Search

The following on the treatments of adenocarcinomas in Barrett's esophagus contains commentaries on endo mucosal resection; choice between other ablative therapies; the remaining genetic abnormalities following stepwise endoscopic mucosal resection and possible recurrences; the Fotelo-Fotesi PDT; the CT TNM classification of early stages of Barrett's carcinoma; the indications of lymphadenectomy in intramucosal cancer; the differences in lymph node yield in transthoracic versus transhiatal dissection; video-assisted lymphadenectomy; and the importance of the length of proximal esophageal resectipon; and indications of sentinel node dissection.

L’œsophage court est une realite anatomique chez pres de 7 % des patients referes pour chirurgie antireflux. L’examen radiologique baryte comportant des cliches pris en position debout est le meilleur moyen d’apprecier l’irreductibilite de la jonction œso-gastrique sous le diaphragme. La fundoplicature intrathoracique selon Nissen realisee par thoracotomie gauche est la meilleure technique chirurgicale pour obtenir un controle permanent et durable du reflux du contenu gastrique dans la lumiere d’un œsophage court. Certains details d’ordre technique doivent etre imperativement respectes pour eviter des complications chirurgicales graves.

This study reviews current data regarding duodenogastric and gastroesophageal bile reflux-pathophysiology, clinical presentation, methods of diagnosis (namely, 24-hour intraluminal bile monitoring) and therapeutic management. Duodenogastric reflux (DGR) consists of retrograde passage of alkaline duodenal contents into the stomach; it may occur due to antroduodenal motility disorder (primary DGR) or may arise following surgical alteration of gastoduodenal anatomy or because of biliary pathology (secondary DGR). Pathologic DGR may generate symptoms of epigastric pain, nausea, and bilious vomiting. In patients with concomitant gastroesophageal reflux, the backwash of duodenal content into the lower esophagus can cause mixed (alkaline and acid) reflux esophagitis, and lead, in turn, to esophageal mucosal damage such as Barrett's metaplasia and adenocarcinoma. The treatment of DGR is difficult, non-specific, and relatively ineffective in controlling symptoms. Proton pump inhibitors decrease the upstream effects of DGR on the esophagus by decreasing the volume of secretions; promotility agents diminish gastric exposure to duodenal secretions by improving gastric emptying. In patients with severe reflux resistant to medical therapy, a duodenal diversion operation such as the duodenal switch procedure may be indicated.

Primary duodenogastric reflux is a rare disorder in adults which has not yet been documented in children. Six young patients, aged 4.5 to 16.5 years (median 13.5 years) presented with atypical reflux symptoms persisting from 1 to 84 months (median 8 months) and unresponsive to classical antacid therapy. In all six patients, 24 h gastric bilimetry showed excessive bile exposures for absorbances ranging from 0.25 to 0.60. The fraction of time (supine period) above the 0.25 absorbance threshold ranged from 30% to 75% while the 95th percentile value for healthy adults is 31%. In all patients tested, hepato-iminodiacetic acid scintigraphy revealed the occurrence of a massive duodenogastric reflux and four out of five patients had an alkaline shift (fraction of time pH >8 on 24 h lower oesophageal pH monitoring) ranging from 4.2% to 20% (control values 0.0% to 2.9%). Endoscopic findings included abundant bilious gastric leak (6/6) and chronic prepyloric Helicobacter pylori negative gastritis (2/6). Daily administration of cisapride, sucralfate with or without omeprazole resulted in an improvement of symptoms in five patients within 15 days. This treatment was ineffective in one patient who became symptom-free only after a surgical duodenal switch with fundoplication was performed. Conclusion: primary duodenogastric reflux is a rare foregut disorder of unknown origin occurring in late childhood. If suspected, 24 h intragastric bilimetry appears to be a useful investigation to confirm the diagnosis.

Incomplete symptomatic relief of pharyngoesophageal (Zenker's) diverticulum after endoscopic stapling or laser division has been reported by some authors. The clinical relevance of cricomyotomy, although supported by experimental data, remains controversial.Operative procedures consisted of transcervical resection (n = 34, group I), transcervical resection plus cricomyotomy (n = 12, group II), transcervical cricomyotomy (n = 8, group III), transcervical cricomyotomy plus diverticulopexy (n = 47, group IV), endoscopic stapling division (n = 31, group V), and endoscopic laser division (n = 55; group VI).The percentage of totally asymptomatic patients was significantly (p0.004) higher after open procedures (combined groups I to IV) than after endoscopic treatment (combined groups V and VI) regardless of the size of the pouch (3 cm, 85% versus 25%;or = 3 cm, 86% versus 50%). The percentage of patients with no or occasional (ie, fewer than twice a week) symptoms was significantly (p0.001) higher after open procedures (98%) than after endoscopic treatment (57%) for less than 3-cm diverticula whereas it was not higher (p = 0.409) for 3-cm or greater pouches (open, 97%; endoscopic, 88%). Furthermore, this percentage was similar (p0.286) after endoscopic stapling division and after endoscopic laser division (3 cm, 50% versus 58%;or = 3 cm, 96% versus 80%). It was also similar (p0.197) after resection alone (group I) and after open operations including myotomy (combined groups II to IV) (3 cm, 100% versus 98%;or = 3 cm, 92% versus 100%). Unlike endoscopic stapling and division, laser division was complicated by mediastinitis (2 patients), and 1 patient was referred because of cervical esophageal disruption during laser division. Five of six postoperative fistulas after resection occurred in patients who did not have myotomy, and 4 patients were referred 12 to 49 years after resection without myotomy for true recurrence of the pouch.Open techniques afford better symptomatic relief than endoscopic techniques, especially in patients with small diverticula. Endoscopic stapling and division is safer than laser division. Although very effective at midterm, resection without myotomy predisposes to the development of postoperative fistula and to recurrence of the pouch after many years.

Background and aims. The rising incidence of Barrett's carcinoma is a matter of major concern in Western societies. We realized a review of the literature to evaluate the impact of antireflux surgery on prevention of Barrett's carcinoma. Methods. We used MedLine- and PubMed-based review of the literature published since 1970 on surgical treatment of Barrett's esophagus. Results. There is no report in the literature that describes de novo development of Barrett's metaplasia after successful antireflux surgery. Compared with medical therapy, the risk for malignant degeneration of Barrett's metaplasia is reduced in surgical patients according to some studies. On the other hand, regression of Barrett's metaplasia after antireflux surgery is rare and Barrett's carcinoma after surgery has been observed repeatedly. The combination of antireflux surgery and ablation of metaplastic mucosa in order to obtain regression has led to encouraging preliminary results; however, experience is still limited and numerous studies currently are underway. Dysplastic Barrett's esophagus (BE) is a precancerosis and should not be treated as BE without dysplasia; strategies other than antireflux surgery need to be discussed. Conclusion. A prophylactic effect of early antireflux surgery upon de novo development of Barrett's metaplasia is probable. The impact of surgery on malignant degeneration of Barrett's epithelium remains uncertain. Data currently available show no clear benefit of antireflux surgery on cancerogenesis in patients with Barrett's metaplasia.

