- Email: [email protected]
Esophagogastrectomy via left thoracophrenotomy
Esophagogastrectomy Via Left Thoracophrenotomy Richard D. Page, FRCS(E), Joseph F. Khalil, FRCS(E), Richard I. Whyte, MD, David K. Kaplan, FRCS(E), and Raymund J. Donnelly, FRCS(E) Regional Adult Cardiothoradc Unit, Broadgreen Hospital, Liverpool, England
Esophagogastrectomy is generally considered to be the treatment of choice for resectable tumors of the esophagus. Although many approaches and techniques have been advocated, since April 1983 we have used a left thoracophrenotomy approach for most lesions of the lower two thirds of the esophagus and gastric cardia. Stapling instruments have been used for mobilization of the stomach and fashioning of the esophagogastric anastomosis. One-hundred fifteen patients undergoing resection of malignant tumors with this technique were retrospectively reviewed. Perioperative mortality was 8.7%
(lOhl5). The rate of anastomotic leakage was 1.7% (2/ 115), and benign narrowing of the anastomosis requiring dilation developed in 16 patients. The rate of recurrent anastomotic tumor was 4.3%. Eighteen patients had complications, and the mean postoperative hospital stay was 13 days. Survival at 3 years was 22.1%. During the period of study, 22 patients underwent esophageal resection by some other approach; the reasons for this are described. The advantages of the left thoracophrenotomy approach are discussed. (Ann Thorac Surg 1990;49:763-6)
E
other approach during this period. The various reasons for this are shown in Table 1. Of the 115 patients who had a left thoracophrenotomy, 80 were men (69.5%)and 35 were women (30.5%).Their mean age was 63.2 years (k10.5 years; range, 33 to 84 years). In addition to a full clinical history and examination, preoperative assessment consisted of barium swallow, biochemical and hematological screening, esophagoscopy, and, in the case of midthoracic lesions, bronchoscopy. Further investigations for evaluation of local or distant metastatic disease were performed only when indicated either clinically or from information derived from the routine preoperative assessment. Histological cell types were as follows: adenocarcinoma, 60.8%; squamous, 35.7%; undifferentiated, 1.7%; small cell, 0.9%; and malignant lymphoma, 0.9%. Eightytwo tumors (71.3%)were located in the lower third of the esophagus and cardia, and 33 (28.7%)were in the middle third. No tumors in the upper third were resected by left thoracotomy. Staging was determined by histopathological examination of the surgical specimen. A double-lumen endotracheal tube is used for anesthesia so that the left lung can be collapsed during operation. A left thoracotomy is made through the bed of the seventh rib. The skin incision extends forward almost to the costal margin, but the anterior costal margin itself is left intact. The esophagus is mobilized from the diaphragmatic hiatus to a point at least 5 cm above the upper level of the tumor. When we attempt a curative resection, we extend mobilization behind and above the aortic arch, taking care to preserve the recurrent laryngeal nerve. If necessary, the arch is mobilized by dividing one or two intercostal branches. We have not found it necessary to resect the seventh rib. The diaphragm is opened in line with the wound, leaving a 1-inch cuff anteriorly to facilitate closure. The stomach is mobilized using the LDS stapler (Auto-
sophageal resection is currently accepted as the best form of treatment for operable tumors affecting the middle and lower thirds of the esophagus. Not only does it provide the best chance of curing the disease, but it is also the most effective method of palliation. The manner in which the operation is approached is controversial, however. Several different incisions, often in combination, are widely used. These include left thoracoabdominal [l]; laparotomy and right thoracotomy [2,3]; three-stage resection comprising laparotomy, right thoracotomy, and a cervical incision [4]; the transhiatal approach [5, 61; bilateral thoracotomies [7]; and left thoracophrenotomy, first described by Churchill and Sweet [8]. The left thoracophrenotomy approach to tumors of the middle and lower thirds of the esophagus and cardia of the stomach is still used by thoracic surgeons but has attracted little attention in the literature in recent years. We believe it offers advantages over other approaches and, since May 1983, have adopted it for routine resections of all esophageal tumors. This report describes our experience with this approach.
Material and Methods The case records of all patients undergoing esophagogastrectomy under the care of one surgeon (R.J.D.) between April 1983 and February 1989 were retrospectively reviewed. During this period, a left thoracophrenotomy incision has been the preferred approach for this procedure and has been used in 115 patients. This group constitutes the major part of this report. Twenty-two patients have undergone esophageal resection by some Accepted for publication Dec 26, 1989. Address reprint requests to Mr Donnelly, Broadgreen Hospital, Thomas Dr, Liverpool, England L14 3LB. Dr Whyte’s present address is Department of Cardiac Surgery, Massachusetts General Hospital, Boston, MA.
