J
THORAC CARDIOVASC SURG
85:330-336, 1983
Original Communications
One-stage reconstruction for pharyngolaryngectomy Esophagectomy and pharyngogastrostomy without thoracotomy From 1969 to 1981, a total of 22 patients underwent laryngopharyngectomy and nonthoracotomy esophagectomy; with immediate pharyngogastrostomy, for hypopharyngeal or postcricoid carcinoma. Thirteen initially had been treated by high-dose radiotherapy, but the tumor had either persisted or recurred. Four patients underwent planned preoperative irradiation on the morning of the operation. Two patients had had previous high-dose local irradiation to the neck for other disease, and three patients had no irradiation. There was one operative death. Anastomotic leaks developed in four patients, but only one of the leaks was considered a serious problem. Three patients had transient dysphagia, but only one required dilatation. Transient delayed gastric emptying was a problem in three other patients. The average postoperative stay was 31 days, with 38% of patients being discharged by 21 days. All patients were discharged eating a normal diet. Fifty percent survived longer than 12 months, with an actuarial survival rate of 30% at 5 years. The patient surviving longest is disease free at 12 years. Palliation was considered excellent in all 21 operative survivors. Immediate pharyngogastrostomy via nonthoracotomy esophagectomy is a safe and excellent means of palliation in this group of patients, for whom palliation is often the only option.
D, W, O. Moores, M.D., F.R.C.S.(C), R. lIves, M.D., F.R.C.S.(C), 1. D. Cooper, M.D., F.R.C.S.(C), T. R. J. Todd, M.D., F.R.C.S.(C), and F. G. Pearson, M.D., F.R.C.S.(C), Toronto, Ontario, Canada
Reconstruction with restoration of swallowing following laryngopharyngectomy is a challenging surgical problem. An adequate procedure to ensure early postoperative feeding and early hospital discharge is essential. The reconstruction techniques currently in use fall into two categories: those employing local procedures
and those involving visceral interposition. Development of deltopectoral' and myocutaneousv ' flaps is the local procedure most commonly used in current practice. Visceral interposition options include mobilization of the whole stomach.v " use of gastric tubes.!" a Roux-en-Y jejunal 100p,!1 colonic interposition, 11, 12 and free jejunal or colonic 100p13-15 with microvascular anastomosis.
From the Division of Thoracic Surgery, Toronto General Hospital, and the Department of Surgery, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada.
Patients
Read at the Eighth Annual Meeting of The Samson Thoracic Surgical Society, San Diego, Calif., June 23-27, 1982. Address for reprints: Dr. R. lives, Toronto General Hospital, Eaton Building North 10-230, Toronto, Ontario, Canada M5G IL7.
330
Between 1969 and 1981 at Toronto General Hospital, 22 patients underwent laryngopharyngectomy with primary pharyngogastrostomy for hypopharyngeal" or postcricoid" carcinoma. This was accomplished by means of whole stomach interposition via the posterior
Volume 85
Number 3 March, 1983
Pharyngolaryngectomy
331
Fig. 1. View intotheesophageal hiatus with retractors in place,showing application of a clip in the mediastinum. mediastinum following thoracic esophagectomy without thoracotomy. There were 12 men and 10 women ranging in age from 35 to 74 years, mean 58 years. Twenty patients had squamous cell carcinoma and two had adenocarcinoma. Thirteen patients were initially treated with high-dose radiotherapy, but the tumor either persisted or recurred. Four patients received deliberate preoperative radiotherapy (700 rads the morning of operation). Two patients had previously been treated with highdose local radiotherapy to the neck for other unrelated diseases, and in three patients no radiotherapy was given.
Operative procedure The operation is performed by two operating teams. An en bloc laryngopharyngectomy and neck dissection, if indicated, are done first. Once resectability is determined, the abdominal part of the operation is undertaken by a second team. The abdomen is opened through a long upper midline incision. If the stomach and duodenum have no abnormalities, the stomach is mobilized with the right gastroepiploic and right gastric arteries left intact. The duodenum is mobilized by the Kocher maneuver. At this stage the Over Hand Retractor* is put into place. The blades of the retractor are secured under the costal margins to retract the ribs in an upward and outward direction. The esophageal hiatus is *J. Hugh Knight Instrument Co., New Orleans, La.
