1978, British Journal of Radiology, 51, 494-497
Subpulmonary pneumothorax By A. Schulman, M.R.C.P., F.R.C.R., and R. B. Dalrymple, M.B., Ch.B., D.M.R.D., F.F.RAD(D) (S.A.) Departments of Radiology, Groote Schuur Hospital and University of Cape Town, South Africa {Received October, 1977 and in revised form February, 1978) ABSTRACT
Seven cases of subpulmonary pneumothorax are presented : four due to penetrating injury, two to blunt trauma and one to osteosarcoma metastasis. The typical and diagnostic appearance is a basal band of radiolucency bounded above by the thin hair-line of visceral pleura paralleling the dome of the hemi-diaphragm. When partially clotted blood is also present, the appearance becomes less typical and has to be differentiated from traumatic diaphragmatic herniation of bowel and from traumatic pneumatocoele by barium studies and by decubitus radiographs respectively. It is the bridge-like disposition of the pleural cavity between the dome of the hemi-diaphragm and the hollowed concavity of the lung base which allows pneumothorax to collect in it. It is rarely seen because blebs and bullae which are the commonest causes of pneumothorax are most often located in the upper zones.
lucency without septa or lung markings paralleling the upper surface of the hemi-diaphragm and separated from the lung by a thin hair-line representing the basal visceral pleura (Figs. 1-3). Two cases, due to penetrating and blunt trauma respectively, were atypical presumably due to the additional presence of blood in the pleural space. In
A recent report of four cases of subpulmonary pneumothorax (Christensen and Dietz, 1976) stressed the importance of pleural adhesions in the upper pleural space causing the pneumothorax to localize in a basal position. We wish to emphasize that subpulmonary pneumothorax can probably occur without pleural adhesions, that it may be especially common with basal penetrating trauma and that the thoracic bases and not simply the apices must be diligently searched for early small pneumothoraces. MATERIALS AND METHODS
During a recent period of eight months, we have seen seven cases of pneumothorax presenting initially in an entirely subpulmonary situation. Four of them were due to penetrating stab wounds of the left lower chest, two to blunt assault on the left side of the chest without radiographic or clinical evidence of fractured ribs, and one to pulmonary metastases from a femoral osteosarcoma. One patient had treated, healed tuberculosis in both upper lobes, but none of the others had clinical or radiographic evidence of pre-existing lung or pleural disease. In five cases, at least some of the subpulmonary air later moved into more common sites, lateral or medial to the lung and even eventually up to the apex; this occurred after lying the patient on his side for decubitus radiographs or simply with the passage of 24 to 48 hours. In five cases, the subpulmonary pneumothorax showed the typical appearance, i.e. a band of radio-
(A) Erect P.A. film.
(B) Erect lateral film. FIG. 1. Case 1. A 26-year old man received two stab wounds in posterior aspect of left hemithorax causing small subpulmonary pneumothorax. See line of visceral pleura (arrow-heads). Later the same day, a repeat film showed that the air had moved to the apex.
494
Subpulmonary pneumothorax
FIG. 2. Case 2. Erect P.A. film. A 31-year old man stabbed in back of left hemithorax causing small subpulmonary pneumothorax. Paralleling upper surface of diaphragm is the thin line of visceral pleura (arrows).
FIG. 3. Case 3. Erect P.A. film. A 17-year old boy seven months after amputation for femoral osteosarcoma. The pneumothorax on the left is purely subpulmonary (white triangles), that on the right is all around the lung. Whole lung tomography confirmed that the opacity at the left base (arrow-heads) was a metastasis projecting inwards from the visceral pleura and that there were multiple metastases in both lungs; proven by thoracotomy and histology. one, the blood and air formed a rounded collection with a fluid level above the stomach bubble (Figs. 4A & B). Barium examination of the stomach was entirely normal, while decubitus radiographs showed that the gas moved freely within the pleural space,
FIG. 4. Case 4. A 25-year old woman received four stab wounds in back of chest inflicted with screwdriver. In A and B, note the air-fluid collection (straight black arrows) at the left base above the stomach bubble (arrow-heads). Diagnosis—subpulmonary haemopneumothorax?, traumatic pneumatocoele?, traumatic hernia? Barium meal showed no gastric herniation. Decubitus film (c) shows the air moving freely upwards in the pleural cavity (open arrows) and the blood freely downwards (curved arrows). Diagnosis—subpulmonary haemopneumothorax.
