Corneal collagen crosslinking (CXL) is a procedure that uses riboflavin (vitamin B₂) as a photosensitizer and irradiates the corneal stroma with UVA light at a wavelength of 365 nm to increase crosslinks between collagen fibers 2). The first clinical report for progressive keratoconus was published in 2003 1), and it has since become widely adopted.
The FDA has approved CXL for progressive keratoconus and post-refractive surgery corneal ectasia in patients aged 14 to 65 years 2). The primary goal of CXL is to halt the progression of corneal ectasia, with improvement in corneal optical properties being a secondary effect 2).
CXL has two main approaches: the Epi-off method (epithelium-off), which removes the corneal epithelium, and the Epi-on method (transepithelial), which preserves the epithelium.
Epi-off Method
Riboflavin absorption: Epithelium removal allows good absorption of riboflavin and UVA.
Effect: Corneal stiffening and flattening are more pronounced, with superior progression inhibition effects 1).
Pain: Postoperative pain is severe, and it takes time for vision to recover.
Epi-on method
Recovery: Quick return to contact lens wear.
Pain: Short duration of discomfort.
Risk: Lower risk of corneal infection and haze, but efficacy may be inferior to the Epi-off method1).
| Item | Details |
|---|---|
| Indications | Progressive keratoconus, post-LASIK corneal ectasia |
| Contraindications | Active infection, history of herpes simplex virus/HZV, corneal stromal scar |
| Corneal thickness | Recommended ≥400 μm after epithelial removal |
The definition of progression is not standardized, but is generally determined by any of the following within 12–18 months: an increase in Kmax of ≥1 D, an increase in mean corneal refractive power, an increase in refractive astigmatism of ≥1 D, or a decrease in corneal thickness of ≥10%1). Caution is required in patients with a history of herpes simplex keratitis due to the risk of viral reactivation by UVA irradiation2).
Conventionally, a corneal thickness of less than 400 μm after epithelial removal has been considered a contraindication for CXL. However, a protocol using a hypotonic riboflavin solution to swell the cornea to 400 μm or more before UVA irradiation (sub400 protocol) has been reported 2). A protocol that individually adjusts UV irradiation time for cases with corneal thickness of 200–400 μm has also been reported 2). In the Epi-on method, evaluation is based on corneal thickness including the epithelium, so cases that would be less than 400 μm after epithelial removal may still be eligible for treatment.
The standard Epi-off method is called the Dresden protocol and is performed as follows 1).
A method using PTK (phototherapeutic keratectomy) for epithelial removal has also been reported. Dextran-free HPMC suspension is also used as the riboflavin solution. HPMC preparations have the advantage of reducing intraoperative corneal dehydration 1).
Accelerated protocols that increase irradiation intensity and shorten irradiation time have also been attempted. While maintaining a total energy of 5.4 J/cm², for example, there are protocols of 9 mW/cm² for 10 minutes or 30 mW/cm² for 3 minutes. Some studies report equivalent efficacy to the standard method, while others report inferior corneal flattening or stiffening, so results are not consistent 1).
Standard Dresden CXL halts the progression of keratoconus (the main purpose of treatment). A meta-analysis of 75 studies with follow-up of 36 months or more showed improvement in uncorrected visual acuity, late reduction in corneal refractive power, and a certain decrease in astigmatism 2).
In an FDA phase III trial (205 patients), the group treated with the Dresden protocol showed a decrease in Kmax of 1.6 ± 4.2 D from baseline at 1 year, while the control group continued to progress 2). In a trial for post-LASIK ectasia (179 patients), Kmax in the treatment group decreased by 0.7 ± 2.1 D 2).
In a prospective RCT by Wittig-Silva et al., Kmax significantly flattened in the treatment group (-1.45 ± 1.00 D, P < 0.002) while progressing in the control group (+1.28 D, P < 0.001). Long-term follow-up showed sustained flattening even after 4 years 1).
The accelerated protocol reduces treatment time by increasing irradiation intensity and shortening irradiation time, but the evidence is inconsistent 1). While some studies show comparable clinical parameters to the standard method, others report inferior corneal flattening and stiffening. Studies comparing the Epi-on accelerated method with the Epi-off standard method have shown equivalent efficacy, but data are limited.
In the Epi-on method, riboflavin is penetrated into the corneal stroma while preserving the corneal epithelium. Because the corneal epithelium is hydrophobic, riboflavin penetration is difficult, and the following techniques are used to promote penetration 1).
Chemical additives: Benzalkonium chloride (BAC), EDTA, trometamol, local anesthetics, etc. are added to increase the permeability of the epithelial barrier.
Iontophoresis: A technique that uses an electrical gradient to move charged riboflavin molecules across the epithelium.
Increased riboflavin concentration: Solutions with increased concentration from 0.1% to 0.25–0.5% may be used.
No standardized protocol has been established, and various riboflavin solutions and additives are used in different studies. Currently, there is no recommended standard Epi-on method 1).
A systematic review of the Epi-on method showed that UDVA improved by 0.22–0.28 logMAR at 3–12 months after CXL. However, changes in CDVA and reduction in Kmax were not statistically significant.
