KERATOCONUS

Keratoconus (from Greek: kerato- horn, cornea; and konos cone), is a degenerative disorder of the eye in which structural changes within the cornea cause it to thin and change to a more conical shape than its normal gradual curve. Keratoconus can cause substantial distortion of vision, with multiple images, streaking and sensitivity to light all often reported by the patient. It is typically diagnosed in the patient's adolescent years and attains its most severe state in the twenties and thirties. If afflicting both eyes, the deterioration in vision can affect the patient's ability to drive a car or read normal print. In most cases, corrective lenses are effective enough to allow the patient to continue to drive legally and likewise function normally. Further progression of the disease may require surgery including intrastromal corneal ring segments, corneal collagen cross-linking, or corneal transplantation. However, despite the disease's unpredictable course, keratoconus can often be successfully managed with little or no impairment to the patient's quality of life. Keratoconus is the most common dystrophy of the cornea, affecting around one person in a thousand. It seems to occur in populations throughout the world, although it occurs more frequently in certain ethnic groups. The exact cause of keratoconus is uncertain, but has been associated with detrimental enzyme activity within the cornea. A genetic link seems likely, as the incidence rate is greater if a family member has been diagnosed. The progression of keratoconus is rapid in patients having Down syndrome. Signs and symptoms A simulation of the multiple images seen by a person with keratoconus. "... a candle, when looked at, appears like a number of lights, confusedly running into one another" Nottingham People with early keratoconus typically notice a minor blurring of their vision and come to their clinician seeking corrective lenses for reading or driving. At early stages, the symptoms of keratoconus may be no different from those of any other refractive defect of the eye. As the disease progresses, vision deteriorates, sometimes rapidly. Visual acuity becomes impaired at all distances, and night vision is often quite poor. Some individuals have vision in one eye that is markedly worse than that in the other eye. The disease is often bilateral, though asymmetrical in many patients. Some develop photophobia (sensitivity to bright light), eye strain from squinting in order to read, or itching in the eye, but there is normally little or no sensation of pain. The classic symptom of keratoconus is the perception of multiple 'ghost' images, known as monocular polyopia. This effect is most clearly seen with a high contrast field, such as a point of light on a dark background. Instead of seeing just one point, a person with keratoconus sees many images of the point, spread out in a chaotic pattern. This pattern does not typically change from day to day, but over time it often takes on new forms. Patients also commonly notice streaking and flaring distortion around light sources. Some even notice the images moving relative to one another in time with their heart beat. The predominant optical aberration of the eye as an optical system in keratoconus is the so-called coma.
1

Diagnosis Prior to any physical examination, the diagnosis of keratoconus frequently begins with an ophthalmologist's or optometrist's assessment of the patient's medical history, particularly the chief complaint and other visual symptoms, the presence of any history of ocular disease or injury which might affect vision, and the presence of any family history of ocular disease. An eye chart, such as a standard Snellen chart of progressively smaller letters, is then used to determine the patient's visual acuity. The eye examination may proceed to measurement of the localised curvature of the cornea with a manual keratometer, with detection of irregular astigmatism suggesting a possibility of keratoconus. Severe cases can exceed the instrument's measuring ability. A further indication can be provided by retinoscopy, in which a light beam is focused on the patient's retina and the reflection, or reflex, observed as the examiner tilts the light source back and forth. Keratoconus is amongst the ophthalmic conditions that exhibit a scissor reflex action of two bands moving toward and away from each other like the blades of a pair of scissors. If keratoconus is suspected, the ophthalmologist or optometrist will search for other characteristic findings of the disease by means of slit lamp examination of the cornea. An advanced case is usually readily apparent to the examiner, and can provide for an unambiguous diagnosis prior to more specialised testing. Under close examination, a ring of yellow-brown to olive-green pigmentation known as a Fleischer ring can be observed in around half of keratoconic eyes. The Fleischer ring, caused by deposition of the iron oxide hemosiderin within the corneal epithelium, is subtle and may not be readily detectable in all cases, but becomes more evident when viewed under a cobalt blue filter. Similarly, around 50% of subjects exhibit Vogt's striae, fine stress lines within the cornea caused by stretching and thinning. The striae temporarily disappear while slight pressure is applied to the eyeball. A highly pronounced cone can create a V-shaped indentation in the lower eyelid when the patient's gaze is directed downwards, known as Munson's sign. Other clinical signs of keratoconus will normally have presented themselves long before Munson's sign becomes apparent, and so this finding, though a classic sign of the disease, tends not to be of primary diagnostic importance.

A handheld keratoscope, sometimes known as Placido's disk, can provide a simple noninvasive visualization of the surface of the cornea by projecting a series of concentric rings of light onto the cornea. A more definitive diagnosis can be obtained using corneal topography, in which an automated instrument projects the illuminated pattern onto the cornea and determines its topology from analysis of the digital image. The topographical map indicates any distortions or scarring in the cornea, with keratoconus revealed by a characteristic steepening of curvature which is usually below the centreline of the eye. The technique can record a snapshot of the degree and extent of the deformation as a benchmark for assessing its rate of progression. It is of particular value in detecting the disorder in its early stages when other signs have not yet presented. Once keratoconus has been diagnosed, its degree may be classified by several metrics: The steepness of greatest curvature from mild (< 45 D), advanced (up to 52 D) or severe (> 52 D); The morphology of the cone: nipple (small: 5 mm and near-central), oval (larger, below-center and often sagging), or globus (more than 75% of cornea affected);
2

The corneal thickness from mild (> 506 m) to advanced (< 446 m). Increasing use of corneal topography has led to a decline in use of these terms.

Cause A specimen of keratoconic cornea taken out six years after diagnosis: thin stroma, wrinkled posterior surface. Despite considerable research, the etiology of keratoconus remains somewhat of a mystery. A number of sources suggest that keratoconus likely arises from a number of different factors: genetic, environmental or cellular, any of which may form the trigger for the onset of the disease. A genetic predisposition to keratoconus has been observed, with the disease running in certain families, and incidences reported of concordance in identical twins. The frequency of occurrence in close family members is not clearly defined, though it is known to be considerably higher than that in the general population, and studies have obtained estimates ranging between 6% and 19%. A responsible gene has not been identified: two studies involving isolated, largely homogenetic communities have contrarily mapped putative gene locations to chromosomes 16q and 20q. However, most genetic studies agree on an autosomal dominant model of inheritance. Keratoconus is also diagnosed more often in people with Down syndrome, though the reasons for this link have not yet been determined. Keratoconus has been associated with atopic diseases, which include asthma, allergies, and eczema, and it is not uncommon for several or all of these diseases to affect one person. A number of studies suggest that vigorous eye rubbing contributes to the progression of keratoconus, and that patients should be discouraged from the practice. Pathophysiology Once initiated, the disease normally develops by progressive dissolution of Bowman's layer, which lies between the corneal epithelium and stroma. As the two come into contact, cellular and structural changes in the cornea adversely affect its integrity and lead to the bulging and scarring that are characteristic of the disorder. Within any individual keratoconic cornea, there may be found regions of degenerative thinning coexisting with regions undergoing wound healing. The visual distortion experienced by the patient comes from two sources, one being the irregular deformation of the surface of the cornea; the other being scarring that occurs on its exposed highpoints. These factors act to form regions on the cornea that map an image to different locations on the retina and give rise to the symptom of monocular polyopia. The effect can worsen in low light conditions as the dark-adapted pupil dilates to expose more of the irregular surface of the cornea. Scarring appears to be an aspect of the corneal degradation; however, a recent, large, multi-center study suggests that abrasion by contact lenses may increase the likelihood of this finding by a factor of over two. A number of patients complain of chronic eye rubbing and also think it as a possible cause to the disease but it is not so however it has been observed that Kerataconus progresses faster due to eye-rubbing. A number of studies have indicated that keratoconic corneas show signs of increased activity by proteases, a class of enzymes that break some of the collagen cross-linkages in the stroma, with a simultaneous reduced expression of protease inhibitors. Other studies have
3

