- Email: [email protected]
Bowel Interposition for Esophageal Replacement: Twenty-Five–Year Experience
ORIGINAL ARTICLES: GENERAL THORACIC
Bowel Interposition for Esophageal Replacement: Twenty-Five–Year Experience Kamal A. Mansour, MD, F. Curtis Bryan, MD, and Grant W. Carlson, MD Joseph B. Whitehead Department of Surgery, Divisions of Cardiothoracic and Plastic and Reconstructive Surgery, Emory University School of Medicine, Atlanta, Georgia
Background. From 1972 to 1996, bowel interposition reconstruction after esophagectomy for benign and malignant conditions was performed in 129 of 131 patients. The indication for operation was benign disease in 94 patients (72.9%) and malignant disease in 35 patients (27.1%). Benign stricture was the most common presentation in the benign group (41 patients), and adenocarcinoma was the most common indication in the malignant group (19 patients). Methods. One hundred thirty-three conduits were performed in the 129 patients. Four patients (3.1%) required reoperative reconstruction. Of the 97 conduits employed for reconstruction of benign disease, the right colon was used in 70 patients, the left colon in 9 patients, and the transverse colon in 4 patients. A jejunal interposition graft was employed in 11 patients and a free jejunal autograft in 3 patients. The right colon was used in 15 patients with malignant disease, the left colon in 9 patients, and the jejunum in 12 patients.
Results. The mean age of the population was 54.5 years (range, 14 to 72 years) with a male-to-female ratio of 1.3:1. The average number of prior thoracic or abdominal procedures was 2.9 (range, 1 to 8) with 50.9% of patients undergoing reoperation. The mean length of stay was 21.7 days (range, 8 to 290 days). Complications occurred in 37.1% of patients with anastomotic leak occurring in 14.8% and ischemic colitis in 3.0% of conduits performed. The in-hospital mortality was 5.9%. Conclusions. Bowel interposition reconstruction after esophagectomy for benign and malignant disease can be performed with an acceptable morbidity and mortality, despite prior operative procedures in the abdomen or chest. Colonic and jejunal conduits, employed alone or in combination, can effectively restore gastrointestinal continuity.
S
Material and Methods
ince the beginning of the century, techniques for reconstruction of the esophagus have evolved to include the use of intestinal and colonic conduits. In 1911, Kelling [1] performed the initial stage of reconstruction of the esophagus employing a subcutaneously tunneled isoperistaltic transverse colon conduit. Subsequently, with improvements in the technical aspects of the operation and in postoperative care, the colonic conduit has become a safer and more widely applied graft for reconstruction of the esophagus. In a similar fashion, techniques employing the small intestine have evolved from the use of interposition segments to that of microvascular jejunal transplantation. A review of the experience of a single surgeon (K.A.M.), who uses all types of conduits, allows for recommendations with regard to conduit selection and management of complications. Consequently, we report an experience of greater than 20 years’ duration with the replacement of the esophagus with pedicled and free bowel interposition grafts.
Presented at the Poster Session at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3–5, 1997. Address reprint requests to Dr Mansour, Division of Cardiothoracic Surgery, The Emory Clinic, 1365 Clifton Rd NE, Atlanta, GA 30322.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 1997;64:752– 6) © 1997 by The Society of Thoracic Surgeons
From July 1972 to June 1996, 131 patients with attempted esophageal reconstruction with a bowel interposition graft were identified from the databases of Emory University Hospital, Crawford Long Hospital, Egleston Childrens Hospital, and the Veterans Administration Medical Center. The charts of the patients were reviewed and analyzed. Complete information regarding hospital course was available for 101 patients. Two patients could not complete the reconstruction because of intraoperative difficulties necessitating abortion of the procedure. Four patients required reoperative reconstruction within 6 months of the hospital admission. The indication for operation was benign disease in 94 patients and malignant disease in 35 patients.
Preoperative Evaluation A contrast esophagogram was performed in all patients preoperatively. Manometry and pH studies were reserved for those patients with motility disorders or reflux esophagitis. A barium enema was performed in all patients for whom a colonic interposition was planned. Angiography of the mesenteric vasculature was reserved for those patients older than 60 years of age and patients with prior intestinal resection or peripheral vascular disease. Angiography was helpful in 3 patients, who had 0003-4975/97/$17.00 PII S0003-4975(97)00616-4
Ann Thorac Surg 1997;64:752– 6
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
753
malignant disease, 24 were reconstructed exclusively with colonic conduits and 11 with jejunal conduits. The right colon was used in 15 patients and the left colon in 9 patients. Of those patients receiving a jejunal graft, 10 were interposition and 1 free. One patient required reconstruction during the same admission with a free jejunal graft after a failed left colon interposition graft. Figure 2 illustrates the distribution of conduits in the group with malignant disease.
Postoperative Management Fig 1. Distribution of conduits in benign disease (n 5 97).
a previous colonic resection, in outlining the vascular arcade of the segment of bowel to be interposed. Mechanical and antibiotic bowel preparation was performed in all patients. Contraindications to bowel interposition reconstruction included intrinsic bowel pathology, mesenteric arterial disease, and subnormal motility.
