Background
A case clusters of pneumonia caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first reported in Wuhan city of China in December 2019, then spread worldwide [
1]. Clinical presentations of coronavirus related disease (COVID-19) vary from asymptomatic to mild influenza-like symptoms to life-threatening pneumonia, whilst multiorgan failure can also occur [
1,
2]. COVID-19 acute lung injury results from exaggerated inflammation started by viral replication and then sustained by a cytokine storm that is associated with disease severity [
3,
4]. Although the physiopathological features of respiratory failure caused by SARS-CoV-2 infection may be heterogeneous, in most severe cases of COVID-19, acute respiratory distress syndrome (ARDS) occurs [
5]. In these patients, high elastance and severe lung inhomogeneity make the parenchyma prone to ventilator-induced lung injury (VILI) when mechanical ventilation (MV) is applied [
6,
7]. Positive pressure ventilation, especially at the interface between structures with different extensibilities, may lead to harmful transpulmonary pressure causing the rupture of extracellular polymers, resulting in “barotrauma”, of which the most common form is pneumothorax [
9,
10]. Despite the onset of pneumothorax requiring chest drainage having been reported in 5.9% of COVID-19 patients undergoing MV [
11], only a few cases of persistent air leak (PAL) have been described to date. PAL is defined as an air leak that persists for more than 5–7 days [
12]. A bronchopleural fistula (BPF) occurs when PALs are supported by a communication between the bronchus and the pleural space, whereas an alveolar-pleural fistula (APF) follows a distal origin of PALs [
12,
13]. PALs are known to be the one of common complications after lung surgery [
14]; however, APF is not frequently associated with an excessive stress/strain applied to the lung parenchyma by MV. Nonetheless, this clinical condition has serious consequences for patient management and outcome resulting in prolonged chest tube maintenance and hospital stay, and in a higher rate of infectious complications due to loss of sterility in the pleural space [
12,
13]. Retrospective studies in patients with PALs have shown mortality rates ranging from 16 to 72% [
11]. Treatment options include conservative management, surgical repair, and minimally invasive interventional pulmonology or bronchoscopic management [
13,
15].
We present two cases of severe COVID-19 pneumonia complicated by bacterial pneumonia and APF secondary to VILI and treated through EBV positioning.
Discussion and conclusion
COVID-19 exhibits a broad spectrum of clinical conditions, including pneumonia, ARDS, pulmonary embolism, and hearth failure, leading to the development of respiratory failure often requiring ICU management and MV (in about 20% of cases) [
16]. The pathogenesis of respiratory failure is complex and covers different clinical scenarios ranging from patients with relatively spared lung compliance to patients with ARDS mechanical features. [
5,
17]. Most patients admitted to the ICU with severe COVID-19 pneumonia fulfill ARDS criteria and require MV, which in turn may lead to a VILI [
18]. The high stress level generated during MV associated with lung tissue inflammation can cause bronchial or alveolar ruptures followed by pneumothorax or pneumomediastinum. Gammon et al. have shown that ARDS per se is an independent risk factor, although the less aggressive use of pressure, in the lung protective strategy era, was associated with a markedly lower incidence of pneumothorax [
8]. Both cases underwent MV with protective strategy (expiratory tidal volume [Vte] < 6 ml/kg of predicted body weight, plateau pressure [P
plat] < 28 cmH
2O, PEEP titration according to FiO
2 schedule) and were not subjected to recruitment maneuvers. In this scenario, it may be speculated that regional factors associated to lung inhomogeneity due to advanced stages of COVID-19 related ARDS and inflammatory related tissue frailty, could have played a major role in exacerbating barotrauma, resulting in APF. In the previous SARS outbreak, barotrauma was a common complication during MV. Indeed, studies in mechanically ventilated SARS patients showed a high incidence of pneumothorax (20–30%) and pneumomediastinum (12%) [
19‐
22]. Patients with a lower PaO
2/FiO
2 ratio, higher PaCO
2, and higher respiratory rate during ICU stay were at a greater risk of developing pneumothorax [
23]. The incidence of SARS-CoV-2-related pneumothorax appears to be significantly lower (5.9%) than reported in patients with SARS-CoV-1, but it can lead to a mortality of 30% [
24]. Moreover, studies performed in non-SARS-CoV-2 patients with ARDS showed that overinflation and distortion of lung structures due to MV could result in emphysema-like lesions, lung cysts, and bronchiectasis [
25]. In the late stages of ARDS, these lesions prevail in nondependent and caudal lung regions as confirmed by chest CT scan [
26]. Furthermore, experimental studies in animals have demonstrated that MV set at high tidal volume or high PEEP and recruitment maneuvers results in air space enlargement after 2 days [
27,
28].
