Special Feature

Managing Bronchopleural Fistula During Mechanical Ventilation

By David J. Pierson MD, Editor, Pulmonary and Critical Care Medicine, Harborview Medical Center, University of Washington, is Editor for Critical Care Alert

Development of a Bronchopleural Fistula (BPF) in a patient receiving mechanical ventilation is a serious complication that causes concern on the part of caregivers and often prompts a variety of changes in management. This essay explains what is meant by a BPF in this context, summarizes the possible ways it may occur in the ventilated patient, discusses the potential implications for the patient, and presents an approach to management based on both logic and the available evidence.1,2

What is a BPF?

With very few exceptions, detection of a pneumothorax in a mechanically ventilated patient should be followed promptly by insertion of a chest tube and the application of external suction. Evacuation of air from the pleural space is shown by bubbling through the water seal of the chest drainage device. If there is no direct communication between the airways and the pleural space, this bubbling ceases within an hour or two once the lung is fully reinflated. When the bubbling continues for 24 hours or more, a BPF may be said to be present. Most such leaks consist of only a few bubbles escaping through the water seal during the ventilator’s inspiratory phase; although in a few cases the leak persists through both inspiration and expiration, and its volume may reach several hundred milliliters per breath.

Persistent bronchopleural air leak would be a more descriptive term, and would avoid confusion from the common implications of inflammation and suppuration associated with the word fistula in surgery and other settings. However, the term BPF has come to denote any air leak during mechanical ventilation, and its use has become so widespread that change is unlikely.

Bronchopleural fistula is a relatively uncommon complication, even in institutions managing large numbers of patients with the acute respiratory distress syndrome (ARDS) and other forms of severe acute respiratory failure. Of 1700 patients ventilated during one 4-year period at a major trauma center in the early 1980s, 39 (2%) developed a bronchopleural air leak that persisted at least 24 hours after chest tube insertion.3 Tidal volumes and minute ventilations were considerably larger when that series was collected than those that are recommended today, but, as noted subsequently, whether this has reduced the incidence of BPF among ventilated patients, remains to be seen.

What are the Possible Causes of a BPF in the

Ventilated Patient?

There are 3 main mechanisms that can result in a BPF. The first is airway disruption or alveolar rupture prior to the initiation of ventilatory support. Chest trauma is a common example, in which the airway injury can be either direct, as with laceration of the lung by fractured ribs or a penetrating object, or as a result of blunt injury with transient alveolar overdistension and rupture. Partial or complete rupture of a central airway is an uncommon but serious form of blunt deceleration injury, usually in patients with other major thoracic trauma such as great vessel injury or fractures of the first rib or scapula; the air leak in such cases is usually (but not always) massive, typically with persistent lung collapse despite pleural suction. Alveolar rupture can also occur during over-zealous mouth-to-mouth resuscitation or manual ventilation. The airway may be lacerated during attempted intubation, or the visceral pleura may be punctured during central line placement, thoracentesis, or tube thoracostomy.

The second potential mechanism leading to a BPF is direct laceration of visceral pleura or airway while the patient is receiving mechanical ventilation. Most commonly this occurs during attempted central line placement, but it can also complicate thoracentesis, tube thoracostomy, transbronchial biopsy or brushing, or other airway procedures.

Finally, BPF in the ventilated patient may follow spontaneous alveolar rupture, the latter either as a manifestation of the primary disease process (as in necrotizing pneumonia) or from inadvertent alveolar overdistension. A classic scenario is accidental intubation of the right mainstem bronchus, with atelectasis of the left lung and overdistension of the right lung leading to pneumothorax and subsequent BPF. Even without right bronchial intubation, overzealous manual ventilation or the use of high levels of positive end-expiratory pressure (PEEP) may distend alveolar regions to the point of rupture, with subsequent BPF.

Does ventilation with lower tidal volumes reduce the incidence of BPF? Although the evidence is compelling that lung-protective, low-tidal-volume ventilation improves survival and other outcomes in patients with acute lung injury and ARDS, evidence that this approach leads to fewer instances of barotrauma, including BPF, is lacking. Boussarsar and colleagues4 reviewed the findings of 11 studies (2270 patients) reporting the incidence of barotrauma in patients with ARDS. These studies varied with respect to patient population and ventilator management, and reported incidences of barotrauma that varied from zero to 76 percent. However, the authors found that end-inspiratory plateau pressure was the only ventilator management-related variable that correlated statistically with the occurrence of barotrauma. There were no significant correlations with tidal volume (either absolute or expressed in mL/kg), PEEP, or peak inspiratory pressure.

