Bleeding
Perforation
Gastric polypectomy
EMR:
Esophagus
1.2% [101]
Gastric
1% [108]
ESD:
Esophageal
Gastric
Infection
Infectious adverse events associated with UGI endoscopy can be from an endogenous source (usually as a result of the procedure itself) or from an exogenous source (secondary to poor reprocessing of endoscopic devices). Infectious complications of UGI endoscopy are rare. The mechanism of endogenous infection is thought to be the result of normal GI flora gaining access to the circulation through areas of mucosal trauma or instrumentation during an endoscopic procedure.
A number of studies have examined the rate of bacteremia following UGI endoscopy. These studies drew blood cultures before and after an endoscopy was performed in an effort to identify bacteremia. The reported rate of bacteremia following diagnostic UGI endoscopy has been reported as 1–8% [11–14]. Bacteremia observed in these studies was usually transient, and the rate of clinically significant infectious symptoms or sequelae (e.g., endocarditis, meningitis, or abscess formation) was extremely low [15]. The current American Heart Association (AHA) and ASGE guidelines do not recommend the use of prophylactic antibiotics to prevent endocarditis in patients undergoing diagnostic UGI endoscopies [16, 17].
Infections from an exogenous source are exceedingly rare and have nearly all occurred as a result of a breach in the current guidelines for cleaning and disinfecting endoscopic equipment. A 2003 review published in Gastrointestinal Endoscopy identified 317 reported episodes of pathogen transmission over a 36-year period. Pathogens most frequently identified in these studies include Pseudomonas, Salmonella, H. Pylori, and Hepatitis B [18]. Recent concerns regarding dissemination of carbapenem-resistant Enterobacteriaceae (CRE) infections during GI procedures have predominantly been limited to ERCP procedures but could theoretically be transmitted by any endoscope [19]. The inherent complexity of the duodenoscope design used during ERCP is one of the primary factors leading to challenges with appropriate CRE disinfection [19]. As a traditional upper endoscope does not have this same level of complexity, transmission of CRE infection during an EGD is much less likely to occur [20].
Bleeding (Table 10.2)
Bleeding is a rare complication of diagnostic UGI endoscopy. Mallory–Weiss tears have been reported following <0.5% of upper endoscopies, usually as a result of the patient having coughing or retching during the examination [21, 22]. Most of these tears do not result in clinically significant bleeding. Management of Mallory–Weiss tears involves admission to the hospital for observation, intravenous fluids, and serial hemoglobin testing. Thrombocytopenia and coagulopathies have been shown to increase the risk of bleeding. There is no agreed upon platelet count recommended prior to performing a diagnostic UGI endoscopy. Some studies have suggested that upper GI endoscopy is safe in patients with platelet counts >20,000/mL [4, 23]. Video 10.1 demonstrated control of hemostasis with endoscopic hemoclips due to biopsy of an antral submucosal mass after unroofing.
Table 10.2
Reported bleeding and perforation rates for various upper endoscopic resection techniques
Risk factors | Odds ratios [95% CI] |
|---|---|
Age > 60 years old | 1.8 [1.6–1.9] |
ASA classification: | |
III | 1.8 [1.6–2.0] |
IV | 3.2 [2.5–4.1] |
V | 7.4 [3.2–17.6] |
Inpatient procedures | 1.5 [1.3–1.7] |
Involvement of trainee | 1.3 [1.2–1.4] |
Perforation (Table 10.2)
While an UGI endoscopy is the most common cause of esophageal perforation, its incidence as a result of a diagnostic UGI endoscopy is overall very rare. The risk of esophageal perforation during diagnostic UGI endoscopy has been reported as <0.04% [8, 9]. Risk factors associated with esophageal perforation include anterior cervical osteophytes, Zenker’s diverticulum, upper gastrointestinal malignancies, and esophageal diverticula. The morality rate of esophageal perforation is reported between 2 and 36% [2]. Early identification of esophageal perforations is essential and has been shown to reduce the morbidity and mortality associated with this complication [24, 25]. The most frequent symptom of esophageal perforation includes neck, chest, or abdominal pain [26–28]. Other frequently reported signs and symptoms include fever, dyspnea, crepitus, and leukocytosis [26]. If esophageal perforation is suspected, initial diagnostics should include thoracic and cervical radiographs, which may reveal mediastinal or subcutaneous air dissection. However, studies have shown that radiographic findings may not be present immediately following perforation [29]. In patients with a high suspicion of esophageal perforation with a normal X-ray, further confirmatory testing with a gastrografin esophagram should be performed. If no site of perforation is identified but clinical suspicion remains high, a dilute barium esophagram or CT scan of the chest should be performed [30, 31]. The management of esophageal perforations depends on the clinical status of the patient and the size/involvement of the perforation. Most perforations can be medically managed with intravenous antibiotics, avoidance of oral intake, and parenteral nutrition [32]. Surgical intervention should be considered for patients who develop sepsis, those with pleural space involvement, and patients who do not improve with medical management [2]. Case reports and studies describing the use of endoscopically placed stents (Fig. 10.1) and clips to treat esophageal perforations have been published, and these techniques are coming into more widespread use [33–35].
