Corneal transplantation has evolved toward selective tissue transplantation over the last decade, with many corneal diseases being treated by a variety of anterior and posterior lamellar graft procedures. Surgeons have adopted these new techniques rapidly because of distinct advantages over penetrating keratoplasty (PK). Simultaneously, the quest for a tissue-engineered corneal substitute has made rapid progress, and implantation of corneal substitutes may become the future of corneal surgery. Despite the exciting progress in corneal surgery, PK remains the most common form of corneal transplantation in the world at present, and although PK may diminish in importance in the future, it will remain the default or procedure of choice for some eyes.

The success of PK generally has been measured by the longevity of graft survival, which is the time to graft failure. A wealth of graft survival and outcomes data have been amassed over the last 3 decades from sources such as individual surgeon series, randomized clinical trials, and corneal graft registries. The Singapore Corneal Transplant Study (SCTS) is one such source of data and is unique because it reflects the indications for and outcomes of PK in an Asian population. Since 1991, all corneal transplantation procedures at the Singapore National Eye Center, which performs 80% to 90% of all corneal transplants in Singapore, have been entered into the SCTS, and postoperative outcomes have been recorded annually. During the first 13 years of the study, there were 1130 PKs, and the initial outcomes of 901 PKs in 901 eyes were reported. Of these, 87% were for optical (visual rehabilitation) indications, whereas 13% were for therapeutic (surgical treatment of infectious disease) and tectonic (preservation of globe integrity) indications. The graft survival rate was highest for optical indications (64% at 5 years and 52% at 10 years) and was lower for therapeutic (37% at 5 years) and tectonic (42% at 3 years) indications. At the time of the first report of the SCTS, 31% of penetrating grafts were known to have failed, and half of these were because of endothelial graft rejection and late endothelial failure. The initial results of optical PK in the SCTS are comparable with those of other large series outside of Asia (there are limited data for outcomes of PK in Asia ) and are a testament to successful corneal transplantation and eye banking in Asia. Not surprisingly, the proportion of therapeutic and tectonic PKs in Asia, accounting for 12% to 13% of all PKs, is much higher than that outside of Asia.

In this issue, Anshu and associates report the postoperative factors that influenced graft survival after PK in the SCTS. Graft survival was excellent (85% at 10 years) in the absence of any postoperative risk factor, but the presence of one postoperative risk factor, including infection, rejection, disease recurrence, eyelid or glaucoma surgery, or a repeat graft, resulted in lower graft survival (34% at 10 years). Anshu and associates also performed a multivariate analysis of preoperative, intraoperative, and postoperative risk factors in the SCTS and found that the postoperative factors listed above were indeed independent risk factors for graft failure, and identified recipient diagnosis, recipient gender, preoperative inflammation, and donor diameter as additional independent risk factors. Thus, although the preoperative risk factors for graft failure cannot be modified, the prevention or early detection of postoperative risk factors for graft failure may be important for prolonging graft survival.

Many of the risk factors for graft survival in the SCTS agree with those identified in other large series. After considering the data from the SCTS and from the other series, several common risk factors become apparent and can be divided into recipient, donor, and postoperative risk factors. Recipient diagnosis is the most important preoperative risk factor, with grafts for keratoconus having the best survival rate, and grafts for endothelial disease, corneal opacity, and ipsilateral graft failure having a worse survival rate. Recipients in their first decade of life have poor graft survival, and survival also is worsened by active or previous intraocular inflammation or ocular hypertension. With the exception of donor diameter, donor and preservation factors seem to have little influence on graft survival, presumably because donors are screened and selected by eye banks for tissue quality. After surgery, allograft rejection consistently has been found to be a leading cause of graft failure ; disease recurrence and infection also are significant postoperative risk factors, and they may have more importance in Asia because of the indications for PK in Asia. There have been conflicting results for corneal neovascularization, glaucoma, donor age, and preoperative donor endothelial cell density as risk factors for graft survival. Discrepancies between studies may arise because of differences in study design, data collection methods, and definitions of outcomes, and because of variations in the populations, inclusion criteria, and postoperative immunosuppression regimens. In addition, studies may be biased by loss to follow-up, which is inevitable with all long-term graft survival studies. In the SCTS, the number grafts analyzed at each follow-up examination was not reported, although attempts were made to recall as many patients as possible.

Considering data from numerous studies over the last 3 decades, graft survival has remained the primary outcome measure of success. However, when assessing the success of corneal transplantation, graft survival should be interpreted with caution and in the context of the study goals. First, there is no consensus for the definition of graft failure; some studies define this as the loss of optical clarity, whereas other studies specify the loss of optical clarity in terms of a variably defined effect on vision. Second, although loss of optical clarity often is considered to be a well-defined end point, there may be subjective variability in studies with multiple observers, and the gradual loss of clarity with late endothelial failure, which is a leading cause of graft failure, could give an uncertain end point. Third, loss of optical clarity does not always convey success, depending on the broad indication (optical, therapeutic, or tectonic) for transplantation. For example, a cloudy graft after a therapeutic PK may be considered successful if all active infection were eradicated; in the SCTS, 13% of PKs were for therapeutic or tectonic indications, yet outcomes other than the time to loss of graft clarity were not reported to indicate whether these grafts served their initial purpose. Similarly, the optical clarity of grafts for optical indications does not directly convey success at restoring functional vision, or perhaps more broadly, quality of life, to the patient. Although visual outcomes after corneal transplantation are important, they can be confounded by other causes of decreased vision and, after PK, are subject to the ability to provide the best refractive correction. Nevertheless, as corneal transplantation techniques continue to evolve, especially for selective endothelial replacement, vision and quality of life outcomes may become more useful determinants of graft success.

Anshu and associates are to be congratulated for their valuable contribution of the risk factors for PK graft survival in an Asian population, which are pertinent to the treatment of corneal blindness in the developing world. As the Singapore Corneal Transplant Study continues to mature, we can anticipate additional relevant outcomes data for various forms of corneal transplantation.