Hemorrhage after pancreaticoduodenectomy: when is surgery still indicated?

BACKGROUND: This study analyzed presentation and management of hemorrhage after pancreaticoduodenectomy (PD) to determine the respective role of surgery and embolization. METHODS: From January 1992 to March 2005, 411 patients underwent PD and were analyzed with regard to postoperative hemorrhage. RESULTS: Hemorrhage occurred in 27 patients (7%), either within the first 3 postoperative days ("early" hemorrhage, n = 11) or after day 8 ("delayed" hemorrhage, n = 16, including 4 with "sentinel" bleeding). At the time of bleeding, 12 patients (44%) (all with delayed hemorrhage) had associated abdominal complications. Two patients had successful conservative treatment. Two stable patients with pseudoaneurysm, diagnosed by computed tomography scan, underwent successful embolization. Four patients with active bleeding underwent unsuccessful angiography. Overall, 23 patients were reoperated on without any completion pancreatectomy, 3 rebled, and 3 (11%) died (including 2 with delayed hemorrhage). CONCLUSIONS: Both embolization and surgery have a role in the management of hemorrhage after PD. For early hemorrhage, reoperation is appropriate. In case of sentinel bleeding, pseudoaneurysms can be detected by computed tomography scan and treated by embolization. For delayed active hemorrhage, reoperation is still indicated.     

Sentinel Bleeding After Pancreaticoduodenectomy: A Disregarded Sign

Introduction Delayed massive hemorrhage induced by pancreatic fistula after pancreaticoduodenectomy is a rare but life-threatening complication. The purpose of this study was to analyze the clinical course of patients with late hemorrhage, with or without sentinel bleeding, to better define treatment options in the future. Material and Methods From April 1998 to December 2006, 189 pancreaticoduodenectomies were performed. Eleven patients, including two patients referred from other hospitals, were treated with delayed massive hemorrhage occurring 5 days or more after pancreaticoduodenectomy. Sentinel bleeding was defined as minor blood loss via surgical drains or the gastrointestinal tract with an asymptomatic interval until development of hemorrhagic shock. The clinical data of patients with bleeding episodes were analyzed retrospectively. Results Eight of the 11 patients had sentinel bleeding, and seven of them had it at least 6 h before acute deterioration. Seven out of 11 patients died, five out of eight with sentinel bleeding. No differences could be detected between patients with or without sentinel bleeding before delayed massive hemorrhage. The only difference found was that non-surviving patients were significantly older than surviving patients. Delayed massive hemorrhage is a common cause of death after pancreaticoduodenostomy complicated by pancreatic fistula formation. The observation of sentinel bleeding should lead to emergency angiography and dependent from the result to emergency relaparotomy to increase the likelihood of survival.     

Delayed hemorrhage after pancreaticoduodenectomy

BACKGROUND: Postoperative hemorrhage, particularly delayed hemorrhage after pancreaticoduodenectomy, is a serious complication and one of the most common causes of mortality after pancreaticoduodenectomy. STUDY DESIGN: The medical records of 500 patients who underwent pancreaticoduodenectomy between October 1994 and December 2002 were analyzed with regard to postoperative hemorrhagic complications. Delayed hemorrhage was defined as bleeding at the operation site after 5 or more postoperative days. RESULTS: Delayed hemorrhage occurred in 22 patients (4.4%), with a median time of 13 days (range 7 to 32 days) after pancreaticoduodenectomy, and developed more frequently (9/77 versus 13/423, p = 0.003) in patients with preceding intraabdominal complications such as pancreatic fistula, bile fistula, and intraabdominal abscess. In 17 of these 22 patients, angiography and laparotomy revealed bleeding foci at 14 arterial and 3 anastomotic sites. In nine patients, hemorrhage developed from pseudoaneurysms of the major arteries around the pancreaticojejunostomy. Hemostatis was attempted by transcatheter arterial embolization in 14 patients and with laparotomy in 4 patients. Four of 14 patients who received transcatheter arterial embolization eventually required laparotomy. Overall, 4 of the 22 delayed hemorrhage patients died (18.2%) of complications related to massive bleeding or transcatheter arterial embolization. CONCLUSIONS: Delayed hemorrhage after pancreaticoduodenectomy is associated with a high mortality. Intraabdominal complications after pancreaticoduodenectomy should be evaluated properly and guidelines for the diagnosis and treatment of delayed hemorrhage should be established in advance. Clinicians must be alert to the possibility of pseudoaneurysm hemorrhage.     