now mobilized and a Penrose drain is passed around the gastroesophageal junction as a sling. The hiatus is enlarged circumferentially by manual stretching or by cutting the diaphragm anteriorly toward the pericardium. Maximum exposure is obtained with the Over Hand Retractor in place, the hooked back end of a narrow Dever retractor holding the hiatus open, and downward retraction on the esophagus by the Penrose sling (Fig. 1). With this exposure it is possible to see well up into the posterior mediastinum to the level of the carina. The pleural surfaces are visible laterally, the aorta posteriorly, and the posterior pericardium anteriorly. All esophageal vessels are then divided under direct vision with either surgical steel clips or electrocoagulation. Following completion of the laryngopharyngectomy, dissection of the cervical esophagus is begun from above. Because of the wide surgical field provided by the laryngopharyngectomy, it is possible to see down into the posterior mediastinum almost to the level of the carina. The esophagus is dissected free from the trachea and its lateral support structures, and again all vessels are divided under direct vision. Extreme care is used to preserve the tracheal blood supply; therefore, no lateral dissection is carried out anteriorly at the level of the trachea. A hand is passed up into the posterior mediastinum from below and down into the posterior mediastinum from above, to facilitate mobilization of the esophagus. In 19 of 22 cases in this series, the esophagus was
The Journal of
332 Moores et al.
Fig. 2. Lateral chest x-ray film with surgical clips indicating direct visualization along the entire esophageal bed.
completely mobilized by this method. The stomach can then be drawn into the neck by applying traction on the esophagus and feeding the stomach manually through the hiatus. In the lateral x-ray film (Fig. 2), the extent of direct visualization is shown by the placement of surgical clips in the posterior mediastinum. The gastroesophageal junction is then divided with the GIA surgical stapling device* and the staple line is oversewn with an absorbable running suture. If the esophagus cannot be fully mobilized under direct vision, then the method of esophageal extraction described by Akiyama, Hiyama, and Miyazono!" may be employed, with a vein stripper used for eversion extraction. This was performed uneventfully three times. In these cases, a gum-tipped bougie is passed from above downward to the surgeon's hand, which is passed through the hiatus into the posterior mediastinum. The end of the bougie is guided down into the abdomen. A figure-of-eight suture is placed at the high point of the gastric fundus and tied around the bougie. The stomach is drawn up through the posterior mediastinum into the neck with *Auto Suture Company Division, United States Surgical Corporation, Norwalk, Conn.
Thoracic and Cardiovascular Surgery
traction on the bougie and manual propulsion through the hiatus. With delivery of the gastric fundus into the neck, extreme care is taken to ensure proper axial alignment of the stomach as it traverses the posterior mediastinum. An end-to-side two-layer pharyngogastrostomy is then performed at the level of the highest point of the greater curve (Fig. 3). The outer layer of sutures incorporates the muscular wall and submucosa, with the inner layer being full thickness. Following completion of the anastomosis, the stomach is fixed in position with several sutures between the posterior wall of the stomach and the anterior longitudinal spinal ligament to relieve tension on the anastomosis. A pyloric drainage procedure, either pyloroplasty or pyloromyotomy, was performed in 11 cases. Initially this was done routinely, but subsequent review showed no difference in delayed gastric emptying whether or not a drainage procedure had been used. Now a drainage procedure is employed if duodenal disease is believed to exist. Thus in 11 cases the pylorus was not divided. A feeding jejunostomy was performed in the majority of cases prior to abdominal closure. The esophagus removed in this operation is a normal esophagus. The surgeon, therefore, is not concerned with wide lymph node dissection, and there are no inflammatory adhesions between the esophagus and the mediastinal structures. Results
Complete follow-up is available on 21 patients. One patient was lost to follow-up at 5 months. There was one operative death (4.5%) from unrecognized hypotension and acidosis, resulting in cardiac arrest. Anastomotic leak occurred in four patients (19%). In only one patient was the leak considered a serious problem necessitating reoperation with revision and closure. The other three fistulas closed spontaneously during the first week. The patients having a small fistula continued to take clear fluids orally and received supplementary nutrition with jejeunostomy feedings. Pneumonia developed postoperatively in three patients but responded rapidly to intravenous antibiotics. Transient dysphagia occurred in three patients, but only one required a single dilatation. Delayed gastric emptying was a transient problem in three patients. All responded well to conservative measure, one having peroral dilatation of the pylorus with a Maloney bougie and the other two treatment with metoclopramide. One of the patients having delayed gastric emptying had had a pyloroplasty performed at the time of operation; the other two patients had no pyloric drainage procedure. Mild transient
Volume 85 Number 3
Pharyngolaryngectomy
March, 1983
333
short' gastries
Fig. 3. The stomach is mobilized with the right gastric and gastroepiploic vessels left intact. The anastomosis is performed at point A. (Published with permission of the University of Toronto.)
