- Email: [email protected]
Indications, risks, and results of completion pneumonectomy
Indications, risks, and results of completion pneumonectomy Completion pneumonectomy refers to an operation intended to remove what is left of a lung partially resected during a previous operation. The procedure is seldom indicated and, according to current medical literature, it carries a higher risk of operative mortality and morbidity than does standard pneumonectomy, especially when done for benign disease. Over the past 20 years, 60 consecutive patients aged 17 to 70 years and having a diagnosis of recurrent lung cancer (n = 28), new primary lung cancer (n = 13), or benign pleuropulmonary disease (n = 19) underwent completion pneumonectomy. The mean interval between the first operation and completion pneumonectomy was 30 months for patients with carcinoma and 215 months for patients with benign disease. For all patients, the previous thoracotomy incision was reopened and maneuvers such as rib resection, intrapericardial blood vessel ligation, division of the bronchus first, local application of glues and hemostatic agents, and bronchial reinforcement were routinely used. Six patients died during (n = 2) or after (n = 4) the operation, for an overall operative mortality of 10 %. The rate was higher for patients with carcinoma (11.6%) than for patients with benign disease (5.9%). Actuarial 5-year survivals from the time of completion pneumonectomy were 48 % for the entire population, 33 % for patients with cancer, and 88 % for patients with benign disease. These results suggest that completion pneumonectomy can be done with an operative risk similar to the one reported for standard pneumonectomy (6% to 10%). In addition, patients undergoing completion pneumonectomy have a reasonable prospect for long-term survival. (J THORAC CARDIOVASC SURG 1993;105:918-24)
Jocelyn Gregoire, MD, FRCS(C) (by invitation), Jean Deslauriers, MD, FRCS(C), Liu Guojin, MD (by invitation), and Jacques Rouleau, MD, FRCP(C) (by invitation), Sainte-Fay, Quebec, Canada
h e first successful pneumonectomy was performed in 1933 by Evarts A. Graham. 1 Since then, pulmonary resection, whether pneumonectomy, lobectomy, or more limited procedures, has become an important therapeutic option not only in the management of primary lung cancer but also in the treatment of less common benign disorders such as cavitary tuberculosis, localized bronchiectasis, or lung abscess. In modern day surgery, operative mortality for resectional surgery of the lung is From the Division of Thoracic Surgery, Le Centre de Pneumologie de Laval, and Laval University, Sainte-Foy, Quebec, Canada. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: Jean Deslauriers, MD, FRCS(C), Head of Division of Thoracic Surgery, Le Centre de Pneumologie de Laval and Professor of Surgery, Laval University, 2725 Chemin Sainte-Foy, Sainte- Foy, Quebec, Canada G 1V 4G5. Copyright © 1993 by Mosby-Year Book, Inc. 0022-5223/93 $\.00
9I8
+ .10
12/6/44496
under 5% for lobectomy and under 10% for pneumonectomy.? Completion pneumonectomy refers to an operation intended to remove what is left of a lung partially resected during a previous operation. The procedure is seldom indicated and, according to current medical literature, it carries a higher risk of operative mortality and morbidity than that of standard pneumonectomy, especially when it is done for benign disease': 4 This report reviews a series of 60 patients who underwent completion pneumonectomy during a 22-year interval. The results show that the operation can be done with an acceptable mortality and, most important, that patients undergoing the operation have a reasonable prospect for long-term survival. Patients and methods First operation. Between January 1969 and September 1991, 60 consecutive patients underwent elective completion
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
Table I. Indications for initial operation Indication
Lung cancer pTNM stage I II IlIa
Unknown Benign disease Pleuropulmonary tuberculosis Bronchiectasis
9I9
Table II. Indications for completion pneumonectomy Completion pneumonectomy
Patients
43 31 5 6
First operation (n)
Indication
Patients
Lung cancer (43)
Local recurrence Secondprimary tumor Bronchial stricture Bronchopleural fistula Hemorrhagicnecrosis of lung Bronchiectasis Bronchopleural fistula Tuberculosis Carcinoma Bronchiectasis Bronchopleural fistula
28
I 17 9
8
pneumonectomy for a variety of pleuropulmonary disorders. There were 42 men and 18 women whose ages at the time of the first operation ranged from 5 to 71 years with a mean of 50 ± 17 years (mean ± standard deviation) (median age 55 years). Indications for initial resection were primary lung cancer in 43 patients and benign disease in 17 patients (Table I). In all patients with carcinoma the disease was staged in accordance with the revised TNM terminology'; and 31 patients had stage I disease, 5 had stage II, 6 had stage IlIa, and in 1 patient the disease could not be staged. Forty-one operations were done on the right side and 19 on the left. These consisted of 45 lobectomies of which 4 were sleeve resections, 14 bilobectornies, and I segmentectomy (lingula). All but I patient had single resection before completion pneumonectomy and all operations done for lung cancer were considered complete on the basis of removal of all gross tumor, a disease-free bronchial resection margin, and highest node free of tumor. Completion pneumonectomy. The interval between the first operation and completion pneumonectomy was less than I year in 16 patients, 1 to 10 years in 31 patients, and more than 10 years in 13 patients. The mean interval was 30 ± 36 months for patients with prior lung tumor (median interval 17 months, range I to 163 months) and 216 ± 146 months for patients with benign disease (median interval 208 months, range 40 to 564 months). Surgical indications for completion pneumonectomy are listed in Table II. Among the 43 patients with initial lung cancer, 28 underwent reoperation for local tumor recurrence, defined as cancer recurring within the ipsilateral hemithorax. Among those 28 individuals, the recurrence was at the bronchial stump in 13 (46%) (Fig. 1), within the hilar nodes with secondary bronchial compression or invasion or both in 5, and in the lung parenchyma in 5 (chest wall invasion in 4 of 5). The remaining 5 patients had recurring multicentric bronchoalveolar carcinoma. Twelve patients underwent completion pneumonectomy for a second primary carcinoma, defined as a tumor of different histologic type or a tumor of similar histologic type if it occurred more than 2 years after the first operation or if its origin could be traced to a carcinoma in situ. 6-8 All ofthese patients had stage I lung cancer (Tis NO, 2 patients; T1 NO, 5 patients; T2 NO, 5 patients), and in every case the completion pneumonectomy was considered to be complete with adequate tumor clearance. The last 3 patients required reoperation for a complication of the previous operation, I for a bronchial stricture documented 6 months after sleeve resection of the right upper lobe, I for a bronchopleural fistula, and I for hemorrhagic necrosis of the lung as a result of pulmonary venous interruption.
