Gastrointestinal Intervention 2018; 7(3): 112-122
Published online October 31, 2018 https://doi.org/10.18528/gii180031
Copyright © International Journal of Gastrointestinal Intervention.
Ki Bae Bang, and Hyun Deok Shin*
Department of Internal Medicine, Dankook University College of Medicine, Cheonan, Korea
Correspondence to:*Department of Internal Medicine, Dankook University Hospital, Dankook University College of Medicine, 201 Manghyang-ro, Dongnam-gu, Cheonan 31116, Korea.
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Endoscopy is a safe and effective modality for the diagnosis and treatment of lesions in the gastrointestinal (GI) tract. During the last few decades, improvements in image quality and technical advances have led to the widespread use of endoscopy in various medical fields. Because it is relatively noninvasive and generally safe, the role of endoscopy has been emphasized in morbid patients with postoperative GI bleeding. However, there has been concern about the safety and complications of endoscopy. Here, we review endoscopic management of GI bleeding as a complication of surgery and therapeutic endoscopy.
Keywords: Endoscopy, Hemorrhage, Hemostasis, Postoperative complications
Endoscopic resection such as endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) can be used to treat various gastrointestinal (GI) tract neoplasms including premalignant and malignant lesions of the esophagus, stomach, duodenum, and colorectum.1 However, postprocedural bleeding or perforation are uncommon complications of endoscopic treatment and can be fatal. Fortunately, most iatrogenic complications such as post EMR or ESD bleeding can be controlled by endoscopic interventions.2 Moreover, endoscopic modalities can be used to treat postoperative complications such as GI bleeding, anastomosis leakage, and perforation and prevent unnecessary reoperations.3 Endoscopy enables identification of the bleeding focus and management of GI bleeding. Hemostasis of suture-line bleeding can be achieved using various endoscopic modalities (e.g., injection, mechanical, and thermal therapy).4 In this article, we review the endoscopic management of procedure-related and postoperative hemorrhage.
Endoscopic resection is minimally invasive and as effective as surgery for early gastric cancer.5 Technical advances have led to the widespread use of ESD as a first-line therapy for early gastric cancer without evidence of metastasis.6 ESD is associated with higher rates of
Post ESD bleeding is the most frequent adverse event associated with ESD.11 The incidence of bleeding after gastric ESD is 0.6% to 15.6%,12–26 and the overall incidence of postprocedural bleeding is 5.1%.27 The risk of bleeding after ESD is higher in the stomach compared to other sites.11 Most instances of post ESD bleeding occurred within 24 hours of ESD, but in some cases occurred up to 4 weeks later.19,24 Comorbidities, lesion size (≥ 4 cm), antithrombotic therapy, lesion location (lesser curvature, upper stomach), and a prolonged procedure are risk factors of delayed bleeding after gastric ESD.12,13,27,28
Intraprocedural bleeding is more frequent in the mid and upper stomach due to the greater diameter of submucosal arteries compared to the lower stomach. Delayed bleeding is thought to be more frequent in lesions in the lower stomach, likely due to antral peristalsis, bile reflux, and more careful hemostasis in lesions in the upper or mid-stomach.15,19,20 However, in a recent meta-analysis, the overall bleeding rate did not significantly differ according to lesion location, but delayed bleeding was associated with lesions in the upper stomach.27 Further studies are required for this issue.