BACKGROUND: Because very poor survival rates were reported after exclusive nonradical surgery, the current opinion in the medical community is that very few esophageal adenocarcinoma patients can anticipate long-term survival after esophagectomy. In the current study the ability of exclusive radical surgery including very extended lymph node dissection to provide a substantial percentage of patients with long-term survival was examined. METHODS: Radical esophagectomy (including removal of the esophageal tube, excision of the potentially involved locoregional lymph nodes, and skeletization of the nonresectable vital organs in the mediastinum and upper abdomen) was attempted in 183 consecutive patients with either Barrett (n = 77) or non-Barrett (n = 106) adenocarcinoma of the esophagus or cardia. Esophagectomy was subtotal (neck anastomosis) or distal (chest anastomosis) in 103 patients and 80 patients, respectively. RESULTS: Radical esophagectomy (Ro resection) was feasible in 137 patients (75%) whereas 46 patients (25%) in whom a part of the neoplastic process was not resectable (R1 or R2 resection) underwent a palliative esophagectomy. The 5-year survival, including in-hospital deaths (4.3%), was 35.3% for the whole series, 48% after Ro resection, and 0% after R1 or R2 resection. The 5-year survival rate after any R resection was 57.2% in patients with Barrett adenocarcinoma compared with 20% in patients with non-Barrett adenocarcinoma (P or = 5 metastastic lymph nodes = 6.8% vs. R1, R2 = 0%; P < 0.0001). CONCLUSIONS: Exclusive radical esophagectomy provides a chance of long-term survival in 35% of esophageal adenocarcinoma patients in whom it is attempted and nearly 50% of those patients in whom it is feasible. The presence of a small number of metastatic lymph nodes does not appear to preclude a long-term favorable outcome.

Previous studies have shown that encircling of the esophagogastric junction by a semiabsorbable scarf effectively prevents gastroesophageal reflux. The present study was performed to assess the long-term safety and biocompatibility of this type of scarf. The semiabsorbable scarf was implanted into 20 dogs either laparoscopically or via laparotomy. Pre- and post-operatively, contrast radiography, esophageal manometry, and upper gastrointestinal endoscopy were performed. No cases of perforation, stricture formation or other adverse effects were found after 1 and 2 years. It is concluded that the new type of scarf is without any adverse side-effects. Functional evaluation in reflux patients appears to be warranted.

Successful management of lymphatic leaks by continuous IV administration of somatostatin was first reported by Ulibarri and coworkers in Spain,(1) and more recently by authors from Italy(2) and Switzerland.(3) The present article reports the clinical history of two patients in whom postsurgical lymphatic leak was successfully treated after the administration of either somatostatin-14 alone (case 1) or combined somatostatin-14 and total parenteral nutrition (TPN; case 2). Although further pathophysiologic studies are needed for the elucidation of its mechanisms of action, somatostatin-14 seems to be an intriguing therapy against postsurgical lymphatic leaks that may make potentially risky transthoracic reoperation unnecessary.

OBJECTIVE: To determine whether early postoperative administration of erythromycin accelerates the spontaneous motor recovery process after elevation of the denervated whole stomach up to the neck. SUMMARY BACKGROUND DATA: Spontaneous motor recovery after gastric denervation is a slow process that progressively takes place over years. METHODS: Erythromycin was administered as follows: continuous intravenous (i.v.) perfusion until postoperative day 10 in ten whole stomach (WS) patients at a dose of either 1 g (n = 5) or 2 g (n = 5) per day; oral intake at a dose of 1 g/day during 1.5 to 8 months after surgery in 11 WS patients, followed in 7 of them by discontinuation of the drug during 2 to 4 weeks. Gastric motility was assessed with intraluminal perfused catheters in these 21 patients, in 23 WS patients not receiving erythromycin, and in 11 healthy volunteers. A motility index was established by dividing the sum of the areas under the curves of >9 mmHg contractions by the time of recording. RESULTS: The motility index after IV or oral administration of erythromycin at and after surgery was significantly higher than that without erythromycin (i.v., 1 g: p = 0.0090; i.v., 2 g: p = 0.0090; oral, 1 g: p = 0.0017). It was similar to that in healthy volunteers (i.v., 1 g: p = 0.2818; oral, 1 g: p = 0.7179) and to that in WS patients with >3 years of follow-up who never received erythromycin (i.v., 1 g: p = 0.2206; oral, 1 g: p = 0.8326). The motility index after discontinuation of the drug was similar or superior to that recorded under medication in four patients who did not experience any modification of their alimentary comfort, whereas it dropped dramatically parallel to deterioration of the alimentary comfort in three patients. CONCLUSIONS: Early postoperative contractility of the denervated whole stomach pulled up to the neck under either i.v. or oral erythromycin is similar to that recovered spontaneously beyond 3 years of follow-up. In some patients, this booster effect persists after discontinuation of the drug.

Background Esophagectomy with extensive lymph node dissection is the best way to give Barrett's patients with locally advanced adenocarcinoma a good chance of cure. Material and Methods Fifty-five patients underwent subtotal (n = 47) or distal (n = 8) esophagectomy for Barrett's adenocarcinoma (n = 43) or high-grade dysplasia (HGD) (n = 12). Thirteen patients (23.6%) never had had any reflux symptom before disclosure of the neoplastic lesion, and 20 patients (36.4%) had esophageal shortening. Ro resections (n = 50) included removal of the esophageal tube en bloc with the locoregional lymph nodes. Results An invasive carcinoma was found in the resected specimen of 4 of the 12 patients operated on for HGD. Two of the 5 patients whose metaplasia was surveyed endoscopically were operated on for an advanced lesion (T2N1, T3N1) because they had not strictly complied with the proposed schedule. One of the 4 patients whose HGD was followed up endoscopically until disclosure of deeper mucosal invasion had positive lymph nodes at operation. The prevalence of early lesions (Tis, T1, T2, No) was 7.4% in patients with tumor-related symptoms versus 85.7% in those having unrelated symptoms ( P = 0.0000), which resulted in a 5-year survival rate of 33.8% and 82.4%, respectively ( P = 0.0012). Five-year survival rate after Ro resection made for invasive carcinoma was 59.3% (all cases), 73.1% (No), 61.5% (≤5 positive lymph nodes), and 0% (>5 positive lymph nodes). Conclusions High-grade dysplasia is an indication for esophageal resection. Early detection of the neoplastic transformation of Barrett's metaplasia prior to the onset of obstructive symptoms gives the best chance of cure. Esophagectomy with radical lymph node clearance is capable of curing a large proportion of the patients having no or a limited number of metastatic lymph nodes.

