Bronchopleural fistula

This topic has now appeared five times in the exam papers:

The first of these, in 2014, asked the candidates to "outline the principles of, and strategies for management of a persisting broncho-pleural fistula (BPF) in a mechanically ventilated patient." The examiners felt the need to comment that overall "candidates had poor knowledge of this topic." The pass rate for Question 4 was 26%, and someone managed to score 6.5 marks. Then, in 2019, when it was repeated for the fourth time, the pass rate was 90%, suggesting that people have been paying attention. And when the pass rate dropped again in 2023 (to 51.5%), the examiners commented that "broncho-pleural fistula has appeared many times in the fellowship examination and candidates are advised to read widely around repeated topics". In short, they know that you know that they repeat this on purpose, and they want you to know that they know. 

As always, the LITFL entry on bronchopleural fistulae is the ideal last-minute revision resource.

The time-rich exam candidate may wish to explore the following published literature:

These resources have been remixed and pruned to form the summary made available below:

Pathophysiology of bronchopleural fistula

The clinical features that would make you suspicious of a fistula include:

  • Air leak (obviously) which is not resolving over more than 24 hrs
  • Pneumothorax on the CXR which is not resolving in spite of a properly placed chest drain (i.e. the lung has failed to reinflate)
  • The development of worsening pneumomediastinum of subcutaneous emphysema
  • Poor alveolar ventilation, i.e. the appearance of a large physiological dead space (as some  expired COescapes via the leak)

Normally, a pneumothorax bubbles for a while, and the the lung reinflates completely and the air leak stops as the pneumothorax is drained completely, and the lung fully re-expands. A bronchopleural fistula is by definition a persisting air leak into the pleural space. The definition of "persisting" is an air leak which continues for longer than 24 hours.

The leak can be assessed by:

  • Measurement of the difference between the inspiratory and expiratory volumes (if your ventilator does that)
  • Qualitatively, by looking at the pattern of bubbling, as in the Pierson article  which stratifies BPF severity according to bubble timing:
    • Bubbling during inspiration only
    • Bubbling during both inspiration and expiration
    • Bubbling during both inspiration and expiration with a detectable volume difference (at least 100-150ml leak per breath)
  • Fancy drains (eg. the Thopaz system) can measure the air leak quantitatively

Clinical features of bronchopleural fistula

Causes of bronchopleural fistula

A good article from 2010 (Lois et al) lists the following aetiologies associated with BPF:

  • Haemophilus influenzae
  • Streptococcus viridans
  • Staphylococcus aureus
  • Pseudomona aeruginosa
  • Klebsiella pneumoniae
  • Pneumococcus
  • Nonhemolytic streptococcus
  • Aspergillus
  • Histoplasma capsulatum
  • Gastroesophageal reflux disease with Barrett esophagus
  • Boerhaave syndrome
  • Broncholithiasis
  • Lung cancer
  • Thyroid cancer
  • Esophageal cancer
  • Lymphomas
Traumatic and iatrogenic
  • Thoracic trauma with tracheobronchial tree disruption
  • Post-surgical (eg. lobectomy or lung biopsy) - by far the most common cause!
  • Bougie intubation
  • Necrotizing lung disease associated with radiation or chemotherapy
  • ARDS
  • Ventilator-induced barotrauma
  • Excessive manual ventilation
  • Central line placement
  • Complication of bronchoscopy

Beyond the various lung-puncturing causes, there are also perpetuating factors. For instance:

  • High suction of the ICC
  • High tidal volumes
  • High inspiratory pressures

The management strategies are discussed below. I have attempted to synthesize the college answer to Question 4 from the first  paper of 2014 with the content of the abovelisted articles.

