INTRODUCTION — Operations that resect the esophagus and restore gastrointestinal continuity are technically challenging procedures. The rates for morbidity and mortality depend on many factors (eg, patient comorbidities, operative approach, hospital/surgeon volume) but vary widely between publications. A review of all publications between 2005 and 2009 found that no single complication was reported in all papers, and in-hospital mortality, the most common term for postoperative death, had six different definitions [1].
The systemic and procedure-specific complications of esophageal resection are reviewed here. Methods by which surgical resection is accomplished are reviewed separately. (See "Surgical management of resectable esophageal and esophagogastric junction cancers".)
INCIDENCE AND RISK FACTORS — The overall incidence of postoperative complications varies widely between 20 and 80 percent and includes systemic complications (eg, pneumonia, myocardial infarction) and complications specific to the surgical procedure (eg, anastomotic leaks, recurrent laryngeal nerve injury) [2-13]. Pulmonary complications are the most common postoperative complications, occurring in 16 to 67 percent of patients [7,14-16], but anastomotic leak is the most dreaded, occurring in 0 to 40 percent of patients [7,17,18]. (See 'Anastomotic leak' below and 'Pulmonary' below.)
A multivariate analysis identified several preoperative factors that increased the risk of complications following esophageal resection and reconstruction [3]. Some of these included increasing age, conditions associated with compromised pulmonary function (eg, chronic obstructive pulmonary disease), malnutrition, renal or hepatic dysfunction, and emergency surgery. The preoperative indication for surgery (ie, malignant or benign disease) was not associated with increased morbidity. Patients with malignant disease had similar 30 day morbidity rates compared with patients with benign disease (49.0 versus 51.1 percent).
Comorbid illnesses increase the risk of postoperative complications (eg, cardiorespiratory complications, anastomotic leakage, reoperation rates, wound infection) and death following esophagectomy [8,19,20]. As an example, in a prospective study of 615 patients, those with comorbid illness had an increased overall rate of postoperative complications or major anastomotic leaks compared with patients without comorbidities (28 versus 18 percent and 11 versus <1 percent, respectively) [8]. In addition, a metaregression from a separate study showed the risk of anastomotic leakage or atrial fibrillation in patients with obesity and diabetes was significantly higher compared with patients with obesity but not diabetes [19].
Prospective and retrospective studies using multivariable logistic regression analysis reveal that increasing age is independently associated with an higher risk of morbidity and mortality [3,8,20-22]. Appropriately selected older adult patients, however, can undergo esophagectomy with comparable outcomes to younger patients. In a prospective analysis of 615 patients undergoing an esophagectomy, patients older than 75 years of age had a similar rate of complications compared with those 75 years and younger (29 of 110 patients [26 percent] versus 125 of 505 patients [25 percent]) [8]. The rate of a major anastomotic complication was also similar (12 patients [11 percent] versus 44 patients [9 percent]).
MORTALITY — The overall in-hospital mortality rates range from 0 to 22 percent [2,3,7,22-27]. The overall 30 day mortality rates (excluding in-hospital deaths) are <1 to 6 percent [23].
Postoperative pulmonary complications are the most significant factor contributing to death following esophageal resection and reconstruction. In a retrospective review of 379 patients, the incidence of death for patients who developed pneumonia was significantly higher compared with those who did not (20 versus 3 percent) [21]. Increasing age was the only other variable independently associated with postoperative mortality. Anastomotic complications were once associated with a 50 percent mortality rate, but with aggressive management, including immediate reoperation for uncontained leaks, or in some cases stent placement, mortality resulting from leak has declined [28]. Notably, a Society of Thoracic Surgery database investigation revealed that anastomotic leak requiring reoperation was associated with an increased mortality (odds ratio [OR] 1.48, 95% CI 1.03-2.14) but medically managed leaks were not [29]. (See 'Pulmonary' below and 'Anastomotic leak' below.)
Other clinical factors that increase the risk of perioperative (30 day) death following esophageal resection and reconstruction included ascites, increasing age, insulin-dependent diabetes, decreased functional status, neoadjuvant therapy, renal dysfunction, excess alcohol use, and hepatic dysfunction [3]. Mortality rates were similar for patients who underwent resection for cancer versus benign disease (9.7 versus 10.1 percent). In a separate retrospective review of 23 studies that included 13 randomized trials, neoadjuvant chemotherapy or neoadjuvant chemoradiation therapy did not increase the risk of postoperative complications or 30 day mortality [30].
