW/cm2 pulsed UVA; dose 15 J/cm2), Elling et al. used a smaller UVA beam diameter (4.0 mm) than we did (6.0 mm). They reported a mean MRSE reduction of 0.90 ± 0.40 D and a mean UDVA improvement of 0.34 ± 0.20 logMAR; the reduction in the MRSE is 0.2 D greater than in our study.

Our prospective pilot study to evaluate high-fluence pupil-centered transepithelial photorefractive intrastromal CXL for the treatment of low myopia in healthy eyes used protocols developed from the modeling framework and from early clinical experience. The purpose was to ascertain whether photorefractive intrastromal CXL could safely flatten the cornea and reduce myopia. The benefit of this CXL technique over the epi-off photorefractive intrastromal CXL technique is the epithelium is not removed, avoiding the risk for complications such as infectious keratitis,22 sterile infiltrates, corneal opacities, herpetic reactivation, and postoperative discomfort resulting from epithelial debridement.23

For ethical reasons, we performed transepithelial photorefractive intrastromal CXL only in the nondominant eye of each patient; thus, the CXL sample was limited to 19 eyes. The dominant eye was treated using PRK. This allowed us to compare transepithelial photorefractive intrastromal CXL outcomes with those of a standard corneal refractive surgery technique. In cases of undercorrection, transepithelial photorefractive intrastromal CXL triggers slight monovision that is well tolerated by most patients with myopia. The CXL treatment protocol in our study decreased the MRSE and improved the UDVA, more in the CXL eyes than in PRK eyes. The standard deviation of the refractive outcome suggests that transepithelial photorefractive intrastromal CXL has lower predictability than conventional refractive procedures. Thus, further customization of the procedure based on individual patient factors (eg, topography) might be needed.

The improvements in the MRSE and UDVA occurred during by the first postoperative month. In addition, there were no significant differences in the central epithelial thickness between preoperatively and 1 month, 3 months, and 6 months postoperatively, indicating that the outcomes were likely the result of stromal remodeling and not epithelial variations. Furthermore, there was no significant change from the first through the sixth month, confirming that there was no early refractive regression. Although some refractive techniques used in the past, such as collagen shrinking, led to early regression, the findings in our study suggest there is a potential for refractive stability after transepithelial photorefractive intrastromal CXL. Longer term assessment is needed in this area.

To our knowledge, ours is the first study to report the results of transepithelial photorefractive intrastromal CXL with pulsed UV light and supplemental oxygen delivery. When the epithelial barrier is left intact, the epithelium slows the diffusion of riboflavin and ambient oxygen into the cornea; this might reduce the efficacy of the procedure compared with that of the standard epi-off technique.24

With the aim of maximizing the efficacy of transepithelial photorefractive intrastromal CXL, we applied a 2-part riboflavin formulation specifically marketed for use in transepithelial CXL procedures (ParaCel Part 1 and 2). This formulation contains a higher concentration of riboflavin than that used in conventional epi-off CXL and contains permeability-enhancing agents (benzalkonium chloride 0.02% in methylcellulose) intended to loosen the epithelial junctions and aid in the delivery of riboflavin to the stroma. The superficial epithelial layer might be damaged, which would enhance riboflavin penetration without the treatment being completely abrasive.14 Several recent publications reported the efficacy of this formulation in combination with accelerated CXL techniques to reduce maximum K values in patients with progressive keratoconus.14,25,26 Because the amount of flattening in these studies is lower than reported in some clinical studies of epi-off CXL, in our study CXL was performed in a humidified high-oxygen (>95%) environment using an oxygen delivery mask placed on the eye to improve the efficiency of the epi-on procedure. The rationale for the addition of pulsed UV and supplemental oxygen is based on the understanding of CXL photochemical reactions, which lead to the formation of crosslink covalent bonds within the collagen and proteoglycans molecules.27

Two types of photochemical reactions have been described. Type 1 reactions predominate under anaerobic conditions; excited riboflavin is used as substrate and generates oxygen-free radicals by riboflavin photolysis. Type 2 reactions occur under aerobic conditions at the beginning of the UVA irradiation and result in the formation of a singlet of oxygen and subsequent photooxidation of stromal proteins. Rapid oxygen depletion occurs within 10 to 15 seconds of the initiation of UV in CXL reactions, while turning the UV light off leads to rapid replenishment of oxygen.28 Hence, pulsing the UV light during CXL treatment theoretically increases the oxygen concentration by slowing the rate of oxygen consumption, while the addition of supplemental oxygen at the surface increases the rate of oxygen diffusion. In silico photochemical kinetic modeling and ex vivo biomechanical analyses have shown that the combination of pulsed irradiation of UV and an oxygen-enriched environment increases the stromal oxygen concentration during CXL and increases corneal stiffness.29 However, the study did not include a control group of eyes treated without oxygen; therefore, we cannot make a conclusion about the effect of the oxygen.