Fig. 10.1
An esophageal perforation caused by balloon dilation of a stricture (a) is treated by placement of a fully covered self-expanding metal stent (b). The stent is anchored to the esophageal wall using an Over-The-Scope Clip
Prevention of Adverse Events: The Use of Routine Testing Prior to Endoscopy
Routine pre-endoscopy laboratory testing is the practice of ordering a set panel of tests on every patient undergoing an endoscopy regardless of the patient’s history, physical, and/or preexisting medical conditions. Routine laboratory testing is often used in an attempt to prevent adverse events that may arise during endoscopy. The use of routine laboratory testing prior to endoscopy has not been validated as an effective way to prevent complications. Many studies have shown that routine laboratory testing rarely influences periprocedural management and that the cost of screening and the expense of following-up on these results outweigh their benefit [36, 37]. The American Society of Gastrointestinal Endoscopy (ASGE) recommends against routine pre-endoscopy testing in healthy patients. They recommend screening patients based on their history, physical, and preexisting medical conditions [37].
Complications of Therapeutic Upper Gastrointestinal Endoscopy
In addition to being an important diagnostic tool, the upper endoscopy procedure has become a widely adopted therapeutic modality for a wide range of GI conditions. As expected, more invasive therapeutic maneuvers during EGD also carry additional complications.
Complications of Dilation with Upper Endoscopy
The overall rate of reported adverse events associated with UGI dilation with endoscopy is between 0.1 and 0.4% [2, 4, 38]. The most common adverse events associated with UGI dilation are perforation, bleeding, aspiration, and bacteremia. The type and frequency of adverse events differ depending on the condition being treated. The most devastating adverse event associated with UGI dilation is esophageal perforation. Esophageal perforation in this setting is associated with a high mortality rate of 4–20% [25, 39].
Dilation of Esophageal Strictures
The most common adverse events associated with dilation of esophageal strictures are bleeding and perforation. The incidence of perforation with dilation of esophageal strictures is low; however, it is largely dependent on the technique used and the etiology of the stricture. The rate of perforation with the dilation of benign esophageal stricture has been reported between 0.1 and 0.3% [4, 40–43]. The rates of adverse events appear to be lower with wire-guided or pneumatic dilation as opposed to that performed with blind passage dilators [42]. Certain stricture types and etiologies that are associated with a higher risk of perforation include complex strictures (angular, tortuous, or long), strictures from caustic ingestion or eosinophilic esophagitis, and malignant or radiation-induced strictures [2].
Dilation for Achalasia
Pneumatic dilation of the lower esophageal sphincter is a frequently used treatment modality for achalasia. Endoscopic balloon dilation in this setting can only be called an aggressive maneuver, with balloons being inflated to 30–40 mm in diameter. The most severe complication of endoscopic balloon dilation for achalasia is perforation [44]. Perforation following pneumatic dilation for achalasia was reported to be 2% in a recent meta-analysis by Katzka et al. Most perforations examined in this study were managed medically with nasogastric decompression, intravenous antibiotics, and nothing by mouth. Only 1% of esophageal perforations required surgical intervention [45].
Dilation for Benign Gastric Outlet Obstruction (GOO)
The most common etiology of benign GOO is PUD. Conservative management with acid suppression, avoidance of NSAIDs, and H Pylori eradication when applicable is the first-line therapy. Endoscopic balloon dilation should be attempted only in patients who fail medical therapy. Perforation rates of pneumatic dilation for benign causes of GOO have been reported from 1.1 to 8.0% [46–49]. The risk of perforation is increased with active ulceration at the site of obstruction and dilation with balloons greater than 15 mm in diameter [47, 50].
Complications of Foreign Body Retrieval
Ingestion of foreign bodies is the second most common endoscopic emergency behind GI hemorrhage [51]. The type of foreign body ingested often varies based upon patient demographics. Fish and chicken bones, as well as impacted meat boluses, are typically seen in adults, while coins and toys are often seen in pediatric populations [52–54]. Intentional ingestion of various potentially obstructive foreign objects can also be seen in psychiatric patients and prisoners [55, 56]. Ingested foreign bodies can cause significant morbidity and mortality when they become impacted in the esophagus. Specific complications resulting from ingested foreign bodies include inflammation, mucosal laceration, perforation, hemorrhage, and even death [51, 57].