Operative Approach All patients brought to the operative suite underwent esophagoscopy to delineate the extent of the pathology. Bronchoscopy was performed in those patients with malignant disease of the upper and middle third of the esophagus, in whom a concern for tracheoesophageal fistula was present. The routes of graft placement included 2 subcutaneous, 41 retrosternal, and 90 posterior mediastinal. The retrosternal route was usually elected when a long segment of bowel was used for bypass without esophageal resection and the posterior mediastinal route was usually employed when esophagectomy was done and a short or long segment of bowel was required for interposition. Apart from the aesthetic aspect of the subcutaneous placement, we found that patients had no difficulty of swallowing related to the route of graft placement. Ninety-seven grafts were performed in 94 patients for benign disease. Of the 83 colon interposition grafts performed, the right colon was employed in 70 cases, the left colon in 9 cases, and the transverse colon in 4 cases. A jejunal interposition graft was used for reconstruction in 11 cases and a free jejunal graft in 3 cases. One patient required reconstruction with both right colon and free jejunal grafts to reestablish esophagogastric continuity. Three patients underwent reoperative reconstruction within 6 months of the primary bowel interposition reconstruction. The left colon was employed in 2 patients and a free jejunal transfer in 1 patient. One of the three patients required a second reoperative procedure employing a free jejunal transfer. Two procedures were aborted in the benign population; 1 patient experienced intraoperative hypotension and the other patient was reconstructed with stomach because of ischemia of the left colonic conduit. Figure 1 illustrates the distribution of conduits in the population of patients with benign disease. In the 35 patients who underwent reconstruction for
All patients were decompressed postoperatively with nasogastric suction. A Gastrografin (E.R. Squibb & Sons, Princeton, NJ) esophagogram, followed by barium swallow if the Gastrografin study showed no leak, was performed between postoperative day 6 and 9. Patients whose study did not demonstrate any evidence of leak or obstruction were given a clear liquid diet and advanced as tolerated. Patients with small cervical leaks were allowed to begin oral alimentation as well. Intrathoracic anastomotic leaks were managed with thoracostomy tube drainage or reoperation as dictated by the clinical course of the patient.
Results Of the 131 patients in whom esophageal reconstruction with bowel interposition was attempted the procedure was successful in 98% of patients. Only 1 patient could not be reconstructed because of inadequate conduit. The mean age of the group was 54.4 years with a range of 14 to 72 years. The male to female ratio was 1.3:1. The indication for operation was benign disease in 94 patients (72.9%) and malignant disease in 35 patients (27.1%). An undilatable benign stricture (n 5 41) was the most common reason for reconstruction in the benign group, followed by advanced motility disorders in which myotomy was not effective or indicated (n 5 11), Barrett’s esophagus with high-grade dysplasia (n 5 9), refractory reflux esophagitis after failed multiple antireflux procedures (n 5 8), and extensive chemical strictures unsuitable for dilation (n 5 8). Other miscellaneous conditions comprised the indication for operation in the remainder of this population of patients. Figure 3 illustrates the
Fig 2. Distribution of conduits in malignant disease (n 5 36).
754
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
Ann Thorac Surg 1997;64:752– 6
Fig 5. Nonfatal postoperative complications in 101 patients. Fig 3. Indications for operation in benign disease (n 5 97).
indications for operation in the population with benign disease. Adenocarcinoma of the distal esophagus and stomach (n 5 19) was the most common indication for operation in the group of patients with malignant disease. Squamous cell carcinoma (n 5 14), lymphoma (n 5 2), and failed interposition graft (n 5 1) were the indications in the remainder of the group. Figure 4 illustrates the indications for operation in patients with malignant disease. Complete data concerning the postoperative course were obtained in 101 patients (78.9%). The mean length of stay postoperatively was 21.7 days with a range of 8 to 290 days. The overall postoperative complication rate was 38.2%. The most common complication was anastomotic leak occurring in 14.8% of patients, followed by wound infections in 5.9%. The overall mortality was 5.9%. Figure 5 describes the distribution of complications. Long-term follow-up from 6 months to 20 years of patients who had bowel interposition for benign disease showed continued improvement of swallowing habits, maintenance of weight, and growth pattern. Stricture formation was rarely encountered; only 3 patients required either dilation or revision of the anastomosis, proximal or distal, within the first 2 years postoperatively. Of 35 patients who underwent bowel interposition for malignant disease, 3 patients were lost to follow-up. Twenty-seven patients died between 2 and 6 years after the operation of causes unrelated to the bowel interposition. The remain-
Fig 4. Indications for operation in malignant disease (n 5 36).
ing 5 patients are alive and doing well between 2 and 10 years postoperatively and have no difficulty swallowing or maintaining their weight; the 10-year survivor required three dilations and none in the past 2 years.
