To evaluate the clinical benefit of the prophylactic use of urate oxidase in children with non-Hodgkin's lymphoma (NHL), we analyzed the incidence and complications of tumor lysis syndrome (TLS) in children with B-cell acute lymphoblastic leukemia (B-ALL) or stage III/IV Burkitt's lymphoma and a lactate dehydrogenase (LDH) level > or =500 U/l before and after the introduction of a protocol amendment to use urate oxidase for the prophylaxis of TLS. Data from 1791 children with NHL enrolled in the two subsequent multicenter studies NHL-BFM 90 and 95 were evaluated. The presence of the side effects TLS, anuria, sepsis, and other complications during the first 2 weeks after admission were registered. Until March 1996, no urate oxidase was used (period 1). From November 1997 all children with B-ALL or stage III and IV B-NHL and LDH > or =500 U/l should receive urate oxidase prophylactically (period 3). In between (period 2), urate oxidase was given in a minority of hospitals therapeutically. Initial chemotherapy was identical. Altogether, 78 children (4.4%) developed a TLS. Patients with B-ALL had the highest risk to develop a TLS (26.4%) followed by B-ALL/Burkitt's lymphoma and a LDH > or =500 U/l (14.9%). In period 1, 16.1% and 9.2% of the latter children developed a TLS or anuria, respectively, compared to 12.3% and 6.2% in period 3 ( p=NS). The incidence of sepsis remained unchanged (5.0% vs 4.6%). In children with B-ALL the differences in the incidence of TLS and anuria between period 3 and period 1 were more pronounced, reaching significance for anuria (15.4% vs 3.8%, p=0.03). Our results suggest that patients with the highest risk to develop a TLS might benefit from the prophylactic use of urate oxidase.