Colon Interposition for Advanced Nonmalignant Esophageal Stricture: Experience with 40 Patients

Transcription

1 Colon Interposition for Advanced Nonmalignant Esophageal Stricture: Experience with 40 Patients Kamal A. Mansour, M.D., Henry A. Hansen, 11, M.D., Theodore Hersh, M.D., Joseph I. Miller, Jr., M.D., and Charles R. Hatcher, Jr., M.D. ABSTRACT This report details our experience in 40 patients with benign strictures of the esophagus who underwent colon (or gastric) interposition with or without esophageal resection between 1972 and There were 23 men and 17 women ranging from 5 to 76 years old. Twenty-seven patients had fibrotic strictures secondary to reflux esophagitis including 12 after failure of antireflux procedures; 4 had caustic strictures; 3 had a Barrett s esophagus; 2 had systemic candidiasis; 2 had scleroderma; and in 2 the etiology was undetermined. The right colon was used in 27 patients, the left colon in 4, the transverse colon in 1, and the jejunum in 1. In 7 patients the stomach was employed because of vascular insufficiency of the colon. Three surgical approaches were utilized. Manometric studies were done postoperatively in 10 patients. Complications occurred in 7 patients: cervical leaks, 4; reflux colitis, 2; and a late cervical stricture, 1. There were 4 deaths, only 1 of which was related directly to technique. This report summarizes the beneficial effect of right colon interposition for long esophageal strictures. Its vasculature is adequate, and it functions properly to propel food into the stomach irrespective of the peristaltic orientation. Esophageal substitution using a variety of techniques and conduits was utilized initially for bypass of malignant lesions [l-41. With the improvement in surgical and anesthetic techniques and increased experience with esophageal reconstruction, lower operative morbidity and mortality have been achieved. This has led to the use of esophageal substitution for relief of nonmalignant strictures and From the Joseph 8. Whitehead Department of Surgery, Division of Cardio-Thoracic Surgery, and the Department of Medicine, Division of Digestive Diseases, Emory University School of Medicine, Atlanta, GA. Presented at the Seventeenth Annual Meeting of The Society of Thoracic Surgeons, Jan 26-28,1981, Los Angeles, CA. Address reprint requests to Dr. Mansour, Emory University Clinic, 1365 Clifton Road NE, Atlanta, GA their attendant nutritional, pulmonary, and psychological complications. The colon [S] has become a favored organ for interposition because of the ease of isolation of a segment and its concomitant vasculature with the marginal artery of Drummond. In 1911, Kelling [91 used the isoperistaltic transverse colon for esophageal substitution, and Vulliet [lo] also described the use of the transverse colon but placed it in an antiperistaltic position. Roith 1111, in 1923, employed the right colon in an isoperistaltic fashion. In 1950, Orsini and Toupet [121 transposed a left colon segment. Four years later, Mahoney and Sherman [11 stimulated interest in the right colon and reported a case of retrosternal colon transplant for reconstruction of the thoracic esophagus. Belsey [13], in 1965, reported his large experience with left colon transplants in the management of irreversible esophageal obstructive lesions. In 1971, Postlethwait and co-workers [141 reported their experience with 17 patients with benign intractable esophageal strictures: the right colon was used in 14 patients and the left, in 3. The purpose of this paper is to review our experience with right colon interposition for benign esophageal strictures of the esophagus. Material and Methods The 40 patients in this study, who were classified as medical failures, underwent colon (mainly right) or gastric interposition with or without esophageal resection between January, 1972, and July, There were 23 men and 17 women ranging from 5 to 76 years old. Twenty-seven patients had end-stage fibrotic stricture secondary to reflux esophagitis including 12 who had failure of primary antireflux procedures. Four patients had caustic strictures, 3 had a Barrett s stricture, 2 had systemic candidiasis, and 2 had scleroderma. In 2 patients the etiology was undetermined. The right colon was used as the interposed by The Society of Thoracic Surgeons

