- Email: [email protected]
Gangrene of the Lung: Treatment in Two Stages
Gangrene of the Lung: Treatment in Two Stages Yael Refaely, MD, and Dov Weissberg, MD Department of Thoracic Surgery, Tel Aviv University Sackler School of Medicine, Tel Aviv, and E. Wolfson Medical Center, Holon, Israel
Background. Pulmonary gangrene is a rare complication of severe lung infection with devitalization of lung parenchyma and secondary infection. If untreated, gangrene of the lung leads to sepsis, multiple-organ failure, and death. Resection of all gangrenous tissue is mandatory and is lifesaving. Pleural empyema commonly accompanies gangrene of the lung; in its presence, dissection of hilar structures for resection can lead to mediastinitis or bronchopleural fistula and should be avoided. Methods. Three patients with pulmonary gangrene were treated in two stages: immediate fenestration first
and then delayed resection of gangrenous lung in a clean field and immediate closure of the pleural window. Results. Two patients underwent pneumonectomy and 1 patient, lobectomy. All patients recovered without complications. Conclusions. Creation of a pleural window (fenestration) for 1 week enables safe and curative resection of a gangrenous lung or lobe in a clean field and in a patient in stable condition. (Ann Thorac Surg 1997;64:970 – 4) © 1997 by The Society of Thoracic Surgeons
P
Patient 1
ulmonary gangrene is a complication of severe lung infection in which thrombosis of pulmonary vessels causes devitalization of pulmonary parenchyma with secondary anaerobic infection and necrosis. Gangrene of the lung is rare; until 1994, only 25 cases had been reported in the English-language literature [1]. During the past 10 years, we have encountered 3 patients with pulmonary gangrene. In all of them, it was complicated by empyema. We present here the case reports of these 3 patients.
Patients and Methods Three patients with pulmonary gangrene were treated in our institutions between 1987 and 1993. Two had fulminant pleural empyema complicating pneumonia, and 1 had multiple lung abscesses complicated by empyema. Treatment with antibiotics and closed thoracostomy drainage was ineffective. Pleuroscopy established the diagnosis in all patients. The patients were treated in two stages: immediate fenestration and 1 week later, resection of gangrenous lung tissue and closure of the pleural window. Our technique of fenestration has been described in detail previously [2]. In brief, with the patient under general anesthesia, segments of two or three ribs are resected with intercostal muscles, thereby creating a wide opening into the pleural cavity. The skin and subcutaneous tissue are preserved, folded into the pleural cavity, and sutured to the parietal pleura. This cutaneous flap is used later for closure of the window. The case reports of the patients follow. Accepted for publication April 7, 1997. Address reprint requests to Dr Weissberg, Department of Surgery, E. Wolfson Medical Center, Holon 58100, Israel.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
A 25-year-old woman was admitted because of a temperature of 39.4°C lasting 1 week. Chest roentgenogram showed complete opacification of the left lung field. Aspiration of the pleura yielded pus from which Streptococcus pneumoniae was grown. While she was being treated with pleural drainage and penicillin, the patient became septic; body temperature rose to 41°C and up to 950 mL of pus was drained every 24 hours. At pleuroscopy, a large amount of malodorous pus and fibrin was evacuated. The lung appeared black, and it did not expand with controlled respiration. The pleuroscopy was converted to a thoracotomy, and gangrene of the whole lung and a fulminant inflammatory process involving the entire pleura were found. Cultures grew Bacteroides fragilis. Pneumonectomy was indicated, but dissection of hilar structures could cause mediastinitis. Therefore, preliminary fenestration was chosen as a temporizing procedure. The pleural cavity was packed with gauze soaked in Eusol (a bactericidal acid solution of sodium hypochlorite), and the chest was left open. Packs with Eusol were changed every 12 hours, and metronidazole was administered intravenously. Within 7 days, the pleural cavity became clean. Pneumonectomy was then performed with immediate closure of the chest. The histologic diagnosis was pulmonary gangrene. The postoperative course was uneventful. Over the next 5 years, the patient married and delivered 2 babies.
Patient 2 A 20-year-old man was hospitalized because of dyspnea, cough, temperature of up to 40.5°C, blood-stained sputum, and right-sided pleuritic pain lasting 10 days. The white blood cell count was 18,600/mL. Chest roentgenogram showed a large amount of liquid surrounding the right lung. Aspiration yielded pus from which Klebsiella 0003-4975/97/$17.00 PII S0003-4975(97)00837-0
Ann Thorac Surg 1997;64:970 – 4
REFAELY AND WEISSBERG GANGRENE OF LUNG
971
tient’s condition deteriorated, with chills, temperature rising to 41°C, and stupor. At pleuroscopy, 250 mL of foul-smelling pus was evacuated. The lung appeared gangrenous. Pneumonectomy was postponed in favor of a temporizing fenestration. Within 24 hours, the patient regained full consciousness and became cooperative. Within 48 hours, body temperature became normal. After 7 days, the open pleural cavity was clean; pneumonectomy was performed, and the chest was closed. Recovery was uneventful. Histologic examination confirmed gangrene of the lung. During the 4 years of follow-up, the patient did well.
