Collagen cross-linking: a new treatment paradigm in corneal disease – a review

Grant R Snibson FRANZCO
Centre for Eye Research Australia, Department of Ophthalmology, University of Melbourne, Royal Victorian Eye & Ear Hospital, East Melbourne, Victoria, Australia
Correspondence: Associate Professor Grant R Snibson, Level 2, 232 Victoria Parade, East Melbourne, Vic. 3002, Australia. Email:


The last 2 years has seen a marked increase in the prominence of corneal collagen cross-linking as a treatment strategy for progressive keratoconus. This interest has arisen from a body of laboratory evidence documenting the biomechanical and cellular changes induced by cross-linking. The findings of this research provide a plausible rationale for its use in keratoconus to retard the progression of this common disease. The rapidly growing number of clinical reports suggests, not only a consistent stabilizing effect of cross-linking, but that a variable improvement in corneal shape and visual function may also occur in some patients. However, the marked variation in the clinical course of keratoconus, together with the challenges of accurately evaluating refractive error, visual acuity and even corneal shape in this condition, demands further evidence from randomized controlled clinical trials. The aim of this review is to summarize the theoretical basis and risks of corneal collagen cross-linking, along with the available evidence for its use in keratoconus and other corneal disease states.

Key words: collagen, corneal dystrophies, corneal topography, cross-linking, keratoconus progression, riboflavin.


Seldom in the professional life of an ophthalmologist is a novel treatment developed that, for the first time, is capable of profoundly impacting on the natural history of a chronic and progressive disease. As has been the case with vascular endothelial growth factor (VEGF) inhibitors in macular disease, the advent of corneal collagen cross-linking (CXL) brings the potential to transform our approach to the management of keratoconus.
Until now, individuals with progressive forms of keratoconus could only look forward to increasing visual incapacity, corneal transplantation or, at best, a lifetime of rigid contact lens wear. It is not surprising therefore that such patients – many of whom are young and well informed – are actively seeking out this treatment that is being rapidly adopted by corneal subspecialists and general ophthalmologists throughout the world. Although most research to date has related to the treatment of keratoconus, a role for CXL has also been suggested for other forms of corneal ectasia and a number of unrelated corneal conditions (Table 1).
But can optimistic claims such as these be justified by the laboratory and clinical data at hand? Although enthusiasm for new and promising treatments is understandable, there is a risk that the pace and diversity of the clinical application of CXL will outstrip the growth of the limited evidence base that supports it. This review is intended to summarize the rationale, evolution, application and current status of this procedure in the light of published research.

Although cross-linking has been used in all these circumstances, other than for progressive keratoconus, there is currently little published evidence to support these applications.


Intermolecular cross-linking to enhance the rigidity of materials is not a new concept, having widespread application in synthetic polymer chemistry, the manufacture of plastics and in other industries. In fact, it was a visit to a dentist who used UV irradiation to harden a synthetic filling material that led to the serendipitous inspiration to apply a similar process to stiffen the human cornea (T Seiler, pers. comm., 2009).
In the normal cornea, covalently bonded molecular bridges or cross-links exist between adjacent tropocollagen helices and between microfibrils and fibrils at intervals along their length.1,2 The alternative mechanisms by which these cross-links may be formed have been recently reviewed by Ashwin and McDonnell.3 With normal aging, the extent of this natural cross-linking increases4,5 with consequential change in the biomechanical properties of the tissue and a measurable increase in Young’s modulus.6 This phenomenon may, to some degree, explain both the flexibility of the infant cornea relative to adult tissue and the slowing of the rate of progression of keratoconus that is observed with increasing age.7 Crosslinking is also enhanced by increased glycation of corneal collagen8 that is consistent with the observation that diabetes may protect against the development of keratoconus.9 Although a paucity of natural cross-links may not be the major factor in the aetiology of keratoconus,10 the mechanical strength of the cornea in this condition is reduced.11 By actively increasing the degree of the bonding between collagen molecules, therapeutic cross-linking could reasonably be expected to enhance corneal rigidity and potentially slow, or even halt, the progression of this disease.
Alternative methods of enhancing the crosslinking of corneal collagen have been entertained, one of the more effective being the use of glutaraldehyde. 12 However, the toxicity and the inability to control the depth and extent of diffusion of glutaraldehyde make the use of this agent impractical in the clinical context. Conversely, the ability to target and titrate the effect is a major advantage of using riboflavin and UV-A to achieve cross-linking of corneal collagen in vivo.
The photosensitizer, riboflavin (vitamin B2), has an absorption peak for UV-A at a wavelength of approximately 370 microns. When the riboflavin saturated cornea is exposed to UV irradiation of this wavelength, the riboflavin molecule fluoresces (Fig. 1) and is excited into a triplet state with subsequent generation of singlet oxygen and superoxide radicals.13 These reactive oxygen species then lead to the formation of covalent bonds between collagen molecules by oxidative desamination, the process being dependent on the presence of oxygen and enhanced by deuterium oxide.14 This photodynamic reaction differs from the mechanism of formation of the cross-links that accrue with increasing age and diabetes (Maillard reaction).15

Figure 1. Fluorescence of the riboflavin and riboflavin saturated corneal stroma during UV-A irradiation.


Spoerl and Huhle et al. were the first to investigate the use of riboflavin and UV-A irradiation to achieve cross-linking of corneal collagen.16 Using this method, Wollensak et al.17 and Kohlhaas et al.18 were able to demonstrate a marked positive effect of CXL on the biomechanical properties of both porcine and human corneal tissue. CXL was shown to increase the rigidity of human corneal tissue (as expressed by Young’s modulus at 6% strain) by a factor of 4.5.17 In a similar experiment, the increase in Young’s modulus (67% at 5% strain) was shown to be confined to the anterior 200 microns of stroma where most of the UV-A radiation is absorbed.18 Contemporary techniques such as supersonic shear imaging have confirmed these findings by demonstrating a 460% increase in Young’s modulus following CXL of porcine corneal tissue.19
Several other animal studies have followed, predominantly from Dresden, examining the various tissue effects of CXL. Keratocyte toxicity studies have shown that (at least in the rabbit model) saturation of the corneal stroma using 0.1% riboflavin in a dextran solution and 30 min of UV-A irradiation with a wavelength of 370 nm and a surface irradiance of 3 mW/cm2 may limit keratocyte death to a depth of approximately 300 microns.20 In this study, there was no observable toxic effect of CXL at a depth beyond this level. The inference that there will therefore be insufficient UV energy and oxygen radical production at the level of the corneal endothelium to cause damage to this cell layer is supported by theoretical constructs and some experimental evidence. 21,22
However, the long-term impact of CXL on the structure and function of the human corneal endothelium has not yet been determined.
Within 1 h of UV-A-riboflavin treatment of eye bank corneae, changes can already be observed in stromal keratocytes with cell shrinkage, chromatin condensation and apoptotic bodies.23 In the rabbit model, this apoptosis is complete by day 3 and early keratocyte repopulation can be observed by day 7. This repopulation commences in the posterior stroma, with activated keratocytes (migratory fibroblasts) visible in the adjacent untreated stroma. By the sixth week, the process appears complete.24
In addition to the tissue stiffening effect of CXL, an alteration in the swelling properties of the porcine cornea has also been described by Wollensak et al.25 Consistent with this observation, compaction of the anterior corneal stroma has been demonstrated using immunofluoresence and confocal microscopy in the same animal model, with this effect being dependent on epithelial removal.26 CXL has been shown to increase stromal shrinking temperatures27 and to increase resistance to enzymatic digestion.28 CXL of corneal collagen may also increase the impedance of the corneal stroma to the diffusion of some solutes, as has been demonstrated for fluorescein.29 The effect (if any) of CXL on the corneal penetration of topically applied therapeutic agents has not been studied.
Other than for these important contributions, relatively little laboratory and preclinical animal research has been published (Table 2).