Comment Early attempts at esophageal resection and reconstruction were limited to the cervical esophagus and consisted primarily of skin tubes to reestablish continuity. In the current era of esophageal reconstruction, several different types of conduits are employed with relative success. Short-segment colon interposition grafts, usually the right or transverse colon, are targeted for reconstruction of the intrathoracic esophagus. Long colon interposition grafts, often the left colon, are employed for anastomosis to the cervical esophagus or pharynx. Jejunal interposition and free jejunal grafts have been used for a variety of reasons, especially in cases of reoperative reconstruction. Classically, isoperistaltic colon interposition grafts are employed in reconstruction of the esophagus when longterm survival of the patient is anticipated. Debate as to which segment of colon is ideally suited for reconstructive purposes has focused primarily on the right versus left colon. The left colon has been considered by many to be a preferable conduit for several reasons. First, the diameter of the left colon is smaller and less prone to dilatation. The blood supply has been shown in anatomic studies of Ventemiglia and colleagues [2] to be more reliable than that of the right colon. The left colon provides adequate length for reconstruction of not only the intrathoracic esophagus but also the cervical esophagus and pharynx. Finally, the left colon is quite effective at propelling a solid bolus. In the current study, the right colon was the conduit of choice for reconstructions employing a colonic graft. Eighty-five right colon conduits were harvested successfully with no intraoperative difficulties caused by arterial anatomy. In fact, all cases of ischemic colitis that occurred intraoperatively presented in left colon conduits. One of the technical considerations that has resulted in successful use of the right colon is the mobilization of the conduit at the beginning of the operation. The vessels that are selected for ligation are occluded with small bulldog clamps while attention is
Ann Thorac Surg 1997;64:752– 6
directed to mobilization of the stomach and esophagus. At the completion of gastric and distal esophageal mobilization, the viability of the colon can be assured before ligation of the arterial branches. Adhering to these intraoperative principles, the right colon can be harvested and employed quite effectively as a conduit. In fact, the results are comparable with those reported in 1988 by DeMeester and colleagues [3], who reconstructed 85 patients with left colon grafts. In their study, 3 of the 85 patients demonstrated intraoperative ischemia. Only seven right colon grafts were employed by this group. Consequently, the current experience at our institution suggests that, at least from a technical standpoint, the right colon, placed in an isoperistaltic fashion, can be employed as effectively as the left colon for reconstruction of the esophagus for benign conditions [4]. When the esophagus is resected for malignant disease, the stomach has been the conduit used most commonly for restoration of gastrointestinal continuity. However, in patients with previous gastric resection, or in those patients who require total gastrectomy, bowel interposition has been used. Recently, Isolauri and colleagues [5] reported the use of colon interposition grafts for restoration of continuity in resection of malignant disease. They retrospectively reviewed 248 patients in whom the diagnosis of squamous cell carcinoma was present in 73%. The left colon was the conduit employed in 54% of patients and the right colon in 27% of patients. The remainder of reconstructions were performed with the transverse colon. The complication rate for this group of patients was 37% with a mortality of 16%. In a similar manner, the colon has been employed in our institution for reconstruction in malignant disease, although less commonly. The right colon (41.7%) was used most often, followed by the left colon (25%). The rate of anastomotic leakage for this subgroup of patients was 4.2% with a mortality of 8.8%. These results are certainly acceptable and comparable to those published for benign disease. Because of the limited number of colonic interposition grafts used for reconstruction for malignant disease in the current study, it is difficult to make any concrete conclusions. The data from this study and published studies, however, suggest that a colon interposition graft is a reasonable alternative when the stomach is not an option for reconstruction. This technique does carry a considerable morbidity and mortality for what is a palliative procedure in most patients. As colonic interposition grafts are employed largely for reconstructions of the esophagus, the jejunum can be used in a variety of circumstances. The jejunal interposition graft, in a Roux-en-Y fashion, is often selected in patients in whom total gastrectomy is planned as part of the operative procedure. The majority of these patients are undergoing resection of malignant disease. On the other hand, reconstruction of the cervical esophagus has seen an evolution of technical advances. In 1942, Wookey [6] described a technique based on the creation of lateral neck flaps. Twenty years later, Bakamjian [7] reported a two-stage reconstruction employing a deltopectoral flap. Despite the improvement these techniques provided
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
755