Anti-acid therapy can reduce the rate of postprocedural bleeding after EMR.29 In a meta-analysis of five studies involving 506 patients, the rate of delayed bleeding was reduced in patients treated with a proton pump inhibitor compared to an H2 receptor antagonist (H2RA) (odds ratio [OR], 0.41; 95% confidence interval [CI], 0.20–0.85). Regarding EMR, there were no significant differences in the rate of delayed bleeding between the two groups.30 Muco-protective agents may enhance artificial ulcer healing after gastric ESD. However, the rate of delayed bleeding was not reduced by faster ulcer healing in several randomized controlled trials (RCTs) and a meta-analysis27,31–33 Routine second-look endoscopy after ESD remains controversial.34–37 However, in the meta-analysis of three RCTs involving 854 patients, second-look endoscopy had no advantage for the prevention of post ESD bleeding.38 Moreover, in a recent meta-analysis of 12 non-randomized studies, second-look endoscopy had no effect in reducing delayed post ESD bleeding.39 However, the role of second-look endoscopy is unclear in high-risk patients according to the Forrest classification.24,35,40
Intraprocedural bleeding is a common complication during ESD that necessitates repeated hemostasis. Identification of vessels during submucosal dissection and prophylactic hemostasis is important to prevent intraprocedural bleeding (Fig. 1). Prophylactic coagulation and minor oozing from small vessels can be managed with dissection knives; hemostatic forceps are useful for controlling active arterial bleeding.19,41–43 If coagulation forceps fail to control bleeding before hemoclip application, the bleeding site must be fully exposed.1 Because their use can hinder further dissection, hemoclips should be reserved for uncontrolled bleeding or after completion of dissection.44 After dissection, routine coagulation of all visible vessels should be performed to reduce the risk of delayed bleeding.19 In most cases, delayed bleeding can be managed using standard endoscopic hemostasis methods (Table 1).12–14,16–20,23,24 The degree of ulcer healing might affect the treatment decision. Hemoclips or coagulation forceps are useful in the early phase of ulcer healing (Fig. 2), whereas injection therapy can be useful in the late phase due to hardening of the ulcer floor over time.45
Globally, colorectal cancer is the third most common cancer.46 Endoscopic treatment of colorectal neoplasm has reduced the incidence of colorectal cancer and the rate of cancer-related mortality.47–49 Endoscopic treatment is now used widely, although there is some concern over complications, including bleeding (the most frequent complication) and perforation. Postprocedural bleeding is potentially life threatening50 and can require hospitalization, repeated colonoscopy, and blood transfusion.51
The incidence of postprocedural bleeding in the colon is 0.6% to 8.1%.52–63 Postprocedural bleeding can occur immediately or be delayed by up to 2 to 4 weeks.64,65 The risk of bleeding is higher in patients with comorbidities (cardiovascular disease or hypertension), large polyps (> 10 mm), polyps located in the right colon, and polyps with a large stalk (≥ 5 mm).66,67 Antiplatelet or anticoagulant therapy is reportedly associated with delayed bleeding.56,68 The rate of delayed bleeding is estimated to increase by 13% for each 1 mm increase in polyp diameter.69 However, despite the larger polyp size, the bleeding risk for ESD is comparable to EMR according to a recent meta-analysis.70
Prophylactic management of postprocedural bleedingAlthough prophylactic clipping after polypectomy has been used to reduce the rate of delayed bleeding, the evidence for the efficacy of this practice is weak, particularly for small polyps. A recent RCT of 3,365 polyps of < 2 cm demonstrated no decrease in bleeding in patients with compared to those without prophylactic clips.71 Moreover, in a recent network meta-analysis, the rate of delayed bleeding was not reduced by prophylactic injection, mechanical, or coagulation therapy, but the risk of early postpolypectomy bleeding was reduced by mechanical or injection therapy.72 However, in two RCTs involving large polyps (≥ 2 cm), mechanical prophylaxis such as a detachable snare or hemoclips in combination with adrenaline injection reduced the rate of delayed bleeding compared to adrenaline injection alone.73,74 Prophylactic epinephrine injection might be useful for large pedunculated polyps.75 Clip closure is a feasible alternative, which is reportedly associated with a decreased frequency of delayed bleeding (1.8% for ≥ 2 cm and 1.1% for 1–4 cm polyps) compared to no-closure (9.7% for ≥ 2 cm and 6.9% for 1–4 cm polyps).76,77 This is in agreement with a report that clip closure of ESD-induced mucosal defects reduces the risk of delayed bleeding (0% vs 8.2%;
In cases of intraprocedural bleeding after polypectomy, hemostasis can be achieved by snare-tip soft coagulation or using coagulation forceps or hemoclips.79,80 Snare-tip soft coagulation achieved effective hemostasis in 91% of intraprocedural bleeding patients. Additional treatment with coagulating forceps is required if hemostasis is not achieved using the snare-tip alone. Hemoclips are not used during the procedure as they could interfere with further resection.79,81 Intraprocedural bleeding after removal of a pedunculated polyp can be managed using hemoclips. Grasping the remaining stalk with the snare is a feasible method to stop bleeding; maintain a clear visual field; and provide time for the application of clips, an injection needle, or hemostatic forceps.44 If intraprocedural bleeding is refractory to conventional modalities, the over-the-scope clip (OTSC) can be used; OTSC showed a 100% success rate for procedure-related bleeding after polypectomy, EMR, or ESD in the colon.82–86 The main disadvantage of OTSC is the need for withdrawal of the endoscope to load the device.1 Delayed bleeding is usually caused by ulceration at the polypectomy site,87 and can in most cases be resolved by standard endoscopic treatment (Table 2).52–62 Hemostasis with clips is preferred to avoid further thermal injury to the muscle layer (Fig. 3).1 Caution should be taken when using thermal therapy in the right colon because of coagulation syndrome and perforation.88