^Bfapi
495
(c) Left lateral decubitus film with barium in stomach.
VOL.
51, N O . 607 A. Schulman and R. B. Dalrymple
thus excluding both traumatic herniation of the gastric fundus and traumatic pneumatocoele (Fig. 4c). In the other case (Fig. 5), the air collection at the base was crossed by numerous septa causing it to resemble bowel. Again, traumatic herniation was considered but barium examinations of stomach and colon were normal. Following drainage of the subpulmonary pneumothorax, chest radiographs were normal. The patient with osteosarcoma metastases had been treated by amputation, radio-therapy and cytotoxic drugs. Routine chest radiographs were normal until seven months after amputation when, although the patient was asymptomatic, there were bilateral pneumothoraces and pulmonary metastases on the
FIG. 5. Case 5. Erect PA film, left basal detail. A 24-year old man was assaulted receiving blunt skull and facial injuries with amnesia. No clinical or radiographic evidence of fractured ribs but this multi-septate air collection appeared at the left base. Contrast studies of stomach, jejunum and colon were normal, excluding traumatic herniation. Following basal pleural aspiration, chest radiographs became normal. Diagnosis—subpulmonary haemopneumothorax.
visceral pleura (Fig. 3). The right sided pneumothorax encircled the lung while that on the left side was purely subpulmonary. DISCUSSION
Radiologists are trained to look meticulously at the apices for the earliest evidence of small pneumothorax, and recent text books (Fraser and Pare, 1970; Crofton and Douglas, 1975; Harris and Harris, 1975) make no mention of pneumothorax located in the initial stages below the lung base even in erect patients. Our series indicates that, especially in cases of penetrating trauma, one should look most carefully for the fine hair-line of visceral pleura pushed upwards from the hemi-diaphragm by a subpulmonary pneumothorax. It is presumably the bridge-like disposition of the pleural space between the dome of the hemidiaphragm and the hollowed out base of the lung that allows pneumothorax, arising from injury or disease of this part of the lung, to collect here. Part of the reason why this is seen so seldom is that the major causes of pneumothorax, i.e. pleural blebs and emphysematous bullae, are situated most commonly in the upper parts (Crofton and Douglas, 1975). Neonatal pneumomediastinum is occasionally associated with air loculated below the lung and this is thought to be extra-pleural, between the parietal pleura and the hemi-diaphragm (Lillard and Allen, 1965; Caffey, 1972). However, none of our patients had pneumomediastinum and in five of them, some or all of the subpulmonary air later moved into more typical intrapleural sites around or above the lung after some time or after lying the patient on his side for decubitus radiographs. Indeed, it would be difficult for air to track far beneath the parietal pleura as it is found to be firmly adherent to underlying tissue during thoracic surgical dissections. It is also unnecessary to postulate the presence of pleural adhesions higher up causing a subpulmonary situation (Christensen and Dietz, 1976) as in six of our cases there was nothing to suggest previous pulmonary or pleural disease and again because of the later movement of some of the air into higher parts of the pleural space. The typical appearance of subpulmonary pneumothorax is a radiolucent zone without bronchovascular markings or septa lying on the hemi-diaphragm and bounded above by the thin hair-line of the visceral pleura paralleling the dome of the hemidiaphragm (Figs. 1-3). This appearance should be diagnostic but can be confirmed by the free movement of the air into other parts of the pleural space after lying the patient on his side.
