The Management of Nonmalignant. intrathoracic esophageal perforation,

Transcription

1 The Management of Nonmalignant Intrathoracic Esophageal Perforations Richard J. Finley, F.R.C.S.(C), F. Griffith Pearson, F.R.C.S.(C), Richard D. Weisel, F.R.C.S.(C), Thomas R. J. Todd, F.R.C.S.(C), Riivo Ilves, F.R.C.S.(C), and Joel Cooper, F.R.C.S.(C) ABSTRACT Eight patients with nonmalignant intrathoracic esophageal perforations recognized more than 48 hours (48 hours to 14 days) after rupture were treated at Toronto General Hospital between 1973 and Perforation was due to postemetic rupture in 7 patients and to instrumentation in 1. The patients were seen with pain (8), vomiting (71,fever (7), shock (4), respiratory insufficiency (51, pleural effusion (7), pulmonary infiltrates (71, and leukocytosis (6). All patients were managed with thoracotomy. Direct suture closure of the perforation was carried out in 4 patients with midesophageal perforations. Postoperative localized leaks developed in 2 of these patients but healed with conservative management. Cervical esophagostomy and esophageal diversion were used in 1 patient in whom a severe empyema developed in the postoperative period. Direct suture closure, reinforced with a gastric patch, was used to close three lower esophageal perforations. None of these patients had a postoperative leak but all developed subsequent reflux esophagitis. All 8 patients survived. In patients with delayed recognition of a nonmalignant intrathoracic esophageal perforation, elimination of continued chemical and bacterial contamination can be achieved by a clear definition and closure of the esophageal mucosal margins. The obliteration of potential pleural spaces by good tube drainage, lung decortication, and the elective use of mechanical ventilation with positive end-expiratory pressure decreases the incidence of uncontrolled intrapleural sepsis. Perforation of the intrathoracic esophagus is a life-threatening condition and requires prompt From the General Thoracic Unit, Toronto General Hospital and the Department of Surgery, University of Toronto, Toronto, Ont, Canada. Presented at the Sixteenth Annual Meeting of The Society of Thoracic Surgeons, Jan 20-23, 1980, Atlanta, GA. Address reprint requests to Dr. Cooper, Department of Thoracic Surgery, Toronto General Hospital, University Ave, Toronto, Ont, Canada. diagnosis and immediate surgical management. When a perforation of the intrathoracic esophagus is untreated, gastric juices and bacteria-laden oropharyngeal secretions leak into the mediastinum and produce a fulminant inflammatory response. At first, the mediastinitis is chemical but within hours of the rupture, bacterial invasion occurs. If the rupture extends into the pleural space, the mediastinitis may be less severe but the pleural cavity becomes extensively infected. In either case, dehydration, bacteremia, and circulatory collapse usually develop within hours. Factors affecting the outcome of esophageal ruptures include the age and general health of the patient, the location, size, and cause of the perforation, the interval between rupture and treatment, the type of treatment, and the presence of preexisting esophageal disease. After 40 years of age, mortality from spontaneous rupture of the esophagus increases progressively with age [12]. Blichert-Toft [2], reporting results between 1944 and 1969, found that patients who had operations within 12 hours had a mortality of 22% and within the next 12 hours, 36%. However, if treatment was delayed beyond 24 hours, the mortality rose to 64%. Primary repair produced a 34% mortality while nonoperative treatment resulted in an 80% rate of mortality. Although it is agreed in general that management of the recently ruptured thoracic esophagus is direct suture with drainage of the pleural space, there is disagreement as to the best method of managing late intrathoracic esophageal perforations [4, 5, 8-10, 13-15]. Most authors concur that tube drainage of the mediastinum and pleural cavity, administration of antibiotics, and adequate nutrition are important components of treatment of esophageal perforations with delayed recognition. However, there is a divergence of opinion con by The Society of Thoracic Surgeons