A stromal demarcation line was detected on AS-OCT at 1 month in all eyes that had transepithelial photorefractive intrastromal CXL. The mean depth of the line was greater (366 ± 104 μm) than that described in the literature on transepithelial CXL (105 ± 15 μm).30 This might have been the result of the increase in ambient oxygen concentration and the higher total UVA dose. There was a statistically significant correlation between the demarcation line at 1 month and the mean MRSE (P < .05) and corneal flattening (P < .01). The deeper the line, the more the cornea was flattened and the greater the reduction in the myopic error. Thus, the line, which was clearly visible at 1 month, seems to be an indirect indicator of CXL efficacy, as reported in a previous study.31 This correlation is supported by the very good refractive result obtained in 6 CXL eyes that had hyperreflectivity of the entire thickness of the central cornea on AS-OCT. However, one might speculate that an even higher UVA dose could achieve a greater refractive effect but could have a deleterious effect on endothelium. Perhaps the demarcation line depth is related to the high oxygen concentration. If so, this concentration with the same UVA total dose could affect the depth of the demarcation line.

The ECC was stable over 6 months in all eyes. However, the image of endothelial cells in CXL eyes obtained by specular microscopy was blurry until the third month. This might be related to the presence of stromal haze in these eyes. Elling et al.21 reported a decrease in cells hexagonality with an energy of 15 J/cm2; they attributed this to poor image quality. The maximum corneal and lenticular UV radiant exposure is established in International Organisation for Standardization 15004-2 (2007),32 which defines light hazard protections for ophthalmic instruments. For exposure times less than 1000 seconds (16.7 minutes), the maximum corneal and lenticular dose is 1 J/cm2 for a Group 2 instrument (ie, ophthalmic instrument for which a potential light hazard exists). The presence of riboflavin in the cornea results in transmission of only 5% of UV light to the endothelium in an epi-off procedure. The transmission is reduced to only 3.5% to 4.0% in an epi-on procedure as a result of additional absorption in the epithelial cells. Therefore, the estimated UV dose reaching the endothelial cells during a 15 J/cm2 CXL treatment is 0.53 to 0.75 J/cm2, depending on the presence or absence of the epithelium; this dose is less than the limit established in the international standard. The standard takes into account the potential for repeated or chronic exposure to the light hazard; therefore, the limits contain a significant additional safety buffer that is not necessarily applicable to a 1-time treatment such as CXL. Comparable studies of rabbit eyes33 found acute damage thresholds of up to 70 J/cm2 for the lens and 42 J/cm2 for the cornea. Longer term monitoring of these patients and of larger samples is needed to establish the safety of this procedure.

No serious side effects occurred in our study. The CDVA remained stable through 6 months postoperatively. The epi-on approach, which limits epithelial damage to superficial punctate staining immediately after the procedure, reduces the risk for infection and significantly increases the patient’s comfort. Although the procedure was relatively easy for the surgeon to perform and comfortable for the patients, it lasts almost 30 minutes while other corneal refractive procedures last just a few minutes. Hence, ways to reduce the time of the procedure should be explored.

We have continued to monitor the patients in this study. Based on our observations, we believe that the residual refractive error after transepithelial photorefractive intrastromal CXL be corrected using PRK retreatment for patients who wish to achieve emmetropia. Eight patients were treated with PRK after the sixth month; the last refraction, measured 6 weeks after PRK, was a lower than before retreatment (plano refraction; mean UDVA 0.0 logMAR). Postoperative biomicroscopy did not show subepithelial haze or other adverse effects after the retreatment.

