We automated two procedures for determination of urinary N-acetyl-beta-D-glucosaminidase (NAG; EC 3.2.1.30) concentrations and evaluated their reliability for detecting drug-induced tubular damage in children receiving cisplatin, methotrexate, or ifosfamide. Results for 174 patient specimens correlated well (r = 0.98), but NAG concentrations determined by the m-cresolsulfonphthaleinyl (MCP) procedure were about 40% lower than those obtained with p-nitrophenyl-N-acetyl-beta-D-glucosaminide substrate. Dialysis and assay of 50 specimens disclosed no evidence of activators or inhibitors of enzymatic activity. Drugs and metabolites added to urine had negligible effect on NAG determinations; however, NAG was unstable in alkaline urine (pH greater than 8) associated with methotrexate therapy. Both procedures detect tubular damage equally well and neither requires laborious sample treatment. The MCP procedure, being more sensitive and not requiring a sample blank, is better suited for rapid automated assays. Comparisons of clinical data obtained by the two procedures require standardization against human NAG.