A retrospective analysis of children with first relapse of acute lymphoblastic leukaemia (ALL), treated on the UKALL R2 protocol at four different hospitals, between June 1995 and December 2002 was performed. Of the 150 children 139 (93%) achieved a second complete remission. The overall survival (OS) and event-free survival (EFS) for the whole group was 56% and 47% respectively. The duration of first complete remission and immunophenotype, but not sites of relapse, were predictive for survival. Using the Berlin-Frankfürt-Münster risk stratification for relapsed ALL, the OS and EFS for standard, intermediate (IR) and high risk (HR) groups were 92% and 92%, 64% and 51%, and 14% and 15%, respectively; P < 0.0001 for both OS and EFS. In the IR group, those with a very early isolated central nervous system relapse also had a significantly worse outcome (P = 0.0001). Given the poor outcome of a second relapse, clear strategies are required to identify those in the IR group who will most benefit from stem cell transplantation (SCT). A higher proportion (16%) of induction failures in the HR group suggest the need for novel agents during this phase of treatment, but SCT was associated with a lower relapse rate and better outcome than those treated with chemotherapy alone.