2 585 Mansour et al: Colon Interposition for Nonmalignant Esophageal Stricture Fig I. Chest roentgenogram showing feeding tube in subcutaneous stomach and decompression tube in cervical esophagus. segment in 27 patients, the left colon in 4, the transverse colon in 1, and the jejunum in 1. The right colon was inspected first in all patients, and if the vascularity appeared insufficient during temporary occlusion, another segment of the gut was used. Isoperistaltic segment was used in 24 patients and antiperistaltic segment in 9. In 7 patients the stomach was employed because of vascular insufficiency of the proposed segment of the bowel. In 1 patient the stomach would not reach the neck, and the gap between the pharynx and subcutaneous stomach was bridged using an antethoracic deltopectoral flap (Figs 1, 2). Preoperative Preparation Barium enema was performed in all patients to ascertain the absence of anatomical abnormalities or pathological changes of the colon. Bowel preparation consisted of three days of appropriate diet and mechanical cleansing followed by oral administration of neomycin or erythromycin in divided doses on the day before operation. No patients underwent arteriography or colonoscopic evaluation. Fig 2. Barium swallow study shows reconstruction of the esophagus using a skin-lined deltopectoral flap. Technical Considerations Endotracheal general anesthesia and one-lung ventilation was used when the esophagocolic anastomosis was performed in the chest. A nasogastric tube was introduced as far as the esophageal stricture permitted. The procedure was performed with the patient in the supine position with the chest elevated 45 degrees. Eighteen patients underwent a combined abdominal and cervical approach for longsegment colon interposition. The strictured esophagus was left in all patients except 1 who had sustained multiple perforations during previous dilations and required total esophagectomy at an earlier stage. Sixteen patients underwent combined abdominal and right chest approach, with resection of the diseased esophagus in all patients except 3 who had had previous Thal-Nissen procedures. In 6 pa-

3 586 The Annals of Thoracic Surgery Vol 32 No 6 December 1981 tients, a left thoracoabdominal approach for short-segment colon interposition was used, and the diseased segment of the esophagus was removed in all patients. A retrosternal tunnel was used when total bypass of the esophagus was contemplated, and the ileum usually was anastomosed to the esophagus. A transhiatal, posterior mediastinal tunneling was used for short-segment colon interposition, and the ileum, cecum, or ascending colon was anastomosed to the esophagus. After the right colon was freed, the mesenteric vessels were examined by transillumination, and the ileocolic artery was test clamped using two vascular clamps near the root of the mesentery. If the blood supply from the right branch of the middle colic and right colic arteries was adequate, the ileocolic artery was doubly ligated and divided, and the colon was passed behind the stomach and through the gastrohepatic ligament either in a retrosternal tunnel or through the esophageal hiatus. Anastomosis was performed usually with two layers of 3-0 silk sutures, an inner mucosal layer with the knots tied on the inside and an outer seromuscular layer with the knots tied on the outside. If the blood supply through the ileocolic artery was predominant, the right branch of the middle colic and the right colic arteries were sacrificed and the right colon was used in an antiperistaltic fashion. If the right colon was not available, the left colon was selected depending on the blood supply from the left branch of the middle colic artery or the left colic artery. If the blood supply to either colon was not satisfactory, the stomach was used as the last choice. Tube gastrostomy was employed in all patients, and pyloroplasty was performed only when esophageal resection was carried out. Results Thirty-six patients survived the operation and had good results with respect to swallowing, digestion, and physical activity. Complications occurred in 7 patients. In 4 patients, a cervical leak developed at the site of the esophagoileal anastomosis but all healed spontaneously within three weeks. Four years postoperatively, a late stricture developed at the cervical anastomosis in 1 patient and required revision. Reflux colitis developed in 2 patients, but resolved with cimetidine and antacid therapy. There has been no instance of colonic necrosis or late fibrosis due to ischemia. There were 4 operative deaths. One was related to leakage of the transplanted colon in the right chest at the site of an anchoring suture to the diaphragmatic hiatus. Sepsis and fatal pulmonary hemorrhage occurred three weeks after operation. One patient died on