Endoscopic therapy is an established treatment modality for early esophageal squamous cell carcinoma and Barrett’s neoplasm.89 The mortality rate of patients with early esophageal cancer after endoscopic resection is comparable to that after surgery.90 Endoscopic resection is primarily indicated for lesions confined to the mucosal layer classified as M1 (epithelium) or M2 (lamina propria), because such lesions rarely show lymph node invasion.89 ESD is preferred over EMR due to its better
Surgical treatment is an essential part for the management of disease. And increasing demand for digestive surgery is estimated. Because the global incidence of digestive cancer has been increased, largely due to the population aging and population growth.95 Moreover, the number of bariatric operations performed globally is increasing due to the increasing prevalence of obesity.96,97 Although postoperative bleeding is an uncommon complication and mostly self-limiting, complications from surgery should not be ignored98 considering the increasing burden of digestive surgery.95 Moreovcer, postoperative bleeding can prolong hospital stay and result in morbidity and mortality, particularly in critically ill patients.99–101
Postoperative bleeding can be intraluminal or extraluminal. Indeed, a significant proportion of hemorrhage is extraluminal102–105 and combined intra/extra-luminal bleeding is not uncommon.106–109 Intraluminal GI bleeding can be caused by a ruptured pseudoaneurysm secondary to an anastomotic leak after pancreaticoduodenectomy. Endoscopy is a reliable modality for the management of intraluminal GI bleeding. However, in case of bleeding involving extraluminal component, endoscopy can fail to detect the exact site and amount of hemorrhage delaying the appropriate intervention such as angiographic embolization or surgery.109,110
Making decisions regarding treatment for postoperative bleeding is complicated and should take into consideration the type of surgery, accessibility of the bleeding site, time after surgery, and comorbidities.110 Because postoperative bleeding is associated with significant morbidity and mortality, the risks and benefits of therapeutic interventions, including conservative modalities, should be considered.4
Postoperative GI bleeding can be caused by surgery itself or a surgery-related condition (Fig. 4). Marginal ulcers may develop from mucosal ischemia originating from perfusion defects, anastomosis tension, or sutures material.111 Although bleeding from suture line is a major source of intraluminal GI bleeding,112 stress-related mucosal damage can cause upper GI bleeding.4 Gastritis, duodenitis, gastric ulcer, and duodenal ulcer can be caused by colorectal or non-GI surgery such as cardiothoracic or vascular surgery.4,99,113 Mallory-Weiss tear can be caused by postoperative nausea and vomiting.114,115 Importantly, comorbidities are closely related to the risk of postoperative bleeding. The number of postoperative complications is associated with the number of comorbidities in a study of gastric cancer patient.116
The incidence of postoperative bleeding varies from 0.22% to 8.4% according to the type of surgery and the definition used.98,102,104–109,117–128 The presentation of postoperative bleeding varies according to its site and extent, from an asymptomatic decrease in hemoglobin level to overt signs of hemorrhage and hemodynamic instability. Hematemesis is a common clinical presentation after GI surgery. Hematemesis was reported in 73% of patients with GI bleeding in a retrospective study involving 933 patients who underwent Roux-en-Y gastric bypass (RYGB) surgery. The frequent hematemesis may be related to the small gastric pouch with a limited gastric reservoir.117
In the literature, endoscopic treatment was performed in 4.5% to 100% of patients with intraluminal GI bleeding and has a success rate of 20% to 100% (Table 3).98,102,104–107,109,117–122,124–128 Hemoclips, epinephrine injection, fibrin glue, argon plasma coagulation, or a coagulation grasper or heater probe may be used to control GI hemorrhage.98,102,104,106,107,109,117–122,124–127
The success rate of endoscopic treatment is related to the type of surgery. The different accessibility to the bleeding focus is a main reason for this (Fig. 5). Endoscopic treatment is relatively difficult for RYGB or pancreaticoduodenectomy compared to gastrectomy or colectomy.106,107,109,118,119 The hemostasis success rate after gastrojejunostomy is favorable.117,118,126 Moreover, a bleeding focus in the anastomotic ring may be associated with the success of hemostasis. In a retrospective study of 16,591 patients who had undergone gastrectomy, the hemostasis success rate was highest on the anterior wall (100%) and lowest on the posterior wall (50%) of the anastomotic ring.120 Therefore, easy accessibility with a right angle is crucial part for successful hemostasis.