0 1990 by
The Society of Thoracic Surgeons
0003-4975/90/$3.50
764
FAGEETAL ESOPHAGOGASTRECTOMY
Ann Thorac Surg 1990;49:76=
Table 1. Incisions Other Than Left Thouacophrenotomy, 1983-1 989
Incision and Reason Left thoracoabdominal Tumor clearance
Previous operation Poor exposure Previous benign perforation Acute perforation Unspecified Total Cervical plus left thoracophrenotomy Small proximal thoracic esophagus Proximal tumor Total
n 5 2 2 1 1 2 13
3 2 5
Right thoracotomy plus laparotomy Tumor at level of archileft main bronchus Previous left thoracic operation Suspicious nodule in right lung Total Total
2 1 1 4 22
suture, UK) around the greater curvature except for the short gastric vessels, for which the Surgiclip instrument is used. After division of the short gastric vessels, the gastrocolic omentum, the left gastric artery, and the lesser omentum, the stomach is delivered into the left hemithorax through the hiatus. It is transected using a linear stapler at a site at least 5 cm distal to the lower margin of the tumor. The esophagus is divided proximally, and the tumor-bearing segment is removed. More recently, however, we have temporarily left the posterior wall intact to facilitate introduction of the detachable anvil of the endto-end anastomotic (EEA) stapling instrument, after which the remaining part of the esophagus is divided. Esophagogastric continuity is normally restored using the EEA stapler (Autosuture) as previously described [9, lo]. The 25-mm head was used in 34 patients, the 28-mm head was used in 64, and the 31-mm head was used in 15. In 3 patients, initial sizing with the 28-mm head led to mucosal tearing of the proximal esophagus and it was necessary to use the 25-mm head to perform the anastomosis more proximally. In 2 patients, a hand-sewn anastomosis was necessary because the proximal esophagus was too small for the 25-mm head to be inserted. Eighty anastomoses (69.5%) were constructed below and 35 (30.5%) were constructed above the aortic arch. In this latter group, the stomach is routed lateral to the arch. After placement of a nasogastric tube, the anterior gastrotomy through which the EEA stapler was inserted is closed with a linear stapler. All straight staple lines are oversewn with a continuous 3-0 Mersilene suture. Neither fundoplication nor pyloroplasty are routinely used. Patients are generally extubated in the operating room and returned directly to the general ward. After operation the nasogastric tube is left open to air until the drainage is minimal, at which time it is removed,
generally on the second postoperative day. Crystalloids are administered through a peripheral vein for fluid maintenence, and patients are allowed to drink water on the fourth day. Oral intake is advanced to soft solids over the next three days. Contrast radiographic studies are obtained only if there is clinical evidence to suspect an anastomotic leak or if there was difficulty performing the anastomosis at the time of operation. After discharge from the hospital, patients are seen at 3-month intervals and assessed for signs of tumor recurrence and anastomotic stricture formation. Patients with dysphagia undergo esophagoscopy and dilation if necessary. Follow-up to May 1989 was complete in 99.1% of patients. Radiotherapy and chemotherapy were not used preoperatively in this series and postoperatively were used only for palliation of recurrent or residual disease. Survival was calculated according to the method described by Grunkemeier and Starr [ll].
Results The TNM status of the tumors as deduced from histological examination of the resected specimens showed that 37 tumors were stage 1, 7 were stage 2A, 70 were stage 3, and 1 was stage 4. The gastric and esophageal resection margins were free of tumor in 108 (93.9%) of the specimens. Operative mortality, defined as deaths within 30 days or during initial hospital stay, was 8.7% (lOD15). The cause of death was anastomotic leakage in 1 patient, pneumonia in 4, pulmonary embolism in 2, chylothorax in 1, septicemia in 1, and myocardial infarction in 1. Eighteen patients (15.6%) had postoperative complications, although these did not delay discharge from the hospital for 8 patients. Complications that could be attributed to operative technique included one anastomotic leak, one gastrotomy leak, and one chylothorax. Patients were discharged from the hospital at an average of 13.1 days (*10.5; range, six to 99 days) after operation. Anastomotic leakage developed in 2 patients (1.7%), 1 of whom died as a result. The leakage was successfully repaired in the other patient three days after operation. Narrowing of the anastomosis sufficient to cause dysphagia and relieved by dilation developed in 16 patients. In 13 of these, a single dilation was sufficient to restore normal swallowing. There was no statistically significant difference between the rates of benign stricture formation of the three sizes of anastomotic stapler. Malignant strictures at the anastomosis occurred in 5 patients (4.3%). Survival is shown in Figure 1 and was 57.0% at 1 year, 32.3% at 2 years, and 22.1% at 3 years.
