Childhood autoimmune haemolytic anaemia (AIHA) requires second-line immunosuppressive therapy in 30-50% of cases. It appears that rituximab is indicated in such circumstances. This prospective national study reports the practice, efficacy and tolerance of rituximab in children with isolated AIHA and AIHA in the setting of Evans syndrome (ES). Sixty-one children were given rituximab between 2000 and 2014. The median interval from diagnosis to rituximab was 9·9 [interquartile range (IQR) 1·6-28·5] months. Forty-six patients responded (75%) and the 6-year relapse-free survival (RFS) was 48%. Twenty patients relapsed at a median interval of 10·8 (IQR 3·9-18·7) months, rituximab allowed steroid withdrawal in 44/61 (72%) of children. In isolated AIHA, complete response and 6-year RFS were significantly higher than in ES (P < 0·05). Ten out of 61 patients were infants, seven of who responded with a 6-year RFS of 71%. Among patients without immunoglobulin substitution before rituximab, 4 are still receiving substitutions. Five patients died, including one potentially attributable to rituximab. This large observational series of childhood AIHA established the rituximab benefit-risk ratio, allowing steroid withdrawal, with 37% of long-term responders, mainly in isolated AIHA. All subgroups of patients drew benefit. Our long-term results indicate the baseline to be challenged by new treatment approaches.