Table 2. Physicochemical changes in the corneal stroma induced by collagen cross-linking



Until recently, the number of studies providing clinical data to support the efficacy and safety of the CXL procedure in keratoconus has also been small, with only two publications prior to 2008. The initial clinical experience in Dresden was reported by Wollensak et al. in 2003 and a report of a smaller Italian study of 10 patients followed in 2006.33,35 In 2008, Raiskup-Wolf et al. described what remains the largest published series comprising 241 eyes followed in Dresden for up to 6 years after CXL.34 This uncontrolled, retrospective study confirmed earlier findings with statistically significant improvements in astigmatism, best-corrected visual acuity (BCVA) and maximum simulated keratometry values (Kmax) at 12 months. Flattening was observed in 54% of eyes with a mean change in Kmax of -1.91 D (P < 0.01). The effects of CXL were maintained over the duration of follow up with progression of the disease documented in only two patients.
Subsequent reports from several other centres have described similar results (Table 3). A recent publication from Vinciguerra et al. describes the outcome of CXL in 28 eyes with progressive keratoconus, using the less affected fellow eye used as the control. This study demonstrated a statistically significant improvement in both uncorrected and BCVA at 12 months along with a reduction in the steepest simulated keratometry value of as much as 6.16 D (P < 0.0011). Improvements were also observed in topographic keratoconus indices and higher order aberrations.43 The benefits attributed to CXL were maintained 2 years after treatment.44
Although the study was not randomized, it is noteworthy that the keratoconus in the untreated contralateral eyes progressed over the same period. Although the outcomes of the studies published to date have been consistently positive, conclusive evidence that CXL can retard the progression of keratoconus can only come from the critical analysis of multiple randomized controlled trials. The design of such trials is critical3 given the variable natural history of keratoconus and the poor reproducibility of most measures of visual acuity, refractive error and corneal shape that characterizes this condition. An effect is more likely to be demonstrated if trial recruitment is restricted to patients with progressive disease. At this time, no meta-analysis is possible as only the preliminary findings of a single randomized trial have been published. In this Melbourne study, statistically significant differences were observed between the treatment and control groups in terms of both changes in maximum (steepest) simulated keratometry values and best spectacle-corrected visual acuity at 3, 6 and 12 months following CXL (Fig. 2).36 A larger, multicentre, randomized trial is also currently in progress in the USA and the findings of that important study are also eagerly awaited.

B[S]CVA = best [spectacle]-corrected visual acuity; Kmax, maximum curvature or steepest simulated keratometry value derived from computerized videokeratography; n, eyes subject to analysis; ns, not statistically significant; SEQ, spherical equivalent; UCVA, uncorrected visual acuity.

Table 3. Published clinical studies of corneal collagen cross-linking in the treatment of keratoconus



The outcome measure common to most clinical studies has been the change in corneal curvature as determined by computerized videokeratography – an increase in corneal curvature being a marker of progression of the keratoconus and a decrease being interpreted as an additional therapeutic effect of CXL. A halting of progression of ectasia following CXL has been a consistent finding in all published studies with very few subjects requiring a second treatment.34 However, given the variability in the rate of progression of keratoconus,7,46 this apparent stabilization should be interpreted with caution in the absence of a contemporaneous control group. A history of progression prior to treatment based on changes in refraction, visual acuity or the need for contact lens refitting is not as reliable or as objective as serial computerized videokeratography. In practice, reliable topographic evidence of progression is commonly unavailable. If the adopted inclusion criteria inadvertently permit recruitment of eyes that are, in reality, stable, a lack of progression following CXL may be falsely attributed to the treatment. On the other hand, a measurable flattening of the cornea – observed in many but not all subjects – is more convincing evidence of a structural effect of CXL and a phenomenon not seen in the natural history of the disease.
Flattening of the cornea, with or without a reduction in the magnitude of the astigmatism, is usually accompanied by an improvement in uncorrected visual acuity. Not infrequently, however, improvements in visual acuity will be observed without a concomitant change in simulated keratometry values. In such cases, the improvement has been construed to occur as a result of a reduction in the irregular component of the astigmatism. A reduction in some higher order aberrations, particularly coma, has been demonstrated after CXL suggesting improved symmetry and homogeneity of the anterior (and possibly posterior) corneal surface topography.33,42,43
Given the stiffening effect that CXL has on the cornea, this treatment might be expected to lead to an overestimation of intraocular pressure when measured by applanation tonometry and there is some laboratory evidence to support this assumption.47 A small (2 mm Hg) increase in intraocular pressure measurements has been observed clinically in one series41 but not in others.34,35,40,48
Transient mild corneal oedema occurs universally during the early postoperative period and begins to resolve after closure of the epithelial defect. The development of mild anterior and mid-stromal haze is also expected. Haze can persist for several months but is rarely visually significant.48 The posterior limit of the corneal haze is often visible on slit-lamp examination as a line or layer of denser haze in the deeper stroma. This demarcation line was described by Seiler and Hafezi who postulate that this represents the transition zone between the treated (crosslinked) stroma and the untreated posterior layers.49

Figure 2. Changes in maximum simulated keratometry values on computerized videokeratography at 12 months compared with preoperative baseline in the first 19 eyes to complete 12-month follow up in a randomized controlled trial of corneal collagen cross-linking in progressive keratoconus.36 (a) Corneal steepening occurred in all 10 control (untreated) eyes at 12 months with a mean increase in Kmax of 1.28 D (P < 0.001). (b) Corneal flattening occurred in all nine treated (cross-linked) eyes at 12 months with a mean reduction in Kmax of 1.45 D (P = 0.002).