suggested that reduced activity by the enzyme aldehyde dehydrogenase may be responsible for a build-up of free radicals and oxidising species in the cornea. It seems likely that, whatever the pathogenetical process, the damage caused by activity within the cornea results in a reduction in its thickness and biomechanical strength. While keratoconus is considered a non-inflammatory disorder, one study shows that rigid contact lens wear by patients leads to overexpression of proinflammatory cytokines, such as IL-6, TNF-alpha, ICAM-1, and VCAM-1 in the tear fluid. Treatment Contact lenses: Rigid gas permeable lens for keratoconus In early stages of keratoconus, spectacles or soft contact lenses can suffice to correct for the mild astigmatism. As the condition progresses, these may no longer provide the patient with a satisfactory degree of visual acuity, and most clinical practitioners will move to managing the condition with rigid contact lenses, known as rigid gas-permeables, or RGPs. RGP lenses provide a good level of visual correction, but do not arrest progression of the condition. In keratoconic patients, rigid contact lenses improve vision by means of tear fluid filling the gap between the irregular corneal surface and the smooth regular inner surface of the lens, thereby creating the effect of a smoother cornea. Many specialized types of contact lenses have been developed for keratoconus, and affected people may seek out both doctors specialized in conditions of the cornea, and contact lens fitters who have experience managing patients with keratoconus. The irregular cone presents a challenge and the fitter will endeavour to produce a lens with the optimal contact, stability and steepness. Some trial-and-error fitting may prove necessary. Traditionally, contact lenses for keratoconus have been the 'hard' or rigid gas-permeable variety, although manufacturers have also produced specialized 'soft' or hydrophilic lenses and, most recently, silicone hydrogel lenses. A soft lens has a tendency to conform to the conical shape of the cornea, thus diminishing its effect. To counter this, hybrid lenses have been developed which are hard in the centre and encompassed by a soft skirt. Soft or hybrid lenses do not however prove effective for every patient. Some patients also find good vision correction and comfort with a "piggyback" lens combination, in which gas permeable rigid lenses are worn over soft lenses, both providing a degree of vision correction. One form of piggyback lens makes use of a soft lens with a countersunk central area to accept the rigid lens. Fitting a piggyback lens combination requires experience on the part of the lens fitter, and tolerance on the part of the keratoconic patient. Scleral lenses are sometimes prescribed for cases of advanced or very irregular keratoconus; these lenses cover a greater proportion of the surface of the eye and hence can offer improved stability. The larger size of the lenses may make them unappealing or uncomfortable to some, however their easier handling can find favour with patients with reduced dexterity, such as the elderly. Surgical options: Corneal transplant
4

Corneal transplant for keratoconus, approximately one week after surgery. Multiple light reflections indicate folds in the cornea which later resolved. Between 10% and 25% of cases of keratoconus will progress to a point where vision correction is no longer possible, thinning of the cornea becomes excessive, or scarring as a result of contact lens wear causes problems of its own, and a corneal transplantation or penetrating keratoplasty becomes required. Keratoconus is the most common grounds for conducting a penetrating keratoplasty, generally accounting for around a quarter of such procedures. The corneal transplant surgeon trephines a lenticule of corneal tissue and then grafts the donor cornea to the existing eye tissue, usually using a combination of running and individual sutures. The cornea does not have a direct blood supply, and so donor tissue is not required to be blood type matched. Eye banks check the donor corneas for any disease or cellular irregularities. Spanish-born eye surgeon Ramon Castroviejo successfully performed keratoplasty as early as 1936 The acute recovery period can take four to six weeks and full post-operative vision stabilization often takes a year or more but most transplants are very stable in the long term. The National Keratoconus Foundation reports that penetrating keratoplasty has the most successful outcome of all transplant procedures, and when performed for keratoconus in an otherwise healthy eye, its success rate can be 95% or greater. The sutures used usually dissolve over a period of three to five years but individual sutures can be removed during the healing process if they are causing irritation to the patient. In the USA, corneal transplants (also known as corneal grafts) for keratoconus are usually performed under sedation as outpatient surgery. In other countries, such as Australia and the UK, the operation is commonly performed with the patient undergoing a general anaesthetic. All cases require a careful follow-up with an eye surgeon (ophthalmologist) for a number of years. Frequently, vision is greatly improved after the surgery, but even if the actual visual acuity does not improve, because the cornea is a more normal shape after the healing is completed, patients can more easily be fitted with corrective lenses. Complications of corneal transplants are mostly related to vascularization of the corneal tissue and rejection of the donor cornea. Vision loss is very rare, though difficult-to-correct vision is possible. When rejection is severe, repeat transplants are often attempted, and are frequently successful. Keratoconus will not normally reoccur in the transplanted cornea; incidences of this have been observed, but are usually attributed to incomplete excision of the original cornea or inadequate screening of the donor tissue. The long-term outlook for corneal transplants performed for keratoconus is usually favorable once the initial healing period is completed and a few years have elapsed without problems. Corneal ring segment inserts: A pair of Intacs inserts A recent surgical alternative to corneal transplant is the insertion of intrastromal corneal ring segments. A small incision is made in the periphery of the cornea and two thin arcs of polymethyl methacrylate are slid between the layers of the stroma on either side of the pupil before the incision is closed. The segments push out against the curvature of the cornea, flattening the peak of the cone and returning it to a more natural shape. The procedure, carried out on an outpatient basis under local anaesthesia, offers the benefit of being reversible and even potentially exchangeable as it involves no removal of eye tissue.
5