Laparoscopic surgery for gastroesophageal reflux disease has replaced the open approach in several institutions, and it is likely to become the "standard" for treatment in the near future. Members of five European surgical centers with extensive experience in pathophysiological research, diagnostic testing, and conventional surgery for esophageal disease met after five years of experience in using laparoscopic antireflux surgery, and established a plan to evaluate the potential for consensus among the centers involved in the surgical management of the disease. The consensus process started with a pathophysiological assessment of the reporting requirements for diagnostic workup. To allow a thorough appreciation of the surgical techniques used by all the participants, experience was exchanged in collaborative operations in an experimental surgical laboratory. It was concluded that the pathophysiological background to the disease is multifactorial, as many publications have shown in recent years. The group's meetings and discussions established a consensus list for the preoperative assessment of patients suspected of having gastroesophageal reflux disease, as well as a common list of operative techniques for successful antireflux surgery.

OBJECTIVES: The specific contribution of the herniation of an abdominal antireflux fundoplication into the chest to symptomatic and therefore surgical failure remains unclear. METHODS: The study was conducted in 189 consecutive fundoplication patients, categorized as patients reoperated on for chest herniation of either an abdominal 360° (Group 1; n = 95) or a partial (Group 2; n = 10) fundoplication, and patients having undergone an intrathoracic 360° fundoplication for short oesophagus (Group 3; n = 84; reference group). There were four subgroups in Group 1: 1A: wrap still complete and perioesophageal; 1B: wrap still complete but perigastric; 1C: wrap still perioesophageal but partially disrupted and 1D: wrap perigastric and partially disrupted. RESULTS: The prevalence of defective symptoms (heartburn and regurgitation) was significantly lower (P < 0.0001) in Group 3 (0.0%) and Subgroup 1A (3.7%) than in Subgroups 1B (84.4%), 1C (86.7%) and 1D (100%) and Group 2 (100%). The prevalence of obstructive symptoms (dysphagia, chest pain, necrosis and perforation) was significantly higher (P < 0.0001) in Subgroup 1A (100%) than in Subgroups 1B (57.8%), 1C (60.0%) and 1D (25.0%). The prevalence of a short oesophagus, an abdominal wall hernia repair and high abdominal pressure episodes in reoperated patients were 13.7, 36.2 and 67.2%, respectively. CONCLUSIONS: Unlike perigastric or partial fundoplication, a 360° perioesophageal abdominal fundoplication, when herniated into the chest, is still effective against reflux. Obstructive symptoms are due to either diaphragmatic strangulation or perigastric migration of the wrap (slipknot effect). Short oesophagus, weakness of the abdominal wall and high abdominal pressure episodes favour the herniation process.

Background : Detection of subclinical recurence after surgical resection of esophageal cancer would allow earlier treatment of recurrent disease and potentially offer a better outcome for rescue therapy. Methods : The utility of serum carcinoembryonic antigen (CEA) assay was evaluated in the management of patients with esophageal cancer. Results : Serum carcinoembryonic antigen was measured preoperatively in 74 patients. Elevation of the CEA level (>5 ng/mL) was present in 14 patients (19%). There was no relationship between preoperative CEA elevation and the stage of the tumor or the patients' survival. Eighty-three patients had CEA assay at regular follow-up intervals after resection. Objective evidence of recurrent disease was determined at similar intervals by chest radiography and abdominal and thoracic computed tomography scans. During follow-up, 53 of 83 patients developed recurrence. Postoperative elevation of CEA levels occurred in 32 patients, resulting in a sensitivity of 55% for detecting recurrent disease. Twenty-nine of the 32 patients who developed CEA elevation had objective evidence of metastatic disease. In 13 patients, the rise in CEA levels predated objective evidence of recurrence by a median of 4 months (range 3 to 35), and in 16 patients, it occurred concomitantly. The specificity with which an elevated postoperative CEA level indicated recurrence was high, 90%, with a positive predictive value of 91%. Conclusions : Postoperative CEA elevation is highly predictive of recurrent disease. In 16% of patients, elevation of CEA was the earliest objective sign of recurrence; such elevation should prompt consideration of adjuvant therapy.

Background. The stomach can be used either in its entirety or as a greater curvature tube for esophageal replacement. Methods. The study compares the gastric tube (group A; n = 112) to the whole stomach whose lesser curvature is denuded (group B; n = 100) in terms of technical complication and alimentary comfort. The clinical results are substantiated by assessment of the eating performance of patients and control subjects at a test meal, measurement of the gastric dimensions before and after both tailoring procedures, and intraarterial staining of the gastric wall. Results. Major differences between the two groups are cervical anastomosis stenoses (22.3% versus 6% [A versus B]; p = 0.008), fistulas (7.9% versus 1%; p = 0.0209), number of meals and snacks per day (4.6 versus 4; p = 0.0275), sensation of early fullness at meals (52.4% versus 17.8%; p < 0.0001), ratings given to the long-term alimentary comfort (presymptomatic condition = 10 points) (7.6 versus 8.8 out of 10 on average; p < 0.0001), and calories consumed in 1 minute at a test meal (59% [p < 0.05] versus 77% of those consumed by control subjects). The volume of the stomach is reduced by a range of 21.4% to 47.2% after tubulization (group A) whereas it increases by a range of 4.9% to 17.4% after denudation of the lesser curve (group B). Intraarterial staining of the gastric wall reveals the poor vascularity of the upper-most segment of the greater curve. Conclusion. Slight increase of the gastric capacity and maintenance of the submucosal vascular network account for the better results achieved with the whole stomach.

From 1976 until April 1989, 31 intrathoracic total fundoplications were performed for reflux esophagitis and irreducible hiatus hernia. In the first 16 patients (group 1) the operation was complicated with acute perforation of the wrap in 4 cases, bronchogastric fistula in 1, and herniation of the wrap higher in the chest in 1. Technical modifications were applied to 15 more recent patients (group 2). These are enlargement of the hiatus, looseness of the wrap and its appropriate anchorage, avoidance of forceps when handling the stomach, care with the vagi, and efficient gastric decompression in the postoperative period. The postoperative course was always uneventful in group 2. Twenty-six patients, who still have their initial wrap, were considered for clinical evaluation: 11 from group 1 (mean follow-up, 81.5 months) and 15 from group 2 (mean follow-up, 32.8 months). All are free from any symptom of reflux; gas-bloat syndrome is infrequent and dysphagia is relieved. Twenty-four-hour pH monitoring, performed in 14 patients (3 from group 1 and 11 from group 2) (mean follow-up, 42 months), was normal in 13; a pathological upright reflux (time pH less than 4, 8.4%) was demonstrated in one symptom-free woman in whom endoscopy was unremarkable. Mechanisms of complications experienced in group 1 are analyzed in the light of the technical evolution of the procedure, and the place of the intrathoracic total fundoplication in the management of short esophagus is defined, considering the other available surgical techniques.

