Transhiatal oesophagectomy for oesophageal carcinoma

Br. J. Surg. 1991, Vol. 78. November, 1348-1 351

N. Gurkan, T. Terzioglu, S. Tezelman and 0. Sasmaz Department of General Surgery, Istanbul Medical Faculty, Istanbul University, Turkey Correspondence to: Associate Professor T. Terzioglu, Istanbul Tip Fakultesi, Gene1 Cerrahi A.B.D., Topkapi, Istanbul, Turkey

Transhiatal oesophagectomy for oesophageal carcinoma Between 1979 and I990 transhiatal oesophagectomy and reconstruction with stomach was performed in 148 patients with carcinoma of the oesophagus. Ninety-seven patients were men and 51 were women; ages ranged f r o m 21 to 88 years with a mean of 57.4 years. Dysphagia and weight loss were the usual clinical symptoms. The mean duration of symptoms was 14 weeks. Squamous cell carcinoma was present in 129 patients (87.2per cent), 18patients (12.2per cent) had adenocarcinoma, and one had lymphoma (0.7 per cent). In two-thirds of the patients tumours were located in the middle thoracic (SO of 148 patients) or distal thoracic oesophagus (59 of 148 patients). Three-quarters of the patients had tumours determined as stage III. The mean length of hospital stay after operation was 12.8 days. Anastomotic leakage occurred in IS cases (10.1 per cent). Pulmonary complications other than pneumothorax were observed in 36 cases (24.3 per cent). The 30-day postoperative mortality rate was 8.1 per cent (12 of 148 patients). Respiratory insuficiency was observed as the major cause of death (six of 12 patients). Mediastinitis due to necrosis of the transposed stomach in the mediastinum was the cause of death in three cases. Two-year actuarial survival rates in patients with cervical, upper, middle and lower thoracic tumours were 20, 22, 26 and 30 per cent respectively. Transhiatal oesophagectomy is safe and ofers limited morbidity and mortality, although pulmonary complications and anastomotic leakage in the early postoperative period still pose a significant risk, especially f o r elderly patients in poor condition.

Patients with oesophageal carcinoma are often elderly an d have poor nutritional a n d pulmonary status. Blunt transhiatal oesophagectomy is advocated for these patients t o reduce morbidity an d mortality rates. This technique has been applied Transhiatal t o tumours a t all levels of the oesophagus'-'. oesophagectomy is a safe, short and physiologically well tolerated technique associated with less blood loss and a lower incidence of pulmonary complications t h a n in the standard transthoracic approach for carcinoma of the oesophagus6. Th e complications of this procedure are few a n d can be treated easily3. O n e major disadvantage of the procedure is that it has been reported as a n inadequate operation for cancer. Thus, some authors prefer radical transthoracic oesophagectomy for the '~. the improvetreatment of oesophageal c a r c i n ~ m a ~ .Despite ments in care before and after operation, a combined thoracoabdominal procedure in these patients may entail extremely high morbidity and mortality rates'.". Some authors have compared oesophagectomy with and without t h o r a c ~ t o m y ' ~ . 'Blunt ~. transhiatal oesophagectomy avoids respiratory complications associated with a thoracotomy incision and removes the risk of mediastinitis from a n intrathoracic anastomotic leakage. For the past 10 years, transhiatal oesophagectomy and reconstruction with stomach has been the usual operative approach in o u r ~ l i n i c ' ~ .This ' ~ . report reviews o u r experience of 148 patients with oesophageal carcinoma analysed according t o morbidity, mortality a n d survival rates.

