Journal Logo

Brief Reports

Hemopneumothorax After Transbronchial Lung Biopsy

Fujiwara, Miwa MD; Smith, Peter R. MD; Herman, Steven D. MD; Mathur, Ambrish MD; Wolf, Lawrence MD

Author Information
doi: 10.1097/LBR.0b013e31816060f1
  • Free


Pneumothorax complicates bronchoscopic transbronchial lung biopsy in 0.68% to 6.0% of procedures.1–5 Hemopneumothorax has not been reported after this procedure except in a single case report in Danish.6 We describe a patient who sustained a pneumothorax after transbronchial biopsy that progressed to a large hemopneumothorax 2 days later. We review the possible mechanisms for delayed occurrence of intrapleural hemorrhage after pneumothorax.


A 48-year-old African American woman was admitted with increasing shortness of breath for 5 months. She also complained of intermittent fevers, dry cough, night sweats, and 30-lb weight loss. Her past medical history included mild intermittent asthma and keloids, which were extensive, disfiguring, and painful. She smoked for a short time as a teenager, and occasionally drank alcohol. Her temperature was 100.2°F. There were bibasilar rales on auscultation of the lungs. Extensive, tender, keloids were present on her neck, chest, and upper back. Complete blood count and coagulation profile were normal. Serum chemistries were normal except for alkaline phosphatase of 531 U/L (30 to 125 U/L), alanine aminotranferease of 61 U/L (0 to 30 U/L), aspartate aminotransferase of 49 U/L (12 to 30 U/L), and lactate dehydrogenase of 472 U/L (100 to 250 U/L). Angiotensin-converting enzyme was 241 U/L (9 to 67 U/L). The chest radiograph showed bilateral hilar adenopathy and diffuse interstitial infiltrates. A computed tomographic scan of the chest showed bilateral, diffuse reticular-nodular densities, and bilateral hilar, mediastinal, and retroperitoneal adenopathy. Tuberculin skin test (5 TU) was negative. Sputum smears and cultures were negative for acid fast bacilli. Pulmonary function studies showed a mild to moderate restrictive ventilatory defect and mild reduction of diffusing capacity.

At bronchoscopy, the mucosa had a finely nodular appearance. Transbronchial biopsies were performed in the left upper and lower lobes with fluoroscopic guidance. Estimated bleeding was less than 10 mL. The histopathology showed fragments of lung parenchyma with numerous non-necrotizing granulomas. Special stains for microorganisms were negative. At the termination of the procedure, repeat fluoroscopy showed a small to moderate left-sided pneumothorax. She was asymptomatic and otherwise stable. Fearing the development of new keloids, she refused tube thorocostomy, and was managed with supplemental oxygen (100% via face mask) and daily chest radiographs.

The pneumothorax remained unchanged until the third day after bronchoscopy when the chest radiograph showed complete collapse of the lung and a large pleural effusion. Thoracentesis yielded grossly bloody fluid with a hematocrit of 18%. Tube thorocostomy was performed with immediate drainage of 1.6 L of bloody fluid. Chest radiograph showed reexpansion of the lung and a small amount of residual pleural fluid.

Over the next several days, she continued to drain 300 to 400 mL of bloody fluid daily. Her religious beliefs precluded transfusion and she received intravenous iron sucrose and darbepoetin. The hemoglobin level declined progressively to 6 mg/dL by the seventh day after bronchoscopy. She was scheduled for video-assisted thoracoscopy but the bleeding slowed and then ceased completely. The chest tube was removed and she went home on the ninth day after bronchoscopy. She continued to do well at follow-up and hemoglobin level returned to normal with supplemental oral iron therapy. She was treated with prednisone, initially 20 mg daily. Dyspnea improved, fevers resolved, and her weight returned to her normal baseline. Liver enzymes and angiotensin-converting enzyme improved. Chest radiographs showed resolution of intrathoracic adenopathy but the interstitial densities remained unchanged. Pulmonary function remained essentially unchanged.