Subtotal esophagectomy may consist of either resection of the lower 90% of the thoracic segment of the esophagus with subsequent esophagogastrostomy at the apex of the chest, or resection of the whole thoracic segment plus the lower segment of the cervical part of the esophagus with subsequent cervical esophagogastrostomy.

OBJECTIVE: To assess the long-term results of the duodenal switch operation made for pathologic transpyloric duodenogastric reflux (DGR). SUMMARY BACKGROUND DATA: DGR symptoms and lesions are poorly responsive to medical treatment. METHODS: A duodenal switch operation was made on 48 patients suffering from pathologic transpyloric DGR either unrelated (n = 28) or secondary (n = 20) to previous upper gastrointestinal (GI) surgery, including cholecystectomy or vagotomy. The diagnosis was based on the combination of several objective arguments: a long history of gastric symptoms (ie, nausea, epigastric pain, and/or bilious vomiting) poorly responsive to medical treatment (48 of 48), gastroesophageal reflux symptoms unresponsive to proton-pump inhibitors (PPI) (23 of 29), gastritis on upper GI endoscopy (37 of 48) and/or at histology (28 of 41), presence of a bilious gastric lake at >1 upper GI endoscopy (30 of 48), DGR at diisopropyl iminodiacetic acid (DISIDA) scintigraphy scanning (7 of 13), pathologic 24-hour intragastric bile monitoring with the Bilitec device (40 of 41), and absence of Helicobacter pylori antral infection (39 of 41). RESULTS: At follow-up (median, 81 months), gastric symptoms were nil, had improved, and remained unchanged in 29 (60.4%), 16 (33.3%), and 2(4.2%) patients, respectively, and 1 patient experienced symptomatic recurrence after a 92-month symptom-free period (2.1%). Among the 44 patients who had postoperative upper GI endoscopy, 42 (95.5%) had no gastritis whereas 5 (11.3%) had an ulcer at the duodenojejunostomy. Gastric exposure to bile at postoperative 24-hour intragastric Bilitec test in 36 patients was nil, within the normal range, and still slightly pathologic in 15 (41.7%), 19 (52.8%), and 2 (5.5%), respectively. CONCLUSIONS: The duodenal switch operation made on patients in whom diagnosis of pathologic transpyloric DGR is supported by several objective arguments provides most of them with symptomatic and endoscopic improvement parallel to abolishment or normalization of gastric exposure to bile. Postoperative PPI therapy during a 2-month period is to be recommended to prevent the development of an anastomotic ulcer.

With introduction of the Bilitec 2000 device, intraluminal bile monitoring has become a standard technique for evaluation of patients with gastroesophageal reflux disease and symptoms of bile reflux. A specific “white diet≓ excluding colored food is necessary to provide correct measurements. The influence of this specific diet on simultaneous esophageal pH monitoring is unknown. Forty patients with reflux symptoms were studied prospectively. Meal times and supine and erect phases of measurement were recorded in a standardized fashion using a patient protocol. Esophageal 24-hour pH monitoring with administration of a “colorless≓ diet (water, milk, potatoes, fish, chicken) was started on day 1, followed by esophageal 24-hour pH-metry with intake of a normal diet on day 2. Data from the two successive pH measurements were compared. The prevalence of a pathologic esophageal pH-metry was significantly higher during intake of a normal diet compared with a colorless diet (P = 0.025). During total and upright phases, administration of a white diet led to significant reduction in the percentage of time with a pH less than 4 (P ≤ 0.01), the total number of reflux episodes (P ≤ 0.001), and the DeMeester's score (P = 0.01). This difference was exclusively found in patients with a normal pH-metry (group 1, n = 13) and pathologic upright reflux (group 2, n = 12). No change in reflux pattern was found in patients with isolated supine reflux (group 3, n = 7) and combined upright and supine reflux (group 4, n = 8). In patients with a pathologic upright reflux pattern, administration of a white diet results in a significant modification of esophageal pH-metry. Data derived from simultaneous esophageal pHand bile monitoring should be interpreted with care.

The paper describes an original technique of gastric tailoring in which the two-thirds of the lesser curvature proximal to the crow's foot are denuded flush with the gastric wall, leaving both nerves of Latarjet and the hepatic branches of the left vagus nerve intact. Maintenance of the vagal supply to the antro-pyloric segment in two patients resulted in the presence of peristaltic contractions sweeping over the antrum on simple observation of the antral wall at the end of the procedure and on both upper G-I series and intragastric manometry tracings 6 weeks postoperatively. Gastric exposure to bile on 24-h gastric bile monitoring was normal 6 weeks after the operation. Neither patient had any gastrointestinal symptoms with the exception of early sensations of postprandial fullness when overeating.

Background The degree which the various reconstruction techniques prevent bile reflux after gastroduodenal surgery has been poorly studied. Methods Bile exposure in the intestinal tract just proximal to the jejunal loop was measured with the Bilitec 2000® device for 24 h after gastroduodenal surgery in three groups of patients. Group 1 comprised 24 patients with a 60-cm Henley's loop after total gastrectomy. Group 2 included 31 patients with a 60-cm Roux-en-Y loop after total (22 patients) or subtotal (nine) gastrectomy. Group 3 contained 21 patients with a 60-cm Roux-en-Y loop anastomosed to the proximal duodenum as part of a duodenal switch operation for pathological transpyloric duodenogastric reflux. Bile exposure, measured as the percentage time with bile absorbance greater than 0·25, was classified as nil, within the range of a control population of healthy subjects, or pathological (above the 95th percentile for the control population). Reflux symptoms were scored and all patients had upper gastrointestinal endoscopy. Results Bile was detected in the intestine proximal to the loop in none of 24 patients in group 1, eight of 31 in group 2 and 12 of 21 in group 3 (P < 0·001). The mean reflux symptom score increased with the degree of bile exposure, and the proportion of patients with oesophagitis or gastritis correlated well with the extent of bile exposure (P < 0·001). Conclusion A long Henley's loop was more effective in preventing bile reflux than a long Roux-en-Y loop. Bilitec® data correlated well with the severity of reflux symptoms and the presence of mucosal lesions.