Confocal microscopy

The confocal microscopic findings in keratoconus have been well documented,50,51 and a small but growing body of work has described the changes that occur following CXL. The presence of stromal oedema, a reduction in keratocyte density, keratocyte apoptosis and a lack of nerve fibres have been observed in the anterior stroma 1 month after CXL.52 In the same study,Mazzotta et al. noted a reduction in oedema, increased stromal density and progressive repopulation of the stroma with activated keratocytes 3–6 months after treatment. The deep stroma and endothelial morphology appeared unaffected. Similar findings were reported by Kymionis et al. who also noted keratocyte repopulation of the anterior and mid-stroma within 6 months.53 However, changes in the structure and cellularity of the cornea can be observed by confocal microscopy as long as 36 months after CXL.54
Using the Heidelberg Retinal Tomography (HRT) II confocal microscope, Mazzotta et al. in a later publication, described needle-shaped hyper-reflective bands or bridges in the anterior mid-stroma 2 years after CXL, which they interpret as new structured collagen.55 These are not dissimilar to the intrastromal striate reflections that we have observed with the Nidek Confoscan 4 confocal microscope as early as 1 month after CXL (Fig. 3).36 The histological and ultrastructural changes correlating with these striate reflections and their time–course remain to be determined.

Figure 3. Intrastromal striate reflections seen on confocal microscopy at a depth of approximately 280 microns 3 months after corneal collagen cross-linking. Arrows show elongated keratocyte nuclei.


Although the apparent simplicity of the CXL procedure is beguiling, the potential for adverse outcomes should not be underestimated. Although the corneal endothelium is relatively resistant to ultraviolet irradiation, inappropriate UV delivery or dosimetry and insufficient stromal riboflavin concentrations may lead to unacceptable irradiance of endothelium and intraocular structures.
The selection of appropriate treatment zone diameters and/or masking to protect the limbus may also be important to protect limbal stem cells from the toxic effects of oxygen radicals generated by the procedure. It is these reactive oxygen species that also pose a threat to the corneal endothelium. The depth at which oxygen radicals are created within the stroma will depend on the relationship between the UV energy applied and the tissue riboflavin concentration achieved, together with any masking effect of residual epithelium. For these reasons, modifications to the treatment protocol should be made with extreme care. To ensure uniformity of cross-linking, it is also important that the UV-A source utilized is capable of delivering uniform irradiance across the treated surface of the cornea.
If the proximity of the deeper limit of the photochemical reaction to the corneal endothelium is to be controlled, knowledge of the minimum corneal thickness is essential. This is best assessed preoperatively by elevation mapping using computerized videokeratography as it is difficult to accurately determine the thinnest point of the cornea by ultrasound pachymetry alone. In order to minimize the risk to the endothelium, a minimum corneal thickness of at least 400 microns after removal of the epithelium has been adopted by many investigators. This minimum thickness should then be maintained throughout the procedure. As thinning of the stroma is common after removal of the epithelium, intraoperative ultrasonic pachymetry immediately prior to commencement of irradiation (and at regular intervals thereafter) is vital. Some researchers have adopted as a threshold for treatment a minimum preoperative corneal thickness of 400 microns inclusive of the epithelium, but this further reduces the safety margin and demands more rigorous intraoperative monitoring of corneal pachymetry.
Unfortunately, many patients with progressive keratoconus have a minimum corneal thickness less than this threshold and would reasonably be excluded from treatment. To address this, hypoosmolar solutions of riboflavin have been used to swell the corneal stroma to greater than 400 microns and thereby allow the treatment of thinner corneae. 56 It is difficult to be certain, however, that the over-hydrated corneal stroma will respond similarly to the CXL procedure and that this modification to the technique affords adequate endothelial protection.
The potential for corneal collagen CXL to result in damage to the corneal endothelium is of concern and warrants close clinical scrutiny. At this time, only four publications have included quantitative specular microscopy data35,36,43,44 and in only 41 eyes have endothelial cell counts been reported beyond 12 months.35,44 Haze may limit cell border recognition and impede quantitative specular microscopy in the early months, but significant changes in endothelial cell density 1 year after CXL have not been reported. As yet, there are no published reports of irreversible endothelial failure complicating collagen CXL, but the potential for the underreporting of this complication should be acknowledged.
In addition to these specific risks to adjacent tissues, there is potential for all of the complications usually associated with the creation and healing of an epithelial defect. Sterile corneal infiltrates may occur during the healing phase57 but usually respond to removal of the bandage contact lens and the introduction of topical corticosteroids. Occasionally, these reactions are more severe and may lead to scarring and loss of vision.58
Infectious keratitis attributed to bacteria,59–61 acanthamoeba62 and Herpes simplex virus63 have each been reported following CXL, and such infections clearly have the potential to compromise corrected visual acuity. This risk is likely to be increased by the use (and misuse) of bandage soft contact lenses for control of postoperative discomfort.59
Although most reports detail improvements in BCVA, Koller et al. observed a loss of two or more lines of BCVA in 2.9% of eyes 1 year after CXL. A patient age in excess of 35 years and good preoperative BCVA (6/7.5 or better) were each identified as risk factors for visual loss. This same study also documented progression of ectasia (treatment failure) in 7.6%.48


Cross-linking technique

Although minor variations are described, most published CXL techniques are based broadly on the methodology developed in Dresden.22 Preoperatively, pilocarpine is used by some to constrict the pupil and limit the exposure of the lens and posterior structures to any transmitted UV-A irradiation. Under topical anaesthesia, the central 7.0–9.0 mm of corneal epithelium is removed by mechanical debridement, with or without the assistance of alcohol. Riboflavin 0.1% solution (usually containing 20% dextran) is then applied topically every 2–3 min for 30 min. After this time, fluorescence can be observed in the anterior chamber on slit-lamp examination. After confirming that the stromal thickness at the thinnest point remains at least 400 microns by ultrasonic pachymetry, UV-A irradiation is commenced using a wavelength of 370 nm, at surface irradiance of 3.0 mW/cm2 for 30 min (surface dose 5.4 J/cm2). Throughout the irradiation phase, riboflavin solution is applied every 2–3 min, ensuring that the stromal surface is kept moist and that the stromal thickness remains above 400 microns. The continuing application of riboflavin also serves to maintain the optimal tissue concentration of riboflavin and ensures that the bleaching effect of the UV-A irradiation does not lead to reduced UV-A absorption within the stroma.64 Topical antibiotics are commenced at the completion of the procedure and a bandage contact lens is inserted. Corticosteroid eye drops are commonly used to minimize the inflammatory response and the contact lens is removed upon epithelial healing – usually on the third postoperative day.