The development of postoperative surgical complications significantly affects longer-term mortality rates. In a study of more than 10,000 esophagectomies from the Society of Thoracic Surgery database, the risk for mortality was increased with each individual postoperative event: acute respiratory distress syndrome (OR 7.48, 95% CI 5.23-10.7), reintubation (OR 6.55, 95% CI 4.61-9.30), new renal failure (OR 5.97, 95% CI 4.08-8.75), central neurologic event (OR 5.66, 95% CI 2.43-12.6), reoperation for bleeding (OR 5.12, 95% CI 2.48-10.6), myocardial infarction (OR 4.93, 95% CI 2.34-10.4), ventricular arrhythmia (OR 4.04, 95% CI 2.26-7.23), reoperation for chylothorax (OR 2.53, 95% CI 1.54-4.17), sepsis (OR 2.41, 95% CI 1.69-3.46), atrial arrhythmia (OR 1.52, 95% CI 1.17-1.98), reoperation for other cause (OR 1.5, 95% CI 1.04-2.17), and reoperation for anastomotic leak (OR 1.48, 95% CI 1.03-2.14) [29].
In a prospective study of 615 patients who underwent an esophagectomy, the median survival was significantly shorter in patients who sustained surgical complications compared with those without complications (15 versus 22 months) [8]. Including only those who survived 90 days or longer postoperatively (n = 567), the risk of mortality was significantly increased in patients who sustained postoperative complications (hazard ratio [HR] 1.29, 95% CI 1.02-1.63). Most patients who die following an esophagectomy have experienced multiple serious complications (median 2.5, range 1 to 8) rather than a single event [15]. (See 'Procedure-specific complications' below.)
Patients who undergo esophagectomy at hospitals where a larger number of esophagectomies are performed have lower perioperative mortality rates and better early clinical outcomes compared with lower-volume institutions [9,23,24,31-38]. The definition of low versus high volume is variable, with most studies defining "low volume" as <4 to <10 procedures and "high volume" as >9 to >40 procedures [23]. Surgeon experience, dedicated intensive care teams, nursing skill, respiratory therapy, resource allocation, multidisciplinary team management, and availability of advanced diagnostic and therapeutic equipment are examples of variables that are likely enhanced at high-volume centers [39-43]. As an example, in one report that used Medicare claims data, the mortality following esophagectomy at the highest-volume hospitals (>19 procedures annually) was significantly lower compared with the lowest-volume hospitals (<2 procedures annually, adjusted OR 0.36, 95% CI 0.26-0.50) [31]. In a later review of nine trials, in-hospital and 30 day mortality rates were significantly lower at high-volume hospitals (2.8 versus 8.5 percent and 0.7 versus 2.0 percent, respectively) [23].
SYSTEMIC COMPLICATIONS
Pulmonary — Pulmonary complications are the most frequent complications following esophageal resection and reconstruction, occurring in approximately 16 to 67 percent of all patients, and account for approximately two-thirds of mortality related to esophageal resection [3,7,14,15,21,44]. Pneumonia is an independent risk factor for postoperative mortality. (See 'Mortality' above.)
Postoperative pulmonary complications include pneumonia, bronchospasm, acute respiratory distress syndrome, acute exacerbation of chronic obstructive pulmonary disease, and pulmonary embolism. The risk of venous thromboembolism is approximately 1.5 to 2.4 percent following an esophagectomy [3,21]. The management of postoperative pulmonary complications is reviewed elsewhere. (See "Overview of the management of postoperative pulmonary complications".)
Preoperative, intraoperative, and postoperative measures that can reduce postoperative pulmonary complications include preoperative respiratory rehabilitation (eg, smoking cessation, inspiratory muscle training), postoperative lung expansion maneuvers, proper perioperative oral hygiene, postoperative pulmonary toilet, aspiration precautions, and adequate pain management. (See "Strategies to reduce postoperative pulmonary complications in adults".)
Initial attempts at oral intake following esophagectomy should be supervised, and liberal use of speech therapy services is encouraged in order to avoid aspiration events, especially in patients with a neck incision where the risk of recurrent laryngeal nerve injury is elevated. (See 'Recurrent laryngeal nerve injury' below.)