Endoscopy for postoperative GI bleeding enables identification of the bleeding site, treatment of the bleeding, and estimation of the risk of rebleeding.114 Although there was concern about a risk of anastomotic disruption and perforation during endoscopic procedures performed immediately after surgery,129 endoscopy was reported to be safe without significant complications even in the early postoperative period in recent studies.117,118,121,124,130
In a randomized controlled study of patients with recurrent ulcer bleeding, the success rates of endoscopic and surgical treatment were not significantly different, and complications were less common after endoscopic treatment.131 Endoscopic hemostasis should initially be attempted for postoperative intraluminal hemorrhage. The hemostatic technique for postoperative bleeding does not different from that of conventional GI bleeding (Fig. 6). Clipping, epinephrine injection, electrocoagulation, and argon plasma coagulation are the available options for hemostasis.114 Hemoclips are more durable for anastomotic bleeding and do not injure tissue, unlike sclerosant injection or electrocoagulation, and can be used to manage anastomotic leaks or iatrogenic perforations.132 However, hemoclips application can be technically difficult due to a poor axis caused by anatomical distortion after surgery. No RCT has compared the therapeutic efficacies of the various modalities for postoperative hemorrhage. However, Lee et al112 reported that the rate of rebleeding was significantly higher after epinephrine and/or heater probe coagulation than after hemoclip application (33% vs 5%). In cases of diffuse bleeding in which hemostasis is not achieved by conventional modalities, hemospray can be used in the early postoperative period.133 For the refractory bleeding, the OTSC has an acceptable success rate.134
Sleeve gastrectomy is the most commonly performed procedure worldwide (45.9%), followed by RYGB (39.6%).96 Bleeding can arise from multiple sites (e.g., anastomoses, staple lines, the pouch, the contiguous small intestine, the excluded stomach, or the bypassed small intestine).135 Early bleeding usually occurs at the staple lines of the gastrojejunal anastomosis and has an incidence of 1% to 5% after RYGB and 0% to 8% after sleeve gastrectomy. Late bleeding is usually secondary to an anastomotic ulcer. Upper endoscopy can be performed to access the gastrojejunal anastomosis.136
However, the diagnosis can be particularly difficult when the source of the bleeding is in the bypassed gastric remnant, proximal duodenum, or biliopancreatic limb.123 Excluded GI segments after Roux-en-Y reconstruction can be accessed using double-balloon enteroscopy.135 Alternatively, laparoscopic- or laparotomy-assisted endoscopy can be performed.137 Bleeding at the gastrojejunal anastomosis can be treated endoscopically in conjunction with standard hemostatic modalities (epinephrine injection, thermal therapy, and hemoclipping). Hemoclipping is the preferred method, and use of thermal therapy at the staple line and anastomosis site requires caution due to the risk of tissue injury.136