In an international study of 5183 mechanically ventilated adult patients with a wide variety of diagnoses, Anzueto et al5 found no correlation between any ventilator setting or pressure measurement and the development of barotrauma. Despite the substantial and increasing number of other studies on ventilator management in ARDS, none to date has directly shown a relationship between how the ventilator is set and the subsequent development of barotrauma, including BPF.

In What Ways is a BPF Harmful?

The presence of a bronchopleural air leak following insertion of a chest tube could potentially create several important clinical problems. If the lung fails to completely re-expand, atelectasis and ventilation-perfusion mismatching may worsen hypoxemia; this is especially a problem with very large leaks (eg, several hundred mL per breath), or underlying pulmonary fibrosis or other condition causing increased elastic lung recoil. Loss of a large proportion of each delivered tidal volume via the BPF may impair expansion and ventilation of other lung areas, thus also contributing to ventilation-perfusion mismatching. Hypoxemia may also be worsened if a large air leak prevents the effective application of PEEP in patients with acute lung injury.

While it is logical to assume that a large BPF would compromise alveolar ventilation and impair CO2 removal, this turns out not to be the case, at least in ARDS. During its passage through the lung, the leaked volume actively participates in gas exchange-and in fact, may contain more CO2 than the gas exiting through the endotracheal tube.6 Thus, except in the case of a large proximal airway disruption, hypercapnia in the presence of a BPF is more a reflection of general gas exchange derangement than of the air leak per se.

Depending on the mode employed and how the ventilator is set, a large air leak can interfere with effective ventilatory support. In the assist-control mode, high pleural suction pressures can be transmitted to the central airways and cause factitious triggering. On the other hand, depending on the ventilator brand and the inspiratory flow cut-off threshold used in the pressure support mode, a large bronchopleural air leak may prevent the inspiratory phase from terminating.

Another way in which a BPF can potentially be harmful to the patient is by permitting airway organisms to reach the normally sterile pleural space, leading to clinical infection and further impeding closure of the leak. Although this is an obvious threat in patients with bacterial pneumonia, airway colonization without overt lung infection could also give pathogenic organisms access to the pleural space. As with other aspects of this topic, there is little or no experimental evidence to support or refute these statements.

How Should a BPF Be Managed?

In this era of evidence-based medicine, there is essentially no high-quality evidence in which specific interventions to reduce the volume of gas leaked through a BPF lead to better patient outcomes. Most of the literature consists of descriptions of techniques and anecdotal descriptions of immediate and short-term physiological changes. Common to much of this literature is a lack of convincing evidence that "conventional" ventilator management had been inadequate before the experimental intervention was used. The subjects of most published reports have been desperately ill patients who died in spite of the short-term physiologic improvements associated with the touted intervention.

Several reports have described the use of independent lung ventilation via a double-lumen endotracheal tube and two ventilators.2 As with the use of this ventilatory technique in other settings, improved gas exchange has generally been achieved. The total reported experience is with only about a dozen patients. Placement and maintenance of a double-lumen endotracheal tube in a critically ill patient require considerable expertise, and the small lumens of these tubes make bronchial hygiene difficult in patients with lots of airway secretions.

Bronchopleural fistula is one of the primary clinical settings in which high-frequency jet ventilation has been used, although enthusiasm has waned and there have been few new reports of the use of this technique in the last 20 years.2 Clinical experience with jet ventilation in patients without underlying lung disease, as in traumatic bronchial disruption or during tracheobronchial surgery, has generally been positive; however, both short-term and outcome results have been discouraging when BPF occurs in ARDS or other diffuse disease. High-frequency oscillatory ventilation has also been tried in BPF, but whether this technique offers any advantage over adjustments in conventional ventilation remains to be seen.

Several approaches have been reported for decreasing the size of the air leak through manipulation of the pleural drainage system, although again in very few patients.2 Their application becomes progressively more difficult as the number of chest tubes increases in a given patient, and incomplete lung expansion on the side of the leak is a common problem. Plugging or otherwise sealing BPFs using the flexible or rigid bronchoscope (with laser therapy, cauterization, mechanical plugs, Gelfoam, tissue glue, and other things) has been reported in numerous case reports and small series.2 However, only about 10% of the patients in these reports have had BPFs in the setting of ventilator-associated barotrauma; most of the reported experience is with BPFs following lung resection.