The principal intrastromal ring available is known by the trade name of Intacs. Internationally, Ferrara Rings are also available. Intacs are a patented technology and are placed outside the optical zone versus the smaller prismatic Ferrara rings that are placed just inside the 5 mm optical zone. Intacs are the only corneal implants that have gone through the FDA Phase I, II and III clinical trials and were first approved by the Food and Drug Administration (FDA) in the United States in 1999 for myopia; this was extended to the treatment of keratoconus in July 2004. Clinical studies on the effectiveness of intrastromal rings on keratoconus are in their early stages, and results have so far been generally encouraging, though they have yet to enter into wide acceptance with the refractive surgery community. In common with penetrating keratoplasty, the requirement for some vision correction in the form of spectacles or hydrophilic contact lenses may remain subsequent to the operation. Potential complications of intrastromal rings include accidental penetration through to the anterior chamber when forming the channel, post-operative infection of the cornea, and migration or extrusion of the segments. The rings offer a good chance of vision improvement even in otherwise hard to manage eyes, but results are not guaranteed and in a few cases may worsen. Early studies on intrastromal corneal rings involved use of two segments to cause global flattening of the cornea. A later study reported that better results could be obtained for those cones located more to the periphery of the cornea by using a single Intacs segment. This leads to preferential flattening of the cone below, but also to steepening the over-flat upper part of the cornea. Corneal collagen crosslinking with riboflavin: Removed corneal epithelium during CCR operation on an eye with post-LASIK complication. A treatment developed at the Technische Universitt Dresden, and which has shown early success is Corneal Collagen Crosslinking with Riboflavin, also known as CXL, CCR, and CCL. A one-time application of riboflavin solution is administered to the eye and is activated by illumination with UV-A light for approximately 30 minutes. The riboflavin causes new bonds to form across adjacent collagen strands in the stromal layer of the cornea, which recovers and preserves some of the cornea's mechanical strength. The corneal epithelial layer is generally removed in order to increase penetration of the riboflavin into the stroma. Clinical trials are ongoing, but crosslinking is seeing increasing adoption by the ophthalmological community, and has shown success in treating early cases of the disease. The procedure, with epithelium removed, is approved for use throughout Europe, and has recently commenced clinical trials in the USA. Early results are very promising with one study reporting stabilization in all treated eyes, and a slight correction in visual acuity in most patients. In some cases, crosslinking may also be successfully combined with other treatment methods such as corneal ring segment inserts. Corrective lenses may still be required after this treatment, but it may play an important role in limiting deterioration of vision and reducing the case for corneal transplantation.
6

Radial keratotomy: Radial keratotomy is a refractive surgery procedure where the surgeon makes a spoke-like pattern of incisions into the cornea to modify its shape. This early surgical option for myopia has been largely superseded by LASIK and other similar procedures. LASIK itself is absolutely contraindicated in keratoconus and other corneal thinning conditions as removal of corneal stromal tissue will further damage an already thin and weak cornea. For similar reasons, radial keratotomy has also generally not been used for keratoconic patients. However, an Italian clinic has reported some success with a modified asymmetric radial keratotomy procedure, in which the incisions are confined to one sector of the eye. The corneal thickness is first measured using a pachymeter, then the surgeon makes cuts to a depth of 70 80% of the measured thickness. The patient may initially experience photophobia and fluctuation of vision, in common with other forms of refractive surgery. This technique has yet to go through the official experimentation and follow-up period which is generally required by the Italian National Health Service to accept a new surgery technique before it can be offered to patients. DALK transplants: One way of reducing the risk of rejection is to use a newer technique called a Deep Anterior Lamellar Keratoplasty, referred to as DALK. In a DALK graft, only the outermost epithelium and the main bulk of the cornea, the stroma, are replaced; the patient's rearmost endothelium layer and the Descemet's membrane are left, giving some additional structural integrity to the post-graft cornea. Because a graft rejection usually begins in the endothelium, the chance of a rejection episode is greatly reduced. Furthermore, it is possible to transplant tissue from a donor which has been freeze-dried. The freeze-drying process ensures that this tissue is dead, so there is no chance of a rejection. Some surgeons prefer to remove the donor epithelium, others leave the donor's cells in place. Removing it can cause a slight improvement in overall vision, but a corresponding increase in visual recovery time. Epikeratophakia: Rarely, a non-penetrating keratoplasty known as an epikeratophakia (or epikeratoplasty) may be performed in cases of keratoconus. The corneal epithelium is removed and a lenticule of donor cornea grafted on top of it. The procedure requires a greater level of skill on the part of the surgeon, and is less frequently performed than a penetrating keratoplasty as the outcome is generally less favorable. It may however be seen as an option in a number of cases, particularly for young patients. Prognosis A small rupture of Descemet's membrane Patients with keratoconus typically present initially with mild astigmatism, commonly at the onset of puberty, and are diagnosed as having the disease by the late teenage years or early
7

20s. In rare cases keratoconus can occur in children or not present until later adulthood. A diagnosis of the disease at an early age may indicate a greater risk of severity in later life. Patients' vision will seem to fluctuate over a period of months, driving them to change lens prescriptions frequently but as the condition worsens, contact lenses become required in the majority of cases. The course of the disorder can be quite variable, with some patients remaining stable for years or indefinitely, while others progress rapidly or experience occasional exacerbations over a long and otherwise steady course. Most commonly, keratoconus progresses for a period of ten to twenty years before the course of the disease generally ceases. ... leads to corneal hydrops. In advanced cases, bulging of the cornea can result in a localized rupture of Descemet's membrane, an inner layer of the cornea. Aqueous humor from the eye's anterior chamber seeps into the cornea before Descemet's membrane reseals. The patient experiences pain and a sudden severe clouding of vision, with the cornea taking on a translucent milky-white appearance known as a corneal hydrops. Although disconcerting to the patient, the effect is normally temporary and after a period of six to eight weeks the cornea usually returns to its former transparency. The recovery can be aided non-surgically by bandaging with an osmotic saline solution. Although a hydrops usually causes increased scarring of the cornea, occasionally it will benefit a patient by creating a flatter cone, aiding the fitting of contact lenses. Occasionally, in extreme cases, the cornea thins to the point that a partial rupture occurs, resulting in a small, bead-like swelling on the cornea that has been filled with fluid. When this occurs, a corneal transplant can become urgently necessary to avoid complete rupture and resulting loss of the eye. Epidemiology The National Eye Institute reports that keratoconus is the most common corneal dystrophy in the United States, affecting approximately 1 in 2,000 Americans, but some reports place the figure as high as 1 in 500. The inconsistency may be due to variations in diagnostic criteria, with some cases of severe astigmatism interpreted as those of keratoconus, and vice versa. A long-term study found a mean incidence rate of 2.0 new cases per 100,000 population per year. It is suggested that males and females, and all ethnicities appear equally susceptible, though some recent studies have cast doubt upon this, suggesting a higher prevalence amongst females; the literature however varying as to its extent. Also, a study carried out in the UK suggests that people of an Asian heritage are 4.4 times as likely to suffer from keratoconus as Caucasians, and are also more likely to be affected with the condition earlier. Keratoconus is normally bilateral (affecting both eyes) although the distortion is usually asymmetric and is rarely completely identical in both corneas. Unilateral cases tend to be uncommon, and may in fact be very rare if a very mild condition in the better eye is simply below the limit of clinical detection. It is common for keratoconus to be diagnosed first in one eye and not until later in the other. As the condition then progresses in both eyes, the vision in the earlier-diagnosed eye will often persist to be poorer than that in its fellow. History Practical observations on conical cornea, 1854