The operative approach itself may influence the risk of respiratory complications [45-47]. As an example, in a trial that randomly assigned patients with adenocarcinoma involving the mid-to-distal esophagus to a transthoracic or transhiatal approach (cervical phase identical), those undergoing a transthoracic esophagectomy were significantly more likely to experience pulmonary complications (57 versus 27 percent) [45]. In another trial that compared open versus minimally invasive esophagectomy following adjuvant chemoradiotherapy, patients undergoing open transthoracic esophagectomy had a significantly higher rate of pulmonary complications (29 versus 9 percent) [46]. There were no significant differences in the numbers of patients in each group with respect to stage of disease. In a review of 1568 patients undergoing elective esophagectomy from the National Surgery Quality Improvement (NSQIP) database, thoracic approaches to esophagectomy increased the risk of pneumonia, ventilator dependence, and septic shock but not mortality [48]. Performance of a McKeown esophagectomy is an independent predictor of combined morbidity and mortality in the risk model created by the Society of Thoracic Surgeons [22]. A description of the esophagectomy procedures is provided in a separate topic. (See "Surgical management of resectable esophageal and esophagogastric junction cancers", section on 'Operative procedures'.)
Cardiac — Cardiac complications include a spectrum of events, such as atrial fibrillation and myocardial infarction, and are variably defined and reported [1].
Atrial fibrillation — Atrial fibrillation is reported to occur in up to 20 percent of patients undergoing an esophagectomy [49,50]. In a retrospective review of 324 esophagectomy patients, the incidences of atrial fibrillation for patients undergoing a transthoracic esophagectomy, transhiatal esophagectomy, and three-field esophagectomy were 16.2, 18.6, and 16.7 percent, respectively [50]. (See "Surgical management of resectable esophageal and esophagogastric junction cancers", section on 'Thoracic cancer resection'.)
Atrial fibrillation is associated with increased overall surgical morbidity and mortality following major thoracic operations [29,49,50]. A retrospective review of 921 patients undergoing esophagectomy found that patients with postoperative atrial fibrillation had significantly higher rates of pulmonary complications, anastomotic leaks, and mortality rates compared with patients without atrial fibrillation (42 versus 17 percent, 6.9 versus 1.4 percent, and 23 versus 6 percent, respectively) [49]. Patients undergoing major thoracic operations (including esophagectomy) with postoperative atrial fibrillation also had a significantly higher mortality rate and longer hospital mean length of stay compared with those patients without atrial fibrillation (7.5 versus 2.0 percent, and [mean ± standard deviation] 8.2±11.0 versus 16.6±18.6 days, respectively) [50].
Management of atrial fibrillation is discussed in separate topics. (See "New onset atrial fibrillation" and "Antiarrhythmic drugs to maintain sinus rhythm in patients with atrial fibrillation: Recommendations" and "Overview of the acute management of tachyarrhythmias".)
Myocardial infarction — Based upon a large prospective cohort study and retrospective reviews, the incidence of a postoperative myocardial infarction ranges from 1.1 to 3.8 percent [3,4,21]. The reported rates of myocardial infarction are similar to those of other thoracic surgical operations, and there are no procedure-specific contributors to this incidence. However, the occurrence of myocardial infarction has significant implications for perfusion of the conduit, and options for management of perioperative myocardial infarction following esophagectomy are more limited (limited anticoagulation, no thrombolysis). Management of a postoperative myocardial infarction is typically done in consultation with a cardiologist and is discussed separately. (See "Perioperative myocardial infarction or injury after noncardiac surgery".)
PROCEDURE-SPECIFIC COMPLICATIONS — The more common complications related to the esophagectomy procedure include conduit complications (eg, anastomotic leak, ischemia, stricture), nerve injury, lymphatic leak, functional disorders, and diaphragmatic hernia, which are discussed below. Other rarer complications include airway injury, tracheoesophageal injury, and splenic injury.
Conduit complications — Ischemia and denervation of the conduit are the inherent complications of creating a neoesophagus. In one study, scanning laser Doppler perfusion measurements were significantly decreased in each of 16 patients after stomach mobilization and gastric tube formation, compared with measurements taken beforehand. The decrease (by approximately 70 percent) was particularly prominent at the proximal end of the gastric conduit [51]. Reduced conduit perfusion can lead to anastomotic breakdown and leak and/or stricture formation, or to functional conduit disorders.
