BACKGROUND Septic shock is frequently complicated by a syndrome of disseminated intravascular coagulation (DIC). Numerous uncontrolled clinical studies have reported that antithrombin III (ATIII) substitution might prevent DIC and death in septic shock. METHODS We conducted a randomized double-blind placebo-controlled trial in patients with a documented septic shock and DIC. The patients received either a placebo or ATIII (90 to 120 IU/kg in loading dose, then 90 to 120 IU/kg/d during 4 days). Administration of fresh frozen plasma, platelets, and fibrinogen concentrates was restricted to patients with hemorrhages and severe decreases in prothrombin time, platelet count, and fibrinogen levels. RESULTS Thirty-five patients entered the study (18 placebo, 17 ATIII). Both groups were well balanced for all demographic, hemodynamic, and biologic data. Three patients were excluded before the treatment allocation code was broken. In the ATIII group, ATIII levels were rapidly corrected and remained over normal levels until day 10; sequential protein C and protein S levels were not modified. The duration of DIC was significantly reduced: in the ATIII group, 64 percent of patients were cured of DIC at day 2, and 71 percent were cured at the end of treatment vs in the placebo group, 11 percent (p < 0.01) and 33 percent (p < 0.05), respectively. In the 32 included patients, the mortality in ICU was reduced by 44 percent in the ATIII group (p = 0.22, NS). Care loads and transfusion requirements were not different. No side effect was observed. CONCLUSIONS Mortality was reduced by 44 percent in this trial, but the difference did not reach the statistical significance. Circulating protein C and protein S levels were not modified by ATIII supplementation. High doses of ATIII concentrates significantly improved sepsis-induced DIC during septic shock. The trend toward improved survival suggests further randomized studies.