Management Strategies for Bronchopleural Fistula
Strategy Advantages Disadvantages
- large-bore drain
- or, multiple drains
- minimise suction
  • easy and readily available
  • Usually well tolerated
  • Does not interfere with weaning of ventilation
  • Risk of damaging more lung and creating larger leaks
  • Potentially, perpetuates the fistula by negative pressure suction
  • Invasive
Ventilator strategy:
- low VT
- low PEEP
- low resp rate
- short insp. time
- tolerate high PCO2
- wean rapidly
- extubate early
  • easy and readily available
  • Usually well tolerated
  • Early extubation is the ideal step to aim for, as spontaneous negative pressure breathing is better for BPF healing than positive pressure ventilation.
  • The BPF itself may frustrate weaning off ventilation
  • Mandatory mode may prolong ventilation time
  • Permissive hypercapnea may lead to respiratory acidosis, which is not ideal for the patient with traumatic brain injury
Independent lung ventilation
- dual-lumen tube
- or, bronch blocker
  • Isolation of one lung permits the selective low-volume low-pressure ventilation of the affected lung, and more rapid higher volume ventilation of the unaffected lung.
  •  PCO2 levels may be easier to control in this manner
  • Technically difficult: DLT insertion is one thing; running two ventilators is another.
  • There may be leak of gas and pressure from one lung to another if the seal is imperfect
  • Sedation requirements will  be higher, to tolerate the larger tube and the very unnatural respiratory pattern
  • Local pressure effects of the DLT are also more problematic
Surgical repair
  • The affected lung can be surgically repaired. USually, this means segmental lobectomy (for alveolar leaks) or patching and oversowing of the bronchial leak
  • Apparently, success rates are between 80 and 95%
  • It may be impossible to find the leak intraoperatively
  • It may be unfeasible to remove so much lung
  • It may be impossible if there are multiple leaks
  • The patient must tolerate one-lung ventilation
  • This approach requires thoracotomy
Bronchial stenting
  • The affected bronchus can be stented over bronchoscopically, thereby blocking the leak.
  • This is a minimally invasive alternative to surgical patch repair
  • You need to be sure of where the leak is
  • The leak must be in an accessible bronchus.
  • This may not work if there are multiple leaks
  • The procedure requires technical expertise
  • The patient must be fit to tolerate the bronchoscopy
Bronchial occlusion
  • Similarly to surgery, the affected pronchus is blocked with either a one-way valve or a plug. In fact, the Lois article lists options such as blood clot, cyanoacrylate glue, fibrin, lead shot,  gel foam, calf bone, and various others.
  • You need to be sure of where the leak is
  • The leak must be in an accessible bronchus
  • A major part of the lung may be sacrificed
  • The atelectatic lung may develop infection
Application of PEEP to the ICC
  • The equal intra and extrathoracic PEEP decreases the leak volume
  • Maintained intra-thoracic PEEP permits higher PEEP levels to be used
  • Drainage is compromised
  • There is a major risk of rapid tension pneumothorax
  • May reduce peak pressures
  • Certainly reduces tidal volume (to ~50ml)
  • Thus, theoretically reduces flow across the BPD, allowing it to heal
  • This is avery unnatural form of ventilation, and may be poorly tolerated
  • Large amounts of sedation or paralysis will be required
  • This may be the only option for severely hypoxic patients
  • With ECMO, one can limit or totally abolish gas flow through the BPF
  • All the risks of ECMO apply, as it is a maximally invasive therapy
  • It is not widely available.
  • There is little experience with this in BPF.

The air leak following lobectomy

 Question 3 from the second paper of 2017 presented the candidates with a post-lobectomy patient who is extubated, and who has a significant air leak. An excellent article by Mueller & Marzluf (2014) is used here as the main guide to the discussion. 

There are only two possible ways the drain could be producing an air leak:

  • It's coming from the patient's lung
  • It's coming from the extrathoracic air

Assessment of a post-lobectomy air leak should therefore consist of the following steps:

  • Examine the patient (an ABC based primary survey, ensuring that the patient is not about to die of pneumothorax or mediastinal shift)
  • Exclude drain dislodgement (examine the drain)
  • Quantify the leak:
    • Ask the patient to cough
    • Ask the patient to speak
    • Observe the quantity of bubbles.
    • If bubbling is constantly happening while the patient is speaking, the leak is significant.
    • If the bubbling is only present with cough and diminishes with ssutained coughing, the leak is probably small.
  • CXR to confirm pneumothorax
  • CT chest if the CXR is uninformative

Management options for a post-lobectomy air leak are numerous, ranging widely in their invasiveness. 