jaundice, which settled rapidly, developed in two patients. The only long-term complication was mild intermittent regurgitation of gastric content, occurring in seven patients (33%) when lying down or bending forward after eating. This was easily managed and not a significant disability in any patient. These patients all have been advised to sleep with the head of the bed permanently elevated on 6 inch blocks, since regurgitation is almost universal following this reconstruction. Postoperative hospital stay ranged from 16 to 86 days, with a mean of 31 days. Eight patients (38%) were discharged in less than 21 days. All patients at discharge were eating a normal diet. Of the 20 patients available for follow-up, 10 (50%) survived for longer than 1 year and three for longer than 4 years. Seven patients are presently alive and well 5 months to 12 years postoperatively. Mean survival time is 22 months, and actuarial survival rate is 30% at 5 years (Fig. 4). Cause of death was tumor related in 10 of the 13 patients, seven deaths resulting from local recurrence and three from metastatic disease. Discussion
Local procedures for reconstruction following laryngopharyngectomy have serious limitations. Use of the deltopectoral flap is associated with a high incidence of stricture and fistula formation. 7, 17 The report by Fredrickson, Wagenfeld, and Pearson 7 from this institution stresses that gastric interposition significantly shortens hospital stay and time from operation to swallowing. Subsequent hospitalizations are minimized and the quality of swallowing is much superior in the gastric interposition group. Reconstruction with a myocutane-
100 90 80 70 60 % SURVIVAL
50 40 30 20 10
YEARS
Fig. 4. Actual survival (solid line) and actuarial survival (broken line).
ous flap is more predictable than with a deltoplectoral flap and is effective for partial pharyngeal reconstruction. However, the flap may be too bulky to be used for total pharyngeal replacement. 2 Mobilization of the whole stomach through the posterior mediastinal route has the following advantages: 1. The right gastroepiploic artery provides an excellent and predictable blood supply for the mobilized stomach. 2. The normal mobilized stomach will extend even to the nasopharynx without excessive tension, in all cases. 3. The reconstruction requires only one anastomosis. 4. The width of the gastric fundus is adequate to accommodate the wide diameter of the divided phar-
The Journal of
334 Moores et at.
Thoracic and Cardiovascular Surgery
less satisfactory than in cases of hypopharyngeal carcinoma. This will be the subject of a future report. During the time of this study, both deltopectoral and myocutaneous flaps have been used by the head and neck surgeons. However, gastric interposition is usually the procedure of choice for reconstruction following pharyngolaryngectomy for carcinoma of the hypopharynx. This is justified by the low mortality and morbidity achieved and by the excellent cosmetic and functional result. . We would like to acknowledge the surgical expertise of those surgeons performing the head and neck portion of these operations. From the Department of Otolaryngology, Toronto General Hospital: Dr. D. P. Bryce, Dr. J. M. Fredrickson, and Dr. R. E. Hayden. From the Department of General Surgery, Toronto General Hospital: Dr. J. A. Palmer.
