We evaluated the ability of three enzymes--N-acetyl-beta-D-glucosaminidase (NAG; EC 3.2.1.30), alanine aminopeptidase (AAP; microsomal aminopeptidase, EC 3.4.11.2), and gamma-glutamyltransferase (GGT; EC 2.3.2.2)--and adenosine deaminase binding protein (ABP) in urine to predict or confirm renal-transplant rejection in patients treated with cyclosporine. We measured the enzymes daily during the early post-transplant hospital stay of 104 renal-transplant recipients (72 men and 32 women). We also measured ABP in 32 of these patients. We analyzed the data by calculating the activity ratio of each day's test value to the previous day's result and optimized the sensitivity (SN) and specificity (SP) to determine the optimal ratio for each test. The results indicate that cyclosporine treatment reduces the optimal sensitivity and specificity of these tests. Three comparable tests (ABP, GGT, and AAP) yield the best optimal values (SN = 0.77, 0.69, 0.77; and SP = 0.71, 0.74, 0.63, respectively), and the NAG test yields the lowest combination of sensitivity and specificity (SN = 0.62, SP = 0.66). All four tests were less sensitive and specific than the plasma creatinine test (optimal day-to-day difference = 5 mg/L). However, the ABP and AAP tests gave indications of rejection at least 24 h before clinical diagnosis for 50% of the patients experiencing rejection, while early plasma creatinine increases of 5 mg/L occurred in only 19% of this group.