Between January 1974 and November 1978, 41 consecutive increased risk (age less than 24 months or less than 120 months, or leukocyte count greater than 30,000/mm3) patients with acute lymphoblastic leukemia (ALL) were entered on two consecutive treatment protocols which employed intermediate dose methotrexate (IDM). IDM was employed for central nervous system prophylaxis and systemic intensification. It was anticipated that the avoidance of prophylactic cranial irradiation would result in a lower incidence of longterm central nervous system sequelae. Twenty-two children and adolescents were entered on the first study (IDM X 3) which employed three courses of IDM (500 mg/m2) and six doses of intrathecal (IT) methotrexate (MTX). Nineteen children and adolescents were entered on the second study (IDM X 6) which employed six courses of IDM (3-500 mg/m2 and 3-1500 mg/m2), six doses of IT MTX and three additional doses of triple IT chemotherapy (MTX, cytosine arabinoside, and hydrocortisone or dexamethasone). The cumulative percentage of patients who remained in continuous complete remission was 30% for IDM X 3 and 57% for IDM X 6. This difference was statistically significant (P = 0.046; and BM + CNS, I-using IDM X 6. The cumulative incidence of primary CNS relapse was 36.4% for IDM X 3 and 29.9% for IDM X 6. This difference was not statistically significant (P = 0.44). The use of more intensive therapy with IDM and triple IT chemotherapy did improve the duration of continuous, complete remission but did not decrease the incidence of primary CNS relapse in increased risk patients.