A cooperative prospective study was set up in seven countries. Very-high-risk acute lymphoblastic leukaemia in first complete remission was defined by the presence of at least one of the following criteria: (1) failure to achieve complete remission after the first four-drug induction phase; (2) t(9;22) or t(4;11) clonal abnormalities; and (3) poor response to prednisone associated with T immunophenotype, white-blood-cell count of 100×109/L or greater, or both. Children were allocated treatment by genetic chance, according to the availability of a compatible related donor, and assigned chemotherapy or haemopoietic-cell transplantation. The primary outcome was disease-free survival and analysis was by intention to treat.
Between April, 1995, and December, 2000, 357 children entered the study, of whom 280 were assigned chemotherapy and 77 related-donor haemopoietic-cell transplantation. 5-year disease-free survival was 40·6 % (SE 3·1) in children allocated chemotherapy and 56·7 % (5·7) in those assigned transplantation (hazard ratio 0·67 [95 % CI 0·46–0·99]; p=0·02); 5-year survival was 50·1 % (3·1) and 56·4 % (5·9), respectively (0·73 [0·49–1·09]; p=0·12).
Children with very-high-risk acute lymphoblastic leukaemia benefit from related-donor haemopoietic-cell transplantation compared with chemotherapy. The gap between the two strategies increases as the risk profile of the patient worsens.