Though complications resulting from the endoscopic retrieval of foreign bodies are rare, they can be difficult to distinguish from the complications that result from ingestion of the foreign body itself. This is demonstrated by the fact that the most commonly reported complications of endoscopic retrieval of foreign bodies are superficial mucosal laceration (≤2%), GI hemorrhage (≤1%), and perforation (≤0.8%) [2]. Factors that have been shown to increase the risk of complications resulting from endoscopic intervention include presentation greater than 24 h after the onset of symptoms, sharp foreign objects, and the presence of multiple objects [51, 55, 58]. Sharp and irregular objects, in particular fish bones and chicken bones, significantly increase the risk of perforation [2, 54].
Special consideration should be given to include careful examination of the cervical esophagus as well as to technique when removing boluses of food or sharp foreign objects in order to minimize the risk of aspiration and perforation. During routine upper endoscopy, passing through the upper esophageal sphincter quickly and reaching the middle esophagus are a common practice. However, in cases when foreign body obstruction is suspected, extra care should be taken to prevent secondary injury to the esophagus, as foreign bodies located at the level of the cervical esophagus are often difficult to remove due to a limited working space for the endoscope [51, 56]. This is compounded by the fact that the majority of foreign bodies are found in the upper esophagus [51, 55, 56, 59, 60]. Currently, the ASGE guidelines regarding the management of food or meat boluses by piecemeal extraction recommends consideration of using an esophageal overtube and/or endotracheal intubation in order to minimize the risk of aspiration [2]. The ASGE has previously advocated against pushing the bolus into the stomach without first examining the esophagus distal to the obstruction by passing the endoscope around the bolus. However, 2 large published series using the push technique reported no perforations, and this technique may minimize the risk of aspiration [55, 56]. Tools such as an overtube or rubber hoods attached to the end of an endoscope are tools that can be used to help minimize the risk of perforation during removal of sharp or pointed objects. Additionally, removing foreign objects so that the sharp end is trailing also helps to minimize the risk of perforation [2, 52, 56].
After endoscopic retrieval of a foreign body, the mucosa should be carefully assessed for complications such as mucosal lacerations, bleeding, and especially perforation [2, 55]. In cases where retrieval was difficult, patients should be watched closely for signs and symptoms of perforation and should be considered for immediate radiographic contrast studies or chest radiograph to detect any evidence of mediastinal air [55, 61]. While the majority of mucosal injuries and bleeding can be managed either conservatively or with standard endoscopic hemostasis techniques, surgery may sometimes be indicated for more serious complications such as perforation [2, 61]. Perforation, when recognized immediately during endoscopy with no evidence of mediastinal contamination, can sometimes be treated with removable plastic or covered metal esophageal stents or with endoscopic clips [35, 61]. These strategies can be especially beneficial in patients who are not good surgical candidates or in patients with an underlying esophageal neoplasm [35, 61, 62]. Perforation, when foreign body induced, recognized later by signs and symptoms such as fever, tachycardia, chest or abdominal pain, and crepitus involving the soft tissue surrounding the neck can be treated with surgery if endoscopic options fail or are not felt to be appropriate [55, 61].
Complications of Percutaneous Endoscopic Gastrostomy (PEG) Placement
Percutaneous endoscopic gastrostomy (PEG) was first introduced in 1980 as a means to provide long-term enteral nutritional support to patients with a functional gastrointestinal system, without the need for surgical laparotomy [63]. Currently, PEG continues to be one of the most common endoscopic procedures worldwide [64]. It is generally considered to be a safe procedure, with an overall rate of adverse events reported to be between 4.9 and 10.3%, and a rate of serious adverse events reported in 1.5–9.4% of cases [2, 64]. Minor complications of PEG placement include tube occlusion, tube migration causing gastric outlet obstruction, granuloma formation, pneumoperitoneum, and peristomal leakage and/or pain. Major complications include aspiration pneumonia, bleeding, internal organ injury, gastric perforation, “buried bumper syndrome,” wound infection, necrotizing fasciitis, and tumor seeding of the stoma [2, 65, 66]. Death resulting from PEG placement is very rare, and according to one meta-analysis of 4194 PEG procedures, PEG-procedure-related mortality was reported to be only 0.53% [65, 66].
Several different strategies for PEG placement have been developed since it was first introduced, such as the “pull” technique, the “push” technique, and the introducer technique. The “pull” technique is currently reported to be the most common technique, though no significant differences in complication and efficacy rates has been reported between the different methods [67, 68]. Tumor seeding of the stoma is a rare complication seen in patients with head and neck or esophageal cancer, with only 22 cases reported [69]. Though hematogenous spread is possible, the exact mechanism of this phenomenon is not well established. It is generally believed that direct seeding of the gastrostomy site can occur when the PEG tube comes into contact with head and neck cancer during the “push” or “pull” technique [64]. For cases of PEG placement in patients with head and neck cancer, greater consideration for the introducer technique can be given as opposed to the more commonly used “push” or “pull” techniques [65].
