
Corneal collagen cross-linking (CXL) treatment, using riboflavin and ultraviolet (UV) light, is a novel procedure that has been developed to address this clinical need. We review the clinical characteristics of keratoconus and provide an evidence-based update on the efficacy and safety of CXL.
Keratoconus, first described in detail in 1854,1 derives from the Greek words Kerato (cornea) and Konos (cone); it is a progressive, bilateral but typically asymmetric, non-inflammatory ectasia of the cornea that is characterised by progressive thinning of the axial corneal stroma. The condition has both physical and chemical markers that arise from the cellular malformation of the corneal stroma. Epidemiological studies indicate that the prevalence of keratoconus is 5.4 per 10,000;2,3 the condition can occur in all ethnic groups with no male or female preponderance.4 While keratoconus typically occurs in isolation, the most common recognised systemic associations are generalised atopy,5 Down syndrome6 and Marfan’s syndrome.7 Ocular comorbidities include Leber’s congenital amaurosis8 and retinitis pigmentosa.9 Keratoconus can vary significantly in its clinical course, but classically manifests at puberty and is progressive until the third or fourth decade, when it usually arrests. It may however, commence later in life and progress or stabilise at any age.4
Approximately 15 per cent of affected individuals will require a corneal transplant, due to inadequate vision correction with contact lenses, contact lens intolerance and/or visually debilitating corneal scarring.10 At present, the aetiology of keratoconus is uncertain but may involve genetic, biochemical and environmental factors.
Keratoconus can vary significantly in its clinical course, but classically manifests at puberty and is progressive until the third or fourth decade, when it usually arrests…
Clinical Features
The clinical hallmark of keratoconus is progressive thinning of the axial corneal stroma, resulting in protrusion of the corneal apex to assume a conical shape. The thinner apex is typically downwardly displaced, leading to irregular astigmatism and visual distortion.5 Although keratoconus affects both eyes it can be highly asymmetric, with the condition far more advanced in one eye compared with the other.
Figure 1 – Biomicroscopic signs of keratoconus. A: Fleischer ring, consisting of an iron line at the base of the cone (arrow). B: Vogt’s striae, as indicated by the presence of finevertical lines in the deep corneal stroma. C: Apical corneal scarring (arrow).
Figure 2 – Corneal axial power maps showing different cone morphologies in keratoconus. A: Centred (nipple) cone consisting of localized steepening at the corneal apex. B: Oval (sagging) cone with corneal steepening located inferior to the visual axis. C: Larger oval cone showing a more substantial area of inferior corneal steepening.
Patient symptoms are variable and depend upon the severity of the disease. Early symptoms typically include monocular diplopia, mild photophobia and a history of deteriorating vision with spectacles.11,12 Advanced keratoconus is associated with more significant visual impairment and displays such symptomology as: poor best-corrected spectacle acuity, haloing around lights, a rapidly changing subjective refraction and generalised asthenopia.
Fortunately, much of this visual distortion can be corrected with appropriately-fitted rigid gas permeable lenses which are the mainstay of optometric management. Hybrid, mini-scleral and scleral contact lenses are further contact lens modalities that may be appropriate for patients with keratoconus.
Biomicroscopic indicators of keratoconus may include one, or more, of the following signs (Figure 1):
- Central or paracentral corneal thinning (usually inferior or infero-temporal);
- A complete, or incomplete, iron line at the base of the cone (Fleischer ring) is observed in approximately half of patients; it represents the accumulation of iron deposits from the tear film onto the cornea as a result of severe corneal curvature changes and/or modification of the normal epithelial slide process;
- Fine vertical lines located in the deep corneal stroma and Descemet’s membrane that parallel the axis of the cone (Vogt’s striae), produced by compression of Descemet’s membrane;
- Increased visibility of the corneal nerves;
- Apical corneal scarring, which may be due to ruptures in Bowman’s membrane or the result of a flat-fitting contact lens;
- A V-shaped deformation of the lower lid produced by an advanced, ectatic cornea in down-gaze (Munson’s sign);
- Corneal hydrops, caused by an acute rupture in Descemet’s membrane which results in a sudden, abnormal accumulation of fluid in the corneal stroma. Corneal hydrops is estimated to occur in 5 to 15 per cent of keratoconus patients. The corneal oedema may take weeks to months to resolve, resulting in residual stromal scarring that is accompanied by a relative flattening of the corneal curvature; this can in some cases, allow contact lenses to be more readily fitted.