In a 1748 doctoral dissertation, the German oculist Burchard Mauchart provided an early description of a case of keratoconus, which he called staphyloma diaphanum. However, it was not until 1854 that British physician John Nottingham clearly described keratoconus and distinguished it from other ectasias of the cornea. Nottingham reported the cases of "conical cornea" that had come to his attention, and described several classic features of the disease, including polyopia, weakness of the cornea, and difficulty matching corrective lenses to the patient's vision. In 1859 British surgeon William Bowman used an ophthalmoscope (recently invented by Hermann von Helmholtz) to diagnose keratoconus, and described how to angle the instrument's mirror so as to best see the conical shape of the cornea. Bowman also attempted to restore the vision by pulling on the iris with a fine hook inserted through the cornea and stretching the pupil into a vertical slit, like that of a cat. He reported that he had had a measure of success with the technique, restoring vision to an 18-year old woman who had previously been unable to count fingers at a distance of 8 inches (20 cm). By 1869, when the pioneering Swiss ophthalmologist Johann Horner wrote a thesis entitled On the treatment of keratoconus, the disorder had acquired its current name. The treatment at that time, endorsed by the leading German ophthalmologist Albrecht von Graefe, was an attempt to physically reshape the cornea by chemical cauterization with a silver nitrate solution and application of a miosis-causing agent with a pressure dressing. In 1888 the treatment of keratoconus became one of the first practical applications of the then newly invented contact lens, when the French physician Eugne Kalt manufactured a glass scleral shell which improved vision by compressing the cornea into a more regular shape. Since the start of the twentieth century, research on keratoconus has both improved understanding of the disease and greatly expanded the range of treatment options. The first successful transplantation of cornea to treat keratoconus was done in 1936 by Ramon Castroviejo. The cornea is the window of the eye. Light travels through the cornea past the lens to the retina and then the brain to form a visual image. The normal corneal surface is smooth and aspheric round in the center, flattening towards its outer edges. Light rays passing through it moves in an undistorted manner to the retina to project a clear image to the brain. In patients with keratoconus the cornea is cone shaped and the surface is also irregular resulting in a distorted image being projected onto the brain. Because the cornea is irregular and cone shaped, glasses do not adequately correct the vision in patients with keratoconus since they cannot conform to the shape of the eye. Patients with keratoconus see best with rigid contact lenses since these lenses provide a clear surface in front of the cornea allowing the light rays to be projected clearly to the retina. Hence the vast majority of patients are treated with rigid contact lenses. There are however some excellent new surgical options for patients with keratoconus who cannot tolerate these lenses, these options are discussed under treatments for keratoconus. Many patients are initially unaware they have keratoconus and see their eye doctor because of increasing spectacle blur or progressive changes in their prescription. In many instances even a good refraction yields poor vision. Keratoconus is most often diagnosed by a cornea specialist who my see typical findings when examining the patient at the slit-lamp. In early forms of the disease there may be no obvious finding on slit-lamp evaluation and the diagnosis is made by computerized videokeratography only. Keratoconus typically commences at puberty and progresses to the mid thirties at which time progression slows and often stops. Between age 12 and 35 it can arrest or progress at any
9

time and there is now way to predict how fast it will progress or if it will progress at all. In general young patients with advanced disease are more likely to progress to the point where they may ultimately require some form of surgical intervention. Keratoconus may occur in one eye only initially but most commonly affects both eyes with one eye being more severely affected than the other. Both males and females are equally affected and there is no ethnic predilection though in some parts of the world such as New Zealand and in certain parts of Finland there is a higher incidence due to genetic factors. Despite millions of dollars being spent on keratoconus no one truly knows the cause of the disease. There have been many interesting theories but none of them have been proven conclusively neither have any of them consistently been reproduced by multiple research groups. For example one theory suggests that there is deficient collagen crosslinking caused by free radicals but there is no scientific reproducible evidence to support such a theory. Others suggest that eye rubbing causes the progression of keratoconus. The evidence for this is however anecdotal based on several case reports, but again there is no reproducible scientific evidence to support this. A research group was the first group to demonstrate that genetic factors play a major role in the development of keratoconus. While our scientific based evidence supports a role for genetic factors this does not mean if you have a child with keratoconus they will necessarily develop keratoconus, since only 13-15% of keratoconus patients have a family history with keratoconus. It does mean however that genes play a role in its development and suggests that any proposed treatment for the disease will either be very temporary or short lived until the genes that contribute to its development are identified and either replaced or suppressed. This is the only potential hope for a permanent cure to stop progression and ensure the cornea will no longer continue to thin.

Yaron S. Rabinowitz M.D. is a corneal surgeon and expert in the treatment and diagnosis of keratoconus. He has published more articles on the diagnosis and treatment of keratoconus than any other practicing eye surgeon in the world. He is Clinical Professor of Ophthalmology at U.C.L.A. School of Medicine and the Director of Eye Research at CedarsSinai Medical Center. His research on the early detection and genetics of keratoconus has received funding from the National Eye Institutes of Health and for the past 15 years his research has provided new insights into the understanding and treatment of keratoconus. Among the insights provided into the understanding of keratoconus are: The first to describe that mild topographic abnormalities occur in family members of patients with keratoconus. Authored one of the first text books on corneal topography. The first to demonstrate through a research study that keratoconus has a genetic basis. The first to demonstrate which keratoconus suspect patterns progress to ultimately develop keratoconus. Most recently published the first article which demonstrates that the femtosecond laser is accurate and preferable for creating channels to insert INTACS a novel new treatment for patients with keratoconus. He has been a recipient of multiple awards to acknowledge his contributions to keratoconus research, these include:
10

The American Academy of Ophthalmology Honor Award. The Jules Stein/UCLA Research Alumni Award. The 3rd American to be the honored guest of the French Eye Society. The recipient of the International Society of Refractive Surgery and the American Academy of Ophthalmology Kritzinger Memorial Research Award for contributions to refractive surgery research.

