- Email: [email protected]
Esophagopleural fistula after pneumonectomy for inflammatory disease
J
THORAC CARDIOVASC SURG
89:77-81, 1985
Esophagopleural fistula after pneumonectomy for inflammatory disease The development of an esopbagopleural fIStula after pneumonectomy is one of the less common complications of pneumonectomy. Herein reported are seven cases overa periodof 14 years, five from a series of 896 pneumonectomies performed for malignant or inflammatory disease in the Department of Thoracic Surgery and two referredafter pneumonectomy elsewhere. The fIStula was demonstrated by the escape of radiographic contrast material, methylene blue,or food particles into the pleural space or was observed at esopbagoscopy. In all, the associated empyema was fust treated by drainage, and surgical repairof the fIStula was. attempted in only three cases. In six cases the fIStula bad closed before the patient left hospital. In the seventh, the patient insisted on leaving the hospital while WIder treatment and before closure had occurred. One patient died of cor pulmonale 2 years after closureof the fIStula. There has been no recurrence of the fIStula in any of the patients observed. Conservative management is at variance with that of many authors.
D. M. Shama, B.A., F.R.C.S., and J. A. Odell, F.R.C.S.(Edin.), Durban, Natal, South Africa
h e development of an esophagopleural fistula after pneumonectomy is an uncommon complication. Over an 18 year period, 896 pneumonectomies for malignant disease, bronchiectasis, and tuberculosis have been performed in the departments of thoracic surgery of the University of Natal and Wentworth Hospital. Five cases of esophagopleural fistula have been recognized from this series. This is an incidence of 0.5%, the same as that reported by Evans! (1972). Two patients were referred to our unit after pneumonectomy elsewhere. All our patients had signs of an empyema within the pneumonectomy space, and the diagnosis of an esophagopleural fistula was made by the observation in the pleural liquid of food particles, of methylene blue after oral intake, by contrast studies, or by the presence of a hole in the esophagus at esophagoscopy. All had undergone pneumonectomy for inflammatory disease, and an analysis of published cases indicates that this concurs with the experience of other authors.' Of our seven cases, two developed after left pneumonectomy. Again, the lower incidence on the left side agreed with published cases, of which four have been left-sided and 48 right-sided. From the Department of Thoracic Surgery, University of Natal, Durban. Natal, South Africa. Received for publication Oct. 7, 1983. Accepted for publication March 4, 1984. Address for reprints: J. Odell, Department of Thoracic Surgery, Wentworth Hospital, Jacobs, Natal, South Africa.
Case reports
The details are summarized in Table I.
CASE 1. In December, 1982, a 60-year-old African woman presented with a 3 day history of offensive sputum and right-sided chest pain. In 1965 she had undergone right pneumonectomy elsewhere for inflammatory diseaseand since then had been well. Chest radiographs showedleft pneumonic changes and a fluid level in the pneumonectomy space. The postpneumonectomy empyema was managed initially by closed intercostaldrainage, when food debris was noted in the drained liquid. An esophagopleural fistula was confirmed by barium swallow (Fig. 1). The fistula was not seen at esophagoscopy. The postpneumonectomy empyema was then managed by open drainage, and a fine-bore nasogastric feeding tube was inserted for alimentation. The esophagopleural fistula healed (Fig. 2) and the empyema was thereafter sterilizedand the drain removed. She was discharged welland asymptomatic 9 weeks after admission. Six months later, she had no recurrence of either empyema or esophageal leak. This case initiated the retrospective analysis of our experience with esophagopleural fistula.This case and Case 2 are not included in our incidence figures, as the initial pulmonary resection was not performed by this Unit. CASE 2. In 1954 a 49-year-old Indian woman underwent right pneumonectomy for pulmonary tuberculosis. In April, 1979, a postpneumonectomy empyema and esophagopleural fistula, confirmed by barium swallow but not by esophagoscopy, was managed by thoracoplasty and an intercostal muscle flap repair of the fistula,which leaked.She also was treated by total parenteral nutrition for 2 weeks, after which the fistula was shown to have closed (Fig. 3). The empyema healed. At her last review 3 years later, she had been asymptomatic and well since then. CASE 3. In 1965 a 40-year-old African man underwent
