Secondary Logo

Journal Logo

Case Reports

Bronchopleural Fistulae and Pulmonary Ossification in Posttraumatic Acute Respiratory Distress Syndrome: Successful Treatment With Extracorporeal Support

Bombino, Michela*; Patroniti, Nicolò*†; Foti, Giuseppe*; Isgrò, Stefano*; Grasselli, Giacomo*; Pesenti, Antonio*†

Author Information
doi: 10.1097/MAT.0b013e31821d8182
  • Free


Bronchopleural fistulae (BPF) in the setting of posttraumatic acute respiratory distress syndrome (ARDS) can result directly from chest trauma or be an aspect of ventilator-induced lung injury (VILI).1

The ventilatory strategy of ARDS, directed to alveolar recruitment, conflicts with the best management of BPF, which requires minimization of airway pressure to decrease the air leak. In ARDS, moreover, the diffuse involvement of both lungs further compromises the chances for a successful control of major air leaks.2,3

Several years ago, we reported4 the case of a patient with bullous emphysema and bilateral air leaks in whom partial extracorporeal carbon dioxide removal (ECCO2R)5 allowed us to restore spontaneous breathing, leading to the resolution of recurrent pneumothoraces. We describe in this study a patient with posttraumatic ARDS complicated by high-output bilateral BPF and pulmonary ossification. In this patient, we applied ECCO2R after the failure of other conservative measures.

Case Report

A 24-year-old man was involved in a motorcycle accident. At the scene, he was tachycardic, hypotensive, tachypnoeic, and hypoxemic (pulse oximetry <70%). He was intubated and transported to the closest hospital. In the emergency department, a right pneumothorax, bilateral lung contusions (Figure 1), left acetabulum and clavicle fractures, and luxation of left shoulder and left femur were diagnosed, and a right chest tube was inserted.

Figure 1.
Figure 1.:
Chest x-ray and CT at admission showing bilateral pulmonary contusion with pneumatoceles and a right pneumothorax (chest tube visible on CT). CT, computed tomography.

At intensive care unit (ICU) admission, blood gases showed severe hypoxemia and respiratory acidosis (Pao2: 64.7 mm Hg, Paco2: 72.5 mm Hg, and pH 7.24) while on pressure control ventilation (PCV) with tidal volume (TV) of 670 ml, positive end-expiratory pressure (PEEP): 11 cm H2O, respiratory rate (RR): 22 breaths/min, and Fio2: 1. An attempt to increase PEEP up to 17 cm H2O failed because of hemodynamic compromise. A Swan-Ganz catheter was inserted and showed pulmonary hypertension (mean pulmonary artery pressure, 39 mm Hg); the intrapulmonary shunt fraction (Qs/Qt) was >0.5. On day 2, another chest tube was inserted on the left side, and the ventilatory modality was changed to pressure support ventilation (PSV), maintaining an elevated mean airway pressure (Paw) of 25–27 cm H2O. Despite this, the patient remained severely hypoxemic, and Paco2 progressively increased up to 80 mm Hg, with persistent air leaks and recurrent bronchial bleedings. On the 6th ICU day, inhaled nitric oxide (iNO) was administered to improve oxygenation after an unsuccessful prone positioning test. To increase CO2 clearance, intratracheal gas insufflations (TGI) was also initiated, but on day 22, because of further gas exchange derangement, the patient was paralyzed and switched to high-frequency PCV. Main gas exchange parameters and ventilatory strategies are reported in Figure 2A.

Figure 2.
Figure 2.:
A: Main gas exchange parameters before ECCO2R (Pao2/FiO2 mm Hg [•]; Paco2 mm Hg [•]; and mean airway pressure [Paw; s]). Vertical lines separate different periods of ventilation. Inhaled nitric oxide (iNO) and Tracheal gas insufflation (TGI) administrations are reported. B: Main oxygenation parameters during ECCO2R. On the x-axis, extracorporeal days are reported with day 0 being before connection. On the y-axis, main determinants of oxygenation during bypass, extracorporeal blood flow, range 2.2–2.7 L/min, expressed in percent of cardiac output (BF%), and mean airway pressure (Paw and cm H2O) are reported altogether with oxygenation parameters (VO2NL%, amount of oxygenation accomplished by the natural lung; VO2 ML%, amount of oxygenation carried out by the artificial oxygenators; Qs/Qt, shunt fraction). PSV, pressure support ventilation; HFV, high frequency ventilation; APRV, airway pressure release ventilation; SIMV, synchronized intermittent mandatory ventilation; PCV, pressure controlled ventilation.

