A group of 43 pediatric patients with standard risk acute lymphoblastic leukemia (ALL) was studied prospectively and treated with a protocol that included adriamycin, vincristine and prednisone (HOP) to induce remission; cranial irradiation and intrathecal methotrexate (MTX) as CNS prophylaxis and mercaptopurine and MTX together with pulses of HOP every three months to maintain remission. Complete remission (CR) was achieved in 95.3% of the group; 5-year survival was 67%. The following variables were analyzed in the outcome to treatment: Age, sex, WBC at diagnosis, FAB morphology, CALLA/CD10 reactivity of the blast cells, lymph node, liver or spleen enlargement, site of treatment (private practice versus city hospital) and malnutrition. None of these variables had a significant impact in survival, but malnutrition. Under-nourished children (UNC) n = 16, had a significant worse outcome than well-nourished children (WNC) n = 27. Although CR was achieved in 98% of WNC versus 94% of UNC, five-year survival was 83% for WNC and 26% for UNC (p less than .001); relapses were observed in 18% of WNC and 75% of UNC (p less than .0005). Relapses presented more frequently in the bone marrow in UNC than in WNC (56% versus 7% p less than .0001). The doses of maintenance chemotherapy had to be reduced in 68% of UNC and 10% of WNC (p less than .005). The poor outcome to treatment observed in UNC was due to systemic relapses, apparently related to a poor tolerance to maintenance chemotherapy. Malnutrition might be included as an adverse prognostic factor in the outcome to treatment of children with ALL, in developing countries.