the eighth postoperative day of massive gastrointestinal hemorrhage related to a retained Nissen gastric wrap in the chest. Another patient died on the tenth postoperative day of an acute myocardial infarction, and 1 patient died at home of massive pulmonary embolism one month after operation. Colon Manometry Manometric studies of emptying of the implanted colon were performed postoperatively in 10 patients three or more months after the surgical intervention [15]. Manometry was performed with four water-filled polyvinyl tubes that had openings 5 cm apart and that were constantly perfused with microquantities of water. Intraluminal pressure was transmitted to external transducers, and pressures were graphed on a multichannel curvilinear inkwriting recorder. As in standard esophageal manometry, the pressures in the interposed segment of the colon were recorded with the patient in the supine position and starting with the distal opening at 40 cm from the incisor teeth. Pneumographs around the chest and neck monitored respiration and swallowing, respectively. A recording was taken first in the basal state without swallow for a period of one hour. Subsequently, with the catheter assembly positioned in the same place, the patient was given swallows of ice water. Usual colonic contractions were recorded in the basal tracing. In each instance, an occasional high-pressure contraction was present, which appeared subsequently in each lead (Fig 3). In none of the patients studied were contractions present that were related to the act of

4 587 Mansour et al: Colon Interposition for Nonmalignant Esophageal Stricture Contraction Contraction 11) ion CmjCont ract RESTING TRACING L--L^urc- -. ContractionL Fig 3. Tracing of colon manometry at resting state showing occasional colonic contractions. swallowing water. The tracing in each patient remained as it had been in the resting state, independent of swallows (Fig 4). Comment The conservative treatment of extensive benign esophageal strictures is limited to repeated dilations [ Esophageal substitution is the treatment of choice for patients considered "medical failures" because the strictures are refractory to dilation. The colon is the most versatile organ for transplantation, and its advantages have been described in previous reports [2, 3, 6, 81. Colon interposition is an operation of considerable magnitude requiring appropriate selection of patients and meticulous sur- Fig 4. Recording of colon manometry showing colonic contractions unrelated to swallows of ice water. gical technique. Most authors [4, 12, 131 have advocated the use of the left colon, but we have utilized the right colon in an isoperistaltic fashion whenever possible. Acceptable morbidity and mortalilty can be achieved, as demonstrated by our series, which parallels the improvement of surgical and anesthetic techniques and our familiarity with the procedure. Recently we extended our indications for right colon interposition to bypass a small atonic and obstructed gastric remnant in a patient who had had multiple procedures for gastric resections, intestinal obstruction, and fecal fistulas (Fig 5). Leaks from cervical anastomosis are the most frequent complication [B, 141. They occur in about 25% of patients and frequently lead to development of late strictures. Mediastinitis and death from cervical leakage have been reported [20], but this has not been our experience. At initial operation, all cervical anastomoses are drained and subsequently treated conservatively by opening the drainage tract; 35 cm Colonic contraction II i i

5 588 The Annals of Thoracic Surgery Vol 32 No 6 December 1981 A B Fig 5. (A) Preoperative flat film of abdomen two weeks after barium swallow showed retained obstructed gastric remnant and internal fistula formation. (B) Postoperative barium swallow showed right colon interposition bekueen the thoracic esophagus and proximal jejunum. feeding is maintained through gastrostomy. Interestingly, all cervical anastomotic leaks were at the site of the esophagoileal anastomoses; no leaks occurred at the esophagocolic anastomotic lines. Strictures have been reported in 12% of patients [51, most often after anastomotic leaks, and dilation of these patients may be difficult and hazardous. If any nonpliable obstruction is encountered, revision is the treatment of choice. Reflux colitis occurred in 2 patients in our series (5%) and responded favorably to cimetidine and antacid therapy. A huge ulcer developed in 1 patient, but has not recurred. Fixation of the colon conduit to the diaphragm