2 576 The Annals of Thoracic Surgery Vol 30 No 6 December 1980 ceming direct suture closure of the perforation [3] and the use of esophageal patches [4, 5, 9, 11,13,14], esophageal resections [7], and esophageal exclusion and diversion [6, 8, 151. This report describes the recent experience at the Toronto General Hospital with 8 consecutive patients who had nonmalignant intrathoracic esophageal perforations with late recognition. They were treated, when possible, by direct closure of the esophageal mucosa and obliteration of potential pleural spaces by good tube drainage, lung decortication, and the elective use of mechanical ventilation with positive end-expiratory pressure (PEEP) to maintain lung expansion. Materials and Methods The hospital records of all patients with a diagnosis of perforation of the intrathoracic esophagus treated at the Toronto General Hospital during the past six years (1973 through 1978) were reviewed. There were 15 patients in whom this diagnosis was established. In the 8 patients analyzed here, the perforations were recognized more than 48 hours (48 hours to 14 days) after rupture. This group consisted of 3 women and 5 men ranging from 34 to 80 years old. Instrumentation accounted for one perforation and postemetic rupture for seven perforations. Five of the 8 patients had an esophageal stricture, 4 as a result of reflux esophagitis and 1 as a result of a lye bum. The patients were seen initially with pain (8 patients), vomiting (7), fever (7), hypotension (4), hypoxemia (5), pleural effusion (7), pulmonary infiltrates (7), and leukocytosis (6). Two patients required preoperative mechanical ventilation. A preoperative diagnosis of esophageal perforation was made in 7 of the 8 patients by esophagogram (6 patients), esophagoscopy (5), and in some patients by both methods. Four of the eight perforations occurred in the midesophagus with the rest in the lower esophagus. All midesophageal ruptures were associated with distal esophageal strictures. All patients had a thoracotomy. The mucosa was closed in 7 patients and reinforced with a gastric patch in 3. The pleural space was drained in all patients. Pulmonary decortication was performed in 2 patients to fully reexpand the lung and obliterate the pleural space. Elective postoperative mechanical ventilation with PEEP was used in 4 patients. Direct Suture of the Perforation Four patients with spontaneous midesophageal ruptures underwent a right thoracotomy and direct suture closure of the perforation. The margins of the mucosa were defined, debrided, and accurately closed. Patient 1 A 69-year-old woman was admitted with a 2-day history of vomiting and retrostemal chest pain, which radiated into the back. She was febrile but normotensive. A contrast esophagogram revealed a small posterior wall perforation at the junction of the lower and middle thirds of the esophagus plus a distal esophageal stricture. Esophagoscopy revealed the perforation as well as a benign esophageal stricture. At thoracotomy, a 2 cm perforation was visualized 7 cm below the azygos vein. The esophagus was dilated with a No. 50 Maloney bougie. The mucosa around the perforation was debrided and closed with interrupted 4-0 chromic catgut sutures. The muscle was closed with interrupted 3-0 chromic catgut sutures. The pleural space was cleansed and drained with one chest tube. No complications occurred in the postoperative period. The patient was discharged 15 days after the operation and could eat a normal diet. She has no esophageal symptoms at present. Patient 2 A 68-year-old man who had epigastric pain, fever, and a right pneumothorax was admitted 6 days after a bout of severe vomiting. A right chest tube drained 2 liters of purulent fluid. A contrast esophagogram revealed a large, midesophageal perforation. After resuscitation, the patient underwent a right thoracotomy. A 4 cm tear in the esophagus was located 10 cm proximal to the gastroesophageal junction. The defect in the muscle layer was surgically extended longitudinally at each end of the perforation in order to visualize

3 577 Finley et al: Nonmalignant Esophageal Perforations the mucosa. The esophagus was closed in two layers and wrapped with pleura. The pleural cavity was decorticated, cleansed, and drained with two chest tubes. A leak from the suture line developed and persisted for two weeks. An esophagoscopy done at that time showed that a chest tube had eroded through the suture line into the esophageal lumen. The chest tube was shortened and since there was no distal esophageal obstruction, the fistula closed in a week. The patient required esophageal dilation for dysphagia on one occasion after the operation. At present, he has no esophageal symptoms. Pa tien t 3 An 80-year-old woman with a history of esophageal obstruction secondary to achalasia was admitted to the hospital with a 3-day history of food impaction and recurrent vomiting. Two pieces of meat were removed by endoscopy and a right posterolateral esophageal perforation could be visualized 30 cm from the teeth. The perforation into a mediastinal abscess cavity was confirmed by contrast esophagogram. After the patient was resuscitated, a large mediastinal abscess was visualized through a right thoracotomy. The abscess cavity was opened and traced to a 2 cm tear in the esophagus just below the azygos vein. The perforation was debrided and closed with interrupted 2-0 chromic catgut sutures. A long myotomy was carried out distally from the perforation to the gastroesophageal