Risk factors for and management of delayed intraperitoneal hemorrhage after pancreatic and biliary surgery

BACKGROUND: Delayed intraperitoneal hemorrhage (DIH) is still an important cause of postoperative mortality in pancreatic and biliary surgery. METHODS: Sixty-nine patients who underwent pancreatic and biliary surgery with skeletonization for lymphadenectomy of the hepatoduodenal ligament between April 2002 and March 2005 were included in this study. Statistical analyses of the risk factors for DIH were performed using both univariate and multivariate modalities. RESULTS: DIH occurred in 4 patients (5.8%) within a median time of 15 days after surgery. Stepwise logistic regression analysis identified intra-abdominal abscess formation as the independent predictor of DIH. All 4 patients had a sentinel bleed before the onset of DIH. Three patients were treated by transarterial embolization and 1 patient was treated by surgical intervention. Three patients had liver abscess after hemostasis of DIH, but all 4 patients recovered and were discharged from the hospital. CONCLUSIONS: A computed tomography angiography should be performed on patients with intra-abdominal abscess formation and sentinel bleed after pancreatic and biliary surgery to check if a pseudoaneurysm has formed.     

Management of postoperative arterial hemorrhage after pancreato-biliary surgery according to the site of bleeding: re-laparotomy or interventional radiology

Background/Purpose  Intra-abdominal arterial hemorrhage is still one of the most serious complications after pancreato-biliary surgery. We retrospectively analyzed our experiences with 15 patients in order to establish a therapeutic strategy for postoperative arterial hemorrhage following pancreato-biliary surgery. Methods  Between August 1981 and November 2007, 15 patients developed massive intra-abdominal arterial bleeding after pancreato-biliary surgery. The initial surgery of these 15 patients were pylorus-preserving pancreatoduodenectomy (PPPD) (7 patients), hemihepatectomy and caudate lobectomy with extrahepatic bile duct resection or PPPD (4 patients), Whipple’s pancreatoduodenectomy (PD) (3 patients), and total pancreatectomy (1 patient). Twelve patients were managed by transcatheter arterial embolization and three patients underwent re-laparotomy. Results  Patients were divided into two groups according to the site of bleeding: SMA group, superior mesenteric artery (4 patients); HA group, stump of gastroduodenal artery, right hepatic artery, common hepatic artery, or proper hepatic artery (11 patients). In the SMA group, re-laparotomy and coil embolization for pseudoaneurysm were performed in three and one patients, respectively, but none of the patients survived. In the HA group, all 11 patients were managed by transcatheter arterial embolization. None of four patients who had major hepatectomy with extrahepatic bile duct resection survived. Six of seven patients (85.7%) who had pancreatectomy survived, although hepatic infarction occurred in four. Conclusions  Management of postoperative arterial hemorrhage after pancreato-biliary surgery should be done according to the site of bleeding and the initial operative procedure. Careful consideration is required for indication of interventional radiology for bleeding from SMA after pancreatectomy and hepatic artery after major hepatectomy with bilioenteric anastomosis.     

Use of a stent graft for bleeding hepatic artery pseudoaneurysm following pancreaticoduodenectomy

Although uncommon, bleeding following pancreaticoduodenectomy is associated with high mortality. Management generally includes surgical reexploration or, alternatively, transarterial embolization. We report the case of a 62-year-old man who presented with massive upper gastrointestinal bleeding 3 weeks after pancreaticoduodenectomy. Selective coeliac angiography revealed a large pseudoaneurysm involving the proper hepatic artery. This was treated successfully with a stent graft. There was no recurrence of bleeding at the 6-month follow-up. To our knowledge, this is the first report of stent graft repair of bleeding hepatic artery pseudoaneurysm following pancreaticoduodenectomy.     

Clinical features and management of pseudoaneurysmal bleeding after pancreatoduodenectomy

A ruptured pseudoaneurysm is the most serious and life-threatening cause of postpancreatoduodenectomy (PD) hemorrhages. We have evaluated the clinical course and management of pseudoaneurysms after PD. Of 586 patients who underwent PD for periampullary tumors in Asan Medical Center between March 2003 and March 2011, 27 experienced pseudoaneurysmal bleeding. Bleeding developed at a median of 21 days (range, 8 to 45 days) after surgery, including 9 patients who developed bleeding more than 4 weeks after surgery. Before development of bleeding, 26 patients showed pancreatic fistula. Bleeding was developed from the gastroduodenal artery stump in 12 patients, the common hepatic artery in eight, the proper hepatic artery in five, and the left hepatic artery in two. Of the angiographic group, 21 patients underwent with microcoil embolization, four underwent stent insertion, and one experienced technical failure. Only one patient required emergent laparotomy without angiography. Of 25 patients with angiographic procedures, all patients achieved hemostasis. The mortality rate was 22.2 per cent (6 patients). Delayed hemorrhage after PD is closely associated with pancreatic fistula and carried a significantly higher mortality rate. The patients with pancreatic fistula should be carefully monitored, even more than 4 weeks after surgery. Selective microcoil embolization or stent graft is effective for pseudoaneurysmal bleeding.     