Peristomal wound infection is the most common infectious complication following PEG placement. According to a meta-analysis of 10 randomized clinical trials, the pooled rate of peristomal wound infection was 26% [2, 64, 70]. Patients undergoing PEG placement are considered to be at higher risk of developing infections as the patient population often have significant underlying comorbidities, poor nutritional intake, and advanced age [71]. Minor infections can often be treated with topical antiseptics and local wound care, while more serious wound infections require systemic antibiotics [64]. A single dose of an intravenous cephalosporin- or penicillin-based antibiotic administered 30 min before the procedure is currently recommended as it has been shown in multiple randomized, controlled trials to be effective in reducing the occurrence of peristomal wound infections [2, 64, 70–73]. In areas where MRSA is endemic, pre-procedure screening of patients for colonization with MRSA, and subsequent decontamination if positive, has been shown to be beneficial in reducing the rate of MRSA wound infections following PEG placement [64, 71, 74, 75].
Necrotizing fasciitis is a rare but potentially life-threatening infectious complication of PEG placement [76, 77]. It is characterized by the necrosis of abdominal fascia due to a rapidly progressing infection along the fascial planes. Two main risk factors in the development of necrotizing fasciitis are traction and pressure on the PEG tube; therefore, maintaining a distance of 1–2 cm between the external bumper and the abdominal wall can help prevent this complication [64, 78]. Additional risk factors for development of necrosis after PEG placement include diabetes mellitus, atherosclerosis, alcoholism, malnutrition, immunosuppression, and older age [2, 64, 77, 79]. Development of necrotizing fasciitis is an emergency and requires immediate surgical consultation for consideration of wide surgical debridement, broad-spectrum empiric antibiotics, and intensive care support [64].
Buried bumper syndrome is another major complication that can occur as early as 3 weeks following PEG placement [80, 81]. It is characterized by ischemic necrosis of the gastric wall, believed to be caused by excessive traction on the internal bumper and migration of the tube out of the gastric lumen and toward the abdominal wall [2, 64]. Buried bumper syndrome is preventable by checking tube position regularly, leaving a small distance between the external bumper and the skin, and rotating the PEG tube 180–360° daily [64]. Treatment involves removal and replacement of the PEG tube, either endoscopically, surgically, or by external traction of the tube [64, 82, 83]. A case of “buried bumper syndrome” is demonstrated in Fig. 10.2.
Fig. 10.2
Patient with a “buried bumper” of his PEG tube. The old PEG tube is removed and replaced with a new PEG tube under wire-guided assistance
Other major complications of endoscopic PEG tube placement include aspiration pneumonia, bleeding, and injury to internal organs. Aspiration directly related to the procedure of PEG placement is reported to be between 0.3 and 1.0%, though it can be difficult to determine whether the aspiration event occurs during the procedure itself or later during feeding via the PEG tube [2, 65, 84, 85]. The risk factors for aspiration include supine position, advanced age, need for sedation, and neurologic impairment. Aspiration risk can be reduced by avoiding over sedation, minimizing air insufflation, and thoroughly aspirating gastric contents before PEG placement [86]. Bleeding can occur either from traumatic erosions of the esophageal or gastric mucosa, as well as from puncturing gastric or abdominal wall vessels, including the gastric artery and splenic or mesenteric veins [2, 64, 65]. Fortunately, acute hemorrhage is a rare complication and occurs in less than 1% of procedures [2, 64, 65]. Bleeding can usually be managed by applying direct pressure over the abdominal wound, though endoscopic or surgical exploration may be necessary in some cases [64]. Correcting coagulation disorders and stopping anticoagulants prior to PEG placement as much as possible are recommended to decrease the risk of significant bleeding [87].
Injuries to internal organs and gastric tears or lacerations are also rare complications of PEG placement, occurring in less than 0.5–1.8% of cases, though elderly patients may have a slightly increased risk of injury to bowels due to laxity of the colonic mesentery [65, 88]. Generally, injury to the colon or small bowel is more common than injury to the spleen or liver [64]. Injury to internal organs often warrants surgical intervention, though specific management has not been well studied [2, 64, 89]. Diagnosis of injury to internal organs can often be complicated by the fact that benign, transient (up to 72 h) pneumoperitoneum is reported to occur in 12–38% of patients undergoing uncomplicated PEG, therefore limiting the reliability of plain films in the diagnosis of suspected perforation of visceral organs [2, 64, 88, 90, 91]. In such cases, using a water-soluble oral contrast with computed tomography (CT) scan is a useful alternative in the diagnosis of possible defects in gastrointestinal integrity [64].