Comment
Fig 1. (Patient 3.) Chest roentgenogram shows multiple infiltrates and areas of destruction in the right lung.
pneumoniae was grown. Treatment with cefotaxime sodium and closed thoracostomy drainage was started, but the patient’s condition deteriorated progressively, and he became stuporous. Five days after initiation of treatment, pleuroscopy was performed, and 620 mL of pus was evacuated. The right lower lobe appeared frankly necrotic. Although resection of the gangrenous lobe was indicated, it was postponed because of fulminant empyema, and fenestration was performed as a temporizing measure. The pleural cavity was filled with gauze soaked in Eusol, which was changed every 12 hours. During the following days, the infection cleared as the patient gradually recovered from sepsis. Eight days after the fenestration, the right lower lobe was resected, and the chest was closed. The postoperative course was uneventful. The patient did well during the 3-year follow-up.
Patient 3 A 40-year-old man who was a heavy smoker with a history of alcohol abuse had had recurrent bouts of pulmonary infections complicated with multiple abscesses of the right lung for 2 years. During the most recent episode of coughing, he expectorated fragments of necrotic lung tissue. On admission, he was in fair general condition with a body temperature of 38.5°C, a breathing rate of 22 per minute, a pulse rate of 84 per minute, and a blood pressure of 125/70 mm Hg. Coarse rhonchi and diminished breath sounds were heard over the right hemithorax; there were no other major findings. Chest roentgenogram and computed tomograms showed massive infiltrates in the right lung, areas of destruction, and a moderate amount of liquid in the pleural cavity (Figs 1, 2). These findings were compatible with necrotizing pneumonitis. Aspiration of pus from the right side of the pleura yielded anaerobes and Aspergillus fumigatus. Closed thoracostomy drainage was instituted, and treatment with metronidazole was started, but the pa-
Gangrene of the lung begins as a pulmonary infectious process, such as lobar pneumonia. Aspiration and devascularization with superimposed anaerobic infection is the presumed usual pathogenic process, and alcohol abuse is implicated as an important contributing factor [3]. The infecting organisms cause vascular thrombosis, resulting in loss of or gross reduction in blood supply to the pulmonary parenchyma. This brings about devitalization of the lung tissue, thus facilitating secondary infection and necrosis [4 –7]. Various pathogens have been found to be causative agents, especially Klebsiella pneumoniae, Friedla¨nder’s bacillus, polymicrobial anaerobic organisms, pneumococcus, and various strains of Aspergillus [1, 4, 8]. In the angioinvasive form of aspergillosis, mycotic invasion of vasa vasorum results in infarction of the media and secondary thrombus formation, leading to pulmonary infarction [8]. A different kind of mechanism has been described in Pseudomonas-induced pulmonary gangrene. Most strains of Pseudomonas produce proteolytic enzymes (exotoxins), which are capable of causing gangrene by a chemical process [8 –11]. When an extensive necrotizing process involves a large mass of pulmonary parenchyma, the gangrenous tissue is partially expectorated and partially accumulates in the cavity created by the process.
Fig 2. (Patient 3.) Computed tomography confirmed the findings shown in Figure 1. In addition, a moderate amount of liquid is visible in the pleural cavity.