over the creation of skin tubes, they did require a second-stage procedure and were not as physiologic as other bowel interposition grafts. The era of the free jejunal graft began in 1907 with the microvascular work of Alexis Carrel, who successfully transplanted small intestine into the neck of a dog [8]. Fifty years after this experimental landmark, Seidenberg and colleagues [9] performed the first jejunal free graft in a patient. With further advances in microvascular surgical techniques, especially during the development of myocutaneous free flaps, jejunal transplantation became more commonly employed for reconstruction of the esophagus, especially the pharyngocervical portion. In 1989, Coleman and others [10], from our institution, reported their experience with the cervical jejunal free autograft in 101 patients, 91 of whom were operated on for malignant disease. Sixty-eight percent of the patients with malignant disease underwent synchronous extirpation of the primary tumor. The remainder of this subgroup of patients were reoperative cases for fistula or stricture after laryngectomy or radiotherapy. Thirteen graft failures were reported as well as 33 pharyngocutaneous fistulas. The mortality rate was 5%. They concluded that, despite the relatively high complication rate, free jejunal autografts can be employed effectively for restoration of gastrointestinal continuity. In the current study, four jejunal autografts were performed for reconstruction of the esophagus. Three of the grafts were performed for failed bowel interposition and required free jejunal transfers between the pharynx or cervical esophagus and the salvaged segment of colonic interposition. We have reported these patients previously [11, 12]. Our experience showed that, unlike colonic interpositions, which recorded ineffective motor contractions on manometric studies [4], jejunal grafts showed vigorous peristaltic waves, and isoperistaltic placement of the graft is essential. Admittedly, reconstruction of the esophagus with bowel interposition grafts in this series resulted in a morbidity rate of 38.2%. The most notable characteristic of the patient population of the current review is the significant proportion of patients who had undergone prior thoracic or intraabdominal procedures, further complicating any operative intervention. Considering the fact that more than half of the patients had undergone a mean of three prior operative procedures, an anastomotic leakage rate of 14.8% is acceptable. This rate of anastomotic leakage compares favorably with other published series [3, 5, 10, 13]. In fact, 44% of the leaks occurred in reoperative cases. Most of these were treated nonoperatively with drainage of the leak either by tube thoracostomy or by opening of the cervical wound. Those not amenable to more conservative means of treatment were salvaged with a reoperative bowel interposition graft. This situation in which gastrointestinal discontinuity exists between the proximal esophagus and the distal interposition remnant is particularly suited for the free jejunal autograft. This conduit can be tailored to fill the intervening defect without the need for a second colon interposition graft. The 3 patients in whom this method
756
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
of salvage was employed experienced no dysphagia, regurgitation, or nutritional difficulties. The mortality rate of 5.9% falls within the range of previously published series. The major cause of death for the population of patients studied was sepsis secondary to nosocomial pneumonia. The last patient mortality occurred in 1989. Since that time, 39 patients have undergone reconstruction for benign and malignant disease without a single mortality. This low mortality, especially in the most recent 7 years, testifies to the improvements in the perioperative management of these very complex patients in the current era of critical care medicine. In conclusion, techniques of bowel interposition reconstruction of the esophagus have evolved to include the jejunum and colon in selected circumstances. Reconstruction employing a colonic conduit is well-suited for reconstruction in patients with benign and, in select cases, malignant disease. The right colon is a suitable conduit and can be harvested as reliably as a left colon conduit if certain principles are followed intraoperatively. Likewise, the jejunal interposition graft is a suitable choice in patients in whom the entire stomach is resected. The free jejunal autograft is a very useful graft, particularly in cases of salvage reconstruction. The microvascular anastomosis for this graft is performed by members of our Division of Plastic and Reconstructive Surgery. Reconstruction with bowel interposition grafts can be performed with a morbidity of approximately 38% and a mortality of approximately 6%, even in a population of patients in which almost half are reoperative procedures. Gastrointestinal continuity can be restored effectively by employing a combination of interposition grafts if necessary. The free jejunal autograft is an excellent choice when primary reconstruction has failed.
Ann Thorac Surg 1997;64:752– 6
References 1. Kelling G. Oesopagoplastik mit Hilfe de Querkolon. Zentralblatt Chir 1911;38:1209–12. 2. Ventemiglia R, Khalil KG, Frazier OH, Mountain CF. The role of preoperative arteriography in colon interposition. J Thorac Cardiovasc Surg 1977;74:98 –104. 3. DeMeester TR, Johansson K-E, Franze I, et al. Indications, surgical technique, and long-term functional results of colon interposition or bypass. Ann Surg 1988;208:460–74. 4. Mansour KA, Hansen HA II, Hersh T, Miller JI, Hatcher CR Jr. Colon interposition for advanced non-malignant esophageal stricture. Ann Thorac Surg 1981;32:584–91. 5. Isolauri J, Markkula H, Autio V. Colon interposition in the treatment of carcinoma of the esophagus and the gastric cardia. Ann Thorac Surg 1987;43:420– 4. 6. Wookey H. The surgical treatment of carcinoma of the pharynx and upper esophagus. Surg Gynecol Obstet 1942; 75:499 –502. 7. Bakamjian VY. A two-stage method for pharyngoesophageal reconstruction with a primary pectoral skin flap. Plast Reconstr Surg 1965;36:173– 84. 8. Paletta CE, Jurkiewicz MJ. Esophageal replacement: microvascular jejunal transplantation. In: Baue AE, Geha AS, Hammond GL, Laks H, Naunheim KS, eds. Glenn’s thoracic and cardiovascular surgery. Stamford, CT: Appleton & Lange, 1996:931– 8. 9. Seidenberg B, Rosenak SS, Hurwitt ES, Som ML. Immediate reconstruction of the cervical esophagus by a revascularized isolated jejunal segment. Ann Surg 1959;149:162– 6. 10. Coleman JJ, Tan K-C, Searles JM, Hester TR, Nahai F. Jejunal free autograft: complications and their resolution. Plast Reconstr Surg 1989;84:589–95. 11. Mansour KA, Picone AL, Coleman JJ III. Surgery for high cervical esophageal carcinoma: experience with 11 patients. Ann Thorac Surg 1990;49:597– 602. 12. Carlson GW, Anderson TM, Galloway JR, Mansour KA. Salvage of colon interposition by antethoracic free jejunal transfer. Ann Thorac Surg 1994;58:1523–5. 13. Neville WE, Najem AZ. Colon replacement of the esophagus for congenital and benign disease. Ann Thorac Surg 1983;36: 626–33.