Development of gastrocolocutaneous fistulas may result if a loop of bowel is inadvertently perforated during PEG placement, or even over time via erosion into adjacent loops of. Asymptomatic or chronic gastrocolocutaneous fistulas can similarly be diagnosed using computed tomography (CT) with water-soluble contrast, and management of these cases includes simple removal of the tube [2, 65]. Surgery is required only in rare cases where a fistula persists after removal of the tube [65]. Using proper technique for PEG placement minimizes the risk of injury to internal organs and includes measures such as adequate gastric trans-illumination, finger indentation, and use of the “safe-tract” method during PEG placement [2, 92].
After successful PEG placement, inadvertent dislodgment of the PEG tube has been reported to occur in 1.6–4.4% of cases [65]. If dislodgement occurs before a mature tract is able to develop (usually 7–10 days), a free intra-abdominal perforation can result as the stomach separates from the anterior abdominal wall. If identified immediately, endoscopic placement of a new tube either via the same opening in the abdominal wall, or near the original site, is appropriate as pulling the stomach back against the anterior abdominal wall will seal the perforation [2, 65]. If tube dislodgement is identified late in a patient with an immature tract, management should include placement of a nasogastric Salem sump tube, broad-spectrum antibiotics, and new PEG placement within 7–10 days, as long as the patient does not show signs of peritoneal inflammation [65]. Patients who have a mature tract that experience tube dislodgement can have a new PEG tube placed safely through the same tract without the need for endoscopy [64].
Complications of Therapeutic Endoscopy
Complications of Polypectomy
While upper GI endoscopy is a routinely performed procedure with a relatively low risk of mortality and adverse events, therapeutic interventions increase the incidence of complications including bleeding, pain, dysphagia, and perforation [4, 5, 8–10, 93]. Snare polypectomy of gastric polyps is frequently performed in order to assess polyp histology for diagnosis [94]. Bleeding is the most common complication with an incidence of 6–7.2% [95–97]. In comparison with colonic polypectomy, gastric polypectomy demonstrates a much higher rate of bleeding (1% vs. 7%, respectively) [98]. Suggested risk factors for post-polypectomy bleeding include large size (greater than 8 mm) and sessile appearance [99]. Bleeding is often effectively controlled with injection of 5–15 mL of 1:10,000 diluted epinephrine followed by bipolar electro-cauterization or hemoclip application [99]. Closure of a gastric polyp defect after polypectomy of a gastric adenoma is demonstrated in Fig. 10.3.
Fig. 10.3
A large gastric adenoma was removed using hot snare polypectomy (a). The ulcerated defect and bleeding from an endoscopic mucosal resection site were closed using endoscopic clips (b, c)
Endoscopic Mucosal Resection (EMR)
EMR is increasingly being used to remove benign and early malignant lesions of the GI tract. Initially developed for the removal of sessile or flat neoplasms confined to the mucosa and submucosa, EMR may involve submucosal injection or ligation assistance to lift the lesion and aid in resection [100]. Bleeding represents the most common adverse event associated with EMR. A single-center study including 681 patients who underwent 2513 EMRs of the esophagus showed a rate of significant bleeding of 1.2%. The authors defined significant bleeding as a drop in hemoglobin greater than 2 mg/dl from baseline, bleeding requiring therapeutic intervention or blood transfusion, and/or bleeding at a later time requiring rehospitalization. In the 8 cases of post-EMR bleeding, seven were treated successfully with epinephrine injection, clips, and thermal coagulation. One patient required surgery for adequate hemostasis. Factors including patient age, length of Barrett’s esophagus, number of EMR performed, and use of anticoagulants were all analyzed, and none of which were found to correlate with post-EMR bleeding. Bleeding occurred at a mean time of 2.5 days following the EMR procedure [101].
Like esophageal EMR, EMR of gastric and duodenal lesions may also be complicated by intra-procedural bleeding with reported rates of 0–11.5% [102–104]. Bleeding following gastric tumor EMR occurred in approximately 5% of patients based on a retrospective study of 472 patients [105]. Bleeding can be effectively controlled with hemostatic clipping, even in cases of spurting blood vessel from the EMR site [102].
EMR of the esophagus may also be complicated by perforation with reported rates ranging from 0.5 to 5% [102, 106]. The rate of perforation appears to be correlated with physician experience. A multicenter randomized clinical trial comparing endoscopic resection to radiofrequency ablation for Barrett’s esophagus with high-grade dysplasia or early cancer demonstrated a perforation rate of 5% in the first 120 esophageal EMRs performed by 6 physicians who were provided with structured training [107]. A prospectively maintained database of patients with Barrett’s esophagus reported no EMR-related perforations in the study period possibly related to operator experience [101].