2
3
Fig. 5. Postoperative appearance at I year. ynx; therefore, the anastomosis is performed without the need to narrow the pharyngeal lumen. 5. The two surgical team approach shortens operative time. 6. The cosmetic appearance is good (Fig. 5). Nonthoracotomy esophagectomy with gastric interposition following pharyngolaryngectomy was first described by Le Quesne and Ranger" and is now used by others. 6-8 This initial report as well as others described the esophagectomy as a blind operation performed by blunt dissection. With the technique outlined in this report, the esophagectomy is almost always performed under direct vision, with minimal blood loss and complications. This method also has been employed for carcinoma of the cervical or high thoracic esophagus following irradiation failure. In several of these cases, major and lethal complications have occurred with tracheal necrosis and fistula formation, presumably secondary to radiation effects. The results of such operations after heavy radiation to the mediastinum are significantly
4
5 6 7
8 9
10 II 12
REFERENCES Bakmjian VY: A two-stage method for pharyngoesophageal reconstruction with a primary pectoral skin flap. Plast Reconstr Surg 36: 173-184, 1965 Theorgaraj OS, Merritt WH, Acharya G, Cohen K: The pectoralis major musculocutaneous island flap in singlestage reconstruction of the pharyngoesophageal region. Plast Reconstr Surg 65:267-276, 1980 Strawberry CW, de Fries HO, Deeb ZE: Reconstruction of the hypopharynx and cervical esophagus with bilateral pectoralis major myocutaneous flaps. Head Neck Surg 4:161-164, 1981 Withers EH, Franklin JD, Madden11, Lynch JB: Immediate reconstruction of the pharynx and cervical esophagus with the pectoralis major myocutaneous flap following laryngopharyngectomy. Plast Reconstr Surg 68:898-904, 1981 Le Quesne LP, Ranger 0: Pharyngolaryngectomy, with immediate pharyngogastric anastomosis. Br J Surg 53: 105-109, 1966 Harrison OF: Surgical management of hypopharyngeal cancer. Arch Otolaryngol105:149-152, 1979 Fredrickson JM, Wagenfeld DJH, Pearson FG: Gastric pull-up vs deltopectoral flap for reconstruction of the cervical esophagus. Arch Otolaryngol 107:613-616, 1981 Orringer MB, Sloan H: Esophagectomy without thoracotomy. J THORAC CARDIOVASC SURG 76:643-654, 1978 Ong GB, Lee TC: Pharyngogastric anastomosis after esophago-pharyngectomy for carcinoma of the hypopharynx and cervical esophagus. Br J Surg 48: 193-200, 1960 Heimlich HJ: Carcinoma of the cervical esophagus. J THORAC CARDIOVASC SURG 59:309-318, 1970 Hanna EA, Harrison AW, Derrick JR: Long-term results of visceral esophageal substitutes. Ann Thorac Surg 3:111-118,1967 Goligher JC, Robin IA: Use of left colon for reconstruc-
Volume 85
Pharyngolaryngectomy
Number 3 March, 1983
13
14
15
16
17
18
tion of pharynx and oesophagus after pharyngectomy, Br J Surg 42:282-290, 1954 Nakayama K, Tamiya T, Yamamoto K, Akimoto S: A simple new apparatus for small vessel anastomosis (free autograft of the sigmoid included). Surgery 52:918-931, 1962 May JW Jr, Bunting PO: Free jejunal transfer for irradiated esophageal reconstruction, Reconstructive Microsurgery, Boston, 1977, Little, Brown & Company McKee OM, Peters CR: Reconstruction of the hypopharynx and cervical esophagus with a microvascular jejunal transplant, Microsurgical Composite Tissue Transplantation, St. Louis, 1979, The C. V. Mosby Company Akiyama H, Hiyama M, Miyazono H: Total esophageal reconstruction after extraction of the esophagus. Ann Surg 182:547-552, 1975 Carpenter RJ, DeSanto LW, Devine KD: Reconstruction after total laryngopharyngectomy . Arch Otolaryngol 105: 417-422, 1979 Steiger Z, Wilson RF: Comparison of the results of esophagectomy with and without a thoracotomy. Surg Gynecol Obstet 153:653-656, 1981
Discussion DR. VICTOR RICHARDS San Francisco. Calif.
Carcinomas of the upper cervical esophagus, with or without laryngeal involvement, are among the most difficult and unsatisfactory cancers to treat. Surgical resection in the past has generally been a complex, tedious, and unsatisfactory multistaged operation with extremely poor results in terms of both palliation and survival. Radiation therapy alone has been equally unsatisfactory, and even multimodality chemotherapy by systemic or regional perfusion has rarely improved the quality of life during the short period of survival. This report presents the results of a single-stage nonthoracotomy esophagectomy with immediate pharyngogastrostomy following laryngopharyngectomy for carcinoma of the hypopharynx in 22 patients. Thirteen had been initially treated by irradiation without success, yet the option of preoperative radiation may be employed. There was only one operative death, three patients had an insignificant anastomotic leak which closed spontaneously, and only one patient had a serious cervical fistula. Transient dysphagia occurred, but only one patient required esophageal dilatation. Insignificant delay in gastric emptying was a minor problem despite lack of a pyloroplasty in most. Ultimately a normal diet was tolerated by all patients, and 50% survived more than 12 months, with 30% surviving 5 years or more. This is excellent palliation by a single-stage safe operation that truly improves the quality of life. The simplicity and merit of this approach were proposed by Orringer in 1978. His first report on 26 patients was sharply criticized as advocating a dangerous operation which violated the basic Halstedian principles of adequate exposure,