77
The Journal of Thoracic and Cardiovascular Surgery
7 8 Shama and Odell
Table I. Esophagopleural fistula after pneumonectomy for inflammatory lung disease Period between pneumonectomy and fistula (yr)
Age (yr)
Sex
60
F
RIght pneumonectomy for inflammatory lung disease
17
2
49
F
Right pneumonectomy for post-tuberculous lung
25
3
40
M
4
8
M
5
26
F
Emergency right pneumonectomy for destructive pneumonia; thoracoplasty for empyema Difficult right pneumonectomy for destroyed lung through an empyema Right pneumonectomy for bronchiectasis
6
8
F
Difficult left pneumonectomy for pulmonary tuberculosis
Early
7
46
F
Emergency left pneumonectomy for haemoptysis; post-tuberculous lung
3
Case
Initial operation
6
Management Drainage of empyema, nasogastric feeding, sterilization of space Drainage of empyema, thoracoplasty and intercostal flap, sterilization of space Direct suture, sterilization of space
Outcome Healing of fistula and empyema
Failure: healed after nasogastric feeding; empyema healed
Failure; healed after nasogastric feeding; empyema healed
Early
Drainage of empyema, nasogastric feeding
Discharged himself
Early
Drainage of empyema, nasogastric feeding and chemical cautery Drainage of empyema, nasogastric feeding, sterilization of space Drainage of empyema, nasogastric feeding, sterilization of space
Failed; healed after thoracoplasty and muscle flap
emergency right pneumonectomy for life-threatening hemoptysis, consequent upon a destructive pneumonia. His postoperative course was complicated by empyema that necessitated thoracoplasty. He was discharged well and returned in 1971 with an esophagopleural fistula. No cause was found. Direct suture failed, but conservative measures, namely, nasogastric tube feeding, resulted in healing of the fistula. The empyema was sterilized and healed. He was discharged well but has not returned for follow-up. CASE 4. In 1974 a 38-year-old African man with an empyema necessitatis was treated initially by open drainage. Pneumonectomy for a destroyed right lung was then performed with difficulty. This was followed by an early empyema resulting from an esophagopleural fistula. The man was treated by open drainage and nasogastric tube feeding, but he discharged himself against medical advice and has never returned for reassessment. CASE 5. A 26-year-old African woman presented with hemoptysis that resulted from bronchiectasis affecting the entire right lung. In 1977 a right pneumonectomy was performed with difficulty, and an early empyema and esophagopleural fistula developed. Initial management was by drainage of the empyema, nasogastric tube feeding, and chemical cautery of the fistula. This approach failed to close the fistula. Thoracoplasty and intercostal muscle flap repair of the fistula then effected closure of the fistula and healing of the empyema. She was discharged in early 1978. In December, 1979, the woman was readmitted and died of cor pulmonale, the esophagus having remained intact. CASE 6. An 8-year-old girl presented with tuberculous
Healing of fistula and empyema
Healing of fistula and empyema
destruction of the left lung. At the age of 5 years she had had a transthoracic decompression of the thoracic spinal cord for tuberculous paraplegia. In mid-1978 a difficult left pneumonectomy was performed, necessitating reopening of the chest for bleeding. An early empyema with esophagopleural fistula followed. This was managed by open drainage and nasogastric tube feeding, which resulted in closure of the fistula. Thereafter the empyema space was sterilized and she was discharged. The last time she was seen, 3 years after closure of the fistula, there had been no recurrence of the esophageal leak or of empyema. CASE 7. In 1978 a 46-year-old African woman had an emergency left pneumonectomy for life-threatening hemoptysis. She was known to have had pulmonary tuberculosis in the past, the most recent positive sputum test having been 10 months previously. She was discharged well but in October. 1981, returned with hemoptysis. An empyema in the left pneumonectomy space and an esophagopleural fistula were present. Treatment was by open drainage and nasogastric tube feeding, which effected closure of the esophageal fistula. The empyema healed after sterilization. She was discharged well after 6 weeks but has declined to return for follow-up.
