Pneumothorax in Critically Ill COVID-19 Patients in the Indian Subcontinent

COVID-19 pneumonia resulting from severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection [1] is a rapidly evolving condition that has become a challenge for the critical care physician owing to its highly infectious nature. Recent evidence suggests pneumothorax (PTX) and pneumomediastinum (PNM) commonly complicate the clinical course in COVID-19 patients in patients with moderate to severe COVID-19 disease.

Development of PTX/PNM is a known complication and is widely reported in ARDS patients receiving both invasive and non-invasive modes of ventilation [2]. The surge of the COVID-19 pandemic also witnessed the development of PTX/PNM in patients with SARS-CoV-2 pneumonia not only under invasive [3,4] or non-invasive [5] modes of ventilation but also in patients receiving conventional oxygen therapy [6,7,8,9,10]. There have also been reports of COVID patients presenting in the emergency department with PTX/PNM [5]. The occurrence of PTX/PNM even in spontaneously breathing COVID patients indicate that there must have been additional factors contributing to their development of PTX/PNM.

Till now four retrospective studies have evaluated patients’ characteristics contributing to development of PTX/PNM in COVID-19 patients [4,5,11–12]. Both Wang et al. [4] and Enhakem et al. [11] concluded that pneumothorax is a frequent and fatal complication of critically ill patients with COVID-19. Wang et al. observed that the reduction of NMB, recruitment manoeuvres, severe cough, and changes in lung structure and function are factors responsible for development of PTX/PNM in COVID-19 patients [5]. Enakem et al. concluded that marked inflammatory response, fibrosis, and need for positive pressure ventilation in COVID-19 pneumonia are likely contributory to the development of pneumothorax in these patients. Martinelli et al. concluded that it was not an independent marker of poor prognosis, and that age >70 years and acidosis to be associated factors responsible for poor prognosis [6]. Two very recent studies concluded that baseline radiological findings of ground-glass opacities and consolidation and mechanical ventilation appeared to predict high risk of developing PTX/PNM [12,13]. The data from all these studies are from the first wave of COVID-19.

However, in our literature search we could not retrieve any study exploring the risk factors predicting the development of PTX /PNM in COVID-19 patients, nor the incidence of PTX/PNM in the Indian subcontinent. Different strains of the COVID-19 virus are prevalent; therefore, it is desirable to have a study exploring their fatality and risk factors in different geographic regions.

There is very scant literature throwing light on mortality in patients who developed PT/PNM during COVID-19 and whether it is an independent marker of poor prognosis in COVID-19 patients, particularly those from the Indian subcontinent.

Therefore, with this background, we designed this retrospective cohort study to evaluate the incidence of mortality among COVID-19 patients with pneumothorax and/or pneumomediastinum and to determine the risk factors contributing to development of pneumothorax/pneumomediastinum in the Indian subcontinent during the second wave of COVID-19.

A total of 934 patients were admitted in various ICUs during the designated period. Out of this, 788 patients reported positive reverse transcriptase polymerase chain reaction for COVID-19, and in the rest of the patients, a clinicoradiological diagnosis was made. Twenty-two patients left against medical advice and no data is available for them. Finally, 912 patients were included in the study. On screening the files of all these patients, it was observed that 18 patients developed PTX and/or PNM, an incidence of 1.9% (18/912). Thirty-four patients were included in the non-exposure group of COVID-19 patients, i.e. who did not develop PTX/PNM.

The clinical characteristics of COVID-19 patients with PTX/PNM.

The baseline characteristics of patients who developed PTX/PNM are shown in Tables 1, 2, and 3. Both groups were comparable with respect to age and gender distribution well as time of onset of the disease; therefore, both the groups were matched. Various comorbidities such as COPD, asthma, hypertension, and diabetes were comparable between the two groups.

Amongst the 18 COVID-19 patients who developed PTX/PNM, only one patient had more than one episode of pneumothorax. This patient had a contralateral pneumothorax as compared to the initial pneumothorax. In terms of the initial pneumothorax of each patient, the majority were right-sided (12/19, 63%), followed by left-sided (5/19, 26%) and bilateral (1/19,5%). pneumomediastinum, in addition to PTX, was observed in 10/19 (53%) of cases.

Out of 18, 13 patients were on non-invasive ventilation (CPAP/BiPAP), four patients were on IPPV, and one was not on oxygen therapy (Table 2). Almost half, 9/19 (47%), of the pneumothoraces received tube thoracostomy for the management of their pneumothorax, whereas the remainder, 10/19 (53%), did not need any intervention.

