Intended for healthcare professionals


Artificial corneas

BMJ 1999; 318 doi: (Published 27 March 1999) Cite this as: BMJ 1999;318:821

Risks of complications are high now, but better materials are on the way

  1. Bruce Allan, Consultant ophthalmologist, director of biomaterials research
  1. Moorfields Eye Hospital, London EC1V 2PD

    Corneal disease is the second commonest cause of world blindness1 and often occurs in the context of a badly damaged ocular surface. Patients with dry eyes and vascularised corneal scarring resulting from, for example, trachoma or chemical injuries have virtually no prospect of retaining a clear corneal allograft.2 Techniques for implanting a synthetic corneal replacement (keratoprosthesis) offer hope to this desperate group of patients, who are often blind for many years. Understandably, press and public imagination is captured by news of possible restoration of sight in these cases. The pioneering spirit of tackling unmet clinical need has been emphasised in recent newspaper reports,35 but uncertainty surrounding outcomes is barely mentioned.

    Keratoprosthesis techniques reported recently in the British press are not new. Implantation methods and the devices themselves have not materially altered since the 1960s. Devices consist of a porous outer skirt element (Dacron for the Pintucci keratoprosthesis recently implanted in Nottingham and autologous tooth for the Strampelli keratoprosthesis used in Brighton) joined to a rigid polymethylmethacrylate lens. These keratoprostheses are initially implanted beneath the facial skin, to allow the porous skirt to populate with autologous fibroblasts, before being transferred to the eye several weeks later. The surface of the eye is reinforced with an autologous buccal mucosal graft before implantation. This is fenestrated over the lens element to allow a pathway for light to focus on the retina, restoring vision. A wound healing reaction to implantation weaves new collagen through the porous skirt to integrate the keratoprosthesis with the scleral tissue surround.

    Various other keratoprosthesis materials and techniques have been tried in the past or are currently being evaluated.6 Most methods can achieve sustained visual rehabilitation in some patients, but serious complications, including infection, extrusion, retroprosthesis membranes, and secondary glaucoma are common. Because of this, surgery is normally offered only to patients with bilateral corneal blindness who are unsuitable for allografting.

    It is currently very difficult to offer accurate prognostic advice to prospective recipients of keratoprostheses or make an informed choice between different techniques on the basis of published results. Reports are generally of retrospective, uncontrolled case series with conflicting data from different centres. Exclusions are often poorly documented, and true complication rates are difficult to determine. Italian results from 20 patients receiving the Pintucci keratoprosthesis published in 1995 suggest visual improvement in 13 (65%) patients after at least two years' follow up.7 A similar design implanted in the United States, also featuring a Dacron skirt element and a rigid polymethylmethacrylate lens, was abandoned in the 1970s with a serious complication rate of 85% after 10 years' follow up.8 Dacron is known to degrade significantly after implantation,9 and this may explain increased complications after longer follow up. Pintucci's observed complications do not include glaucoma, yet this was a complication in over 50% of cases for the Strampelli keratoprosthesis.10 Both these keratoprostheses require iris and lens removal for implantation, making damage to the trabecular meshwork (the natural fluid outflow pathway from the eye) and glaucoma very likely.

    Prospective data are beginning to emerge from Paris11 and Perth, Australia,6 for a newer generation of keratoprostheses, which use soft materials, resemble the dimensions of the normal cornea more closely, and do not require removal of the iris and lens for implantation. Importantly, the French keratoprosthesis is CE marked, indicating adherence to good manufacturing practice. Complications remain frequent, however. From an initial series of 13 patients in France, five patients had serious complications after 3-9 months.10 Clinical results for the Australian keratoprosthesis are not yet published.

    Driven by the health burden from third world corneal blindness, research into improved keratoprostheses continues to advance. No current device supports an epithelium. Cytokine cross talk between the corneal epithelium and stromal fibroblasts (keratocytes) is important in regulating corneal tissue metabolism.12 If a normal corneal epithelial phenotype can be sustained on a keratoprosthesis, this may protect against tissue necrosis and device extrusion. As an avascular organ, the cornea can be modelled accurately in organ culture and is in many ways an ideal environment for tissue engineering research. Existing laboratory results suggest that the development of an epithelialised corneal replacement may be close. 13 14 Recent experiments showing adult stem cell reprogramming15 indicate that it may also be possible to take autologous stem cells from outside the eye to regenerate the ocular surface.

    Setting these exciting possibilities against a background of uncertainty and generally poor results for existing techniques defines the dilemma for doctors and patients alike. Visual potential is often lost when keratoprosthesis surgery fails. The chance of regaining vision now must be balanced against the possibility that techniques may soon improve. For younger patients in particular, this equation must be carefully considered. For the present, there are strong arguments for concentrating on keratoprostheses that do not require destructive surgery during implantation. International collaboration is required to improve the quality of outcome data.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.
    View Abstract