972
REFAELY AND WEISSBERG GANGRENE OF LUNG
Danner and associates [4] described radiologic findings typical for pulmonary gangrene. The process begins as dense consolidation and enlargement of the affected lobe, causing outward bulging of the interlobar fissures. This is followed by breakdown into many small cavities, which in turn coalesce into a single large cavity occupying the entire lobe. The cavity is filled with fluid in which irregular pieces of sloughed lung parenchyma float like icebergs [4]. These radiologic findings have been observed also by others [1, 7, 8, 12] but are not invariably present. Hammond and co-workers [3] presented the cases of 2 patients with massive pulmonary gangrene. In both, there was extensive unilateral involvement, with initial dense consolidation followed by cavitation. However, the feature of coalescence into a large cavity with free-floating slough was not seen. Nor were these radiologic features observed in our 3 patients. In 2, the roentgenographic findings were those of empyema. In the other, multiple abscesses presented as areas of necrosis, but the coalescence into a single large cavity did not occur. Accordingly, although the radiologic features are helpful when present, they should not be considered an absolute requirement for the diagnosis of pulmonary gangrene. In their absence, gangrene of the lung can be difficult to diagnose, particularly in the presence of empyema, when opacification of the lung field obstructs changes in the pulmonary parenchyma. This difficulty can easily be overcome by pleuroscopy. We [13] have used this procedure for many years both to obtain information and to aid in treatment of patients with empyema resistant to treatment, and the results have nearly always been gratifying. In fact, pleuroscopy was the determining diagnostic procedure in the 3 patients described here. At pleuroscopy, the gangrenous lung appears black and does not expand with respiration. Once diagnosed, pulmonary gangrene must be treated promptly by resection of all necrotic tissue. Failure to do so is likely to result in sepsis, multiple-organ failure, and death [4]. In the collected series reviewed by Danner and colleagues [4], all 6 patients who were treated surgically, either by resection or by open drainage with spontaneous extrusion of the necrotic lobe, survived, whereas 4 patients treated medically died. Reports of medically treated patients surviving are an exception [12]. When the entire lung or lobe is involved, curative resection requires dissection of hilar vessels and bronchi. This is hazardous in the presence of severe purulent infection involving the entire pleura and can lead to mediastinitis or bronchopleural fistula. In addition, such patients are often too ill to undergo lobectomy or pneumonectomy at this stage. These difficulties can be overcome by dividing treatment into two stages: take care of the overwhelming infection first and then resect the gangrenous lobe or lung in a clean field. This approach was reported by Young and Samson [14] in 1980. The patient had gangrene of the right lower lobe, was close to death, and was unable to tolerate a lobectomy. She was treated by open drainage, with unroofing
Ann Thorac Surg 1997;64:970 – 4
of the abscess cavity and loose packing with iodoformimpregnated gauze. The wound was left open. On the 17th postoperative day, right lower lobectomy and decortication of the remaining lobes were performed, and the patient eventually recovered. Management of pulmonary gangrene by open drainage with resection of three ribs was also reported by Juettner and associates [15]. On the other hand, Hammond and colleagues [3] discussed 2 patients in whom gangrenous lung was resected without preceding open drainage. In both patients, empyema developed after the seemingly successful resection. We believe that by reversing the order of treatment and resecting the gangrenous lung in a clean field, postoperative empyema could have been avoided. Our 3 patients were treated in two stages, and all had good results. Postponement of resection for 1 week with the chest cavity open led to recovery from empyema and permitted resection of the gangrenous lung or lobe in a patient in stable condition and in a clean field. The operative risk was thus minimized, and the risk of causing mediastinitis was avoided. On the basis of our findings and those of others, we draw the following conclusions: 1. Untreated pulmonary gangrene leads to multipleorgan failure, sepsis, and death. Resection of all necrotic tissue is mandatory. 2. Dissection of hilar structures in the presence of pleural empyema can lead to mediastinitis or bronchopleural fistula and should be avoided. 3. Fenestration followed by delayed resection in a clean field and in a patient in stable condition is curative.
References 1. Penner C, Maycher B, Long R. Pulmonary gangrene. A complication of bacterial pneumonia. Chest 1994;105:567–73. 2. Weissberg D. Empyema and broncho-pleural fistula: experience with the open window thoracostomy. Chest 1982;82: 447–50. 3. Hammond JMJ, Lydell C, Potgieter PD, Odell J. Severe pneumococcal pneumonia complicated by massive pulmonary gangrene. Chest 1993;104:1610–2. 4. Danner PK, McFarland DR, Felson B. Massive pulmonary gangrene. Am J Roentgenol Radium Ther Nucl Med 1968; 103:548–54. 5. Zagoria RJ, Choplin RM, Karstaedt N. Pulmonary gangrene as a complication of mucormycosis. AJR 1985;144:1195– 6. 6. Chargan NB, Turk MG, Dhand R. The role of bronchial circulation in lung abscess. Am Rev Respir Dis 1985;131: 121– 4. 7. Phillips LG, Rao KVS. Gangrene of the lung. J Thorac Cardiovasc Surg 1989;97:114– 8. 8. Reich JM. Pulmonary gangrene and the air crescent sign. Thorax 1993;48:70– 4. 9. Hitschman F, Kreilich K. Zur Pathogenese des Bacillus pyocyaneus und zur Aetiologie des Ekthyma Gangraenosum. Wien Klin Wochenschr 1897;10:1093–107. 10. Liu PV. The roles of various fractions of Pseudomonas in its pathogenesis. II. Effects of lecithinase and protease. J Infect Dis 1966;116:112– 6. 11. Liu PV. Extracellular toxins of Pseudomonas aeruginosa. J Infect Dis 1974;130(Suppl O):S94 –9. 12. Padmanabhan K, Rajgopalan K, Yeo K, Dhar SR. Intracavity
Ann Thorac Surg 1997;64:970 – 4
mass in a patient with Klebsiella pneumonia. Chest 1988;93: 187– 8. 13. Weissberg D, Refaely Y. Pleural empyema: 24-year experience. Ann Thorac Surg 1996;62:1026–9. 14. Young JN, Samson PC. Pseudomonas aeruginosa septicemia
REFAELY AND WEISSBERG GANGRENE OF LUNG
973
with gangrene of the lung and empyema. Ann Thorac Surg 1980;29:254–7. 15. Juettner FM, Arian-Schad K, Kraus I, Gallhofer G, Popper H, Friehs G. Total unilateral lung gangrene in Hodgkin’s disease: treatment by thoracostomy. Ann Thorac Surg 1991;51:302–3.