has been described as an integral part of esophageal substitution [7]. However, in our experience, 1 patient died secondary to disruption of an anchoring stitch. We believe the problem is kinetic rather than technical, and we no longer attach the colon, a relatively fixed organ, to the diaphragm, which is a vigorously contractile structure. We encountered no instances of diaphragmatic herniation in this series. Colon bypass for caustic strictures of the esophagus was performed without removal of the strictured esophagus. Although there may be some increased risk of malignant degeneration within the scarred esophagus, it is not great enough to warrant esophagectomy L21, 221. The reported cases of esophageal cancer occurred in corrosive strictures of the esophagus that were repeatedly dilated but never bypassed [23, 241. However, in cases of stricture due to reflux esophagitis, the diseased segment of the esophagus should be removed because of hazardous complications that might occur in the retained segment and particularly when the diagnosis of carcinoma cannot be ruled out. References Mahoney EB, Sherman CD: Total esophagoplasty using intrathoracic right colon. Surgery 35:937, 1954 May IA, Samson PC: Esophageal reconstruction and replacement (collective review). Ann Thorac Surg 7:249, 1969 Wilkins EW, Burke JF: Colon esophageal bypass. Am J Surg 129:394, 1975 Skinner DB: Esophageal reconstruction. Am J Surg 139:810, 1980

6 589 Mansour et al: Colon Interposition for Nonmalignant Esophageal Stricture 5. Hong PW, See1 DJ, Dietrick RB: The use of colon in the repair of benign stricture of the esophagus. Pacif Med Surg 75:148, Orringer MB, Kirsh MM, Sloan H: Esophageal reconstruction for benign disease: technical considerations. J Thorac Cardiovasc Surg 73:807, Gross RE, Firestone FN: Colonic reconstruction of the esophagus in infants and children. Surgery 61:955, Shackelford RT: Surgery of the Alimentary Tract. Second edition. Philadelphia, Saunders, 1978, vol 1 9. Kelling G: Osophagoplastik mit Hilfe des Querkolon. Zentralbl Chir 38:1209, Vulliet H: De l esophagoplastie et des diverses modifications. Sem Med 31:529, Roith 0: Die einzeitige antethorakale Oesophagoplastik aus dem Dickdarm. Dtsch Z Chir 183:419, Orsini P, Toupet A: Utilization of descending colon and left portion of transverse colon in prethoracic esophagoplasty. Presse Med 58:804, Belsey R: Reconstruction of the esophagus with left colon. J Thorac Cardiovasc Surg 49:33, Postlethwait RW, Sealy WC, Dillon ML, Young WG: Colon interposition for esophageal substitution (current review). Ann Thorac Surg 12:89, Clark J, Moraldi H, et al: Functional evaluation of the interposed colon as an esophageal substitute. Ann Surg 183:93, Benedict EB, Nardi GL: The Esophagus. Boston, Little, Brown, Bayless TM: Management of esophageal disease. Mod Treatm 7:1081, Burkhart KL, Sullivan BH: Course and treatment of benign esophageal strictures. Am J Gastroenterol 58:531, Kongtahworn C, Rossi NP: Dilatation for severe esophageal stricture. Ann Thorac Surg 14:678, Reichle R: Suppurative mediastinitis as a late complication following antethoracic esophagoplasty. (Eitrige mediastinitis als Spaetkomplikation nach antethorakaler Oesophaguplastic.) Thoraxchirurgie 4:471, Joske RA, Benedict EB: The role of benign esophageal obstruction in the development of carcinoma of the esophagus. Gastroenterology 36:749, Kiviranta UK: Corrosion carcinoma of the esophagus. Acta Otolaryngol42:89, Carver CM, Sealy WC, Dillon JJ: Management of alkali burns of the esophagus. JAMA 160:447, Marchand P: Caustic strictures of the esophagus. Thorax 10:171, 1956 Discussion DR. DAVID B. SKINNER (Chicago, IL): I greatly enjoyed this fine presentation by Dr. Mansour and appreciated the opportunity to study his manuscript. It stimulated Dr. DeMeester and me to review our experience with esophageal replacement for benign disease in 77 patients. This serves as the basis for my comments. First, I have a few remarks about the indications for operation. Colon interposition is a very large operation with a significant risk, as Dr. Mansour s series indicates, and it should be reserved only for those patients who cannot be treated by lesser operations. There were 27 patients in the series who had reflux-induced strictures but only 12 of them had undergone previous antireflux operations. In our experience, the vast majority of patients with reflux strictures can be treated by intensive medical therapy, dilation, and then one of the several effective antireflux operations. Less than 10% of such patients