  • Do nothing. Keep the drain on suction and wait for it to settle down. Most do within 3 days of the surgery. Generally, about 50% of post-lobectomy patients will have some sort of air leak postoperatively, but this tends to settle down by day 3 or so. A "persistent air leak"  is then defined as an air leak after Day 5, which is the average length of stay post lobectomy. In  Question 3 from the second paper of 2017 the college did not give us a timeframe of how many days post-op the patient is, but the wording of the question sounds as if the patient has just returned from theatre, which would make waiting and watching a completely reasonable option. But, let's say it continues.
  • Decrease the level of suction. Cerfolio et al (1998) dropped the suction down from the usual 20cm H2O down to 10cm H2O,
  • Provocative chest drain clamping: the air leak may settle down more easily if the pressure in the pleural space is less negative. Clamping the drain and putting the patient on a high FiO2 may help denitrogenate the pneumothorax and allow the leak to heal.
  • Permissive chest tube removal: essentially an irreversible alternative to clamping the chest drain; the advantage is that by doing this one eliminates another hole in the chest cavity which might be bringing in extrathoracic air. 
  • Send them home like that. Outpatient management with a chest tube and a Heimlich valve is a legitimate option for patients whose survival from cancer (or existing comorbidities) are such that further major surgery or prolonged hospital stay would be counterproductive. It may be in their best interest to get discharged with a chest drain and a palliative care referral.
  • Blood patch is analogous to the same procedure performed for post-epidural CSF leaks. One ends up having to administer about 100ml of autologous blood via the chest drain. The sclerosant effect of blood is not as great as that of talc, and there is a risk of empyema. Combined with the dodgy evidence from small studies, these features make this an unpopular option
  • Talc pleurodesis via the chest drain  is an option for persisting air leaks. Cerfolio et al (1998) gave this to patients with an air leak around day 7 post op. 60ml of a water/talc slurry is injected via the chest drain. Alternative agents mentioned in the literature included tetracycline and silver nitrate
  • Pneumoperitoneum  is an option discussed by Mueller & Marzluf (2014); essentially the dome of the diaphragm on the affected side is punctured so that some gas can escape into the peritoneal cavity, thus (hopefully) obliterating the potential space in the chest cavity. It is actually a preventative measure which is most effective when performed intraoperatively, immediately following lobectomy. If not, then it is best done early, "when the lung still is mobile enough to shift to the apex of the chest cavity."
  • Intrabronchial valve which is placed bronchoscopically: essentially it is a one-way valve which allows gas to escape from the cavity but not into it. This is invasive, has the potential to dislodge, and can act as a nidus for infection. Moreover, you'd need to remove it a few weeks later.
  • Surgical revision is the ultimate solution for a leak which is not resolving with conservative measures in a patient suitable for surgery. The most aggressive solution (hopefulyl avoidable) would be a pneumonectomy. 

Cerfolio et al (1998) give a nice algorithm:

algorithm for management of post-lobectomy air leak (Cerfolio et al, 1998)

Basically, this team would keep the drain on suction (20cm H2O) until day 2 post op, and then turn the suction off. A CXR was then performed to assess for pneumothorax. Then, if there was a pneumothorax with an air leak, they would take the level of suction down to 10cm H2O.


Lois, Manuel, and Marc Noppen. "Bronchopleural fistulas: an overview of the problem with special focus on endoscopic management." CHEST Journal 128.6 (2005): 3955-3965.

Baumann, Michael H., and Steven A. Sahn. "Medical management and therapy of bronchopleural fistulas in the mechanically ventilated patient." CHEST Journal 97.3 (1990): 721-728.

Pierson, David J., et al. "Management of bronchopleural fistula in patients on mechanical ventilation." (2012) - from UpToDate.

Mueller, Michael Rolf, and Beatrice A. Marzluf. "The anticipation and management of air leaks and residual spaces post lung resection." Journal of thoracic disease 6.3 (2014): 271.

Cerfolio, Robert J., et al. "A prospective algorithm for the management of air leaks after pulmonary resection." The Annals of thoracic surgery 66.5 (1998): 1726-1730.