496
JULY 1978
Subpulmonary pneumothorax A subpulmonary pneumothorax crossed by strands of partially clotted blood producing a multiseptate appearance (Fig. 5) can resemble basal bullae or the fold or haustral pattern of bowel herniated through a diaphragmatic laceration. The latter is easily excluded by barium study of stomach and colon (Fataar and Schulman, 1978). If the blood produces a fluid level within the loculated air (Figs. 4A and B) it can again resemble traumatic herniation as well as traumatic pneumatocoele (Fagan and Swischuk, 1976; Freed, 1977) but decubitus radiographs show independent movement of the air upward and the blood downward to opposite extremities of the pleural space thus proving that the collection is intrapleural and not intrapulmonary. The visceral pleura is much thinner than the hemi-diaphragm so that subpulmonary air should not be mistaken for subdiaphragmatic air. We are unable to account for the fact that all seven of our cases were left sided. It can be argued that when facing a right handed assailant, the left side of the chest is more likely to receive the stab wound. However, in three of the four cases of stab wounds, they were in fact in the back of the chest. Pulmonary metastases are a known cause of pneumothorax which for obvious reasons may be bilateral (Janetos and Ochner, 1963; Wright, 1976; Winter, 1976; Kai-Yui Yeung and Bonnet, 1977). The majority of published cases have been due to sarcomas of various origins, the commonest being osteosarcoma and most frequently in children and adolescents. Subpulmonary pneumothorax has not been reported. Our two cases due to blunt trauma had a pneumothorax in spite of no evidence of rib fracture or penetrating injury. The presence of rib fractures is
underestimated by radiological examination but even in the absence of fractures, blunt trauma can produce both lung damage and pneumothorax (Gerblich and Kleinerman, 1977). REFERENCES
CAFFEY, J., 1972. Pediatric x-ray diagnosis, Vol. I, 6th edn. pp. 431-434 (Year Book Medical Publishers). CHRISTENSEN, E. E., and DIETZ, G. W., 1976. Subpulmonic
pneumothorax in patients with chronic obstructive pulmonary disease. Radiology, 121, 33—37. CROFTON, J., and DOUGLAS, A., 1975. Respiratory diseases,
2nd edn. pp. 350-359 and 478-487 (Blackwell Scientific Publications, Oxford). FAGAN, C. J., and SWISCHUK, L. E., 1976. Traumatic lung
and para-mediastinal pneumatocoeles. Radiology, 120, 11-18. FATAAR, S., and SCHULMAN, A., 1978. The diagnosis of
diaphragmatic tears. British Journal of Radiology (in press). FRASER, R. G., and PARE, J. A. P., 1970. Diagnosis of diseases
of the chest, Vol. 1, pp. 371-376 (W. B. Saunders Company, Philadelphia). FREED, C , 1977. Traumatic lung cysts after penetrating chest injury (report of three cases). South African Medical Journal, 51,720-722. GERBLICH, A. A., and KLEINERMAN, J., 1977. Blunt chest
trauma and the lung (editorial). American Review of Respiratory Diseases, 115, 369-371. HARRIS, J. H., and HARRIS, W. H., 1975. The radiology of
emergency medicine, pp. 226-230. (Williams and Wilkins, Baltimore). JANETOS, G. P., and OCHNER, S. F., 1963. Bilateral pneu-
mothorax in metastatic osteogenic sarcoma. American Review of Respiratory Diseases, 88, 73-76. KAI-YUI YEUNG and BONNET, J. D., 1977. Spontaneous
pneumothorax with metastatic malignant melanoma. Chest, 71, 435-436. LILLARD, R. L., and ALLEN, R. P., 1965. The extrapleural
air sign in pneumomediastinum. Radiology, 85, 10931098. WINTER, W. G., 1976. Spontaneous pneumothorax heralding metastasis of adamantinoma of tibia. The Journal of Bone and Joint Surgery (American Volume), 52, 416-417. WRIGHT, F. W., 1976. Spontaneous pneumothorax and pulmonary malignant disease—a syndrome sometimes associated with cavitating tumours. Clinical Radiology, 27, 211-222.
Book review Computerised cranial tomography. Edited by B. Felson, pp. 146, illus., 1977 (New York; Grune and Stratton Inc.) $18.5O/£13-15. This book presents a collection of papers which were originally published in seminar in Radiolgoy Vol. 12. Each of the major papers is a good, well illustrated review of its subject. The presentation on equipment and physics by Tor-Pogossian stands out as a model of clarity in explaining this difficult field to diagnostic radiologists. This book is not meant to be comprehensive and there are obvious omissions as for example, the lack of consideration of hydrocephalus or the effects of surgery and other therapy. The defects are covered to some extent in the good selection of references. It is recommended as an authorative introduction to cranial computed tomography. B. E. KENDALL.