Tuberculosis (9)
Bronchiectasis (8)
12 I I I
4 3 I I 6
2
Among the 17 patients with initial benign disease, 16 required completion pneumonectomy for another benign problem (Fig. 2). The remaining patient had a completion pneumonectomy for a squamous cell carcinoma pTNM T2 NO MO that occurred 30 years after initial lobectomy for tuberculosis. Operative technique. The first task of the surgeon was to carefully read the surgical report on the patient's previous procedure, not only to learn about problems that may have been encountered at the time but also to find out if that first operation had been done extrapleurally.? All operations were done by reopening of the previously used thoracotomy incision. In about 25% of cases, the fifth or sixth rib had to be removed to provide easier access to the pleural space. Excessive bleeding from the chest wall was prevented by mobilization of the lung through the intrapleural plane whenever possible. This part of the procedure was particularly difficult when the first operation had been done extrapleurally, such as in the majority of patients with prior pleuropulmonary tuberculosis. Once the lung was freed, the hilum was avoided and the pulmonary blood vessels were ligated intrapericardially. If the pericardial cavity was completely obliterated, we tried to isolate and divide the bronchus first and then ligate the pulmonary blood vessels. The bronchus was transected as close to the tracheal bifurcation as possible and the stump often was reinforced with viable tissue such as pericardium, pericardial fat mobilized from the cardiophrenic angle, intercostal muscle, or transposed thoracic skeletal muscle. Before closure, fibrin glues and hemostatic agents were generously applied over the raw surfaces and the pleural space was always drained. The mean operative time was 209 ± 68 minutes (median time 203 minutes, range 90 to 370 minutes) and the mean blood loss during the procedure was 2019 ± 1885 ml (median blood loss 1500 ml, range 190 to 8500 rnl).
Results Intraoperative morbidity. We encountered significant intraoperative difficulties in 16 patients: laceration of major blood vessels in 7 instances (pulmonary artery, 4; azygos vein, I; pulmonary vein, I; right atrium, I) and laceration of the esophagus and of the diaphragm in I instance each. The pleural space was contaminated by
The Journal of Thoracic and Cardiovascular Surgery May 1993
9 2 0 Gregoire et al.
Fig. 1. A, Posteroanterior chest radiograph of a 65-year-old woman with previous right lower and middle lobectomy for lung cancer. The radiograph shows an air fluid level in the upper lobe caused by local recurrence. B, Posteroanterior chest radiograph after completion pneumonectomy.
Table III. Completion pneumonectomy and causes of death Cause of death
No. ofpatients
Intraoperative (n = 2) Hemorrhage Hemorrhagic necrosis of lung Postoperative (n = 4) Cardiac herniation Myocardial infarction Respiratory failure Bronchopleural fistula
spillage from the lung in 6 patients and the resection was considered incomplete in 4 patients with cancer, all of them with local recurrence. Two patients died during the operation. One died of uncontrollable hemorrhage and the second of heart failure caused by septic shock, acute coagulopathy, and hemorrhagic necrosis of the lung. Postoperative mortality and morbidity. In addition to the 2 individuals who died during the operation, 4 others died during the immediate postoperative period for an overall 30-day mortality of 10% (6/60). The operative mortality was higher for patients with carcinoma (12.2%,
5/41) than for patients with benign disease (5.3%, 1/19). The causes of death are listed in Table III. Sixteen patients (27%) had one major nonfatal complication or more. These included bronchopleural fistula with or without empyema (n = 8), hemorrhage necessitating reoperation (n = 4), respiratory failure (n = 3), and pneumonia (n = 2). The morbidity was twice as high for patients with benign disease (41.1%) as for patients with lung cancer (20.9%). Follow-up. Follow-up was complete for all 54 patients who survived completion pneumonectomy. The range was I month to 20 years with a mean of 4.5 years. The analysis of survival was based on deaths from all causes including operative deaths. Using the actuarial or life-table method for study of survival.!" we analyzed the available data with respect to the initial disease, and all survival figures were calculated from the time of completion pneumonectomy (Fig. 3). For all 60 patients, 5-year actuarial survival was 48%. For patients with initial lung cancer, 5-year survival was 33%, and for patients with initial benign disease it was 88%. We also evaluated survival in relation to the indication for completion pneumonectomy (Fig. 4). When the operation was done for benign disease (n = 19), which includ-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
92 I
Fig. 2. A, Posteroanterior chest radiograph of a 63-year-old woman with previous right upper and middle lobectomyfor bronchiectasis. The radiograph shows an aspergilloma in the remaining lower lobe. B,Posteroanterior chest radiograph after completion pneumonectomy.