Discussion After analyzing the published cases with regard to the etiology of postpneumonectomy esophagopleural fistula, we concluded that a distinction must be made between early and late presentation." Early development of the fistula would appear to be associated with a technically
Volume 89 Number 1
Esophagopleural fistula
79
January, 1985
Fig. 1. Barium swallow demonstrating barium in thedrained right pneumonectomy space. Pneumonectomy had been undertaken 17 years previously (Case 1).
Fig. 2. Barium swallow in thesame patient (Case 1),4 weeks after the commencement of fine-bore nasogastric feeding. There is no leak of barium into the pneumonectomy space.
difficult operation and dissection. This was certainly the case in our experience, as the operative notes recorded difficulty in every instance. A difficult dissection may impair the blood supply to the esophagus. The poorest supply is in the region just distal to the main carina.v? the site where subcarinal nodes lying close to the esophagus are frequently dissected. This also is the most common site of occurrence of esophagopleural fistulas. Other causes for early occurrence are direct injury to the esophagus or a preexistent esophageal diverticulum. Late presentation may be due to tumor recurrence, a continuing inflammatory process in the bronchus itself or in the mediastinal nodes, or chronic suppuration in the pneumonectomy space. The majority of esophagopleural fistulas reported have occurred after right pneumonectomy, only four left-sided fistulas having been previously reported. This bias would appear to be due to the anatomy of the esophagus, as the organ lies closer to the right hemithorax. On the left it is separated from the pleural cavity, and partially supported, by the aorta.
Van Den Bosch and associates,' in a review of the literature, noted that an esophagopleural fistula is most likely to occur after resection for inflammatory disease, proposing that this is consequent upon a difficult dissection. Our experience is in agreement with this suggestion, but it remains unclear to us why our four late-presentingpatients (one after 25 years) should have developed an esophagopieural fistula. No cause was found in any of them, and malignancy was excluded. The correlation between neoplastic disease and early presentation may be a reflection on an aggressive approach to removesubcarinal nodes, an area where the blood supply is poorest. Early diagnosis is essential to reduce morbidity and mortality. Symptoms and signs of an empyema are always present. Analysis of the empyema liquid may reveal food particles. Eriksen" described the presence of desquamated cells in the empyema liquid as diagnostic, but we have never employedthis test. Methylene blue, a swallow of contrast medium, and esophagoscopy are all helpful to confirm the diagnosis, but at esophagoscopy
The Journal of
80 Shama and Odell
Fig. 3. Barium swallow demonstrating no esophagopleural fistula after 2 weeks of total parenteral nutrition (Case 2).
the fistula is not always seen. A helium test' can be used in diagnosis, but again we have not found it necessary to employ this investigation. Our policy has been to drain the associated empyema and to institute enteral alimentation by the nasogastric route. Only once have we had to resort to intravenous parenteral hyperalimentation to maintain the nutritional status of the patient. None of our patients had septicemia or evidence of acute mediastinitis, as was the case with Richarson, Campbell, and Trinkle.' Mediastinal drainage is not necessary, as the esophagus perforates freely into the pneumonectomy space. Once the patient's condition is stable, other surgical maneuvers to close the fistula can be attempted. The published literature suggests that conservative measures are unlikely to succeed. This is in direct variance to our experience, wherein the fistula was closed in five of our seven cases by prolonged nasogastric tube feeding. Our success may indicate that the fistulas were small, as in only one case was the fistula visualized directly despite esophagoscopy in all patients. Because of
Thoracic and Cardiovascular Surgery
the small size, the fistula may be more likely to heal with conservative measures; similarly, the use of conservative measures, after thoracoplasty and intercostal flap repair had failed to heal the fistula, may affect healing because the size of the fistula had been reduced. In none of the patients was their health or recovery compromised by prolonged nasogastric feeding. Direct suture of the fistula seems unlikely to succeed, as exemplified by Case 3, especially in the early phase when the fistula is edematous and acutely inflamed. Muscle flaps or pleural flaps seem to have a high success rate. In combination with these procedures, thoracoplasty may be added to obliterate the pneumonectomy space and thereby diminish the likelihood of empyema and possible recurrent fistula formation from chronic sepsis. Esophageal bypass procedures