junction. The abscess cavity was debrided and drained with two chest tubes. A gastrostomy tube was brought out through a stab wound in the abdominal wall. There was no leak from the esophageal suture line. The patient was discharged from the hospital and could eat a solid diet; she had no dysphagia. Patient 4 A 79-year-old man with a history of reflux esophagitis was admitted with a 4-day history of vomiting and upper abdominal pain. He was hypotensive and dyspneic and had a right pleural effusion. A tube placed in the right chest drained 2 liters of purulent material. A contrast esophagogram showed a midesopha- geal perforation draining into the right chest. Esophagoscopy revealed gastric-lined mucosa below the perforation but no evidence of an esophageal stricture. After resuscitation, the patient underwent a right thoracotomy. A 3 cm tear in the esophagus was visualized 4 cm below the azygos vein. The esophagus was mobilized, and the defect in the muscle layer was surgically extended at either end to ensure adequate exposure of the mucosal margins. The mucosa was debrided and was closed with interrupted 3-0 chromic catgut sutures. The muscle was then approximated with interrupted 3-0 silk sutures. The pleural cavity was decorticated and drained with two chest tubes. The patient was placed on elective mechanical ventilation with PEEP for 2 days. Eight days after the operation, a small, asymptomatic leak at the suture line was documented radiographically. It had closed by 14 days after the operation. The patient was discharged 20 days after the operation. Esophageal dilation was carried out for dysphagia one month after the operation. The patient has no dysphagia at present, and the reflux esophagitis is being treated conservatively. COMMENT. The esophageal suture line leaked in 2 out of these 4 patients. The leaks were well controlled, and the patients showed no evidence of sepsis. In Patient 2, the chest tube eroded into the esophagus since it was placed too close to the suture line. Dilation of the distal esophageal stricture and simple suture closure of the perforation resulted in a low incidence of postoperative esophageal symptoms in these patients. Exclusion and Diversion of the Esophagus Patient 5 A 53-year-old woman with a lye stricture of the esophagus was admitted 2 days after undergoing esophagoscopy for removal of a food bolus stuck at the lower end of the esophagus. On admission she was dyspneic and febrile, and she had cervical crepitus and a large right pleural effusion. A right chest tube drained 1,200 ml of purulent fluid. A contrast esophagogram showed a large perforation into the right chest just above the gastroesophageal junction.

4 578 The Annals of Thoracic Surgery Vol 30 No 6 December 1980 After resuscitation, the patient underwent an esophagoscopy, which revealed a midesophageal stricture as well as a stricture just distal to the esophageal rupture. An 8 cm perforation was visualized at the lower end of the esophagus through a right thoracotomy. Since there was not sufficient esophageal tissue left to close the perforation, the proximal and distal esophagus were oversewn. The pleural cavity was debrided, cleansed, and drained. The esophagus was decompressed with a cervical esophagostomy and gastrostomy. In the postoperative period, the patient remained febrile and required a Clagett procedure for drainage of a right empyema. Thirty-five days after the operation, she was discharged home but had to have gastrostomy tube feedings. A left colon interposition was done three months after the esophageal perforation. As a result of the complications from the colon interposition, the patient required tube feedings for ten months after the initial operation. COMMENT. Although this operation may be necessary in poor-risk patients with large perforations, the magnitude of the operation and the duration of postoperative morbidity do not warrant its use in most patients with esophageal perforations. Direct Suture Reinforced with Gastric Patch Three patients with distal esophageal perforations had direct closure of the perforation and reinforcement with a partial [21 or complete wraparound of stomach. Patient 6 A 65-year-old man was admitted 3 days after undergoing a laparotomy for epigastric pain and vomiting. There were no abnormalities at the time of laparotomy. The patient s condition gradually deteriorated; respiratory failure, bilateral serous pleural effusions, renal failure, and hypotension developed, and he became obtunded. Fourteen days after laparotomy, material like coffee grounds was noted to be draining from the right chest tube and a contrast esophagogram showed a gross leak from the gastroesophageal junction into the left chest. After resuscitation, the patient underwent a left thoracotomy. The gastroesophageal junction was mobilized, and the muscle at each end of the perforation