ed operations done for complications of previous cancer operations, the survival at 5 years was 70%. For patients with second primary tumors (n = 13) and local recurrences (n = 28), the 5-year survivals were 46% and 24%, respectively.
Discussion Completion pneumonectomy refers to a procedure in which the surgeon removes what is left of a lung partially resected during a previous operation. It is a widely known intervention, but yet there are only two reports discussing in detail the rationale, indications, and results of the operation.': 4 In 1988, McGovern and colleagues' reported a series of 113 consecutive patients who underwent completion pneumonectomy at the Mayo Clinic for lung cancer (n = 64), pulmonary metastasis (n = 20), and benign lung disease (n = 29). There were 14 operative deaths (operative mortality 12.4%), and the 5-year actuarial survival for all 113 patients was 28.4%. In 1990, Oizumi and associates" reported a series of 29 patients who had completion pneumonectomy between 1962 and 1988 at the National Cancer Center Hospital in Tokyo. These operations were done for lung cancer (n = 21),
complications of the initial operation (n = 7), and pulmonary arterial injury during a second operation (n = I). The operative mortality was 13.8% and the 5-year survival was 32.9% for patients with lung cancer. Completion pneumonectomy may be indicated for second primary carcinomas, which are expected to occur in about 5% to 10% of patients having had a complete resection before,6-8, II local recurrent tumor, pulmonary metastasis, complications of a previous operation, or recurrent benign inflammatory disease of the lung. Six reports- 4, 12-15 describe the results of operation for patients with recurrent tumor or a second primary tumor (Table IV). The average 30-day operative mortality of 10% in those series and the mortality of 12.2% reported in this series are comparable with modern standards for pneumonectomy. Ginsberg and associates with the Lung Cancer Study Group- have shown that the 30-day mortality for simple or radical pneumonectomy was 6.2%, with a wide variation among centers (0% to 17.5%) and hospitals (0% to 25%) within the Lung Cancer Study Group. The 5-year survival of the 41 patients with lung cancer at the time of completion pneumonectomy is 33%. This survival is comparable with the one reported by
922
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gregoire et at.
100 90 80
'if!.
Cii >
...
'S;
::::I
en
70 60 50 40 30 20
Benign (N=17)
--4--
10
All cases (N=60) Malignant (N=43)
--+-
---0--
0 0
2
3
4
5
Years Fig. 3. Analysis of survival related to indication for first operation.
100 90 80
~ 0
70
Cii
60
...::::I
50
>
'S;
en
40 30 20
Benign (N=19)
--4--
10
Second primaries (N=13) Local Recurrences (N=28)
--+-
0 0
---0--
3
2
4
5
Years Fig. 4. Analysis of survival related to indication for completion pneumonectomy.
Mcfiovern.' Oizumi' and their colleagues but less than the one expected after the first operation at our institution (number of patients, 1021; 5-year survival, 44%16),a difference likely the result of a lower operative mortality the first time. When completion pneumonectomy is done for benign disease, patients appear to be at a higher risk for complications and death. McGovern and colleagues' have operative mortality figures of 27.6% with an additional 55.2% of patients having major complications. Oizumi and associates" also reported 2 operative deaths among 7 patients who underwent completion pneumonectomy for
complications of their initial operation (operative mortality 28.5%). In our series, the operative mortality associated with completion pneumonectomy for benign disease is low (5.3%) but the morbidity is significant. The reason for this difference in operative mortality appears to be the selection of patients for surgical treatment. In the current series no patient had prior thoracic irradiation, whereas in the Mayo Clinic series 4 of the 6 patients who died intraoperatively had received prior adjuvant radiation therapy. The possibility of having to reoperate on a patient who has had complete resection of lung cancer should therefore be taken into consideration before postoperative
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
923
Table IV. Reports from the literature concerning completion pneumonectomy for patients with bronchogenic carcinoma (local recurrence and second primary tumor) First author
Year
Neptune'? Mathisen!'