497
Subpulmonary pneumothorax By A. Schulman, M.R.C.P., F.R.C.R., and R. B. Dalrymple, M.B., Ch.B., D.M.R.D., F.F.RAD(D) (S.A.) Departments of Radiology, Groote Schuur Hospital and University of Cape Town, South Africa {Received October, 1977 and in revised form February, 1978) ABSTRACT
Seven cases of subpulmonary pneumothorax are presented : four due to penetrating injury, two to blunt trauma and one to osteosarcoma metastasis. The typical and diagnostic appearance is a basal band of radiolucency bounded above by the thin hair-line of visceral pleura paralleling the dome of the hemi-diaphragm. When partially clotted blood is also present, the appearance becomes less typical and has to be differentiated from traumatic diaphragmatic herniation of bowel and from traumatic pneumatocoele by barium studies and by decubitus radiographs respectively. It is the bridge-like disposition of the pleural cavity between the dome of the hemi-diaphragm and the hollowed concavity of the lung base which allows pneumothorax to collect in it. It is rarely seen because blebs and bullae which are the commonest causes of pneumothorax are most often located in the upper zones.
lucency without septa or lung markings paralleling the upper surface of the hemi-diaphragm and separated from the lung by a thin hair-line representing the basal visceral pleura (Figs. 1-3). Two cases, due to penetrating and blunt trauma respectively, were atypical presumably due to the additional presence of blood in the pleural space. In
A recent report of four cases of subpulmonary pneumothorax (Christensen and Dietz, 1976) stressed the importance of pleural adhesions in the upper pleural space causing the pneumothorax to localize in a basal position. We wish to emphasize that subpulmonary pneumothorax can probably occur without pleural adhesions, that it may be especially common with basal penetrating trauma and that the thoracic bases and not simply the apices must be diligently searched for early small pneumothoraces. MATERIALS AND METHODS
During a recent period of eight months, we have seen seven cases of pneumothorax presenting initially in an entirely subpulmonary situation. Four of them were due to penetrating stab wounds of the left lower chest, two to blunt assault on the left side of the chest without radiographic or clinical evidence of fractured ribs, and one to pulmonary metastases from a femoral osteosarcoma. One patient had treated, healed tuberculosis in both upper lobes, but none of the others had clinical or radiographic evidence of pre-existing lung or pleural disease. In five cases, at least some of the subpulmonary air later moved into more common sites, lateral or medial to the lung and even eventually up to the apex; this occurred after lying the patient on his side for decubitus radiographs or simply with the passage of 24 to 48 hours. In five cases, the subpulmonary pneumothorax showed the typical appearance, i.e. a band of radio-
(A) Erect P.A. film.
(B) Erect lateral film. FIG. 1. Case 1. A 26-year old man received two stab wounds in posterior aspect of left hemithorax causing small subpulmonary pneumothorax. See line of visceral pleura (arrow-heads). Later the same day, a repeat film showed that the air had moved to the apex.
494
Subpulmonary pneumothorax
FIG. 2. Case 2. Erect P.A. film. A 31-year old man stabbed in back of left hemithorax causing small subpulmonary pneumothorax. Paralleling upper surface of diaphragm is the thin line of visceral pleura (arrows).
FIG. 3. Case 3. Erect P.A. film. A 17-year old boy seven months after amputation for femoral osteosarcoma. The pneumothorax on the left is purely subpulmonary (white triangles), that on the right is all around the lung. Whole lung tomography confirmed that the opacity at the left base (arrow-heads) was a metastasis projecting inwards from the visceral pleura and that there were multiple metastases in both lungs; proven by thoracotomy and histology. one, the blood and air formed a rounded collection with a fluid level above the stomach bubble (Figs. 4A & B). Barium examination of the stomach was entirely normal, while decubitus radiographs showed that the gas moved freely within the pleural space,
FIG. 4. Case 4. A 25-year old woman received four stab wounds in back of chest inflicted with screwdriver. In A and B, note the air-fluid collection (straight black arrows) at the left base above the stomach bubble (arrow-heads). Diagnosis—subpulmonary haemopneumothorax?, traumatic pneumatocoele?, traumatic hernia? Barium meal showed no gastric herniation. Decubitus film (c) shows the air moving freely upwards in the pleural cavity (open arrows) and the blood freely downwards (curved arrows). Diagnosis—subpulmonary haemopneumothorax.