Some comparative studies found differences in Kmax reduction between the Epi-off and Epi-on groups. In a study by Al Fayez et al. (36-month follow-up), Kmax decreased by -2.4 D in the Epi-off group, whereas it increased by +1.1 D in the Epi-on group (P < 0.0001) 1). These results suggest that the efficacy of the Epi-on method may be inferior to that of the Epi-off method 1).
Comparative studies consistently show less pain with the Epi-on method. In a study of 70 patients, the mean pain score on a 1–5 scale was 2 in the Epi-on group and 4 in the Epi-off group (P = 0.0035).
| Parameter | Epi-off method | Epi-on method |
|---|---|---|
| Kmax change | Significant flattening | Limited effect |
| CDVA | Significant improvement | Improvement unclear |
A meta-analysis including 8 comparative studies found no significant difference in UDVA and CDVA at 1-year follow-up. However, the reduction in mean K value was significantly greater with the Epi-off method (standardized mean difference 0.28, P = 0.03). The difference was even larger when compared with the Epi-on method using iontophoresis (standardized mean difference 0.43, P = 0.01).
The KERALINK trial was an RCT involving 60 patients aged 10–16 years with progressive keratoconus 3). In the CXL group, K2 (steep keratometry) at 18 months was 49.7 ± 3.8 D, compared with 53.4 ± 5.8 D in the standard treatment group, with an adjusted mean difference of -3.0 D (95% CI: -4.9 to -1.1, P = 0.002), significantly better in the CXL group 3). Progression occurred in 7% of the CXL group versus 43% in the standard treatment group, and CXL reduced the risk of progression by 90% (OR 0.1, P = 0.004) 3).
CXL halts the progression of keratoconus in the majority of young patients. CXL should be considered as a first-line treatment for progressive disease 3).
In a prospective cohort study with 5-year follow-up of 78 pediatric eyes, the Epi-off group achieved a mean corneal flattening of 3.18 D, whereas the accelerated Epi-on group achieved only 0.09 D. Progression was observed in 9.4% (3/32 eyes) of the accelerated Epi-on group, but no progression was observed in the Epi-off group at 5 years.
At present, evidence has accumulated that the Epi-off method is superior in terms of corneal flattening and progression suppression 1). In particular, long-term data show that progression suppression persists for 5 years with Epi-off, whereas some cases of progression are observed with Epi-on. On the other hand, Epi-on has the advantage of less pain and faster recovery. There are also reports that the chemically enhanced Epi-on method is equivalent to Epi-off in terms of K-value reduction compared to the iontophoresis method. The final choice of technique should be made on a case-by-case basis.
Complications of Epi-off
Corneal haze: Appears 1–2 months after surgery and usually resolves within 6–12 months. The incidence of permanent stromal scarring has been reported to be up to 8.6% 1).
Infectious keratitis: Reported in 0.0017% of cases. All were Epi-off procedures 1).
Sterile infiltrates: Appear early postoperatively and resolve with steroid eye drops 1).
Persistent epithelial defect: Delayed epithelial healing may pose a risk of corneal melting.
Complications of Epi-on
Hyperemia and foreign body sensation: Mostly transient findings that resolve within 24 hours after surgery.
Photophobia: Reported as an early transient symptom.
Corneal epithelial damage: In accelerated Epi-on trials, photokeratitis with diffuse punctate superficial keratopathy has been reported.
Overall trend: Haze, scarring, and microbial infection have only been reported in studies of the Epi-off method.
Other complications include corneal edema, hyperopia due to excessive corneal flattening, and endothelial damage 1). The risk of endothelial damage increases when UVA irradiation is applied to corneas thinner than 400 μm. Transient decrease in endothelial cell count is observed but usually normalizes within 6 months 2).
CXL uses riboflavin as a photosensitizer to induce localized photopolymerization in the corneal stroma, increasing biochemical bonds between collagen fibers 2). This stiffens the structurally weak cornea and suppresses the progression of ectasia.
Irradiation with UVA light (365 nm) excites riboflavin, producing reactive oxygen species. These reactive oxygen species form new covalent bonds between collagen molecules, leading to an increase in collagen fiber diameter and inter-fiber spacing. This change is concentrated in the anterior 300 μm of the corneal stroma.
The corneal epithelium acts as a barrier to riboflavin penetration. In the Epi-off method, epithelial removal allows riboflavin to directly reach the corneal stroma. In the Epi-on method, the amount of riboflavin passing through the epithelium is limited, reducing crosslink formation in the stroma and potentially decreasing efficacy 1). Chemical additives and iontophoresis are attempts to overcome this limitation, but no standardized method has been established 1).
Biomechanical parameters such as corneal resistance factor and corneal hysteresis show small changes after CXL, but custom variables derived from the same device have shown changes suggesting stiffening 2).
The KERALINK trial was an RCT targeting young keratoconus patients aged 10–16 years, showing that CXL halted keratoconus progression in the majority of young patients at 18 months 3). The progression rate was 7% in the CXL group versus 43% in the standard treatment group, representing a 90% reduction in progression risk 3). These results support early consideration of CXL as a first-line treatment, especially in young patients with rapid progression 3).
