Heparin is commonly used to treat neonatal thrombosis. Drug monitoring often involves assays that measure the inhibition of added factor Xa or thrombin by the antithrombin III (AT-III) heparin complex. We examined whether the neonatal AT-III deficiency affects heparin recovery in such assays at therapeutic drug concentrations (0.1 to 0.6 U/ml). The chromogenic anti-factor Xa assay (Teien AN, Lie M, Abildgaard U. Thromb Res 1976;8:413-6) and the protamine titration test were performed in eight pooled cord plasma samples and in normal adult plasma. Only 65% to 70% of heparin activity detected in adult plasma was recovered by the assay of Teien et al. In cord plasma; recovery by the protamine test was 80% or less. Heparin recovery in cord plasma was significantly improved by raising the AT-III concentration to normal adult levels in both assays. We conclude that heparin assays underestimate drug levels in neonatal plasma unless the neonatal AT-III deficiency is fully corrected in the test system. The use of standard assays may lead to over-heparinization of newborn infants, thereby placing them at a higher risk of bleeding.