Introduction

Esophagectomy is a procedure associated with significant morbidity. Early detection of malignant lesions and improvement of perioperative management, as well as surgical techniques, have resulted in significantly improved survival of patients after esophagectomy for esophageal cancer. Currently, postoperative hospital mortality in centers with adequate experience is well below 5%. Overall 5-year survival rates as high as 30–40% have been reported after resection with intent to cure [1]. This enhanced survival has focused attention on the quality of life of patients after esophagectomy.

Problems that affect the quality of life after esophagectomy include dysphagia, heartburn, regurgitation, early satiety, fatigue, and psychological concerns [2]. Among these, reflux and stenosis have serious implications. Reflux of duodenal and gastric contents is a significant problem after subtotal esophagectomy and reconstruction with a gastric conduit; it has been documented to be present in 60–80% of patients [3]. These patients remain at increased risk for prolonged esophageal exposure to gastric acid and duodenal secretions because normal antireflux mechanisms (the lower esophageal sphincter, angle of His, and phrenoesophageal ligament) have been resected or disrupted. In addition, gastric drainage procedures may promote duodenogastric reflux, leading in turn to bile reflux into the esophagus. Furthermore, negative intrathoracic pressure and positive intra-abdominal pressure act together to promote reflux across the anastomosis [4]. The devastating consequences may include aspiration pneumonia, bleeding, recurrent stricture, and columnar epithelial metaplasia (Barrett’s esophagus), which can give rise to adenocarcinoma [2].

The incidence of anastomotic stenosis remains relatively high, between 10 and 56% [5]. The presence of ulceration involving more than 50% of the anastomotic circumference has been found to be the most important factor predicting an increased risk for the development of an anastomotic stricture [6]. The patients with esophagogastric stenosis suffer from dysphagia, aspiration, and weight loss; these patients may require repeated endoscopic dilatation.

The postesophagectomy quality of life of patients may be very much influenced by the quality of their esophageal anastomosis [1], which can be evaluated by the presence of reflux and stenosis. In the present study the quality of the esophagogastric anastomosis was evaluated by determination of the frequency of postoperative reflux and stenosis.

Materials and methods

Seventy-four patients underwent esophagectomy and gastric tube reconstruction for esophageal cancer between January 1995 and December 2004. Among this group, 53 patients (45 men and 8 women) underwent endoscopic procedures. The mean age at the time of esophagectomy was 60.3 ± 8.87 (range = 39–81) years. Twenty-one patients who did not undergo endoscopy because of early death (within 1 month, 3 patients) and early loss to follow-up (within 1 year, 18 patients) were excluded from this study. Upper gastrointestinal endoscopic examinations were immediately performed in patients with reflux symptoms and routinely performed annually at follow-up, with or without symptoms. We reviewed retrospectively the clinical records of these patients and recorded the postoperative medication, anastomosis site, symptoms of reflux esophagitis, presence of anastomotic stenosis, and endoscopic interventions. The Ethics Committee of the Guro Hospital, Korea University Medical Center approved this retrospective study and waived the need for patient consent.

After an upper midline incision, Kocher’s mobilization of the duodenum was performed; the gastric tube was trimmed along the greater curvature of the stomach. Digital dilatation of the pyloric ring was performed. For patients with middle and lower esophageal cancer (27 patients), intrathoracic esophagogastrostomy was performed; the procedure was carried out 7 cm or more from the proximal margin of the esophageal cancer with a right posterolateral thoracotomy. In the intrathoracic anastomosis group, a stapler anastomosis was performed in 16 patients and a manual anastomosis was performed in 11 patients. In patients with upper esophageal cancer (26 patients), esophagogastrostomy was performed through a left cervical incision. In the cervical anastomosis group, a stapler anastomosis was performed in 13 patients and a manual anastomosis was performed in 13 patients. Twenty-three patients had hand-sewn anastomoses and 30 patients had circular stapler anastomoses (EEA, Autosuture, U.S. Surgical Corp., Norwalk, CT). Hand-sewn anastomoses were end-to-side and constructed with a single-layer interrupted 4–0 suture (polyglactin 910). A 28-mm EEA surgical stapler was used in 27 patients and a 25-mm stapler in 3 patients. Beginning January 2000, the circular stapler anastomosis procedure was the only method used in all patients.

We assessed reflux symptoms such as pharyngeal regurgitation, cervical heartburn, pain, throat disturbance, or nocturnal cough and the symptoms of stenosis such as dysphagia. All information was obtained from the patients’ medical records. Data on medication with antacid agents such as proton pump inhibitors (PPIs) and histamine receptor-2 blockers (H2-blockers) were obtained from the patients’ clinical records.