Transarterial embolization for postoperative hemorrhage after abdominal surgery

The study goal was to evaluate the efficacy, safety, and clinical outcome of transarterial embolization for postoperative hemorrhage after abdominal surgery. Thirty-three patients were referred for angiography because of gastrointestinal or intra-abdominal bleeding after abdominal surgery. Urgent angiography and transarterial embolization was performed in all 33 patients. The clinical and angiographic features were retrospectively reviewed. Angiography revealed a discrete bleeding focus in 26 (79%) of 33 patients. Transarterial embolization was technically successful in 24 (92%) of 26 patients with a discrete bleeding focus. Rebleeding occurred in four (17%) of 24 patients. They were successfully managed with repeat embolization. There was no procedure-related complication during follow-up period. Angiography has a high detection rate of bleeding site in patients with postoperative hemorrhage after abdominal surgery. Transarterial embolization is considered to be an effective and safe means in the management of postoperative hemorrhage.     

经导管选择性动脉栓塞治疗创伤性假性动脉瘤

目的探讨经导管选择性动脉栓塞治疗损伤性假性动脉瘤的疗效。方法 16例损伤性出血患者,其中肾脏出血5例,肝脏出血3例,剖宫产后出血4例,切口妊娠流产1例,宫颈癌放疗后出血1例,骨外伤术后出血2例。采用经股动脉穿刺选择性插管进行靶动脉造影及栓塞治疗。采用明胶海绵和（或）弹簧钢圈栓塞瘤腔及供血动脉。结果全部患者动脉造影均见假性动脉瘤,其中2例伴有动静脉瘘。栓塞成功率100%,栓塞后即刻造影示假性动脉瘤消失,止血成功率93.75%。术后均未发生严重并发症,随防6个月均无复发。结论经导管选择性动脉栓塞治疗损伤性假性动脉瘤性出血安全有效、创伤小、并发症少,是可靠的治疗方法 。     

外伤性肾动脉假性动脉瘤的诊断和治疗(附2例报告及文献复习)

目的:探讨外伤性肾动脉假性动脉瘤(Renal artery pseudoaneurysm,RAP)的诊断与治疗方法。方法:回顾性分析2例RAP患者的临床资料:均为男性,年龄分别为17岁和42岁,临床主要表现为持续性出血或迟发性肉眼血尿。血管造影示损伤动脉活动性出血,肾内见充盈造影剂的囊腔状结构。2例患者均行超选择性血管栓塞治疗。结果:2例患者术后血尿或出血均消失,无并发症发生。术后3个月以上行CT检查,肾功能均未见异常。结论:RAP临床表现为持续性出血或迟发性肉眼血尿;血管造影在诊断RAP中具有重要价值;超选择性动脉栓塞是治疗RAP的安全、有效方法。     

腹部手术后腹腔干分支假性动脉瘤消化道瘘的临床分析：附5例分析

目的:总结分析腹部手术后腹腔干分支假性动脉瘤消化道瘘导致的迟发性消化道大出血的诊断和治疗。方法:回顾性分析自2013年1月—2014年9月腹部肿瘤术后上消化道大出血经造影证实腹腔干分支假性动脉瘤消化道瘘的5例患者的临床资料。结果:5例患者消化道出血时间平均为术后53.6 d;假性动脉瘤位于脾动脉2例,位于肝总动脉2例,位于左肝动脉1例;造影后行栓塞治疗4例,行覆膜支架置入1例。无术后严重并发症及围手术死亡病例。随访时间6~16个月,无再次假性动脉瘤破裂出血,肝总动脉覆膜支架置入患者于8个月猝死,原因未明。结论:腹腔干分支假性动脉瘤消化道瘘是腹部手术后罕见而又致命的并发症,应提高该病的认识,其诊断及治疗首选动脉造影及血管腔内治疗,避免医源性损伤可能是减少该病发生的关键。     

