PURPOSE: To analyse the causes of coagulopathy and determine the effect of embolotherapy on acute gastrointestinal(GI) bleeding coexisting with coagulopathy. MATERIALS AND METHODS: Between June 1991 and December 1998, 29 patients with acute GI bleeding (M:F =21:8, mean age, 57.8 years) underwent percutaneous embolotherapy and immediate cessation of bleeding was confirmed. The patients were divided into two groups: control (n =16) and those with coagulopathy (n =13), group membership being determined according to the criteria of >+/-2SD of normal prothrombin time (PT) and activated partial thromboplastin time (aPTT) ( PT >23 seconds, aPTT >40 seconds) at the time at which embolization was requested. Embolotherapy was, defined as clinically successful, if the patient was stable for at least three days, without bleeding, after technically successful embolization. The clinical success rate of embolization and the mortality rate were compared between the two groups, and the causes of coagulopathy statistically analysed. RESULTS: The clinical success rate of embolization was 75% (n =12) in the control group, compared with 38.5% (n =5) in the coagulopathic group (p < 0.05), while the mortality rate for the two groups was 6.3% (n =1) and 53.8% (n =7), respectively (p < 0.005). Statistically, massive transfusion and sustained shock before embolization were the causes of coagulopathy (p < 0.05). CONCLUSION: In coagulopathic patients with acute GI bleeding, embolotherapy induces transient bleeding control, but is unlikely to save lives.