In conclusion, high-dose pupil-centered transepithelial photorefractive intrastromal CXL reduced low myopia. Or results indicate that this treatment is minimally invasive in the treatment of mild refractive error. It seems to be a good option for patients with thin corneas (virgin or after refractive surgery) or suspicious corneal topography, which are contraindications to conventional refractive surgery. However, larger, more diverse cohorts with long-term data are needed to confirm the safety of the procedure and the stability of the results.

The efficiency of transepithelial photorefractive intrastromal CXL is still limited and inferior to that of epi-off CXL and PRK. It could be improved by modulating the various parameters, which is now possible because of the development of numeric modeling of the cornea and CXL. Alternate beam patterns or integration of corneal topography data might improve treatment accuracy, in particular in the correction of astigmatism.

What was known

  • Photorefractive intrastromal corneal crosslinking (CXL) is a new approach for the correction of low myopic errors based on flattening of the central cornea after CXL.
  • Epithelium-off photorefractive intrastromal CXL has been studied in small cohorts and has been reported to reduce myopia and improve uncorrected distance visual acuity.
  • Transepithelial photorefractive intrastromal CXL, although not widely studied, is thought to reduce the risk for complications.

What this paper adds

  • Transepithelial photorefractive intrastromal CXL was effective in reducing the low myopic error in healthy eyes, and no complications occurred during the 6-month follow-up.
  • In the future this noninvasive refractive procedure, with some improvements, can be used to safely correct low myopia, especially in eyes with a thin cornea.