Protocol variations

Although the Dresden treatment protocol has been widely adopted internationally and appears to be safe with appropriate case selection, alternative protocols may be equally effective. Certainly, the ability to shorten treatment time would be a major advantage. This may be achievable with alternative photoactive cross-linking agents that are effective with much abbreviated UV-A exposures. Surface wave elastometry in the in vitro porcine cornea suggests that ‘flash-linking’ employing a different photosensitizer has similar efficacy in stiffening the cornea as does conventional CXL,65 but clinical data are as yet unavailable. Alternatively, it may be possible to induce a similar cross-linking response using higher UV-A energy levels with shorter or fractionated exposure.With any significant change in the treatment protocol, safety studies should be repeated to confirm the depth of the resulting keratocyte apoptosis and ensure that endothelial cells and intraocular structures are adequately protected.
The ability to achieve predictable cross-linking without epithelial removal would be another desirable modification in order to lessen discomfort and shorten recovery time. However, complete removal of the epithelium would seem necessary to permit adequate and uniform saturation of the stroma with riboflavin (molecular weight: 376.36).26,66–69 Some practitioners have sought to minimize the diffusion barrier effect of the epithelium by the use of local anaesthetic preparations containing benzalkonium chloride or by partial epithelial ablation with the Excimer laser. The latter method has been shown, in one study, to increase pain and slow stromal saturation with riboflavin when compared with full thickness epithelial removal.70 Although CXL treatments are currently being performed in some centres without complete (full thickness) epithelial removal, there is currently little published information that would attest to the efficacy of this approach.
Unfortunately, it is not yet possible to directly observe or measure the increase in cross-linking that is induced by this application of riboflavin and UV-A, or to accurately measure the depth of the changes occurring in the cornea in vivo. Although the observations that have been made using in vivo confocal microscopy may to some extent act as surrogate evidence of cross-linking, future protocol development will rely heavily on animal modelling and cautious clinical validation.

Figure 4. Patient view of the seven light emitting diodes of a UV-A delivery device (UV-X, Peschkemed Meditrade).

UV-A delivery

Most of the laboratory studies of CXL reported to date have been performed using unsophisticated UV-A sources fabricated by (or for) the investigators.17 Many of the clinical procedures reported in the literature were also performed using non-proprietary devices. The use of direct LED sources for the delivery of UV-A leads to the potential for ‘hot-spots’ at the corneal surface and variations in intensity with changes in working distance. UV-A sources for the clinical application of CXL should be based on the principle of Koehler illumination with optical homogenization to ensure uniform and consistent beam intensity across varying working distances.64,71
Over recent years, a number of UV-A delivery systems have been developed specifically for this purpose and are now available commercially. Although a comparison of the commercially available UV-A sources for performing the CXL procedure is beyond the scope of this review, it is noteworthy that there are significant differences between these devices. All utilize light emitting diodes to generate a UV-A output of a wavelength of 360–380 nm, but vary in the number of diodes used (5–25), focusing systems, working distance, beam diameter and beam uniformity, together with the extent to which the operator is able to alter the delivery parameters (Fig. 4). These differences may not have a major impact on clinical outcomes, but should be borne in mind when evaluating the published results of CXL.


Riboflavin has an essential role as a photosensitizer, absorbing the energy of the UV-A irradiation and generating the reactive oxygen species that catalyse formation of covalent cross-links between adjacent collagen molecules. In order for cross-linking to occur to the desired depth and to limit the UV-A irradiance of deeper tissues by riboflavin absorption, sufficient time must be allowed for riboflavin to saturate the stroma. To achieve a minimum safe concentration throughout the stroma is believed to take approximately 30 min with regular topical application following epithelial removal.22 Inadequate riboflavin concentrations will fail to generate sufficient crosslinks and will expose the endothelium and intraocular structures to higher levels of UV-A irradiation.
The riboflavin solutions currently available commercially have a riboflavin concentration of 0.1%. Most also contain a high molecular weight dextran and are formulated at an osmolarity that is intended to prevent excessive swelling of the corneal stroma. However, thinning of the corneal stroma during the saturation and/or treatment phases of the CXL procedure is not uncommon and may necessitate the use of hypoosmolar riboflavin solutions. The use of sterile water or diluted saline solutions to swell the cornea during treatment may cause unpredictable dilution of the riboflavin within the stroma and should be avoided.



The available evidence at this time would suggest that CXL does have a role in the management of keratoconus which can be shown to be progressing, particularly in patients with less advanced disease and relatively preserved corrected visual acuity. Although the risks of CXL have not yet been fully quantified, the potential for some risk would seem justified in the context of progressive disease that is otherwise likely to result in further impairment of corrected vision or even corneal transplantation. If progression can be reliably documented, then early intervention is likely to be most beneficial, arguably even before the onset of visual impairment or the development of obvious clinical signs. It follows then, that serial computerized videokeratography will come to form an important part of the management of patients with newly diagnosed keratoconus.
Although flattening of the corneal curvature and improved unaided and spectacle-corrected visual acuity is often observed, such improvement is unpredictable and may not impact appreciably on binocular visual function and quality of life. The possibility of such improvement may not, in its own right justify the procedure in keratoconus that is not progressing. Further accretion of data will ultimately answer this question.
In patients who undergo implantation of intrastromal corneal ring segments for more advanced forms of keratoconus, the addition of collagen crosslinking may enhance the flattening effect of the implant.72,73 It may also lessen the corneal steepening that can occur in the years following ring segment implantation.74 Consensus has not yet been achieved as to the optimal timing of CXL in relation to the insertion of the ring segments or as to the impact of CXL on current ring selection nomograms.
Not infrequently, patients will present with more severe disease characterized by marked thinning of the cornea and significant apical scarring. Such patients would appropriately be excluded from conventional CXL of the central cornea. However, it is conceivable that cross-linking the peripheral cornea prior to penetrating keratoplasty in this setting may stabilize the disease in the peripheral host tissue, thereby potentially reducing post-keratoplasty astigmatism and the need for repeat transplantation.