Pancreaticoduodenectomy is a complex, high-risk surgery for tumors of the pancreatic head and other periampullary structures. The incidence of postpancreaticoduodenectomy hemorrhage is 2.5% to 20.2%,106 and the incidence of postoperative intraluminal GI bleeding is 2.2% to 8.4%.105–109 Standard pancreaticoduodenectomy involves distal gastrectomy, duodenectomy, partial pancreatectomy, partial choledochectomy, cholecystectomy, and proximal jejunectomy. Therefore, the procedure involves three anastomoses (gastrojejunostomy, hepaticojejunostomy, and pancreaticojejunostomy). Pylorus-preserving pancreaticoduodenectomy preserves the stomach, pylorus, and the proximal duodenum, which is anastomosed to the jejunum.138
Bleeding can occur at the gastrojejunostomy, hepaticojejunostomy, pancreaticojejunostomy, and/or duodenojejunostomy suture line. In cases of suspected GI bleeding, endoscopy is used to identify the bleeding site and achieve hemostasis. However, the success rate of endoscopic treatment after pancreatoduodenectomy is less than 60%,106,107,109 possibly due to interference by blood clots in the stomach and because of difficulty accessing the pancreaticojejunostomy suture line if the field of view is obscured by active bleeding.139
Postoperative and procedure-related bleeding are uncommon complications, but can prolong the hospital stay and result in significant morbidity and mortality. Endoscopic modalities are first-line options for surgery- or procedure-related GI bleeding as they enable localization and the control of bleeding. However, data on the efficacy of the available treatment modalities are insufficient. Moreover, the efficacy of these modalities is affected by the postoperative anatomy and bleeding location. Treatment decisions should take into consideration, among other factors, the type of surgery or endoscopic procedure. Hemostasis can be achieved using various endoscopic modalities, which have acceptable success rate and safety profiles.
No potential conflict of interest relevant to this article was reported.
Studies Reporting on the Procedure Related Bleeding in the Stomach
Study | No. of lesions | Indication | Treatment | Incidence (%) | Mortality (%)* | Risk factors | Bleeding control | Success (%)† | Rescue therapy (n) |
---|---|---|---|---|---|---|---|---|---|
Yano et al13 (2017) | 1,767 | EGC | ESD | 8.5 | 0 | Specimen ≥ 40 mm, antithrombotic agents | Coagulation forceps and epinephrine injection | 100 | |
Na et al24 (2015) | 706 | EGC, adenoma | ESD | 13 | 0 | Specimen size | Electrocoagulation and/or hemoclipping | 100 | |
Koh et al12 (2013) | 1,166 | EGC, adenoma | ESD | 5.3 | 0 | Specimen ≥ 40 mm, antithrombotic agent | Coagulation forceps or hemoclipping | 100 | |
Lim et al23 (2013) | 1,461 | EGC, adenoma | ESD | 4.4 | 0 | Cumulative ESD time | Endoscopy | 100 | |
Yoshio et al18 (2013) | 1,310 | EGC | ESD | 5.3 | 0 | Heparin replacement | Endoscopy | 98.5 | Surgery (1) |
Chung et al14 (2009) | 1,000 | EGC | ESD | 16.2 | 0 | Upper stomach, tumor > 40 mm, recurrent lesion, flat lesion | Endoscopy | 99 | Surgery (1) |
Lim et al16 (2012) | 1,591 | EGC, adenoma | ESD | 5.9 | 0 | EGC, comorbidity, specimen ≥ 40 mm | Endoscopy | 96.8 | Embolization (3) |
Toyokawa et al17 (2012) | 1,123 | EGC, adenoma | ESD | 5.0 | 0 | Age ≥ 80, long procedure time | NA | NA | |
Takizawa et al19 (2008) | 1,083 | EGC | ESD | 5.8 | 0 | Upper stomach, no post-ESD coagulation | Endoscopy | 100 | |
Miyahara et al20 (2012) | 1,190 | EGC, adenoma | ESD | 6.9 | 0 | Lower stomach, large resection size, scar in the lesion | Hemoclipping, electrocoagulation | 98.7 | Surgery (1) |
EGC, early gastric cancer; ESD, endoscopic submucosal dissection; NA, not applicable.
Bleeding related mortality.
Success rate for endoscopic treatment.
