Background: Delayed massive bleeding is one of the leading causes of mortality after pancreaticoduodenectomy (PD) and is often preceded by sentinel bleed. Immediate and accurate diagnosis of sentinel bleed is essential to save patients from a delayed massive hemorrhage. Angiography is probably the procedure of choice for patients with sentinel bleed after PD, as it will localize the bleeding point and provide interventional embolization. The purpose of this study is to test the efficiency of angiography as the initial management for patients with sentinel bleed after pancreaticoduodenectomy.
Methods: The study group consisted of 283 patients who underwent PD from July 2002 to June 2007. Angiography and arterial embolization were performed for every sentinel bleed and detected pseudoaneurysm. Patients (n = 311) from a previous study (July 1996-June 2002) were used as a historical control group.
Results: Sentinel bleed was detected in 20 patients in study group. Of these, angiography-detected pseudoaneurysm was evident in seven (35%); all were successfully embolized. Delayed massive hemorrhage occurred in three of 13 patients with sentinel bleed but negative angiography. All three were operated on; one died of uncontrolled bleeding. The number of hemodynamically unstable patients before transfusion, units of transfused packed cells, and bleeding related mortalities were significantly less in study group than the control group.
Conclusions: Institution of angiography for every detected sentinel bleed after PD enabled us to embolize seven pseudoaneurysms before massive hemorrhage. Most importantly, bleeding-related mortality was significantly less than in the absence of angiography.