Perforations due to EMR of the stomach and duodenum appear to be uncommon. A systematic review demonstrated the risk of perforation after EMR to be 1% [108]. A prospectively maintained database of patients who underwent endoscopic resection of duodenal adenomas or laterally spreading tumors demonstrated a duodenal perforation rate after EMR to be 2% [109]. Perforation rates of the duodenum due to EMR have been shown to be related to physician experience. Perforations can be effectively closed using endoscopic clips in cases in which the perforation was recognized at the time of occurrence. In delayed perforations, the patients should be managed with surgical repair [109]. Closure of a gastric perforation using an Over-The-Scope Clip (OTSC) is shown in Fig. 10.4.
Fig. 10.4
Closure of a gastric perforation after endoscopic mucosal resection of a polyp with an Over-The-Scope Clip (OTSC; Ovesco Endoscopy GmbH, Tuebingen, Germany)
Complications of Endoscopic Submucosal Dissection (ESD)
ESD is a technique of endoscopic resection that allows for en bloc removal of lesions in the epithelium. Intra-procedural bleeding is common and may be treated with coagulation current via the ESD knife or with hemostatic forceps [114]. Occurring in 4.5–15.6% of cases, post-procedure bleeding is a known complication, which occurs more frequently with gastric resections as compared to esophageal resections [108, 115]. Risk factors for delayed bleeding include lesion size greater than 40 mm and resumption of antithrombotic therapy [116]. A meta-analysis of 6 studies demonstrates a reduced incidence of delayed bleeding after gastric ESD in patients treated with a proton pump inhibitor compared to those treated with an H2 receptor antagonist [117].
A meta-analysis of gastric ESD reports a perforation rate of 4.5%, while a review of a series of studies of esophageal ESD reports a pooled rate of perforation of 2.3% [108, 118, 119]. It should be noted that a meta-analysis comparing adverse event rates for ESD and EMR for superficial esophageal cancers demonstrated a significantly higher rate of perforation in the ESD group [120]. Treatment of perforations may be non-operative with the use of clip closure. A Japanese study of 10 years of ESD/EMR gastric perforations describes successful closure of perforations in 98% of cases [121].
Like EMR, ESD may be complicated by post-ESD strictures, which most often occur in the esophagus. Reported rates of stricture formation are 12–17% with greater circumference of resection and length of resection being known risk factors [122–125]. Reported treatments include serial dilation, intra-lesional steroid injection, topical steroid application, radial electroincision, and prophylactic placement of self-expandable metal stents [126–128].
Complications of Endoscopic Eradication Therapy (EET) and Coagulation With an increasing incidence of esophageal adenocarcinoma in the Western world, there has been heightened interest in EET in the treatment of Barrett’s esophagus [4]. Also, hemostasis during upper endoscopy is increasingly accomplished by both contact and non-contact thermal devices.
Complications of Argon Plasma Coagulation (APC)
APC is a non-contact thermocoagulation modality often used to eradicate mucosal lesions. Randomized trials with APC report bleeding rates as high as 4%, esophageal perforation rates as high as 2% and strictures in up to 6% of patients, all of which appears to be higher than other ablative modalities [129–132]. However, a Cochrane randomized control trial comparing APC and multipolar electrocoagulation for the treatment of Barrett’s esophagus demonstrated no serious adverse events [129]. More commonly reported are events of upper GI discomfort including pain, dysphagia, and nausea [129].
APC may also be used in cases of gastrointestinal bleeding and adenoma eradication. There have been case reports of pneumoperitoneum following APC [133]. Pneumoperitoneum following APC may not be a sign of perforation and may simply be due to argon gas passing through the GI tract wall into the abdomen. APC-induced ulcers may result in gastrointestinal bleeding [134]. A large series of 2193 sessions of APC in 1062 patients demonstrated a perforation rate of 0.2% [135].
Complications of Photodynamic Therapy (PDT)
PDT is another ablative modality that uses porfimer sodium with a photosensitizing agent. Similar to APC, the most commonly reported adverse event is chest discomfort and photosensitivity. A randomized, multicenter study conducted over 5 years reported resolution of all cases of photosensitivity [136]. Another study reported pleural effusions and fever in patients who underwent PDT [137]. Of the ablative modalities, PDT appears to have the highest incidence of post-procedure esophageal strictures with rates up to 35%. However, after being followed for 5 years post-treatment for any complications, none of the study patients reported any long-term adverse events [136].
Complications of Radiofrequency Ablation (RFA)
RFA is commonly used in the ablation of Barrett’s esophagus. RFA involves the delivery of a preset amount of radiofrequency energy via a balloon, resulting in circumferential superficial tissue destruction [138, 139]. RFA may be associated with chest discomfort in up to 2% of patients, which resolves within 1 week of the procedure [140]. Superficial lacerations from the procedure were reported in 6% of patients in a single trial [141]. Hemodynamically significant bleeding is relatively rare, occurring in less than 2% of procedures [140, 142]. A multicenter study of 429 patients who underwent RFA reported no serious adverse events. Strictures were seen in 1.8% of the participants with other trials reporting a stricture rate of 2–8% [140–142]. These strictures are easily treated using endoscopic balloon dilation (Fig. 10.5). Esophageal perforation successfully treated with an endoprosthesis has been reported [143].
