Twenty-seven children with systemic lupus erythematosus (SLE) were tested for antiphospholipid antibodies (APLA), i.e. lupus anticoagulant and immunoglobulin (Ig)G or IgM anticardiolipin antibodies (ACLA), with beta-2 glycoprotein I as cofactor, in a single-centre, prospective study over 2 years. Eighteen patients (67%) tested positively for one or the other APLA during the course of the study. Twelve children (44%) tested positively for IgG ACLA and ten (37%) for IgM ACLA, whereas eight (30%) were positive for lupus anticoagulant. In two patients with thrombosis, IgG anticardiolipin positivity was seen to be variable. Unlike the results of most other reports in the literature, lupus anticoagulant positivity was not consistently associated with thrombosis. A majority of the children (83%) tested positively for ACLA during disease activity. Immunoglobulin G and IgM ACLA positivity did not correlate significantly with disease status. The results of this prospective study would indicate that, though frequently present, APLA may be unable to be predictive of disease behaviour in children with SLE.