Introduction
Collagen cross-linking (CXL) was proposed by Wollensak et al. as a new possibility to stabilize progressive keratoconus, preventing some of the underlying pathophysiologic mechanisms of the disease. This shows promise in the attempt to stop progressive visual loss due to the evolution of the pathology and to delay or avoid invasive surgical procedures such as corneal transplantation, which is usually required in advanced cases.
Since this first report by Wollensak et al. in 2003, numerous publications have been added to the peer-reviewed literature over the last decade addressing safety and efficacy of CXL in treating adult eyes with keratoconus and in other corneal ectatic conditions. These studies have provided sufficient evidence that CXL is successful in slowing or halting keratoconus progression and may even yield visual, topographic, and aberrometric improvement by induced corneal flattening and reduction in irregular astigmatism. It is important to note that medium and long-term studies have validated an excellent safety profile for standard CXL (epi-off Dresden protocol), with respect to the health of the corneal endothelium, lens, and retina despite the potential cytotoxic effect of ultraviolet A (UVA) light. Since no permanent side effects and an acceptable complication rate were observed in adults when strict inclusion criteria were adhered to, the introduction of corneal collagen cross-linking (CXL) in routine clinical practice has changed the management of keratoconus in both the adult and pediatric populations.
Keratoconus is a progressive, frequently asymmetric, noninflammatory corneal dystrophy, characterized by changes in the corneal collagen structure and organization, causing a biomechanical instability that leads to irregular astigmatism, progressive myopia, corneal thinning, and central corneal scarring, with subsequent mild to marked impairment in visual quality. Many studies on keratoconus epidemiology from different countries reported an incidence of 1.3 to 22.3 per 100,000 and a prevalence of 0.4 to 86 cases per 100,000. The incidence of corneal ectasia after refractive surgery is still unknown, but it has been estimated to be 0.04% to 0.6% after laser in situ keratomileusis (LASIK). Post-LASIK ectasia represents about 96% of all secondary ectasias after refractive surgery, while 4% are related to photorefractive keratectomy (PRK).
The disease affects both male and female. However, not all age groups are affected equally, since the onset of the disease is typically during adolescence and puberty. Keratoconus etiology is not yet completely understood and includes genetic, biochemical, and physical factors. It usually appears as an isolated condition but has been associated with several ocular and systemic disorders. A reduced number of collagen cross-links and a pepsin digestion higher than normal have been suggested as possible explanations for an overall structural weakness of the corneal tissue in keratoconus, resulting in a stiffness that is only 60% of the normal cornea. However, its etiologic basis remains poorly understood: a defective formation of extracellular constituents of corneal tissue, fewer collagen lamellae, less collagen fibrils per lamella, and closer packing of collagen fibrils could all decrease mechanical corneal stability.
The incidence of keratoconus is higher in relatives of patients with the disorder than in the general population. Retinitis pigmentosa, blue sclera, magnesium deficiency, Down syndrome, Turner syndrome, Marfan syndrome, Ehlers-Danlos syndrome, mental retardation, Leber congenital amaurosis, osteogenesis imperfecta, and pseudoxanthoma elasticum have been reported to be correlated with keratoconus. Approximately 6% to 24% of cases demonstrate clinically recognized familial aggregation. Both dominant and recessive models have been observed in individual keratoconus pedigrees. In addition, segregation analyses, twin studies, and gene mapping studies have also indicated the important role of genetic factors.
In the vast majority of patients (> 90%) keratoconus is bilateral. However, the eyes are affected with different severity. In many cases, the disorder may start unilaterally and involve the other eye over time. This bilaterality supports the assumption of a genetic basis for this disease.
In addition to genetic factors, many pediatric keratoconus patients show ocular comorbidities, such as surface allergy, atopic dermatitis, and especially vernal keratoconjunctivitis (VKC). VKC compounds the problems of keratoconus, as continued surface inflammation and the tendency to rub the eyes accelerate keratoconus degeneration. Therefore it is recommended that children with atopy be referred to a comprehensive ophthalmic examination, even in the apparent absence of visual symptoms, to ensure the timely diagnosis and management of any atopy-associated ocular disease. Prompt referral is particularly essential for pediatric corneal ectasia, in which the rapidity of progression may preclude stabilizing treatments and may result in significant childhood visual impairment. In addition, VKC should be controlled aggressively prior to CXL and patients and their parents should be counseled about avoiding eye rubbing. Care should be taken to protect limbal stem cells during irradiation with UV-light during CXL. Although keratoconus is frequently diagnosed after adolescence, the corneal ectasia process usually starts at a much younger age.
Recent studies conducted in Germany suggest that thyroid gland dysfunction due to inflammatory or immunologic causes is associated with keratoconus and might correlate with the onset and progression of the disease. Good medical practice aims to screen patients with hypothyroidism using corneal topography to detect early stage keratoconus. Conversely, serologic tests aimed to identify any malfunctioning thyroid should be recommended to each patient with keratoconus. Other reports suggest that hormonal changes occurring regularly during gestation may modify the function of the thyroid gland and may affect corneal biomechanics negatively and may have a severe impact on the progression of keratoconus.
