[Implementation of a clinical prediction tool in renal transplant recipients]

Kasiske developed a tool for predicting the risk of 5-year graft loss. We analyzed our results using this model.

109 deceased donor kidney transplants were included. 5-year probability of graft survival was calculated during transplantation, seven days after transplantation and 1-year after transplantation. Z-test and ROC curves were used for proportion differences and discrimination ability.

Mean age of donor and recipient was 33.7 and 33.9 years, respectively. 59.6% died due to trauma. Mean of years on dialysis was 3.7. 22.9% of patients had delayed graft function (DGF). Calculated 5-year probability of graft survival during transplantation time was 74.1%; 7 days after transplantation, 74.9%; and one year after transplantation, 76.4%. 5-year death censored graft survival was 64.9%. There were no differences between death-censored graft survival and calculated probabilities (Z-test), with a C-statistic value of 0.54 ± 0.6 (95%CI 0.42-0.65, p = 0.5) and 0.51 ± 0.6 (0.39-0.63, 95% CI, p = 0. 7) for transplant time and seven days after. C-statistic value 1-year after transplantation was 0.68 ± 0.8 (95%CI 0.52-0.84, p = 0.02).

Only calculated 5-year graft survival one year after transplantation had modest prediction ability.