Materials and methods

The KC propositi, index cases through whom the families were recruited, were sequentially ascertained at the Cornea Genetic Eye Institute at Cedars-Sinai Medical Center from 1992 to 1996. The first-degree relatives of propositi were recruited for the study. Normal controls, with no known clinical evidence or family history of KC, were recruited from spouses or acquaintances of KC patients, as well as employees of Cedars-Sinai Medical Center. The diagnosis of KC was based on clinical examination. Any patient who had one or more of the following clinical signs with no other pathology in one eye was classified as KC: obvious corneal stromal thinning, Vogt striae, or a Fleischer ring detected by slit-lamp examination; obvious scissoring of the red reflex or the Charleaux oil droplet sign was identified by retinoscopy. Propositi with a family history of Down syndrome, Leber congenital amaurosis, or any other recognized genetic disorder were excluded from this study in order to dissect out the contribution of heredity to the development of isolated KC. Eyes with complicating factors, such as cornea transplantation, scarring, or contact lenses wearers, were also excluded from the analysis to avoid the videokeratography artifacts. Moreover, individuals age 13 years or younger were not included in the analysis because they may not be old enough to develop KC. In total we ascertained 539 propositi, 1,226 firstdegree relatives, and 268 controls. We interviewed each individual with a questionnaire including information on demography, medical history, and family history of KC. Of these subjects, 381 KC propositi, 373 of their first-degree relatives, and 252 controls underwent both clinical and videokeratography evaluation. The clinical examinations included slit-lamp biomicroscopy, retinoscopy examinations, and fundus evaluation. The slit-lamp biomicroscope was used to examine whether there was stromal corneal thinning, Vogt striae, or a Fleischer ring. Retinoscopy was performed with a fully dilated pupil (20 minutes after phenylephrine 2.5% and cyclopentolate 1% drops had been instilled in the eye) to determine the presence or absence of retroillumination signs of keratoconus, such as the oil droplet sign and the scissorring. Videokeratography evaluation was also performed on each eye. The study protocol was approved by the Institutional Review Board of Cedars-Sinai Medical Center.

Videokeratography measurements All eyes were studied using the Topographic Modeling System (TMS-1) (Computed Anatomy, Inc., New York, NY), a computerized corneal topographic analysis system (Software
11

Version 1.61). Indices of CK, IS, and KISA on each eye were generated based on this system, which has been described in detail [Rabinowitz, 1995; Rabinowitz and Rasheed, 1999]. CK, a measure of central corneal steepness, is calculated by averaging the dioptric power points on rings 2, 3, and 4 of the videokeratograph. In total 768 data points were evaluated. The IS value, a measure of corneal asymmetry, is calculated by subtracting the superior (S) value from the inferior (I) value. The I value is calculated by averaging 15 data points on rings 14, 15, and 16 of the videokeratographs approximately 3.0 mm inferior to the center of the cornea at 30 degree intervals (i.e., at axes 210, 240, 270, 300, and 330 degree); the S value is derived from 15 data points at 30, 60, 90, 120, and 150 degrees on the superior cornea. A positive IS value indicates a steeper inferior cornea, and a negative IS value indicates a steeper superior cornea. The KISA index is derived from the product of four indices that include CK, IS, and two other measurements (AST and SRAX) that quantify the irregular shape and astigmatism of the cornea, i.e., KISA 4 (CK) (I S) (AST) (SRAX)/3, in which AST (astigmatism) quantifies the degree of regular cornea astigmatism [Wilson et al., 1991] and SRAX (skewness of steepest radial axes) quantifies irregular corneal astigmatism [Rabinowitz and Rasheed, 1999]. At least four pictures were taken to ensure the reproducibility of video images. The best videokeratograph of the four was selected based on the quality of the keratoscope photos by visual inspection. We used the largest value of each index for each pair of eyes in our analysis. Because IS values can be positive or negative and a high value for either can suggest keratoconus, we took the absolute higher IS value for each pair of eyes as the IS value for that subject. Statistical analysis

Log transformation of index KISA was used for all subsequent analyses. The empiric risk of KC in firstdegree relatives was estimated by calculating the prevalence of KC and its 95% confidence interval (CI) in first-degree relatives (propositus parents, sibs, and offspring). We compared both differences between propositi and controls and between clinically unaffected relatives and controls for each of three indices. In order to perform valid statistical tests, we compared independent unaffected parents with controls for each of the three indices. Students ttests implemented in the SAS software were used [SAS Institute Inc., 1990] for comparing these quantitative indices between two groups. Familial correlation and segregation analysis

We examined familial correlations of three indices among all unaffected individuals using the FCOR (Familial Correlation) program of S.A.G.E (Statistical Analysis for Genetic Epidemiology) [Department of Epidemiology and Biostatistics, 1997]. We also performed complex segregation analysis to examine whether major gene effects play an important role in the underlying mechanism of KC development. To increase statistical power, instead of using clinical KC status as the trait, we performed segregation analysis for a quantitative index, KISA. As reviewed above, KISA is an index that integrates the information on both steepness (CK and IS) and asymmetric shape of a cornea (AST and SRAX), which are the typical characteristics of
12

KC. A total of 770 individuals from 95 families with at least three individuals in whom the KISA was measured was included in the segregation analysis. A series of class-A regressive models of REGC (Regressive Models for Continuous Traits) implemented in S.A.G.E. were tested [Bonney, 1984; Department of Epidemiology and Biostatistics, 1997]. The models test whether variation among individuals for a quantitative trait is due to a major gene effect, polygenes, or environmental factors. A general model consists of a major gene effect, polygenic, and environmental components. Two alleles at a single locus (denoted A and B) were assumed as three possible types of individuals, namely AA, AB, and BB. The mean of KISA associated with each underlying type was denoted as AA, AB, and BB, with variance 2AA, 2AB, and 2BB, respectively. The type effects may come from a genetic or environmental factor. The parameter P was defined as the frequency of allele A, and q = (1 P) as the frequency of allele B. Hardy-Weinberg equilibrium was assumed. Three arbitrary transmission parameters (tAA, tAB, tBB) represented the probability that an individual of a given type transmits an allele A to the offspring. These parameters were all estimated under the general model. The parameters rsp and rpo denoted spouse-spouse and parent-offspring correlation, respectively. The sporadic model assumed no familial correlation, i.e., parameters rsp and rpo were fixed to be 0. The environmental model restricted tAA = tAB = tBB = P. In this model, the prior probability of an offspring having a certain type was completely independent of parental types, i.e., the environment is the only determinant. Mendelian models restricted the transmission probabilities tAA, tAB, and tBB to be 1, 0.5, and 0 respectively. The dominant model further fixed AA = AB and the recessive model fixed AB = BB. Under the most general Mendelian model, the three m were free and were estimated. Hypotheses were tested by comparing the general model with various restricted models. Two criteria were used to compare these models. For hierarchical models, the likelihood ratio test (LRT) was used. Twice the difference in log-likelihoods (2Ln(L)) between a restricted and an unrestricted model can be treated as a chi-square statistic with degrees of freedom equal to the difference in the number of parameters being estimated under the two competing models. For non-hierarchical models, Akaikes Information Criterion (AIC) was used [Akaike, 1974]. AIC 4 2Ln(L) + 2k, where k is the number of parameters estimated in the models. Because this statistic combines measures of parsimony and goodness of fit, the hypothesis with the minimum AIC best fits the data. Since families were recruited through the propositi, ascertainment correction was necessary in order to estimate the meaningful parameters for the population. We corrected for ascertainment by conditioning on the propositi in the analysis. Results