1966 1984 1984 1985 1988 1990
Nielsen 14 Dartevelle 15 Mcfiovern' Oizumi"
No. of resections
8 17
4 14 64 21
radiation therapy is given. Indeed, the Lung Cancer Study Group has shown that postoperative adjuvant mediastinal radiation therapy did not improve survival for patients with completely resected stage II and stage III epidermoid carcinomas.!? If radiation therapy is nevertheless indicated, it is worth remembering that doses higher that 4000 to 4500 rad given 12 months or more before operation substantially increase the risk of operative complications and death. For those patients with benign disease, it may also be wise to consider nonsurgical treatment options that may not be optimal in terms of result but may have the advantage of a lower operative risk. The 88% 5-year survival for patients with benign disease included in this series appears to justify the risks involved in the operation. Completion pneumonectomy is a technically difficult operation for which the surgeon must use all of his or her experience and skills. Among the important steps in the procedure, the first one has to be the careful reading of the surgical report of the previous procedure. This is where one learns about prior difficulties, potential hazards, and if that first operation was done extrapleurally. In such cases, mobilization of the lung can be expected to be much more difficult and bloody and the surgeon must be aware of potential injuries to important structures such as the superior vena cava, diaphragm, esophagus, and thoracic duct. Intraoperative bleeding can be a difficult problem which may, on occasion, lead to immediate death or late disastrous complications such as empyema. Chances of major bleeding can be reduced if the lung is mobilized intrapleurally and if the pulmonary blood vessels are ligated intrapericardially. If the pericardial cavity is obliterated, the bronchus should be mobilized from behind and divided first before the pulmonary artery and veins. To prevent postoperative bleeding, we also recommend the liberal use of glues and hemostatic agents over the parietal and hilar surfaces. In summary, this report shows that completion pneumonectomy carries an operative risk similar to that of
Operative mortality (%)
12.5 11.8 0 7.1 9.4 9.5
Survival
4/8 (5-42 mo) 1/4 (4 mo) 2/14 (84, 156 mo) 26.4% (5 yr) 32.9% (5 yr)
standard pneumonectomy. Long-term survivals of 48% appear to justify the risks of the operation. REFERENCES I. Graham EA, Singer JJ. Successful removal of entire lung for carcinoma of the bronchus. JAMA 1933;10I:1371-4. 2. Ginsberg RJ, Hill LD, Eagan RT, et al. Modern thirty-day operative mortality for surgical resectionsin lung cancer. J THORAC CARDIOVASC SURG 1983;86:654-8. 3. McGovern EM, Trastek VF, Pairolero PC, Payne WS. Completion pneumonectomy: indications, complications, and results. Ann Thorac Surg 1988;46:141-6. 4. Oizumi H, Naruke T, Watanabe H, et al. Completion pneumonectomy,a reviewof29 cases.Nippon Kyobu Geka Gakkai Zasshi 1990;38:72-7 (in Japanese). 5. Mountain CF. A new international staging system for lung cancer. Chest 1986 suppl;89:2255-335. 6. Martini N, Melamed MR. Multiple primary lung cancers: J THORAC CARDIOVASC SURG 1975;70:606-12. 7. Martini N, Ghosn P, Melamed MR. Local recurrence and new primary carcinoma after resection. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery. Vol I. Philadelphia: WB Saunders, 1985:164-9. 8. Mousset X, Deslauriers J, Beaulieu M, et al. Le cancer broncho-pulmonaire successif. Importance du diagnostic precoce et survie apres nouvelle exerese chirurgicale. Ann Chir 1989;43:658-62 (in French). 9. Deslauriers J. Indications for completion pneumonectomy [Editorial]. Ann Thorac Surg 1988;46:133. 10. Anderson RP, Bonchek LI, Grunkemeier GL, Lambert LE, Starr A. The analysis and presentation of surgical results by actuarial methods. J Surg Res 1974;16:224-30. II. Pairolero PC, Williams DE, Bergstralh EJ, Piehler JM, Bernatz PE, Payne WS. Postsurgical stage I bronchogenic carcinoma: morbid implications of recurrent disease. Ann Thorac Surg 1984;38:331-8. 12. Neptune WB, Woods FM, Overholt RH. Reoperation for bronchogenic carcinoma. J THORAC CARDIOVASC SURG 1966;52:342-50. 13. Mathisen DJ, Jensik RJ, Faber LP, Kittle CF. Survival following resection for second and third primary lung cancers. J THoRAc CARDIOVASC SURG 1984;88:502-10.
924
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gregoire et ai.
14. Nielsen OS, Boas J, Bertelsen S. Reoperation for recurrent bronchogenic carcinoma. Scand J Thorac Cardiovasc Surg 1984;18:249-50. 15. Dartevelle P, Khalife J. Surgical approach to local recurrence and the second primary lesion. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery. Vol I. Philadelphia: WB Saunders, 1985:156-63. 16. Deslauriers J, Brisson J, Cartier R, et al. Carcinoma of the lung: evaluation of satellite nodules as a factor influencing prognosis after resection. J THORAC CARDIOVASC SURG 1989;97:504-12. 17. The Lung Cancer Study Group: Effects of post-operative mediastinal radiation on completely resected stage II and stage III epidermoid cancer of the lung. N Engl J Med 1986;315: 1377-81.