Although it is tempting to think that the ultimate therapy for a BPF would be direct surgical closure, this is seldom possible for technical reasons in settings other than acute traumatic tracheobronchial disruption. If the fistula is localized and associated with a necrotizing pneumonia, resection of the affected lobe may be feasible. A BPF after open lung biopsy or other procedure may be amenable to surgical repair. In most instances of barotrauma, however, the air leak is a manifestation of the severity of the underlying pulmonary disease, and direct suture or cautery of the leak or leaks is simply not technically feasible.

For most patients, ventilatory management in the presence of a BPF is the same as if the air leak were not there, provided that a functioning chest tube and pleural drainage system are in place. In the majority of instances, a BPF can be thought of as an indicator or manifestation of the severity of a patient’s illness rather than as a condition requiring specific treatment. Sound general management is more important than any specific measure directed at the leak itself. The accompanying table provides a guide to management, based on pathophysiology, the author’s experience, and the available literature, that incorporates both general measures and an approach to specific manipulations for controlling the leak.1,2

Bronchopleural Fistula in the Ventilated Patient: 10 Principles of Management
  1. Base ventilator management primarily on the patient’s overall condition, rather than specifically on the BPF, since the latter will nearly always resolve as the former improves
  2. Use the lowest number of mechanical breaths that permits acceptable alveolar ventilation
    1. Reduce both mean airway pressure and number of high-pressure breaths per minute
    2. Discontinue ventilatory support completely if possible, even if patient must remain intubated
    3. Consider pressure support or other form of partial (as opposed to full) ventilatory support
    4. Avoid respiratory alkalosis (PaCO2 < 40 mm Hg), in order to minimize minute ventilation
    5. Unless contraindicated (eg, by intracranial hypertension, ongoing cardiac ischemia, or serious arrhythmias), consider reducing minute ventilation even further, with permissive hypercapnia
  3. Use low tidal volumes (eg, 6 mL/kg returned volume)
  4. Minimize inspiratory time
    1. Keep inspiration:expiration ratio low (eg, 1:2)
    2. Use high inspiratory flow (eg, >70 L/min)
    3. Avoid end-inspiratory pause and inverse-ratio ventilation
    4. Use low-compressible-volume, non-disposable ventilator circuit, to reduce the tidal volume the ventilator must generate to produce
    target (corrected) tidal volume
  5. Minimize total PEEP (both set PEEP and auto-PEEP)
  6. Use the least amount of chest tube suction that maintains lung inflation
  7. Explore positional differences, and avoid placing patient in positions that exacerbate the leak
  8. Treat bronchospasm and other causes of expiratory airflow obstruction
  9. Consider specific or unconventional measures only if the patient remains unstable, develops clinically harmful, uncorrectable respiratory acidosis despite above measures, or if the lung fails to completely re-expand
    1. Independent lung ventilation
    2. Surgical closure (not usually technically feasible)
    3. Endobronchial measures (see text)
  10. Treat underlying cause of respiratory failure, maintaining nutritional and other support, with goal of discontinuing mechanical ventilation as soon as possible

Adapted from references 1 and 2


  1. Pierson DJ. Management of bronchopleural fistula in patients on mechanical ventilation. UpToDate On Line, version 13.2, 2005. (www.uptodate.com)
  2. Pierson DJ. Barotrauma and bronchopleural fistula. In: Tobin MJ, ed: Principles and practice of mechanical ventilation. New York, McGraw-Hill, 1994:813-836. [2nd edition in press]
  3. Pierson DJ, et al. Persistent bronchopleural air leak during mechanical ventilation: A review of 39 cases. Chest 1986;90:321-323.
  4. Boussarsar M, et al. Relationship between ventilatory settings and barotrauma in the acute respiratory distress syndrome. Intensive Care Med. 2002.28:406-413.
  5. Anzueto A, et al. Incidence, risk factors and outcome of barotrauma in mechanically ventilated patients. Intensive Care Med. 2004 Apr;30(4):612-619.
  6. Bishop MJ, et al. Carbon dioxide excretion via bronchopleural fistulas in adult respiratory distress syndrome. Chest 1987;91:400.