Patients and methods Between 1979 and 1990,148 patients with carcinoma of the oesophagus underwent transhiatal oesophagectomy and reconstruction with

stomach. This was our standard approach during this period for oesophageal carcinoma. Patients with tumours of the cardia that infiltrated the fundus of the stomach and who underwent transthoracic oesophagectomy with intrathoracic oesophagogastric anastomosis and patients with irresectable tumours in whom other procedures were performed have been excluded. Between 1965 and 1990, 1125 patients with carcinoma of the oesophagus were admitted to our clinic. Of these, 488 (43.4 per cent) patients were irresectable. The operative approaches performed in 637 (56.6 per cent) cases considered as resectable were as follows: transthoracic oesophagectomy and gastric reconstruction, 178 cases; transthoracic oesophagectomy and colonic reconstruction, 227 cases; transthoracic oesophagectomy and jejunal reconstruction, 62 cases; transhiatal oesophagectomy and reconstruction with stomach, 148 cases; and cervical jejunal autograft interposition, 22 cases. Of those who underwent transhiatal oesophagectomy there were 97 men (65.5 per cent) and 51 women (34.5 per cent) (M:F ratio 1.9:l). The ages of the patients ranged from 21 to 88 years with a mean of 57.4 years (Figure I ) . One hundred and six patients (71.6 per cent) were aged over 50 years. Dysphagia and weight loss were the usual symptoms. Twenty-four patients had dysphagia for less than 30 days, 105 patients for between 1 and 6 months, and the remaining 19 patients had dysphagia for more than 6 months before the diagnosis of oesophageal carcinoma. The mean duration of symptoms was 14 weeks. Tumours were located in the cervical oesophagus in 24 (16.2 per cent) patients, in the upper thoracic oesophagus in 15 (10.1 per cent) patients, in the middle thoracic part in 50 (33.8 per cent), and in the lower part of the oesophagus in 59 (39.9 per cent) patients. Squamous cell carcinoma was present in 129 (87.2 per cent) cases, adenocarcinoma in 18 (12.2 per cent) cases and lymphoma in one (0.7 per cent) case. One hundred and twelve (75.7 per cent) cases were in stage 111, 32 cases were in stage I1 (21-6 per cent), three cases were in stage IV (2.0 per cent) and only onecase was in stage I (0.7 per cent)16. Preoperative management

This work wws presented at the 1st Eurosurgery Congress, held on 1-4 October 1990 in Paris. France

After physical examination, investigation included routine haematological and biochemical tests, electrocardiography, liver and lung function tests, and arterial blood gas analysis. Localization and possible spread of the tumour was investigated by chest radiography, barium

1348

0007-1323/91/11134844

6 1991 Butterworth-Heinemann Ltd

Transhiatal oesophagectomy for oesophageal carcinoma: N. Gurkan et al.

6o

t

50 -

In c

40-

-2

30-

c

-

48

a, .-

0 0

=

20-

"

21-30

31-40

41-50

51-60

61-70

71-

Age group (years) Figure 1 Age distribution of 148 patients with oesophageal malignancy treated by transhiatal oesophagectomy

swallow, endoscopy and biopsy in all cases. Indirect laryngoscopy and bronchoscopy were performed as indicated by symptoms. Ultrasonographicexamination of the liver and coeliac lymph nodes was performed routinely, but computed tomography (CT) was performed only in the last 48 cases. Parenteral nutrition was administered before operation if the serum albumin level was less than 3.0 g/l. Operation Transhiatal oesophagectomy with cervical oesophagogastrostomy was performed in all patients. The stomach was passed through the posterior mediastinum in the original oesophageal bed. During the preparation of the stomach in tubular form, the vascular arcade of the greater curvature was maintained to the level of the short gastric vessels. The anastomosis was made in the cervical region with a hand-sewn technique using two layers of sutures. A chest drain was inserted intraoperatively if pleural rupture was observed during the procedure. Antibiotic prophylaxis was administered to all patients perioperatively. Postoperative management Nasogastric tube decompression and infusion of saline and dextrose solutions or parenteral nutrition were used in all cases. Total parenteral nutrition delivered 3000 kcal/day (12.5 MJ/day). On the seventh day after operation, before commencing oral feeding, anastomotic integrity was studied by radiographic examination with Gastrografin@(Schering AG, Berlin, Germany). Clinical leakage, associated with local and systemic signs of sepsis, was diagnosed when there was discharge of saliva or food substance through the cervical wound. In addition, when the radiological examination revealed contained extravasation of the contrast in an asymptomatic patient, the leakage was defined as subclinical. Patients with stage 111 and IV tumours received radiotherapy after discharge from our clinic. Statistical analysis

Survival curves were constructed by the Kaplan-Meier method".

