Background: Despite expectant management, preeclampsia remote from term usually results in preterm delivery. Antithrombin, which displays antiinflammatory and anticoagulant properties, may have a therapeutic role in treating preterm preeclampsia, a disorder characterized by endothelial dysfunction, inflammation, and activation of the coagulation system. Objective: This randomized, placebo-controlled clinical trial aimed to evaluate whether intravenous recombinant human antithrombin could prolong gestation and therefore improve maternal and fetal outcomes. Study Design: We performed a double-blind, placebo-controlled trial at 23 hospitals. Women were eligible if they had a singleton pregnancy, early-onset or superimposed preeclampsia at 23 0/7 to 30 0/7 weeks’ gestation, and planned expectant management. In addition to standard therapy, patients were randomized to receive either recombinant human antithrombin 250 mg loading dose followed by a continuous infusion of 2000 mg per 24 hours or an identical saline infusion until delivery. The primary outcome was days gained from randomization until delivery. The secondary outcome was composite neonatal morbidity score. A total of 120 women were randomized. Results: There was no difference in median gestational age at enrollment (27.3 weeks’ gestation for the recombinant human antithrombin group [range, 23.1–30.0] and 27.6 weeks’ gestation for the placebo group [range, 23.0–30.0]; P=.67). There were no differences in median increase in days gained (5.0 in the recombinant human antithrombin group [range, 0–75] and 6.0 for the placebo group [range, 0–85]; P=.95). There were no differences between groups in composite neonatal morbidity scores or in maternal complications. No safety issues related to recombinant human antithrombin were noted in this study, despite the achievement of supraphysiological antithrombin concentrations. Conclusion: The administration of recombinant human antithrombin in preterm preeclampsia neither prolonged pregnancy nor improved neonatal or maternal outcomes.