was extended longitudinally. The mucosa was debrided and closed with interrupted 3-0 chromic catgut sutures. The muscle was closed with interrupted 3-0 silk sutures. The perforation, which was approximately 2 cm from the gastroesophageal junction, was then partially wrapped with the piece of stomach brought up through the hiatus. A gastrostomy tube was left in place. The patient s postoperative course was complicated with episodes of atrial fibrillation, acute tubular necrosis, grand ma1 seizures, and a staphylococcal septicemia. However, he gradually improved and received nutritional support by gastrostomy feedings. The atrial fibrillation and seizures were controlled. His serum creatinine stabilized to 2 mg per 100 ml, and he was able to maintain a normal fluid and electrolyte balance. A contrast esophagogram done 11 days after the operation revealed no leak from the lower end of the esophagus. Twenty-one days after thoracotomy, a small bowel obstruction developed and required a laparotomy for lysis of adhesions. The patient recovered from this operation without any difficulty and was discharged home six weeks after admission. Symptoms of severe reflux esophagitis developed, which had not been present preoperatively. The patient required strict medical management with antacids, elevation of the head of his bed, and frequent small meals. He also required intermittent bougienage for dysphagia. Patient 7 A 66-year-old man was admitted with a 3-day history of retrosternal and epigastric pain. The pain started during an episode of vomiting, 6 hours after a suprapubic prostatectomy. The patient had also undergone an abdominal hiatus hernia repair for ulcerative esophagitis five years prior to this admission. On admission, he was hypotensive and dyspneic, and he had bilateral pleural effusions. Bilateral chest tubes drained purulent fluid. A contrast esophagogram revealed a perforation 2 cm above the diaphragmatic hiatus.

5 579 Finley et al: Nonmalignant Esophageal Perforations After the patient was resuscitated, the stomach was mobilized and wrapped 360 degrees around the site of the perforation in the chest. The pleural cavity was decorticated, cleansed, and drained with two chest tubes. The patient required mechanical ventilation for 4 days after the operation. There was no leak from the suture line in the postoperative period. The patient was discharged 30 days after the operation. He could take a normal diet but experienced occasional dysphagia after eating dry bread. Three fears after the repair of the esophageal perforation, reflux esophagitis and a bleeding gastric ulcer in the fundoplication developed. The gastric ulcer was oversewn, and a transthoracic gastroplasty and hiatus hernia repair were done. In the postoperative period, a recurrent gastric hemorrhage developed and required a total gastrectomy and Roux-en-Y repair. There was a leak from the upper anastomosis, and a left empyema developed. The patient died in septic shock two weeks after the last operation. Patient 8 A 34-year-old man was admitted with a 2-day history of epigastric pain following a bout of vomiting. On admission he was dyspneic and febrile and had a left pleural effusion. A left chest tube drained 1 liter of bile-stained fluid. A contrast esophagogram revealed a perforation 3 cm above the gastroesophageal junction. Through a left thoracotomy, the esophageal perforation was debrided and closed in two layers. The stomach was mobilized and partially wrapped around the perforation in the chest. The patient received elective mechanical ventilation for 48 hours. There were no leaks from the esophageal suture line in the postoperative period. The patient required a gastroplasty and hiatus hernia repair for severe reflux esophagitis eighteen months after the operation. COMMENT. Although the use of intrathoracic gastric patches to reinforce the esophageal suture line prevented esophageal leakage in this study, the enlargement of the hiatus, the intrathoracic positioning of the stomach, and the failure to do a standard antireflux operation re- sulted in a high incidence of reflux esophagitis. Comment Despite awareness of the seriousness of intrathoracic esophageal ruptures, delay in diagnosis is the major stumbling block to lowering morbidity and mortality. In this study, 6 of the 8 patients saw their primary physician within 24 hours of experiencing symptoms of esophageal rupture. Initial diagnoses included myocardial infarction, pulmonary embolus, spontaneous pneumothorax, and perforated duodenal ulcer. Rupture of the esophagus should always be considered in patients with severe chest or upper abdominal pain and signs of infection. Early in the course of an esophageal perforation, the diagnosis may only be suspected from a characteristic history. However, once mediastinitis becomes established, chest or abdominal