ultimately require resection or interposition. I wonder if some of the patients in the present series could have been treated by lesser operations. Second, what can we learn from the 4 deaths in this series? The mortality of 10% is rather high. A number of reports in the literature and our own experience indicate mortality should be less than 5%. In fact, I have had no hospital death in 40 patients undergoing resection and interposition for benign reflux strictures. One of the deaths in Dr. Mansour s series was due to hemorrhage from a retained esophageal segment. If the esophageal disease is so severe that an interposition is required, then the disease process should be resected to avoid further complications of the disease. This should be standard practice. One death was caused by a technical error in the anchoring of the colon segment to the hiatus. This is not a reason to abandon this important step in the operation, because failure to anchor the colon to the hiatus may result in herniation of the small bowel into the chest or upward migration of the colon leading to redundancy and a sink trap effect, which prevents effective colon emptying. The solution is careful suturing of only the seromuscular layer to the hiatus. The death from myocardial infarction indicates the overlap between chest pain caused by both esophageal disease and coronary disease. In patients with chest pain, a full coronary workup is indicated prior to the esophageal operation. If coronary disease is present, then a coronary operation should take precedence. Finally, the death due to pulmonary embolism might have been obviated by the use of prophylactic anticoagulants. We use them routinely in all of our patients undergoing thoracic operations. My final comments concern the choice of organ for esophageal replacements. As we reported to this So-

7 590 The Annals of Thoracic Surgery Vol 32 No 6 December 1981 ciety nearly ten years ago, these colon segments do experience peristalsis, as Dr. Mansour confirmed again today, so there should be isoperistalsis. We prefer the left colon not only because of its accessibility through a left thoracotomy, but because it has a consistently reliable arterial blood supply. We have seen only 1 patient in whom the arterial supply was inadequate to do a left colon interposition. This is in contrast apparently to 10 patients in this series in whom the blood supply to the right colon proved inadequate. Finally, the left colon has the advantage of being nearly the same diameter as the esophagus for an anastomosis. The Atlanta group and others have used the ileum to get around the problem of size discrepancy, but the normal function of the ileocecal valve is to impede the emptying of the ileum into the colon. I have seen 1 patient in whom the interposed ileocecal valve proved completely obstructive to solid boluses of swallowed food, and resection and reanastomosis were necessary. All four leaks in this series occurred when the ileum had been anastomosed to the esophagus. I wonder if the partially obstructing function of the ileocecal valve might have contributed to this incidence of leakage. DR. PAUL A. KIRSCHNER (New York, NY): I enjoyed this paper, but I wonder why more consideration is not being given to use of the stomach. I noted that the authors used it in 7 patients. Recently, I used the stomach in a patient with no pulsatile circulation in the arcades of the colon or middle colic artery. The patient had extensive arterial disease including previous myocardial infarction and stroke. It was easy to mobilize the stomach, divide the esophagogastric junction, and bring the stomach up to the neck substernally where it was anastomosed to the cervical esophagus. Thus, only one anastomosis was required, which healed promptly, instead of three when the colon interposition is done. On other occasions a side-to-side anastomosis in continuity, between the stomach and the esophagus proximal to the stricture in the chesteven as high as the apex of the chest-can be done again with only one anastomosis. This technique can also be used to bypass an unresectable carcinoma of the esophagus. I wonder whether Dr. Mansour has noted any leaks when the stomach was used. It is my impression that leakage would be much less frequent than when the colon was used, because of the better blood supply of the stomach. DR. ROBERT w. RIEMER (Providence, RI): In our experience with this interposition, we used a substernal tunnel in an effort to avoid a leak into a pleural cavity, which is a most disastrous complication of this