The diagnosis of chronic allograft nephropathy (CAN) is based on pathological examination according to Banff 97 schema. The aim of the study was to evaluate the usefulness of tubular and glomerular proteinuria for noninvasive recognition of CAN. One hundred and thirty renal allograft recipients (at least 90 days after transplantation) who had undergone diagnostic allograft biopsy were included in the study. Beta2-microglobulin, alpha1-microglobulin, albumin, immunoglobulin G, total protein, and creatinine concentrations were obtained from the second morning urine specimen. Raw data and values calculated per 1 g of creatinine excreted in urine along with time after transplantation, serum creatinine, and its change over a period of 2 months prior to biopsy were taken for analysis. Urine proteins were measured using a nephelometric method. Statistical calculations were performed using MANOVA and stepwise discriminant analysis (SDA). Statistical diagnosis and staging of CAN matched the pathological method in 68% of a preliminary SDA. Therefore patients were divided into normoalbuminuric, microalbuminuric, and macroalbuminuric groups. There was no significant differences between protein excretion, except alpha1-microglobulinuria (CAN 0 vs 2, P = .018; CAN 1 vs 2, P = .041), beta2-microglobulinuria (CAN 0 vs 2, P = .026; CAN 1 vs 2, P = .0033), and total proteinuria (CAN 0 vs 2, P = .042) in the normoalbuminuric group. Nevertheless, diagnoses obtained using SDA were 89%, 91%, and 92% identical to the results of pathological examinations, for normoalbuminuric, microalbuminuric, and macroalbuminuric groups, respectively. In conclusion, tubular and glomerular proteinuria measurements may be useful for a noninvasive CAN diagnosis and staging only with regard to degree of urinary albumin excretion.