Current Developments in Keratoconus Research

New findings shed light on a possible treatment for keratoconus and on the histopathology of the disease


Current Developments in Keratoconus Research

New findings shed light on a possible treatment for keratoconus and on the histopathology of the disease.

By Jan P.G. Bergmanson, OD, PhD, PhD h.c, DSc, & Jessica H. Mathew, OD

Dr. Bergmanson is a professor of Optometry at the University of Houston College of Optometry and founding director of the Texas Eye Research and Technology Center. He is a Foundation Fellow of the College of Optometry in the United Kingdom and a Diplomate in the Cornea and Contact Lens Section of the American Academy of Optometry.

Dr. Mathew is currently working toward a PhD degree with a focus on the histopathology of keratoconus. She is a two-time recipient of the William C. Ezell Fellowship (2007–2009) and the NIH Loan Repayment Grant (2007–2010).

Keratoconus is an elusive — if not perplexing — corneal disease. We have known about this disease for more than 150 years. Early descriptions appeared in the literature in the mid-1800s (Nottingham, 1854). Yet we do not have a cure, we do not understand its cause, and we do not know exactly where it all starts. Does a contact lens cause the scarring we see in our keratoconus patients? Does unilateral keratoconus really exist? Which transplant surgery, when indicated, is the best option? These are some of the many questions that are begging for answers and, indeed, are often posed by keratoconus patients.

A few new developments in the field of keratoconus research have been reported recently, and these may have significance to clinical practice. The purpose of the present article is to discuss this new information, which incidentally will address some of the above questions.

Corneal Stromal Crosslinking

A few years ago the combined corneal exposure to a riboflavin solution and ultraviolet radiation (UVR) was launched in Europe as a treatment for keratoconus. This procedure, utilizing ultraviolet A (UVA) at 370nm wavelength, is currently under investigation in a Food and Drug Administration (FDA) clinical trial. A number of clinics are gearing up to meet the potential public demand. It is important for primary care eyecare professionals to know where we are in this promising but not fully explored new field, so that we may keep our patients informed without raising hopes that later may be dashed.

The principal goal of the riboflavin-UVR procedure is to use irradiation, possibly in association with a photosensitizer (riboflavin), to achieve crosslinking of the stroma. The collagen crosslinking is intended to create a stiffening of the cornea, thereby counteracting the tendency for ectasia of the tissue. To have the desired effect from riboflavin, corneal debridement is essential. Other methods such as topical application of anesthetic to inhibit the epithelial tight junctions will not permit sufficient quantities of riboflavin to flow into the cornea (Hayes et al, 2008).

As mentioned earlier, the wavelength chosen for the irradiation is 370nm, which is within the UVA band. UVA is the longer wavelength UVR and is not believed to be as toxic as UVB. However, the cornea is 90 percent transparent to the wavelength of 370nm, which means that rays of this wavelength will pass right through the cornea and be almost completely absorbed by the crystalline lens (Bergmanson 2009) (Figure 1).

Figure 1. Ocular UVR transmittance. Published with permission from Walsh, Bergmanson, Harmey in: Bergmanson JPG. Clinical Ocular Anatomy and Physiology, 16th edition, 2009. ISBN-13: 978-0-9800708-1-1.

This is where 0.1% riboflavin in dextran T500 20% solution enters the picture. Riboflavin is an absorber of the 370nm wavelength and when presented topically to the cornea, it will facilitate the necessary absorption to accomplish the desired collagen crosslinking. Incidentally, riboflavin may have a dual function in this treatment modality. In addition to its absorption role, riboflavin may also serve as a photosensitizer, producing oxygen-free radicals participating in the crosslinking process.

Is the UVA Dosage Safe? Despite the absorptive properties of the riboflavin, its presence will not make the cornea opaque to UVA of 370nm. As a consequence, perhaps 50 percent of the UVA will reach and be absorbed by the crystalline lens. This is problematic because it may be a suprathreshold dose resulting in cataract formation that may take years to develop. Scientific data detailing the toxicity or nontoxicity of this irradiation is essential.

There are a few questions that need to be resolved. The dose selected is based on a 3mW/cm2 energy level delivered for 30 minutes, which translates into a dose of 5.4J/cm2 (Wollensak et al, 2003; Coskunseven et al, 2009; Vinciguerra et al, 2009). This is 30 times greater than what the sun emits at this wavelength. Compare this to the corneal trauma threshold for more toxic UVB (300nm), which is 0.08J/cm2 (Pitts et al, 1987) or a 68-times weaker dose than 5.4J/cm2. A histological study on in-vivo exposed rabbits showed that UVA irradiation in this wavelength and dose caused a complete endothelial and keratocyte loss in the exposed zone (Wollensak et al, 2007). The rabbit cornea may have had less than 400μm stromal thickness, but this study clearly demonstrated the corneal toxicity of this dose of 370nm wavelength radiation.

Of course UVB is more toxic than UVA, but unfortunately, we do not have reliable data on UVA toxicity. Therefore, there are questions regarding the safety of this procedure. To counter the possibility of UVA leaking through the cornea, the general rule applied is that the cornea must have a stromal thickness of 400μm, which will itself contraindicate this procedure for a substantial proportion of keratoconic corneas.