Study Population The study included individuals between age 13 and 97 years. Table I shows the age and gender distributions in propositi, their first-degree relatives, and controls. The mean age for propositi, their first-degree relatives, and controls were 36.3, 45.4, and 39.4 years, respectively. More males were observed in propositi (60.1%) than in controls (46.2%) in this study population. The number of males and females was approximately the same among first-degree relatives (49.7% were males).
13

TABLE I. Characteristics of the Study Population Parameters Propositi First-degree relatives N 539 1,226 Age (years) Mean (SD) 36.3 (13.0) 45.4 (18.6) Range (14.097.0) (13.093.0) Male (%) 60.1 49.7 Quantitative measurements

Controls 268 39.4 (12.5) (14.081.0) 46.2

Indices CK, IS, and KISA were normally distributed in controls. The KISA index exhibited a bimodal distribution in the whole sample, including propositi, relatives, and controls. The means of the two peaks were 3.5 D and 7.9 D (Fig. 1a). Indices CK and IS also showed similar distributions, but the bimodality was not as obvious as with KISA (Fig. 1b and c). The bimodal distribution of these traits is consistent with the major gene involvement hypothesis, especially for the KISA index. The means of three quantitative indices in KC propositi and normal controls are compared in Table II. The average values of CK, IS, and KISA were significantly elevated in propositi (51.56 D, 7.51 D, and 7.88 D) compared with controls (43.95 D, 0.41 D, and 2.92 D, respectively) (all P < 0.0001). These data, together with previously reported discriminant analysis in a subset of the current sample [Rabinowitz et al., 1998], indicate that the quantitative videokeratography indices can be used to distinguish KC from normal individuals. TABLE II. Comparisons of KC Indices Between Propositi and Controls Propositi Controls N Mean (SD) N Mean (SD) P value 381 51.56 (3.50) 252 43.95 (1.44) 0.0001 381 7.51 (4.04) 251 0.41 (0.50) 0.0001 372 7.88 (1.59) 247 2.92 (0.96) 0.0001

Indices CK (D) IS (D) KISA (D)

Familial aggregation

The estimated KC prevalence in first-degree relatives was 3.34% (41/1,226, 95% CI42.334.35%). This is 15 to 67 times higher than the general population prevalence (0.23 0.05%) [Rabinowitz, 1998]. Among first-degree relatives, the empiric risk of KC was 3.78% (23/609, 95% CI = 2.285.29%) in sibs, and 2.92% (18/ 617, 95% CI = 1.584.22%) in parents and offspring. The increased empiric risks in relatives of KC patients suggest a strong familial aggregation of KC affection status. For quantitative indices of KC, we compared clinically unaffected relatives with controls. Table III summarizes the means and standard deviations of the three indices for the clinically unaffected relatives, unaffected parents only, and controls. A consistent trend was observed in that clinically unaffected relatives had higher mean values than controls for all three indices. Clinically unaffected parents had significantly increased mean values as compared with controls for all three indices (all P < 0.016 or less). Such results provide further support not only for the familial aggrega-tion of KC, but also for KC
14

related quantitative traits. Because the relatives of KC patients have a greater risk to develop KC than the general population, the increased quantitative indices in relatives provide evidence that these indices may serve as subclinical phenotypes of KC. Table IV shows the familial correlation of the three indices among all unaffected relatives. The correlations of KISA between sib and parent-offspring pairs (r = 0.30 and 0.22, respectively) are significantly greater than that in marital pairs (r = 0.14) (the latter is not significantly different from zero). The estimated upperbound heritability for KISA was 0.6. These results suggest that the familial aggregation in KC is more likely due to genetic factors. TABLE III. Comparison of KC Indices Between Relatives and Controls Unaffected relativesa N Mean (SD) 346 44.73 (1.63) 346 0.74 (0.98) 328 2.93 (1.10) Unaffected parents N Mean (SD) 142 44.70 (1.69) 142 0.89 (1.27) 134 3.20 (1.14) Controls Mean (SD) 43.95 (1.44) 0.41 (0.50) 2.92 (0.96)

Indices CK (D) IS (D) KISA (D)

N 252 251 247

P value* 0.0001 0.0001 0.016

Parents are included. *P value is based on a comparison of unaffected parents versus controls. Segregation analysis

In total, 95 families with at least three individuals with the KISA measurement were used in the segregation analysis. The estimated parameters, 2Ln(L), x2, and the AIC statistics for each model are summarized in Table V. First, we compared the sporadic model (model 2), the environmental model (model 3), and the Mendelian model (model 7) to the general model (model 1). The Mendelian model was not rejected compared with the general model (model 7 versus model 1, P > 0.1). In contrast, both the sporadic and environmental models were significantly different from the general model (model 2 and 3 versus model 1) and were rejected at P < 0.001. Second, we compared data within the major gene models. The major gene-only model was not significantly different from the major gene-plus-environmental model, the major genepluspolygene model, and the major gene with polygene and environmental model (model 7 versus models 4, 5, and 6, all P > 0.05, P-values not shown in the table). This suggested that a major gene could play a prominent role in the transmission of KISA. Finally, we further tested the mode of inheritance under the assumption of Mendelian segregation, the result indicated that the recessive model was the best fitting model, it was not significantly different from the major gene model (model 9 versus model 7, P > 0.1). In contrast, both the dominant and additive models were significantly different from the major gene model (models 8 and 10 versus model 7, both P < 0.01). The most parsimonious model was the Mendelian recessive model (AIC = 1,006.381, the smallest value). The estimated frequency of allele A was 0.37, and the estimated KISA means were 3.93, 2.83, and 2.83 D for AA, AB, and BB genotypes, respectively. Thus, the complex segregation analysis demonstrated that the autosomal-recessive model fit the data best among the tested models

