Hyperleukocytosis (greater than or equal to 100 x 10(9) leukocytes/L) was identified at diagnosis of acute lymphoblastic leukemia in 64 of 358 patients enrolled on St Jude Total Therapy Study XI from February 1984 to September 1988. These children received a seven-drug induction regimen followed by high-dose methotrexate, cranial irradiation at 1 year of remission, and 120 weeks of continuation therapy with rotational administration of four drug pairs. The 27 patients with leukocyte counts greater than or equal to 200 x 10(9)/L underwent initial cytoreduction via leukapheresis or exchange transfusions. The complete remission rate for patients with hyperleukocytosis (94%) was similar to that for the overall series (96%). Stepwise regression analysis showed that hyperleukocytosis was significantly associated with age less than 1 year at diagnosis, T-cell immunophenotype, leukemic cell ploidy less than or equal to 50 chromosomes, organomegaly, and elevated lactic dehydrogenase. The 27 patients with extreme hyperleukocytosis (greater than 200 x 10(9)/L) different from the other 37 children only in a higher frequency of French-American- British (FAB) L2 morphology. Estimated 4-year event-free survival (EFS) was 52% +/- 8% (SE) for patients with hyperleukocytosis versus 79% +/- 4% for patients with leukemic counts less than 100 x 10(9)/L (P less than .0001). Patients with leukocyte counts of 100 to 200 x 10(9)/L had a significantly better EFS than those with counts greater than 200 x 10(9)/L (64% +/- 10% v 34% +/- 14%; P = .04). Thus, the therapy in this trial proved satisfactory for children with leukocyte counts of 100 to 200 x 10(9)/L; further study is needed to improve the outlook for children with counts greater than 200 x 10(9)/L.