Why should we attempt to develop keratoprostheses (KPro) for corneal opacity when standard corneal transplantation is so well established and relatively safe? If the latter procedure were uniformly successful, there certainly would be no need for a KPro, with its even higher risks and greater demands on health resources. The answer lies in the stubborn failure rate of keratoplasty, in spite of a century of development. The outcome of keratoplasty varies markedly with underlying disease, geography, availability of donor tissue, health budget, and other factors, and it is still poorly documented. Some statistics are beginning to emerge, however, even if they are not population based but instead drawn from individual surgeons’ office practices—which automatically excludes the “inoperable” cases. For instance, the Eye Bank Association of America’s report for 2002 indicates that of a total of 33,000 transplants in the United States, more than 13% were done for failed grafts (1). (It is not known how many graft failures were not reoperated.) In a large outcome study, only 20% of regrafts for all causes remained clear for 5 years (2). In another group of first regrafts, again only 20% survived 5 years, whereas all repeat regrafts failed within the same time period (3). Moreover, on a world-wide basis, a large number of people with corneal blindness have no access to corneal surgery. The World Health Organization reports that approximately 6 million people are bilaterally blind from trachoma alone (4). Few of these patients receive a corneal transplant, and when it is performed, the failure rate is reportedly high.

The fact that standard keratoplasties have failed or have been deemed hopeless in such a large cohort of corneal patients does not automatically mean that a KPro is the immediate next step. In spite of efforts by fearless surgeons over 200 years, progress in developing long-lasting and safe KPro techniques has been slow. Creation of meaningful animal models has been difficult, and the long delay in the development of postoperative complications, many of them catastrophic, has made changes in technique and design slow in coming and hard to interpret. In spite of these impediments, however, very substantial progress has been made, especially during the past few decades. Rapid rehabilitation and sometimes spectacular vision in KPro eyes after corneal graft failures is now no longer a rarity. The long history of KPro development will not be detailed here—the reader is referred to earlier reviews (5,6).

The current KPro effort is maintained primarily in approximately a dozen centers worldwide and encompasses diverse approaches. One technique has been inspired by Strampelli (7), and a number of followers have modified the original principles further [Falcinelli et al. (8), Marchi et al. (9), Temprano (10), Grabner et al. (11), Liu et al. (12), Hille (13), and others]. This technique involves harvesting a tooth from the patient and preparing a slice of osteodental lamina to be used as a skirt for a cylinder of polymethylmethacrylate (PMMA), which is to be inserted (in two steps) into the patient’s cornea and covered by lid skin or buccal mucosal graft. The procedure is somewhat invasive but it has a reputation for stability and low rate of infection. Pintucci et al. (14), Girard et al. (15), Legeais et al. (16), and others have replaced the autologous tooth-derived skirt with “biocolonizable” porous plastic materials. In Russia, a large number of patients, especially with chemical burns, have been implanted with devices of different designs, but the optical core is still being made of PMMA (17,18). Yet another approach has been to develop a hydrogel sheet with porous edges to serve as an “artificial cornea.” It is implanted intrastromally and covered by a conjunctival flap. The center of the device is exposed later. This technique appears to be of value for graft failures in less inflamed eyes (19). Many smaller-scale efforts, some ingenious, are also currently under way (20).

In Boston, we have for a number of years used a PMMA KPro of a double-plated “collar button” design (21). Such a general configuration has been suggested in the past (22,23), and has been modified by us as described in the following section. Design and material of KPros are undoubtedly important in themselves, but it should be emphasized that the health of the surrounding tissue is vitally important. Without claiming superiority, we devote this chapter to the devices, surgical techniques, and follow-up routines that we have gradually developed.