Fig. 10.5
Patient underwent a radiofrequency ablation for Barrett’s esophagus (a). Patient developed a esophageal stricture (b) four weeks later. This was treated successfully with balloon dilation (c)
Complications of Cryotherapy
Cryotherapy in which liquid nitrogen is applied to an area of Barrett’s esophagus is a relatively novel idea. Initially reports demonstrate that the procedure is relatively well tolerated. In a study of 333 treatments performed on 98 patients, there were no serious adverse events. Two percentage of patients reported chest discomfort, which resolved after a brief treatment course with narcotics. Strictures were seen in 3% of patients, all of whom underwent successful esophageal dilation thereafter [144]. Perforation was reported in 1 patient who had Marfan’s syndrome [145].
Complications of Enteral Stenting of the Upper Gastrointestinal Tract
Esophageal, gastric, and duodenal stenting have become common procedures for the management of benign and malignant strictures [35, 146, 147]. Esophageal stents are also used to treat benign esophageal perforations and anastomotic leaks [148]. Features and design of these stents are variable with complication rates frequently correlating with design specifics of the stent [149]. Historically, the use of rigid esophageal stents carried a complication rate of 20% and a mortality rate of 9%, respectively. Complications with these older devices included bleeding, fistula formation, stent migration, food impaction, and tissue overgrowth [150]. These devices are now obsolete and are not commercially available.
With the advent of self-expandable metal stents (SEMS), complication rates have significantly dropped leading to the demise of rigid stents [151]. SEMS may be partially covered or fully covered. Partially covered stents are more frequently subject to tissue ingrowth and, to a lesser extent, overgrowth, while fully covered stents are more likely to migrate as they cannot embed in the mucosa [152–155]. Significant immediate adverse events following SEMS placement may occur in 2–12% of patients and include aspiration, respiratory compromise, stent mal-positioning, and perforation [156–159]. Most post-procedure adverse events are self-limiting and include chest pain and nausea [156, 160]. More significant adverse events include tumor overgrowth, stent migration, luminal perforation, and bleeding. Stents may migrate and tumor overgrowth may occur in up to 27% of patients [159, 161]. A Swedish study of 152 patients who underwent SEMS placement for esophageal strictures reported transient chest/pharyngeal discomfort in all patients, stent migration in 5%, perforation in 1%, and stent occlusion in 10% [160]. Stents placed across the esophagogastric junction may result in increased rates of gastroesophageal reflux [162]. Patients with stents that cross the esophagogastric junction are often placed on prophylactic acid suppression medication to good effect.
Stent placement is also a well-established palliative treatment modality for malignant gastric outlet obstruction [163]. Nonetheless, gastroduodenal stents are frequently placed in older patients who often have multiple comorbidities [164]. Severe early adverse events including bleeding and perforation are reported in 1–5% of patients [163, 165, 166]. A prospective study of 108 patients with malignant gastric outlet obstruction who underwent stent placement reported no procedure-related mortality. The most common adverse event was stent occlusion, which was reported in 14.8% of patients. It must be noted that the stent used in this study was uncovered. Other reported adverse events were GI bleeding (3.7%) and stent migration (1.9%) [164]. Precautions must be taken to avoid aspiration during placement of the stent as this represents a significant periprocedural complication [167].
Complications of Upper Endoscopy Performed for Evaluation of Upper Gastrointestinal Bleeding
Gastrointestinal bleeding (GIB) is one of the most common diagnoses leading to hospitalization in the USA. GIB is the principal diagnosis in up to 182 per 100,000 adults [168] and in the top 10 GI diagnoses within a hospitalization and causes of GI mortality [1]. Upper GIB has traditionally been described as bleeding within the gastrointestinal tract proximal to the ligament of Treitz. Upper endoscopy is frequently used to diagnose and potentially treat the source of UGIB. Based on source of bleeding and intervention performed, the complications and management can vary widely.
Complications of Endoscopic Variceal Hemostasis
Acutely bleeding esophageal varices (EV) or gastric varices (GV) are a common indication for performing an EGD and carry a relatively high mortality if not intervened on in a timely manner. Prior to performing an EGD in a patient who potentially has bleeding EV or GV, certain steps should be considered prior to starting the endoscopy in order to minimize complications.
Patients should be stabilized in an intensive care unit or other monitored setting with appropriate intravenous access in order to maximize hemodynamic stability. Goal hemoglobin concentration is typically 7–8 g/dL [169, 170]. Caution must be used to not over-resuscitate a cirrhotic patient with blood products or crystalloid solutions since this may lead to increased portal pressures resulting in increased risk of rebleeding and mortality [170].