Discussion Dr. Marvin Pomerantz (Denver, Colo.). I have one question for Dr. Gregoire. Three of your patients died of respiratory distress. In my own series of completion pneumonectomies, acute adult respiratory distress syndrome developed within the first 48 hours in three patients. There was no fluid overload or fluid imbalance in these patients. Is this just an unusual finding in my patients or is there something comparable in your three deaths from respiratory causes? Dr. Gregoire. In most patients who had respiratory distress syndrome after completion pneumonectomy, this complication developed 2 to 3 days after the operation. In every case, the clinical presentation was similar to the one described by Peters, and it was attributed to postpneumonectomy pulmonary edema.
Jocelyn Gregoire, MD, FRCS(C) (by invitation), Jean Deslauriers, MD, FRCS(C), Liu Guojin, MD (by invitation), and Jacques Rouleau, MD, FRCP(C) (by invitation), Sainte-Fay, Quebec, Canada
h e first successful pneumonectomy was performed in 1933 by Evarts A. Graham. 1 Since then, pulmonary resection, whether pneumonectomy, lobectomy, or more limited procedures, has become an important therapeutic option not only in the management of primary lung cancer but also in the treatment of less common benign disorders such as cavitary tuberculosis, localized bronchiectasis, or lung abscess. In modern day surgery, operative mortality for resectional surgery of the lung is From the Division of Thoracic Surgery, Le Centre de Pneumologie de Laval, and Laval University, Sainte-Foy, Quebec, Canada. Read at the Seventy-second Annual Meeting of The American Association for Thoracic Surgery, Los Angeles, Calif., April 26-29, 1992. Address for reprints: Jean Deslauriers, MD, FRCS(C), Head of Division of Thoracic Surgery, Le Centre de Pneumologie de Laval and Professor of Surgery, Laval University, 2725 Chemin Sainte-Foy, Sainte- Foy, Quebec, Canada G 1V 4G5. Copyright © 1993 by Mosby-Year Book, Inc. 0022-5223/93 $\.00
9I8
+ .10
12/6/44496
under 5% for lobectomy and under 10% for pneumonectomy.? Completion pneumonectomy refers to an operation intended to remove what is left of a lung partially resected during a previous operation. The procedure is seldom indicated and, according to current medical literature, it carries a higher risk of operative mortality and morbidity than that of standard pneumonectomy, especially when it is done for benign disease': 4 This report reviews a series of 60 patients who underwent completion pneumonectomy during a 22-year interval. The results show that the operation can be done with an acceptable mortality and, most important, that patients undergoing the operation have a reasonable prospect for long-term survival. Patients and methods First operation. Between January 1969 and September 1991, 60 consecutive patients underwent elective completion
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
Table I. Indications for initial operation Indication
Lung cancer pTNM stage I II IlIa
Unknown Benign disease Pleuropulmonary tuberculosis Bronchiectasis
9I9
Table II. Indications for completion pneumonectomy Completion pneumonectomy
Patients
43 31 5 6
First operation (n)
Indication
Patients
Lung cancer (43)
Local recurrence Secondprimary tumor Bronchial stricture Bronchopleural fistula Hemorrhagicnecrosis of lung Bronchiectasis Bronchopleural fistula Tuberculosis Carcinoma Bronchiectasis Bronchopleural fistula
28
I 17 9
8
pneumonectomy for a variety of pleuropulmonary disorders. There were 42 men and 18 women whose ages at the time of the first operation ranged from 5 to 71 years with a mean of 50 ± 17 years (mean ± standard deviation) (median age 55 years). Indications for initial resection were primary lung cancer in 43 patients and benign disease in 17 patients (Table I). In all patients with carcinoma the disease was staged in accordance with the revised TNM terminology'; and 31 patients had stage I disease, 5 had stage II, 6 had stage IlIa, and in 1 patient the disease could not be staged. Forty-one operations were done on the right side and 19 on the left. These consisted of 45 lobectomies of which 4 were sleeve resections, 14 bilobectornies, and I segmentectomy (lingula). All but I patient had single resection before completion pneumonectomy and all operations done for lung cancer were considered complete on the basis of removal of all gross tumor, a disease-free bronchial resection margin, and highest node free of tumor. Completion pneumonectomy. The interval between the first operation and completion pneumonectomy was less than I year in 16 patients, 1 to 10 years in 31 patients, and more than 10 years in 13 patients. The mean interval was 30 ± 36 months for patients with prior lung tumor (median interval 17 months, range I to 163 months) and 216 ± 146 months for patients with benign disease (median interval 208 months, range 40 to 564 months). Surgical indications for completion pneumonectomy are listed in Table II. Among the 43 patients with initial lung cancer, 28 underwent reoperation for local tumor recurrence, defined as cancer recurring within the ipsilateral hemithorax. Among those 28 individuals, the recurrence was at the bronchial stump in 13 (46%) (Fig. 1), within the hilar nodes with secondary bronchial compression or invasion or both in 5, and in the lung parenchyma in 5 (chest wall invasion in 4 of 5). The remaining 5 patients had recurring multicentric bronchoalveolar carcinoma. Twelve patients underwent completion pneumonectomy for a second primary carcinoma, defined as a tumor of different histologic type or a tumor of similar histologic type if it occurred more than 2 years after the first operation or if its origin could be traced to a carcinoma in situ. 6-8 All ofthese patients had stage I lung cancer (Tis NO, 2 patients; T1 NO, 5 patients; T2 NO, 5 patients), and in every case the completion pneumonectomy was considered to be complete with adequate tumor clearance. The last 3 patients required reoperation for a complication of the previous operation, I for a bronchial stricture documented 6 months after sleeve resection of the right upper lobe, I for a bronchopleural fistula, and I for hemorrhagic necrosis of the lung as a result of pulmonary venous interruption.
