The efficacy and limitations of transarterial embolization for severe hepatic injury

BACKGROUND: The efficacy of transarterial embolization (TAE) for severe blunt hepatic injury has been reported. We performed a prospective study evaluating the efficacy and the limitation of TAE from January 1996 to December 2000. METHODS: All patients with blunt abdominal injury who could be stabilized by fluid resuscitation underwent computed tomographic (CT) scan examinations. Patients with CT scan evidence of hepatic injury were classified into five grades according to CT scan findings on the basis of the injury scale of the American Association for the Surgery of Trauma (Mirvis classification). All patients with CT scan grade 3 to 5 injury underwent angiography. When angiography showed extravasation of contrast medium extending from hepatic arterial branches, TAE was performed. RESULTS: Of 612 patients with blunt abdominal trauma, 51 had CT scan grade 3 to 5 injury. Thirty-seven of these patients had a CT scan grade 3 injury and 18 underwent TAE. One of 19 patients who did not undergo TAE developed a delayed hemorrhage on day 6 and required a laparotomy. All 13 patients with a CT scan grade 4 injury had angiographic findings of the extravasation. TAE was successful in 11 patients and unsuccessful in 2. Five patients with a CT scan grade 4 injury required laparotomy. One developed a delayed hemorrhage on day 4. The remaining four patients had a major venous injury (a right lobectomy was performed in two with inferior vena cava injury, and a gauze packing in two with hepatic venous injury). One patient with a CT scan grade 5 injury underwent immediate laparotomy after TAE. Laparotomy revealed inferior vena cava injury and a right lobectomy was performed. Only two patients who underwent a lobectomy died of an uncontrollable hemorrhage. All CT scans of patients with hepatic venous or inferior vena cava injury showed a large low-density area (> or = 10 cm) with involvement of these vessels. The volumes of fluid resuscitation needed from admission until TAE ranged from 2,109 to 2,638 mL/h. CONCLUSION: It was considered that the combination of the presence of a CT scan grade 4 or 5 lesion and the fluid requirements of more than 2,000 mL/h to maintain normotension indicated the absolute necessity of surgery. We felt that these patients were not candidates for TAE, and should undergo immediate laparotomy.