The prognostic value of systemic methotrexate clearance (ClMTX) during high-dose therapy was evaluated in a cohort of 42 children with acute lymphocytic leukemia (ALL). As part of an extensive chemotherapy protocol, they had received a total of 293 methotrexate (MTX) infusions in the 6-8 g/m2 dose range. At the termination of the study, when they had all been followed up for 3.5 years or more, 26 of these patients were still in continuous complete remission, whereas 16 had suffered relapse. The intrapatient variability in ClMTX during the eight courses was up to six-fold. In 67% of the patients, the maximum level of ClMTX reached at least twice the minimum value. The coefficients of variation for the intra- and interindividual variability in ClMTX were 9-57% and 26-41%, respectively. The cumulative probability of relapse, estimated by the Kaplan-Meier procedure, was increased for patients with a high ClMTX during the initial treatment course, but the difference was not significant on a 5% level. There was no significant relationship between high individual median ClMTX and subsequent relapse of ALL. However, ClMTX during the initial infusion, the time-dependent mean for ClMTX, and the individual patient's median ClMTX, were significant predictors for event-free survival in a Cox proportional hazards regression analysis. The present study demonstrates gross pharmacokinetic variability and unpredictable values of ClMTX in subsequent courses after standardized administration of MTX to paediatric patients with ALL. In spite of the association between ClMTX and prognosis shown by some of the analyses, estimates of ClMTX rates may not necessarily be related to disease outcome in a way that can be exploited to the benefit of the individual patient.