Comment Controversy persists regarding the optimal surgical approach for esophagogastrectomy, and a variety of techniques have been advocated. These include left thoracophrenotomy [8, 12-14], right thoracotomy and laparotomy [2-4, 151, left thoracoabdominal [ l , 161, left thoracotomy and midline laparotomy [ 171, bilateral thora-
PAGEETAL ESOPHAGOGASTRECTOMY
Ann Thorac Surg 1990;49:7634
XswIVMNo
3 YEARS
Fig 1. Actuarial 3-year survival of 115 patients undergoing esophagogasfrectomyfor malignant disease via left fhoracophrenofomy.
cotomies [7], and laparotomy and cervical incision without thoracotomy [5, 6, 181. A neck incision with cervical anastomosis may be added to any of these approaches. None of these various methods have made any significant difference in long-term survival, although some investigators consider subtotal esophagectomy to be important in this regard. Alternative techniques have generally been recommended and found favor among different surgeons for their ease of performance of a difficult operation and the consequent reduction in morbidity and mortality that should result. The left thoracophrenotomy approach is historically important as the method advocated by Sweet and others in the pioneering days of esophageal surgery. Although this approach is still practiced by some thoracic surgeons on both sides of the Atlantic, little appears in the literature about this technique. We believe that it offers several advantages over other methods for tumors of the middle third of the esophagus and below; for this reason, we report our experience with the technique. If the thoracotomy incision is made at the correct level through the sixth interspace, good access is available to the entire intrathoracic and abdominal esophagus, including the portion behind the aortic arch where dissection can be facilitated if necessary by mobilization of the aorta. Care must obviously be taken with the recurrent laryngeal nerve, and the course of the thoracic duct must be borne in mind. The supraaortic esophagus can be visualized and mobilized by dividing the pleura lateral to the left subclavian artery, and esophagogastric continuity can be reestablished by anastomosis at the apex of the thorax. We have not found it necessary to resect the seventh rib [14], but very rarely we make a second incision through the fourth interspace. This has been necessary only in patients requiring a thoracoabdominal incision. The left thoracophrenotomy incision offers adequate exposure of the stomach and excellent access to the short and left gastric arteries through the opening in the left hemidiaphragm. Even performance of pyloroplasty is often possible, although we do not perform it routinely. Opening and closing of the wound is rapid and simple and, although we have not attempted to quantify them, it is not unreasonable to assume potential benefits in terms
765
of reduced operating time and a smaller incision with consequent greater patient comfort and reduced wound morbidity. Stapling instruments have proved to be very useful in performing esophagogastrectomy through the left side of the chest. Not only do they facilitate mobilization of the stomach through the diaphragm, but use of the EEA stapler allows rapid and reliable creation of the anastomosis [9, 10, 19-21] at the apex of the left thorax, so that it is nearly cervical in location. Proponents of the cervical anastomosis argue that, should a leak occur, extravasation is confined to the neck, causing less morbidity than an intrapleural leak. Our rate of anastomotic leakage is low, and we do not believe that there is sufficient justification for a separate neck incision with its consequent cosmetic effects and prolongation of operating time. We have not found the left thoracophrenotomy approach to be associated with a significantly higher incidence of postoperative complications as compared with other approaches [22]. The rate of stricture formation was similar to that previously reported 1231. Some of the more frequently used alternatives to the left thoracophrenotomy approach are the right thoracotomy plus laparotomy or the left thoracoabdominal incision, either of which can be combined with a cervical incision, and the transhiatal esophagectomy. Arguments in favor of the right thoracotomy plus laparotomy approach are that a more complete tumor resection can be obtained and that a higher anastomosis can be created. We believe that an equally complete resection is possible through a left thoracotomy and that, particularly with use of the EEA stapler, an anastomosis at the apex of the chest is relatively straightforward. The left thoracoabdominal approach, as used by us before 1983, provides little added exposure over a left thoracophrenotomy and poses potential problems associated with division of the costal margin and anterior abdominal wall musculature. The transhiatal approach, as advocated by Orringer and associates [5, 61, has been criticized for violating the basic surgical principles of adequate exposure and hemostasis and for an excessive mortality and morbidity. Although Orringer and associates have argued these points well, Shahian and colleagues concluded [24], and we agree, that this approach should be viewed as an addition to the surgical armamentarium rather than a replacement of standard techniques. During the course of this series, we have occasionally elected to use one of the alternative approaches for the reasons listed in Table 1. In addition, we have not hesitated during the course of operation either to extend the incision to the abdominal wall because of difficultiesof access or tumor bulk or to perform the anastomosis in the neck if the proximal esophagus is too small or involved by tumor. Our results with esophagogastrectomy by left thoracophrenotomy are comparable with those of published series using alternative methods. We believe that this technique has advantages of its own and is worthy of its place in modern thoracic surgery.
