Polyethylene glycol-conjugated (PEG) asparaginase is approved for use in patients who develop allergy to other forms of asparaginase, although its ability to deplete asparagine systemically in patients with hypersensitivity has not been well elucidated. In 53 children with newly diagnosed acute lymphoblastic leukemia, we serially assessed asparagine concentrations in cerebrospinal fluid (CSF) and plasma as well as serum anti-asparaginase antibodies. All patients received native Escherichia coli (Elspar) asparaginase during induction therapy; patients received PEG asparaginase during reinductions when available, and those who developed allergy received Erwinia asparaginase. All eight patients who developed clinical evidence of allergy to asparaginase had anti-asparaginase antibodies. Among patients who had no antibodies, those who received E. coli had lower mean (+/-s.d.) CSF asparagine (0.29+/-0.63, n=9) than those who received PEG (0.77+/-0.82, n=4) (P=0.007). Results were similar for plasma asparagine. There was no situation where asparagine concentrations were more effectively depleted by PEG than by other preparations. None of the five patients who developed thrombosis had an allergy or antibodies to asparaginase at the time of the thrombosis. We conclude that asparagine concentrations were less effectively depleted by PEG than by E. coli asparaginase at the doses commonly used. The risk of thrombosis may be affected by the intensity of asparaginase exposure.