This randomised trial was designed to compare two groups treated with different G-CSF administration schedules with a third group receiving no G-CSF, after autologous peripheral blood stem cell transplantation (APBSCT). Children and adults with haematological malignancies or solid tumours were randomly assigned to receive either 150 microg/m2/day of Lenograstim starting on day 1 (G1) or on day 5 (G5) post APBSCT, or no Lenograstim (G0). Randomisation was stratified according to the conditioning regimen (Busulfan vs TBI vs no Busulfan and no TBI) and the graft CD 34+ cell count. A total of 240 patients were randomised; 239 were evaluable. All three patient groups were comparable. Median duration of neutropenia was 9 days (4-40), and 10 days (5-15) in the G1 and G5 groups, respectively, significantly shorter than in the G0 group, 13 days (7-36) (P < 0.0001). No difference was observed in the duration of thrombocytopenia, transfusion support and extra-haematological complications. The duration of post transplant hospitalisation was significantly shorter in adults who received G-CSF. Clinical and cost arguments favour the initiation of G-CSF on day 5 in adults. The same policy could be applied in children given that clinical management is easier and costs are similar.