pain, dyspnea, fever, hydropneumothorax, or mediastinal emphysema are highly suggestive of an esophageal rupture. These symptoms and signs demand that a Hypaque (diatrizoate meglumine) swallow be performed. Esophagoscopy helps confirm the location of the perforation and establish the caliber of the distal esophagus. Patients who are seen late after esophageal rupture are dyspneic, dehydrated, and very often hypotensive. After fluid resuscitation, antibiotics, and tube drainage of the infected pleural spaces, the patient s condition improves and there is reluctance to repair the esophageal rupture by operation. However, if an operation is not done, soilage of the mediastinum and pleural cavity progresses to severe empyema and the patient dies of the complications of uncontrolled sepsis. The well-documented high mortality associated with nonoperative management of old esophageal ruptures has led to more aggressive surgical management in most centers [21. As time goes by after the rupture, the esophageal muscle around the perforation becomes inflamed and necrotic. The inability of this tissue to hold sutures was recognized in the past, and recurrent leakage often has been reported with direct suture closure of old esophageal

6 580 The Annals of Thoracic Surgery Vol 30 No 6 December 1980 tears. Consequently, complex procedures have been designed for the management of patients with esophageal perforations. Use of Silastic T-tubes or the Celestin tube [l] to exteriorize or put a stent on the perforation has met with favorable results in selected patients. Resection of the rupture and esophagogastrostomy may be indicated for perforated neoplasms [7]. However, the risk of such an extensive operation, if not done for cancer, is prohibitive. Johnson and co-workers [61 recommended complete exclusion of the esophageal perforation by suture closure of the esophagus at the cardia and cervical esophagostomy with closure of the distal end. This technique was modified by Menguy [81 and later Urschel and colleagues [151 who placed a tape around the cardia to prevent reflux and performed a loop cervical esophagostomy to divert the oropharyngeal secretions. These techniques may be useful for very large esophageal perforations that cannot be closed directly. The 1 patient in this study who was treated by this technique required a rib resection and drainage for uncontrolled intrapleural sepsis. She underwent a left colon interposition three months after the operation but required gastrostomy feedings for ten months after the rupture. In patients with delayed recognition of nonmalignant intrathoracic esophageal perforation, elimination of continued chemical and bacterial contamination can be achieved by primary closure of the esophageal perforation. Although the esophageal muscle is often inflamed and necrotic, the mucosa usually remains viable and can hold sutures securely. The mucosal tear is often longer than the muscle tear so it is important to mobilize the esophagus and to extend the muscle tear at each end of the perforation in order to define, debride, and close the mucosa accurately with interrupted sutures. The inner aspect of the esophagus should be carefully inspected before closure of the rupture in order to rule out the presence of cancer, stricture, or another tear. The muscle layer should be debrided and closed with interrupted sutures. Various tissues have been used to reinforce the esophageal suture line and thereby avoid leaks. Thal and Hatafuku [141 used the gastric fundus to reinforce the closure or to bridge the defect in the esophagus. Although the gastric patch successfully prevented esophageal leakage in 3 patients in this study, neither partial nor complete gastric wraparound of the intrathoracic esophagus prevented gastric reflux. In these patients, enlargement of the hiatus, intrathoracic positioning of the cardia, and inability to do a standard antireflux procedure may have contributed to the reflux. Two of the 3 patients required a subsequent antireflux operation for severe esophagitis. Pericardial[9] and diaphragmatic tissue [13] also have been used. The benefits of operations using these tissues are offset by the inevitable contamination of the pericardial and peritoneal cavities. The drawbacks of the preceding techniques make the use of thickened parietal pleura [51, lung [lll, or intercostal muscle [41 for support of the esophageal suture line the most appealing techniques since healthy local tissue is used without added risk to the patient. Potential intrapleural dead space is eliminated by adequate pleural drainage, pulmonary decortication, and the elective use of mechanical ventilation with PEEP in order to maintain full expansion of the lung. Finally, the use of total parenteral nutrition for maintenance of the protein and caloric needs of these catabolic patients is mandatory for increased survival [ 101. References 