Studies Reporting on the Procedure Related Bleeding in the Colorectum
Study | No. of patients (no. of lesions) | Indication | Treatment | Incidence (%) | Mortality (%)* | Risk factors | Bleeding control | Success (%)† | Rescue therapy (n) |
---|---|---|---|---|---|---|---|---|---|
Okamoto et al52 (2017) | 451 (509) | Colorectal tumor | ESD | 3.1 | NA | Antithrombotic therapy, rectal lesion | Endoscopy (100%) | 92.9 | Surgery (1) |
Ogasawara et al53 (2016) | 124 (124) | Large lesion which |
ESD | 8.1 | 0 | Rectal lesion, arterial bleeding during ESD (≥ 3) | Endoscopy (60%), hemoclips, APC, coagulation forcep, conservative (40%) | 100 | |
Suzuki et al54 (2014) | 317 (327) | Lesion > 20 mm, lesion with fibrosis, residual cancer, tumor with chronic inflammation | ESD | 4.4 | 0 | Cecal lesion, intraprocedrual bleeding | Hemoclips | 100 | |
Terasaki et al55 (2014) | 363 (377) | Lesion > 20 mm, lesion with fibrosis, residual cancer, tumor with chronic inflammation | ESD | 6.6 | 0 | Rectal lesion | Hemoclips, coagulation forcep | 100 | |
Park et al56 (2018) | 3,887 (8,175) | Polys ≥ 5 mm | Polypectomy | 3.4 | 0 | Age < 50, immediate bleeding | Hemoclips, APC | 100 | |
Kwon et al57 (2015) | 1,745 | Polyps | Polypectomy | 1.2 | 0 | Polyps > 10 mm, pedunculated polyp, right side colon, high BMI | Hemoclips, APC | NA | |
Choung et al58 (2014) | 3,788 (5,981) | Polyps > 5 mm | Polypectomy | 1.1 | NA | Polyps > 10 mm, right side colon, endoscopist’s experience (<300 cases="" td=""> | NA | NA | |
Consolo et al60 (2008) | 1,038 (1,354) | Polyps | Polypectomy | 1.3 | 0 | Cardiac disease, tubular adenoma, polyp size | Endoscopy (100%), adrenalin injection, hemoclips, APC | 100 | |
Kim et al59 (2013) | 3,253 (7,447) | Polyps | Polypectomy, EMR, ESD | 1.3 | 0 | Polyps > 10 mm, pedunculated polyp, right side colon | Endoscopy (68%), conservative (32%) | 100 | |
Burgess et al61 (2014) | 1,172 | Polyps ≥ 2 cm | EMR | 6.2 | 0 | Proximal colon, IPB | Endoscopy (100%), thermal therapy, hemoclips, epinephrine | 100 | |
Metz et al62 (2011) | 288 (302) | LST > 2 cm | EMR | 7 | 0 | Right colon, use of aspirin, age | Endoscopy (48%), hemoclip, coagulation forcep, conservative (52%) | 80 | Embolization (1), surgery (1) |
EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; NA, not applicable; APC, argon plasma coagulation; BMI, body mass index; IPB, intraprocedural bleeding; LST, laterally spreading tumor.
Bleeding related mortality.
Success rate for endoscopic treatment.