How Effective is the Procedure? Clinical results from studies on this procedure are reporting short-term results, although some patients have been followed for up to four years (Tables 1 and 2). We will have to wait for long-term studies of much longer duration than this.

What have we learned so far? A small decrease in myopia and a modest flattening of curvature appears to occur consistently. However, for a −8.00D myope who has a K reading of 52.00D, one or two diopters of myopia reduction or corneal flattening will not have much impact. The initial reports indicate the effect of crosslinking to be stable. If this turns out to not be the case and the increased corneal rigidity is gradually lost, re-exposures may be necessary.

In conclusion, the riboflavin-UVA crosslinking treatment of keratoconus shows some promise, but also raises some significant concerns. At this point it does appear to arrest progression of the disease, but does not offer a meaningful improvement of the condition. Nevertheless, this is an option our keratoconus patients should have if efficacy and especially safety can be proven. Should this therapy become a viable and popular procedure, the question will arise as to whether the combined delivery of a radiation wavelength (370nm) used in suntanning parlors and a vitamin B2 topical formulation meets the definition of surgery. That should be an interesting debate.

Figure 2. Light micrograph showing the loss of the anterior limiting lamina (ALL) over a wide area. An arrow indicates where this layer is lost. The removal of ALL typically leads to epithelial thickness variations that are also evident in this view. Magnification approximately X200.

Keratoconus Histopathology

Surprisingly, the literature is devoid of systematic, morphometric descriptions of keratoconus at the microscopic level. Most available histopathological reports are either simple case reports or they lack morphometry. Histopathology utilizing transmission electron microscopy allows an observer to view the abnormal tissue at a cellular and subcellular level. When this methodology is employed in combination with careful measurements on multiple samples, we stand to learn important lessons about a disease. We have pursued this hitherto unexplored avenue and made useful progress, on which we can here report. This new understanding can be applied directly to clinical practice.

Our work has clearly shown that at least initially, keratoconus is a corneal disease with an anterior corneal focus. However, eventually the entire cornea becomes involved. Although the literature describes cracks, breaks, and ruptures of the anterior limiting lamina (ALL) (Leibowitz and Waring, 1998; Rabinowitz, 2005), our research has revealed a far more extensive involvement of the ALL in the pathophysiology of keratoconus. Centrally in the cone, more than 50 percent of the ALL is affected, and over expansive areas it is completely lost (Figure 2) (Horne et al, 2007).

The ALL is the foundation on which the epithelium is attached. This explains why keratoconus patients more often than not have epithelial defects that will stain with fluorescein. In addition to this foundation dilemma is the epithelial involvement in this corneal pathology — indeed, it was proposed about 50 years ago that keratoconus starts in the epithelium (Teng, 1963). We are not ready to say that, but it certainly is part of the pathology. The presence of unhealthy, degenerative cells in the epithelium can only add to the fragility of the ocular surface in keratoconus patients. This is why we advocate apical clearance when possible in fitting GP contact lenses on keratoconic corneas. A scleral lens that typically vaults the cone region is another good way to manage these patients. In this manner we reduce the mechanical stress on an already weakened, pathological area.

The actual removal of the entire ALL corneal layer is accomplished with the aid of cells recruited from areas beyond the cornea (Figure 3) (Mathew et al, 2009). What these cells are, where they come from and what exactly they are doing is the focus for future studies in our laboratory.

Figure 3. Transmission electron micrograph of a non-corneal cell (triangle) in the anterior stroma. This undocumented corneal visitor has an intimate relationship with a keratocyte (arrow).

When the removal is complete and the ALL is gone, scar tissue is frequently laid down to serve as an interface between the stroma and the epithelium. This is our explanation for the anterior, subepithelial scarring so typically observed in keratoconus. It is our opinion that this corneal scarring has nothing to do with contact lens wear but is a result of the disease process. Again, it is advisable to try to stay clear of this part of the cornea in your contact lens fitting. After all, we do not wish to put further stress on this already severely challenged corneal region.

Because most of the events in keratoconus occur anteriorly in the cornea (Horne et al, 2007), it may be advantageous for a patient requiring transplant surgery to consider lamellar keratoplasty (LKP). Currently, penetrating keratoplasty (PKP) appears to be the standard procedure for keratoconus. A recent Texas Eye Research and Technology Center study demonstrated no visual advantage of PKP over LKP (Nielson et al, 2009). But LKP surgery offers other important advantages in that rejection is a highly unlikely event, global strength is maintained, and healing time is reduced.

Looking Ahead

Currently, we are also researching lamellar organization and are close to explaining how ectasia occurs. We know that the disease process causes a loss of ALL and the anterior interweaving stromal lamellae, all of which contribute to corneal thinning. However, even with these losses, we have reported a surprising increase in the overall number of lamellae found in the keratoconic cornea (Mathew et al, 2009). The reason for this increase is that the lamellae have broken down into smaller units. This discovery of the unraveling of stromal lamellae could be the key missing factor that explains ectasia in keratoconus and post-refractive surgery, which cannot be explained by simple thinning alone. This lamellar fragmentation possibly leads to a general weakening of the corneal infrastructure, provoking ectasia. CLS

For references, please visit and click on document #166.