Clinical indicators that can assist with a diagnosis of keratoconus include:
- Subjective refraction: myopia with, or without, high astigmatism (typically oblique or against the rule) and near acuity better than expected from refraction and age, due to the multifocality of the cornea;
- Retro-illumination: scissoring reflex on retinoscopy and/or the ‘Charleux’ oil-droplet sign evident with ophthalmoscopy;
- Pachymetry: reduced central corneal thickness (< 450µm is suspicious of keratoconus, however due to significant variability of central pachymetry measures within the normal population it cannot be solely relied upon);
- Photokeratoscopy: distortion or steepening of keratometry mires centrally or inferiorly;
- Videokeratoscopy (corneal topography): is regarded as useful for both the detection and monitoring of keratoconus. This technique, which provides a graphical representation of the physical configuration of the cornea, can accurately reveal the shape, size and location of the cone (Figure 2).
While corneal topography is the most commonly used device to accurately detect and monitor keratoconus, the Pentacam instrument (Oculus, Wetzlar, Germany) is an alternative optical instrument for assessing the anterior ocular surface. The Pentacam takes multiple images of the cornea at different angles using a rotating camera. As such, it allows for the evaluation of disease severity and progression based upon changes in corneal volume and anterior chamber angle, depth and volume.
Classification of Keratoconus
Several classification systems for keratoconus have been proposed in the literature. These systems categorise the condition based upon different criteria, including corneal morphology, disease evolution, ocular signs and corneal indexes.
Morphological Classification
The classification of keratoconus using morphological criteria is based upon corneal topographical data; three major sub-groups of keratoconus have been described (Figure 2):
– Centred (nipple) cones are small (i.e. a cone diameter ≤ 5mm), round in shape and are positioned central or just inferior to the visual axis; the peripheral cornea remains relatively normal in curvature. Approximately 45 per cent of cones are reported to be of this morphology. Correction with rigid gas permeable lenses is relatively straightforward owing to the centralised location of the corneal steepening.
– Oval (sagging) cones are larger in size (i.e. a cone diameter >5mm), and are displaced inferonasally or inferotemporally, inducing high degrees of irregular astigmatism. It is recognised that approximately half of cones are of an oval morphology. Since contact lenses tend to naturally centre over the apex of the cone, centration and adequate pupillary coverage can be more difficult to achieve than for the centred cones.
Globus cones are the least common morphology, constituting less than five per cent of presentations. As the cone involves at least 75 per cent of the cornea, these are the most challenging cones to fit with contact lenses; typically, large intra-limbal or sclera lenses will be required.
Disease evolution
Disease evolution refers to the classification of keratoconus based upon the severity of the clinical signs.13
Stage one is defined as forme fruste or sub-clinical keratoconus and is characterised by normal slit lamp findings, best-corrected spectacle acuity of 6/6 (or better) combined with the presence of subtle corneal irregularity on corneal topography. Stage two represents early keratoconus, in which mild corneal thinning may be evident on biomicroscopy, but corneal scarring is absent. In Stage three (moderate keratoconus), slit lamp signs such as Vogt’s striae and Fleischer’s ring are more common. Best-corrected spectacle acuity is reduced to below 6/6, with irregular astigmatism between 2.00 and 8.00 dioptres. Stage four is the most severe form of keratoconus, with a maximal curvature value in excess of 55.00D, the presence of stromal corneal scarring, severe corneal ectasia and acuity below 6/7.5 even with contact lens correction.
Index-based systems
Several index-based systems have been described for the detection of keratoconus (Table 1). These systems have the advantage of being objective and are designed to be sensitive for detecting, early sub-clinical forms of the disease.
Corneal Collagen Cross-linking
Corneal collagen cross-linking (CXL) is the most recent clinical intervention that demonstrates promise for arresting the progression of keratoconus. The treatment was inspired by the German ophthalmologist Professor Theo Seilor in the 1990s, whom during a visit to his dentist noted the use of ultraviolet (UV) radiation to harden a synthetic filling and proposed that a similar process may have the potential to stiffen a weakened keratoconic cornea. The procedure aims to increase the biomechanical stability of the keratoconic cornea, thereby potentially slowing, or even halting progressive corneal ectasia and postponing, or even negating, the need for future corneal transplantation.
Mechanism of Action
In the normal cornea, covalent bonds (or cross-links) exist between collagen fibrils, imparting structural integrity and rigidity to the tissue. In keratoconus, a reduced number of cross-links between the collagen layers reduce the mechanical strength of the cornea by up to 30 per cent.19 CXL involves the use of the photosensitisation agent riboflavin (Vitamin B2) and ultra-violet A (UVA) irradiation, to induce photo-oxidative cross-linking of the collagen within the corneal stroma in vivo.