Other ectatic diseases

Other ectatic conditions of the peripheral cornea, such as pellucid marginal degeneration, may also benefit from CXL75 provided that any risk to the limbal stem cells is adequately addressed. Lateral diffusion of the singlet oxygen and superoxide radicals is limited by their short lifespan,76 but the limbal basal epithelial layer is presumed to be susceptible to the cytotoxic effects of locally created oxygen radicals generated by direct irradiation of the limbus. The extent to which limbal stem cells are vulnerable and capable of being protected by the masking of the limbus from UV-A irradiation is yet to be determined experimentally.
Iatrogenic keratectasia is a feared complication of corneal refractive surgery (laser in situ keratomileusis) which, in many respects, mimics the behaviour of progressive keratoconus and can result in the need for keratoplasty. Early case reports of the use of CXL in the management of this condition, with and without insertion of intrastromal corneal ring segments, have been encouraging with stabilization being observed in the short to medium term.77–79 The confocal microscopic findings after CXL are similar in the two conditions.53
The detection of topographic abnormalities resembling those seen in forme-fruste keratoconus in patients requesting refractive surgery is regarded as an important risk factor for the development iatrogenic keratectasia.80 Prophylactic CXL has been advocated in such patents prior to (or at the time of) photorefractive keratectomy (PRK) in an attempt to prevent this complication.81 Such a thesis ignores, however, the impact that CXL itself may have on both corneal curvature and refractive error. CXL could also conceivably alter both the stromal ablation rate and the biomechanical response of the cornea to the ablation and further reduce the refractive predictability of PRK. The use of CXL as means of extending the range of patients eligible for laser vision correction would be difficult to justify on the basis of current evidence.

Bullous keratopathy

Corneal collagen cross-linking has also been suggested as a treatment for corneal oedema. This concept is supported by changes in the hydration behaviour of the porcine cornea after CXL25 and the observation that stromal compaction follows CXL in a similar experimental model.26 Ehlers and Hjortdal reported a reduction in corneal thickness in 10 of 11 eyes treated with CXL with the majority experiencing some improvement in vision.82 Thinning was also observed by Wollensak et al. in three eyes with corneal oedema.83 In order to achieve more effective cross-linking of the deeper stroma in the thickened cornea, the staged intrastromal delivery of riboflavin and subsequent UV-A irradiation has been suggested and demonstrated in eye bank eyes and one patient.84
Further evaluation of CXL in this context is clearly necessary. However, this putative application for CXL is attractive in that it offers the potential to reduce the need for corneal transplantation in a condition other than keratoconus. It may also offer another means of controlling pain in patients with bullous keratopathy who are either unsuitable for or awaiting keratoplasty.

Microbial keratitis

The antimicrobial effects of the photoactivation of riboflavin may also be harnessed to treat infections of the cornea. Riboflavin has a modest affinity for nucleic acid and its absorption of UV-A leads to the oxidation of guanine bases, thus preventing the replication of the viral and bacterial genome.85 This effect is synergistic with any direct antimicrobial effect of UV-A irradiation itself and with any damage to microbial cell membranes and DNA caused by oxygen radicals. Currently, this phenomenon is being utilized to inactivate a variety of pathogens in blood products prior to transfusion.86,87 The antimicrobial efficacy of the combination of riboflavin and UV-A against a range of common bacterial pathogens causing infectious keratitis has now been well demonstrated in vitro.88–90
However, convincing clinical evidence for the effectiveness of CXL in treating corneal infections is currently lacking. Recent case reports describe successful outcomes of CXL in one case of corneal infection with Escherichia coli91 and in another keratitis caused by an unknown pathogen.92 We have had similar anecdotal success in the treatment of recalcitrant fungal keratitis. CXL has also been employed with positive outcomes in five eyes with refractory infection and stromal ulceration.93 Although such reports are encouraging, the contribution of CXL to observed clinical improvement in the context of simultaneous or recent intensive antimicrobial therapy is open to challenge.
Should CXL be proven capable of sterilizing the cornea to a diameter of at least 8 mm and to a depth that includes the anterior and mid-stroma, it follows that many corneal infections may be controlled with a single treatment. Given the potential benefits of this approach in recalcitrant infections with fungi, acanthamoeba and antibiotic resistant bacteria, further study is clearly warranted.

Stromal ulceration

Experimental evidence that CXL increases the resistance of porcine corneal stroma to digestion by collagenase and other proteolytic enzymes28 has led to the suggestion that CXL may have a role in slowing or preventing stromal ulceration resulting from infective, traumatic or immune-mediated corneal disease.94 However, other than for a single report of cases with concomitant infection,93 there is no published clinical evidence to support this inference. Unlike EDTA and other anti-collagenases that rely on the chelation of metal cofactors, the resistance to collagenases afforded by CXL is thought to be due to changes in the tertiary structure of the collagen fibrils that impede access of proteolytic enzymes to their specific cleavage sites.95

Donor cornea modification

As CXL is capable of depleting the stroma of keratocytes and other antigen-presenting cells, it has been suggested that pretreatment of donor corneal tissue may prevent or reduce the severity of allograft reactions following penetrating keratoplasty and thereby prolong graft survival.96


The advent of corneal collagen CXL is one of the more promising developments of this decade in the management of keratoconus. It appears, on the basis of information currently available, to have the potential to reduce the morbidity of progressive forms of this disease and may ultimately reduce the need for corneal transplantation. There is also a theoretical basis for the use of collagen CXL in several other corneal disease states.
Until recently, enthusiasm for its application has been tempered by the existence of only limited laboratory and clinical data supporting its efficacy, particularly in relation to conditions other than keratoconus. For the most part, the clinical evidence in support of CXL has been limited to small uncontrolled retrospective series with relatively short follow up. However, the last 2 years have seen rapid growth in interest in this treatment worldwide with clinical outcomes now reported from at least eight countries – the findings of all studies being remarkably consistent. The publication of the results of the randomized controlled trials currently in progress will enable a more confident assessment of the efficacy of this procedure.
The safety of corneal collagen CXL – particularly with regard to the corneal endothelium – will also take some time to fully ascertain and long-term observations will be required. In the meantime, any variations in the treatment protocol, riboflavin solution formulation and UV-A delivery should be informed and made with care.


The author would like to acknowledge the assistance of Dr Christine Wittig-Silva of the Centre for Eye Research Australia, who provided the photographs and confocal microscopy image. I would also like to express my appreciation to both Dr Wittig-Silva and Mr Kent Snibson for their thoughtful critique of the manuscript.



1. Sung HW, Chang WH, Ma CY, Lee MH. Crosslinking of biological tissues using genipin and/or carbodiimide. J Biomed Mater Res A 2003; 64: 427–38.

2. Robins SP. Biochemistry and functional significance of collagen cross-linking. Biochem Soc Trans 2007; 35: 849– 52.

3. Ashwin PT, McDonnell PJ. Collagen cross-linkage: a comprehensive review and directions for future research. Br J Ophthalmol 2009; 9: 9.

4. Malik NS, Moss SJ, Ahmed N, Furth AJ, Wall RS, Meek KM. Ageing of the human corneal stroma: structural and biochemical changes. Biochim Biophys Acta 1992; 1138: 222–8.