Bowel Interposition for Esophageal Replacement: Twenty-Five–Year Experience Kamal A. Mansour, MD, F. Curtis Bryan, MD, and Grant W. Carlson, MD Joseph B. Whitehead Department of Surgery, Divisions of Cardiothoracic and Plastic and Reconstructive Surgery, Emory University School of Medicine, Atlanta, Georgia
Background. From 1972 to 1996, bowel interposition reconstruction after esophagectomy for benign and malignant conditions was performed in 129 of 131 patients. The indication for operation was benign disease in 94 patients (72.9%) and malignant disease in 35 patients (27.1%). Benign stricture was the most common presentation in the benign group (41 patients), and adenocarcinoma was the most common indication in the malignant group (19 patients). Methods. One hundred thirty-three conduits were performed in the 129 patients. Four patients (3.1%) required reoperative reconstruction. Of the 97 conduits employed for reconstruction of benign disease, the right colon was used in 70 patients, the left colon in 9 patients, and the transverse colon in 4 patients. A jejunal interposition graft was employed in 11 patients and a free jejunal autograft in 3 patients. The right colon was used in 15 patients with malignant disease, the left colon in 9 patients, and the jejunum in 12 patients.
Results. The mean age of the population was 54.5 years (range, 14 to 72 years) with a male-to-female ratio of 1.3:1. The average number of prior thoracic or abdominal procedures was 2.9 (range, 1 to 8) with 50.9% of patients undergoing reoperation. The mean length of stay was 21.7 days (range, 8 to 290 days). Complications occurred in 37.1% of patients with anastomotic leak occurring in 14.8% and ischemic colitis in 3.0% of conduits performed. The in-hospital mortality was 5.9%. Conclusions. Bowel interposition reconstruction after esophagectomy for benign and malignant disease can be performed with an acceptable morbidity and mortality, despite prior operative procedures in the abdomen or chest. Colonic and jejunal conduits, employed alone or in combination, can effectively restore gastrointestinal continuity.
S
Material and Methods
ince the beginning of the century, techniques for reconstruction of the esophagus have evolved to include the use of intestinal and colonic conduits. In 1911, Kelling [1] performed the initial stage of reconstruction of the esophagus employing a subcutaneously tunneled isoperistaltic transverse colon conduit. Subsequently, with improvements in the technical aspects of the operation and in postoperative care, the colonic conduit has become a safer and more widely applied graft for reconstruction of the esophagus. In a similar fashion, techniques employing the small intestine have evolved from the use of interposition segments to that of microvascular jejunal transplantation. A review of the experience of a single surgeon (K.A.M.), who uses all types of conduits, allows for recommendations with regard to conduit selection and management of complications. Consequently, we report an experience of greater than 20 years’ duration with the replacement of the esophagus with pedicled and free bowel interposition grafts.
Presented at the Poster Session at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3–5, 1997. Address reprint requests to Dr Mansour, Division of Cardiothoracic Surgery, The Emory Clinic, 1365 Clifton Rd NE, Atlanta, GA 30322.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
(Ann Thorac Surg 1997;64:752– 6) © 1997 by The Society of Thoracic Surgeons
From July 1972 to June 1996, 131 patients with attempted esophageal reconstruction with a bowel interposition graft were identified from the databases of Emory University Hospital, Crawford Long Hospital, Egleston Childrens Hospital, and the Veterans Administration Medical Center. The charts of the patients were reviewed and analyzed. Complete information regarding hospital course was available for 101 patients. Two patients could not complete the reconstruction because of intraoperative difficulties necessitating abortion of the procedure. Four patients required reoperative reconstruction within 6 months of the hospital admission. The indication for operation was benign disease in 94 patients and malignant disease in 35 patients.
Preoperative Evaluation A contrast esophagogram was performed in all patients preoperatively. Manometry and pH studies were reserved for those patients with motility disorders or reflux esophagitis. A barium enema was performed in all patients for whom a colonic interposition was planned. Angiography of the mesenteric vasculature was reserved for those patients older than 60 years of age and patients with prior intestinal resection or peripheral vascular disease. Angiography was helpful in 3 patients, who had 0003-4975/97/$17.00 PII S0003-4975(97)00616-4
Ann Thorac Surg 1997;64:752– 6
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
753
malignant disease, 24 were reconstructed exclusively with colonic conduits and 11 with jejunal conduits. The right colon was used in 15 patients and the left colon in 9 patients. Of those patients receiving a jejunal graft, 10 were interposition and 1 free. One patient required reconstruction during the same admission with a free jejunal graft after a failed left colon interposition graft. Figure 2 illustrates the distribution of conduits in the group with malignant disease.
Postoperative Management Fig 1. Distribution of conduits in benign disease (n 5 97).
a previous colonic resection, in outlining the vascular arcade of the segment of bowel to be interposed. Mechanical and antibiotic bowel preparation was performed in all patients. Contraindications to bowel interposition reconstruction included intrinsic bowel pathology, mesenteric arterial disease, and subnormal motility.