The pneumothorax complicating bronchoscopic transbronchial lung biopsy in our patient was identified at the time of the procedure and was clearly iatrogenic. The hemothorax did not appear until 48 hours later and was accompanied by enlargement of the pneumothorax. The delayed occurrence of intrapleural hemorrhage makes it unlikely that the transbronchial biopsies per se were the cause of bleeding. It is far more likely that one of the mechanisms that have been implicated in the syndrome of spontaneous hemopneumothorax was involved.

Spontaneous hemopneumothorax occurs in 1% to 12% of all spontaneous pneumothoraces.7,8 Three mechanisms have been postulated as operative in this condition.8–10 The first results from a torn adhesion between the parietal and visceral pleurae. This is believed to be the most common mechanism accounting for 75% of cases.9 The site of bleeding is usually a small noncontractile vessel on the parietal pleura where vascular adhesions exist.11 The second mechanism is due to rupture of vascularized bullae and the underlying lung parenchyma. The third postulated cause is bleeding from torn, congenitally aberrant vessels between the parietal pleura and bullae.

In our patient, the etiology of hemothorax was probably similar to the first of the three mechanisms described above, that is, rupture of a vessel associated with a visceral-parietal adhesion. For the first 2 days after the pneumothorax occurred, there was no reduction in its size despite therapy with 100% oxygen. On the third day after bronchoscopy, the pneumothorax had progressed and this most likely led to intrapleural bleeding from disruption of an adhesion-associated small vessel. It is unlikely that either of the other 2 mechanisms for spontaneous hemopneumothorax were responsible. Both involve parenchymal bullous disease and this was not seen on computerized tomography.

In summary, this is the second reported case of hemopneumothorax complicating bronchoscopic transbronchial lung biopsy and the first in the English language literature. Because the onset of intrapleural hemorrhage was delayed by 2 days after the initial procedure-related pneumothorax, the mechanism for bleeding was most likely disruption of a small vessel in a torn visceral-parietal adhesion. This etiology is the most common cause of spontaneous hemopneumothorax.


1. Izbicki G, Shitrit D, Yarmolovsky A, et al. Is routine chest radiography after transbronchial biopsy necessary? Chest. 2006;129:1561–1564.
2. Pue CA, Pacht ER. Complications of fiberoptic bronchoscopy at a university hospital. Chest. 1995;107:430–432.
3. Hernández Blasco L, Sánchez Hernández IM, Villena Garrido V, et al. Safety of the transbronchial biopsy in outpatients. Chest. 1991;99:562–565.
4. Shure D. Transbronchial biopsy and needle aspiration. Chest. 1989;95:1130–1138.
5. Ahmad M, Livingston DR, Golish JA, et al. The safety of outpatient transbronchial biopsy. Chest. 1986;90:403–405.
6. Jacobsen VB, Laoutliev BN, Milman N. A rare complication to flexiblebronchoscopy with transbronchial biopsy. Ugeskr Laeger. 2001;163:7074–7075.
7. Hwong MT, Ng CSH, Lee TW, et al. Video-assisted thoracic surgery in spontaneous pneumothorax. Eur J Cardiothorac Sur. 2004;26:893–896.
8. Tatebe S, Kanazawa H, Yamazaki Y, et al. Spontaneous hemopneumothorax. Ann Thorac Surg. 1996;62:1011–1015.
9. Hsu NY, Shih CS, Hsu CP, et al. Spontaneous hemopneumothorax revisited: clinical approach and systemic review of the literature. Ann Thorac Surg. 2005;80:1859–1863.
10. Wu YC, Lu MS, Yeh CH, et al. Justifying video-assisted thoracic surgery for spontaneous hemopneumothorax. Chest. 2002;122:1844–1847.
11. Barry JJ, Pfitzner J, Peacock MJ. Video-assisted thoracoscopy for spontaneous haemothorax. Anaesth Intensive Care. 1995;23:354–357.

hemopneumothorax; pneumothorax; transbronchial lung biopsy; flexible bronchoscopy

© 2008 Lippincott Williams & Wilkins, Inc.