Background: Recombinant activated factor VII (rfVIIa) is widely used as a prohemostatic agent for patients with congenital and acquired haemophilia. In the last few years it has been increasingly used ‘off label’ in medical and surgical patients to minimise or control severe or refractory bleeding, in particular when conventional therapy failed or was judged to be insufficient. Methods: A computer-assisted search of the MEDLINE, EMBASE, and CENTRAL (Cochrane Central Register of Controlled Trials) electronic databases up to July 2006 to identify published randomised controlled trials (RCTs) that evaluated the effect of rfVIIa as prophylaxis and/or treatment of bleeding episodes in patients without inherited haemostatic diathesis Results: Sixteen published double-blind RCTs met the inclusion criteria, in six main clinical settings: surgery in patients with normal preoperative haemostasis, acute intracerebral haemorrhage, liver disease, trauma, haematopoietic stem cell transplantation, and Dengue hemorrhagic fever. Overall, RCTs showed limited efficacy of rfVIIa in several clinical settings but in patients with spontaneous intracerebral haemorrhage. Conclusions: Published data suggest that rfVIIa should not be routinely used as a general haemostatic agent. Until approval, if any, will be given, if ever, by regulatory agencies, the compassionate administration of rfVIIa as a general haemostatic agent should always be based on individual clinical conditions, patient risk/benefit profile and published literature.