Pneumomediastinum is a Poor Prognostic Factor in COVID-19 Patients – Case Series and Review

Abstract

During the second wave of COVID-19, a number of patients requiring invasive or noninvasive pressure support ventilation rose significantly. The cases of pneumomediastinum (PM), pneumothorax (PTX), and subcutaneous emphysema (SCE) were also reported, largely attributed to barotrauma with observation of some spontaneous cases. Here, we present seven moderate-to-severe COVID-19 cases with variable severity developing PM with or without PTX and SCE during the course of hospitalization. Five received noninvasive ventilation, of whom two required invasive mechanical ventilation. The remaining two patients received oxygen therapy alone. Two patients required insertion of intercostal chest drainage tubes for pneumothoraces. Appropriate medical management was given to all patients. Three patients improved and were discharged from the hospital, while four expired. The case series highlights need of early diagnosis, constant monitoring, and judicious use of pressure support ventilation to prevent air leak complications in COVID-19 patients in order to improve their prognostic outcomes.

INTRODUCTION

Pneumomediastinum (PM) is a life-threatening condition, leading to accumulation of air within the mediastinum, usually due to barotrauma, injury to tracheobronchial tree, lung, pleura, and esophagus. During the second wave of the COVID-19 pandemic in India, a large number of patients required pressure support ventilation through either invasive or noninvasive support. Cases of PM, pneumothorax (PTX), and subcutaneous emphysema (SCE), occurring either iatrogenically or spontaneously, have been reported across the world during the COVID pandemic. As per the POETIC survey, a structured survey of 58,484 COVID pneumonia patients, recently done in the United Kingdom, the incidence of PM in COVID-19 was 0.64% with 51.7% of overall 120-day mortality in such patients.[¹] Acute respiratory distress syndrome (ARDS) patients have been reported to develop PTX and other air leaks in 6.9% and 10.6% of cases, respectively.[²] Patients with severe acute respiratory syndrome (SARS) during 2003 also demonstrated an increased rate of barotrauma resulting in PM and PTX.[³⁴]

COVID-19 patients who already have reduced physiological reserve may be prone to the pressure effects of mediastinal emphysema on the great vessels and thus may mimic the presentation of cardiac tamponade in severe cases.[⁵] Indian data on the subject are scanty. We report a case series of seven COVID-19 patients who developed PM which is the largest Indian series on the subject to the best of our knowledge. The series is likely to add further information to the available literature in terms of patients' characteristics, nature of severity, nature of management intervention in the form of ventilator/noninvasive ventilation (NIV)/oxygen requirement, and the prognosis/mortality, which would warrant a judicious use of the application of positive end-expiratory pressure (PEEP)/pressure support ventilation/NIV/high-flow nasal cannula (HFNC) and would require to keep the air leak complication in mind during the management planning.

CASE SERIES

We present a case series of seven patients with proven COVID-19 pneumonia, who developed PM with SCE with or without PTX during the course of hospitalization. At the time of diagnosis of PM, two out of seven patients had a history of intubation and mechanical ventilation, three were on NIV support, and two were only on high-flow oxygen therapy (thereby having a spontaneous PM).

We recorded baseline patient clinical characteristics including comorbidities, use of pressure support, chest X-ray (CXR)/computed tomography (CT) chest, and management in Table 1.

Table 1: Clinical characteristics of patients

Among seven patients, two were females and five were males with a median age of 45 years. In two patients (S. Nos. 1 and 2) having ARDS, who were invasively ventilated after NIV failure, mode of the mechanical ventilation was pressure-regulated volume control with an initial PEEP range from 5 to 10 cm of H₂O with peak airway pressure of 25–40 cm of H₂O. Both patients expired (one patient within 2 days and another after 10 days of diagnosis). Of three patients (S. Nos. 3, 4, and 7) on the NIV and oxygen support, one could be discharged after 20 days of hospital stay on room air. The other remained on high-flow oxygen support and died after 12 days of diagnosis presumably due to refractory hypoxemia and aspiration pneumonitis despite the standard medical management of COVID-19. The third patient on NIV support was discharged after 45 days of hospital management on domiciliary oxygen therapy with an oxygen flow rate of 2 L/min (FiO₂ of 28%) through nasal prong. Among the two patients with spontaneous PM receiving only oxygen therapy (S. Nos. 5 and 6), one continued to have high oxygen requirement which was given through high-flow nasal cannula for a period of 15 days before deterioration culminating in endotracheal intubation and development of PTX that was managed with intercostal chest drainage (ICD) tube unsuccessfully. The other patient receiving oxygen therapy was discharged on room air after appropriate management. The days on which the events and the patient's outcomes occurred are shown in the swimmer plot [Figure 1]. The consent for the use of data and imaging was obtained from patients or close relatives.