BACKGROUND: The management of dysplasia arising in Barrett's esophagus is controversial. STUDY DESIGN: Twenty patients (group 1, prompt attitude) underwent operation as soon as high-grade dysplasia (HGD) was discovered (n = 8) or just after either the presence of HGD was confirmed (n = 9) or invasive carcinoma (IC) was found (n = 3) in a second set of biopsy samples taken soon after HGD had been discovered. In contrast, esophagectomy in 13 patients (group 2, expectant attitude) was performed only because HGD persisted (n 4) or turned into IC (n = 4) at endoscopic followup (7 to 23 months) (subgroup 2a, n 8) or because HGD (n = 2) or low-grade dysplasia (LGD) (n = 3) was disregarded until dysphagia and IC developed (12 to 70 months) (subgroup 2b, n = 5). Skeletonizing en-bloc esophagectomy was performed in 29 patients and four patients (three with HGD and one with mucosal IC in the resected specimen) underwent vagus-sparing esophagectomy. RESULTS: Invasive carcinoma was found in 11 of 24 patients (45.8%) supposed. to have only HGD (in repeat biopsies in 3 patients from group 1 and in the resected specimen in eight of 21 patients (38%) operated on for HGD. Metastatic lymph nodes were found in the resected specimen of seven patients (group 1: one of 20 or 5%, versus subgroup 2a: two of eight or 25%, versus subgroup 2b: four of five or 80%; p = 0.001). Unlike none of the 26 patients (0%) with an intramural process, five of the seven patients (71.4%) with an extramural process (one had had disregarded LGD) developed neoplastic recurrence at followup (p < 0.0001). Cancer-related survival in the long term was 100% in group 1 versus 52.5% in group 2 (p = 0.0094). CONCLUSIONS: Invasive carcinoma is present in almost one half of patients with HGD within a Barrett's area. Promptness in the decision regarding an esophageal resection as soon as HGD is found is much safer than expectant observation. Not enrolling a patient with LGD in an endoscopic surveillance program can lead to the development of extramural IC with poor outcomes. (C) 2003 by the American College of Surgeons.

The main principles for optimal management of HGD arising in Barrett's esophagus are that unequivocal diagnosis of HGD is a prerequisite for making the decision of any kind of treatment. HGD must be resected because of the presence of neoplastic cells in the lamina propria in 40% of patients. No reliable endoscopic or endosonographic feature exists that allows accurate prediction of the existence of neoplastic cells within the lamina propria of a patient having HGD in endoscopic biopsy material. Prompt decision to remove an HGD lesion as soon as unequivocal histologic diagnosis has been settled prevents the development of extraesophageal neoplastic spread. Esophagectomy is preferable to endoscopic mucosal excision because approximately 20% of patients who have HGD in preoperative biopsy material carry neoplastic cells beyond the muscularis mucosae. Esophagectomy can be limited to the removal of the esophageal tube without extended lymphadenectomy because 96% of patients who have HGD in endoscopic biopsy samples have a neoplastic process confined to the esophageal wall. Esophageal resection must encompass all the Barrett's area because of the risk for the further development of a second cancer in the metaplastic remnant. Vagus-sparing esophagectomy with colon interposition or elevation of the antrally innervated stomach up to the neck is preferable to conventional esophagectomy with gastric pull up because the former procedure maintains gastric function intact, whereas the latter exposes patients to the risk for the long-term development of reflux esophagitis and even of metaplastic transformation of the proximal esophageal remnant. Subtle details in the understanding of a given patient's clinical course may be critical for making the decision of the most relevant mode of therapy; therefore, patients who have HGD should be treated in dedicated centers, the experience of which offers the best chances of uneventful recovery if the surgical option is retained.

BACKGROUND: Both the supine position and the existence of a gastric drainage procedure are suspected to promote reflux of duodenal juice into the denervated intrathoracic stomach. Erythromycin has been shown to weaken pyloric resistance to gastric outflow and to enhance antral motility, gastric emptying, and gallbladder contractility. METHODS: The presence of bile in the gastric transplant of 79 patients was monitored over a 24-hour period with use of the Bilitec 2000 optoelectronic device 3 to 195 months after subtotal esophagectomy. Ten patients were reinvestigated after a 3-year period. Five groups were studied: group I: n = 12, no gastric drainage, never given erythromycin, group 2: n = 40, gastric drainage, never given erythromycin, group 3: n = 7, no gastric drainage, given erythromycin, group 4: n = 13, gastric drainage, given erythromycin, and group 5: n = 7, no longer given erythromycin (with or without gastric drainage). The percentage of time gastric bile absorbance was more than 0.25 was calculated for the total, supine, and upright periods of recording in reference to data from 25 healthy volunteers. RESULTS: The Bilitec test was pathologic in 9 of the 12 patients of group 1 whereas it was normal in three. Gastric exposure to bile was longer in group I patients than in controls for the total (p = 0.012) and supine (0.036) periods, but the difference did not reach statistical significance for the upright period (p = 0.080). Bile exposure in group 4 did not significantly differ from controls (total: p = 0.701; supine: p = 0.124; upright: p = 0.712). Bile exposure for the total period did not significantly differ whether patients were taking erythromycin or the drug had been discontinued at the time of the study (p = 0.234); and it tended to decrease with time in patients investigated twice (p = 0.046). CONCLUSIONS: Gastric exposure to bile after truncal vagotomy and transposition of the stomach up to the neck is pathologic in three quarters of patients. It is more marked in the supine than in the upright position and tends to decrease with time. The addition of a gastric drainage procedure in combination with erythromycin therapy tends to normalize gastric exposure to bile. The effects of erythromycin may persist after discontinuation of the drug.

OBJECTIVE: To evaluate the long-term outcome of patients with esophageal cancer after resection of the extraesophageal component of the neoplastic process en bloc with the esophageal tube. SUMMARY BACKGROUND DATA: Opinions are conflicting about the addition of extended resection of locoregional lymph nodes and soft tissue to removal of the esophageal tube. METHODS: Esophagectomy performed en bloc with locoregional lymph nodes and resulting in a real skeletonization of the nonresectable anatomical structures adjacent to the esophagus was attempted in 324 patients. The esophagus was removed using a right thoracic (n = 208), transdiaphragmatic (n = 39), or left thoracic (n = 77) approach. Lymphadenectomy was performed in the upper abdomen and lower mediastinum in all patients. It was extended over the upper mediastinum when a right thoracic approach was used and up to the neck in 17 patients. Esophagectomy was carried out flush with the esophageal wall as soon as it became obvious that a macroscopically complete resection was not feasible. Neoplastic processes were classified according to completeness of the resection, depth of wall penetration, and lymph node involvement. RESULTS: Skeletonizing en bloc esophagectomy was feasible in 235 of the 324 patients (73%). The 5-year survival rate, including in-hospital deaths (5%), was 35% (324 patients); it was 64% in the 117 patients with an intramural neoplastic process versus 19% in the 207 patients having neoplastic tissue outside the esophageal wall or surgical margins (P

