Research Article: Corneal Absorption of a New Riboflavin-Nanostructured System for Transepithelial Collagen Cross-Linking

Date Published: June 13, 2013

Publisher: Public Library of Science

Author(s): Katia M. Bottos, Anselmo G. Oliveira, Patrícia A. Bersanetti, Regina F. Nogueira, Acácio A. S. Lima-Filho, José A. Cardillo, Paulo Schor, Wallace Chamon, Che John Connon.


Corneal collagen cross-linking (CXL) has been described as a promising therapy for keratoconus. According to standard CXL protocol, epithelium should be debrided before treatment to allow penetration of riboflavin into the corneal stroma. However, removal of the epithelium can increase procedure risks. In this study we aim to evaluate stromal penetration of a biocompatible riboflavin-based nanoemulsion system (riboflavin-5-phosphate and riboflavin-base) in rabbit corneas with intact epithelium. Two riboflavin nanoemulsions were developed. Transmittance and absorption coefficient were measured on corneas with intact epithelia after 30, 60, 120, 180, and 240 minutes following exposure to either the nanoemulsions or standard 0.1% or 1% riboflavin-dextran solutions. For the nanoemulsions, the epithelium was removed after measurements to assure that the riboflavin had passed through the hydrophobic epithelium and retained within the stroma. Results were compared to de-epithelialized corneas exposed to 0.1% riboflavin solution and to the same riboflavin nanoemulsions for 30 minutes (standard protocol). Mean transmittance and absorption measured in epithelialized corneas receiving the standard 0.1% riboflavin solution did not reach the levels found on the debrided corneas using the standard technique. Neither increasing the time of exposure nor the concentration of the riboflavin solution from 0.1% to 1% improved riboflavin penetration through the epithelium. When using riboflavin-5-phosphate nanoemulsion for 240 minutes, we found no difference between the mean absorption coefficients to the standard cross-linking protocol (p = 0.54). Riboflavin nanoemulsion was able to penetrate the corneal epithelium, achieving, after 240 minutes, greater stromal concentration when compared to debrided corneas with the standard protocol (p = 0.002). The riboflavin-5-phosphate nanoemulsion diffused better into the stroma than the riboflavin-base nanoemulsion.

Partial Text

Corneal collagen cross-linking (CXL) has emerged as a promising and innovative treatment for progressive keratoconus and surgically-related ectasia [1]. Recently, a significant number of studies have used this treatment for a wider range of applications, including microbial keratitis [2] and prevention of keratectasia prior to refractive surgery [3]. The standard CXL protocol proposed by Wollensak et al. recommended removal of the epithelium before treatment to allow penetration of riboflavin into the stroma [1]. Experimental and clinical investigations have shown that the intact epithelium does not block the effect of ultraviolet A (UVA) light [4], but decreases the effectiveness of the treatment by impairing the adequate stromal diffusion of riboflavin [5]–[11]. However, de-epithelialization is not risk-free and can increase the rate of corneal infections, haze, scarring, and infiltrates [12]–[16]. Moreover, it can cause postoperative pain, photophobia, and delayed visual rehabilitation [1]. Some modifications of the initial treatment have been suggested to overcome these disadvantages. Studies reported the use of benzalkonium chloride (BAK) to loosen tight junctions of corneal epithelial cells and facilitate riboflavin stromal diffusion [9], [17]. However, both in vitro and in vivo studies have demonstrated adverse effects of BAK on epithelial cell populations [18]–[21]. Recently, BAK-induced corneal neurotoxicity has been demonstrated [21]. Riboflavin in association with anesthetic drops poorly penetrated the cornea [7], and even superficial epithelial trauma was not sufficient to permit the adequate penetration of riboflavin into the stroma. Samaras et al. confirmed these findings not being able to increase penetration of riboflavin by using 20% alcohol in one group, and a grid pattern epithelial debridement in another group [6]. Neither groups had an acceptable and homogeneous riboflavin penetration. Furthermore, prolonging the exposure time to standard hydrophilic riboflavin did not lead to increased stromal saturation in corneas with intact epithelium [22]. Other alternatives to avoid complete epithelial debridement include a central corneal pocket created by a femtosecond laser that allows for direct intrastromal drug administration [23], [24]. However, one should be concerned regarding biomechanical stability after stromal dissection by the femtosecond laser. A riboflavin complex, with edetate disodium (EDTA) and tromethamine, has been tested for transepithelial CXL. However, due to its limited permeation through the epithelium, the CXL effect was reduced [25]. Recently, cyclodextrins were successfully used to enhance riboflavin solubility in water and to improve its permeability through fresh and cryopreserved bovine corneas [26]. In summary, none of the present modalities has proven to be clinically effective in delivering an adequate intrastromal concentration of riboflavin, hence a better method still needs to be developed.

The average central corneal thickness measured with ultrasound was 580.0±52.8 µm and 529.0±49.3 µm, before and after epithelium removal, respectively. Both formulations of riboflavin tested showed absorption peaks between 360 and 375 nm (Figure 2). Thus, we used a solid-state device, available in our laboratory, set at 365 nm.

During CXL, stromal diffusion of riboflavin is necessary because it is the basis of CXL safety and effectiveness. Riboflavin is not only important as a photosensitizer, but also because it shields UVA light from damaging the underlying ocular structures such as the endothelium, lens, and retina. UVA radiation decreases as it passes through the cornea, because riboflavin in the cornea absorbs the highest amount of UVA, preventing high doses of this radiation from reaching the posterior ocular structures [36], [37]. Until now, epithelium removal is the standard method to facilitate penetration of riboflavin into the corneal stroma. In the present study, we describe a new option to achieve stromal diffusion of riboflavin without removal of the epithelium and without the use of benzalkoniun chloride.