Figure 1: Swimmer plot showing events in order

DISCUSSION

PM, PTX, and SCE are known complications of ARDS from viral infections including COVID-19 disease. The POETIC survey observed a 0.64% incidence of PM in COVID-19.[¹] The spontaneous as well as iatrogenic barotrauma-induced PM, PTX, and SCE have been increasingly described and also seen during the second wave (April–May 2021) of the COVID-19 pandemic. Other air leak syndromes such as pneumoperitoneum and pneumopericardium have also been reported in COVID patients.[⁶] Incidence of air leak was found to be 13.6% in COVID-19-associated ARDS which was significantly higher than 1.9% in the non-COVID ARDS.[⁷⁸]

The etiology of PM may be divided into three groups as shown in Table 2.[⁵] The first is subpleural alveolar rupture leading to free air leak which tracks proximally after dissecting peri-bronchovascular sheath, which is known as the Macklin effect.[⁹] The second is barotrauma which makes the patients more vulnerable to the rupture of alveoli due to increased transalveolar pressure which crosses local stress and strain limit for integrity between epitheliums and interstitial tissue.[¹⁰] Both of them may have an etiological linkage for the causation of PM in COVID-19. Because airway pressure was normal in our ventilated patients, the Macklin phenomenon appears to be more responsible in the causation of air leaks in them rather than the barotrauma.

Table 2: Etiology of pneumomediastinum

Patients of COVID pneumonitis, who are not ventilated, developed spontaneous alveolar rupture because of increased transalveolar pressure leading to air leak and tracking of air into the peri-bronchovascular sheath till the proximal mediastinum. However, few studies did not approve the pressure effects in patients.[¹⁰] Lemmers et al. in their study found that PM developed at the time of low airway pressure and attributed such finding because of lung frailty which is caused by ongoing inflammatory process of COVID-19 as opposed to barotrauma.[⁸] Further, Kangas-Dick et al.[¹¹] have reported the occurrence of PM with the use of a median PEEP of only 12 in their cohort of intubated patients with PM. Therefore, a high pressure may not be the only factor contributing to the development of PM.

Other causes as mentioned in Table 2 are also important in the occurrence of air leaks. Wali et al.[¹²] showed a multifactorial cause for the increase in susceptibility to barotrauma related to inflammation although there is no robust evidence. Baek et al.[¹³] found a correlation between steroid use and PTX in ARDS which may be due to the confounder such as long duration of illness. Villar et al. in their study in 277 ARDS cases found barotrauma in 10% of cases in the dexamethasone group versus 7% in the placebo group.[¹⁴] Larger studies are needed to establish the role of steroids if any in the development of PM. Procedural airway injury or barotrauma does not completely explain the PM in many case series and case reports because many of them developed PM in the absence of intubation or use of positive pressure ventilation. Ye et al. postulated that the cytokine storm of severe SARS-CoV-2 infection might result in the release of air contained within ruptured alveoli tracking into mediastinum consistent with the previously described Macklin phenomenon.[¹⁵] This theory is further strengthened by the finding of a recent autopsy series by Fox et al.[¹⁶]

Figure 2: (a) Pneumomediastinum and subcutaneous emphysema along with severe ARDS, (b) Rim of air alongside heart border suggesting pneumomediastinum, (c) Mild pneumomediastinum with pulmonary fibrosis, (d) Pneumomediastinum and subcutaneous emphysema in RA-ILD and COVID patient, (e) Patient developed spontaneous pneumomediastinum and pneumothorax, (f) Milder disease with small pneumomediastinum, (g) Patient developed PM, PTX, and SCE after initial recovery. ARDS: Acute respiratory distress syndrome, RA-ILD: Rheumatoid arthritis-interstitial lung disease, PM: Pneumomediastinum, PTX: Pneumothorax, SCE: Subcutaneous emphysema