^Bfapi
495
(c) Left lateral decubitus film with barium in stomach.
VOL.
51, N O . 607 A. Schulman and R. B. Dalrymple
thus excluding both traumatic herniation of the gastric fundus and traumatic pneumatocoele (Fig. 4c). In the other case (Fig. 5), the air collection at the base was crossed by numerous septa causing it to resemble bowel. Again, traumatic herniation was considered but barium examinations of stomach and colon were normal. Following drainage of the subpulmonary pneumothorax, chest radiographs were normal. The patient with osteosarcoma metastases had been treated by amputation, radio-therapy and cytotoxic drugs. Routine chest radiographs were normal until seven months after amputation when, although the patient was asymptomatic, there were bilateral pneumothoraces and pulmonary metastases on the
FIG. 5. Case 5. Erect PA film, left basal detail. A 24-year old man was assaulted receiving blunt skull and facial injuries with amnesia. No clinical or radiographic evidence of fractured ribs but this multi-septate air collection appeared at the left base. Contrast studies of stomach, jejunum and colon were normal, excluding traumatic herniation. Following basal pleural aspiration, chest radiographs became normal. Diagnosis—subpulmonary haemopneumothorax.
visceral pleura (Fig. 3). The right sided pneumothorax encircled the lung while that on the left side was purely subpulmonary. DISCUSSION
Radiologists are trained to look meticulously at the apices for the earliest evidence of small pneumothorax, and recent text books (Fraser and Pare, 1970; Crofton and Douglas, 1975; Harris and Harris, 1975) make no mention of pneumothorax located in the initial stages below the lung base even in erect patients. Our series indicates that, especially in cases of penetrating trauma, one should look most carefully for the fine hair-line of visceral pleura pushed upwards from the hemi-diaphragm by a subpulmonary pneumothorax. It is presumably the bridge-like disposition of the pleural space between the dome of the hemidiaphragm and the hollowed out base of the lung that allows pneumothorax, arising from injury or disease of this part of the lung, to collect here. Part of the reason why this is seen so seldom is that the major causes of pneumothorax, i.e. pleural blebs and emphysematous bullae, are situated most commonly in the upper parts (Crofton and Douglas, 1975). Neonatal pneumomediastinum is occasionally associated with air loculated below the lung and this is thought to be extra-pleural, between the parietal pleura and the hemi-diaphragm (Lillard and Allen, 1965; Caffey, 1972). However, none of our patients had pneumomediastinum and in five of them, some or all of the subpulmonary air later moved into more typical intrapleural sites around or above the lung after some time or after lying the patient on his side for decubitus radiographs. Indeed, it would be difficult for air to track far beneath the parietal pleura as it is found to be firmly adherent to underlying tissue during thoracic surgical dissections. It is also unnecessary to postulate the presence of pleural adhesions higher up causing a subpulmonary situation (Christensen and Dietz, 1976) as in six of our cases there was nothing to suggest previous pulmonary or pleural disease and again because of the later movement of some of the air into higher parts of the pleural space. The typical appearance of subpulmonary pneumothorax is a radiolucent zone without bronchovascular markings or septa lying on the hemi-diaphragm and bounded above by the thin hair-line of the visceral pleura paralleling the dome of the hemidiaphragm (Figs. 1-3). This appearance should be diagnostic but can be confirmed by the free movement of the air into other parts of the pleural space after lying the patient on his side.