Upper gastrointestinal endoscopy was performed after the esophagectomy (the mean interval between esophagectomy and endoscopic examination at the last follow-up was 29.2 ± 23.67 months, range = 5–111). The severity of reflux esophagitis was graded using the Los Angeles Classification System [7], and anastomotic stenosis was diagnosed in patients who required anastomotic dilatation. The frequency of reflux and stenosis was evaluated in association with the anastomotic site and the surgical technique used.

The data were expressed as the mean ± standard deviation. A binary logistic regression was used to perform univariate and multivariate risk factor analyses for esophagitis and anastomotic stenosis. The Spearman rank correlation coefficient test was used to evaluate the correlation between reflux symptoms and esophagitis. A p value less than 0.05 was considered to indicate statistical significance. The statistical software package SPSS 12.0 (SPSS Inc., Chicago, IL) was used for all calculations using a standard personal computer.

Results

A cervical anastomotic leak developed in three hand-sewn patients and in one of the patients who had a stapler procedure. The leak closed spontaneously in all four patients. Fifteen of 53 patients (28.3%) had symptoms of reflux; endoscopy revealed that 32.1% of these patients (17/53) had reflux esophagitis in the remaining esophagus. In 3 of 17 patients (17.6%), severe esophagitis (i.e., grade C or D) was found (Table 1). There was a significant correlation between the reflux symptoms and the endoscopic findings of reflux esophagitis (p = 0.000) (Table 2). Antacid agents such as PPIs and H2 blockers were administered to all patients.

Table 1 Incidence and severity of reflux esophagitis and stenosis
Full size table
Table 2 Relationship between reflux symptoms and endoscopic findings
Full size table

Univariate and multivariate analyses of risk factors (gender, age, PPI use, BMI, anastomotic location, and anastomotic technique) that predisposed to esophagitis and stenosis were performed. Analysis of reflux esophagitis by the anastomotic site showed that the incidence of reflux esophagitis was 12% (3/26) in the cervical anastomosis group and 52% (14/27) in the intrathoracic anastomosis group; these differences were significant (p = 0.041). Assessment of reflux esophagitis by the anastomotic technique showed that the incidence of reflux esophagitis was 26.1% (6/23) in the manual anastomosis group and 36.7% (11/30) in the stapler anastomosis group; the difference between the two was not statistically significant (p = 0.879) (Table 3).

Table 3 Risk factors for predisposition to reflux esophagitis
Full size table

Follow-up endoscopy in ten patients (18.5%) showed anastomotic stenosis; two of these patients had local disease recurrence at the anastomotic site. In seven of the ten patients, anastomotic dilatation was required. The two patients with local disease recurrence required insertion of an esophageal stent under endoscopic guidance. Three of the ten patients refused further procedures.

Analysis of esophageal stenosis by the anastomotic site showed that the incidence of anastomotic stenosis was 23% (6/26) in the cervical anastomosis group and 14% (4/27) in the intrathoracic anastomosis group; these differences were not significant (p = 0.829). Evaluation of the association of esophageal stenosis to the anastomotic technique showed that the incidence of anastomotic stenosis in the manual anastomosis group was 21.7% (5/23) and 16.7% (5/30) in the stapler anastomosis group; the difference was not statistically significant (p = 0.689) (Table 4).

Table 4 Risk factors for predisposition to anastomotic stenosis
Full size table

Discussion

After esophagogastrostomy procedures, early postoperative complications such as anastomotic leaks are serious risks that increase morbidity and mortality. Late complications such as anastomotic reflux or stenosis are factors that decrease the effectiveness of esophagogastrostomy for palliation in cancer patients. Prevention of reflux or stenosis may improve the postoperative quality of life.

Between 60 and 80% of patients have symptoms of reflux esophagitis after esophagectomy [3]; esophagitis occurs in 27–35% of patients, and columnar metaplasia, including intestinal (Barrett’s) metaplasia, may be found in the remnant esophagus after esophagectomy [4]. It has been reported that there is no correlation between reflux symptoms and reflux esophagitis [2]. Shibuya et al. [8] reported that the occurrence of reflux esophagitis in their patients was as high as 71.6%, higher than that reported in previous studies. They explained their higher incidence of reflux by the increased number of long-term survivors after esophagectomy as well as careful endoscopic gastrointestinal examination. In addition, they reported that the reflux esophagitis in the intrathoracic anastomosis group was significantly higher than in the cervical group.