He was transferred to our institution 24 days after admission, while in PCV (RR: 60, PEEP: 23 cm H2O, Pinsp: 35 cm H2O, Fio2: 1, iNO: 15 ppm; Pao2: 73.5 mm Hg, Paco2: 84.9 mm Hg, and pH 7.39). Massive subcutaneous emphysema was obvious, with four chest tubes in place, and air leaks from BPF approached 70%–80% of the inspired TV. Initially, we modified the modality of ventilation using low pressure control (PC) (+8–10 cm H2O), very high RR (110/min), and high PEEP (25 cm H2O), with the aim of keeping an elevated mean Paw of 30–32 cm H2O. This strategy led to an improvement in oxygenation (which allowed us to reduce Fio2 to 0.6 and to discontinue iNO) but was associated with hyperinflation of localized lung areas and further increase of BPF air leak.

Subsequent attempts (airway pressure release ventilation [APRV], synchronized intermittent mandatory ventilation [SIMV] + pressure support [PS] with a bias flow to avoid autotriggering, and PCV with reduced PEEP) to decrease Paw failed due to worsening hypercapnia and acidosis(Figure 2A). To decrease the air leaks, we tried at last to apply a positive pressure on the chest drains, without any significant improvement. At this time, a computed tomography (CT) scan showed the development of pulmonary ossifications, mainly located in the ventral part of the apexes (Figure 3A).

Figure 3.
Figure 3.:
A: CT taken 1 week before ECCO2R showing hyperinflation of bullae and pulmonary ossifications prominent in the upper lung fields. B: CT during ECCO2R showing a decrease in size of the bullae but also collapse of lung parenchyma. C: CT at 6 month follow-up showing reduction of calcified areas. ECCO2R, extracorporeal carbon dioxide removal; CT, computed tomography.

On the 70th day postinjury, we decided to connect the patient to an ECCO2R circuit, and the goal was to support lung function while allowing restoration of spontaneous breathing at lower airway pressures, thus creating optimal conditions for lung healing and BPF closure. Venovenous ECCO2R was instituted by percutaneous cannulation of the femoral veins (21 F BioMedicus cannulae, peristaltic SARNS pump, 35 ml Avecor R-38 reservoir, raceway supertygon Avecor TBNG #565, Medtronic Carmeda® bonded tubing, and two Avecor Membrane Oxygenators with integral heat exchanger I-4500-2A), and anticoagulation was maintained by heparin infusion titrated to maintain activated clotting time (ACT) (Hemochron, Cremascoli & Iris srl, Milan) in the range 180–220 s.5 Blood gases before ECCO2R institution showed Pao2: 85, Paco2: 104, and pH 7.33 while on PCV, Fio22: 0.9, RR: 35, TV: 420, and Paw: 12 cm H2O. During the first week, PEEP (initially set equal to Paw) was progressively decreased from 12 to 4 cm H2O; paralytic drugs were suspended and sedatives reduced to allow resumption of spontaneous breathing, and the modality of ventilation was switched to PSV, whereas the patient's respiratory drive was controlled by modulating the amount of CO2 removal through the artificial lung (membrane lung). This led to an almost complete resolution of air leaks without adverse consequences on oxygenation (Figure 2B).

On the 14th ECCO2R day, a new thoracic CT scan showed substantial derecruitment of the lung (Figure 3B). Despite this, we observed a constant improvement in lung function (Figure 2B), with the patient becoming able to clear increasing amounts of CO2 through his natural lung without distress. He was disconnected from bypass after 23 days, and in two additional weeks, he was completely weaned from the ventilator. On day 110 posttrauma, the patient was transferred to the ward and subsequently to a respiratory rehabilitation center, from which he was discharged after 3 months.


We describe in this study the clinical course of a patient with severe posttraumatic ARDS complicated by major BPFs. A key point in the patient's history is the late referral to our institution and the attempt to maintain spontaneous breathing in the acute phase, with high transpulmonary pressures despite low TV, which could have played a role in the development of persistent BPF and VILI. The decision to initiate a respiratory extracorporeal support was delayed due to the initial rapid improvement in oxygenation observed during the first weeks. In the majority of extracorporeal membrane oxygenation (ECMO) centers, a duration of MV of more than 7 days is considered a contraindication to initiate the extracorporeal support. We agree that early patient's referral is a crucial point in the management of ARDS, but we also believe that the duration of MV is only one of the factors that have to be considered. The decision to institute an extracorporeal respiratory support is really challenging and should be made on a patient-by-patient basis, weighting the expected benefits against the possible complications (mainly hemorrhagic5) associated with the technique. In the past 10 years, improvement in ECMO technology simplified its application. Interventional lung assist (iLA) device was not available at the time we treated this patient, but also now we prefer to have a pump-driven venovenous ECMO circuit because of the possible changes in support requirement.

