We hypothesized that infusion of recombinant human antithrombin without concomitant heparin would have dose-dependent anticoagulant properties and potentially decrease endotoxin (lipopolysaccharide [LPS])-induced cytokine production. This was a randomized, double-blind, placebo-controlled study in parallel groups enrolling 30 healthy male volunteers. The active treatment groups received infusions of recombinant human antithrombin to increase antithrombin levels to 200% and 500% before infusion of 2 ng/kg endotoxin (LPS). Infusion of antithrombin dose-dependently decreased coagulation (P < .01 by repeated-measures ANOVA): peak levels of prothrombin fragment (1.8 nmol/L [95% confidence interval (CI), 1.3-2.3 nmol/L] in the 500% antithrombin group and 4.4 nmol/L [95% CI, 2.7-6.2 nmol/L] in the placebo group at 4 hours), thrombin antithrombin complexes (12 microg/L [95% CI, 8-16 microg/L] in the 500% antithrombin group and 34 microg/L [95% CI, 20-48 microg/L] in the placebo group at 4 hours), and D-dimer (0.2 microg/L [95% CI, 0.1-0.2 microg/L] in the 500% antithrombin group and 0.5 microg/L [95% CI, 0.4-0.7 microg/L] in the placebo group). Recombinant human antithrombin decreased peak interleukin-6 levels by 40% (222 pg/mL [95% CI, 148-295 pg/mL] and 216 pg/mL [95% CI, 112-320 pg/mL] in the 500% and 200% antithrombin groups, respectively, versus 357 pg/mL [95% CI, 241-474 pg/mL] in the placebo group; P < .001 by ANOVA). Finally, infusion of recombinant human antithrombin rapidly and transiently decreased neutrophil counts (by 19% [95% CI, 8%-30%] in the 500% antithrombin group versus 6% [95% CI, 1%-10%] in the placebo group, P = .002 by Kruskal-Wallis ANOVA) and monocyte counts (by 30% [95% CI, 16%-44%] in the 500% antithrombin group and 18% [95% CI, 9%-28%] in the 200% antithrombin group versus 8% [95% CI, 5%-20%] in the placebo group, P = .04) before LPS challenge, indicating that recombinant human antithrombin directly interacts with these leukocyte subsets. In summary, recombinant human antithrombin dose-dependently inhibited tissue factor-triggered coagulation. Effects on leukocytes and inhibition of interleukin-6 release seem to represent specific pharmacodynamic properties of recombinant human antithrombin.