Endoscopic clipping for gastrointestinal bleeding: emergency and prophylactic indications

Bleeding peptic ulcer

Bleeding peptic ulcers are the main cause of upper gastrointestinal bleeding [8]. Early endoscopy should be performed within 24 h of the patient’s admission [9-11]. Endoscopic intervention reduces the risk of rebleeding and the requirement for surgery, and improves mortality. Endoscopic therapy options include: injectable methods (dilute epinephrine solution, ethanol, ethanolamine and polidocanol); thermocoagulation methods (heater probes, electrocautery probes, coagulation forceps and argon plasma coagulation); and mechanical hemostasis (clipping and elastic banding). Established indications for endoscopic hemostasis include actively bleeding ulcers and ulcers with a non-bleeding visible vessel (Forrest classes Ia, Ib and IIa). Ulcers with an adherent clot (Forrest IIb) should receive individualized management, and may or may not be suitable for endoscopic hemostasis [12].

Epinephrine injection alone is less effective than thermocoagulation or clipping as monotherapy [9,10]. A randomized controlled trial (RCT) with 47 patients showed that endoscopic clipping as monotherapy is comparable to epinephrine injection combined with thermocoagulation via a bipolar cautery probe [13]. A meta-analysis of prospective RCTs reinforced that epinephrine injection alone results in higher rates of rebleeding and requirement for surgical intervention compared with clipping as monotherapy, and also failed to show a clear benefit of endoclipping compared to thermocoagulation [14,15]. Another RCT showed that monopolar hemostatic forceps with soft coagulation, a relatively new contact thermal method, was associated with a higher hemostasis success rate, lower recurrent bleeding rate and a shorter procedure time compared to endoclip methods [16]. According to the European Society for Gastrointestinal Endoscopy (ESGE) guidelines, endoclip can be used as monotherapy for ulcers with a non-bleeding visible vessel. In ulcers with active bleeding, combination therapy (injection of epinephrine in conjunction with either a mechanical or thermal method) is recommended [10].

Rates of hemostatic success using endoclip technology vary from 85-100% and a failure rate of up to 20% has been described in those patients with chronic ulcers [2]. Other factors that increase the difficulty of endoclip placement and predict failure of hemostasis include: large ulcers (over 2 cm), a large non-bleeding visible vessel within the lesion, ulcers located in the lesser gastric curvature or in the posterior duodenal wall, and excessive blood in the stomach [12].

Mallory-Weiss tear

A Mallory-Weiss tear or Mallory-Weiss syndrome (MWS) is a laceration of the mucosa at the esophagogastric junction, caused by repetitive vomiting, usually treated conservatively [17]. Endoscopic treatment is indicated in cases of active bleeding on examination or presence of a deep laceration, and may consist of injection of dilute epinephrine, thermocoagulation or placement of endoclips [9]. A study published in 2016, including 168 patients with MWS, demonstrated that endoclip-based therapy and endoscopic band ligation achieved higher success rates than injection therapy in preventing re-bleeding (96.4%, 88.9% and 71.4%, respectively, P=0.013) [18]. Although there are no RCTs demonstrating superiority of one particular method over another [9], endoscopic clipping causes less tissue damage, reducing the risks of tissue necrosis and perforation [19]. Despite this, in a prospective RCT, Young-Seok et al demonstrated that endoscopic band ligation and endoclipping are equally effective and safe for the management of active bleeding in patients with MWS [20].