Anastomotic leak — The incidence of anastomotic leak ranges from 5 to 40 percent following esophageal resection and anastomosis, and the mortality associated with leak is between 2 and 12 percent [7,17,18,28,52-60]. Perianastomotic leaks may also occur due to gastric staple line failure or gastric tip necrosis, which can be differentiated with endoscopic and radiographic studies. Factors that influence the incidence of anastomotic leak include:
●Anastomotic technique (hand sewn versus stapled versus hybrid) [61]
●Location of the anastomosis (neck versus chest)
●Type of conduit (stomach versus colon versus small bowel)
●Location of the conduit (orthotopic versus heterotopic)
●Creation of the conduit (degree of devascularization and tubularization)
In a single-center retrospective review of 393 esophagectomy patients, risk factors for development of an anastomotic leak include conduit ischemia (odds ratio [OR] 5.5, 95% CI 2.5-12.2), neoadjuvant therapy (OR 2.2, 95% CI 1.1-4.5), and comorbid conditions (OR 2.1, 95% CI 1.1-3.9) [57]. A review of the Society of Thoracic Surgery (STS) database that included 7595 esophagectomies identified heart failure, hypertension, renal insufficiency, and type of procedure as significant factors associated with anastomotic leak [60].
A thoracic anastomosis, due to a shorter conduit length and less tension with resultant improved proximal conduit perfusion, is less prone to leak [7,60]. In the STS study, the leak rate was 9.3 percent in patients with intrathoracic anastomosis and 12.3 percent for those who had a cervical anastomosis. In a prior systematic review, the odds ratio of developing an anastomotic leak for those undergoing a cervical anastomosis relative to a thoracic anastomosis was 3.43 (95% CI 1.09-10.78) [7]. No difference in mortality was identified when comparing cervical with thoracic leaks. However, the morbidity of pleural and mediastinal soilage is theoretically higher than for cervical leaks, which are generally contained by the neck anatomy.
Management — In general, cervical anastomotic leaks can be managed with drainage of the neck wound with subsequent wet-to-dry dressing changes [16].
Thoracic anastomotic leaks are more likely to require re-exploration for appropriate control [62], although endoscopic stenting or transluminal vacuum therapy may provide acceptable outcomes in selected circumstances [63]. As an example, 49 patients were treated endoscopically for esophageal anastomotic leaks after cancer surgery [64]. Thirty-one patients had a covered stent placed across the leaks; three patients had the leaks closed with clips. After a median follow-up of 83 days, 88 percent of patients achieved healing of their leaks. Of the 23 patients who received more than one endoscopic intervention, 96 percent healed their leaks. However, stent migration remains a challenge, especially in the thoracic anastomotic position. Furthermore, there remains concern that the radial forces exerted by expandable stents could worsen regional ischemia and thus cause more significant tissue loss. Finally, stent erosion into surrounding structures such as the aorta and airway remains a risk with prolonged use of this technology [65]. Endoluminal vacuum therapy has been demonstrated as an alternative in difficult situations. The technique applies negative pressure to the wound site by means of a transnasal tube affixed to a vacuum sponge [66]. (See "Endoluminal vacuum therapy".)
Myocutaneous free flaps such as a radial forearm free flap are gaining popularity for repair of conduit defects in the neck or chest. This requires a multidisciplinary evaluation including a microvascular surgeon experienced in these procedures, which are more commonly done for pharyngeal reconstruction related to head and neck cancers [67].
Other basic principles of anastomotic leak management include [28]:
●Blood flow to the esophageal conduit is vulnerable to hypotension. Thus, all patients should have adequate hemodynamic monitoring with maintenance of euvolemia and avoidance of vasopressors if possible.
●All leaks and contaminated spaces must be adequately drained, whether by wound opening or percutaneous drainage. When an anastomotic leak is suspected, a computed tomography (CT) scan can locate extraluminal collections.
●A nasogastric tube may be used to prevent continued contamination of surrounding structures while maintaining the patient on "nothing by mouth" (NPO, nil per os).
●Endoscopy is useful to assess conduit viability in the clinical setting of progressive hypotension, oliguria, and acidosis.
●Systemic antibiotics are administered once there is a clinical suspicion of a leak. (See "Surgical management of esophageal perforation", section on 'Initial management'.)