are other alternatives, but they are associated with a high mortality and are probably unjustified in frail patients. Palliative intubation of the esophagus has been tried in one reported case' for tumor recurrence, but the patient died 12 days later. The exact cause of death was unknown, as no postmortem examination was obtained. From our experience with esophagopleural fistula after pneumonectomy for inflammatory disease, we would recommend drainage of the empyema and prolonged nasogastric feeding. Should there be no sign of healing, then a thoracoplasty and muscle flap repair would appear to be the best choice. We are not in accord with Sethi and Takaro" or Evans, I who condemn nasogastric feeding as doomed to failure or as having a high mortality. We have had no deaths, and healing was achieved in five of seven cases. When considering the management of the esophagopleural fistula, the problem of postpneumonectomy empyema must also be tackled. Our initial management has been to afford drainage by rib resection at the most dependent point. Once the fistula has healed, then definitive attempts to sterilize the space are made. The space is irrigated daily with hydrogen peroxide and povidone-iodine. Once the drainage becomes minimal the size of the drainage tube is slowly reduced and the empyema space is irrigated with the appropriate antibiotic. Once a small size drain is reached, repeat cultures are taken and if negative the drain is removed. If the empyema recurs after attempts to sterilize the space, then thoracoplasty is undertaken. Two of the seven patients in our series required thoracoplasty, one as part of definitive management of the fistula and the other for persisting empyema. Three of the remaining five fistulas healed, with no evidence of recurrent empyema, and did not necessitate thoracoplasty. Patient 3 had had a thoracoplasty done 6 years prior to the occurrence of the esophagopleural fistula, the
Volume 89
Esophagopleural fistula
Number 1 January. 1985
reason at that stage being a postpneumonectomy empyema not responding to sterilization. This case also demonstrates that an esophagopleural fistula can develop despite thoracoplasty, and it is another point in favor of attempting conservative measures first. One of our patients (Patient 4) discharged himself against medical advice, and no further comment can be made regarding the outcome, as he never returned for reassessment. It would appear from the results that it is reasonable to attempt to sterilize the space first and to reserve thoracoplasty for recurrent or persistent empyemas. The empyema space is thus managed on its own merits and should not be considered in conjunction with attempts to close the fistula, as we have shown that the majority heal with conservative measures. REFERENCES Evans JP: Post-pneumonectomy oesophageal fistula. Thorax 27:674-677, 1972 2 Symes JM, Page AJF, Flavell G: Esophagopleural fistula.
3 4 5 6 7
8 9
81
A late complication after pneumonectomy. J THORAC CARDIOVASC SURG 63:783-786, 1972 Van Den Bosch JMM, Swierenga J, Gelissen HJ, Laros CD: Postpneumonectomy oesophagopleural fistula. Thorax 35:865-868, 1980 Benjamin I, Olsen AM, Ellis FH: Esophagopleural fistula. Ann Thorac Surg 7:139-144,1969 Sethi GK, Takaro T: Esophagopleural fistula following pulmonary resection. Ann Thorac Surg 25:74-81, 1978 Takara T, Walkup HE, Okano T: Esophagopleural fistula as a complication of thoracic surgery. J THORAC CARDIOVAse SURG 40:179-193,1960 Dumont A, De Graef, J. La fistule oesophagopleurale, complication tardive de la pneumonectomie. Lyon Chir 57:481-488, 1961 Eriksen KR: Oesophagopleural fistula diagnosed by microscopic examination of pleural fluid. Acta Chir Scand 128:771-777,1964 Richardson JD, Campbell 0, Trinkle JK: Esophagopleural fistula after pneumonectomy. Chest 69:795-797, 1976
THORAC CARDIOVASC SURG
89:77-81, 1985
Esophagopleural fistula after pneumonectomy for inflammatory disease The development of an esopbagopleural fIStula after pneumonectomy is one of the less common complications of pneumonectomy. Herein reported are seven cases overa periodof 14 years, five from a series of 896 pneumonectomies performed for malignant or inflammatory disease in the Department of Thoracic Surgery and two referredafter pneumonectomy elsewhere. The fIStula was demonstrated by the escape of radiographic contrast material, methylene blue,or food particles into the pleural space or was observed at esopbagoscopy. In all, the associated empyema was fust treated by drainage, and surgical repairof the fIStula was. attempted in only three cases. In six cases the fIStula bad closed before the patient left hospital. In the seventh, the patient insisted on leaving the hospital while WIder treatment and before closure had occurred. One patient died of cor pulmonale 2 years after closureof the fIStula. There has been no recurrence of the fIStula in any of the patients observed. Conservative management is at variance with that of many authors.