Dieulafoy’s lesion

Dieulafoy’s lesion is a vascular abnormality in which there is exposure of a dilated submucosal vessel that erodes, causing massive bleeding. These lesions are relatively rare, usually located in the stomach and duodenum, and have a high risk of rebleeding (19%) [21]. Mechanical hemostasis with endoclipping is superior to injection methods and is considered effective for both the control of acute bleeding and reducing the rate of rebleeding [22]. In addition, the metal clip has the advantage of marking the lesion site in the event of rebleeding and subsequent need for further endoscopic, radiological or surgical intervention [9]. Mechanical methods, including clipping or elastic band ligation, should be the first choice in this type of lesion [21], although there is no consensus regarding their superiority in relation to thermocoagulation. A meta-analysis that included 5 studies compared primary hemostasis and rebleeding rates in 162 patients with upper gastrointestinal bleeding caused by a Dieulafoy’s lesion treated with endoscopic band ligation (n=75) or endoclipping (n=87). The results showed that both mechanical methods achieved similar rates of primary hemostasis (0.96 vs. 0.91, relative risk [RR] 1, 95% confidence interval [CI] 0.96-1.05) and rebleeding (0.06 vs. 0.17, RR 0.37, 95%CI 0.12-1.09) [23]. ESGE guidelines recommend that endoscopic hemostasis of these lesions should be achieved using thermal or mechanical methods, either alone or in combination with epinephrine injection [10].

Colonic diverticular bleeding

Arterial diverticular bleeding typically occurs at the base or edge of the diverticulum and presents as painless hematochezia. Diverticular bleeding is the most common cause of lower gastrointestinal bleeding. Colonoscopy with endoscopic hemostasis is safe and effective in 87-100% of cases [24]. Patients with diverticular bleeding are candidates for endoscopic treatment if there is active bleeding, a non-bleeding visible vessel or adherent clot. If untreated, these bleeding stigmata predict a high risk for rebleeding (84%, 60% and 43%, respectively) [25]. Endoclipping and thermocoagulation are effective treatments for diverticular bleeding; however, mechanical therapy is preferable for vessels located at the base of the diverticulum, which carry a lower risk of perforation [26]. Direct clipping onto the vessel, when feasible, is the treatment of choice in colonic diverticular hemorrhage. The clip may also serve as an anatomical marker for angiography or surgery in the case of recurring bleeding. If clipping is not possible, elastic band ligation is an acceptable alternative [27]. A recent cohort study enrolling 108 patients evaluated the risk of recurrent bleeding up to a year post elastic band ligation versus endoscopic clipping. The authors concluded that band ligation had better outcomes than clipping during long-term follow up [28]. There are no RCTs available that compared these 2 endoscopic modalities.

OTSC

The OTSC, like a “bear-trap”, is a novel endoscopic clipping device that has generally been used in the setting of recurrent or refractory bleeding that cannot be controlled by TTS endoclips [29]. Kirschniak et al demonstrated the clinical utility of the OTSC system for the first time in 2007 to close a gastric perforation [30]. Its main advantages include the strong and sustained compression forces that can be exerted and the capacity to capture a larger volume of tissue compared with standard endoclips [31]. Other applications reported in retrospective studies and case series include the closure of perforations, fistulas and anastomotic leaks [32-34].

In a prospective multicenter trial, 67 patients with recurrent peptic ulcer bleeding following initially successful hemostasis were enrolled and randomized, with 34 receiving standard therapy (TTS clips or thermal therapy plus injection with diluted adrenaline) and 33 receiving OTSC. Treatment with OTSC was significantly superior to standard therapy in terms of bleeding control. Clinical success (i.e., no further bleeding) was also significantly higher in the OTSC versus the standard therapy group (84.9% vs. 42.4%; P=0.001), with a relative risk reduction for further bleeding of 73.6%. There were no significant differences between the groups in mortality or requirement for surgical or angiographic salvage therapy [35]. A recent review study of 1517 cases summarized the clinical results of 30 articles. The average clinical success rate of the OTSC for refractory bleeding was 85% (473/559 cases), 85% (297/351) for perforation, 52% (200/388) for fistulas, and 66% (64/97) for anastomotic dehiscence [7]. A disadvantage of this device in active bleeding is that the endoscope must be removed to mount the OTSC system distally, with reintroduction required to deploy the clips, potentially impairing a good view of the target lesion. Some important clinical studies reviewing endoclip treatment for bleeding lesions in the gastrointestinal tract are listed in Table 1.