Postoperative complications Table I shows the postoperative complications. Anastomotic leakage occurred in 15 (10.1 per cent) cases between 3 and 16 days (mean 8.4 days) after operation. Ten cases were designated as clinical leaks and the others were established as subclinical. All patients had total parenteral nutrition in the initial phase of management. Enteral feeding by nasogastric tube was substituted for parenteral nutrition in four patients who had prolonged fistulae. One patient had a persistent fistula due to recurrence of the tumour which did not respond to conservative treatment. One patient died from a cerebrovascular embolus during conservative treatment of a fistula. The fistulae closed spontaneously in the other eight patients. Eleven (7.4 per cent) cases of this series developed anastomotic stenosis which was treated by mechanical dilatation after operation. Pyloric stenosis occurred in four (2.7 per cent) cases, and wound infection was observed in four (2.7 per cent) cases. Postoperative pulmonary complications occurred in 36 (24.3 per cent) cases (Table I). Eight (5.4 per cent) cases had respiratory infection which was diagnosed as a febrile episode that was accompanied by radiological evidence of pneumonia. Pleural effusion was observed in ten (6.8 per cent) cases. Minimal serous effusions were treated conservatively while clinically symptomatic effusions were treated by surgical drainage. Severe atelectasis requiring mechanical ventilation occurred in 1 1 (7.4 per cent) cases. Acute respiratory failure developed in seven (4.7 per cent) cases, all of which were treated by endotracheal intubation and ventilatory support in the intensive care unit. Six of these patients died. Mortality Twelve patients died within 30 days of operation, resulting in a hospital mortality rate of 8.1 per cent. The cause of death was cerebrovascular embolus in three cases, mediastinitis in three cases, and respiratory insufficiency in six cases. The three deaths from mediastinitis occurred after necrosis of the transposed stomach in the mediastinum. No deaths were recorded from anastomotic leakage or from respiratory infection or atelectasis. Survivd The 136 surviving patients were discharged after a mean hospitalization period after operation of 12.8 days. Two-year actuarial survival rates in patients with cervical, upper, middle and distal thoracic tumours were 20, 22, 26 and 30 per cent respectively (Figure 2 ) .

Discussion Transhiatal oesophagectomy and reconstruction with stomach can be performed for tumours at all levels of the oesophagus. This type of oesophageal resection is a safe, well tolerated and rapid technique, and the stomach is an excellent

Results No intraoperative deaths were recorded. The mean operative blood loss was 898 ml (range 550-1800 ml). The mean length of hospital stay after operation was 12.8 days. intraoperative ~ o ~ ~ ~ i c ~ t i o n s Sixty-six (44.6 per cent) cases had pneumothorax due to rupture of the pleura during transhiatal blunt resection of the oesophagus. This minor complication was treated by introducing a chest drainage tube(s) and establishing an underwater seal drain system. Tracheal or bronchial tear was not observed in our series. Ten patients required splenectomy following intraoperative injury to the spleen. A thoracotomy for intrathoracic haemorrhage was not required in any case. Eleven (7.4 per cent) patients developed hoarseness due to injury to the recurrent laryngeal nerve.

Br. J. Surg., Vol. 78, No. 11, November 1991

Table 1 Postoperative complications in 148 cases No. of cases Surgical complications Anastomotic leakage Stenosis of anastomosis Stenosis of pylorus Wound infection Pulmonary complications Acute respiratory insufficiency Respiratory infection Pleural effusion Atelectasis

I5 11 4 4

(10.1)

(7.4) (2.7) (2.7)

7 (4.7) 8 (5.4) 10 (6.8) 11 (7.4)

Values in parentheses are percentages

1349

Transhiatal oesophagectomy for oesophageal carcinoma: N . Giirkan et al.