Operative Approach All patients brought to the operative suite underwent esophagoscopy to delineate the extent of the pathology. Bronchoscopy was performed in those patients with malignant disease of the upper and middle third of the esophagus, in whom a concern for tracheoesophageal fistula was present. The routes of graft placement included 2 subcutaneous, 41 retrosternal, and 90 posterior mediastinal. The retrosternal route was usually elected when a long segment of bowel was used for bypass without esophageal resection and the posterior mediastinal route was usually employed when esophagectomy was done and a short or long segment of bowel was required for interposition. Apart from the aesthetic aspect of the subcutaneous placement, we found that patients had no difficulty of swallowing related to the route of graft placement. Ninety-seven grafts were performed in 94 patients for benign disease. Of the 83 colon interposition grafts performed, the right colon was employed in 70 cases, the left colon in 9 cases, and the transverse colon in 4 cases. A jejunal interposition graft was used for reconstruction in 11 cases and a free jejunal graft in 3 cases. One patient required reconstruction with both right colon and free jejunal grafts to reestablish esophagogastric continuity. Three patients underwent reoperative reconstruction within 6 months of the primary bowel interposition reconstruction. The left colon was employed in 2 patients and a free jejunal transfer in 1 patient. One of the three patients required a second reoperative procedure employing a free jejunal transfer. Two procedures were aborted in the benign population; 1 patient experienced intraoperative hypotension and the other patient was reconstructed with stomach because of ischemia of the left colonic conduit. Figure 1 illustrates the distribution of conduits in the population of patients with benign disease. In the 35 patients who underwent reconstruction for
All patients were decompressed postoperatively with nasogastric suction. A Gastrografin (E.R. Squibb & Sons, Princeton, NJ) esophagogram, followed by barium swallow if the Gastrografin study showed no leak, was performed between postoperative day 6 and 9. Patients whose study did not demonstrate any evidence of leak or obstruction were given a clear liquid diet and advanced as tolerated. Patients with small cervical leaks were allowed to begin oral alimentation as well. Intrathoracic anastomotic leaks were managed with thoracostomy tube drainage or reoperation as dictated by the clinical course of the patient.
Results Of the 131 patients in whom esophageal reconstruction with bowel interposition was attempted the procedure was successful in 98% of patients. Only 1 patient could not be reconstructed because of inadequate conduit. The mean age of the group was 54.4 years with a range of 14 to 72 years. The male to female ratio was 1.3:1. The indication for operation was benign disease in 94 patients (72.9%) and malignant disease in 35 patients (27.1%). An undilatable benign stricture (n 5 41) was the most common reason for reconstruction in the benign group, followed by advanced motility disorders in which myotomy was not effective or indicated (n 5 11), Barrett’s esophagus with high-grade dysplasia (n 5 9), refractory reflux esophagitis after failed multiple antireflux procedures (n 5 8), and extensive chemical strictures unsuitable for dilation (n 5 8). Other miscellaneous conditions comprised the indication for operation in the remainder of this population of patients. Figure 3 illustrates the
Fig 2. Distribution of conduits in malignant disease (n 5 36).
754
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
Ann Thorac Surg 1997;64:752– 6
Fig 5. Nonfatal postoperative complications in 101 patients. Fig 3. Indications for operation in benign disease (n 5 97).
indications for operation in the population with benign disease. Adenocarcinoma of the distal esophagus and stomach (n 5 19) was the most common indication for operation in the group of patients with malignant disease. Squamous cell carcinoma (n 5 14), lymphoma (n 5 2), and failed interposition graft (n 5 1) were the indications in the remainder of the group. Figure 4 illustrates the indications for operation in patients with malignant disease. Complete data concerning the postoperative course were obtained in 101 patients (78.9%). The mean length of stay postoperatively was 21.7 days with a range of 8 to 290 days. The overall postoperative complication rate was 38.2%. The most common complication was anastomotic leak occurring in 14.8% of patients, followed by wound infections in 5.9%. The overall mortality was 5.9%. Figure 5 describes the distribution of complications. Long-term follow-up from 6 months to 20 years of patients who had bowel interposition for benign disease showed continued improvement of swallowing habits, maintenance of weight, and growth pattern. Stricture formation was rarely encountered; only 3 patients required either dilation or revision of the anastomosis, proximal or distal, within the first 2 years postoperatively. Of 35 patients who underwent bowel interposition for malignant disease, 3 patients were lost to follow-up. Twenty-seven patients died between 2 and 6 years after the operation of causes unrelated to the bowel interposition. The remain-
Fig 4. Indications for operation in malignant disease (n 5 36).
ing 5 patients are alive and doing well between 2 and 10 years postoperatively and have no difficulty swallowing or maintaining their weight; the 10-year survivor required three dilations and none in the past 2 years.