15

Discussion

We have demonstrated strong evidence for familial aggregation of KC and its quantitative measures. Segregation analysis of KISA index showed that a major gene model best fits these data. Autosomal-recessive inheritance for the KISA index was suggested. Thus, our study has provided some insights into the understanding of familiality in KC. This study consistently showed that the videokeratography indices, CK, IS, and KISA, were good quantitative measures for KC affection [Rabinowitz et al., 1998]. Both KC and its quantitative measurements showed familial aggregation in this study population. This is consistent with results of a previous study [Tretter et al., 1995] in which family history was identified as one of the risk factors for KC. The estimated KC prevalence in first-degree relatives was 3.34%, which is much higher than the prevalence in the general population (0.050.23%). The significantly increased correlations of KISA in sib and parentoffspring pairs as compared with marital pairs not only support the familial aggregation of KC but also suggest the importance of genetic factors. To our knowledge, the present study is the first to apply complex segregation analysis to KC pedigree data to investigate the existence of a major gene effect and the mode of inheritance of a KC related index. Segregation analysis of KISA suggested Mendelian inheritance, the major gene model was not rejected from the most general model. An autosomal recessive transmission model fits our data best. In contrast, hypotheses of sporadic and environmental models were strongly rejected in this study. Rejection of the sporadic model was also consistent with the findings of familial aggregation demonstrated by increased empiric risks in relatives and increased correlation of quantitative indices in relative pairs. The bimodal distribution of KC subclinical indices, the familial correlation of both clinical and subclinical measures of KC, and the segregation analysis of KISA all together support the hypothesis of major gene involvement in the genetic mechanisms of KC. Although the recessive model is suggested by these data, there may be genetic heterogeneity as well. For example, we have four large pedigrees referred from other clinics (not included in this study) with a clearly dominant inheritance pattern. Furthermore, segregation analysis does not resolve whether one or more loci are involved. These considerations suggest that nonparametric or model-free methods will likely be the most useful approach in linkage analysis to identify disease susceptibility gene or genes. The data presented in this study showed strong evidence for familial aggregation of KC with respect to affection status and KC related quantitative indices. Segregation analysis provided evidence that this aggregation was well explained by a major gene effect with recessive transmission. Our study also suggests that such indices might have potential as predictors for the early detection of KC in a quantifiable and reproducible manner. Moreover, these quantitative indices may be used as subclinical phenotypes for KC to facilitate the gene mapping endeavors for this disorder. Our findings warrant further investigation of the genetic mechanisms in KC and provide a foundation for future gene mapping studies.

16

KC Indices IS (D) CK (D) KISA (D)

a

TABLE IV. Familial Correlation of KC Indices in Unaffected Individuals Marital Parent offspring Sib a N r P value* N r Pvalue N r Pvalue 49 0.12 ns** 249 0.16 0.01 187 0.24 0.0009 49 0.09 ns 249 0.06 ns 187 0.15 0.04 49 0.14 ns 249 0.22 0.0005 187 0.30 0.00003

Correlation coefficient. *P value was based on comparison of each r to r = 0. **Statistically not significant. Collagen cross-linking

Despite intensive biochemical and genetic research, we do not know what causes keratoconus, although research in these disciplines will ultimately provide answers which could lead to a medical therapy. We do know that keratoconus typically commences at puberty, progressing until the patient's mid thirties, when it typically arrests. During this period, progression may stop and start at any time. It is not possible to predict when keratoconus will progress or stop without intervention. A variant of this disorder, pellucid marginal degeneration, typically starts in the mid thirties and can progress or arrest any time thereafter. Post-LASIK ectasia results in corneal thinning. It is most common in individuals who had forme fruste keratoconus or were predisposed to developing keratoconus prior to their LASIK procedures. For all these ectatic disorders, the treatment of choice is rigid contact lenses. Intacs implants (Addition Technology) and corneal transplants are surgical alternatives for individuals who are contact-lens intolerant. We also know that in keratoconus the cornea is biomechanically unstable. Some literature suggests this may be the result of a decrease in cross-links between and within collagen fibers in the anterior corneal stroma. If there were a way to increase these cross-links, it could provide a means to stiffen the cornea and confer biomechanical stability with the potential to slow the progression of the disorder. In this article, I will detail a procedure that shows promise in achieving these goals. Cross-linking 101 Collagen cross-linking provides a potential means for stiffening the cornea but what is it? Well known in material sciences, it is an enzymatic process whereby there is an increase of molecular bonds to increase the mechanical strength of tissue. Cross-links can be induced enzymatically by means of aldehydes, chemical fixatives and by means of photosensitizing radiation. In vivo experiments have revealed that a combination of UV radiation and riboflavin is the most effective and least harmful procedure for inducing collagen cross-linking in the human cornea (Figure 1). In 2003 Wollensak, Spoerl and Seiler reported on 22 patients with progressive keratoconus upon whom they performed collagen cross-linking and followed for 2 to 4 years.
17

They demonstrated no progression of keratoconus in all eyes, improved visual acuity in 15 of 22 eyes and flattening of K-max by 2 diopters in 16 of 22 eyes. Figure 1. Combining UV radiation and riboflavin is the most effective method to induce collagen cross linking. The procedure as initially described by this group is relatively simple. Using topical anaesthesia, a corneal abrasion is created to facilitate riboflavin diffusion into the cornea. One drop of riboflavin 0.1% ophthalmic solution is instilled topically in the eye every 2 minutes for 30 minutes. At the end of the 30-minute pretreatment period, the eye is examined with blue light for the presence of a yellow flare in the anterior chamber, indicating adequate riboflavin saturation of the corneal tissue. If the yellow flare is not detected, riboflavin is continued at one drop every 2 minutes for an additional two to three drops, then the anterior chamber is rechecked for yellow flare. This process is repeated as necessary. After the presence of the yellow flare is confirmed on slit lamp examination, ultrasonic pachymetry is performed. If the corneal thickness is less than 400 microns, two drops of hypotonic riboflavin 0.1% are instilled every 10 to 15 seconds until the corneal thickness increases to at least 400 microns. During irradiation, instillation of riboflavin is continued every 2 minutes. When the yellow flare in the anterior chamber is confirmed, the eye is aligned under the UV-X light (Figure 2) with the treatment plane at 50 mm from the UV-X beam aperture. The correct aperture setting is selected for the size of the eye; the eye is irradiated for 30 minutes, during which time instillation of riboflavin is continued (one drop every 2 minutes). Figure 2. The eye is irradiated for 30 minutes as riboflavin is instilled. After completion of the procedure, an antibiotic drop is given and a bandage contact lens is placed on the eye. The contact lens is removed once the abrasion has healed. Postoperative medications include an antibiotic and a steroid for 2 weeks postoperatively.

Safety and efficacy There are reports of the procedure being performed without removing the epithelium. This is attractive to patients since they would forgo the pain caused by the abrasion, as well as decrease their risk for infection due to an open wound. However, recent immunofluorescent confocal microscopy studies by Bottos et al. demonstrate that the epithelium is a barrier to crosslinking and very little cross-linking occurs in the presence of epithelium (Figure 3). These findings suggest that for the treatment to be effective, the epithelium should always be removed as initially described by Seiler's group. Figure 3. The epithelium has been found to act as a barrier to cross linking. The procedure appears to be relatively safe. The only adverse event reported to date after cross-linking has been corneal edema in an eye with a pretreatment corneal thickness of less than 400 microns, presumably caused by UV damage to the corneal endothelium. Subsequent experiments led to the conservative recommendation that corneas not be treated with UVA/riboflavin unless they are thicker than 400 microns after epithelial debridement. Other complications reported in the literature are a case of HSV keratitis and DLK in a case of postLASIK ectasia. Both resolved without any long term-effects on the patients.
18

Prior to clinical application of this procedure, there was significant laboratory investigation to demonstrate the safety of the procedure and determine the optimal photosensitizing agent/dosage and irradiance/exposure time. Laboratory studies also focused on characterizing the biomechanical and biochemical effects of cross linking on corneal collagen tissue using rabbit, porcine and human corneal tissue models. Other studies demonstrated toxicity to keratocytes and endothelial cells as well as thermal effects and effects on collagenase resistance. Wollensak's group, using stress-strain measurements on porcine and human corneas, demonstrated a 328% increase in rigidity of the human cornea after being cross-linked with UVA and riboflavin (Figure 4). Subsequent experiments on rabbit corneas demonstrated that the crosslinking effect can last up to 8 months. Confirmation of cross-linking was demonstrated with gel electrophoresis, which demonstrated a significant increase in the diameter of collagen fibres in rabbit corneas. The presence of the cross-linking effect has also been confirmed with thermochemical studies on cross-linked rabbit corneas.