COVID-19 patients who developed pneumothorax and/or pneumomediastinum demonstrated mortality of 72.2%; among the patients who did not develop PTX/PNM mortality was observed to be 26.5% (Table 5). COVID patients who developed PTX/PNM had 2.7 times higher risk of mortality as compared to the patients who did not develop PTX/PNM. P-value was observed to be 0.001 (<0.05) which signifies that the difference in mortality between the two groups was statistically significant. The 95% C.I. was observed to be 1.45–5.12 (Table 5). Median survival time in non-survivors who developed PTX/PNM was observed to be 5.61–8.38 days (Table 6).

With respect to age and sex distribution, no statistically significant difference was observed between the two groups (Table 1). Incidence of systemic comorbidities, diabetes mellitus, hypertension, previous lung disease such as bronchial asthma, and chronic obstructive lung disease was observed to be comparable between the two groups (Table 1). CT severity scores were found to be higher in the group which developed PTX/PNM (19.9 ± 3.5) as compared to the non-exposure group (19.0 ± 4.3); however, the difference was observed to be statistically non-significant (p>0.05) (Table 1). With respect to modes of ventilation, 17 (94.4%) patients received positive pressure ventilation in the exposure group, out of which 13 (72.2%) of patients were administered non-invasive mechanical ventilation in the form of CPAP and 4 patients (22.2%) received invasive mechanical ventilation. This is in contrast to the non-exposure group, where 25 patients (63.5 %) received positive pressure ventilation, with 24 patients administered noninvasive mechanical ventilation whereas only one patient (2.9%) was administered invasive mechanical ventilation.

Interestingly, the proportion of patients who received oxygen therapy via simple face masks was found to be comparable between the two groups (5.6% vs. 5.9%) (Table 2).

Driving pressures in patients of the exposure group who were administered invasive mechanical ventilation were found to be higher (20.9 ± ) as compared to the non-exposure group (18.3 ± ); the difference was observed to be statistically significant (p=0.02) (Table 2). This denotes poorer respiratory mechanics in COVID-19 patients who developed PTX/PNM.

With respect to lab parameters, in both the groups haemoglobin concentration was observed to decline with time. The decline was statistically significant in both the groups at day 7, day 15, and day 28, as compared to the baseline (Table 3).

No statistically significant difference was observed with respect to total leukocyte count and platelet count levels. However, the fall in serum albumin levels in the PTX group was found to be closely significant (p=0.06) at 15–20 days of illness (Table 3). Baseline values of S. interleukin-6 levels in PTX vs. non-PTX group (41.1 vs. 12.6), S. ferritin values (1425.0 vs. 916.5), and C-reactive protein levels (79.0 vs. 40.0) in PTX group were found to be higher as compared to the non-PTX group (Table 4). The difference between the CRP values of the two groups was observed to be closely significant (p=0.06).

Incidence of pneumothorax/pneumomediastinum was observed to be 1.9% among the critically ill COVID-19 patients admitted in the ICUs of our tertiary care hospital (18/934). COVID-19 patients who developed PTX/PNM had 2.7 times higher risk of mortality as compared to the patients who did not develop PTX/PNM (p=0.001), reflecting significantly higher mortality in PTX/PNM group than in the non-PTX/PNM group. A mortality rate of 72.2% was observed in COVID-19 patients who developed pneumothorax and /or pneumomediastinum. Median survival time in COVID-19 patients who developed PTX/PNM was observed to be between 5.61–8.38 days.

In a retrospective study conducted by Wang et al. evaluating the outcomes in COVID-19 patients who developed PTX/PNM, mortality of 80% was observed [3]. The sample size taken in this study was smaller [5]. The results of our study are in concordance with Wang et al., as we observed a mortality of 72.2%. McGuiness et al. identified 89 patients with barotrauma and reported an overall incidence of 15% in mechanically ventilated COVID-19 patients and concluded that barotrauma was associated with increased risk of death [14]. Perhaps the high mortality observed in our study was due to the fact that 17 out of 18 patients were administered positive pressure ventilation, 13 patients were on non-invasive ventilation, and 4 were on invasive mechanical ventilation. Associated barotrauma as a result of PPV in COVID-19 patients might have contributed to the high mortality of 72% in our study cohort.