Relationship between pseudoaneurysm formation and biloma after successful transarterial embolization for severe hepatic injury: permanent embolization using stainless steel coils prevents pseudoaneurysm formation

OBJECTIVE: The purpose of this study was to determine the association between bilomas and pseudoaneurysm complications after severe hepatic injury. METHODS: Angiography was performed in patients with American Association for the Surgery of Trauma grade > or = III hepatic injury on contrast-enhanced computed tomographic scanning. When contrast extravasation was observed, transarterial embolization (TAE) was performed. After TAE, technetium-99m pyridoxyl-5-methyl-tryptophan cholescintigraphy was performed to detect the coexistence of bilomas. Follow-up angiography was performed when a biloma was detected. Eighty consecutive patients underwent angiography; after angiography, five patients died. The remaining 75 patients who underwent cholescintigraphy were included in this study. RESULTS: All 11 patients who had bilomas had angiographic evidence of contrast extravasation. The biloma frequency was higher in patients with grades IV and V injuries than in those with grade III injury (p = 0.024). Follow-up angiography revealed pseudoaneurysms in 7 of these 11 patients. All six patients in whom only gelatin sponge pledget injection was used to embolize had pseudoaneurysms. Among them, two patients had computed tomographic evidence of massive intra-abdominal fluid collection. In contrast, only one of five patients who received the combination of gelatin sponge pledget injection and stainless steel coils to permanently embolize injured arteries had a pseudoaneurysm. In this patient, the pseudoaneurysm was found in the peripheral part of the collateral vessels. All patients with pseudoaneurysms underwent repeat TAE and were discharged from the hospital uneventfully. CONCLUSION: In patients with high-grade hepatic injury and arterial bleeding who developed biloma, use of a gelatin sponge, an absorbable embolic material, is associated with a risk of pseudoaneurysm formation. Permanent arterial embolization using stainless steel coils is indicated to decrease this risk.     

Delayed epistaxis resulting from external carotid artery injury requiring embolization: a rare complication of transsphenoidal surgery: case report

OBJECTIVE AND IMPORTANCE: Delayed epistaxis resulting from trauma to branches of the external carotid artery is an infrequent but potentially serious complication of transsphenoidal surgery. We report two cases of severe, delayed epistaxis in patients who had undergone transsphenoidal surgery. In both cases, noninvasive treatment failed, necessitating endovascular intervention. CLINICAL PRESENTATION: The first patient, a 52-year-old woman with a prolactinoma, underwent a second transsphenoidal resection 18 months after the first surgery. She was readmitted on postoperative Day 15 with massive epistaxis. The second patient, a 40-year-old woman, had undergone two transsphenoidal surgeries, 14 years apart, for an adrenocorticotropic hormone-secreting adenoma. She was readmitted with massive epistaxis on postoperative Day 17. INTERVENTION: Both patients were initially treated with nasal balloon packing but experienced recurrent hemorrhage when the balloon was deflated, necessitating referral to the interventional radiology department for embolization. At arteriography, the first patient was found to have a pseudoaneurysm of the medial branch of the left internal maxillary artery, which was subsequently embolized. Arteriography in the second patient revealed an abnormally dilated midline branch of the right internal maxillary artery in the nasal septum; this vessel was occluded at arteriography. CONCLUSION: Delayed massive epistaxis is a rare but significant complication of transsphenoidal surgery. Injury to branches of the external carotid artery, along with injury to the internal carotid artery, should be suspected in patients who present with delayed epistaxis after transsphenoidal surgery. Angiography performed in patients with refractory bleeding should include selective external carotid injections. Epistaxis that is refractory to anterior and posterior nasal packing may be effectively treated with endovascular embolization.     

