Several clinical and cell biological features, such as sex, age, leukemic cell burden, morphologic FAB type, and immunophenotype, have prognostic value in childhood acute lymphoblastic leukemia (ALL). The explanation for their prognostic significance is unclear, but might be related to cellular drug resistance. We prospectively studied the relation between the above mentioned features with resistance to 13 drugs in 144 childhood ALL samples obtained at initial diagnosis. The MTT assay was used for drug resistance testing. The interindividual differences in drug resistance were very large and exceeded those between the several subgroups. There was generally no significant relation between sex, leukemic cell burden, and FAB type with drug resistance. However, subgroups with a worse prognosis as defined by age (< 18 months and > 120 months at diagnosis) or immunophenotype (pro-B ALL and T-ALL) did show relatively resistant drug resistance profiles as compared to the subgroups with a better prognosis (age 18-120 months, common and pre-B ALL). Within the group of common and pre-B ALL and compared to the intermediate age-group, samples of the younger children were significantly more resistant to daunorubicin, mitoxantrone and teniposide, and samples of the older children were significantly more resistant to prednisolone and mercaptopurine. Pro-B ALL samples were significantly more resistant to 1-asparaginase and thioguanine, and T-ALL samples were significantly more resistant to prednisolone, dexamethasone, 1-asparaginase, vincristine, vindesine, daunorubicin, doxorubicin, teniposide, and ifosfamide, than the group of common and pre-B ALL cases. We conclude that the prognostic significance of age and immunophenotype in particular may be explained, at least partly, by its relation with resistance to certain drugs. The results of this study may be useful for future rational improvements of chemotherapeutic regimens in childhood ALL.