•Broad-spectrum intravenous antibiotics that provide coverage for aerobes and anaerobes are administered intravenously, such as ampicillin/sulbactam (3 grams every six hours), piperacillin/tazobactam (3.375 grams every six hours), or a carbapenem. In the setting of beta lactam hypersensitivity, use of clindamycin (900 mg every eight hours) plus a fluoroquinolone, such as ciprofloxacin (400 mg every 12 hours), is acceptable.
•Most of these patients have chronic gastroesophageal reflux and are on acid reduction therapy. Thus, they should empirically receive antifungal therapy (eg, fluconazole 400 mg daily).
Conduit ischemia — Ischemia of the conduit (gastric and colonic) occurs in approximately 9 percent of patients undergoing an esophagectomy [57]. This can vary from minor anastomotic breakdown to, rarely, complete loss of the conduit. In a retrospective review of 393 consecutive esophagectomy patients, the presence of comorbid illness increased the risk of conduit ischemia twofold (OR 2.2, 95% CI 1.1-4.3), and patients with conduit ischemia had a higher 30 day mortality rate compared with those without ischemia (13.9 versus 3.9 percent) [57]. The rate of ischemia, as identified by endoscopy, was similar for gastric pull-up compared with a colonic interposition graft (10.4 versus 7.4 percent).
Total conduit ischemia can present as a rapidly deteriorating course with evidence of septic shock. Endoscopy can be a useful tool in such a rapidly deteriorating patient to quickly assess for total conduit ischemia [68], which mandates surgical removal and proximal esophageal diversion. For gastric conduit ischemia, a gastrostomy tube may be left in place if there is reasonable-appearing distal stomach. If not, total gastrectomy is warranted. Important additional components of this operation include adequate drainage of all infected spaces, aggressive resuscitation, broad-spectrum antibiotic coverage (similar to that for anastomotic leak), and closure of the hiatal defect to prevent herniation. (See "Sepsis syndromes in adults: Epidemiology, definitions, clinical presentation, diagnosis, and prognosis".)
Anastomotic stricture — Anastomotic stricture occurs in 9 to 40 percent of patients following esophageal resection and reconstruction [17,18,54,57,69-71]. Stricture can be due to conduit ischemia or, with later presentation, recurrent disease at the anastomosis. Anastomotic strictures can cause symptoms of dysphagia, odynophagia, and aspiration and can lead to inadequate dietary intake and malnutrition. Patients with symptoms suggestive of postoperative stricture should undergo endoscopy.
Anastomotic stricture is often closely linked to conduit malperfusion/ischemia or surgical technique. In a retrospective review of 393 consecutive esophagectomies, the incidence of stricture was higher in patients with an ischemic conduit versus those that were not ischemic (48 versus 5 percent). Patients who had an anastomotic leak were also significantly more likely to develop an anastomotic stricture compared with patients without a leak (47 versus 19 percent) [57]. However, of the 80 patients who developed a stricture, 65 percent had no evidence of an anastomotic leak or conduit ischemia. These patients were more likely to have undergone a gastric pull-up procedure rather than a colonic interposition graft (25 versus 7 percent) [57]. Comparison of anastomotic techniques would suggest that the modified Collard or hybrid staple technique has a lower incidence of anastomotic stricture compared with hand-sewn or circular stapler techniques [52-54,61,72].
Benign strictures can be effectively managed by endoscopic dilatation (ie, balloon or Savary dilators). (See "Endoscopic interventions for nonmalignant esophageal strictures in adults".)
In one study, the success rate for relieving symptoms was 93 percent [57,73]. Dilation is performed as soon as the patient presents with dysphagia and is repeated as often as necessary to control symptoms.
In a review of 27 patients with postoperative dysphagia following an esophagectomy and reconstruction using a gastric conduit and cervical anastomosis, most patients had mild-to-moderate stricture diameter (5 to 12 mm in 93 percent) [73]. All patients experienced symptomatic improvement with subsequent endoscopic dilation, and 77 percent had sustainable relief of dysphagia following a median of two dilation sessions.
Surgical revision of the anastomosis is rarely indicated and may require resection and diversion. In the clinical setting of locally recurrent esophageal disease, surgical resection should be considered in patients who are able to tolerate an additional surgical procedure. For a recurrent lesion confined to a small area, a local resection and reanastomosis can be performed; however, a completion esophagogastrectomy and colonic or jejunal interposition is often required [74]. Chemotherapy and radiation therapy are also used to manage local recurrence. (See "Management of locally advanced, unresectable and inoperable esophageal cancer".)