5. Daxer A, Misof K, Grabner B, Ettl A, Fratzl P. Collagen fibrils in the human corneal stroma: structure and aging. Invest Ophthalmol Vis Sci 1998; 39: 644–8.

6. Elsheikh A, Wang D, Brown M, Rama P, Campanelli M, Pye D. Assessment of corneal biomechanical properties and their variation with age. Curr Eye Res 2007; 32: 11–19.

7. Tuft SJ, Moodaley LC, Gregory WM, Davison CR, Buckley RJ. Prognostic factors for the progression of keratoconus. Ophthalmology 1994; 101: 439–47.

8. Chace KV, Carubelli R, Nordquist RE. The role of nonenzymatic glycosylation, transition metals, and free radicals in the formation of collagen aggregates. Arch Biochem Biophys 1991; 288: 473–80.

9. Seiler T, Huhle S, Spoerl E, Kunath H. Manifest diabetes and keratoconus: a retrospective case-control study. Graefes Arch Clin Exp Ophthalmol 2000; 238: 822–5.

10. Cannon DJ, Foster CS. Collagen crosslinking in keratoconus. Invest Ophthalmol Vis Sci 1978; 17: 63–5.

11. Andreassen TT, Simonsen AH, Oxlund H. Biomechanical properties of keratoconus and normal corneas. Exp Eye Res 1980; 31: 435–41.

12. Spoerl E, Huhle M, Seiler T. Induction of cross-links in corneal tissue. Exp Eye Res 1998; 66: 97–103.

13. Wollensak G. Crosslinking treatment of progressive keratoconus: new hope. Curr Opin Ophthalmol 2006; 17: 356–60.

14. McCall AS, Kraft S, Edelhauser HF et al. Mechanisms of Corneal Tissue Cross-Linking in Response to Treatment with Topical Riboflavin and Long Wavelength Ultraviolet Radiation (UVA). Invest Ophthalmol Vis Sci 2009; 30: 30.

15. Sady C, Khosrof S, Nagaraj R. Advanced Maillard reaction and crosslinking of corneal collagen in diabetes. Biochem Biophys Res Commun 1995; 214: 793–7.

16. Sporl E, Huhle M, Kasper M, Seiler T. [Increased rigidity of the cornea caused by intrastromal crosslinking]. Ophthalmologe 1997; 94: 902–6.

17. Wollensak G, Spoerl E, Seiler T. Stress-strain measurements of human and porcine corneas after riboflavinultraviolet- A-induced cross-linking. J Cataract Refract Surg 2003; 29: 1780–5.

18. Kohlhaas M, Spoerl E, Schilde T, Unger G, Wittig C, Pillunat LE. Biomechanical evidence of the distribution of cross-links in corneas treated with riboflavin and ultraviolet A light. J Cataract Refract Surg 2006; 32: 279–83.

19. Tanter M, Touboul D, Bercoff J, Fink M. High Resolution Quantitative Imaging of Cornea Elasticity Using Supersonic Shear Imaging. IEEE Trans Med Imaging 2009; 5: 5.

20. Wollensak G, Spoerl E, Wilsch M, Seiler T. Keratocyte apoptosis after corneal collagen cross-linking using riboflavin/UVA treatment. Cornea 2004; 23: 43–9.

21. Wollensak G, Sporl E, Reber F, Pillunat L, Funk R. Corneal endothelial cytotoxicity of riboflavin/UVA treatment in vitro. Ophthalmic Res 2003; 35: 324–8.

22. Spoerl E, Mrochen M, Sliney D, Trokel S, Seiler T. Safety of UVA-riboflavin cross-linking of the cornea. Cornea 2007; 26: 385–9.

23. Dhaliwal JS, Kaufman SC. Corneal collagen crosslinking: a confocal, electron, and light microscopy study of eye bank corneas. Cornea 2009; 28: 62–7.

24. Wollensak G, Iomdina E, Dittert DD, Herbst H.Wound healing in the rabbit cornea after corneal collagen cross-linking with riboflavin and UVA. Cornea 2007; 26: 600–5.

25. Wollensak G, Aurich H, Pham DT, Wirbelauer C. Hydration behavior of porcine cornea crosslinked with riboflavin and ultraviolet A. J Cataract Refract Surg 2007; 33: 516–21.

26. Bottos KM, Dreyfuss JL, Regatieri CV et al. Immunofluorescence confocal microscopy of porcine corneas following collagen cross-linking treatment with riboflavin and ultraviolet A. J Refract Surg 2008; 24: S715–19.

27. Spoerl E, Wollensak G, Dittert DD, Seiler T. Thermomechanical behavior of collagen-cross-linked porcine cornea. Ophthalmologica 2004; 218: 136–40.

28. Spoerl E,Wollensak G, Seiler T. Increased resistance of crosslinked cornea against enzymatic digestion. Curr Eye Res 2004; 29: 35–40.

29. Stewart JM, Schultz DS, Lee OT, Trinidad ML. Collagen cross-links reduce corneal permeability. Invest Ophthalmol Vis Sci 2009; 50: 1606–12.

30. Mencucci R, Mazzotta C, Rossi F et al. Riboflavin and ultraviolet A collagen crosslinking: in vivo thermographic analysis of the corneal surface. J Cataract Refract Surg 2007; 33: 1005–8.

31. Wollensak G, Wilsch M, Spoerl E, Seiler T. Collagen fiber diameter in the rabbit cornea after collagen crosslinking by riboflavin/UVA. Cornea 2004; 23: 503–7.

32. Wollensak G, Redl B. Gel electrophoretic analysis of corneal collagen after photodynamic cross-linking treatment. Cornea 2008; 27: 353–6.

33. Caporossi A, Baiocchi S, Mazzotta C, Traversi C, Caporossi T. Parasurgical therapy for keratoconus by riboflavin-ultraviolet type A rays induced crosslinking of corneal collagen: preliminary refractive results in an Italian study. J Cataract Refract Surg 2006; 32: 837–45.

34. 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.

35. Wollensak G, Spoerl E, Seiler T. Riboflavin/ ultraviolet-a-induced collagen crosslinking for the treatment of keratoconus. Am J Ophthalmol 2003; 135: 620–7.

36. Wittig-Silva C, Whiting M, Lamoureux E, Lindsay RG, Sullivan LJ, Snibson GR. A randomized controlled trial of corneal collagen cross-linking in progressive keratoconus: preliminary results. J Refract Surg 2008; 24: S720–5.

37. Jankov MR 2nd, Hafezi F, Beko M et al. [Corneal Cross-linking for the treatment of keratoconus: preliminary results]. Arq Bras Oftalmol 2008; 71: 813– 18.