766
PAGEETAL ESOFHAGOGASTRECTOMY
References 1. Heitmiller RF. The left thoracoabdominal incision. Ann Thorac Surg 1988;46:25&3. 2. Lewis I. The surgical treatment of carcinoma of the esophagus with special reference to a new operation for growths of the middle third. Br J Surg 1946;34:18. 3. Tanner NC. The present position of carcinoma of the esophagus. Postgrad Med J 1947;23:109. 4. McKeown KC. Trends in esophageal resection for carcinoma with special reference to total oesophagectomy. Ann R Coll Surg Engl 1972;51:21349. 5. Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978;7664&54. 6. Orringer MB, Orringer J. Esophagectomy without thoracotomy: a dangerous operation? J Thorac Cardiovasc Surg 1983;85:72-80. 7. Dark JF, Mousalli H, Vaughan R. Surgical treatment of carcinoma of the esophagus. Thorax 1981;36:891-5. 8. Churchill ED, Sweet RH. Transthoracic resection of tumours of the stomach and esophagus. Ann Surg 1942;115:897-917. 9. Fabri BM, Donnelly RJ. Esophagogastrectomyusing the end to end anastomotic stapler. Thorax 1982;37296-9. 10. Donnelly RJ, Sastry MR, Wright CD. Esophagogastrectomy using the end to end anastomotic stapler: results of the last 100 patients. Thorax 1985;4095&9. 11. Grunkemeier GL, Starr A. Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 1977;24 404-8. 12. Lam KH, Cheung HC, Wong J, Ong GB. The present state of surgical treatment of carcinoma of the esophagus. J R Coll Surg Edinburgh 1982;27315-26. 13. Adkins PC. Tumours of the esophagus. In: Sabiston DC, ed.
Ann Thorac Surg 1990:497634
Davis-Christopher textbook of surgery. 12th ed. Philadelphia: WB Saunders, 1981:84&55. 14. Pradhan GN, Eng J, Sabanathan S. Left thoracotomy approach for resection of carcinoma of the esophagus. Surg Gynecol Obstet 1989;168:49-53. 15. Mitchell RL. Abdominal and right thoracotomy approach as standard procedure for esophagogastrectomy with low morbidity. J Thorac Cardiovasc Surg 1987;93:20.511. 16. Wong J. Stapled esophagogastric anastomosis in the apex of the right chest after subtotal esophagectomy for carcinoma. Surg Gynecol Obstet 1987;164:569-72. 17. West PN, Marbarger JP, Martz MN, Roper CL. Esophagogastrectomy with the EEA stapler. Ann Surg 1981;193:76-81. 18. Baker JW, Schechter GL. Management of panesophageal cancer by blunt resection without thoracotomy and reconstruction with stomach. Ann Surg 1986;203:491-9. 19. Steichen FM, Ravitch MM. Mechanical sutures in esophageal surgery. Ann Surg 1980;191:37?-91. 20. Hopkins RA, Alexander JC, Postlethwaite RW. Stapled esophagogastric anastomosis. Am J Surg 1984;147283-7. 21. Sugimachi K, Ikeda M, Ueo H, Kai H, Okudaira Y, Inokuchi K. Clinical efficacy of the stapled anastomosis on esophageal reconstruction. Ann Thorac Surg 1981;33:374-8. 22. Muehrcke DD, Donnelly RJ. Complications after esophagogastrectomy using stapling instruments. Ann Thorac Surg 1989;48:257-62. 23. Muehrcke DD, Kaplan DK, Donnelly RJ. Anastornotic narrowing after esophagogastrectomy with the EEA stapling device. J Thorac Cardiovasc Surg 1989;97434-8. 24. Shahian DM, Neptune WB, Ellis FH. Transthoracic versus extrathoracic esophagectomy: mortality, morbidity, and long-term survival. Ann Thorac Surg 1986;41:23746.