Radiologic studies in COVID-19 with ARDS have shown that patients can develop cystic lung parenchyma changes and bullae formation during viral infection [
29]. These parenchymal changes may predispose patients undergoing MV to develop pneumothorax and PALs, especially in the advanced stage when the lung has possibly undergone fibrotic changes. Since the fibrotic lung is particularly subject to static strain, the hyperinflation achieved with the addition of PEEP can facilitate enlargement of the air spaces in non-dependent areas of the lung [
30]. Moreover, both cases showed the isolation of
Klebsiella pneumoniae and
Enterococcus faecalis from bronchoalveolar lavage (Case 1) and pleural fluid (Case 2). A bacterial super-infection may have enhanced the inflammatory mechanism of lung injury already triggered by SARS-CoV-2 infection, thus increasing the susceptibility to PALs development. Indeed, necrotizing infections of the lung caused by gram negative bacteria (namely
Klebsiella pneumoniae) have been widely reported [
31]. Given the lack of evident signs of fistulisation at bronchoscopic examination we concluded for APF diagnosis for both cases.
Traditional approaches to PALs range from watchful waiting with prolonged chest tube placement to more invasive measures including surgery. There is a recent report of two COVID-19 patients with PALs successfully treated with thoracoscopy, blebs resection, and pleurectomy [
32]. Different less invasive options have been tested, such as autologous blood pleurodesis, Heimlich valve positioning, and albumin-glutaraldehyde tissue adhesives [
13]. Recently a BPF related PAL was successfully treated with EBV placement [
33]. To our knowledge, these are the first two cases documenting the effective use of EBVs to treat PALs due o APF in patients infected with novel SARS-CoV-2 who required invasive MV.
EBV was first administered for lung volume reduction in patients with severe emphysema [
34], and it is currently the most actively performed procedure for PALs, when surgical options are not viable [
13,
15]. EBV has been approved by the Food and Drug Administration to treat PALs under humanitarian use regulations, reducing airflow across the fistula and allowing healing and resolution [
35]. In a multicenter study with 40 patients, Travaline et al. showed that Zephyr EBV is an effective minimally invasive device for completely treating PALs in around half of the treated patients, and complications such as valve malposition and/or expectoration, and pneumonia are quite rare [
36].
Localization of PALs is challenging, and the balloon occlusion method is currently the most used approach [
13]. Our two patients experienced continuous bubbling during inspiration/expiration, i.e. grade 4 according to the Cerfolio classification of air leaks [
13]. After the EBV procedure, both cases showed complete resolution of PALs from APF without any need for surgery.
Although, our reported experience in COVID-19 patients with iatrogenic PALs refractory to conservative strategy has demonstrated that the EBV procedure is feasible, safe, and effective, some limitations should be highlighted. First in severely hypoxic patients the bronchoscopic interventional procedure may worsen respiratory failure. Secondly, in patients with bacterial super-infection valve positioning may facilitate post-obstructive pneumonia. Third, considering the risk of pneumothorax relapse, the timing of EBV removal should be carefully planned as data about the optimal management of these patients once air leaks resolution is achieved are lacking. However, in the most severe cases with an advanced stage of ARDS, at higher risk for iatrogenic PALs, the EBV procedure is likely to offer a reasonable minimally invasive therapeutic option.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.