Clinical monitoring and radiology is cornerstone of diagnosis in COVID-19 patients who develop PM. Most of these patients worsened clinically or developed new-onset respiratory distress. Continued clinical monitoring, high index of suspicion, and serial radiology (CXR and CT chest if required) are necessary for early diagnosis.[⁶] Although clinical signs and X-rays showed multiple cases of PM in COVID-19 cases, CT scan remains the definitive diagnostic tool. It can reveal SCE, pneumopericardium, and posterior tracheobronchial injuries alongside the bilateral infiltrate typical of COVID-19.[¹⁷] CXRs will demonstrate SCE as well as the rim of air which is often prominent on the left border of the heart which suggests a pneumopericardium. A similar outline along the left hemidiaphragm and descending aorta is called the Naclerio V sign.[⁵] A lateral film along with the CXR PA view may be necessary half of the time to identify PM correctly. Air anterior to the mediastinum and 'Ring around the artery sign' may be indicators of air in the mediastinum in the lateral film.[¹⁸]

Management in most cases remains largely conservative and so we also followed a conservative approach. Most of the recent case reports and case series have also followed a conservative management approach. Other options include needle aspiration, chest drain insertion, fenestrated catheter insertion, and vacuum-assisted closure therapy. Due to the extension of air distally, the most favored location of decompression is cranially adjacent to the thoracic inlet.[¹²¹⁹] Management of various air leak syndromes in such cases usually depends on severity as well as type of air leak. In a study done by Singh et al., about 67% of patients required ICD tube. There are no established guidelines for managing such patients. In addition, pleurodesis and surgical interventions are important options.[²⁰] However, Brown et al. in their recent trial showed conservative management is noninferior to active intervention.[²¹]

Bilateral pleural drains with or without subcutaneous drain are sometimes required to prevent life-threatening complications in worsening cases of PM. About 45.6% of patients of COVID-19 with PM required mechanical ventilation[¹] and these patients receiving positive pressure ventilation (IMV or NIV) may require ICD placement at the earliest.

PM, though, a rare clinical finding, may pose a serious challenge in the management of already compromised lungs of COVID-19 patients. Although the conservative approach is all that is most emphasized, close clinical monitoring with serial CXRs is essential to look for PTX and cardiorespiratory compromise so that in case of worsening clinical scenario, active intervention can be done on time. Duration of hospitalization increases significantly in COVID-19 patients with air leak syndromes. Miró et al. in their study reported a 4.2-fold increase in hospital stay of COVID-19 patients with PTX as compared to COVID-19 alone.[²²] Since all our COVID cases in the present series had PM, we cannot compare the hospital stay with those not having the air leak complication.

PM has traditionally been the initial manifestation of barotrauma in intubated and mechanically ventilated patients and a precursor of PTX. PM is a bad prognostic factor in intubated patients.[²³] In ARDS, barotrauma overall increases mortality.[²⁴] The recently conducted POETIC survey showed 51.7% of overall 120-day mortality in such patients. Mechanical ventilation, diabetes mellitus, and old age were important risk factors for mortality. One important finding was that switching patients from positive pressure ventilation to oxygen or high-flow nasal cannula therapy did not make difference in terms of mortality.[¹] Kangas-Dick et al.[¹¹] reported 66.7% of mortality in their study of 36 patients, while Wali et al.[¹²] reported 40% of mortality in their case series of 5 patients. In our case series also, the mortality of 57.14% was high owing to various factors such as the nonresponse of ARDS patients to HFNC therapy, existence of comorbidities, and hospital-acquired infection with MDR pathogens. Thus, the presence of PM is a poor prognostic factor in the management of COVID-19 patients.

CONCLUSION

To conclude, the presence of PM either spontaneous or due to barotrauma does complicate the course of illness of COVID patients in terms of management, mortality, and hospital stay. There is a need to keep air leak complications in mind during the management of critically ill COVID patients. Accordingly, judicious use of PEEP/positive pressure ventilation/NIV is warranted. More studies are required to establish the exact cause of air leaks in COVID patients.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.