D. M. Shama, B.A., F.R.C.S., and J. A. Odell, F.R.C.S.(Edin.), Durban, Natal, South Africa
h e development of an esophagopleural fistula after pneumonectomy is an uncommon complication. Over an 18 year period, 896 pneumonectomies for malignant disease, bronchiectasis, and tuberculosis have been performed in the departments of thoracic surgery of the University of Natal and Wentworth Hospital. Five cases of esophagopleural fistula have been recognized from this series. This is an incidence of 0.5%, the same as that reported by Evans! (1972). Two patients were referred to our unit after pneumonectomy elsewhere. All our patients had signs of an empyema within the pneumonectomy space, and the diagnosis of an esophagopleural fistula was made by the observation in the pleural liquid of food particles, of methylene blue after oral intake, by contrast studies, or by the presence of a hole in the esophagus at esophagoscopy. All had undergone pneumonectomy for inflammatory disease, and an analysis of published cases indicates that this concurs with the experience of other authors.' Of our seven cases, two developed after left pneumonectomy. Again, the lower incidence on the left side agreed with published cases, of which four have been left-sided and 48 right-sided. From the Department of Thoracic Surgery, University of Natal, Durban. Natal, South Africa. Received for publication Oct. 7, 1983. Accepted for publication March 4, 1984. Address for reprints: J. Odell, Department of Thoracic Surgery, Wentworth Hospital, Jacobs, Natal, South Africa.
Case reports
The details are summarized in Table I.
CASE 1. In December, 1982, a 60-year-old African woman presented with a 3 day history of offensive sputum and right-sided chest pain. In 1965 she had undergone right pneumonectomy elsewhere for inflammatory diseaseand since then had been well. Chest radiographs showedleft pneumonic changes and a fluid level in the pneumonectomy space. The postpneumonectomy empyema was managed initially by closed intercostaldrainage, when food debris was noted in the drained liquid. An esophagopleural fistula was confirmed by barium swallow (Fig. 1). The fistula was not seen at esophagoscopy. The postpneumonectomy empyema was then managed by open drainage, and a fine-bore nasogastric feeding tube was inserted for alimentation. The esophagopleural fistula healed (Fig. 2) and the empyema was thereafter sterilizedand the drain removed. She was discharged welland asymptomatic 9 weeks after admission. Six months later, she had no recurrence of either empyema or esophageal leak. This case initiated the retrospective analysis of our experience with esophagopleural fistula.This case and Case 2 are not included in our incidence figures, as the initial pulmonary resection was not performed by this Unit. CASE 2. In 1954 a 49-year-old Indian woman underwent right pneumonectomy for pulmonary tuberculosis. In April, 1979, a postpneumonectomy empyema and esophagopleural fistula, confirmed by barium swallow but not by esophagoscopy, was managed by thoracoplasty and an intercostal muscle flap repair of the fistula,which leaked.She also was treated by total parenteral nutrition for 2 weeks, after which the fistula was shown to have closed (Fig. 3). The empyema healed. At her last review 3 years later, she had been asymptomatic and well since then. CASE 3. In 1965 a 40-year-old African man underwent
77
The Journal of Thoracic and Cardiovascular Surgery
7 8 Shama and Odell
Table I. Esophagopleural fistula after pneumonectomy for inflammatory lung disease Period between pneumonectomy and fistula (yr)
Age (yr)
Sex
60
F
RIght pneumonectomy for inflammatory lung disease
17
2
49
F
Right pneumonectomy for post-tuberculous lung
25
3
40
M
4
8
M
5
26
F
Emergency right pneumonectomy for destructive pneumonia; thoracoplasty for empyema Difficult right pneumonectomy for destroyed lung through an empyema Right pneumonectomy for bronchiectasis