Additional blood products should be available before, during, and after the endoscopy. Fresh-frozen plasma (FFP) and/or platelets can be administered in patients with significant coagulopathy. The data regarding the use of recombinant factor VIIa (rFVIIa) in cirrhotic patients with UGIB are somewhat controversial and need further elucidation before its use can be definitively recommended [170].
Intravenous or oral fluoroquinolone (norfloxacin 400 mg po bid or ciprofloxacin 500 mg bid) or intravenous ceftriaxone 1 g daily should be administered to the cirrhotic patient with or without ascites with suspected UGIB for seven days. This short-term antibiotic prophylaxis in cirrhotic patients with or without ascites reduces all-cause mortality, mortality secondary to infection, rebleeding events, hospitalization length, and over bacterial infection rate [169–171]. Cirrhotics that are hospitalized have been demonstrated to have a bacterial infection rate of 20%, and up to 50% can develop an infection while hospitalized with a GIB [170, 172]. Aside from bacterial peritonitis, these patients also are at more risk of respiratory infections, UTIs, and bacteremia [170]. The use of the EGD procedure without the proper antibiotic prophylaxis and/or treatment doses could potentially lead to the complication of increased risk of infection.
If an acute esophageal variceal bleed (EVB) is suspected, intravenous octreotide should be initiated with a bolus dose of 50 μg followed by an infusion of 50 μg/h. If the source of the UGIB is confirmed to be from bleeding EV, then octreotide should be continued for 3–5 days post-EGD [170].
The use of a high-dose IV infusion of proton pump inhibitor should be considered if other etiologies of UGIB outside of an acute variceal bleed are possible or until confirmation of EVB during EGD.
EGD should be performed within 12 h of admission [170]. In addition to the medical management mentioned above, endotracheal intubation (EIT) should be considered prior to performing an EGD. The 2014 ASGE guidelines for endoscopic management of variceal hemorrhage suggest that “intubation of patients before endoscopy to prevent aspiration during the procedures, especially in patients with encephalopathy” [169]. Similarly, AASLD guidelines state that “intubation may be required for airway protection prior to endoscopy” [170]. Despite these statements, the data are not entirely clear on benefits of prophylactic intubation prior to emergent endoscopy, in both variceal and non-variceal UGIB. Some clinical studies actually suggest an increased risk of aspiration pneumonia in patients prophylactically intubated prior to EGD [173].
Despite maximal medical and endoscopic management, variceal bleeding may not be controlled during initial or rebleeding episodes. Patients who survive an episode of acute variceal hemorrhage have a median rebleeding rate in untreated patients of approximately 60% with 1–2 years of initial bleed with a mortality of 33% [170]. The risk of rebleeding is multifactorial and may occur as a complication of EGD. Endoscopic and/or pharmacological treatments may not control variceal bleeding on initial or recurrent episodes in up to 10–20% of patients [170]. In these situations, TIPS should be considered as a salvage therapy, and surgical consultation may be warranted [170].
Complications of Endoscopic Variceal Sclerotherapy (EVS)
EVS is successful in controlling active EVB in more than 90% of patients [169]; however, its use has been primarily supplanted by esophageal variceal ligation (EVL) based on its adverse event profile.
The most common sclerosing agents used during an EGD include ethanolamine oleate, cyanoacrylate, polidocanol, absolute alcohol, sodium tetradecyl sulfate [2]. No agent has been shown to be more efficacious or safer [2]. Overall adverse event rate with EVS ranges from approximately 35 to 78% along with a mortality rate between 1 and 5% [174, 175]. Minor, temporary complications that may be encountered in the first 24–48 h post-EVS include: low-grade fevers, chest pain, and dysphagia [169, 176]. These transient symptoms typically do not require any treatment and are managed conservatively with symptomatic control.
Injecting the sclerosing agent can be technically challenging as the sclerosing agent needs to be injected directly into the varix, compared with banding where the band can be placed in the vicinity of the bleeding varix. Placement of the sclerosing agent into the surrounding tissue can lead to further complications and tissue damage.
Esophageal ulcerations are sometimes deemed a “complication” of EVS or EVL (Fig. 10.6); however, they are expected phenomena after successful endoscopic treatment of bleeding EV. EVS-associated esophageal ulcers are deeper and heal slower compared with those secondary to post-EVL [177]. The severity of these ulcers may also be worsened when EVS is repeated within 1 week of the initial session [178, 179]. These ulcers cannot be prevented with agents such as sucralfate, H2 receptor antagonists, although PPIs may promote ulceration healing [180]. Post-EVS ulcers can bleed in up to 20% of patients, during which traditional endoscopic treatments can be performed for hemostasis, such as clipping [176].
