This randomised study was performed to assess the anti-emetic efficacy and tolerability of two-dose regimens of granisetron in children with leukaemia. 49 children with leukaemia were treated with three consecutive courses of high-dose methotrexate or cytarabine regimen. During the first course, patients were evaluated regarding the emetogenicity of each regimen. They were randomised in a crossover manner to receive 20 or 40 micrograms/kg of granisetron before the second and third course of chemotherapy. Neither emesis nor severe appetite loss were observed in over 80% of patients within the first 24 h in all treatment groups. There was no significant difference in the anti-emetic efficacy between the two-dose regimens of granisetron. However, complete protection was achieved less frequently on days 2 and 3. Older children and girls appeared to be less well protected. No adverse events attributable to granisetron were observed. Granisetron dose regimens of 20 and 40 micrograms/kg are, comparably, well tolerated and effective in controlling chemotherapy-induced emesis in the first 24 h, though this protection fails thereafter, particularly in older patients and girls.