肾部分切除术后迟发性出血原因分析及防治

目的 探讨肾部分切除术后迟发性出血原因及其防治方法.方法 1998-2007年行肾部分切除术382例,发生迟发性出血5例(1.3%).男4例,女1例.平均年龄51(42～63)岁.原发病均为肾癌,肿瘤平均直径2.8(2.3～4.2)cm.2例肿瘤直径＞3.0 cm者手术时阻断肾蒂,3例肿瘤直径＜3.0 cm者游离肾脏后以手握控制肾脏出血;切除范围距离肿瘤边缘0.5～1.0 cm正常肾实质,肾创面"8"字或"U"形对合缝合.5例术后出血时间为6 d～3个月;出血量平均2300(1000～4500)ml.患者均表现为反复肉眼血尿伴患侧腰背部胀痛不适,伴休克表现3例,接受输血治疗3例.5例患者肾动脉造影发现出血原因均为创面肾动脉分支残端形成假性动脉瘤.其中3级肾动脉分支出血4例,4级肾动脉分支出血1例.结果 1例再次手术行患肾切除术,4例行经皮选择性肾动脉栓塞治疗.5例术后出血皆停止.未发生高血压、尿瘘等并发症.4例随访2～9年,肿瘤无复发,未再出血.结果 肾部分切除术后迟发性出血原因包括切面肾动脉小分支未结扎或部分结扎,缝合肾脏创面时缝针贯穿肾实质内动脉等.一旦出现此并发症,需要及时治疗,经皮选择性肾动脉栓塞是有效的治疗方法.     

肾出血急诊肾动脉造影特点及栓塞策略

目的探讨肾出血急诊肾动脉造影特点及栓塞策略。方法回顾性分析63例接受急诊肾动脉造影的肾出血患者,其中医源性操作致出血43例(医源性组)、非医源性操作出血20例(非医源性组),比较2组肾动脉造影表现、所用栓塞材料和止血效果。结果 45例造影可见阳性表现,18例造影阴性。医源性组造影阳性率[79.07%(34/43)]高于非医源性组[55.00%(11/20),P=0.049];2组造影阳性表现差异有统计学意义(P=0.001),医源性组以假性动脉瘤最常见,所用栓塞材料差异无统计学意义(P=0.090)。45例造影阳性病例中,41例经一次栓塞治疗后出血停止,一次栓塞成功率91.11%(41/45),2例经二次栓塞后出血停止,栓塞总成功率95.56%(43/45)。结论医源性操作所致肾动脉出血是急诊肾动脉造影的主因,造影阳性率高,其中假性动脉瘤最常见。选择性肾动脉栓塞治疗肾出血疗效较好。     

超选择性肝动脉插管造影栓塞术治疗胆道大出血

目的 对超选择性肝动脉插管造影栓塞术治疗胆道大出血进行疗效评估。方法 回顾性分析1993年3月至2001年7月胆道大出血8例的临床资料，采用放射介入技术行腹腔动脉和选择性肝右或肝左动脉造影，随即行肝动脉分支出血点近端栓塞。结果 2例患者由于再出血而行第2次栓塞，所有患者均存活，随访2个月至7年无再出血。结论 超选择性肝动脉造影及栓塞安全，可靠微创，可作为胆道大出血的重要诊断及治疗依据。     

Delayed haemorrhage after laparoscopic partial nephrectomy: frequency and angiographic findings

Study Type – Therapy (case series)
Level of Evidence 4 What’s known on the subject? and What does the study add? Delayed haemorrhage has not been well defined in the literature, clinical presentation has not been well described and treatment algorithms are lacking. From our experience we have shown that patients presenting with delayed bleeding after laparoscopic partial nephrectomy will need definitive rather than conservative treatment and angiographic findings demonstrate definitive lesions in all cases. Potential benefits include faster diagnosis and initiation of definitive treatment (angiography with embolization), avoiding repeat computed tomography imaging as it adds little in such patients and reducing need for prolonged hospitalization.

OBJECTIVES

? To determine the frequency of delayed postoperative haemorrhage requiring selective angioembolization (SAE) after laparoscopic partial nephrectomy (LPN). ? To describe the clinical presentation and characterize the angiographic findings encountered in this setting.

PATIENTS AND METHODS

? Prospective data from 640 LPNs performed between August 1993 and May 2009 were retrospectively analyzed, from which patients with delayed postoperative haemorrhage (defined as ‘gross haematuria ≥7 days postoperatively that persists for more than 24 h’) and requiring SAE were identified. ? Clinicopathological, preoperative and perioperative factors were reviewed. ? Selective catheterization and angiography of the renal artery was performed for persistent gross haematuria and for haemodynamic instability associated with a significant drop in haematocrit level. ? Arteries feeding the bleeding site were identified and embolized with endovascular coils.

