Objectives: Progression of chronic allograft nephropathy (CAN) is associated with a progressive decrease in graft function. Prediction of the Banff CAN grade on the basis of correlation between the grade of histological changes and Scr is difficult because of the big spread of individual values. This study sought to predict the Banff CAN grade based on Scr, Ccr and proteinuria using ROC analysis.
Methods: Graft protocol biopsy and functional testing (Scr, Ccr and proteinuria) were performed in 77 subjects (43 men, 34 women, mean age 48.4 +/- 12.8 years) at 33.8 +/- 1.0 months after their first renal transplantation. Immunosuppression was provided with the triple combination of cyclosporin A, prednisone and azathioprine (or mycophenolate mofetil). Statistical evaluation was performed using receiver-operating curve (ROC) analysis. The cut-off value of the Banff CAN score was set at 1.
Results: The mean values and SD of the investigated functional parameters in study subjects were as follows: Scr = 201.5 (+/- 100.0) mumol/l Ccr = 48.1 (+/- 21.2) ml/min/1.73 m2, proteinuria = 0.89 (+/- 1.96) g/24 h. ROC analysis showed the highest AUC (+/- SEM) for Scr 0.806 (+/- 0.063). The respective values were 0.790 (+/- 0.053) for Ccr and 0.643 (+/- 0.075) for proteinuria. The AUC (area under the ROC curve) for Scr was significantly higher (P < 0.043) compared with proteinuria. The values for sensitivity (specificity) were as follows: Scr 65.0 (91.2). Ccr 75.0 (82.5), proteinuria 60.0 (68.4). The best fit values (best combination of sensitivity and specificity) were 257.2 umol/l for Scr, 33.6 ml/min/1.73 m2 for Ccr and 0.40 g/24 hr for proteinuria.
Conclusions: Our findings support the assumption that Scr > 275 mumol/l and Ccr < 33.6 ml/min/1.73 m2 suggest a Banff CAN grade higher than 1 (P < 0.001). Proteinuria had the lowest predictive values. Values > 0.40 g/24 hr were probably associated with a Banff CAN grade higher than 1 (p < 0.05).