Riboflavin has an absorption peak for UVA at a wavelength of approximately 370nm. When the riboflavin-saturated cornea is exposed to radiation of this wavelength, the riboflavin molecules fluoresce, resulting in the generation of singlet oxygen and superoxide radicals;20 these reactive oxygen species lead to the formation of covalent bonds between collagen molecules. Studies indicate that CXL induces physical changes within the cornea that include: an increase in Young’s modulus,21 increased bending stiffness,21 larger collagen fibre diameter22 and an enhanced resistance to enzymatic degradation.23
Clinical Technique
The CXL treatment is conducted under sterile conditions in an operating theatre. Pre-operatively, the patient’s eye is anaesthetised with topical anaesthetic (e.g. Proxymetacaine hydrochloride 0.5 per cent drops) and pilocarpine is instilled to induce pupillary constriction in order to minimize UV exposure to the lens and retina. The central 7-9mm of corneal epithelium is then debrided to allow the diffusion of riboflavin into the corneal stroma.
A 0.1 per cent riboflavin solution (10mg riboflavin-5-phosphate in 10ml dextran 20 per cent solution) is applied to the eye approximately every five minutes, commencing five minutes prior to the first irradiation. The irradiation is performed from a one centimetre distance for 30 minutes using UVA at 370nm and an irradiance of 3mW/cm2. The required irradiance is carefully controlled in each patient directly before the treatment to avoid a potentially dangerous UVA overdose.
At the completion of the procedure, the patient is commenced upon topical, broad-spectrum antibiotics and a
bandage contact lens is applied to the eye. To minimise the ocular inflammatory response, topical corticosteroid eye drops are commonly used upon the third post-operative day, when the bandage contact lens is also removed.
Clinical Results
Cross-linking pre-clinical studies began in 1993 and included laboratory work and experimental trials. The first in vivo clinical study to be published on CXL was conducted by the German research group, Wollensak and colleagues in 2003.24
Since this pilot study, a growing number of papers have described the clinical efficacy of CXL for keratoconus (Table 2). Although differences exist in the study methodology, inclusion and exclusion criteria, treatment parameters and outcome measures, data from all of these studies consistently demonstrate varying degrees of improvement in visual acuity and a reduction in maximal keratometry values with CXL treatment.
The world’s first prospective, randomised, controlled clinical trial on CXL for keratoconus was conducted by the Centre for Eye Research Australia (CERA) at the Royal Victorian Eye and Ear Hospital, Melbourne, Victoria.
Commencing in 2006, this study involved the recruitment of one hundred patients who had demonstrated clinically-significant progression of their keratoconus over the prior six to twelve months. The preliminary findings of this study, published in 2008, indicated statistically significant differences between the CXL treatment and control (untreated) groups for changes to maximum (steepest) simulated keratometry values and best spectacle-corrected acuity.26 A larger, multi-centre treatment trial is currently being conducted in the USA; the findings of this important study are eagerly anticipated.
Authors | Index | Description of diagnostic index |
Rabinowitz/McDonnell14 |
K value I-S value |
Expression of central corneal steepening Inferior-superior asymmetry in keratometry power |
Maeda/Klyce15 | KPI | Keratoconus Prediction Index (derived from eight quantitative topographical indices) |
Smolek/Klyce16 | KSI | Keratoconus Severity Index (used to detect keratoconus, with level of severity assessed using an artificial intelligence system) |
Rabinowitz/Rasheed17 | KISA% | Derived from product of 4 indices: the K value, I-S value, the AST index (quantifies the degree of regular corneal astigmatism) and the skewed radial axis index (a measure of irregular astigmatism) |
McMahon et al18 | KSS | Keratoconus Severity Score (based upon slit-lamp findings, corneal topography, corneal power and higher-order first corneal surface wave-front root mean square error) |
Table 1: Summary of index-based symptoms for detecting keratoconus
Author(s) | Year | Country | Study Design | Number of eyes | Duration (months) |
Wollensak et al24 | 2003 | Germany | Case series | 23 | 3-47 |
Caporrosi et al25 | 2006 | Italy | Prospective, non-randomized | 10 | 6 |
Wittig-Silva et al26 | 2008 | Australia | Prospective, randomized, controlled | 66 | 3-12 |
Raiskup-Wolf et al27 | 2008 | Germany | Retrospective case series | 241 | 6-72 |
Hoyer et al28 | 2008 | Germany | Retrospective case series | 153 | >12 |
Jankov et al29 | 2008 | Serbia | Retrospective case series | 25 | 4-7 |
Vinciguerra et al30 | 2009 | Italy | Prospective, non-randomized | 28 | 12 |
Agrawal31 | 2009 | India | Retrospective case series | 37 | >12 |
Grewal et al.32 | 2009 | India | Prospective, non-randomised | 102 | 12 |
Coskunseven et al.33 | 2009 | Turkey | Prospective, non-randomised | 19 | 5-12 |
Fourni et al.34 | 2009 | France | Uncontrolled, prospective | 20 | 3-18 |
Koller et al.35 | 2009 | Switzerland | Controlled, non-randomised | 21 | 12 |
Saffarian et al.36 | 2010 | Iran | Retrospective, non-randomised | 92 |
12 |
Table 2: Summary of published clinical studies on corneal collagen cross-linking and keratoconus
Patient Selection
With the advent of CXL, it is now more important than ever that optometric management of keratoconus involves the careful documentation and monitoring of younger patients (less than 35 years or age) with established disease; these patients are at the highest risk of progression and therefore may benefit most from CXL. Moreover, as primary eye care providers, optometrists play a pivotal role in identifying patients with the early signs and symptoms of keratoconus, including sub-clinical manifestations, through appropriate clinical examination and corneal topography. Within practices that are not equipped with a corneal topographer, referral for corneal mapping is warranted in patients with a family history of the condition in order to identify keratoconus suspects at an early age. Younger patients may require reviews every few months if there is a suspicion of progression or incipient keratoconus.