1. Berger RL, Donato AT: Treatment of esophageal disruption by intubation: a new method of management. Ann Thorac Surg 13:27, Blichert-Toft M: Spontaneous oesophageal rupture: an evaluation of the results of treatment of Scand J Thorac Cardiovasc Surg 5:111, Bradham RR, Bridgman AH, Scott SM, Betts RH: Spontaneous esophageal perforation: management of the "intermediate" phase. Ann Thorac Surg 3:6, Dooling JA, Zick HR: Closure of an esophagopleural fistula using onlay intercostal pedicle graft. Ann Thorac Surg 3:553, Grillo HC, Wilkins EW Jr: Esophageal repair following late diagnosis of intrathoracic perforation. Ann Thorac Surg 20:387, Johnson J, Schwegman CW, Kirby CK: Esophageal exclusion for persistent fistula following

7 551 Finley et al: Nonmalignant Esophageal Perforations spontaneous rupture of the esophagus. J Thorac Surg , Kerr WF: Emergency oesophagectomy. Thorax 23:204, Menguy R: Near-total esophageal exclusion by cervical esophagostomy and tube gastrostomy in the management of massive esophageal perforation: report of a case. Ann Surg 173:613, Millard AH: Spontaneous perforation of the oesophagus treated by utilization of a pericardial flap. Br J Surg 58:70, Miller HAB, Taylor GA: Management of late cases of esophageal disruption with intravenous hyperalimentation. Can J Surg 18:41, Moore TC, Goldstein J, Teramoto S: Use of intact lung for closure of full-thickness esophageal defects. J Thorac Cardiovasc Surg 41:336, Postlethwait RW: Surgery of the Esophagus. New York, Appleton-Century-Crofts, Rao KVS, Mir M, Cogbill CL: Management of perforations of the thoracic esophagus. Am J Surg 127:609, Thal AP, Hatafuku T: Improved operation for esophageal rupture. JAMA 188:826, Urschel HC Jr, Razzuk MA, Wood RE, et al: Improved management of esophageal perforation: exclusion and diversion in continuity. Ann Surg 179:587, 1974 Discussion DR. P. RICHARD CARTER (West Covina, CA): I compliment Dr. Finley and his distinguished colleagues for an excellent presentation-a reappraisal of primary closure for delayed, nonmalignant esophageal perforation. The most important factor contributing to the continuing high mortality of this devastating lesion is failure to make an early diagnosis. If recognition is prompt, successful treatment can be carried out. It consists of intensive supportive measures, direct transthoracic repair by primary suture, and adequate drainage of the mediastinal and pleural spaces. When recognition is delayed, however, perforating esophageal wounds produce a fulminant mediastinitis, which severely damages the esophagus and often makes it either unsuturable or precariously reparable. Suture line leakage with recurrent perforation often follows, resulting in prolonged morbidity or death of the patient. Not only does delay in diagnosis markedly increase the death rate, but direct operative repair is much more likely to be successful if done as soon as possible after injury. Thus, greater emphasis on earlier diagnostic studies is important in a catastrophic situation in which survival is directly related to prompt surgical intervention. Any patient with an unexplained pleural effusion should have an esophagogram, but a negative study by no means excludes the possibility of esophageal rupture. Direct endoscopic visualization of the perforation should then be done, as was used by the Toronto group. In addition, endoscopic stricture dilation under fluoroscopy, combined with a simultaneous contrast esophagogram, facilitates immediate recognition of instrumental perforation. Multiple, more complex techniques-many authored by members of this Society-are usually necessary after delayed diagnosis. Buttressing a precarious primary closure is advisable, and coexisting distal obstruction, if present, must be corrected. Many techniques of reinforced patch closures are advocated, but the use of parietal pleura or gastric wraparound is especially helpful. Although based on a small but well-managed series of patients, this study suggests that delayed esophageal perforation may be closed primarily in selected patients, when reinforced by patching techniques. The key to proper surgical treatment is the condition of the esophagus at the time of operation. In closing, I ask the authors to define their indications for esophageal exclusion and diversion, and to comment on their use of PEEP in these patients. DR. J. B. SHAMMASH (Springfield, MA): I endorse early operative management of benign perforations, preferably by primary suture, even when diagnosis may be delayed a number of days. Two patients were treated in this manner. One was seen 65 hours after perforation and the other, 10 days after perforation when the thoracostomy tube, inserted to drain an empyema, did not control his septic course. The perforation in the distal esophagus led to a large collection in the mediastinum, which poured out through a small