6
8
F
Difficult left pneumonectomy for pulmonary tuberculosis
Early
7
46
F
Emergency left pneumonectomy for haemoptysis; post-tuberculous lung
3
Case
Initial operation
6
Management Drainage of empyema, nasogastric feeding, sterilization of space Drainage of empyema, thoracoplasty and intercostal flap, sterilization of space Direct suture, sterilization of space
Outcome Healing of fistula and empyema
Failure: healed after nasogastric feeding; empyema healed
Failure; healed after nasogastric feeding; empyema healed
Early
Drainage of empyema, nasogastric feeding
Discharged himself
Early
Drainage of empyema, nasogastric feeding and chemical cautery Drainage of empyema, nasogastric feeding, sterilization of space Drainage of empyema, nasogastric feeding, sterilization of space
Failed; healed after thoracoplasty and muscle flap
emergency right pneumonectomy for life-threatening hemoptysis, consequent upon a destructive pneumonia. His postoperative course was complicated by empyema that necessitated thoracoplasty. He was discharged well and returned in 1971 with an esophagopleural fistula. No cause was found. Direct suture failed, but conservative measures, namely, nasogastric tube feeding, resulted in healing of the fistula. The empyema was sterilized and healed. He was discharged well but has not returned for follow-up. CASE 4. In 1974 a 38-year-old African man with an empyema necessitatis was treated initially by open drainage. Pneumonectomy for a destroyed right lung was then performed with difficulty. This was followed by an early empyema resulting from an esophagopleural fistula. The man was treated by open drainage and nasogastric tube feeding, but he discharged himself against medical advice and has never returned for reassessment. CASE 5. A 26-year-old African woman presented with hemoptysis that resulted from bronchiectasis affecting the entire right lung. In 1977 a right pneumonectomy was performed with difficulty, and an early empyema and esophagopleural fistula developed. Initial management was by drainage of the empyema, nasogastric tube feeding, and chemical cautery of the fistula. This approach failed to close the fistula. Thoracoplasty and intercostal muscle flap repair of the fistula then effected closure of the fistula and healing of the empyema. She was discharged in early 1978. In December, 1979, the woman was readmitted and died of cor pulmonale, the esophagus having remained intact. CASE 6. An 8-year-old girl presented with tuberculous
Healing of fistula and empyema
Healing of fistula and empyema
destruction of the left lung. At the age of 5 years she had had a transthoracic decompression of the thoracic spinal cord for tuberculous paraplegia. In mid-1978 a difficult left pneumonectomy was performed, necessitating reopening of the chest for bleeding. An early empyema with esophagopleural fistula followed. This was managed by open drainage and nasogastric tube feeding, which resulted in closure of the fistula. Thereafter the empyema space was sterilized and she was discharged. The last time she was seen, 3 years after closure of the fistula, there had been no recurrence of the esophageal leak or of empyema. CASE 7. In 1978 a 46-year-old African woman had an emergency left pneumonectomy for life-threatening hemoptysis. She was known to have had pulmonary tuberculosis in the past, the most recent positive sputum test having been 10 months previously. She was discharged well but in October. 1981, returned with hemoptysis. An empyema in the left pneumonectomy space and an esophagopleural fistula were present. Treatment was by open drainage and nasogastric tube feeding, which effected closure of the esophageal fistula. The empyema healed after sterilization. She was discharged well after 6 weeks but has declined to return for follow-up.