concomitant disease, particularly of the cardiorespiratory system, is common. The mean age of the patients who underwent transhiatal oesophagectomy for oesophageal carcinoma was 62 years in the series reported by Orringer3 and 65 years in the series of Barbier et a[.'. In our series, the mean age was 57.4 years, but 71.6 per cent of our patients were more than 50 years old. Therefore old age cannot be claimed as a contraindication for this procedure. The inability to eat causes a rapid weight loss, nutritional deficiency and impaired immunocompetence". Malnutrition is a significant risk factor in patients undergoing major surgical procedures. The beneficial effect of preoperative parenteral nutrition on postoperative morbidity and mortality rates has been reported by some We administered parenteral nutrition to patients with evidence of malnutrition. I I I I I Anastomotic leakage and pulmonary complications are the 6 12 18 24 30 36 most common causes of postoperative death. In one series, more Time after operation (months) anastomotic leakages were observed in bypass procedures Figure 2 Survival rates for patients Mith runlours located in the cervical (34.1 per cent) than in palliative (7.6 per cent) or curative region(---)aniiupper(-).middle( ' ' ~..)undcli.stril(-----)third.s resections (3.7 per cent)". This complication occurred in of the thoracic oesophagus 10.1 per cent of cases in our study. Orringer and Orringer2 reported an anastomotic leak in 12 per cent of 143 patients, of whom 43 had benign disease. One year later, Orringer observed organ for oesophageal reconstruction, providing good function I 3.5.14.15.1 8.19 anastomotic leak in 5 per cent of patients with carcinoma of the thoracic oesophagus3. Others have reported similar figures7. If the oesophageal carcinoma is located at the cervical or The most serious consequence of anastomotic leakage is distal part of the oesophagus, transhiatal resection of the mediastinitis, but this is very rare after transhiatal oesophagecoesophagus can be performed because the tumour can be tomy owing to the positioning of the anastomosis in the cervical dissected under direct vision. In patients with carcinoma of the region and because the transposed stomach blocks the entrance thoracic oesophagus, some technical difficulties may develop to the mediastinum. We have confirmed the low incidence of because of possible adhesions between trachea, bronchus or anastomotic leakage after transhiatal oesophagectomy, and the a~rta~.~'. benign course of such leakage in most cases. In one-third of our patients, tumours were located in the Respiratory complications are the other main problem after middle thoracic part of the oesophagus (33.8 per cent). The use surgery for oesophageal carcinoma. Fok et a/.'3 compared of preoperative CT of the chest in carcinoma of the oesophagus transhiatal and transthoracic resection for oesophageal can predict extraoesophageal spread of the tumour, infiltration carcinoma and found that pleural effusion, atelectasis and of surrounding tissues and resectability. This is especially true empyema thoracis occurred more frequently after a transin tumours of the midthoracic oesophagus. C T has been thoracic procedure. They concluded that a transhiatal approach reported to determine accurately the stage of the tumour in should reduce pulmonary complications. A surgical procedure 88 per cent of cases and infiltration of surrounding tissues in involving an upper abdominal incision with a thoracotomy 89 per cent. Diagnosis of lymph node metastases by C T is wound has been shown to produce a greater disturbance in achieved in only 68 per cent of cases2'. Thus, C T seems to be respiratory function than abdominal incision alonez4. Also, sensitive in identifying tracheobronchial involvement in extended lymphadenectomy can result in stromal oedema of thoracic oesophageal carcinomas. We performed C T routinely the lung due to interference with lymph flowz5and can decrease in our last 48 patients, and C T results correlated well with the expectoration of bronchial secretions because injury to the the postoperative histopathological findings, with accurate bronchial branch of the vagus can impair the cough reflex. prediction of stage in 40 of 48 cases (83.3 per cent) (Table 2 ) . The hospital mortality rate for transthoracic procedures Transhiatal oesophagectomy is a more rapid procedure than ' ~ , 'main ~ . causes of ranges from 11 to 17.5 per ~ e n t ~ , ' ~ . The thoracoabdominal resection, with less blood loss and less death are respiratory failure, sepsis, anastomotic leakage and physiological disturbance. The reported morbidity and malignant ~ a c h e x i a ' ~ .The operative mortality rate of mortality rates with this procedure are lower than after oesophagectomy has been reported to range from standard transthoracic o e s o p h a g e c t ~ m y ~ ~ *~, I*9 . ~ ~The ' ~ ~ ' ~ transhiatal ~~ Fok et a/.I3 compared 1.4 to 7.7 per ~ e n t ~ . ~ , ~ .. I"n. contrast, " most frequent peroperative complication of transhiatal the 30-day mortality rate for transthoracic and transhiatal oesophagectomy is pneumothorax due to rupture of the pleura. of the two procedures and reported that the mortality rates Orringer3 reported that this complication occurred in procedures were not statistically different. In our series, the two-thirds of his patients, and in our series it was observed in hospital mortality rate was 8.1 per cent, and respiratory failure 44.6 per cent of patients. This minor complication can be treated was the major cause of death (six of 12 deaths). Mediastinitis easily by insertion of chest drains. We have not encountered the due to necrosis of the transposed stomach in the mediastinum major intraoperative complications of transhiatal oesophagectomy which are reported as tracheal laceration, chylothorax and uncontrollable haemorrhage3. Dissection of the oesophagus high in the neck sometimes causes injuries to the left Table 2 Comparison of preoperaiive staying by C T with postoperative recurrent laryngeal nerve, and occurred in 1 1 (7.4 per cent) of histopathological sraging in 48 patients our cases. Postoperative histopathological staging A patient with oesophageal carcinoma who has adequate Preoperative staging cardiorespiratory function and no evidence of distant spread by CT I I1 111 IV can be treated palliatively by transhiatal oesophagectomyZ2. Three-quarters of our patients had tumours in stage 111 I Stage I (n = 1) (75.7 per cent) and IV (2.0 per cent). Altogether, 77.7 per cent 11 4 Stage I1 (n= 15) of patients in our series had palliative resections, but 24 Stage 111 (n=28) 4 22.3 per cent were treated curatively on the basis of the staging 4 Stage IV ( n = 4 ) information available. CT, computed tomography Oesophageal carcinoma occurs mainly in older patients, and ~