INVITED COMMENTARY Livid, bloody, foul-smelling or billious sputum supervening in cases of continued fever is of bad significance. However if such expectoration removes the diseased tissues all may be well Hippocrates [1]
The association of lung abscess with a poor prognosis has been known since the age of Hippocrates. A lung abscess usually follows aspiration or a pneumonia occurring in a compromised patient. Sequelae depend on the volume and composition of the inoculum and the resistance of the host, which may be modified by malnutrition, anemia, the presence of underlying disease or damaged lung, and the adequacy of antibiotic therapy. Microabscesses or larger abscesses, single, multiple, or multiloculated may then develop depending on the extent of necrotizing gangrenous infection. The visceral pleura and a thin layer of subjacent lung usually survive owing to independent blood supply from the subpleural vascular plexus. A pleuritis usually results in the visceral pleura adhering to the parietal pleura, but occasionally progression is so rapid that an associated empyema develops as in the patients described by Refaely and Weissberg. Death of tissue or gangrene occurs to a variable extent in every lung abscess; when it is confined to a whole lobe or lung the term “pulmonary gangrene” may be applied. Use of this term does not imply that treatment is different, only that it is more extensive. When large areas of pulmonary gangrene are present, antibiotic therapy is ineffective because of poor diffusion of the drug into an avascular region; thus alternative therapy is necessary. In the early decades of this century when thoracic surgery was finding its feet, Neuhoff and Tourhoff [2] as well as Shaw [3] advocated drainage rather than resection of lung abscess with excellent results. With the advent of antibiotics, immunization, and chest physical therapy, lung abscess today is an infrequent problem in the First World. Open drainage as a form of management has been largely forgotten but periodically surfaces in the literature as advocated by Refaely and Weissberg. One of the largest series on lung abscess and drainage was published in the South African Journal of Surgery [4] by Postma and Le Roux of Durban, South Africa, from an institution where I trained. Two periods were compared: a series where resection was emphasized (268 patients) and a later period (417 patients) where drainage followed by later resection in selected patients was the usual first form of management. Roughly the same percentages (27.5% and 29%) had an operation. The mortality in the first series (resection) was 15.4% and the rate of complications was 34.6%; in the second series (drainage) the
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
corresponding results were 0.9% and 15.7%, strongly suggesting that drainage should be the first procedure in a toxically ill patient and that pulmonary resection should only be undertaken in the quiescent phase of the disease. The excellent results of Refaely and Weissberg confirm this and emphasize the importance of drainage in a sick septic patient before any major thoracic procedure. Although excellent results have been achieved in this small series of 3 patients, one must be careful not to extrapolate all of their conclusions into general practice. The statement that resection of all gangrenous tissue is mandatory and is life-saving implies that the second stage after drainage is necessary in all patients with this condition. This is not so: in some of our patients, especially those with Klebsiella pneumonia, drainage of the gangrenous lobe resulted in a large cavity with bronchi appearing like a stunted tree. Occasionally very rapid improvement in symptoms with a chest radiograph demonstrating either an empty space and in others the destroyed lobe obliterated by expansion of the remaining lobes allowed discharge. We used to call this autoamputation of the lobe. We did not proceed with pulmonary resection, indicating that effective drainage can sometimes result in cure. Obviously, if the whole lung is involved this is unlikely to occur. What is life-saving is the initial drainage procedure. Effective drainage of the empyema in the series of Refaely and Weissberg resulted in rapid improvement of their patients, but the source of infection— gangrenous lung—was left unmanaged in a desperately ill patient, begging the question, which can only be unanswered, whether the gangrenous lung should have been additionally surgically managed. Fenestration, or the creation of a large defect by multiple rib resection and folding skin inward, is usually reserved for patients with a chronic empyema, those unfit for any major procedure, or patients with a postpneumonectomy empyema, yet in Refaely and Weissberg’s patients pulmonary resection and closure of the defect occurred within 8 days. Although Refaely and Weissberg had no indication of response to drainage and the timing of the second procedure it does appear an excessive procedure and that similar drainage, with the same objective of complete evacuation of pus and fibrinous debris, may have been achieved with simple rib resection. It is stated that dissection of hilar structures in the presence of severe infection can lead to mediastinitis and should be avoided. There is no evidence to support such a statement. Although most patients with pulmonary gangrene have, because of the rapidity of the disease process, a free