38. Hoyer A, Raiskup-Wolf F, Sporl E, Pillunat LE. [Collagen cross-linking with riboflavin and UVA light in keratoconus – results from Dresden.]. Ophthalmologe 2009; 106: 133–40.

39. Fournie P, Galiacy S, Arne JL, Malecaze F. [Corneal collagen cross-linking with ultraviolet-A light and riboflavin for the treatment of progressive keratoconus]. J Fr Ophtalmol 2009; 32: 1–7.

40. Grewal DS, Brar GS, Jain R, Sood V, Singla M, Grewal SP. Corneal collagen crosslinking using riboflavin and ultraviolet-A light for keratoconus: one-year analysis using Scheimpflug imaging. J Cataract Refract Surg 2009; 35: 425–32.

41. Coskunseven E, Jankov MR 2nd, Hafezi F. Contralateral eye study of corneal collagen cross-linking with riboflavin and UVA irradiation in patients with keratoconus. J Refract Surg 2009; 25: 371–6.

42. Agrawal VB. Corneal collagen cross-linking with riboflavin and ultraviolet – a light for keratoconus: results in Indian eyes. Indian J Ophthalmol 2009; 57: 111–14.

43. Vinciguerra P, Albe E, Trazza S et al. Refractive, topographic, tomographic, and aberrometric analysis of keratoconic eyes undergoing corneal cross-linking. Ophthalmology 2009; 116: 369–78.

44. Vinciguerra P, Albe E, Trazza S, Seiler T, Epstein D. Intraoperative and postoperative effects of corneal collagen cross-linking on progressive keratoconus. Arch Ophthalmol 2009; 127: 1258–65.

45. Koller T, Iseli HP, Hafezi F, Vinciguerra P, Seiler T. Scheimpflug imaging of corneas after collagen crosslinking. Cornea 2009; 28: 510–15.

46. Wagner H, Barr JT, Zadnik K. Collaborative Longitudinal Evaluation of Keratoconus (CLEK) Study: methods and findings to date. Cont Lens Anterior Eye 2007; 30: 223–32.

47. Romppainen T, Bachmann LM, Kaufmann C, Kniestedt C, Mrochen M, Thiel MA. Effect of riboflavin-UVA-induced collagen cross-linking on intraocular pressure measurement. Invest Ophthalmol Vis Sci 2007; 48: 5494–8.

48. Koller T, Mrochen M, Seiler T. Complication and failure rates after corneal crosslinking. J Cataract Refract Surg 2009; 35: 1358–62.

49. Seiler T, Hafezi F. Corneal cross-linking-induced stromal demarcation line. Cornea 2006; 25: 1057–9.

50. Ku JY, Niederer RL, Patel DV, Sherwin T, McGhee CN. Laser scanning in vivo confocal analysis of keratocyte density in keratoconus. Ophthalmology 2008; 115: 845– 50.

51. Weed KH, MacEwen CJ, Cox A, McGhee CN. Quantitative analysis of corneal microstructure in keratoconus utilising in vivo confocal microscopy. Eye (London, England) 2007; 21: 614–23.

52. Mazzotta C, Balestrazzi A, Traversi C et al. Treatment of progressive keratoconus by riboflavin-UVA-induced cross-linking of corneal collagen: ultrastructural analysis by Heidelberg Retinal Tomograph II in vivo confocal microscopy in humans. Cornea 2007; 26: 390–7.

53. Kymionis GD, Diakonis VF, Kalyvianaki M et al. Oneyear follow-up of corneal confocal microscopy after corneal cross-linking in patients with post laser in situ keratosmileusis ectasia and keratoconus. Am J Ophthalmol 2009; 147: 774–8.

54. Croxatto JO, Tytiun AE,Argento CJ. Sequential in Vivo Confocal Microscopy Study of CornealWound Healing After Cross-Linking in Patients with Keratoconus. J Refract Surg 2009; 1–8. Epub ahead of print.

55. Mazzotta C, Traversi C, Baiocchi S et al. Corneal healing after riboflavin ultraviolet-A collagen crosslinking determined by confocal laser scanning microscopy in vivo: early and late modifications. Am J Ophthalmol 2008; 146: 527–33.

56. Hafezi F, Mrochen M, Iseli HP, Seiler T. Collagen crosslinking with ultraviolet-A and hypoosmolar riboflavin solution in thin corneas. J Cataract Refract Surg 2009; 35: 621–4.

57. Angunawela RI, Arnalich-Montiel F, Allan BD. Peripheral sterile corneal infiltrates and melting after collagen crosslinking for keratoconus. J Cataract Refract Surg 2009; 35: 606–7.

58. Koppen C, Vryghem JC, Gobin L, Tassignon MJ. Keratitis and corneal scarring after UVA/riboflavin cross-linking for keratoconus. J Refract Surg 2009; 25: S819–23.

59. Zamora KV, Males JJ. Polymicrobial keratitis after a collagen cross-linking procedure with postoperative use of a contact lens: a case report. Cornea 2009; 28: 474–6.

60. Pollhammer M, Cursiefen C. Bacterial keratitis early after corneal crosslinking with riboflavin and ultraviolet-A. J Cataract Refract Surg 2009; 35: 588–9.

61. Perez-Santonja JJ, Artola A, Javaloy J, Alio JL, Abad JL. Microbial keratitis after corneal collagen crosslinking. J Cataract Refract Surg 2009; 35: 1138–40.

62. Rama P, Di Matteo F, Matuska S, Paganoni G, Spinelli A. Acanthamoeba keratitis with perforation after corneal crosslinking and bandage contact lens use. J Cataract Refract Surg 2009; 35: 788–91.

63. Kymionis GD, Portaliou DM, Bouzoukis DI et al. Herpetic keratitis with iritis after corneal crosslinking with riboflavin and ultraviolet A for keratoconus. J Cataract Refract Surg 2007; 33: 1982–4.

64. Sporl E, Raiskup-Wolf F, Pillunat LE. [Biophysical principles of collagen cross-linking]. Klin Monatsbl Augenheilkd 2008; 225: 131–7.

65. Rocha KM, Ramos-Esteban JC, Qian Y, Herekar S, Krueger RR. Comparative study of riboflavin-UVA cross-linking and ‘flash-linking’ using surface wave elastometry. J Refract Surg 2008; 24: S748–51.

66. Hayes S, O’Brart DP, Lamdin LS et al. Effect of complete epithelial debridement before riboflavinultraviolet- A corneal collagen crosslinking therapy. J Cataract Refract Surg 2008; 34: 657–61.