opening in the mediastinal pleura to form a right-sided empyema. A 3 cm oval perforation was closed, the mediastinum widely drained, and the pleura decorticated. The patient did well. A more rare and less understood type of limited tear deserves mention-the so-called spontaneous intramural hematoma, which does well without operative intervention. A 65-year-old woman noted sudden substernal pain radiating to the interscapular area associated with hematemesis after she ate a piece of toast. She had no previous dysphagia, and she was not on a regimen of anticoagulants. Chest roentgenogram on admission was normal. A few hours later, bilateral pleural effusions developed. A barium swallow revealed extravasation of barium intramurally throughout the length of the thoracic esophagus, markedly narrowing the lumen. Esophagoscopy showed submucosal hemorrhage obstructing the lumen. Serous fluid, which was positive for salivary amylase, drained from a left thoracostomy. The patient did well on conservative management and was drinking in a week. Of particular note, there was a lack of "systemic" or marked febrile response that is usually seen with frank mediastinal-pleural perforation of the esophagus.

8 582 The Annals of Thoracic Surgery Vol 30 No 6 December 1980 Dr. Finley, have you encountered similar cases of this entity? If so, how were they handled? DR. ANTHONY s. PATTON (Peabody, MA): Thank you very much for the opportunity to discuss this most interesting paper. I think it is a very important subject to most of us in thoracic surgery. We reviewed our experience from Salem Hospital at a meeting in New England in At that time, we had seen 14 patients with spontaneous perforations of the esophagus, and since then have added 4. Nine of these eighteen perforations were the very late variety discussed by Dr. Finley. The main point of our presentation was that these patients often are not seen with the typical Boerhaave s spdrome but often have most peculiar symptoms, usually in association with other gastrointestinal diseases. We had patients who were seen with unusual problems, such as pericardial tamponade, due to an unrecognized spontaneous esophageal perforation that had occurred up to a month earlier. I emphasize that esophageal perforation must be considered in the diagnosis of unusual thoracic problems. I take minor issue with Dr. Finley s group for negating, I think, the concept of esophageal exclusion. We have found this concept to be a good one. I do not mean that we do tapings or primary resections for benign lesions, as some authors have suggested, but we do simple drainage from above with large Levine tubes and from below with a gastrostomy. A jejunostomy for feeding is usually added. A primary attack on the lesion, if it is feasible, is also done, and extensive mediastinal drainage is carried out. We have found this method of esophageal exclusion very successful and have had few mortalities and troubles when these steps are used with vigor. DR. HERMES c. GRILLO (Boston, MA): The principal point of contention in handling perforations recognized late has been whether or not to exclude the esophagus, because of the problem of releakage. I want to update the report that Dr. Wilkins and I [l] made a few years ago about using one of the several tissue patches, a pedicled pleural flap, to avoid this problem. Everyone seems to agree on the other principles-decortication, expansion of lungs, drainage, and elimination of distal obstruction. We now have treated 9 of these patients. All have done well, except for the last 1 who had a pseudodiverticulum, which healed spontaneously. In patients with perforations, the pleura is very thickened due to inflammation. It is easily accessible either for flap coverage or circumferential wrapping. I agree with the argument of Dr. Finley that anatomical exclusion is unnecessary. Reference 1. Grillo HC, Wilkins EW Jr: Esophageal repair following late diagnosis of intrathoracic perforation. Ann Thorac Surg 20:387, 1975 DR. PETER c. PAIROLERO (Rochester MN): I, too, congratulate the authors for their excellent presentation. We at the Mayo Clinic are in agreement with many of the points that they have stressed. However, in addition to the elimination of both bacterial and chemical contamination, an additional goal of treatment must be to eliminate any distal esophageal obstruction. This is especially true in those patients with instrumental perforations secondary to stricture. As already stated, the ideal treatment time is immediately after perforation before sepsis has deveioped. However, once the perforation becomes chronic and the patient has survived the hazards of septic shock, treatment at this late stage also has a good chance of success. We recently reviewed our series of 11 patients with spontaneous rupture of the esophagus seen during the past ten years. Two patients were