1. Wollensak G, Spoerl E, Seiler T. Riboflavin/ultraviolet-A–induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol. 2003;135:620-627.
2. Spoerl E, Huhle M, Seiler T. Induction of cross-links in corneal tissue. Exp Eye Res. 1998;66:97-103.
3. Wollensak G. Crosslinking treatment of progressive keratoconus: new hope. Curr Opin Ophthalmol. 2006;17:356-360.
4. Raiskup-Wolf F, Hoyer A, Spoerl E, Pillunat LE. Collagen crosslinking with riboflavin and ultraviolet-A light in keratoconus: long-term results. J Cataract Refract Surg. 2008;34:796-801.
5. Caporossi A, Mazzotta C, Baiocchi S, Caporossi T. Long-term results of riboflavin ultraviolet A corneal collagen cross-linking for keratoconus in Italy: the Siena Eye Cross Study. Am J Ophthalmol. 2010;149:585-593.
6. Cassagne M, Pierné K, Galiacy SD, Asfaux-Marfaing M-P, Fournié P, Malecaze F. Customized topography-guided corneal collagen cross-linking for keratoconus. J Refract Surg. 2017;33:290-297.
7. Seiler TG, Fischinger I, Koller T, Zapp D, Frueh BE, Seiler T. Customized corneal cross-linking: one-year results. Am J Ophthalmol. 2016;166:14-21.
8. Dawson DG, Randleman JB, Grossniklaus HE, O’Brien TP, Dubovy SR, Schmack I, Stulting RD, Edelhauser HF. Corneal ectasia after excimer laser keratorefractive surgery: histopathology, ultrastructure, and pathophysiology. Ophthalmology. 2008;115:2181-2191.
9. Randleman JB, Caster AI, Banning CS, Stulting RD. Corneal ectasia after photorefractive keratectomy. J Cataract Refract Surg. 2006;32:1395-1398.
10. Malecaze F, Coullet J, Calvas P, Fournié P, Arné J-L, Brodaty C. Corneal ectasia after photorefractive keratectomy for low myopia. Ophthalmology. 2006;113:742-746.
11. Guedj M, Saad A, Audureau E, Gatinel D. Photorefractive keratectomy in patients with suspected keratoconus: five-year follow-up. J Cataract Refract Surg. 2013;39:66-73.
12. Koller T, Iseli HP, Donitzky C, Papadopoulos N, Seiler T. Topography-guided surface ablation for forme fruste keratoconus. Ophthalmology. 2006;113:2198-2202.
13. Shetty R, Nuijts RMMA, Nicholson M, Sargod K, Jayadev C, Veluri H, Sinha Roy A. Cone location–dependent outcomes after combined topography-guided photorefractive keratectomy and collagen cross-linking. Am J Ophthalmol. 2015;159:419-425.
14. Stojanovic A, Chen X, Jin N, Zhang T, Stojanovic F, Raeder S, Utheim TP. Safety and efficacy of epithelium-on corneal collagen cross-linking using a multifactorial approach to achieve proper stromal riboflavin saturation. J Ophthalmol 2012. article ID:498435.
15. Bikbova G, Bikbov M. Transepithelial corneal collagen cross-linking by iontophoresis of riboflavin. Acta Ophthalmol. 2014;92:e30-e34.
16. Tomita M, Mita M, Huseynova T. Accelerated versus conventional corneal collagen crosslinking. J Cataract Refract Surg. 2014;40:1013-1020.
17. Roy AS, Dupps WJ. Patient-specific computational modeling of keratoconus progression and differential responses to collagen cross-linking. Invest Ophthalmol Vis Sci. 2011;52:9174-9187.
18. Kling S, Hafezi F. An algorithm to predict the biomechanical stiffening effect in corneal cross-linking. J Refract Surg. 2017;33:128-136.
19. Seven I, Sinha Roy A, Dupps WJ Jr. Patterned corneal collagen crosslinking for astigmatism: computational modeling study. J Cataract Refract Surg. 2014;40:943-953.
20. Kanellopoulos AJ. Novel myopic refractive correction with transepithelial very high-fluence collagen cross-linking applied in a customized pattern: early clinical results of a feasibility study. Clin Ophthalmol. 2014;8:697-702.
21. Elling M, Kersten-Gomez I, Dick HB. Photorefractive intrastromal corneal crosslinking for the treatment of myopic refractive error: findings from 12-month prospective study using an epithelium-off protocol. J Cataract Refract Surg. 2018;44:487-495.
22. Schallhorn JM, Schallhorn SC, Hettinger K, Hannan S. Infectious keratitis after laser vision correction: incidence and risk factors. J Cataract Refract Surg. 2017;43:473-479. erratum 2018; 44:120.
23. Koller T, Mrochen M, Seiler T. Complication and failure rates after corneal crosslinking. J Cataract Refract Surg. 2009;35:1358-1362.
24. Wollensak G, Iomdina E. Biomechanical and histological changes after corneal crosslinking with and without epithelial debridement. J Cataract Refract Surg. 2009;35:540-546.
25. Heikal MA, Soliman TT, Fayed A, Hamed AM. Efficacy of transepithelial corneal collagen crosslinking for keratoconus: 12-month follow-up. Clin Ophthalmol. 2017;11:767-771.
26. Lesniak SP, Hersh PS. Transepithelial corneal collagen crosslinking for keratoconus: six-month results. J Cataract Refract Surg. 2014;40:1971-1979.
27. Kamaev P, Friedman MD, Sherr E, Muller D. Photochemical kinetics of corneal cross-linking with riboflavin. Invest Ophthalmol Vis Sci. 2012;53:2360-2367.
28. Richoz O, Hammer A, Tabibian D, Gatzioufas Z, Hafezi F. The biomechanical effect of corneal collagen cross-linking (CXL) with riboflavin and UV-A is oxygen dependent. Transl Vis Sci Technol. 2013;2(7):6.
29. Kamaev P, Friedman MD, Sherr E, Muller D. Photochemical kinetics of corneal cross-linking with riboflavin. Invest Ophthalmol Vis Sci. 2012;53:2360-2367.
30. Filippello M, Stagni E, O’Brart D. Transepithelial corneal collagen crosslinking: bilateral study. J Cataract Refract Surg. 2012;38:283-291. erratum, 1515.
31. Kymionis GD, Tsoulnaras KI, Grentzelos MA, Plaka AD, Mikropoulos DG, Liakopoulos DA, Tsakalis NG, Pallikaris IG. Corneal stroma demarcation line after standard and high-intensity collagen crosslinking determined with anterior segment optical coherence tomography. J Cataract Refract Surg. 2014;40:736-740.
32. International Organization for Standardization., 2007. Ophthalmic Instruments – Fundamental Requirements and Test Methods – Part 2: Light Hazard Protection, ISO, Geneva, Switzerland, 15004-2:2007.
33. Spoerl E, Mrochen M, Sliney D, Trokel S, Seiler T. Safety of UVA–riboflavin cross-linking of the cornea. Cornea. 2007;26:385-389.


None of the authors has a financial or proprietary interest in any material or method mentioned.

© 2019 by Lippincott Williams & Wilkins, Inc.