To discover whether the human stomach contracts every 20 seconds or not.Manometric study.Teaching hospital, Belgium.10 healthy volunteers, and 31 patients who had had the whole stomach denervated and pulled up to the neck for oesophageal replacement.Analysis of selected strips of manometric tracings obtained with intraluminal perfused catheters. 13 patients were given erythromycin (1g/day) by mouth.Estimation of the rate and frequency distribution according to amplitude of intraluminal pressure waves with the vertical axis of the tracings scaled up to reflect contractions within the gastric wall.Microwaves (9 mmHg) that came in between conventional macrowaves (9 mmHg) were found, showing that the human stomach undergoes mechanical activity (amplitude ranging from 0.2-310 mmHg) at the pacemaker's rate which varied from 2.43 to 3.60 cycles/minute from one subject to another. Phase I of the interdigestive motor complex contained microwaves only, phase II and the fed pattern consisted of a mixture of microwaves and macrowaves, and phase III contained macrowaves only. The fasting rate of mechanical activity was lower in patients who were given erythromycin than in those not given erythromycin (p = 0.003) and in healthy volunteers (p = 0.002), and it increased significantly after a meal (p0.0001). Microwaves in strips in which they were the most prominent were of higher amplitude in patients than in healthy volunteers (median: 3.5 compared with 2.5 mmHg; p0.0001).The human stomach has mechanical activity at the rate at which the pacemaker generates electrical slow waves. The classic phases of the gastric motor activity seem to differ from each other by the frequency distribution of pressure waves according to amplitude rather than by the contraction rate. Weak mechanical activity is much more readily detectable after the stomach has been denervated and tailored for oesophageal substitution.

Objective To determine whether the denervated stomach as an esophageal substitute recovers normal intraluminal acidity with time, Summary Background Data Bilateral truncal vagotomy to the stomach as an esophageal substitute reduces both gastric acid production and antral motility, but a spontaneous motor recovery process takes place over years. Methods Intraluminal gastric pH and bile were monitored during a 24-hour period 1 to 195 months after transthoracic elevation of the stomach as esophageal replacement in 91 and 76 patients, respectively. Nine patients underwent a second gastric pH monitoring after a 3-year period. The percentages of time that the gastric pH was less than 2 and bile absorbance exceeded 0.25 were calculated in reference to values from 25 healthy volunteers. Eighty-nine upper gastrointestinal endoscopies were performed in 83 patients. Patients were divided into three groups depending on length of follow-up: group 1, less than 1 year; group 2, 1 to 3 years; group 3, more than 3 years. Results The prevalence of a normal gastric pH profile was 32.3% in group 1, 81.5% in group 2, and 97.6% in group 3. The percentage of time that the gastric pH was less than 2 increased from group 1 (27.3%) to group 2 (56.1%) and group 3 (70.5%), parallel to an increase in the prevalence of cervical heartburn and esophagitis. The percentage of time that the gastric pH was less than 2 increased from 28.7% to 81.2% in the nine patients investigated twice. Exposure of the gastric mucosa to bile was 12.8% in patients with a high gastric pH profile versus 19.3% in those with normal acidity. In the esophageal remnant in six patients, Barrett's metaplasia developed, intestinal (n = 2) or gastric (n = 4) in type. Conclusions Early after vagotomy, intraluminal gastric acidity is reduced in two thirds of patients, but the stomach recovers a normal intraluminal pH profile with time, so that in more than one third of patients, disabling cervical heartburn and esophagitis develop. The potential for the development of Barrett's metaplasia in the esophageal remnant brings into question the use of the stomach as an esophageal substitute in benign and early neoplastic disease.

Various esophageal procedures are now feasible through a few holes in the thoracic or abdominal wall, or through the working channel of an endoluminal endoscope1 rather than through a large parietal incision. This is the case for antireflux fundoplication2,3, myotomies4, enucleation of a benign tumor5, clippage of the thoracic duct6, and so on ... The esophagus itself can be removed by so-called minimally invasive approaches that are the right thoracoscopy7–14, the transcervical mediastinoscopy combined with conventional transhiatal dissection by laparotomy15–17, and the laparoscopic transhiatal esophagogastric mobilization in combination with conventional cervicotomy for esophageal extraction and esophagogastric anastomosisl8–23. All those new surgical modalities are not only very attractive because they obviously reduce the parietal damage related to the classic incisions but they constitute also a very appealing technical challenge for the surgeon.

The role of excessive duodenogastric reflux (DRG) in the genesis of gastric symptoms in patients primarily referred for both gastroesophageal reflux (GER) symptoms and esophagitis is poorly understood.The study is based on the clinical, endoscopic, histologic, and 24-hour gastric data from the Bilitec optoelectronic device (Prodotec, Florence, Italy, licensed by Synectics Medical, Stockholm, Sweden) from 49 patients having both typical GER symptoms and gastric symptoms suggestive of excessive DGR (i.e., epigastric pain, nausea, or bilious vomiting) in the absence of previous esophageal or gastric surgery (group 1). Helicobacter pylori organisms were searched for on antral biopsy specimens with use of the Giemsa method. The percentages of total, upright, and supine time during which absorbance exceeded various thresholds through all the working range of the Bilitec device were calculated. Bilitec data from group 1 were compared with those from 16 patients with endoscopic esophagitis and GER symptoms only (group 2) and 25 healthy subjects (group 3).The prevalence of an abnormal Bilitec test result in group 1 increased from 27% (13/49) at the 0.25 absorbance threshold to 36% (18/49) at thresholds ranging from 0.40 to 0.60 and to 41% (20/49) when multiple thresholds ranging from 0.25 to 0.60 were considered. In group 2 one patient had an abnormal Bilitec test result at the 0.25 to 0.30 threshold, whereas the other 15 patients had a normal test result. H pylori antral infection was present in 14 group 1 patients. None of these had an abnormal Bilitec test result, whereas the test was positive in 40% of the H pylori-negative patients without endoscopic gastritis and in 70% of H pylori-negative patients with endoscopic gastritis (P = .001).Twenty-four-hour intragastric bile monitoring provides the clinician with unequivocal evidence of excessive DGR in 41% of patients with an intact stomach having endoscopic esophagitis, GER symptoms, and gastric symptoms suggestive of DGR. The most dependable data are obtained when absorbance thresholds higher than 0.40 are considered. H pylori antral infection and excessive DGR at 24-hour intragastric bile monitoring are mutually exclusive.