Basic Principles of Corneal Cross-Linking
The primary aim of corneal CXL is to stop the progression of corneal ectasia. To obtain a strengthening of corneal tissue, the use of riboflavin is combined with UVA irradiation. Riboflavin plays the role of a photosensitizer in the photopolymerization process and, when combined with UVA irradiation, increases the formation of intrafibrillar and interfibrillar carbonyl-based collagen covalent bonds through a molecular process that has still not been completely elucidated. It was shown that during the early aerobic phase of the process of CXL, riboflavin molecules are excited to a single or triplet state and stromal proteins undergo a photosensitized oxidation via interaction with reactive oxygen species. During the second anaerobic phase, when oxygen is depleted, corneal stroma interacts with reactive species of radical ions. This photochemical reaction results in an increased corneal rigidity, increased collagen fiber thickness, and increased resistance to enzymatic degradation.
Basic Research Results
Currently, the photochemically induced effect of CXL in the cornea cannot be observed directly by staining methods or microscopic techniques. However, CXL induces several changes to collagen-containing tissue from which indirect signs of the CXL effect can be deduced. In fact, stress–strain measurements performed on human and porcine corneas documented an increased corneal rigidity after CXL treatment. The strengthening effect seems to be more evident in corneas with higher collagen content and in older tissue. Corneal CXL occurs physiologically with aging via natural enzymatic pathways such as transglutaminase and lysyl oxidase.
Moreover, it has been reported that porcine cross-linked corneas showed a reduced tendency to swelling and hydration when compared to untreated controls. Ex vivo studies on corneas of humans and rabbits indicated an increase of collagen fiber thickness after CXL treatment. Results of basic research studies showed that CXL improves the corneal resistance to degradation processes mediated by pepsin, trypsin, and collagenase with lengthening of the turnover time of collagen.
Standard Cross-Linking Procedure: Epi-Off Cross-Linking ( )
Prior to CXL, it is necessary to obtain documented evidence of keratoconus progression. In order to reduce the risk of endothelial damage, it must be ensured that minimum corneal thickness preoperatively is greater than 400 µm.
Pediatric keratoconus (keratoconus manifesting in patients younger than 18 years of age) exhibits several unique characteristics. Studies have shown that pediatric keratoconus is often more advanced at diagnosis than in adults, with 27.8% being stage 4 vs 7.8% in adults. In addition, pediatric keratoconus demonstrates a higher rate (88% of keratoconic eyes) and speed of progression than adult keratoconus. The biomechanical rigidity of the cornea is related to age and children with keratoconus are frequent eye rubbers, especially the subgroup of children with coexisting VKC. Since the progression of the disease can be dramatically fast in children, early detection of the disease and close monitoring are crucial in young patients.
CXL is generally performed under topical anesthesia except in noncompliant patients, in whom general anesthesia may be necessary. Thirty minutes before the procedure, systemic pain medication is administered and pilocarpine 2% drops are instilled in the eye to be treated. After topical anesthesia with two applications of lidocaine 4% and oxybuprocaine hydrochloride 0.2% drops, the eye is draped, the ocular surface is rinsed with balanced salt solution, and a lid speculum is applied. The corneal epithelium is abraded in a central 9-mm-diameter area with the aid of an Amoils brush. A riboflavin 0.1% solution (10 mg riboflavin-5-phosphate in 20% dextran-T-500) is applied onto the cornea every minute for 30 minutes to achieve adequate penetration of the solution. Using a slit lamp with blue filter, the surgeon confirms the presence of riboflavin in the anterior chamber. Then, the cornea is exposed to an ultraviolet light emanating from a solid-state device emitting a wavelength of 370.5 nm and an irradiance of 3 mW/cm 2 . A calibrated UVA meter is used before treatment to check the irradiance. The cropped light beam has a 7.5-mm diameter. Exposure time is 30 minutes. During irradiation, riboflavin solution is applied once every 5 minutes to avoid desiccation of the cornea. Intraoperative pachymetry is performed throughout the procedure. In case corneal thickness goes below 400 µm, hypotonic riboflavin solution can be used to swell the cornea. Fixation is achieved by instructing the patient to focus on the central LED of the probe. During the procedure, the surgeon controls for centration of treatment. Both topical anesthetics are added as needed during irradiation. After surgery, patients receive cyclopentolate and levofloxacin drops. A soft bandage contact lens is applied until reepithelialization is complete. Topical levofloxacin is given 4 times daily for 7 days, dexamethasone 21-phosphate 0.15% drops are administered 3 times daily for 20 days, and sodium hyaluronate 0.15% drops are applied 6 times daily for 45 days. In addition, patients receive oral amino-acid supplements for 7 days.
