Since it is unclear whether methotrexate and cytarabine are synergistic or antagonistic in the treatment of acute lymphoblastic leukemia, the Pediatric Oncology Group studied the prognostic significance of a potential interaction between these agents. RBC methotrexate concentrations were compared from 140 patients at lower risk of relapse randomized to two treatment groups: one receiving six methotrexate infusions with overlapping cytarabine; the other, six methotrexate infusions alone. Samples from 248 patients from all risk groups were studied to determine whether patients with extremely low RBC methotrexate concentrations had inferior outcomes. Among low-risk patients studied 3 weeks after the sixth infusion, median RBC methotrexate concentrations were 0.13 nmol/ml RBCs (n = 71) for the methotrexate-only group and 0.02 nmol/ml RBCs (n = 69) for the methotrexate/cytarabine-treated low-risk patients, P < 0.001 by the two-sided Wilcoxon test. For low- and high-risk patients receiving methotrexate/cytarabine infusions, event-free survival at 1 and 3 years after RBC sampling was 97 +/- 2% and 90 +/- 3% for patients with concentrations greater than the median, and 88 +/- 3% and 78 +/- 4% for those with concentrations at or below the median. Log rank comparisons of event-free survival in the first year and overall yielded P = 0.005 and P = 0.04, respectively. Cytarabine altered methotrexate pharmacology when the drugs were infused together. Patients whose levels were extremely low had an adverse prognosis. Although this study could not assess efficacy of the methotrexate/cytarabine combination, it appears that concurrent administration is not optimal.