BACKGROUND: The current opinion is that the reflux of jejunal juice over the whole length of a long Roux-en-Y jejunal loop is very uncommon. We aimed to challenge this concept by monitoring the presence of bile in the organ proximal to a 60-cm loop during a 24-hour period with use of the Bilitec device, an optoelectronic instrument capable of measuring absorbance of a beam of light, the wavelength of which is close to the absorbance peak of bilirubin. PATIENTS AND METHODS: Forty-one patients, 8 of whom had been cholecystectomized, were investigated after total gastrectomy (group I, n = 17), distal gastrectomy (group II, n = 7), or duodenal switch (group III, n = 17). The percentage of recording time absorbance >0.25 (absorbance scale ranging from 0 to 1) was calculated in reference to data from healthy subjects. RESULTS: Bile was detected in 17 patients (41%), 5 belonging to group 1, 2 to group II, and 10 to group III (P = 0.165). Bile exposure remained within the range of controls in 14 patients whereas it was above this range in 3 patients, 2 of whom had disabling heartburn and severe esophagitis. The percentage of time absorbance >0.25 did not significantly differ from one group to another (P = 0.257) or according to whether patients had been cholecystectomized or not (P = 0.439). However, unlike cholecystectomized patients, patients still having their gallbladder refluxed predominantly during postprandial periods. Lengthening of the loop from 60 cm to 110 cm in the 2 symptomatic patients with a pathologic bile reflux resulted in relief of heartburn and healing of esophagitis in both while bile reflux was abolished in 1 and dramatically reduced in the other. CONCLUSIONS: Bile refluxes over the whole length of 4 Roux-en-Y loops out of 10. In most patients, bile reflux remains within the range of healthy subjects, producing neither symptoms nor mucosal damage; and it occurs independently of the organ proximal to the loop, but its timing of occurrence is modified by cholecystectomy. Although only for exceptional indications, lengthening of an incompetent loop is effective in patients with excessive bile reflux and severe related symptoms and lesions. Am J Surg. 2000;179:298-303. (C) 2000 by Excerpta Medica, Inc.

Duodenogastric reflux (DGR) is suspected to be an aetiological factor in the pathogenesis of foregut disease. The `Bilitec' bile probe allows continuous detection of bilirubin, based on spectrophotochemical properties. We aimed to describe duodenogastric bile reflux in healthy, normal volunteers in a Western European population, as a basis for the future study of DGR in disease. An international multicentre study was established. DGR was measured using 24 h ambulatory bile and pH monitoring in the proximal stomach, in 43 normal volunteers from the third to the seventh decades. Subjects adhered to a standard protocol. The total test period, supine and upright components, were analysed. The 90th percentile values for absorbance thresholds of 0.14, 0.25, 0.3, 0.4 and 0.5 were 40.5%, 20.9%, 19.6%, 11.6% and 4.6% of the total time respectively. There was a wide range of absorbance within each threshold. Supine DGR was greater than upright, and associated with an alkaline tide. The upright phase was further subdivided into upright fasting, prandial and post-prandial phases, and ranges for these periods are also described. No relationship between age, weight, or body mass index and duodenogastric reflux was seen. The results of this study form a range which allows further investigation into the contribution of duodenogastric bile reflux in the pathogenesis of foregut disease.

A 60-year-old woman developed a gastrodiaphragmatic fistula as a late complication of transabdominal Nissen fundoplication. For 6 years, she had complained of chest pain and was considered to have pericarditis. At reoperation the fistulous track was found and easily resected. No other causative factor could be identified.

Background. The classic manual end-to-side technique of esophagogastrostomy after gastric pull-up to the neck carries a rather high risk of fistula and stricture. Methods. A terminalized semimechanical side-to-side technique of cervical esophagogastrostomy was performed in 16 patients by the application of an Endo-GPA stapler across the gastric and esophageal walls placed side by side, so as to create a V-shaped posterior opening between the two lumina. The anterior aspect of the anastomosis was hand-sewn using a classic running suture. The cross-sectional area of the semimechanical anastomoses was estimated by barium swallow study 2 months after operation and compared with that of 24 manual end-to-side esophagogastrostomies. Results. The cross-sectional area was 225 +/- 15.7 mm(2) (mean a standard error of the mean) or the 16 semimechanical anastomoses versus 136 +/- 15 mm(2) for the 24 manual anastomoses (p = 0.0001). The anastomotic area decreased from 206.6 +/- 13.5 mm(2) in 29 patients without dysphagia to 107.5 +/- 4.7 mm(2) in 7 patients with moderate dysphagia for solids that did not require endoscopic dilation and to 55.7 +/- 16 mm(2) in 4 patients with severe dysphagia that required dilation (p = 0). The anastomotic area in 6 of the 7 patients with initial moderate dysphagia for solids increased spontaneously with time from 107.3 +/- 5.5 mm(2) to 174.6 +/- 8.1 mm(2), with concomitant symptomatic relief (p = 0.0277). Conclusions. The terminalized semimechanical side-to-side suture technique produces a larger anastomosis than the classic end-to-side esophagogastrostomy technique. inflammatory changes related to the operation may cause transient narrowing of a cervical esophagogastrostomy. (C) 1998 by The Society of Thoracic Surgeons.

OBJECTIVE: To determine whether the denervated stomach as an esophageal substitute is an inert conduit or a contractile organ. SUMMARY BACKGROUND DATA: The motor response of gastric transplants to deglutition suggests that the stomach pulled up to the neck acts as an inert organ. METHODS: The gastric motility of 11 healthy volunteers and 33 patients having either a gastric tube (GT) (n = 10) or their whole stomach (WS) (n = 23) as esophageal replacement was studied with perfused catheters during the fasting state, after a meal, and after intravenous administration of erythromycin lactobionate. A motility index was established for each period of recording by dividing the sum of the areas under the curves of all contractions of >9 mmHg by the time of recording. RESULTS: Over years, the denervated stomach recovers more and more motor activity, even displaying a real phase 3 motor pattern in 6 of the 10 WS patients and 1 of the 7 GT patients with >3 years of follow-up. Erythromycin lactobionate generates a phase 3-like motor pattern regardless of the length of follow-up. Extrinsic denervation of the whole stomach does not significantly modify the fasting motility index established >3 years after surgery (+17% on average, p > 0.05), but it reduces that in the fed period by an average of 62% (p = 0.0016). Tubulization of the denervated whole stomach lowers the fasting motility index by an average of 60% (p = 0.0248) and further impairs that in the fed period by an average of 67% (p = 0.0388). CONCLUSIONS: The denervated stomach as an esophageal substitute is a contractile organ that may even generate complete migrating motor complexes. Motor recovery is better in the fasting than in the fed period, and it is more marked in WS patients than in GT patients.