Postoperative CXL follow-up includes daily examinations until reepithelialization occurs (on average, on the third or fourth day). The patient should avoid dusty and windy places to minimize the risk of corneal infection. The use of amino acid supplements and antioxidants in the immediate preoperative and postoperative period is recommended for regular corneal reepithelialization. For the first month after treatment, the patient should avoid saunas, swimming pools and baths, and direct sunlight exposure without appropriate sunglasses. Studies have shown that the corneal epithelium is a significant barrier for penetration of both UVA light and riboflavin, which is a hydrophilic molecule that cannot easily pass the tight junctions of the intact epithelial barrier. A variety of approaches—including transepithelial procedures (using 20% alcohol solutions or tetracaine 1% to loosen epithelial tight junctions), partial epithelial removal, or femtosecond (FS) laser–created intrastromal pockets—have been attempted to improve riboflavin penetration in the presence of intact epithelium, reduce postoperative discomfort, and accelerate visual recovery. Novel formulations of riboflavin (by adding trometamol and sodium ethylene-diaminetetraacetic acid or sodium or benzalkonium chloride) have been developed to facilitate transepithelial diffusion. However, to date, none has been close to reaching the efficacy of the epithelium off (epi-off) technique. Raiskup-Wolf et al. reported no changes in the biomechanical properties of corneal tissue after CXL was performed with intact epithelium, confirming the need for complete epithelium removal to allow sufficient stromal uptake of riboflavin.
Rapid Accelerated Technique
Following the Bunsen-Roscoe law of reciprocity, the same UVA dosage can be administered by increasing the UVA fluence while simultaneously reducing the exposure time, maintaining efficacy and safety of the technique with a substantial reduction of treatment time. Preclinical in vivo studies have been encouraging. However, a sudden decrease of efficacy has been observed using UV light with very high intensity (> 45 mW/cm 2 ) probably due to a reduced availability of oxygen, which has been shown to limit the photochemical CXL process. So far, few studies with a limited follow-up of 6 months have demonstrated the same efficacy of accelerated CXL to standard protocol of 3 mW/cm 2 UVA and 30 minutes of exposure.
Transepithelial Cross-Linking
Transepithelial CXL has been adopted in various reports, especially in the pediatric keratoconic population, to reduce postoperative pain and the risk of corneal infections and corneal opacities in addition to reducing potential harmful effects on the endothelium in this young population. The most promising technique is enhanced transepithelial riboflavin absorption using iontophoretic delivery . Riboflavin is a small negatively charged molecule at physiologic pH and is easily soluble in water; therefore it is a suitable molecule for iontophoretic transfer. It has been shown that, using iontophoresis, an imbibition time of only 5 minutes achieves a sufficient riboflavin concentration in the corneal stroma for CXL treatment while preserving epithelial integrity.
Numerous ex vivo studies confirmed the effectiveness of iontophoresis imbibition in obtaining an adequate riboflavin concentration in the stroma and the induction of important biomolecular and structural modifications of corneal tissue. Ex vivo biomechanical studies on rabbit and human cadaveric corneas showed that transepithelial CXL with iontophoresis imbibition induced an increase of the biomechanical resistance comparable to that obtained with the standard CXL procedure.
Preliminary clinical results of iontophoresis-assisted corneal CXL are promising. The technique halts keratoconus progression without significant complications. However, longer follow-up and studies with larger patient populations are needed.
Ultrasound, nanoemulsion systems and other epithelial permeation enhancers such as vitamin E-TPGS are currently under preclinical investigation to facilitate transepithelial riboflavin penetration.
Cross-Linking Results
Several published studies present outcomes of CXL treatments (standard epi-off and transepithelial) in adult and pediatric keratoconic patients.
Standard epi-off CXL induces a significant improvement of both uncorrected visual acuity (UCVA) and best spectacle-corrected visual acuity (BSCVA) during the first year after CXL, thereafter remaining unchanged up to 3 years after the procedure. Visual acuity improves due to a progressive topographic flattening of the cornea over time with a reduction of simulated keratometry, minimum keratometry, mean average corneal power, and asymmetry indices. Soeters et al. observed that, before CXL, cones of pediatric keratoconic corneas were located more centrally than in the older age group.
Significant reductions in mean spherical equivalent were observed, especially during the first year after CXL, with a reduction in corneal aberrations, including coma. Minimum corneal thickness is typically reduced during the first 6 months after CXL, recovering to preoperative values within 1 year of the procedure. No endothelial cell loss was observed within the first 4 years after the procedure. Abrasion-related discomfort was reported by most patients in the immediate postoperative period. No ocular or systemic adverse events were noted apart from a low incidence of blepharitis and photophobia up to 4 months after the procedure. No significant intraocular pressure change was seen. In most of the eyes, CXL-specific golden striae developed and in some eyes a moderate haze was observed, which disappeared after the use of topical steroids. Transient haze appearing at 2 to 6 weeks and clearing at 9 to 12 months is the result of an increased density of extracellular matrix and arises at a depth of 300 to 350 µm. It forms the demarcation line that can be seen at slit lamp examination and with optical coherence tomography (OCT; Figs. 22.1 and 22.2 ). Persistent haze has been observed in eyes with corneal apex power higher than 72 diopters (D) and central pachymetry thinner than 420 µm ( Fig. 22.3 ). Sterile infiltrates may occur in the early postoperative period and usually resolve with the use of topical steroids ( Figs. 22.4 and 22.5 ).