It is recommended that patients should be referred for ophthalmologic assessment for CXL if there is demonstrable progression in their keratoconus, as evident clinically through one or more of: topographical/ keratometric changes (≥ 1.0 dioptral change in maximum keratometry value), repeated changes of contact lens base curve (if topography is unavailable) and/ or deteriorating vision.
Treatment is currently contraindicated in the following scenarios:
- Corneal thickness less than 400μm
- Prior ocular herpetic infection
- Concurrent corneal infection
- Severe corneal scarring or opacification
- A history of poor wound healing
- Severe ocular surface disease (excluding dry eye)
- Auto-immune disorders
- Pregnancy
Careful ophthalmologic management of- CXL patients is required over the first few weeks post-treatment. A return to contact lens wear is recommended approximately four to six weeks after CXL, at which time the patient is generally returned to optometric care for ongoing monitoring at 6-12 monthly intervals.
Safety
CXL involves a highly localised photopolymerisation reaction that is generated in vivo, within the corneal stroma. Its application therefore requires consideration with regard to the potential effect on surrounding ocular structures, in particular the corneal endothelium, lens and retina.
Guidelines regarding optimal practice for CXL have recently been published.37 The safety of the procedure is primarily dependent upon the appropriate delivery of the irradiation and ensuring a sufficient stromal riboflavin concentration. Of particular concern is the potential for oxygen radical damage to the corneal endothelium. Control of the depth of the reaction is primarily governed by the pre-operative corneal thickness. It is accepted that in order to minimise risk of permanent endothelial damage, a minimum corneal thickness of 400μm is necessary following the debridement of the epithelium.38 The use of hypo-osmolar solutions of riboflavin can temporarily enhance corneal thickness in patients who do not meet the minimum thickness requirement39, however further research is necessary to determine if this procedure provides adequate endothelial protection.
Although the theoretical risk exists, there is currently no evidence for CXL causing a reduction in corneal endothelial cell density. Published studies that have examined the endothelium using specular microscopy, one-year after CXL treatment have reported no quantitative change to endothelial cell numbers.24,26,30
Side-effects and Risks
Mild anterior- and mid-stromal corneal haze is the most common side-effect of CXL; the effect has been documented to persist for up to twelve months but has been noted to not significantly affect visual acuity.27 Complications related to the debridement of the corneal epithelium include sterile corneal infiltrates40 and infectious keratitis secondary to bacteria41, acanthamoeba42 and Herpes simplex virus43; all of these complications were reported within the first week post-treatment.
While most of the published literature describes relative improvements in visual acuity (VA) post-CXL, a loss of two lines or more best-corrected VA was described in 2.9 per cent of treated eyes, one year post-operatively;35 this should however be taken in context with the overall results of this study which reported progressive corneal ectasia (treatment failure) in 7.6 per cent of eyes.
Other Applications
CXL may also have a therapeutic role in the treatment of other corneal conditions, including pellucid marginal degeneration, post-LASIK keratoecastia, bullous keratopathy and microbial keratitis; definitive published evidence to support these applications is still required.38
Until recently, the management of keratoconus has been limited to supportive optical measures, primarily spectacles and contact lenses. The current scientific literature demonstrates that CXL has a significant arresting effect on the progression of keratoconus. Although further research into the long-term effects of CXL is still warranted, this exciting new treatment demonstrates the potential to transform the lives of patients with keratoconus, reducing their risk of progressive ectasia and visual impairment.
Dr. Laura Downie, BOptom, PhD(Melb), PGCertOcTher, DipMus(Prac), AMusA is an optometrist who specialises in contact lenses. She has been published in scientific journals and is a clinical instructor to undergraduate optometry students.