496
JULY 1978
Subpulmonary pneumothorax A subpulmonary pneumothorax crossed by strands of partially clotted blood producing a multiseptate appearance (Fig. 5) can resemble basal bullae or the fold or haustral pattern of bowel herniated through a diaphragmatic laceration. The latter is easily excluded by barium study of stomach and colon (Fataar and Schulman, 1978). If the blood produces a fluid level within the loculated air (Figs. 4A and B) it can again resemble traumatic herniation as well as traumatic pneumatocoele (Fagan and Swischuk, 1976; Freed, 1977) but decubitus radiographs show independent movement of the air upward and the blood downward to opposite extremities of the pleural space thus proving that the collection is intrapleural and not intrapulmonary. The visceral pleura is much thinner than the hemi-diaphragm so that subpulmonary air should not be mistaken for subdiaphragmatic air. We are unable to account for the fact that all seven of our cases were left sided. It can be argued that when facing a right handed assailant, the left side of the chest is more likely to receive the stab wound. However, in three of the four cases of stab wounds, they were in fact in the back of the chest. Pulmonary metastases are a known cause of pneumothorax which for obvious reasons may be bilateral (Janetos and Ochner, 1963; Wright, 1976; Winter, 1976; Kai-Yui Yeung and Bonnet, 1977). The majority of published cases have been due to sarcomas of various origins, the commonest being osteosarcoma and most frequently in children and adolescents. Subpulmonary pneumothorax has not been reported. Our two cases due to blunt trauma had a pneumothorax in spite of no evidence of rib fracture or penetrating injury. The presence of rib fractures is
underestimated by radiological examination but even in the absence of fractures, blunt trauma can produce both lung damage and pneumothorax (Gerblich and Kleinerman, 1977). REFERENCES
CAFFEY, J., 1972. Pediatric x-ray diagnosis, Vol. I, 6th edn. pp. 431-434 (Year Book Medical Publishers). CHRISTENSEN, E. E., and DIETZ, G. W., 1976. Subpulmonic
pneumothorax in patients with chronic obstructive pulmonary disease. Radiology, 121, 33—37. CROFTON, J., and DOUGLAS, A., 1975. Respiratory diseases,
2nd edn. pp. 350-359 and 478-487 (Blackwell Scientific Publications, Oxford). FAGAN, C. J., and SWISCHUK, L. E., 1976. Traumatic lung
and para-mediastinal pneumatocoeles. Radiology, 120, 11-18. FATAAR, S., and SCHULMAN, A., 1978. The diagnosis of
diaphragmatic tears. British Journal of Radiology (in press). FRASER, R. G., and PARE, J. A. P., 1970. Diagnosis of diseases
of the chest, Vol. 1, pp. 371-376 (W. B. Saunders Company, Philadelphia). FREED, C , 1977. Traumatic lung cysts after penetrating chest injury (report of three cases). South African Medical Journal, 51,720-722. GERBLICH, A. A., and KLEINERMAN, J., 1977. Blunt chest
trauma and the lung (editorial). American Review of Respiratory Diseases, 115, 369-371. HARRIS, J. H., and HARRIS, W. H., 1975. The radiology of
emergency medicine, pp. 226-230. (Williams and Wilkins, Baltimore). JANETOS, G. P., and OCHNER, S. F., 1963. Bilateral pneu-
mothorax in metastatic osteogenic sarcoma. American Review of Respiratory Diseases, 88, 73-76. KAI-YUI YEUNG and BONNET, J. D., 1977. Spontaneous
pneumothorax with metastatic malignant melanoma. Chest, 71, 435-436. LILLARD, R. L., and ALLEN, R. P., 1965. The extrapleural
air sign in pneumomediastinum. Radiology, 85, 10931098. WINTER, W. G., 1976. Spontaneous pneumothorax heralding metastasis of adamantinoma of tibia. The Journal of Bone and Joint Surgery (American Volume), 52, 416-417. WRIGHT, F. W., 1976. Spontaneous pneumothorax and pulmonary malignant disease—a syndrome sometimes associated with cavitating tumours. Clinical Radiology, 27, 211-222.
Book review Computerised cranial tomography. Edited by B. Felson, pp. 146, illus., 1977 (New York; Grune and Stratton Inc.) $18.5O/£13-15. This book presents a collection of papers which were originally published in seminar in Radiolgoy Vol. 12. Each of the major papers is a good, well illustrated review of its subject. The presentation on equipment and physics by Tor-Pogossian stands out as a model of clarity in explaining this difficult field to diagnostic radiologists. This book is not meant to be comprehensive and there are obvious omissions as for example, the lack of consideration of hydrocephalus or the effects of surgery and other therapy. The defects are covered to some extent in the good selection of references. It is recommended as an authorative introduction to cranial computed tomography. B. E. KENDALL.
497