Discussion After analyzing the published cases with regard to the etiology of postpneumonectomy esophagopleural fistula, we concluded that a distinction must be made between early and late presentation." Early development of the fistula would appear to be associated with a technically
Volume 89 Number 1
Esophagopleural fistula
79
January, 1985
Fig. 1. Barium swallow demonstrating barium in thedrained right pneumonectomy space. Pneumonectomy had been undertaken 17 years previously (Case 1).
Fig. 2. Barium swallow in thesame patient (Case 1),4 weeks after the commencement of fine-bore nasogastric feeding. There is no leak of barium into the pneumonectomy space.
difficult operation and dissection. This was certainly the case in our experience, as the operative notes recorded difficulty in every instance. A difficult dissection may impair the blood supply to the esophagus. The poorest supply is in the region just distal to the main carina.v? the site where subcarinal nodes lying close to the esophagus are frequently dissected. This also is the most common site of occurrence of esophagopleural fistulas. Other causes for early occurrence are direct injury to the esophagus or a preexistent esophageal diverticulum. Late presentation may be due to tumor recurrence, a continuing inflammatory process in the bronchus itself or in the mediastinal nodes, or chronic suppuration in the pneumonectomy space. The majority of esophagopleural fistulas reported have occurred after right pneumonectomy, only four left-sided fistulas having been previously reported. This bias would appear to be due to the anatomy of the esophagus, as the organ lies closer to the right hemithorax. On the left it is separated from the pleural cavity, and partially supported, by the aorta.
Van Den Bosch and associates,' in a review of the literature, noted that an esophagopleural fistula is most likely to occur after resection for inflammatory disease, proposing that this is consequent upon a difficult dissection. Our experience is in agreement with this suggestion, but it remains unclear to us why our four late-presentingpatients (one after 25 years) should have developed an esophagopieural fistula. No cause was found in any of them, and malignancy was excluded. The correlation between neoplastic disease and early presentation may be a reflection on an aggressive approach to removesubcarinal nodes, an area where the blood supply is poorest. Early diagnosis is essential to reduce morbidity and mortality. Symptoms and signs of an empyema are always present. Analysis of the empyema liquid may reveal food particles. Eriksen" described the presence of desquamated cells in the empyema liquid as diagnostic, but we have never employedthis test. Methylene blue, a swallow of contrast medium, and esophagoscopy are all helpful to confirm the diagnosis, but at esophagoscopy
The Journal of
80 Shama and Odell
Fig. 3. Barium swallow demonstrating no esophagopleural fistula after 2 weeks of total parenteral nutrition (Case 2).
the fistula is not always seen. A helium test' can be used in diagnosis, but again we have not found it necessary to employ this investigation. Our policy has been to drain the associated empyema and to institute enteral alimentation by the nasogastric route. Only once have we had to resort to intravenous parenteral hyperalimentation to maintain the nutritional status of the patient. None of our patients had septicemia or evidence of acute mediastinitis, as was the case with Richarson, Campbell, and Trinkle.' Mediastinal drainage is not necessary, as the esophagus perforates freely into the pneumonectomy space. Once the patient's condition is stable, other surgical maneuvers to close the fistula can be attempted. The published literature suggests that conservative measures are unlikely to succeed. This is in direct variance to our experience, wherein the fistula was closed in five of our seven cases by prolonged nasogastric tube feeding. Our success may indicate that the fistulas were small, as in only one case was the fistula visualized directly despite esophagoscopy in all patients. Because of
Thoracic and Cardiovascular Surgery