OBJECTIVE: The authors aim to substantiate, with objective arguments, potential advantages of laparoscopic versus open antireflux surgery in the light of the recent crude experience of the Louvain Medical School Hospital. METHODS: Seventy-two consecutive patients with disabling gastroesophageal reflux disease ([GERD], n = 56), symptomatic hiatal hernia without GERD (n = 5), or unsatisfactory outcome after unsuccessful antireflux procedure (n = 11) were operated on by laparotomy (n = 28), laparoscopy (n = 39), or thoracotomy (n = 5). The antireflux procedure was a subdiaphragmatic Nissen fundoplication (n = 60), an intrathoracic Nissen fundoplication (short esophagus, n = 3), a subdiaphragmatic 240 degrees fundoplication (severe motility disorders, n = 3), a Lortat-Jacob repair (hiatal hernia without GERD, n = 5), and a duodenal diversion (delayed gastric emptying, n = 1). RESULTS: Major postoperative morbidity included two pulmonary embolisms (one laparoscopy patient and one laparotomy patient), and one hemothorax (one thoracotomy patient). Mean hospital stay was 6.4 days for laparoscopy, 7.8 days for laparotomy, and 12.5 days for thoracotomy. Postoperative morphine consumption (patient-controlled analgesia) averaged 47 mg/48 hrs (laparoscopy) versus 46 mg/48 hrs (laparotomy with primary antireflux surgery) (p > 0.05). Although 93% of the laparoscopy patients returned to work within 3 weeks after surgery, 92% of the laparotomy and thoracotomy patients resumed their activity after more than 6 weeks. At follow-up, 87.5% of the patients were asymptomatic or had inconsequential symptoms, 9.8% had disabling side effects, and 2.7% had persistent or recurring esophageal symptoms. There were four parietal herniations, i.e., one incisional hernia and one recurrence of a repaired umbilical hernia in the laparotomy group, and two herniations of the wrap into the chest--probably related to a premature return to manual work--in the laparoscopy group. Three laparoscopy patients were dissatisfied with the esthetics of their scars. Lower esophageal sphincter pressure and esophageal acid exposure in the laparoscopy patients who were investigated were normal in 100% and 95%, respectively. CONCLUSIONS: Laparoscopy is a good approach for achieving successful antireflux surgery in selected cases. However, its fails to substantially reduce postoperative complication rate and discomfort, duration of the hospital stay, and the risk of esthetic sequela. Early return to work is questionable for manual workers. The subdiaphragmatic Nissen fundoplication is not an all-purpose antireflux procedure.

We present an endoscopic technique of division of the common wall between the esophagus and the hypopharyngeal (Zenker's) diverticulum. The novelty of the technique, as compared with endoscopic sutureless coagulating methods, consists of stapling the esophageal to the diverticular wall using the Endo-GIA 30 stapler (US Surgical Corp, Norwalk, CT), which protects the neck from any contamination from the digestive lumen and ensures optimal hemostasis of the wound edges. The stapler has been designed such that perforation of the bottom of the diverticulum is not likely. The technique has been applied to 6 patients.

From 1984 to 1989, 175 esophageal cancer patients, 10 patients admitted for severe caustic esophagitis, and 1 patient with pyothorax due to iatrogenic perforation of the esophagus underwent an esophageal resection or bypass operation. One hundred sixty-eight esophageal resections were performed on 167 patients; 13 were total, 106 subtotal and 49 distal. Nineteen digestive transplants were pulled up to the neck to bypass the esophagus or re-establish continuity after an esophagectomy made elsewhere. Digestive continuity was restored by a long gastric transplant in 120 patients, a colon segment in 17, a jejunal loop in 35, and a short gastric transplant after limited esophago-gastrectomy in 14 patients. Thirty day mortality was 0 in the whole group. Hospital mortality was 1.2% in the resection group and 10.5% in the bypass group (p = 0.048). Nonfatal postoperative complications consisted of respiratory distress in 33 patients, recurrent nerve palsy in 10, anastomotic fistula in 10 (cervical in 8 and intrathoracic in 2) and anastomotic stenosis in 18 patients. Respiratory complications were more frequent in patients with a cancer of the thoracic esophagus (29/111) than in those operated on for a cancer located in the esophago-gastric junction (4/50) (p < 0.01). Anastomotic stenosis occurred more frequently in the neck (17/137) than in the chest (1/49) (p < 0.05). Nine patients were reoperated on for a technical complication; intraabdominal hemorrhage (1), thoracic duct injury (2), acute cholecystitis (1), tight stricture of the esophageal anastomosis (2), jejuno-duodenal anastomotic fistual (2), or stridor related to recurrent nerve palsy (1). Low postoperative mortality may be achieved after esophageal surgery, even in a nonselected Caucasian population. However, nonfatal complications indicate that an esophageal resection or bypass operation remains a major surgical procedure. The surgical principles which we have applied to minimize postoperative mortality and severe complications are presented.

Subtotal esophagectomy was attempted by right thoracoscopy on 13 patients, 10 having cancer and 3 long caustic stenosis. Thoracoscopy was converted into thoracotomy in 2 patients, owing to loss of selectivity in one-lung ventilation in 1 and injury to a right intercostal artery flush to the aorta in the other. One patient with cancer underwent an esophageal bypass operation only, owing to tumor invasion into the lung at exploratory thoracoscopy. The ten esophagectomies that could be performed in totality by thoracoscopy consisted of seven en bloc resections of the esophagus with extensive lymph node clearance in the posterior mediastinum, and three standard resections without any lymph node dissection. Postoperative complications included one death due to hepatic failure, two cases of acute pneumonitis, and one persistent chest wall discomfort at the trocar sites. Up to 51 lymph nodes were found in the resected specimens of the cancer patients. Six of the 7 cancer patients who were discharged from the hospital after esophagectomy completed by thoracoscopy were alive at 2 to 20 months of follow-up. Five of them were disease free. The study shows that esophageal resections as extensive as those carried out by thoracotomy can be performed by thoracoscopy. It suggests that prompt management of untoward injury to any mediastinal structure adjacent to the esophagus is less easy by thoracoscopy than by thoracotomy, and that classic complications of open thoracic surgery may occur after thoracoscopy as well.

Two patients with a duodenal lipoma had severe gastrointestinal bleeding. Duodenal lipoma is relatively rare and an even rarer source of severe gastrointestinal bleeding. Endoscopy and routine radiological examination of the gastrointestinal tract may suggest the diagnosis, but computed tomography allows a definitive preoperative diagnosis. Surgery is indicated in order to avoid complications.

Cysts arising from embryonic remnants of the primitive foregut include bronchogenic cyst, oesophageal duplication and neurenteric cyst (Kirwan et al, 1973). The place of computed tomography (CT) and needle biopsy in the diagnosis and the management of foregut cysts of the mediastinum has been demonstrated (Schwartz et al, 1985, 1986; Kuhlman et al, 1988). We report a case of a foregut cyst diagnosed and treated by these methods, but without long-term symptomatic relief inasmuch as symptoms recurred 3 years later with re-accumulation of fluid. In 1986, a 27-year-old man complained of vague right thoracic pain. A right mediastinal mass was discovered on chest radiographs (Fig. la, b). A CT scan showed a tubular cystic mass (internal density + 30 HU). This had regular and smooth borders and was located on the right side of the entire thoracic oesophagus (Fig. lc, d). There was no subdiaphragmatic extension and a CT scan of the upper abdomen did not show any abnormality. In particular, no pancreatic lesion wa...