Esophagogastrectomy is generally considered to be the treatment of choice for resectable tumors of the esophagus. Although many approaches and techniques have been advocated, since April 1983 we have used a left thoracophrenotomy approach for most lesions of the lower two thirds of the esophagus and gastric cardia. Stapling instruments have been used for mobilization of the stomach and fashioning of the esophagogastric anastomosis. One-hundred fifteen patients undergoing resection of malignant tumors with this technique were retrospectively reviewed. Perioperative mortality was 8.7%
(lOhl5). The rate of anastomotic leakage was 1.7% (2/ 115), and benign narrowing of the anastomosis requiring dilation developed in 16 patients. The rate of recurrent anastomotic tumor was 4.3%. Eighteen patients had complications, and the mean postoperative hospital stay was 13 days. Survival at 3 years was 22.1%. During the period of study, 22 patients underwent esophageal resection by some other approach; the reasons for this are described. The advantages of the left thoracophrenotomy approach are discussed. (Ann Thorac Surg 1990;49:763-6)
E
other approach during this period. The various reasons for this are shown in Table 1. Of the 115 patients who had a left thoracophrenotomy, 80 were men (69.5%)and 35 were women (30.5%).Their mean age was 63.2 years (k10.5 years; range, 33 to 84 years). In addition to a full clinical history and examination, preoperative assessment consisted of barium swallow, biochemical and hematological screening, esophagoscopy, and, in the case of midthoracic lesions, bronchoscopy. Further investigations for evaluation of local or distant metastatic disease were performed only when indicated either clinically or from information derived from the routine preoperative assessment. Histological cell types were as follows: adenocarcinoma, 60.8%; squamous, 35.7%; undifferentiated, 1.7%; small cell, 0.9%; and malignant lymphoma, 0.9%. Eightytwo tumors (71.3%)were located in the lower third of the esophagus and cardia, and 33 (28.7%)were in the middle third. No tumors in the upper third were resected by left thoracotomy. Staging was determined by histopathological examination of the surgical specimen. A double-lumen endotracheal tube is used for anesthesia so that the left lung can be collapsed during operation. A left thoracotomy is made through the bed of the seventh rib. The skin incision extends forward almost to the costal margin, but the anterior costal margin itself is left intact. The esophagus is mobilized from the diaphragmatic hiatus to a point at least 5 cm above the upper level of the tumor. When we attempt a curative resection, we extend mobilization behind and above the aortic arch, taking care to preserve the recurrent laryngeal nerve. If necessary, the arch is mobilized by dividing one or two intercostal branches. We have not found it necessary to resect the seventh rib. The diaphragm is opened in line with the wound, leaving a 1-inch cuff anteriorly to facilitate closure. The stomach is mobilized using the LDS stapler (Auto-
sophageal resection is currently accepted as the best form of treatment for operable tumors affecting the middle and lower thirds of the esophagus. Not only does it provide the best chance of curing the disease, but it is also the most effective method of palliation. The manner in which the operation is approached is controversial, however. Several different incisions, often in combination, are widely used. These include left thoracoabdominal [l]; laparotomy and right thoracotomy [2,3]; three-stage resection comprising laparotomy, right thoracotomy, and a cervical incision [4]; the transhiatal approach [5, 61; bilateral thoracotomies [7]; and left thoracophrenotomy, first described by Churchill and Sweet [8]. The left thoracophrenotomy approach to tumors of the middle and lower thirds of the esophagus and cardia of the stomach is still used by thoracic surgeons but has attracted little attention in the literature in recent years. We believe it offers advantages over other approaches and, since May 1983, have adopted it for routine resections of all esophageal tumors. This report describes our experience with this approach.
Material and Methods The case records of all patients undergoing esophagogastrectomy under the care of one surgeon (R.J.D.) between April 1983 and February 1989 were retrospectively reviewed. During this period, a left thoracophrenotomy incision has been the preferred approach for this procedure and has been used in 115 patients. This group constitutes the major part of this report. Twenty-two patients have undergone esophageal resection by some Accepted for publication Dec 26, 1989. Address reprint requests to Mr Donnelly, Broadgreen Hospital, Thomas Dr, Liverpool, England L14 3LB. Dr Whyte’s present address is Department of Cardiac Surgery, Massachusetts General Hospital, Boston, MA.
