We studied the disposition pharmacokinetics of methotrexate (MTX) given orally to 16 children with acute lymphoblastic leukemia (ALL) and its relation to the pharmacokinetics of 6-mercaptopurine (6MP) in the same children. There was an eightfold variability in area-under-concentration time-curve (AUC) of MTX achieved by the same dose. Excellent correlation existed between peak concentrations and AUC0----infinity (r = 0.95, P less than 0.001). Elimination T1/2 was between 1.34 and 5 hours (mean 2.16 +/- 0.23 hr, mean +/- SE). A weak correlation existed between AUC achieved by 1 mg/m2 MTX and patients' age or body weight. Weak but significant correlation existed between AUC achieved by 1 mg/m2 of MTX vs. 6MP (r = 0.54, P less than 0.05). In 13/16 patients peak concentrations were achieved at 60 minutes. There was a significantly larger AUC of 6MP achieved by a standardized dose in longer therapy (greater than 15 mo) vs. short therapy (less than 12 mo) (462 +/- 75 and 246 +/- 58 ng.ml-1.min.mg-1.m2, P less than 0.025). No statistical differences in AUC of MTX were found between short and long therapy. The large interpatient variability in MTX pharmacokinetics supports the possibility that differences in absorption and/or clearance of the drug may affect the clinical response. Because of the excellent correlation between peak and AUC of MTX, and because 3 measurements, at 30, 60, and 90 minutes will almost invariably identify the peak, this measurement can be used to estimate AUC for purpose of correlation with clinical outcome.