procedure. Another consideration is a study of the bowel before it is used. We routinely do a barium enema to rule out any particular disease in the colon. I was wondering whether we should go one step further and do abdominal angiography to make sure that the proper blood supply is available for using either the descending or ascending colon. DR. PHILIP w. WRIGHT (Long Beach, CA): I appreciated the presentation and wonder if the authors would comment upon two points. First, there are reports on the incidence of carcinoma that develops in the retained esophagus with benign chronic stricture. Is this a concern when operation is done in young individuals, as in the present series? Second, could the limitation of vascular supply that Dr. Mansour described be due to inadequate length of the right ileocolic vessels or perhaps to some other arcade deficiency. The latter would seem to be one of the shortcomings of right colon interposition. DR. MANSOUR: I am delighted to have my paper discussed by Dr. Skinner, and I appreciate the cornments of the other discussants. I will address myself to the last remark about the incidence of carcinoma. If you review the literature, the incidence of carcinoma was reported in patients who had dilations and in whom the esophagus was not bypassed. A paper from Finland reported a 1,000-fold higher incidence of carcinoma in patients with corrosive strictures of the esophagus, but these strictures were not bypassed [l]. The authors blamed it on "Virchow's irritation theory"; we believe repeated dilations could be an etiological factor. Concerning the blood supply, we use the right colon first. If the blood supply is not good, we use the left colon, and if not the left colon, we use the stomach. The blood supply was not sufficient in that the arcades were interrupted; the incidence of this is 5% in the literature. Shiu and Ong [2], in a study of 250 colons taken postmortem, found that 5% of the colons had inadequate arterial blood supply. Dr. Riemer, one of your comments concerned the use of a substernal tunnel, not in the chest. We use substernal reconstruction when we have decided on a long segment of colon interposition for caustic strictures. However, for ulcerative esophagitis, especially when carcinoma cannot be ruled out, it would be a dangerous undertaking to ignore the disease process and just bypass it in the neck. As for your other comment, we do study the colon preoperatively, mainly with barium enema. We do not do routine angiography nor do we do colonoscopy. Dr. Kirschner, you asked why the stomach is not used more frequently. We should not use the stomach as the primary organ for bypass in patients with benign disease. This does not include patients with cancer. For cancer of the esophagus, I agree with Dr. Kirschner; we use the stomach, as does al-

8 591 Mansour et al: Colon Interposition for Nonmalignant Esophageal Stricture most everybody. However, for benign disease, we prefer not to, because of the dreaded complications of reflux esophagitis. Dr. Skinner s remarks are well taken. The indications, as you noticed, are end-stage fibrous strictures of the esophagus whether the treatment was delayed or whether antireflux procedures were not done in time. The esophagi were burned out, so dilation was not indicated. If dilation had been performed, it would have failed definitely. The mortality is not lo%, Dr. Skinner. I am talking about mortality related to the technique, or 1 death in 40 patients. This is a mortality of 2.5%. Only 1 patient died as a result of the technique. One patient died of a pulmonary embolus at home, and the report of the postmortem examination came from another state. One patient died of an acute myocardial infarction, which developed on the tenth postoperative day. Why didn t we remove the distal esophagus in 3 patients? We did remove it in 13 out of 16, but in the 3 who had had previous Thal-Nissen procedures, we went into the right chest to avoid adhesions and extensive scarring. As far as using the ileum is concerned, there were no problems with swallowing, and we had no difficulty putting the ileum in. I agree that the incidence of leak is higher with ileal anastomoses than colonic anastomoses, but there is no trouble with emptying of the ileum. References 1. Kiviranta UK: Corrosion carcinoma of the esophagus. Acta Otolaryngol42:89, Shiu MH, Ong GB: Blood supply of the colon in relation to oesophageal replacement: post-mortem study of 250 autopsy cases. Presented at the 3rd Malaysian Congress of Medicine, Kuala Lumpur, Malaysia, August 11-13, 1967