Tuberculosis (9)
Bronchiectasis (8)
12 I I I
4 3 I I 6
2
Among the 17 patients with initial benign disease, 16 required completion pneumonectomy for another benign problem (Fig. 2). The remaining patient had a completion pneumonectomy for a squamous cell carcinoma pTNM T2 NO MO that occurred 30 years after initial lobectomy for tuberculosis. Operative technique. The first task of the surgeon was to carefully read the surgical report on the patient's previous procedure, not only to learn about problems that may have been encountered at the time but also to find out if that first operation had been done extrapleurally.? All operations were done by reopening of the previously used thoracotomy incision. In about 25% of cases, the fifth or sixth rib had to be removed to provide easier access to the pleural space. Excessive bleeding from the chest wall was prevented by mobilization of the lung through the intrapleural plane whenever possible. This part of the procedure was particularly difficult when the first operation had been done extrapleurally, such as in the majority of patients with prior pleuropulmonary tuberculosis. Once the lung was freed, the hilum was avoided and the pulmonary blood vessels were ligated intrapericardially. If the pericardial cavity was completely obliterated, we tried to isolate and divide the bronchus first and then ligate the pulmonary blood vessels. The bronchus was transected as close to the tracheal bifurcation as possible and the stump often was reinforced with viable tissue such as pericardium, pericardial fat mobilized from the cardiophrenic angle, intercostal muscle, or transposed thoracic skeletal muscle. Before closure, fibrin glues and hemostatic agents were generously applied over the raw surfaces and the pleural space was always drained. The mean operative time was 209 ± 68 minutes (median time 203 minutes, range 90 to 370 minutes) and the mean blood loss during the procedure was 2019 ± 1885 ml (median blood loss 1500 ml, range 190 to 8500 rnl).
Results Intraoperative morbidity. We encountered significant intraoperative difficulties in 16 patients: laceration of major blood vessels in 7 instances (pulmonary artery, 4; azygos vein, I; pulmonary vein, I; right atrium, I) and laceration of the esophagus and of the diaphragm in I instance each. The pleural space was contaminated by
The Journal of Thoracic and Cardiovascular Surgery May 1993
9 2 0 Gregoire et al.
Fig. 1. A, Posteroanterior chest radiograph of a 65-year-old woman with previous right lower and middle lobectomy for lung cancer. The radiograph shows an air fluid level in the upper lobe caused by local recurrence. B, Posteroanterior chest radiograph after completion pneumonectomy.
Table III. Completion pneumonectomy and causes of death Cause of death
No. ofpatients
Intraoperative (n = 2) Hemorrhage Hemorrhagic necrosis of lung Postoperative (n = 4) Cardiac herniation Myocardial infarction Respiratory failure Bronchopleural fistula
spillage from the lung in 6 patients and the resection was considered incomplete in 4 patients with cancer, all of them with local recurrence. Two patients died during the operation. One died of uncontrollable hemorrhage and the second of heart failure caused by septic shock, acute coagulopathy, and hemorrhagic necrosis of the lung. Postoperative mortality and morbidity. In addition to the 2 individuals who died during the operation, 4 others died during the immediate postoperative period for an overall 30-day mortality of 10% (6/60). The operative mortality was higher for patients with carcinoma (12.2%,
5/41) than for patients with benign disease (5.3%, 1/19). The causes of death are listed in Table III. Sixteen patients (27%) had one major nonfatal complication or more. These included bronchopleural fistula with or without empyema (n = 8), hemorrhage necessitating reoperation (n = 4), respiratory failure (n = 3), and pneumonia (n = 2). The morbidity was twice as high for patients with benign disease (41.1%) as for patients with lung cancer (20.9%). Follow-up. Follow-up was complete for all 54 patients who survived completion pneumonectomy. The range was I month to 20 years with a mean of 4.5 years. The analysis of survival was based on deaths from all causes including operative deaths. Using the actuarial or life-table method for study of survival.!" we analyzed the available data with respect to the initial disease, and all survival figures were calculated from the time of completion pneumonectomy (Fig. 3). For all 60 patients, 5-year actuarial survival was 48%. For patients with initial lung cancer, 5-year survival was 33%, and for patients with initial benign disease it was 88%. We also evaluated survival in relation to the indication for completion pneumonectomy (Fig. 4). When the operation was done for benign disease (n = 19), which includ-
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
92 I
Fig. 2. A, Posteroanterior chest radiograph of a 63-year-old woman with previous right upper and middle lobectomyfor bronchiectasis. The radiograph shows an aspergilloma in the remaining lower lobe. B,Posteroanterior chest radiograph after completion pneumonectomy.
ed operations done for complications of previous cancer operations, the survival at 5 years was 70%. For patients with second primary tumors (n = 13) and local recurrences (n = 28), the 5-year survivals were 46% and 24%, respectively.
