To assess the maternal response to low molecular weight heparin during pregnancy, by estimation of plasma anti-Xa activity, at three specified gestation points and in the nonpregnant state. A longitudinal, prospective, observational study was set in a tertiary referral recurrent miscarriage clinic. Twenty-four women, attending consecutively, were invited to participate and gave informed consent. Each woman had a history of recurring pregnancy loss and positive preconception screening for antiphospholipid syndrome. After confirmation of a viable pregnancy all subjects began taking 5000 IU of dalteparin once daily subcutaneously. Serial measurement of plasma anti-Xa activity after administration of dalteparin was performed at three standard gestation points (12, 24, and 36 weeks) and in the nonpregnant state (6 weeks postpartum). Peak anti-Xa levels occurred at 4 hours postbolus in pregnancy, as compared with 2 hours in the nonpregnant state. The mean anti-Xa levels at 12, 24, and 36 weeks' gestation were significantly reduced, at 2 hours postinjection, as compared with the nonpregnant state (P <.001, P <.01, P <.001, respectively). The lowest dose-response curve was at 36 weeks' gestation. A repeated-measures analysis of variance found a significant difference (P <.05) between the 36-week group and the postterm group but not between any of the other groups. During pregnancy, differences in the pharmacokinetics of low molecular weight heparin were observed, with an overall reduction in anti-Xa activity. On the basis of this study it is questionable to extrapolate dosing and lack of dose monitoring, in pregnant women, using data derived from a nonpregnant population.