RESULTS

? Patients presented with delayed haemorrhage between 7 and 30 days after surgery. SAE was required in 13 patients (2%) for delayed postoperative bleeding. ? Of the 640 LPNs, 68 (10.6%) were performed without hilar occlusion (‘off‐clamp’) of whom one (1.5%) had a delayed haemorrhage, which was successfully embolized. ? For patients with and without delayed haemorrhage after LPN, the mean tumour size was 2.7 cm and 3.3 cm (P= 0.31), the mean warm ischaemia time was 28.2 min and 14.3 min (P < 0.001), and the mean estimated blood loss 403.8 mL and 308.2 mL (P= 0.26), respectively. ? Percutaneous angiography showed renal artery pseudoaneurysm in 10 patients and arterial contrast extravasation in three patients, two of whom also had an arteriovenous fistula. ? Following embolization, creatinine levels remained stable in all patients.

CONCLUSIONS

? Clinically significant delayed postoperative bleeding after LPN occurs in a small percentage of patients. ? Angiography will accurately make the diagnosis of RAP or AVF and SAE is safe and effective procedure that allows for preservation of renal function.     

经皮肾取石术后严重出血的治疗时机分析

目的 探讨经皮肾取石术后严重出血的治疗时机.方法 回顾性分析2012年7月至2015年6月695例经皮肾取石术患者中20例术后出现严重出血并行介入治疗的临床资料.结果 经皮肾取石术严重出血的发生率为2.9％.术后至介入治疗的时间为7 ～ 36d.急性出血者9例,常发生在术后7d后,表现肾造瘘拔后短时间内出血迅速,持续鲜红色血尿;缓慢失血8例,表现为术后7d内出现尿液、引流管液暗红,反复或持续性发生,最长至术后15d,1例伴感染性休克.急诊返院2例,表现出院后突发肉眼血尿伴血块.肾造瘘管周渗血1例,发生在术后1周左右.肾动脉造影结果阳性19例,假性动脉瘤占66.1％,小动脉分支出血、肾动静脉瘘分别为5例和2例,均行肾动脉栓塞治疗成功,其中1例2次栓塞成功;1例未发现异常,保守治疗成功.结论 根据术后出血的时间、特点、原因和量化指标,有助于确定最佳的治疗时机并指导临床实践.     

Angiographic embolization for arrest of bleeding after penetrating trauma to the abdomen

BACKGROUND: Angiographic embolization is an effective technique to control bleeding after blunt trauma to the liver or pelvis. Its role in penetrating trauma to the abdomen has not been studied. METHODS: From January 1992 to May 1998, 40 patients underwent angiography for bleeding resulting from intra-abdominal penetrating injuries (33 gunshot wounds, 7 stab wounds). Angiographic embolization of intraperitoneal or retroperitoneal vessels was performed by standard angiographic techniques with gelatin sponge and/or coils. Data were extracted from medical records, radiology data bank, trauma registry, and morbidity/mortality records, and compared by Student's t test and chi-square test. The main outcome measures were failure of angiographic embolization to control bleeding and complications of angiographic embolization. RESULTS: Angiography was performed during a course of nonoperative management in 6 patients (group A), because of failure to control bleeding surgically in 23 (group B), and because of late vascular complications after an initially successful operation in 11 more (group C). In 32 patients, angiography revealed active bleeding; 29 (91 %) underwent successful angiographic embolization. Of the remaining 3 patients, 2 were successfully managed surgically (1 each from groups A and B) and 1 died despite multiple surgical maneuvers (group B). One patient who developed postoperatively a large, bleeding superior mesenteric artery pseudoaneurysm, suffered extensive bowel necrosis after angiographic embolization. No other significant complication was related to angiographic embolization. CONCLUSIONS: Angiographic embolization after penetrating injuries to the abdomen is safe and effective for a small number of selected patients. It is a valuable tool for bleeding control when surgery has failed. It may be ideal for control of late vascular complications when reoperation is not desirable. It may prove to be a useful adjunct in the nonoperative treatment of selected injuries.     

Angiography in determining the cause and treatment of gastrointestinal bleeding.

Angiography is useful in the diagnosis of active gastrointestinal bleeding if the rate is greater than 0.5 mL/min. For upper gastrointestinal bleeding, endoscopy is the preferred initial investigation and angiography is used for diagnosis only if the site of bleeding is still obscure. Angiography is the preferred method for investigation of massive lower gastrointestinal bleeding if results of sigmoidoscopy are negative. Vasopressin infusion is most useful for control of bleeding from esophageal varices, erosive gastritis and diverticular disease of the colon. Embolization with Gelfoam or clot is possible for massive hemorrhage from a single source in poor-risk patients. This is most successful for gastric or duodenal bleeding since the collateral blood supply prevents infarction. Some of the methods and complications of embolization are discussed and examples are given. Standard surgical principles should still apply in most cases.