UV concerns A significant concern is the effect of the UV light on the cornea and other ocular structures. This has been studied in detail. The potential cytotoxicity of UVA light and the UVA/riboflavin exposure on keratocytes and endothelial cell function have been characterized in a series of in vitro experiments. In each of these, UVA exposure (370 nm, 30 minutes) and riboflavin (0.025% solution; equivalent to the corneal concentration after diffusion of a 0.1% solution) were administered to mimic conditions of clinical usage. Irradiance levels were varied to determine the irradiance threshold for cytotoxic effects. Keratocyte toxicity was evaluated in porcine keratocyte cell cultures after exposure to riboflavin alone, UVA light alone (irradiance range 2 to 9 mW/cm2), and UVA light plus riboflavin (irradiance range 0.4 to 1 mW/cm2). Riboflavin alone had no cytotoxic effect on keratocytes. The cytotoxic threshold for inducing cellular necrosis or apoptosis was 5 mW/cm2 for UVA light alone and 0.5 mW/cm2 for the UVA/riboflavin treatment. Using the Lambert-Beer equation, in human corneas the cytotoxic keratocyte UVA irradiance of 0.5 mW/cm2 is reached at a stromal depth of 300 microns. The potential for endothelial cell toxicity was evaluated on endothelial cell cultures obtained from porcine cornea that were exposed to riboflavin alone and to various UVA irradiances (range 0.1 to 1.6 mW/cm2) with and without riboflavin. An abrupt cytotoxic threshold was observed at an irradiance of 4 mW/cm2 for UVA light alone and was 10-fold lower with an irradiance threshold of 0.35 mW/cm2 for the UVA/riboflavin treatment. No endothelial cell damage was observed in the cells treated with riboflavin alone. Endothelial cell damage in the UVA groups is believed to be due to oxidative damage caused by the oxygen reactive-free radicals that are generated by the UV light. The lower cytotoxic thresholds observed for the UVA/riboflavin combination in the keratocyte and endothelial cell toxicity studies is consistent with the increase in UVA absorption in the presence of riboflavin. These studies suggest that this treatment is safe in corneas that have been adequately saturated with riboflavin and are at least 400 m in diameter (Figure 5).

19

Research results Subsequent to Wollensak's initial report in 2003, several other studies have suggested the efficacy and safety of this procedure for treating patients with progressive keratoconus. Caporossi et al. reported a 3.6-line increase in uncorrected visual acuity, a 1.66 line improvement in BSCVA, a mean reduction in K-max of 2.1 D (0.13), and a 2.5 D reduction in MRSE at 3 months after cross-linking in a series of 10 eyes in 10 patients with progressive keratoconus. There were no changes in endothelial cell density. Raiskup Wolf et al. reported on 7-year results at the University of Dresden. They noted a decrease in maximum keratometry of 2.7D at 1 year; 2.2D at 2 years and 4.8D at 3 years. Visual acuity improved by one line per year in 54% of patients in the first 3 years. Two patients had continued progression and had to undergo repeat cross linking procedures. In the only randomized prospective controlled clinical trial of collagen cross-linking in progressive keratoconus published to date, Wittig-Silva et al. reported on 66 eyes of 49 patients with documented progression of keratoconus. Interim analysis of treated eyes showed a flattening of the steepest simulated keratometry value (K-max) by an average of 0.74 D at 3 months, 0.92 D at 6 months, and 1.45 D at 12 months. A trend toward improvement in BSCVA was also observed. In the control eyes, mean K-max steepened by 0.60 D after 3 months, by 0.60 D after 6 months, and by 1.28 D after 12 months. BSCVA decreased by logMAR 0.003 over 3 months, 0.056 over 6 months, and 0.12 over 12 months. No statistically significant changes were found for spherical equivalent or endothelial cell density. Other potential applications for cross-linking have been reported. Chan et al. suggests that combining cross-linking with Intacs achieves a greater reduction in cylinder and steep K than by using Intacs alone for treating keratoconus. This is of particular interest to our group. We are planning a prospective study to determine if Intacs and cross-linking are more efficacious than cross-linking alone in young patients with progressive keratoconus. Another report by Hafezi et al. cites 10 eyes treated with cross-linking for ectasia after LASIK with 2 years follow up and no progression of the ectasia. In yet another study Kannelopolous reports on PRK on patients with ectasia after they have been cross-linked with excellent results, suggesting that it might be safe to do PRK on these patients to correct their residual refractive error since the cornea is biomechanically stable after cross-linking. Laboratory studies by Spoerl et al. suggest that cross linking results in increased resistance vs. collagen-digesting enzymes such as collagenase. A later clinical report by Iseli demonstrates its efficacy in a patient treated with a corneal ulcer and in whom corneal melting was arrested. Other potential applications that have been discussed but not reported on include: crosslinking the cornea in place after orthokeratology and pretreating patients with cross-linking prior to PRK in those with keratoconus "suspect" topography. While the procedure appears to be safe and its many potential applications have generated excitement in the ophthalmic community, its efficacy in the treatment of progressive keratoconus needs to be interpreted with a certain level of caution. Currently there is only one prospective randomized clinical trial on which preliminary data on only a small number of patients have been reported. An FDA clinical trial has begun but no data from this study is yet available. All reports to date use a single K reading as an endpoint, which can be highly variable, particularly in keratoconus. The sum of multiple data points averaged over the apex as an initial data point and subsequent data points tracking progression over time will provide more accurate
20

data to show lack of progression. Since keratoconus usually progresses between puberty and age 30, an analysis of patients in this age group who are progressing compared to age-matched controls are needed to demonstrate lack of progression. Patients treated in the older age groups are less likely to progress without treatment; if this group is included in the analysis, there is potential to skew the data in favour of no progression. The potential is there In the United States this procedure is not FDA approved, but clinical trials are currently underway and are likely to generate additional useful data with regard to efficacy and safety. These hopefully will allow for FDA approval in the very near term. In the interim patients will be treated, studies will be performed and long-term data analysis will allow us to evaluate whether this will become a commonly accepted treatment for young patients with keratoconus and ultimately lead to a significant decrease in the number of penetrating keratoplasties performed. Based on currently available data, cross-linking clearly has the potential to become an exciting new treatment for patients with keratoconus. Whether the treatment will be temporary or permanent will become evident as more data on young patients who have been treated in the progressive phase of the disease are tracked over time.

21