treated early, that is, less than 24 hours after perforation. Both were treated with thoracotomy, esophageal closure, and pleural drainage, and both survived. An additional 5 patients were treated later than 5 days following perforation, and 4 survived. Two of these patients were treated by thoracotomy, closure, and drainage; 2 patients by chest drainage only; and 1 by esophageal exclusion. The single death in this group occurred in 1 of the patients treated by chest drainage only. On the other hand, of the 4 patients treated in the intermediate period, that is, between 1 to 3 days following perforation, only 1 patient survived. Three of these patients were treated by thoracotomy, closure, and drainage, with 1 survivor, and 1 by esophageal exclusion. Thus, it may be not so much what we do but rather the timing of the operation that is all important, provided the goals of treatment are met. Dr. Finley, what were your results in those patients in this intermediate zone? Also, what are your recommendations if thoracotomy, closure, and drainage fail? DR. KAMAL A. MANSOUR (Atlanta, GA): I compliment the authors for their magnificent results. However, since their patients survived the acute episode of rupture, I wonder how many would have made it without any surgical intervention. In our series of 47 patients with spontaneous rupture of the esophagus reported some years ago, 3 survivors were seen by us initially for management of residual complications late in their disease. Two patients had leakage into the subphrenic space and 1 was seen with an esophagopleural cutaneous fistula some six months after the onset of the acute episode. This paper supports the fact that primary repair is the treatment of choice whenever possible. However, it does not exclude other modalities of treatment such as T-tube drainage, esophageal exclusion, and even esophageal resection in late cases of esophageal rupture. What about patients with stricture of the esophagus distal to the perforation? We have seen 7 such patients. And how about patients with massive

9 583 Finley et al: Nonmalignant Esophageal Perforations inflammation and necrosis of the esophageal wall, or patients with severe malnutrition and alcoholism in whom healing capacity is very poor? We agree with the authors on the importance of postoperative mechanical ventilation to minimize pleural dead space and overcome ventilatory deficiency. I have one question for the authors. How would you explain the high incidence of postoperative dysphagia-6 out of your 8 patients? DR. LYMAN A. BREWER, 111 (Loma Linda, CA): The Egyptians described an esophageal wound in the Smith Papyrus, BC. One case titled "Instruction concerning a wound of the throat" states: "If thou examines a man having a gaping in his throat, piercing through his gullet; if he drinks water he chokes, it comes out of the mouth of the wound; it is greatly inflamed so that he develops fever from it, thou should draw together that wound with stitching." This is the first recorded use of sutures. In World War 11, 4,444 years later, we began to repair esophageal wounds; no repairs were reported in World War I. In 1944, a casualty with a wound of the esophagus and empyema 7 days after wounding showed esophageal contrast medium entering the pleural cavity. After esophagus repair and pulmonary decortication, the right lung completely expanded. Pulmonary decortication was revived in World War I1 by Drs. Burford, Samson, and myself. We successfully treated 3 other patients with esopha- geal wounds with delayed suture. This technique, first used by the ancient Egyptians BC, was modernized and repoxted in volume two of Surge y in World War 11, Thoracic Surge y (pp ) by Dr. Burford and myself. DR. PINLEY: I thank all the discussants for their comments. I'm sorry that I missed Dr. Brewer's first reference in my literature search. A number of good points have been raised concerning this very controversial subject, and I will talk first about esophageal diversion. In most instances we try to close the esophageal wall where possible unless we are going to completely obstruct the esophagus. We try to divert or decompress the esophagus by using a sump nasogasttic tube above and a gastrostomy tube below the perforation. This technique appears to protect against mediastinal contamination as well as cervical esophagostomy and ligation of the gastroesophageal junction. Most of our patients were diagnosed 2 to 5 days after perforation but a few were seen late and did very well. We used the pleural patch, as described by Dr. Grillo, with success. Six of our 8 patients had postoperative dysphagia. Four improved with dilation early in the postoperative period. Two patients had prolonged dysphagia. One had stenosis at the esophagocolonic anastomosis, and the second had a column-lined esophagus with recurrent peptic stricture of the esophagus.