Recurrent laryngeal nerve injury — Recurrent laryngeal nerve (RLN) injury may present as hoarseness, dyspnea, and/or aspiration pneumonia. Once RLN injury is suspected, laryngoscopy and esophageal swallow evaluation should be promptly obtained. Based upon a systematic review of four trials including 267 patients, patients undergoing a cervical anastomosis are significantly more likely to have recurrent nerve injuries, which is not surprising (OR 7.14, 95% CI 1.75-29.14) [7]. Furthermore, three-field lymphadenectomy was associated with a higher risk of nerve injury [75].
Management of a laterally paralyzed cord requires vocal cord injection, usually performed by an otolaryngologist, for temporary vocal cord medialization. This will enhance the cough effort and therefore improve pulmonary hygiene as well as help prevent aspiration and improve voice quality [76]. If the traumatized nerve was not severed, the patient may recover function. A retrospective review of 51 patients with vocal cord paralysis following an esophagectomy found 41 percent recovered function within one year and four additional patients recovered within two years [77]. If vocal cord paralysis persists, permanent vocal cord medialization should be performed [76].
Chylothorax — The proximity of the thoracic duct to the esophagus translates into a relatively high rate of chyle leaks when compared with other thoracic operations. Rates of injury range from 0 to 8 percent and are associated with mortality rates as high as 18 percent and a 30 day major complication rate of 85 percent [78,79]. Many practitioners empirically ligate the thoracic duct to prevent this complication. However, a study raises concern about this practice [80]. This retrospective review of 1804 patients undergoing esophagectomy for treatment of esophageal cancer found that prophylactic ligation of the thoracic duct did not reduce the incidence of postoperative chyle leak but, on multivariate analysis, was associated with significantly worse long-term survival. Clearly, this single-center study requires further evaluation but raises concern regarding potential long-term detrimental effects of routine thoracic duct ligation.
The diagnosis of a chyle leak is based upon an increase in chest tube output with enteral alimentation and a change in nature of the output from serosanguinous to a milky appearance. Pleural fluid triglyceride level >110 mg/dL or presence of chylomicrons is generally diagnostic of a chyle leak [81].
Once diagnosis is confirmed, management includes elimination of enteral nutrition, parenteral nutrition support, close observation of chest tube output, octreotide, and fluid resuscitation. A short course of conservative management is warranted; however, if the chyle leak persists at a rate of >10 mL/kg for five days, patients most likely will fail conservative management [82]. (See "Management of chylothorax".)
Early surgical intervention (within 14 days from diagnosis) is favored by many surgeons because of the high mortality rate associated with the resultant immunologic and nutritional depletion in an already compromised patient population [83,84]. The standard operative approach is a right thoracotomy with gentle retraction of the conduit and ligation of the thoracic duct. When available, a lymphangiogram can be performed prior to the operation to further define thoracic duct anatomy [85].
An attempt to identify the source of the leak can be made by administering 25 cc/hour of cream or vegetable oil enterally via a feeding tube for six hours prior to the operative procedure. The high-fat diet enhances flow in the lymphatic system. If the site of the leak is identified, the duct should be ligated proximally and distally, using ties or clips. If the site of the leak is not identified, ligation of all tissue between the spine and the aorta is performed as caudal as possible in the right hemithorax [83]. The azygos vein can be ligated en bloc or preserved, based upon surgeon preference. Pleurodesis may also be entertained as a supplementary procedure to prevent effusion recurrence. (See "Management of chylothorax", section on 'Pleurodesis'.)
Functional disorders — Normal digestive function is reported in <20 percent of patients following esophagectomy [71,86], although the majority of patients report minor symptoms and quality-of-life measures similar to those of matched controls [71,86,87]. Functional disorders following esophageal resection and reconstruction include dysphagia, reflux, dumping syndrome, and delayed gastric emptying.
Dysphagia — Dysphagia occurs in up to 65 percent of patients following esophagectomy, with only 3 to 4 percent experiencing severe dysphagia [73,88]. Although the most common etiology of dysphagia following esophagectomy is anastomotic stricture (ischemic or local recurrence), dysphagia can also be functional. (See 'Anastomotic stricture' above.)
