The ideal keratoprosthesis (KPro) should have all the advantages of but none of the problems associated with allografting. It should have specifiable aspheric optical variables, it should block ultraviolet radiation, and it should allow full visual field. The optical zone should be sufficiently rigid to avoid optical aberrations and astigmatism from buckling but sufficiently elastic to allow measurement of intraocular pressure by applanation.1 Full wound healing (biointegration—a form of “biocompatibility”)2 3 should take place at least at the periphery of the artificial cornea, allowing defence against intraocular infection, epithelial downgrowth, as well as eye rubbing and minor trauma. Artificial materials used should be non-toxic and not degrade in the lifetime of the patient.

It should allow host corneal epithelium to grow over its anterior surface and to adhere to it thus making a wettable and self renewing surface; therefore, proteinaceous materials would not be deposited onto its anterior surface and giant papillary conjunctivitis would not develop. Penetration by topical medication is important as steroids, pupil dilating and constricting drops, as well as antiglaucoma drops are used frequently. The formation of a retroprosthetic membrane has been a major cause of failure of many KPros, so its posterior surface should be highly polished and non-sticky. The prosthesis should be flush with the rest of the ocular surface to enhance comfort and to reduce mechanical shearing forces on it. It should be sufficiently soft for suture needles to pass though but strong enough so that suture materials do not cheesewire. Last but not least, the prosthesis should be inexpensive.

This is a tall order indeed, but if met, would solve the problem of worldwide shortage of corneal donors, as well as the two major complications of allografting—graft astigmatism and rejection. How close are we to this? The related field of contact lenses, where safe extended wear has proved to be a tantalisingly difficult goal that has not yet been achieved despite considerable commercial investment, suggests just how complex the problem is likely to be.

In this issue, Hicks et al (p 18) demonstrate to us the need for a multidisciplinary approach to the development of such a biomedical device. Previous workers have used a range of biomaterials incorporated into a range of design features. The common feature that characterises almost all these attempts, however, is the fact that the materials selected have been chosen because of their availability and their susceptibility to fabrication techniques appropriate for the specific KPro design employed, rather than for their ability to interact appropriately with the specific biological environment in question. One significant feature that sets the development work on the Chirila prosthesis apart from other published work in this area is the fact that in its initial stages it sought to make use of a polymer that is significantly more hydrophilic than the generality of those previously employed in KPro design—namely, poly(2-hydroxyethyl methacrylate) or poly-HEMA. Then, having elegantly harnessed established principles of behaviour of the material in the formation of a two part prosthesis with porous skirt and clear core,4the group has recognised that the literature on cellular interaction with hydrogels predicts that some modification to the chemical structure of the porous skirt will inevitably be necessary in order to manipulate and optimise biological integration.5 This paper contains the first report of the initial exploration of the effect of small changes in the chemical structure of the prosthesis. It represents a significant stage in what will inevitably be a long road towards the achievement of a KPro whose chemical and physical features produce an acceptable level of biocompatibility to support long term success.

In terms of long term retention results, we must not forget the osteo-odonto-keratoprosthesis (OOKP) which has been in existence for over 30 years.6-8 Following modifications by Falcinelli, the modern technique has a track record of close to 20 years. Well over 90% of Falcinelli OOKPs are retained in the long term and three quarters of patients achieve 6/12 vision or better.9-11OOKP is not without its problems—its complexity, relatively poor visual field (30°–40°), and having to sacrifice oral structures. However, it can withstand a hostile ocular environment and future generations of KPros will inevitably be judged against the grandfather of KPros when they come to clinical use.