0003-4975/97/$17.00 PII S0003-4975(97)00894-1
Background. Pulmonary gangrene is a rare complication of severe lung infection with devitalization of lung parenchyma and secondary infection. If untreated, gangrene of the lung leads to sepsis, multiple-organ failure, and death. Resection of all gangrenous tissue is mandatory and is lifesaving. Pleural empyema commonly accompanies gangrene of the lung; in its presence, dissection of hilar structures for resection can lead to mediastinitis or bronchopleural fistula and should be avoided. Methods. Three patients with pulmonary gangrene were treated in two stages: immediate fenestration first
and then delayed resection of gangrenous lung in a clean field and immediate closure of the pleural window. Results. Two patients underwent pneumonectomy and 1 patient, lobectomy. All patients recovered without complications. Conclusions. Creation of a pleural window (fenestration) for 1 week enables safe and curative resection of a gangrenous lung or lobe in a clean field and in a patient in stable condition. (Ann Thorac Surg 1997;64:970 – 4) © 1997 by The Society of Thoracic Surgeons
P
Patient 1
ulmonary gangrene is a complication of severe lung infection in which thrombosis of pulmonary vessels causes devitalization of pulmonary parenchyma with secondary anaerobic infection and necrosis. Gangrene of the lung is rare; until 1994, only 25 cases had been reported in the English-language literature [1]. During the past 10 years, we have encountered 3 patients with pulmonary gangrene. In all of them, it was complicated by empyema. We present here the case reports of these 3 patients.
Patients and Methods Three patients with pulmonary gangrene were treated in our institutions between 1987 and 1993. Two had fulminant pleural empyema complicating pneumonia, and 1 had multiple lung abscesses complicated by empyema. Treatment with antibiotics and closed thoracostomy drainage was ineffective. Pleuroscopy established the diagnosis in all patients. The patients were treated in two stages: immediate fenestration and 1 week later, resection of gangrenous lung tissue and closure of the pleural window. Our technique of fenestration has been described in detail previously [2]. In brief, with the patient under general anesthesia, segments of two or three ribs are resected with intercostal muscles, thereby creating a wide opening into the pleural cavity. The skin and subcutaneous tissue are preserved, folded into the pleural cavity, and sutured to the parietal pleura. This cutaneous flap is used later for closure of the window. The case reports of the patients follow. Accepted for publication April 7, 1997. Address reprint requests to Dr Weissberg, Department of Surgery, E. Wolfson Medical Center, Holon 58100, Israel.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
A 25-year-old woman was admitted because of a temperature of 39.4°C lasting 1 week. Chest roentgenogram showed complete opacification of the left lung field. Aspiration of the pleura yielded pus from which Streptococcus pneumoniae was grown. While she was being treated with pleural drainage and penicillin, the patient became septic; body temperature rose to 41°C and up to 950 mL of pus was drained every 24 hours. At pleuroscopy, a large amount of malodorous pus and fibrin was evacuated. The lung appeared black, and it did not expand with controlled respiration. The pleuroscopy was converted to a thoracotomy, and gangrene of the whole lung and a fulminant inflammatory process involving the entire pleura were found. Cultures grew Bacteroides fragilis. Pneumonectomy was indicated, but dissection of hilar structures could cause mediastinitis. Therefore, preliminary fenestration was chosen as a temporizing procedure. The pleural cavity was packed with gauze soaked in Eusol (a bactericidal acid solution of sodium hypochlorite), and the chest was left open. Packs with Eusol were changed every 12 hours, and metronidazole was administered intravenously. Within 7 days, the pleural cavity became clean. Pneumonectomy was then performed with immediate closure of the chest. The histologic diagnosis was pulmonary gangrene. The postoperative course was uneventful. Over the next 5 years, the patient married and delivered 2 babies.
Patient 2 A 20-year-old man was hospitalized because of dyspnea, cough, temperature of up to 40.5°C, blood-stained sputum, and right-sided pleuritic pain lasting 10 days. The white blood cell count was 18,600/mL. Chest roentgenogram showed a large amount of liquid surrounding the right lung. Aspiration yielded pus from which Klebsiella 0003-4975/97/$17.00 PII S0003-4975(97)00837-0
Ann Thorac Surg 1997;64:970 – 4
REFAELY AND WEISSBERG GANGRENE OF LUNG
971
tient’s condition deteriorated, with chills, temperature rising to 41°C, and stupor. At pleuroscopy, 250 mL of foul-smelling pus was evacuated. The lung appeared gangrenous. Pneumonectomy was postponed in favor of a temporizing fenestration. Within 24 hours, the patient regained full consciousness and became cooperative. Within 48 hours, body temperature became normal. After 7 days, the open pleural cavity was clean; pneumonectomy was performed, and the chest was closed. Recovery was uneventful. Histologic examination confirmed gangrene of the lung. During the 4 years of follow-up, the patient did well.
