The antileukaemic enzyme L-asparaginase is used to achieve the greatest possible reduction in blood levels of the amino acid asparagine, an essential factor for the growth of leukaemic blasts. There are two main sources of the enzyme, E. coli and Erwinia. Faced with increasing reports of treatment complications, we established a programme to monitor enzyme activity and asparagine levels in serum, in children receiving treatment for acute lymphoblastic leukaemia (ALL) and non-Hodgkin's lymphoma (NHL). Trough asparagine and asparaginase levels were measured in 49 children on induction treatment with different E. coli preparations (Asparaginase medac, Crasnitin) and in 52 children on re-induction (Asparaginase medac, Crasnitin, and, in the event of allergic reactions, Erwinase) just prior to each sequential application of 10000 U/m2 of asparaginase. Measurements were made by an enzyme assay and an HPLC method. During induction, both Escherichia coli preparations induced the desired reduction in asparagine, but the asparaginase activity with Asparaginase medac was significantly higher than with Crasnitin (median of trough levels 475 versus 74 U/l). Under re-induction treatment (median, Asparaginase medac 528 U/l, Crasnitin 49 U/l, and Erwinase < 20 U/l) complete asparagine depletion was recorded on day 3 in more than 90% of Asparaginase medac samples, more than 60% of Crasnitin samples and in 26% of Erwinase samples. The latter two groups included some children with unchanged asparagine levels and no measurable enzyme activity. Different asparaginase preparations are not readily interchangeable. When Asparaginase medac is used instead of Crasnitin, and identical dose will be associated with significantly higher enzyme activity, well above the level required for complete asparagine depletion. Clinical studies will need to specify both the preparation and the dose to be used. When substitution of an alternative drug is mandatory owing to allergic reactions, monitoring is advisable.