The decrease in the antithrombin III activity is thought to result from consumption by ongoing coagulation, degradation by neutrophil elastase, capillary leak syndrome, and impaired synthesis. A retrospective data analysis of patients with sepsis was conducted to investigate the response of antithrombin III activity after supplementation in patients with sepsis, and to determine what factors affect the response of antithrombin III activity. The study included 42 patients with sepsis, 75 patients with severe sepsis, and 65 patients with septic shock, who were administered antithrombin III. Antithrombin III activity, platelet counts, coagulation, and fibrinolytic markers were collected before administration and 24 h after the supplementation. In the patients with septic shock, the response of antithrombin III activity after supplementation was 0.37% ± 1.21%/IU per kg body weight, which was significantly lower in comparison with those in the patients with sepsis (1.81 ± 1.75; P < 0.001) or severe sepsis (1.36 ± 1.65; P < 0.001). The patients with liver dysfunction had significantly lower response to antithrombin III activity than that of the patients without liver dysfunction (P < 0.0001). A stepwise multiple linear regression analysis revealed that the severity of sepsis and liver function were independent predictors for the response to antithrombin III activity. These results suggest that the response to antithrombin III supplementation may be affected by both a systemic inflammation and impaired synthesis in patients with sepsis.