0 1990 by
The Society of Thoracic Surgeons
0003-4975/90/$3.50
764
FAGEETAL ESOPHAGOGASTRECTOMY
Ann Thorac Surg 1990;49:76=
Table 1. Incisions Other Than Left Thouacophrenotomy, 1983-1 989
Incision and Reason Left thoracoabdominal Tumor clearance
Previous operation Poor exposure Previous benign perforation Acute perforation Unspecified Total Cervical plus left thoracophrenotomy Small proximal thoracic esophagus Proximal tumor Total
n 5 2 2 1 1 2 13
3 2 5
Right thoracotomy plus laparotomy Tumor at level of archileft main bronchus Previous left thoracic operation Suspicious nodule in right lung Total Total
2 1 1 4 22
suture, UK) around the greater curvature except for the short gastric vessels, for which the Surgiclip instrument is used. After division of the short gastric vessels, the gastrocolic omentum, the left gastric artery, and the lesser omentum, the stomach is delivered into the left hemithorax through the hiatus. It is transected using a linear stapler at a site at least 5 cm distal to the lower margin of the tumor. The esophagus is divided proximally, and the tumor-bearing segment is removed. More recently, however, we have temporarily left the posterior wall intact to facilitate introduction of the detachable anvil of the endto-end anastomotic (EEA) stapling instrument, after which the remaining part of the esophagus is divided. Esophagogastric continuity is normally restored using the EEA stapler (Autosuture) as previously described [9, lo]. The 25-mm head was used in 34 patients, the 28-mm head was used in 64, and the 31-mm head was used in 15. In 3 patients, initial sizing with the 28-mm head led to mucosal tearing of the proximal esophagus and it was necessary to use the 25-mm head to perform the anastomosis more proximally. In 2 patients, a hand-sewn anastomosis was necessary because the proximal esophagus was too small for the 25-mm head to be inserted. Eighty anastomoses (69.5%) were constructed below and 35 (30.5%) were constructed above the aortic arch. In this latter group, the stomach is routed lateral to the arch. After placement of a nasogastric tube, the anterior gastrotomy through which the EEA stapler was inserted is closed with a linear stapler. All straight staple lines are oversewn with a continuous 3-0 Mersilene suture. Neither fundoplication nor pyloroplasty are routinely used. Patients are generally extubated in the operating room and returned directly to the general ward. After operation the nasogastric tube is left open to air until the drainage is minimal, at which time it is removed,
generally on the second postoperative day. Crystalloids are administered through a peripheral vein for fluid maintenence, and patients are allowed to drink water on the fourth day. Oral intake is advanced to soft solids over the next three days. Contrast radiographic studies are obtained only if there is clinical evidence to suspect an anastomotic leak or if there was difficulty performing the anastomosis at the time of operation. After discharge from the hospital, patients are seen at 3-month intervals and assessed for signs of tumor recurrence and anastomotic stricture formation. Patients with dysphagia undergo esophagoscopy and dilation if necessary. Follow-up to May 1989 was complete in 99.1% of patients. Radiotherapy and chemotherapy were not used preoperatively in this series and postoperatively were used only for palliation of recurrent or residual disease. Survival was calculated according to the method described by Grunkemeier and Starr [ll].
Results The TNM status of the tumors as deduced from histological examination of the resected specimens showed that 37 tumors were stage 1, 7 were stage 2A, 70 were stage 3, and 1 was stage 4. The gastric and esophageal resection margins were free of tumor in 108 (93.9%) of the specimens. Operative mortality, defined as deaths within 30 days or during initial hospital stay, was 8.7% (lOD15). The cause of death was anastomotic leakage in 1 patient, pneumonia in 4, pulmonary embolism in 2, chylothorax in 1, septicemia in 1, and myocardial infarction in 1. Eighteen patients (15.6%) had postoperative complications, although these did not delay discharge from the hospital for 8 patients. Complications that could be attributed to operative technique included one anastomotic leak, one gastrotomy leak, and one chylothorax. Patients were discharged from the hospital at an average of 13.1 days (*10.5; range, six to 99 days) after operation. Anastomotic leakage developed in 2 patients (1.7%), 1 of whom died as a result. The leakage was successfully repaired in the other patient three days after operation. Narrowing of the anastomosis sufficient to cause dysphagia and relieved by dilation developed in 16 patients. In 13 of these, a single dilation was sufficient to restore normal swallowing. There was no statistically significant difference between the rates of benign stricture formation of the three sizes of anastomotic stapler. Malignant strictures at the anastomosis occurred in 5 patients (4.3%). Survival is shown in Figure 1 and was 57.0% at 1 year, 32.3% at 2 years, and 22.1% at 3 years.
