The safety and efficacy of recombinant DNA-produced factor VIIa (rFVIIa) was investigated in 15 haemophilic patients in non-bleeding states and during bleeding episodes (mild to moderate joint bleed). Patients with severe haemophilia A without inhibitors (n = 4), haemophilia A with inhibitors (n = 10), and haemophilia B with inhibitor (n = 1) received one or more doses of rFVIIa during 32 non-bleeding study episodes and 23 bleeding episodes. The study was an open, uncontrolled, dose-escalation (17.5 micrograms/kg, 35 micrograms/kg, 70 micrograms/kg) trial. Physical evaluation, laboratory assessment, and immunology testing were conducted at baseline, monthly for 3 months and every 3 months thereafter. The immediate safety of rFVIIa was assessed by monitoring of D-dimer, fibrinogen, platelet count, antithrombin III, thrombin-antithrombin complex, and alpha 2-antiplasmin 5 min before and at multiple times throughout the following 24 h. Prothrombin time (PT) and activated partial thromboplastin time (aPTT) values were also obtained. Pain, swelling, joint circumference, and range of motion were recorded before administration of the initial dose of rFVIIa in bleeding patients and at 6, 12, and 24 h. Haemostatic response to rFVIIa was observed in patients with severe VIII and IX deficiency with and without inhibitors. Therapy with rFVIIa was judged effective in 19 of the 22 evaluable bleeding episodes at one or more time points. The 35 micrograms/kg and 70 micrograms/kg doses were associated with higher response rates at 6 and 12 h compared to the 17.5 micrograms/kg dose level. A second dose of rFVIIa was administered in 20 of the 22 bleeding episodes.(ABSTRACT TRUNCATED AT 250 WORDS)