At ECCO2R connection, the air leaks from chest tubes were 70%–80% of TV, PaCO2 was steadily increasing, and oxygenation was deteriorating. Furthermore, attempts to reduce the BPF output with the application of a glue on the leaking points under bronchoscopic guidance could not be done as bronchoscopy showed that both the main superior and middle bronchi would have to be sacrificed resulting in a severe oxygenation impairment. Airway stenting was not in use at that time, but it works only with proximal BPF. Peripheral BPF might benefit of a surgical approach, but the available literature data are scarce and limited to postthoracotomy ARDS (sometimes ECMO adjunct has been used during rethoracotomy). The presence of bilateral BPFs in our case precluded a surgical approach; moreover, surgical treatment of BPFs can be considered only after the control of pleuropulmonary infections, and this was not the case of our patient who had a documented fungal superinfection (confirmed by isolation of Cunighamella bertholletiae in bronchoalveolar lavage [BAL]).

Therefore, we decided to apply ECCO2R mainly to avoid worsening of lung damage due to injurious mechanical ventilation and to buy time for superinfection cure. After connection, the patient was allowed to regain spontaneous breathing: our aim was to minimize mean Paw to favor the closure of BPF. This low airway pressure strategy was associated with suboptimal recruitment of the lung (Figure 3B) and consequently with a transient decrease in the natural lung oxygenation capacity. On the other hand, however, it had a major effect in promoting lung healing, ultimately leading to a progressive improvement in oxygenation.

The patient's attempts to breath were initially ineffective in terms of CO2 clearance. The passage from total to partial ECCO2R6 was very slow to avoid respiratory fatigue and to optimize patient-ventilator interaction. Careful nursing, extended verbal contact, and prolonged presence of the relatives were the necessary key to patient's reassurance in a very distressful environment. At the end of this process, the patient, still on partial extracorporeal support, could watch television and read cartoons.

Another aspect of this patient's history that deserves a comment is the appearance of pulmonary ossification. Ossification of lung parenchyma is a rare evolution in ARDS. We can speculate that in our patient several factors might have contributed to its development, such as the presence of huge amount of blood in the airways, a documented fungal infection, and the milieu related to the huge air leaks itself.7 During rehabilitation, a 99mTechnetium-Methyl diphosphonate bone Single Photon Emission Computed Tomography (SPECT) showed active accumulation of the osteotropic tracer in the lung parenchyma. A follow-up CT scan obtained at 6 months documented restoration of a quite normal lung parenchyma and partial regression of the ossifications (Figure 3C).

In conclusion, we applied ECCO2R in this patient with severe ARDS as we were otherwise unable to maintain viable gas exchange because of the large BPF. ECCO2R allowed us to reinstitute spontaneous breathing at low airway pressure and to achieve a complete closure of the air leaks.


1. Dreyfuss D, Ricard JD, Saumon G: Ventilator-induced injury, in Tobin MJ (ed), Principles and Practice of Mechanical Ventilation, 2nd ed. New York, McGraw-Hill, 2006, pp. 903–930.
2. Pierson DJ: Management of bronchopleural fistula in the adult respiratory distress syndrome. New Horiz 1: 512–521, 1993.
3. Lois M, Noppen M: Bronchopleural fistulas: An overview of the problem with special focus on endoscopic management. Chest 128: 3955–3965, 2005.
4. Pesenti A, Rossi GP, Pelosi P, et al: Percutaneous extracorporeal CO2 removal in a patient with bullous emphysema with recurrent bilateral pneumotoraces and respiratory failure. Anesthesiology 72: 571–573,1990.
5. Pesenti A, Bombino M, Gattinoni L: Extracorporeal carbon dioxide removal, in Marini JJ, Slutsky AS (eds), Physiological Basis of Ventilatory Support, 2nd ed. New York, Marcel Dekker, 2006, pp. 501–512.
6. Marcolin R, Mascheroni D, Pesenti A, et al: Ventilatory impact of partial extracorporeal CO2 removal (PECO2R) in ARF patients. ASAIO Trans 32: 508–510, 1986.
7. Chan ED, Morales DV, Welsh CH, et al: Calcium deposition with or without bone formation in the lung. Am J Respir Crit Care Med 165: 1654–1669, 2002.
Copyright © 2011 by the American Society for Artificial Internal Organs