Comment Controversy persists regarding the optimal surgical approach for esophagogastrectomy, and a variety of techniques have been advocated. These include left thoracophrenotomy [8, 12-14], right thoracotomy and laparotomy [2-4, 151, left thoracoabdominal [ l , 161, left thoracotomy and midline laparotomy [ 171, bilateral thora-
PAGEETAL ESOPHAGOGASTRECTOMY
Ann Thorac Surg 1990;49:7634
XswIVMNo
3 YEARS
Fig 1. Actuarial 3-year survival of 115 patients undergoing esophagogasfrectomyfor malignant disease via left fhoracophrenofomy.
cotomies [7], and laparotomy and cervical incision without thoracotomy [5, 6, 181. A neck incision with cervical anastomosis may be added to any of these approaches. None of these various methods have made any significant difference in long-term survival, although some investigators consider subtotal esophagectomy to be important in this regard. Alternative techniques have generally been recommended and found favor among different surgeons for their ease of performance of a difficult operation and the consequent reduction in morbidity and mortality that should result. The left thoracophrenotomy approach is historically important as the method advocated by Sweet and others in the pioneering days of esophageal surgery. Although this approach is still practiced by some thoracic surgeons on both sides of the Atlantic, little appears in the literature about this technique. We believe that it offers several advantages over other methods for tumors of the middle third of the esophagus and below; for this reason, we report our experience with the technique. If the thoracotomy incision is made at the correct level through the sixth interspace, good access is available to the entire intrathoracic and abdominal esophagus, including the portion behind the aortic arch where dissection can be facilitated if necessary by mobilization of the aorta. Care must obviously be taken with the recurrent laryngeal nerve, and the course of the thoracic duct must be borne in mind. The supraaortic esophagus can be visualized and mobilized by dividing the pleura lateral to the left subclavian artery, and esophagogastric continuity can be reestablished by anastomosis at the apex of the thorax. We have not found it necessary to resect the seventh rib [14], but very rarely we make a second incision through the fourth interspace. This has been necessary only in patients requiring a thoracoabdominal incision. The left thoracophrenotomy incision offers adequate exposure of the stomach and excellent access to the short and left gastric arteries through the opening in the left hemidiaphragm. Even performance of pyloroplasty is often possible, although we do not perform it routinely. Opening and closing of the wound is rapid and simple and, although we have not attempted to quantify them, it is not unreasonable to assume potential benefits in terms
765
of reduced operating time and a smaller incision with consequent greater patient comfort and reduced wound morbidity. Stapling instruments have proved to be very useful in performing esophagogastrectomy through the left side of the chest. Not only do they facilitate mobilization of the stomach through the diaphragm, but use of the EEA stapler allows rapid and reliable creation of the anastomosis [9, 10, 19-21] at the apex of the left thorax, so that it is nearly cervical in location. Proponents of the cervical anastomosis argue that, should a leak occur, extravasation is confined to the neck, causing less morbidity than an intrapleural leak. Our rate of anastomotic leakage is low, and we do not believe that there is sufficient justification for a separate neck incision with its consequent cosmetic effects and prolongation of operating time. We have not found the left thoracophrenotomy approach to be associated with a significantly higher incidence of postoperative complications as compared with other approaches [22]. The rate of stricture formation was similar to that previously reported 1231. Some of the more frequently used alternatives to the left thoracophrenotomy approach are the right thoracotomy plus laparotomy or the left thoracoabdominal incision, either of which can be combined with a cervical incision, and the transhiatal esophagectomy. Arguments in favor of the right thoracotomy plus laparotomy approach are that a more complete tumor resection can be obtained and that a higher anastomosis can be created. We believe that an equally complete resection is possible through a left thoracotomy and that, particularly with use of the EEA stapler, an anastomosis at the apex of the chest is relatively straightforward. The left thoracoabdominal approach, as used by us before 1983, provides little added exposure over a left thoracophrenotomy and poses potential problems associated with division of the costal margin and anterior abdominal wall musculature. The transhiatal approach, as advocated by Orringer and associates [5, 61, has been criticized for violating the basic surgical principles of adequate exposure and hemostasis and for an excessive mortality and morbidity. Although Orringer and associates have argued these points well, Shahian and colleagues concluded [24], and we agree, that this approach should be viewed as an addition to the surgical armamentarium rather than a replacement of standard techniques. During the course of this series, we have occasionally elected to use one of the alternative approaches for the reasons listed in Table 1. In addition, we have not hesitated during the course of operation either to extend the incision to the abdominal wall because of difficultiesof access or tumor bulk or to perform the anastomosis in the neck if the proximal esophagus is too small or involved by tumor. Our results with esophagogastrectomy by left thoracophrenotomy are comparable with those of published series using alternative methods. We believe that this technique has advantages of its own and is worthy of its place in modern thoracic surgery.