Comment Early attempts at esophageal resection and reconstruction were limited to the cervical esophagus and consisted primarily of skin tubes to reestablish continuity. In the current era of esophageal reconstruction, several different types of conduits are employed with relative success. Short-segment colon interposition grafts, usually the right or transverse colon, are targeted for reconstruction of the intrathoracic esophagus. Long colon interposition grafts, often the left colon, are employed for anastomosis to the cervical esophagus or pharynx. Jejunal interposition and free jejunal grafts have been used for a variety of reasons, especially in cases of reoperative reconstruction. Classically, isoperistaltic colon interposition grafts are employed in reconstruction of the esophagus when longterm survival of the patient is anticipated. Debate as to which segment of colon is ideally suited for reconstructive purposes has focused primarily on the right versus left colon. The left colon has been considered by many to be a preferable conduit for several reasons. First, the diameter of the left colon is smaller and less prone to dilatation. The blood supply has been shown in anatomic studies of Ventemiglia and colleagues [2] to be more reliable than that of the right colon. The left colon provides adequate length for reconstruction of not only the intrathoracic esophagus but also the cervical esophagus and pharynx. Finally, the left colon is quite effective at propelling a solid bolus. In the current study, the right colon was the conduit of choice for reconstructions employing a colonic graft. Eighty-five right colon conduits were harvested successfully with no intraoperative difficulties caused by arterial anatomy. In fact, all cases of ischemic colitis that occurred intraoperatively presented in left colon conduits. One of the technical considerations that has resulted in successful use of the right colon is the mobilization of the conduit at the beginning of the operation. The vessels that are selected for ligation are occluded with small bulldog clamps while attention is
Ann Thorac Surg 1997;64:752– 6
directed to mobilization of the stomach and esophagus. At the completion of gastric and distal esophageal mobilization, the viability of the colon can be assured before ligation of the arterial branches. Adhering to these intraoperative principles, the right colon can be harvested and employed quite effectively as a conduit. In fact, the results are comparable with those reported in 1988 by DeMeester and colleagues [3], who reconstructed 85 patients with left colon grafts. In their study, 3 of the 85 patients demonstrated intraoperative ischemia. Only seven right colon grafts were employed by this group. Consequently, the current experience at our institution suggests that, at least from a technical standpoint, the right colon, placed in an isoperistaltic fashion, can be employed as effectively as the left colon for reconstruction of the esophagus for benign conditions [4]. When the esophagus is resected for malignant disease, the stomach has been the conduit used most commonly for restoration of gastrointestinal continuity. However, in patients with previous gastric resection, or in those patients who require total gastrectomy, bowel interposition has been used. Recently, Isolauri and colleagues [5] reported the use of colon interposition grafts for restoration of continuity in resection of malignant disease. They retrospectively reviewed 248 patients in whom the diagnosis of squamous cell carcinoma was present in 73%. The left colon was the conduit employed in 54% of patients and the right colon in 27% of patients. The remainder of reconstructions were performed with the transverse colon. The complication rate for this group of patients was 37% with a mortality of 16%. In a similar manner, the colon has been employed in our institution for reconstruction in malignant disease, although less commonly. The right colon (41.7%) was used most often, followed by the left colon (25%). The rate of anastomotic leakage for this subgroup of patients was 4.2% with a mortality of 8.8%. These results are certainly acceptable and comparable to those published for benign disease. Because of the limited number of colonic interposition grafts used for reconstruction for malignant disease in the current study, it is difficult to make any concrete conclusions. The data from this study and published studies, however, suggest that a colon interposition graft is a reasonable alternative when the stomach is not an option for reconstruction. This technique does carry a considerable morbidity and mortality for what is a palliative procedure in most patients. As colonic interposition grafts are employed largely for reconstructions of the esophagus, the jejunum can be used in a variety of circumstances. The jejunal interposition graft, in a Roux-en-Y fashion, is often selected in patients in whom total gastrectomy is planned as part of the operative procedure. The majority of these patients are undergoing resection of malignant disease. On the other hand, reconstruction of the cervical esophagus has seen an evolution of technical advances. In 1942, Wookey [6] described a technique based on the creation of lateral neck flaps. Twenty years later, Bakamjian [7] reported a two-stage reconstruction employing a deltopectoral flap. Despite the improvement these techniques provided
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
755