Miro et al. in a recent study compared the outcome of non-intubated patients with COVID-19 infection who presented to the Emergency Department with a pneumothorax to those without a pneumothorax and observed that patients who developed pneumothorax had four times higher risk of death as compared to patients who did not [15]. The findings of our study are concordant with the results of Miro et al., as we observed 2.7 times higher risk of death in critically ill COVID-19 patients who developed PTX/PNM as compared to patients who did not develop PTX/PNM. Therefore, we believe that PTX/PNM is an independent marker of poor prognosis in COVID-19 patients and is associated with significantly high mortality.

In a systematic review conducted by Chong et al., the authors concluded that age, pre-existing lung diseases, and active smoking status are not shown to be risk factors. There is perhaps an additional mechanism specific to COVID-19 disease that precipitated the development of PTX and PNM in these patients. In this context, various authors have observed that this disease is characterised radiographically by ground glass opacities, evolving into consolidative changes and in late stages of the disease, fibrotic changes, bullous formation as well as diffuse cyst formation [16,17,18,19]. Similar changes in the form of diffuse alveolar damage have also been observed in patients infected with SARS-CoV2 who developed spontaneous pneumothorax; none of the patients received positive pressure ventilation [20]. Of patients who developed PTX/PNM in our study, 94.4% received positive pressure ventilation (PPV) in the exposure group, as compared to COVID-19 patients who did not develop PTX/PNM, where a lower proportion of patients (63.5%) received positive pressure ventilation. We also observed significantly high inspiratory airway pressures in the patient population that developed PTX/PNM. We believe that instituting positive pressure ventilation with less than optimum ventilator management in the COVID-19 patients who have developed the aforementioned pathological changes might have precipitated the development of PTX and PNM.

It is now known that an excessive host immune response can be triggered in COVID-19, leading to extensive alveolar damage. Interleukin-6 (IL-6) is considered a main trigger of such an immune response. Higher values of CRP levels as well as IL-6 levels at different time intervals in the PTX group as compared to non-PTX denote higher magnitudes of systemic inflammation and more severe disease involvement leading to more extensive alveolar damage; however, the difference was statistically non-significant (p>0.05). High levels of serum CRP and IL-6 are strongly associated with venous thromboembolism, acute kidney injury, critical illness, and mortality in COVID-19 disease [21]. This also additionally explains the poorer prognosis and significantly higher mortality (2.7 times) in COVID-19 patients who developed PTX/PNM.

Initial studies from China reported an incidence of pneumothorax of approximately 1% in all patients infected with COVID-19 [22,23]. A large multicentre study was conducted across 61 emergency departments (ED) in Spain, involving 71,904 COVID-19 patients, which observed the pneumothorax incidence to be 0.56% [15]. In the systematic review by Chong et al., the incidence was observed to 0.3% among hospitalised COVID patients in different centres in the United States. However, in our literature search we could not retrieve any study assessing the incidence of PTX/PNM in critically ill COVID-19 patients in the Indian subcontinent. We observed the incidence of PTX/PNM in COVID-19 patients to be higher in the Indian population (1.9%). The higher incidence can also be attributed to the high virulence of the Delta strain of SARS-CoV-2 virus, which was the prevalent strain in the second wave of COVID-19 pandemic in our nation.

There were several limitations of our study. As it was a retrospective analysis, we could not record the ventilatory parameters in real time. Second, it is a single-centre study. Third, the available therapy and outcomes of the COVID-19 disease have been changing drastically over time [13]. So, a possibility that our results, including mortality as well as risk factors, could not be extrapolated to future critically ill COVID-19 infected patients, cannot be ruled out.

Nonetheless, our study is the first valid attempt at the evaluation of incidence, risk factors, as well as mortality in the subset of Indian patients who developed PTX/PNM during the COVID-19 disease.

To summarise, 1.9% of critically ill COVID-19 patients were admitted in the intensive care units, in the subset of the Indian population. COVID-19 patients who developed PTX/PNM had 2.7 times higher risk of mortality as compared to the patients who did not develop PTX/PNM. A mortality rate of 72.2% was observed in COVID-19 patients who developed pneumothorax and/or pneumomediastinum. A higher proportion of patients requiring PPV signifies more severe disease involvement; patients who developed PTX/PNM had more severe disease. Taking into account the associated pathological changes in lung parenchyma, stricter implementation of lung protective ventilation strategy in COVID-19 patients will reduce the incidence of PTX/PNM.