ABSTRACT
Immunologic rejection of the transplanted cornea is the major cause of human allograft failure with several risk factors contributing to it. Since in the corneal graft, most individuals do not reject the graft, we used the survival analysis with cure rate for the assessment of the factors influencing the cure rate at the time of data analysis. The main aim of this study was to evaluate the cure rate and assess the risk factors for corneal graft rejection in the keratoconus disease in Labafinejad Hospital, Tehran, Iran. This was a routine data base study in which the data were gathered from keratoconus patients' files that had undergone penetrating keratoplasty operation. In the survival analysis, individuals who didn't reject corneal were considered cured. To study the factors influencing the cure rate, we used the Weibull distribution for survival function and the logistic link function for the cure rate because of their tractability and accuracy. Out of 119 patients 31 patients [26%] rejected grafts. Among the factors influencing cure rate, only in vascularization and in persons older than 25 years of age was ameaningful effect on decreasing cure rate. With this cure model, the expected cure rate in the non-vascularization and less than 25 year- old patients was 81, in non-vascularization and more than 25 year- olds it is 64, in the vascularization and less than 25 year- olds, the cure rate is 19 and in the vascularization and more than 25 years of age, the cure rate is 9 percent and the observed cure rate for Kaplan-Meier product limit estimator was 79, 61, 27 and 0 percent, respectively. The results showed that the estimate of cure rate in the survival analysis was near the Kaplan-Meier product-limits estimator One of the benefits of modeling is its ability to generalize the results; using them in the prediction. According to the results obtained from the fitting cure model the vascularization is a major risk factor for decreasing cure rate in the keratoconus disease. We could use the cure rate model in the analysis of survival data with cure fraction