Table 1 Relevant clinical studies of endoclip treatment for bleeding lesions in the gastrointestinal tract

Esophagus and stomach

The use of the endoclip in resection of esophageal and gastric lesions is justified in situations post endoscopic resection where an active uncontrolled bleeding vessel or non-bleeding protuberant vessel is present at the resection site, or in suspected perforation. In other situations, such as patients on anticoagulation or antiplatelet therapy, in the presence of portal hypertension [36] or for lesions with a large pedicle or with large vessels, prophylactic endoclipping can also be considered in the esophagus or stomach. There are currently no high quality controlled clinical trials that recommend the preventive use of endoclips in these segments.

Duodenum

Duodenal polyps or lesions, although rare, are increasingly being identified by endoscopy. The most frequent duodenal lesions are lipomas, gastrointestinal stromal tumors, neuroendocrine tumors, Brunner’s gland hamartomas, Peutz-Jeghers polyps and adenomas [37]. Endoscopic resection of duodenal lesions is technically more difficult and has a higher risk of immediate and delayed bleeding complications (14%) and perforation (1.9%) compared to elsewhere in the gastrointestinal tract. Most studies involve resection of superficial non-ampullary epithelial tumors (0.1-0.4%) [38]. Duodenal polypectomy or mucosectomy is technically challenging, because of the narrow lumen, the presence of peristalsis, loop formation and difficulty in device positioning, especially for lesions distal to the ampulla of Vater. Other contributing factors that render treatment difficult include the presence of a thinner muscularis propria layer, an extensive vascular plexus supplied by the gastroduodenal artery and the pancreatoduodenal arch, and the presence of Brunner’s glands in the submucosa, which may hinder the lifting of the target lesion during catheter injection [37,38].

Multiple biopsies of superficial duodenal lesions should be avoided, as the resulting fibrosis may impair the lifting and removal of the target lesion during endoscopic treatment [39]. In the case of endoscopic therapy in the duodenum, polypectomy may be indicated in lesions ≤10 mm, and mucosectomy in lesions >10 mm.

The effect of prophylactic mucosal closure after endoscopic resection is unclear. A study of 91 lesions enrolling 91 patients who underwent endoscopic resection for superficial non-ampullary duodenal tumors investigated the efficacy of prophylactic clipping for the prevention of late complications. Individuals were allocated to either an immediate clipping group or to a non-clipping group. Delayed bleeding occurred in 1 patient (2.1%) and delayed perforation occurred in 1 patient (2.1%) within the clipping group, whereas within the non-clipping group, delayed bleeding occurred in 6 patients (13.6%, P=0.053) and delayed perforation occurred in 3 patients (6.8%, P=0.350). Therefore, prophylactic clipping showed a tendency towards lower complication rates [40]. In view of the above, for resected duodenal lesions >10 mm, the recommendation is to close the resection site using endoclips to prevent delayed bleeding and perforation [38-41]. Notably, RCTs are needed to assess whether systematic clipping prophylaxis reduces the incidence of late complications after endoscopic resection. Given the infrequency of duodenal lesions and the difficulty of endoscopic resection, there have been few large-scale studies to date.

Colon

Various polypectomy and mucosectomy techniques within the colon have been developed with advances in imaging technology and devices [42]. As a result, adverse events, especially bleeding and perforation, have also increased in absolute numbers. A meta-analysis that included 6529 colonoscopies involving mucosectomies showed a delayed bleeding rate of 4% and a perforation rate of 1.1% [43]. Patient-related perforation risk factors include the presence of diverticulosis, inflammatory bowel disease and corticosteroid use. The endoscopist’s experience and polyp morphology also influence the rate of adverse events. Risk factors described for bleeding after polypectomy include polyp size above 2 cm, right sided colonic polyps, intraprocedural bleeding, use of antithrombotic agents, and exposed vessels in the submucosa after resection [43,44]. The use of microprocessor-controlled electrocautery confers a lower risk of bleeding and perforation after polypectomy [45,46].