Comment
Fig 1. (Patient 3.) Chest roentgenogram shows multiple infiltrates and areas of destruction in the right lung.
pneumoniae was grown. Treatment with cefotaxime sodium and closed thoracostomy drainage was started, but the patient’s condition deteriorated progressively, and he became stuporous. Five days after initiation of treatment, pleuroscopy was performed, and 620 mL of pus was evacuated. The right lower lobe appeared frankly necrotic. Although resection of the gangrenous lobe was indicated, it was postponed because of fulminant empyema, and fenestration was performed as a temporizing measure. The pleural cavity was filled with gauze soaked in Eusol, which was changed every 12 hours. During the following days, the infection cleared as the patient gradually recovered from sepsis. Eight days after the fenestration, the right lower lobe was resected, and the chest was closed. The postoperative course was uneventful. The patient did well during the 3-year follow-up.
Patient 3 A 40-year-old man who was a heavy smoker with a history of alcohol abuse had had recurrent bouts of pulmonary infections complicated with multiple abscesses of the right lung for 2 years. During the most recent episode of coughing, he expectorated fragments of necrotic lung tissue. On admission, he was in fair general condition with a body temperature of 38.5°C, a breathing rate of 22 per minute, a pulse rate of 84 per minute, and a blood pressure of 125/70 mm Hg. Coarse rhonchi and diminished breath sounds were heard over the right hemithorax; there were no other major findings. Chest roentgenogram and computed tomograms showed massive infiltrates in the right lung, areas of destruction, and a moderate amount of liquid in the pleural cavity (Figs 1, 2). These findings were compatible with necrotizing pneumonitis. Aspiration of pus from the right side of the pleura yielded anaerobes and Aspergillus fumigatus. Closed thoracostomy drainage was instituted, and treatment with metronidazole was started, but the pa-
Gangrene of the lung begins as a pulmonary infectious process, such as lobar pneumonia. Aspiration and devascularization with superimposed anaerobic infection is the presumed usual pathogenic process, and alcohol abuse is implicated as an important contributing factor [3]. The infecting organisms cause vascular thrombosis, resulting in loss of or gross reduction in blood supply to the pulmonary parenchyma. This brings about devitalization of the lung tissue, thus facilitating secondary infection and necrosis [4 –7]. Various pathogens have been found to be causative agents, especially Klebsiella pneumoniae, Friedla¨nder’s bacillus, polymicrobial anaerobic organisms, pneumococcus, and various strains of Aspergillus [1, 4, 8]. In the angioinvasive form of aspergillosis, mycotic invasion of vasa vasorum results in infarction of the media and secondary thrombus formation, leading to pulmonary infarction [8]. A different kind of mechanism has been described in Pseudomonas-induced pulmonary gangrene. Most strains of Pseudomonas produce proteolytic enzymes (exotoxins), which are capable of causing gangrene by a chemical process [8 –11]. When an extensive necrotizing process involves a large mass of pulmonary parenchyma, the gangrenous tissue is partially expectorated and partially accumulates in the cavity created by the process.
Fig 2. (Patient 3.) Computed tomography confirmed the findings shown in Figure 1. In addition, a moderate amount of liquid is visible in the pleural cavity.