Delayed gastric emptying — Delayed gastric emptying occurs in up to 50 percent of patients following esophageal resection and reconstruction, due to truncal vagotomy performed during esophageal resection and anatomic compression of the stomach that is inherent to the orthotopic route of restoration of gastrointestinal continuity performed with a gastric pull-up procedure [88]. (See "Pathogenesis of delayed gastric emptying", section on 'Control of gut motor function' and "Management of superficial esophageal cancer", section on 'Esophagectomy' and "Barrett's esophagus: Surveillance and management", section on 'Esophagectomy'.)
The use of a gastric outlet procedure (eg, pyloromyotomy, pyloroplasty) during the esophagectomy is generally accepted, although it is debated as a means to enhance gastric emptying. Proponents of gastric outlet procedures emphasize the low morbidity of the procedure and potential benefits with regard to reducing early pulmonary complications and preventing early conduit dilation, which may compromise anastomotic healing. Opponents emphasize the long-term benefits of an intact pylorus, including diminished biliary reflux, diminished dumping, avoidance of complications at the site of the procedure, and the ability to address gastric outlet issues endoscopically. Alternatively, botulinum toxin (Botox) may be injected into the pylorus, thus achieving a temporary gastric outlet procedure that may achieve the early benefits of a gastric outlet procedure while avoiding later complications [89,90].
Pyloric stenosis, however, is a delayed complication of a gastric outlet procedure that can result in conduit dilation and poor emptying. This can be addressed endoscopically with balloon dilation, Botox, or peroral endoscopic myotomy (POEM) [91], depending on which prior pyloric procedures have been performed and surgeon experience. Conduit malposition, dilation, and/or tortuosity may require surgical revision.
Reflux — Reflux (from the duodenum or stomach) occurs in approximately 20 to 80 percent of patients following esophagectomy [70,71,92-94]. In a retrospective review of the literature, factors that contribute to reflux following esophageal reconstruction include [92]:
●Disruption and/or loss of the normal antireflux mechanisms when using a gastric conduit (ie, lower esophageal sphincter, angle of His, diaphragmatic sling)
●Direct anastomosis with no sphincter-like mechanism to prevent reflux
●Positive intra-abdominal pressure promotes reflux proximally through the anastomosis
●Negative intrathoracic pressure
●Impaired conduit motility
●Impaired esophageal remnant motility, possibly related to denervation
As with other causes of reflux, it is appropriate to treat symptomatic patients with reflux following esophagectomy with proton pump inhibitors and motility agents. Recalcitrant symptoms or those leading to respiratory complications, such as recurrent infection, should lead to endoscopic and functional evaluation to exclude obstruction [79]. (See "Medical management of gastroesophageal reflux disease in adults" and "Surgical management of gastroesophageal reflux in adults".)
Dumping syndrome — Dumping syndrome, secondary to vagotomy, occurs in up to 50 percent of patients following esophagectomy, with 1 to 5 percent having disabling symptoms [88].
The majority of patients (either type of reconstruction) have an early onset, within 10 to 30 minutes, most likely due to rapid transit of hyperosmolar gastric contents into the small bowel. Late-onset postprandial dumping syndrome (one to three hours) occurs in 25 percent of patients and is attributed to hypoglycemia secondary to a profound insulin response to a carbohydrate challenge [95].
Management generally begins with dietary alterations, including reduction of simple sugars, increased frequency and decreased size of meals, and fluid restriction around meals to slow gastric transit. Octreotide may be considered in patients who are refractory to dietary modification [96,97].
In appropriate clinical settings such as Barrett's syndrome and intramucosal carcinoma, vagal-sparing procedures reduce the rate of postoperative dumping due to maintenance of gastric innervation [98].
Hiatal hernia — Hiatal hernias are an uncommon but challenging problem following esophagectomy and reconstruction. It has been proposed that the risk of hiatal hernia is greater after minimally invasive esophagectomy (MIE) than open esophagectomy. Suturing the hiatus is more difficult with MIE, especially Ivor Lewis MIE; there are also fewer adhesions after MIE, which permits greater movement and potential herniation of abdominal contents.