766
PAGEETAL ESOFHAGOGASTRECTOMY
References 1. Heitmiller RF. The left thoracoabdominal incision. Ann Thorac Surg 1988;46:25&3. 2. Lewis I. The surgical treatment of carcinoma of the esophagus with special reference to a new operation for growths of the middle third. Br J Surg 1946;34:18. 3. Tanner NC. The present position of carcinoma of the esophagus. Postgrad Med J 1947;23:109. 4. McKeown KC. Trends in esophageal resection for carcinoma with special reference to total oesophagectomy. Ann R Coll Surg Engl 1972;51:21349. 5. Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978;7664&54. 6. Orringer MB, Orringer J. Esophagectomy without thoracotomy: a dangerous operation? J Thorac Cardiovasc Surg 1983;85:72-80. 7. Dark JF, Mousalli H, Vaughan R. Surgical treatment of carcinoma of the esophagus. Thorax 1981;36:891-5. 8. Churchill ED, Sweet RH. Transthoracic resection of tumours of the stomach and esophagus. Ann Surg 1942;115:897-917. 9. Fabri BM, Donnelly RJ. Esophagogastrectomyusing the end to end anastomotic stapler. Thorax 1982;37296-9. 10. Donnelly RJ, Sastry MR, Wright CD. Esophagogastrectomy using the end to end anastomotic stapler: results of the last 100 patients. Thorax 1985;4095&9. 11. Grunkemeier GL, Starr A. Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 1977;24 404-8. 12. Lam KH, Cheung HC, Wong J, Ong GB. The present state of surgical treatment of carcinoma of the esophagus. J R Coll Surg Edinburgh 1982;27315-26. 13. Adkins PC. Tumours of the esophagus. In: Sabiston DC, ed.
Ann Thorac Surg 1990:497634
Davis-Christopher textbook of surgery. 12th ed. Philadelphia: WB Saunders, 1981:84&55. 14. Pradhan GN, Eng J, Sabanathan S. Left thoracotomy approach for resection of carcinoma of the esophagus. Surg Gynecol Obstet 1989;168:49-53. 15. Mitchell RL. Abdominal and right thoracotomy approach as standard procedure for esophagogastrectomy with low morbidity. J Thorac Cardiovasc Surg 1987;93:20.511. 16. Wong J. Stapled esophagogastric anastomosis in the apex of the right chest after subtotal esophagectomy for carcinoma. Surg Gynecol Obstet 1987;164:569-72. 17. West PN, Marbarger JP, Martz MN, Roper CL. Esophagogastrectomy with the EEA stapler. Ann Surg 1981;193:76-81. 18. Baker JW, Schechter GL. Management of panesophageal cancer by blunt resection without thoracotomy and reconstruction with stomach. Ann Surg 1986;203:491-9. 19. Steichen FM, Ravitch MM. Mechanical sutures in esophageal surgery. Ann Surg 1980;191:37?-91. 20. Hopkins RA, Alexander JC, Postlethwaite RW. Stapled esophagogastric anastomosis. Am J Surg 1984;147283-7. 21. Sugimachi K, Ikeda M, Ueo H, Kai H, Okudaira Y, Inokuchi K. Clinical efficacy of the stapled anastomosis on esophageal reconstruction. Ann Thorac Surg 1981;33:374-8. 22. Muehrcke DD, Donnelly RJ. Complications after esophagogastrectomy using stapling instruments. Ann Thorac Surg 1989;48:257-62. 23. Muehrcke DD, Kaplan DK, Donnelly RJ. Anastornotic narrowing after esophagogastrectomy with the EEA stapling device. J Thorac Cardiovasc Surg 1989;97434-8. 24. Shahian DM, Neptune WB, Ellis FH. Transthoracic versus extrathoracic esophagectomy: mortality, morbidity, and long-term survival. Ann Thorac Surg 1986;41:23746.