3 35
gentle handling of tissues, and complete hemostasis. At the 1982 meeting of The American Association for Thoracic Surgery, Orringer reported on transhiatal esophagectomy without thoracotomy and cervical esophagogastrostomy in 134 patients: 40 with benign disease and 94 with carcinomas at various levels of the esophagus. Ten of these carcinomas were pharyngeal and 20 were cervicothoracic. The colon was used as the conduit rather than the stomach in 10 patients. There were 11 postoperative deaths, an 8.2% mortality. None of these deaths resulted from the esophagectomy directly, but were due to myocardial infarction in three, pneumonia in three, innominate artery rupture in two, pulmonary embolus in one, and mediastinal infection or retroperitoneal infection in the remaining two. Average intraoperative blood loss was 1,150 ml (1,100 ml in the group with benign disease, 1,800 ml in the group with malignant disease requiring concurrent laryngectomy, but only 900 ml for those with carcinomas not requiring laryngectomy). Therefore, the entire esophagus can be resected safely without a thoracic incision. However, this approach does not permit mediastinal node dissection, which is still accepted as a requisite part of resection of esophagogastric and lower esophageal carcinomas. Removal of the regional lymph nodes does not seem as important to success of the operation in high thoracic or cervical esophageal cancers. Recently, Sasaki, Baker, McConnell, and Veto recommended free jejunal mucosal patch grafts for reconstruction of large oropharyngeal defects incident to resection of high cervical carcinomas. Again, this is a one-stage surgical procedure, provides a cosmetically acceptable result, preserves the mobility of the tongue, restores function, and preserves normal anatomic relationships. Among seven patients there was only one death, which could have been avoided by a tracheostomy to prevent aspiration of excessive initial oral secretions. Microvascular techniques are utilized to insert a free jejunal loop into the cervical esophageal defect. The advantages of this technique are that it is a single-stage procedure and that the transplanted jejunal loop provides abundant donor tissue with characteristics similar to oral mucosa. Maximum tongue function with near normal facial appearance is maintained. However, this operation requires inordinate skill in microsurgery beyond the scope of most thoracic surgeons, and it lacks the simplicity of direct removal and replacement of the defect by the stomach for all cervical and upper thoracic esophageal carcinomas. Free jejunal transplants will be a valuable adjunct in selected cases of high cervical esophageal carcinoma where the anastomosis can be done easily in the neck, but the majority of high esophageal cancers in the neck and upper thorax will be more simply treated by the technique proposed by Dr. Moores' group. DR. MANJIT S. BAINS New York. N. Y.
At Memorial Sloan-Kettering Institute in New York, we have an almost identical experience. Fifty-six of our patients have undergone nonthoracotomy esophagectomy. Twenty-six
The Journal of
336 Moores et at.
had lesions of the hypopharynx. We have restricted the use of nonthoracotomy esophagectomy to patients with hypopharyngeal, cervical esophageal, or thoracic inlet cancers. We also have a two team approach, with the two teams working simultaneously. This approach has reduced the operating time to 3 to 4 hours unless additional procedures (such as radical neck dissection or thyroidectomy) are performed in the neck. I would emphasize that nonthoracotomy esophagectomy is not a blind procedure. After dilatation of the hiatus, one can easily dissect the thoracic esophagus almost up to the carina under direct vision. Instruments that have been very helpful to us have been a pistol-grip clip applier and a snake light. The morbidity and mortality from this operation have been minimal. Most of the patients start eating by the seventh day and all by the fifteenth day. Most continue to be able to eat until their death. The hospital stay is acceptable. The mortality in our series is 8%, which is similar to that reported by Dr. Moores. The survival achieved with this technique is comparable to that obtained with other methods, if not better. Long-
Thoracic and Cardiovascular Surgery
term survivorship has been obtained, but, most importantly, excellent palliation has been achieved for these patients. DR. MOO RE S (Closing) I would like to thank Dr. Richards and Dr. Bains for their comments. Dr. Richards, pyloroplasty or pyloromyotomy was performed on half of these patients, and in the other half (predictably in the more recent years) no drainage procedure was performed. There were three instances of delayed gastric emptying. Two occurred in the group that had no pyloric drainage procedure, and the other occurred in a patient who had pyloromyotomy. The free jejunal transplant may make a significant contribution in this field. The problems, as Dr. Richards has said, are that the regular thoracic surgeon is not well versed in the microvascular technique and the free jejunal transplant requires at least five anastomoses for completion. The described procedure requires only one anastomosis.