the small size, the fistula may be more likely to heal with conservative measures; similarly, the use of conservative measures, after thoracoplasty and intercostal flap repair had failed to heal the fistula, may affect healing because the size of the fistula had been reduced. In none of the patients was their health or recovery compromised by prolonged nasogastric feeding. Direct suture of the fistula seems unlikely to succeed, as exemplified by Case 3, especially in the early phase when the fistula is edematous and acutely inflamed. Muscle flaps or pleural flaps seem to have a high success rate. In combination with these procedures, thoracoplasty may be added to obliterate the pneumonectomy space and thereby diminish the likelihood of empyema and possible recurrent fistula formation from chronic sepsis. Esophageal bypass procedures are other alternatives, but they are associated with a high mortality and are probably unjustified in frail patients. Palliative intubation of the esophagus has been tried in one reported case' for tumor recurrence, but the patient died 12 days later. The exact cause of death was unknown, as no postmortem examination was obtained. From our experience with esophagopleural fistula after pneumonectomy for inflammatory disease, we would recommend drainage of the empyema and prolonged nasogastric feeding. Should there be no sign of healing, then a thoracoplasty and muscle flap repair would appear to be the best choice. We are not in accord with Sethi and Takaro" or Evans, I who condemn nasogastric feeding as doomed to failure or as having a high mortality. We have had no deaths, and healing was achieved in five of seven cases. When considering the management of the esophagopleural fistula, the problem of postpneumonectomy empyema must also be tackled. Our initial management has been to afford drainage by rib resection at the most dependent point. Once the fistula has healed, then definitive attempts to sterilize the space are made. The space is irrigated daily with hydrogen peroxide and povidone-iodine. Once the drainage becomes minimal the size of the drainage tube is slowly reduced and the empyema space is irrigated with the appropriate antibiotic. Once a small size drain is reached, repeat cultures are taken and if negative the drain is removed. If the empyema recurs after attempts to sterilize the space, then thoracoplasty is undertaken. Two of the seven patients in our series required thoracoplasty, one as part of definitive management of the fistula and the other for persisting empyema. Three of the remaining five fistulas healed, with no evidence of recurrent empyema, and did not necessitate thoracoplasty. Patient 3 had had a thoracoplasty done 6 years prior to the occurrence of the esophagopleural fistula, the
Volume 89
Esophagopleural fistula
Number 1 January. 1985
reason at that stage being a postpneumonectomy empyema not responding to sterilization. This case also demonstrates that an esophagopleural fistula can develop despite thoracoplasty, and it is another point in favor of attempting conservative measures first. One of our patients (Patient 4) discharged himself against medical advice, and no further comment can be made regarding the outcome, as he never returned for reassessment. It would appear from the results that it is reasonable to attempt to sterilize the space first and to reserve thoracoplasty for recurrent or persistent empyemas. The empyema space is thus managed on its own merits and should not be considered in conjunction with attempts to close the fistula, as we have shown that the majority heal with conservative measures. REFERENCES Evans JP: Post-pneumonectomy oesophageal fistula. Thorax 27:674-677, 1972 2 Symes JM, Page AJF, Flavell G: Esophagopleural fistula.
3 4 5 6 7
8 9
81
A late complication after pneumonectomy. J THORAC CARDIOVASC SURG 63:783-786, 1972 Van Den Bosch JMM, Swierenga J, Gelissen HJ, Laros CD: Postpneumonectomy oesophagopleural fistula. Thorax 35:865-868, 1980 Benjamin I, Olsen AM, Ellis FH: Esophagopleural fistula. Ann Thorac Surg 7:139-144,1969 Sethi GK, Takaro T: Esophagopleural fistula following pulmonary resection. Ann Thorac Surg 25:74-81, 1978 Takara T, Walkup HE, Okano T: Esophagopleural fistula as a complication of thoracic surgery. J THORAC CARDIOVAse SURG 40:179-193,1960 Dumont A, De Graef, J. La fistule oesophagopleurale, complication tardive de la pneumonectomie. Lyon Chir 57:481-488, 1961 Eriksen KR: Oesophagopleural fistula diagnosed by microscopic examination of pleural fluid. Acta Chir Scand 128:771-777,1964 Richardson JD, Campbell 0, Trinkle JK: Esophagopleural fistula after pneumonectomy. Chest 69:795-797, 1976