In a retrospective review of 2182 esophagectomy patients, the risk of a hiatal hernia was <1 percent [99], with a higher rate after MIE compared with an open approach (2.8 versus 0.8 percent) [100]. In another European multicenter retrospective study of 6608 patients, hiatal hernia occurred in 0.73 percent after open esophagectomy but 1.4 percent after MIE [101]. About half of the patients presented with the hiatal hernia acutely. By contrast, a retrospective review at two institutions revealed an overall hiatal hernia rate, detected by computed tomography, of 6.8 percent, with no difference between MIE and open techniques (8 versus 5 percent) [102]. In a separate retrospective review of 440 esophagectomy patients, 15 percent of patients were radiographically diagnosed with hiatal hernia; however, only 2 percent required intervention due to symptoms or progression [103]. The median time to diagnosis was two years.
Symptomatic hiatal hernias are associated with nausea, vomiting, progressive chest pain, and unexplained weight loss. Surgical repair should be considered in all patients unless the hernia is small and asymptomatic or the patient is at prohibitive risk for surgical intervention [99]. In general, the repair is performed by laparotomy or thoracotomy with reduction of hernia contents (eg, colon, small bowel), primary repair of the hernia defect, and avoiding injury to the vasculature of the neoesophagus conduit (eg, the gastroepiploic artery). Laparoscopic repair is feasible in nonurgent situations when performed by experienced esophageal surgeons [101]. (See "Surgical management of paraesophageal hernia".)
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Gastrointestinal perforation" and "Society guideline links: Esophagectomy".)
SUMMARY AND RECOMMENDATIONS
●Operations that resect the esophagus and restore gastrointestinal continuity are technically challenging procedures. Morbidity and mortality rates vary widely and depend on many factors (eg, patient comorbidities, operative approach, hospital/surgeon volume). (See 'Incidence and risk factors' above.)
●Complications include systemic complications (eg, pneumonia, myocardial infarction) and complications specific to the surgical procedure (eg, anastomotic leaks, recurrent laryngeal nerve injury). Atrial fibrillation, reported to occur in approximately 20 percent of patients undergoing an esophagectomy, is a surrogate marker for surgical morbidity and mortality following major thoracic operations. (See 'Systemic complications' above and 'Procedure-specific complications' above.)
●Risk factors for postoperative complications include increasing age, conditions associated with compromised pulmonary function (eg, chronic obstructive pulmonary disease), malnutrition, renal or hepatic dysfunction, comorbid illness, emergency surgery, and treatment at a low-volume institution. The indication for esophageal resection (malignant versus benign disease) has not been associated with increased morbidity. A thoracic anastomosis, due to a shorter conduit length and less anastomotic tension, thus improving proximal conduit perfusion, is less prone to leak compared with a cervical anastomosis. (See 'Incidence and risk factors' above.)
●The overall 30 day mortality rate (excluding in-hospital deaths) is generally <6 percent. Mortality rates are similar for patients who undergo resection for cancer compared with benign disease. The development of postoperative surgical complications significantly affects longer-term mortality rates. Most patients who die following an esophagectomy have experienced multiple serious complications rather than a single event. (See 'Mortality' above.)
●The most common serious complications are postoperative pulmonary complications and anastomotic leak. Other procedure-specific complications include recurrent laryngeal nerve injury, chylothorax, diaphragmatic hernia, and functional gastrointestinal disorders (dysphagia, delayed gastric emptying, reflux, postgastrectomy dumping syndrome). (See 'Pulmonary' above and 'Anastomotic leak' above.)
•Pulmonary complications are the most frequent complications following esophageal resection and reconstruction, accounting for approximately two-thirds of deaths related to esophageal resection and reconstruction. Patients undergoing a transthoracic esophagectomy are more likely to experience pulmonary complications.
•Reduced conduit perfusion can lead to anastomotic breakdown and leak or loss of the conduit, or later stricture formation. Total conduit ischemia can present as a rapidly deteriorating course with evidence of septic shock. Endoscopy quickly assesses for total conduit ischemia in a rapidly deteriorating patient; treatment involves surgical removal and proximal esophageal diversion. The patient with an anastomotic leak is initially treated with broad-spectrum antibiotics and drainage and then observed for evidence of progressive infection or conduit ischemia. Patients with a thoracic anastomosis are likely to require re-exploration for appropriate control of the leak. Cervical anastomotic leaks can be generally managed with drainage of the neck wound without operative intervention.