Studies Reporting on the Postoperative Bleeding
Study | No. of lesions | Indication | Types of surgery | Bleeding site | Incidence (%) | Mortality (%)* | Risk factors | Bleeding control | Success (%)† |
---|---|---|---|---|---|---|---|---|---|
Lee et al120 (2017) | 16,591 | Gastric cancer | Gastrectomy | Anastomosis | 0.22 | 0.02 | Type of gastrectomy (subtotal) | Endoscopy (69%), surgery (17%), conservative (14%) | 64 |
Kim et al118 (2012) | 2,031 | Gastric cancer | Gastrectomy | Gastrojejunostomy (71%), gastroduodenostomy, esophagojejunostomy | 0.3 | 0 | B-II anastomosis, manual anastomosis | Endoscopy (85%), conservative (14%) | 100 |
Tanizawa et al121 (2010) | 1,400 | Gastric cancer | Gastrectomy | Gastroduodenostomy (50%), gastrojejunostomy, staple line of stomach, esophagojejunostomy | 0.43 | 0 | Lymph node dissection (≤ D1) | Endoscopy (83%), surgery (17%) | 100 |
Lim et al122 (2012) | 393 | Gastric cancer | Gastrectomy | Anastomosis | 2.8 | 0 | NA | Endoscopy (82%), conservative (18%) | 89 |
Jeong et al98 (2011) | 1,027 | Gastric cancer | Gastrectomy | Anastomosis (81%), pseudoaneurysm (19%) | 1.6 | 0 | Operating time ≥ 3 hr, BMI ≥ 26 kg/m2 | Endoscopy (12.5%), surgery (19%), conservative (69%) | 100 |
Kim et al104 (2008) | 1,485 | Gastric cancer | Gastrectomy | NA | 1.3 | 0 | Comorbidity, surgeon’s experience | Endoscopy (25%), conservative (75%) | 100 |
Park et al102 (2014) | 5,739 | Gastric cancer | Gastrectomy | Anastomosis | 0.8 | 0 | Male, comorbidity, previous abdominal-operation, palliative surgery | Endoscopy (28.6%), surgery (28.6%), conservative (42.9%) | 50 |
Fernández-Esparrach et al124 (2008) | 381 | Bariatric | RYGB | Anastomosis | 5.8 | 0 | NA | Endoscopy (27%), conservative (73%) | 100 |
Jamil et al117 (2008) | 933 | Bariatric | RYGB | Gastrojejunostomy (100%) | 3.2 | 0.1 | NA | Endoscopy (80%), conservative (20%) | 100 |
Rabl et al126 (2011) | 742 | Bariatric | RYGB | Gastrojejunostomy | 0.54 | 0 | NA | Endoscopy (75%) | 100 |
Fernández de Sevilla Gómez et al125 (2014) | 2,069 | Malignancy, IBD | Colectomy, ileal resection | Anastomosis | 3.17 | 0 | NA | Endoscopy (4.5%), surgery (13.6%), angiography (13.6%), conservative (27%) | 100 |
Lou et al128 (2014) | 2,181 | Rectal cancer | Anterior resection | Anastomosis | 0.3 | 0 | NA | Endoscopy (100%) | 100 |
Besson et al119 (2016) | 729 | Cancer, diverticular disease | Left colectomy | Anastomosis | 6.4 | 0 | Stapled anastomosis, diverticular disease | Endoscopy (78.7%), conservative (24.3%) | 100 |
Linn et al127 (2008) | 143 | Cancer, benign lesion | Left colectomy, anterior resection | Anastomosis | 4 | 0 | Surgery for benign lesion | Endoscopy (17%), conservative (83%) | 100 |
Feng et al106 (2014) | 840 | Malignancy, benign tumor | Pancreaticoduodenectomy | Gastrojejunostomy (29%), marginal ulcer (25%), cholangiojejunostomy, pancreaticojejunostomy | 3.3 | 1.1 | Male, end to side pancreaticojejunostomy, small pancreatic duct | Endoscopy (60%) | 47 |
Yekebas et al107 (2007) | 1,669 | Pancreas neoplasm, pancreatitis | Pancreaticoduodenectomy, pancreatectomy | Gastrojejunostomy, enteroenteric anastomosis | 2.2 | 0.1 | Pancreatic fisula | Endoscopy (42%) | 20 |
Wei et al105 (2009) | 628 | Periampullary lesion | Pancreaticoduodenectomy | Pancreaticogastrostomy (33%) | 2.2 | 0.9 | Pancreatic leakage intraabdominal-abscess | Endoscopy (38.1%) | NA |
Chen et al109 (2015) | 703 | Periampullary cancer | Pancreaticoduodenectomy | NA | 3.8 | NA | Pancreatic fistulae, abdominal infections | Endoscopy (26%) | 57.1 |
B-II, Billroth-II; NA, not applicable; BMI, body mass index; RYGB, Roux-en-Y gastric bypass; IBD, inflammatory bowel disease.
Bleeding related mortality.
Success rate for endoscopic treatment.
© The Society of Gastrointestinal Intervention. Powered by INFOrang Co., Ltd.