Endoclipping is not routinely indicated in the resection of colonic polyps <20 mm, because its use does not reduce the risk of postprocedural bleeding [42,47]. In an RCT enrolling 1499 patients, prophylactic clipping was not necessary to prevent post-polypectomy delayed bleeding for polyps <2 cm [48]. In another study, 1098 patients who had polyps ≥10 mm were randomly divided into 2 groups (clipping or non-clipping groups). The authors found that placement of prophylactic endoclips did not affect the incidence of delayed post-polypectomy bleeding [49]. On the other hand, a recent multicenter RCT with 919 patients, which evaluated the influence of endoscopic clipping of post-polypectomy defects >20 mm, showed that clipping reduced the overall risk of delayed hemorrhage. This benefit was limited to lesions in the proximal colon, where the bleeding risk was significantly lower when clips were applied compared to the control group (3.3% compared to 9.9%; P<0.001) [50].

A recent meta-analysis of 9 RCTs (7197 colorectal lesions) assessed the efficacy of clipping in preventing bleeding after polypectomy. The results showed that the benefit of clipping in reducing bleeding was significant for large polyps (clipped vs. unclipped, odds ratio [OR] 0.54, 95%CI 0.30-0.97; P=0.041), and proximal lesions >20 mm (clipped vs. unclipped, OR 0.34, 95%CI 0.19-0.65; P=0.021). Thus if such lesions were not clipped, there would be 4-fold increase in the baseline risk of post-polypectomy bleeding [51]. According to 2 American guidelines, in non-pedunculated polyps >20 mm, whose Paris classification is of type 0-Is or type 0-II [52], the closure of the endoscopic resection area with clips decreases the incidence of late bleeding, especially in the right colon [42,43] (Fig. 1).

Figure 1 Underwater mucosectomy of a 22-mm lateral spreading tumor in right colon. Three hemoclips were positioned as post-polypectomy bleeding prophylaxis

In pedunculated polyps (Paris classification Ip), prophylaxis for bleeding post-polypectomy with endoclips is recommended in those in which the polyp head is ≥20 mm, or if the pedicle is ≥10 mm [46,53]. The American Society for Gastrointestinal Endoscopy (ASGE) guidelines recommend the prophylactic application of endoclips for resection of polyps with a pedicle greater than 5 mm [40]. ESGE guidelines also suggest the use of a diluted epinephrine injection as an alternative or in association with mechanical hemostasis in the pretreatment of polyps with a large pedicle [46].

The SCALP (Study on ComplicAtions of Large Polypectomy) study, which evaluated 1504 patients, found that prophylactic endoscopic clipping was cost-effective for polyps >10 mm in patients on anticoagulant or antiplatelet therapy. Anticoagulant therapy was the risk factor most associated with delayed bleeding [53]. In the case of periprocedural bleeding or exposure of submucosal vessels, endoscopic coagulation (snare-tip soft coagulation or coagulation forceps) or mechanical therapy (clips or endoloop) is recommended, with or without the combined use of dilute adrenaline injection [46] (Fig. 2).

Figure 2 Polypectomy of a pedunculated polyp with bleeding during the procedure and an exposed vessel. Diluted epinephrine injection (1:10000) and endoscopic clipping were performed

Signs predictive of perforation, such as the “target sign” (representing injury to the colon’s muscularis propria after mucosectomy and identified as a white to grey central circular disk on the transected surface of the specimen) should prompt consideration of endoscopic closure of the defect (Fig. 3).

Figure 3 “Target sign” in a resected specimen of colonic mucosectomy represents a predictive sign of perforation. Closure of the defect was achieved with 6 metal clips

Although the direct costs related to delayed bleeding are largely unreported, a Spanish prospective series of 1424 endoscopic mucosal resections analyzed the cost-effectiveness of prophylactic clipping of large colorectal lesions (≥2 cm). In this study, bleeding costs were higher in relation to the prophylactic strategy in average and high-risk settings (age ≥75 years, American Society of Anesthesiologists classification III-IV, lesion size ≥40 mm, aspirin use, location of the lesion proximal to transverse colon) [54]. A recent publication evaluated the economic impact of prophylactic endoscopic clipping on large colonic polyps. It suggested that prophylactic clip closure after endoscopic resection of large colonic polyps, particularly those within the right colon, is cost-effective, but only if the clip device costs under $100.00 (USD) [55].