972
REFAELY AND WEISSBERG GANGRENE OF LUNG
Danner and associates [4] described radiologic findings typical for pulmonary gangrene. The process begins as dense consolidation and enlargement of the affected lobe, causing outward bulging of the interlobar fissures. This is followed by breakdown into many small cavities, which in turn coalesce into a single large cavity occupying the entire lobe. The cavity is filled with fluid in which irregular pieces of sloughed lung parenchyma float like icebergs [4]. These radiologic findings have been observed also by others [1, 7, 8, 12] but are not invariably present. Hammond and co-workers [3] presented the cases of 2 patients with massive pulmonary gangrene. In both, there was extensive unilateral involvement, with initial dense consolidation followed by cavitation. However, the feature of coalescence into a large cavity with free-floating slough was not seen. Nor were these radiologic features observed in our 3 patients. In 2, the roentgenographic findings were those of empyema. In the other, multiple abscesses presented as areas of necrosis, but the coalescence into a single large cavity did not occur. Accordingly, although the radiologic features are helpful when present, they should not be considered an absolute requirement for the diagnosis of pulmonary gangrene. In their absence, gangrene of the lung can be difficult to diagnose, particularly in the presence of empyema, when opacification of the lung field obstructs changes in the pulmonary parenchyma. This difficulty can easily be overcome by pleuroscopy. We [13] have used this procedure for many years both to obtain information and to aid in treatment of patients with empyema resistant to treatment, and the results have nearly always been gratifying. In fact, pleuroscopy was the determining diagnostic procedure in the 3 patients described here. At pleuroscopy, the gangrenous lung appears black and does not expand with respiration. Once diagnosed, pulmonary gangrene must be treated promptly by resection of all necrotic tissue. Failure to do so is likely to result in sepsis, multiple-organ failure, and death [4]. In the collected series reviewed by Danner and colleagues [4], all 6 patients who were treated surgically, either by resection or by open drainage with spontaneous extrusion of the necrotic lobe, survived, whereas 4 patients treated medically died. Reports of medically treated patients surviving are an exception [12]. When the entire lung or lobe is involved, curative resection requires dissection of hilar vessels and bronchi. This is hazardous in the presence of severe purulent infection involving the entire pleura and can lead to mediastinitis or bronchopleural fistula. In addition, such patients are often too ill to undergo lobectomy or pneumonectomy at this stage. These difficulties can be overcome by dividing treatment into two stages: take care of the overwhelming infection first and then resect the gangrenous lobe or lung in a clean field. This approach was reported by Young and Samson [14] in 1980. The patient had gangrene of the right lower lobe, was close to death, and was unable to tolerate a lobectomy. She was treated by open drainage, with unroofing
Ann Thorac Surg 1997;64:970 – 4
of the abscess cavity and loose packing with iodoformimpregnated gauze. The wound was left open. On the 17th postoperative day, right lower lobectomy and decortication of the remaining lobes were performed, and the patient eventually recovered. Management of pulmonary gangrene by open drainage with resection of three ribs was also reported by Juettner and associates [15]. On the other hand, Hammond and colleagues [3] discussed 2 patients in whom gangrenous lung was resected without preceding open drainage. In both patients, empyema developed after the seemingly successful resection. We believe that by reversing the order of treatment and resecting the gangrenous lung in a clean field, postoperative empyema could have been avoided. Our 3 patients were treated in two stages, and all had good results. Postponement of resection for 1 week with the chest cavity open led to recovery from empyema and permitted resection of the gangrenous lung or lobe in a patient in stable condition and in a clean field. The operative risk was thus minimized, and the risk of causing mediastinitis was avoided. On the basis of our findings and those of others, we draw the following conclusions: 1. Untreated pulmonary gangrene leads to multipleorgan failure, sepsis, and death. Resection of all necrotic tissue is mandatory. 2. Dissection of hilar structures in the presence of pleural empyema can lead to mediastinitis or bronchopleural fistula and should be avoided. 3. Fenestration followed by delayed resection in a clean field and in a patient in stable condition is curative.
References 1. Penner C, Maycher B, Long R. Pulmonary gangrene. A complication of bacterial pneumonia. Chest 1994;105:567–73. 2. Weissberg D. Empyema and broncho-pleural fistula: experience with the open window thoracostomy. Chest 1982;82: 447–50. 3. Hammond JMJ, Lydell C, Potgieter PD, Odell J. Severe pneumococcal pneumonia complicated by massive pulmonary gangrene. Chest 1993;104:1610–2. 4. Danner PK, McFarland DR, Felson B. Massive pulmonary gangrene. Am J Roentgenol Radium Ther Nucl Med 1968; 103:548–54. 5. Zagoria RJ, Choplin RM, Karstaedt N. Pulmonary gangrene as a complication of mucormycosis. AJR 1985;144:1195– 6. 6. Chargan NB, Turk MG, Dhand R. The role of bronchial circulation in lung abscess. Am Rev Respir Dis 1985;131: 121– 4. 7. Phillips LG, Rao KVS. Gangrene of the lung. J Thorac Cardiovasc Surg 1989;97:114– 8. 8. Reich JM. Pulmonary gangrene and the air crescent sign. Thorax 1993;48:70– 4. 9. Hitschman F, Kreilich K. Zur Pathogenese des Bacillus pyocyaneus und zur Aetiologie des Ekthyma Gangraenosum. Wien Klin Wochenschr 1897;10:1093–107. 10. Liu PV. The roles of various fractions of Pseudomonas in its pathogenesis. II. Effects of lecithinase and protease. J Infect Dis 1966;116:112– 6. 11. Liu PV. Extracellular toxins of Pseudomonas aeruginosa. J Infect Dis 1974;130(Suppl O):S94 –9. 12. Padmanabhan K, Rajgopalan K, Yeo K, Dhar SR. Intracavity
Ann Thorac Surg 1997;64:970 – 4
mass in a patient with Klebsiella pneumonia. Chest 1988;93: 187– 8. 13. Weissberg D, Refaely Y. Pleural empyema: 24-year experience. Ann Thorac Surg 1996;62:1026–9. 14. Young JN, Samson PC. Pseudomonas aeruginosa septicemia
REFAELY AND WEISSBERG GANGRENE OF LUNG
973
with gangrene of the lung and empyema. Ann Thorac Surg 1980;29:254–7. 15. Juettner FM, Arian-Schad K, Kraus I, Gallhofer G, Popper H, Friehs G. Total unilateral lung gangrene in Hodgkin’s disease: treatment by thoracostomy. Ann Thorac Surg 1991;51:302–3.