Discussion Completion pneumonectomy refers to a procedure in which the surgeon removes what is left of a lung partially resected during a previous operation. It is a widely known intervention, but yet there are only two reports discussing in detail the rationale, indications, and results of the operation.': 4 In 1988, McGovern and colleagues' reported a series of 113 consecutive patients who underwent completion pneumonectomy at the Mayo Clinic for lung cancer (n = 64), pulmonary metastasis (n = 20), and benign lung disease (n = 29). There were 14 operative deaths (operative mortality 12.4%), and the 5-year actuarial survival for all 113 patients was 28.4%. In 1990, Oizumi and associates" reported a series of 29 patients who had completion pneumonectomy between 1962 and 1988 at the National Cancer Center Hospital in Tokyo. These operations were done for lung cancer (n = 21),
complications of the initial operation (n = 7), and pulmonary arterial injury during a second operation (n = I). The operative mortality was 13.8% and the 5-year survival was 32.9% for patients with lung cancer. Completion pneumonectomy may be indicated for second primary carcinomas, which are expected to occur in about 5% to 10% of patients having had a complete resection before,6-8, II local recurrent tumor, pulmonary metastasis, complications of a previous operation, or recurrent benign inflammatory disease of the lung. Six reports- 4, 12-15 describe the results of operation for patients with recurrent tumor or a second primary tumor (Table IV). The average 30-day operative mortality of 10% in those series and the mortality of 12.2% reported in this series are comparable with modern standards for pneumonectomy. Ginsberg and associates with the Lung Cancer Study Group- have shown that the 30-day mortality for simple or radical pneumonectomy was 6.2%, with a wide variation among centers (0% to 17.5%) and hospitals (0% to 25%) within the Lung Cancer Study Group. The 5-year survival of the 41 patients with lung cancer at the time of completion pneumonectomy is 33%. This survival is comparable with the one reported by
922
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gregoire et at.
100 90 80
'if!.
Cii >
...
'S;
::::I
en
70 60 50 40 30 20
Benign (N=17)
--4--
10
All cases (N=60) Malignant (N=43)
--+-
---0--
0 0
2
3
4
5
Years Fig. 3. Analysis of survival related to indication for first operation.
100 90 80
~ 0
70
Cii
60
...::::I
50
>
'S;
en
40 30 20
Benign (N=19)
--4--
10
Second primaries (N=13) Local Recurrences (N=28)
--+-
0 0
---0--
3
2
4
5
Years Fig. 4. Analysis of survival related to indication for completion pneumonectomy.
Mcfiovern.' Oizumi' and their colleagues but less than the one expected after the first operation at our institution (number of patients, 1021; 5-year survival, 44%16),a difference likely the result of a lower operative mortality the first time. When completion pneumonectomy is done for benign disease, patients appear to be at a higher risk for complications and death. McGovern and colleagues' have operative mortality figures of 27.6% with an additional 55.2% of patients having major complications. Oizumi and associates" also reported 2 operative deaths among 7 patients who underwent completion pneumonectomy for
complications of their initial operation (operative mortality 28.5%). In our series, the operative mortality associated with completion pneumonectomy for benign disease is low (5.3%) but the morbidity is significant. The reason for this difference in operative mortality appears to be the selection of patients for surgical treatment. In the current series no patient had prior thoracic irradiation, whereas in the Mayo Clinic series 4 of the 6 patients who died intraoperatively had received prior adjuvant radiation therapy. The possibility of having to reoperate on a patient who has had complete resection of lung cancer should therefore be taken into consideration before postoperative
The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 5
Gregoire et al.
923
Table IV. Reports from the literature concerning completion pneumonectomy for patients with bronchogenic carcinoma (local recurrence and second primary tumor) First author
Year
Neptune'? Mathisen!'