67. Baiocchi S, Mazzotta C, Cerretani D, Caporossi T, Caporossi A. Corneal crosslinking: riboflavin concentration in corneal stroma exposed with and without epithelium. J Cataract Refract Surg 2009; 35: 893–9.

68. Wollensak G, Iomdina E. Biomechanical and histological changes after corneal crosslinking with and without epithelial debridement. J Cataract Refract Surg 2009; 35: 540–6.

69. Samaras K, O’Brart DP, Doutch J, Hayes S, Marshall J, Meek KM. Effect of epithelial retention and removal on riboflavin absorption in porcine corneas. J Refract Surg 2009; 25: 771–5.

70. Bakke EF, Stojanovic A, Chen X, Drolsum L. Penetration of riboflavin and postoperative pain in corneal collagen crosslinking: excimer laser superficial versus mechanical full-thickness epithelial removal. J Cataract Refract Surg 2009; 35: 1363–6.

71. Bueeler M, Spoerl E, Seiler T, Mrochen M. UV collagen cross-linking of the cornea – safety aspects and design of a UV illumination system. 2008; 1–11.

72. Chan CC, Sharma M, Wachler BS. Effect of inferiorsegment Intacs with and without C3-R on keratoconus. J Cataract Refract Surg 2007; 33: 75–80.

73. Ertan A, Karacal H, Kamburoglu G. Refractive and topographic results of transepithelial cross-linking treatment in eyes with intacs. Cornea 2009; 1: 1.

74. Alio JL, Shabayek MH, Artola A. Intracorneal ring segments for keratoconus correction: long-term follow-up. J Cataract Refract Surg 2006; 32: 978–85.

75. Kymionis GD, Karavitaki AE, Kounis GA, Portaliou DM, Yoo SH, Pallikaris IG. Management of pellucid marginal corneal degeneration with simultaneous customized photorefractive keratectomy and collagen crosslinking. J Cataract Refract Surg 2009; 35: 1298–301.

76. Dissertation. zur Erlangung des Doktorgrades der Naturwissenschaften (Dr. rer. nat.) der Naturwissenschaftlichen Fakultat II – Physik der Universitat Regensburg. vorgelegt von Jurgen Baier aus Bamberg November 2005.

77. Kamburoglu G, Ertan A. Intacs implantation with sequential collagen cross-linking treatment in postoperative LASIK ectasia. J Refract Surg 2008; 24: S726–9.

78. Kohlhaas M, Spoerl E, Speck A, Schilde T, Sandner D, Pillunat LE. [A new treatment of keratectasia after LASIK by using collagen with riboflavin/UVA light cross-linking]. Klin Monatsbl Augenheilkd 2005; 222: 430–6.

79. Hafezi F, Kanellopoulos J, Wiltfang R, Seiler T. Corneal collagen crosslinking with riboflavin and ultraviolet A to treat induced keratectasia after laser in situ keratomileusis. J Cataract Refract Surg 2007; 33: 2035–40.

80. Randleman JB. Post-laser in-situ keratomileusis ectasia: current understanding and future directions. Curr Opin Ophthalmol 2006; 17: 406–12.

81. Kanellopoulos AJ, Binder PS. Collagen cross-linking (CCL) with sequential topography-guided PRK: a temporizing alternative for keratoconus to penetrating keratoplasty. Cornea 2007; 26: 891–5.

82. Ehlers N, Hjortdal J. Riboflavin-ultraviolet light induced cross-linking in endothelial decompensation. Acta Ophthalmol 2008; 86: 549–51.

83. Wollensak G, Aurich H, Wirbelauer C, Pham DT. Potential use of riboflavin/UVA cross-linking in bullous keratopathy. Ophthalmic Res 2009; 41: 114–17.

84. Krueger RR, Ramos-Esteban JC, Kanellopoulos AJ. Staged intrastromal delivery of riboflavin with UVA cross-linking in advanced bullous keratopathy: laboratory investigation and first clinical case. J Refract Surg 2008; 24: S730–6.

85. Corbin F 3rd. Pathogen inactivation of blood components: current status and introduction of an approach using riboflavin as a photosensitizer. Int J Hematol 2002; 76 (Suppl. 2): 253–7.

86. Kumar V, Lockerbie O, Keil SD et al. Riboflavin and UV-light based pathogen reduction: extent and consequence of DNA damage at the molecular level. Photochem Photobiology 2004; 80: 15–21.

87. Ruane PH, Edrich R, Gampp D, Keil SD, Leonard RL, Goodrich RP. Photochemical inactivation of selected viruses and bacteria in platelet concentrates using riboflavin and light. Transfusion 2004; 44: 877–85.

88. Martins SA, Combs JC, Noguera G et al. Antimicrobial efficacy of riboflavin/UVA combination (365 nm) in vitro for bacterial and fungal isolates: a potential new treatment for infectious keratitis. Invest Ophthalmol Vis Sci 2008; 49: 3402–8.

89. Makdoumi K, Backman A, Mortensen J, Crafoord S. Evaluation of antibacterial efficacy of photo-activated riboflavin using ultraviolet light (UVA). Graefes Arch Clin Exp Ophthalmol 2009; 248: 207–12.

90. Schrier A, Greebel G, Attia H, Trokel S, Smith EF. In vitro antimicrobial efficacy of riboflavin and ultraviolet light on Staphylococcus aureus, methicillin-resistant Staphylococcus aureus, and Pseudomonas aeruginosa. J Refract Surg 2009; 25: S799–802.

91. Micelli Ferrari T, Leozappa M, Lorusso M, Epifani E, Micelli Ferrari L. Escherichia coli keratitis treated with ultraviolet A/riboflavin corneal cross-linking: a case report. Eur J Ophthalmol 2009; 19: 295–7.

92. Moren H, Malmsjo M, Mortensen J, Ohrstrom A. Riboflavin and ultraviolet A collagen crosslinking of the cornea for the treatment of keratitis. Cornea 2009; 31: 31.

93. Iseli HP, Thiel MA, Hafezi F, Kampmeier J, Seiler T. Ultraviolet A/riboflavin corneal cross-linking for infectious keratitis associated with corneal melts. Cornea 2008; 27: 590–4.

94. Schnitzler E, Sporl E, Seiler T. [Irradiation of cornea with ultraviolet light and riboflavin administration as a new treatment for erosive corneal processes, preliminary results in four patients]. Klin Monatsbl Augenheilkd 2000; 217: 190–3.

95. Berman MB. Collagenase inhibitors: rationale for their use in treating corneal ulceration. Int Ophthalmol Clin 1975; 15: 49–66.

96. Wang F. UVA/riboflavin-induced apoptosis in mouse cornea. Ophthalmologica 2008; 222: 369–72.