In the present study there was a significant correlation between reflux symptoms and reflux eosphagitis; the incidence of reflux symptoms and reflux esophagitis was lower in our study compared that in prior reports. This may have been due to performing the anastomosis of the conduit at the highest intrathoracic portion, within the limits of possibility, in the intrathoracic group. It may also have been due to all patients being treated with antacid agents throughout the day. An anastomosis below the level of the aortic arch was thought to be “reflexogenic,” while one at the supra-aortic level was less so [9]. The physiologic explanation for this is that with a lower anastomosis, more of the stomach is subject to positive intra-abdominal pressure, which promotes greater reflux [10]. In addition, the results of this study showed that the incidence of reflux esophagitis was significantly lower in the cervical anastomosis group, consistent with previous reports [8]. Cervical anastomoses have shorter proximal esophageal segments compared to the longer intrathoracic segments, which are subject to weakened propulsion. Pulling the stomach all the way up to the neck abolishes the effect of positive intra-abdominal pressure squeezing the fluids upward. The presence of the anastomotic line in the neck does not expose it to the negative pressure that enhances the development of reflux. The mere presence of refluxed gastroduodenal fluid in the proximal esophagus is not pathognomonic. The duration of contact of the acidic fluid with the esophageal mucosa is important for the development of pathologic reflux esophagitis, with its characteristic endoscopic features and histologic changes.

There are two types of anastomotic stricture. First, benign strictures due to wound retraction are readily treated by dilation. Second, strictures secondary to severe anastomotic failure, usually related to problems with vascularization and ischemia, are more likely to be resistant to serial dilations [6]. However, there is no generally accepted definition to classify anastomotic strictures [11]. Dysphagia is the most often-cited symptom associated with an anastomotic stricture. Therefore, we diagnosed patients with an anastomotic stenosis when they required anastomotic dilatation because of dysphagia; this was confirmed endoscopically. However, some patients may have problems with swallowing due to weak propulsive waves of the remaining part of the esophagus; this cannot be detected by endoscopy, in contrast to a mechanical stricture of the anastomosis. This may be caused by denervation of the proximal esophagus during manipulation prior to resection and/or its malalignment or a sagging gastric conduit. Weak propulsion leads to improper esophageal body clearance of refluxed gastroduodenal fluid, enhanced by sluggish gastric tube contractions.

When an anastomotic stricture develops, the recurrence of dysphagia defeats one of the main aims of surgery, i.e., to restore normal swallowing function to improve the patient’s quality of life [12]. Anastomotic stricture formation was correlated with a preceding anastomotic leak, intraoperative blood loss, and poor vascularization of the gastric tube noted at surgery [13]. Honkoop et al. [14] reported that cardiac disease, postoperative leakage at the anastomosis site, and a stapled rather than a hand-sewn anastomosis were found to be independent risk factors for the development of a stricture. From the available data it appears that anastomotic stenosis occurs more frequently after a circular stapler anastomosis than after a hand-sewn anastomosis [1]. In the present study the incidence of anastomotic stenosis was not significantly different according to the anastomotic site or the anastomosis techniques used. These findings might be due to the small sample size. Orringer et al. [5] reported that use of the semimechanical anastomosis techniques may be a promising approach to reduce the incidence of anastomotic stricture formation. In addition, Raz et al. [15] demonstrated a significant reduction in the incidence of leaks and stenosis in stapled side-to-side intrathoracic anastomoses. Chronic PPI therapy combined with pneumatic dilatation are likely key factors for successful treatment of anastomotic strictures. Recently, a number of surgical maneuvers have been described for the prevention of reflux after esophagectomy and intrathoracic anastamoses. Aly et al. [4] reported that a modified fundoplication at and above the anastomosis site was a simple technique; preliminary studies have shown that it appears to offer reasonable control of reflux.

Conclusions

Cervical esophagogastrostomy was found to have a lower incidence of reflux esophagitis and a similar incidence of anastomotic stenosis compared to the intrathoracic approach. In addition, the cervical anastomosis provides a more generous resection margin, which is important in esophageal cancer surgery. Therefore, cervical anastomosis is the preferred approach for esophagectomy with esophagogastrostomy. Although this study did not reveal any differences in reflux esophagitis and stenosis according to the anastomotic technique used, the stapler anastomosis is preferred because of its technical simplicity [16], which is associated with reduced operating time and decreased anastomotic leakage [17]. Continued improvement in surgical techniques along with perioperative management makes the quality of the anastomosis more important for the patient’s quality of life. Approaches that favor no development of reflux should continue to be actively considered.

There were some limitations to this study. Eighteen patients were lost to follow-up during the first year. They accounted for 25% of the patient population. This reduced the sample size of this study. Further study with a larger sample size is needed to confirm our findings.