1350

Br. J. Surg., Vol. 78, No. 11, November 1991

Transhiatal oesophagectomy for oesophageal carcinoma: N. Giirkan et al.

and cerebovascular embolus were the other two causes of postoperative death. Barbier et al.' reported a survival rate of 34 per cent a t 2 years and 23 per cent a t 3 years. T h e median survival time is reported variously as 13.1, 15.7 and 19 month^^^*^^^. In the series of Akiyama et a[.'' the overall 5-year survival rate was 34.6 per cent by the transthoracic procedure. Skinnerg performed radical oesophagectomy with en bloc lymph node dissection for treatment of oesophageal carcinoma, and reported a 3-year actuarial survival rate of 14 per cent for midthoracic carcinomas and of 33 per cent for distal-third tumours. These results appear t o offer little advantage over those obtained with transhiatal oesophagectomy. Orringer's report3 showed a 3-year survival rate of 31 per cent for distal-third carcinomas a n d of 17 per cent for midthoracic carcinomas. O u r 2-year survival rates in midthoracic and distal-third tumours were 26 a n d 30 per cent respectively. In conclusion, transhiatal blunt oesophagectorny and reconstruction with stomach is a physiologically well tolerated technique. CT of the mediastinum should be used before operation, especially in cases with tumours of the middle and upper thoracic oesophagus. If CT demonstrates involvement of surrounding tissues such as the tracheobronchial tree, aorta or pericardium, transhiatal oesophagectomy is not suitable. This safe approach can be considered t o be curative in patients with stage I and I1 tumours. It is a palliative procedure in stage 111 a n d IV disease, but will result in restoration of normal swallowing. Despite improvements in evaluation before operation a n d in care after surgery, pulmonary complications in the early postoperative period still remain a problem.

References 1.

2. 3. 4. 5. 6. 7.

Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978; 76: 643-54. Orringer MB, Orringer JS. Esophagectomy without thoracotomy : a dangerous operation? J Thorac Cardiovasc Surg 1983; 85: 72-80. Orringer MB. Transhiatal esophagectomy without thoracotomy for carcinoma of the thoracic esophagus. Ann Surg 1984; 200: 282-8. Tryzelaar JF, Neptune WB, Ellis FH. Esophagectomy without thoracotomy for carcinoma of the esophagus. Am J Surg 1982; 143: 4869. Baker JW, Schechter GL. Management of panesophageal cancer by blunt resection without thoracotomy and reconstruction with stomach. Ann Surg 1986; 203: 491-9. Wong J. Transhiatal oesophagectomy for carcinoma of the thoracic oesophagus. Br J Surg 1986; 73: 89-90. Barbier PA, Becker CD, Wagner HE. Esophageal carcinoma: patient selection for transhiatal esophagectomy. A prospective analysis of 50 consecutive cases. World J Surg 1988; 12: 263-9.

Er. J. Surg., Vol. 78, No. 11, November 1991

8.

9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21.

22. 23. 24.

Barbier PA, Luder PJ, Schuepfer G, Becker CD. Quality of life and patterns of recurrence following transhiatal esophagectomy for cancer: results of a prospective follow-up in 50 patients. World J Surg 1988; 12: 270-6. Skinner DB. En-bloc resection for neoplasms of the esophagus and cardia. J Thorac Cardiovasc Surg 1983; 85: 59-71. Desa L, Ragnunath AS, Chawla SL, Peel ALG, Dellipiani AW. Treatment policy for the management of carcinoma of the oesophagus. Br J Surg 1988; 75: 275-8. Finley RJ, Grace M, Duff JH. Esophagectomy without thoracotomy for carcinoma of the cardia and lower part of the esophagus. Surg Gynecol Obstet 1985; 160: 49-56. Steiger 2, Wilson RF. Comparison of the results of esophagectomy with and without thoracotomy. Surg Gynecol Obster 1981; 153: 653-6. Fok M, Siu KF, Wong J. A comparison of transhiatal and transthoracic resection for carcinoma of the thoracic esophagus. Am J Surg 1989; 158: 414-19. Giirkan N, Avci C, Terzioglu T, Sasmaz 0. L'oesophagectomie sans thoracotomie dans le traitement du cancer de I'oesophage. Analyses de 47 cas. Chirurgie 1984; 110: 60-14. Giirkan N, Avci C, Demirkol K, Eldegez U, Terzioglu T, Sasmaz 0.Experience with esophagectomy without thoracotomy in 67 cases. In: Siewert JR, Hoelscher AH, eds. Diseases of the Esophagus. Berlin: Springer-Verlag, 1987: 403-6. Japanese Committee for Registration of Esophageal Carcinoma. A proposal for new TNM classification of esophageal carcinoma. Jpn J Clin Oncol 1985; 14: 625-36. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Statist Assoc 1958; 53: 457-81. Akiyama H, Tsurumaru M, Kawamura T, Watanabe C, Ono Y. Principles of surgical treatment for carcinoma of the esophagus: analysis oflymph node involvement. Ann Surg 1981; 194:43846. Postlethwait RW. Complications and deaths after operations for esophageal carcinoma. J Thorac Cardiouasc Surg 1983; 85: 827-3 1 . Goldfaden D, Orringer MB, Appelman HD, Kalish R. Adenocarcinoma of the distal esophagus and gastric cardia. J Thorac Cardiovasc Surg 1986; 91: 242-7. Lefor AT, Merino MM, Steinberg SM et al. Computerized tomographic prediction of extraluminal spread and prognostic implications of lesion width in esophageal carcinoma. Cancer 1988; 62: 1287-97. Hennessy TPJ. Choice of treatment in carcinoma of the oesophagus. Br J Surg 1988; 75: 1934. Daly JM, Massar E, Giacco G et al. Parenteral nutrition in esophageal cancer patients. Ann Surg 1982; 1%: 203-8. Bishop DGM, McKeown KC. Postoperative hypoxaemia: oesophagectomy with gastric replacement. Br J Surg 1979; 66: 810-12.

25.

Sugimaci K, Veo H, Natsuda Y, Kai H, Inokuchi K, Zaitsu A. Cough dynamics in oesophageal carcinoma: prevention of postoperative complications. Br J Surg 1982; 69: 7344.

Paper accepted 8 July 1991

1351