INVITED COMMENTARY Livid, bloody, foul-smelling or billious sputum supervening in cases of continued fever is of bad significance. However if such expectoration removes the diseased tissues all may be well Hippocrates [1]
The association of lung abscess with a poor prognosis has been known since the age of Hippocrates. A lung abscess usually follows aspiration or a pneumonia occurring in a compromised patient. Sequelae depend on the volume and composition of the inoculum and the resistance of the host, which may be modified by malnutrition, anemia, the presence of underlying disease or damaged lung, and the adequacy of antibiotic therapy. Microabscesses or larger abscesses, single, multiple, or multiloculated may then develop depending on the extent of necrotizing gangrenous infection. The visceral pleura and a thin layer of subjacent lung usually survive owing to independent blood supply from the subpleural vascular plexus. A pleuritis usually results in the visceral pleura adhering to the parietal pleura, but occasionally progression is so rapid that an associated empyema develops as in the patients described by Refaely and Weissberg. Death of tissue or gangrene occurs to a variable extent in every lung abscess; when it is confined to a whole lobe or lung the term “pulmonary gangrene” may be applied. Use of this term does not imply that treatment is different, only that it is more extensive. When large areas of pulmonary gangrene are present, antibiotic therapy is ineffective because of poor diffusion of the drug into an avascular region; thus alternative therapy is necessary. In the early decades of this century when thoracic surgery was finding its feet, Neuhoff and Tourhoff [2] as well as Shaw [3] advocated drainage rather than resection of lung abscess with excellent results. With the advent of antibiotics, immunization, and chest physical therapy, lung abscess today is an infrequent problem in the First World. Open drainage as a form of management has been largely forgotten but periodically surfaces in the literature as advocated by Refaely and Weissberg. One of the largest series on lung abscess and drainage was published in the South African Journal of Surgery [4] by Postma and Le Roux of Durban, South Africa, from an institution where I trained. Two periods were compared: a series where resection was emphasized (268 patients) and a later period (417 patients) where drainage followed by later resection in selected patients was the usual first form of management. Roughly the same percentages (27.5% and 29%) had an operation. The mortality in the first series (resection) was 15.4% and the rate of complications was 34.6%; in the second series (drainage) the
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
corresponding results were 0.9% and 15.7%, strongly suggesting that drainage should be the first procedure in a toxically ill patient and that pulmonary resection should only be undertaken in the quiescent phase of the disease. The excellent results of Refaely and Weissberg confirm this and emphasize the importance of drainage in a sick septic patient before any major thoracic procedure. Although excellent results have been achieved in this small series of 3 patients, one must be careful not to extrapolate all of their conclusions into general practice. The statement that resection of all gangrenous tissue is mandatory and is life-saving implies that the second stage after drainage is necessary in all patients with this condition. This is not so: in some of our patients, especially those with Klebsiella pneumonia, drainage of the gangrenous lobe resulted in a large cavity with bronchi appearing like a stunted tree. Occasionally very rapid improvement in symptoms with a chest radiograph demonstrating either an empty space and in others the destroyed lobe obliterated by expansion of the remaining lobes allowed discharge. We used to call this autoamputation of the lobe. We did not proceed with pulmonary resection, indicating that effective drainage can sometimes result in cure. Obviously, if the whole lung is involved this is unlikely to occur. What is life-saving is the initial drainage procedure. Effective drainage of the empyema in the series of Refaely and Weissberg resulted in rapid improvement of their patients, but the source of infection— gangrenous lung—was left unmanaged in a desperately ill patient, begging the question, which can only be unanswered, whether the gangrenous lung should have been additionally surgically managed. Fenestration, or the creation of a large defect by multiple rib resection and folding skin inward, is usually reserved for patients with a chronic empyema, those unfit for any major procedure, or patients with a postpneumonectomy empyema, yet in Refaely and Weissberg’s patients pulmonary resection and closure of the defect occurred within 8 days. Although Refaely and Weissberg had no indication of response to drainage and the timing of the second procedure it does appear an excessive procedure and that similar drainage, with the same objective of complete evacuation of pus and fibrinous debris, may have been achieved with simple rib resection. It is stated that dissection of hilar structures in the presence of severe infection can lead to mediastinitis and should be avoided. There is no evidence to support such a statement. Although most patients with pulmonary gangrene have, because of the rapidity of the disease process, a free
0003-4975/97/$17.00 PII S0003-4975(97)00894-1