over the creation of skin tubes, they did require a second-stage procedure and were not as physiologic as other bowel interposition grafts. The era of the free jejunal graft began in 1907 with the microvascular work of Alexis Carrel, who successfully transplanted small intestine into the neck of a dog [8]. Fifty years after this experimental landmark, Seidenberg and colleagues [9] performed the first jejunal free graft in a patient. With further advances in microvascular surgical techniques, especially during the development of myocutaneous free flaps, jejunal transplantation became more commonly employed for reconstruction of the esophagus, especially the pharyngocervical portion. In 1989, Coleman and others [10], from our institution, reported their experience with the cervical jejunal free autograft in 101 patients, 91 of whom were operated on for malignant disease. Sixty-eight percent of the patients with malignant disease underwent synchronous extirpation of the primary tumor. The remainder of this subgroup of patients were reoperative cases for fistula or stricture after laryngectomy or radiotherapy. Thirteen graft failures were reported as well as 33 pharyngocutaneous fistulas. The mortality rate was 5%. They concluded that, despite the relatively high complication rate, free jejunal autografts can be employed effectively for restoration of gastrointestinal continuity. In the current study, four jejunal autografts were performed for reconstruction of the esophagus. Three of the grafts were performed for failed bowel interposition and required free jejunal transfers between the pharynx or cervical esophagus and the salvaged segment of colonic interposition. We have reported these patients previously [11, 12]. Our experience showed that, unlike colonic interpositions, which recorded ineffective motor contractions on manometric studies [4], jejunal grafts showed vigorous peristaltic waves, and isoperistaltic placement of the graft is essential. Admittedly, reconstruction of the esophagus with bowel interposition grafts in this series resulted in a morbidity rate of 38.2%. The most notable characteristic of the patient population of the current review is the significant proportion of patients who had undergone prior thoracic or intraabdominal procedures, further complicating any operative intervention. Considering the fact that more than half of the patients had undergone a mean of three prior operative procedures, an anastomotic leakage rate of 14.8% is acceptable. This rate of anastomotic leakage compares favorably with other published series [3, 5, 10, 13]. In fact, 44% of the leaks occurred in reoperative cases. Most of these were treated nonoperatively with drainage of the leak either by tube thoracostomy or by opening of the cervical wound. Those not amenable to more conservative means of treatment were salvaged with a reoperative bowel interposition graft. This situation in which gastrointestinal discontinuity exists between the proximal esophagus and the distal interposition remnant is particularly suited for the free jejunal autograft. This conduit can be tailored to fill the intervening defect without the need for a second colon interposition graft. The 3 patients in whom this method
756
MANSOUR ET AL BOWEL INTERPOSITION FOR ESOPHAGEAL REPLACEMENT
of salvage was employed experienced no dysphagia, regurgitation, or nutritional difficulties. The mortality rate of 5.9% falls within the range of previously published series. The major cause of death for the population of patients studied was sepsis secondary to nosocomial pneumonia. The last patient mortality occurred in 1989. Since that time, 39 patients have undergone reconstruction for benign and malignant disease without a single mortality. This low mortality, especially in the most recent 7 years, testifies to the improvements in the perioperative management of these very complex patients in the current era of critical care medicine. In conclusion, techniques of bowel interposition reconstruction of the esophagus have evolved to include the jejunum and colon in selected circumstances. Reconstruction employing a colonic conduit is well-suited for reconstruction in patients with benign and, in select cases, malignant disease. The right colon is a suitable conduit and can be harvested as reliably as a left colon conduit if certain principles are followed intraoperatively. Likewise, the jejunal interposition graft is a suitable choice in patients in whom the entire stomach is resected. The free jejunal autograft is a very useful graft, particularly in cases of salvage reconstruction. The microvascular anastomosis for this graft is performed by members of our Division of Plastic and Reconstructive Surgery. Reconstruction with bowel interposition grafts can be performed with a morbidity of approximately 38% and a mortality of approximately 6%, even in a population of patients in which almost half are reoperative procedures. Gastrointestinal continuity can be restored effectively by employing a combination of interposition grafts if necessary. The free jejunal autograft is an excellent choice when primary reconstruction has failed.
Ann Thorac Surg 1997;64:752– 6
References 1. Kelling G. Oesopagoplastik mit Hilfe de Querkolon. Zentralblatt Chir 1911;38:1209–12. 2. Ventemiglia R, Khalil KG, Frazier OH, Mountain CF. The role of preoperative arteriography in colon interposition. J Thorac Cardiovasc Surg 1977;74:98 –104. 3. DeMeester TR, Johansson K-E, Franze I, et al. Indications, surgical technique, and long-term functional results of colon interposition or bypass. Ann Surg 1988;208:460–74. 4. Mansour KA, Hansen HA II, Hersh T, Miller JI, Hatcher CR Jr. Colon interposition for advanced non-malignant esophageal stricture. Ann Thorac Surg 1981;32:584–91. 5. Isolauri J, Markkula H, Autio V. Colon interposition in the treatment of carcinoma of the esophagus and the gastric cardia. Ann Thorac Surg 1987;43:420– 4. 6. Wookey H. The surgical treatment of carcinoma of the pharynx and upper esophagus. Surg Gynecol Obstet 1942; 75:499 –502. 7. Bakamjian VY. A two-stage method for pharyngoesophageal reconstruction with a primary pectoral skin flap. Plast Reconstr Surg 1965;36:173– 84. 8. Paletta CE, Jurkiewicz MJ. Esophageal replacement: microvascular jejunal transplantation. In: Baue AE, Geha AS, Hammond GL, Laks H, Naunheim KS, eds. Glenn’s thoracic and cardiovascular surgery. Stamford, CT: Appleton & Lange, 1996:931– 8. 9. Seidenberg B, Rosenak SS, Hurwitt ES, Som ML. Immediate reconstruction of the cervical esophagus by a revascularized isolated jejunal segment. Ann Surg 1959;149:162– 6. 10. Coleman JJ, Tan K-C, Searles JM, Hester TR, Nahai F. Jejunal free autograft: complications and their resolution. Plast Reconstr Surg 1989;84:589–95. 11. Mansour KA, Picone AL, Coleman JJ III. Surgery for high cervical esophageal carcinoma: experience with 11 patients. Ann Thorac Surg 1990;49:597– 602. 12. Carlson GW, Anderson TM, Galloway JR, Mansour KA. Salvage of colon interposition by antethoracic free jejunal transfer. Ann Thorac Surg 1994;58:1523–5. 13. Neville WE, Najem AZ. Colon replacement of the esophagus for congenital and benign disease. Ann Thorac Surg 1983;36: 626–33.