1966 1984 1984 1985 1988 1990
Nielsen 14 Dartevelle 15 Mcfiovern' Oizumi"
No. of resections
8 17
4 14 64 21
radiation therapy is given. Indeed, the Lung Cancer Study Group has shown that postoperative adjuvant mediastinal radiation therapy did not improve survival for patients with completely resected stage II and stage III epidermoid carcinomas.!? If radiation therapy is nevertheless indicated, it is worth remembering that doses higher that 4000 to 4500 rad given 12 months or more before operation substantially increase the risk of operative complications and death. For those patients with benign disease, it may also be wise to consider nonsurgical treatment options that may not be optimal in terms of result but may have the advantage of a lower operative risk. The 88% 5-year survival for patients with benign disease included in this series appears to justify the risks involved in the operation. Completion pneumonectomy is a technically difficult operation for which the surgeon must use all of his or her experience and skills. Among the important steps in the procedure, the first one has to be the careful reading of the surgical report of the previous procedure. This is where one learns about prior difficulties, potential hazards, and if that first operation was done extrapleurally. In such cases, mobilization of the lung can be expected to be much more difficult and bloody and the surgeon must be aware of potential injuries to important structures such as the superior vena cava, diaphragm, esophagus, and thoracic duct. Intraoperative bleeding can be a difficult problem which may, on occasion, lead to immediate death or late disastrous complications such as empyema. Chances of major bleeding can be reduced if the lung is mobilized intrapleurally and if the pulmonary blood vessels are ligated intrapericardially. If the pericardial cavity is obliterated, the bronchus should be mobilized from behind and divided first before the pulmonary artery and veins. To prevent postoperative bleeding, we also recommend the liberal use of glues and hemostatic agents over the parietal and hilar surfaces. In summary, this report shows that completion pneumonectomy carries an operative risk similar to that of
Operative mortality (%)
12.5 11.8 0 7.1 9.4 9.5
Survival
4/8 (5-42 mo) 1/4 (4 mo) 2/14 (84, 156 mo) 26.4% (5 yr) 32.9% (5 yr)
standard pneumonectomy. Long-term survivals of 48% appear to justify the risks of the operation. REFERENCES I. Graham EA, Singer JJ. Successful removal of entire lung for carcinoma of the bronchus. JAMA 1933;10I:1371-4. 2. Ginsberg RJ, Hill LD, Eagan RT, et al. Modern thirty-day operative mortality for surgical resectionsin lung cancer. J THORAC CARDIOVASC SURG 1983;86:654-8. 3. McGovern EM, Trastek VF, Pairolero PC, Payne WS. Completion pneumonectomy: indications, complications, and results. Ann Thorac Surg 1988;46:141-6. 4. Oizumi H, Naruke T, Watanabe H, et al. Completion pneumonectomy,a reviewof29 cases.Nippon Kyobu Geka Gakkai Zasshi 1990;38:72-7 (in Japanese). 5. Mountain CF. A new international staging system for lung cancer. Chest 1986 suppl;89:2255-335. 6. Martini N, Melamed MR. Multiple primary lung cancers: J THORAC CARDIOVASC SURG 1975;70:606-12. 7. Martini N, Ghosn P, Melamed MR. Local recurrence and new primary carcinoma after resection. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery. Vol I. Philadelphia: WB Saunders, 1985:164-9. 8. Mousset X, Deslauriers J, Beaulieu M, et al. Le cancer broncho-pulmonaire successif. Importance du diagnostic precoce et survie apres nouvelle exerese chirurgicale. Ann Chir 1989;43:658-62 (in French). 9. Deslauriers J. Indications for completion pneumonectomy [Editorial]. Ann Thorac Surg 1988;46:133. 10. Anderson RP, Bonchek LI, Grunkemeier GL, Lambert LE, Starr A. The analysis and presentation of surgical results by actuarial methods. J Surg Res 1974;16:224-30. II. Pairolero PC, Williams DE, Bergstralh EJ, Piehler JM, Bernatz PE, Payne WS. Postsurgical stage I bronchogenic carcinoma: morbid implications of recurrent disease. Ann Thorac Surg 1984;38:331-8. 12. Neptune WB, Woods FM, Overholt RH. Reoperation for bronchogenic carcinoma. J THORAC CARDIOVASC SURG 1966;52:342-50. 13. Mathisen DJ, Jensik RJ, Faber LP, Kittle CF. Survival following resection for second and third primary lung cancers. J THoRAc CARDIOVASC SURG 1984;88:502-10.
924
The Journal of Thoracic and Cardiovascular Surgery May 1993
Gregoire et ai.
14. Nielsen OS, Boas J, Bertelsen S. Reoperation for recurrent bronchogenic carcinoma. Scand J Thorac Cardiovasc Surg 1984;18:249-50. 15. Dartevelle P, Khalife J. Surgical approach to local recurrence and the second primary lesion. In: Delarue NC, Eschapasse H, eds. International trends in general thoracic surgery. Vol I. Philadelphia: WB Saunders, 1985:156-63. 16. Deslauriers J, Brisson J, Cartier R, et al. Carcinoma of the lung: evaluation of satellite nodules as a factor influencing prognosis after resection. J THORAC CARDIOVASC SURG 1989;97:504-12. 17. The Lung Cancer Study Group: Effects of post-operative mediastinal radiation on completely resected stage II and stage III epidermoid cancer of the lung. N Engl J Med 1986;315: 1377-81.
Discussion Dr. Marvin Pomerantz (Denver, Colo.). I have one question for Dr. Gregoire. Three of your patients died of respiratory distress. In my own series of completion pneumonectomies, acute adult respiratory distress syndrome developed within the first 48 hours in three patients. There was no fluid overload or fluid imbalance in these patients. Is this just an unusual finding in my patients or is there something comparable in your three deaths from respiratory causes? Dr. Gregoire. In most patients who had respiratory distress syndrome after completion pneumonectomy, this complication developed 2 to 3 days after the operation. In every case, the clinical presentation was similar to the one described by Peters, and it was attributed to postpneumonectomy pulmonary edema.