Prevalence and risk factors of barotrauma in Covid-19 patients admitted to an intensive care unit in Kuwait; a retrospective cohort study.

BACKGROUND: The development of barotrauma has been suggested to complicate the management of mechanically ventilated COVID-19 patients admitted to the intensive care unit (ICU). This study aims to identify potential risk factors associated with the development of barotrauma related complications in COVID-19 patients receiving mechanical ventilation.
METHODS: A retrospective cohort study was carried out in a single COVID-19 designated center in Kuwait. Three hundred and forty-three confirmed COVID-19 patients transferred and/or admitted to our institution between February 26, 2020 and June 20, 2020 were included in the study. All patients were admitted into the ICU with the majority being mechanically ventilated (81.3%).
RESULTS: Fifty-four (15.4%) patients developed barotrauma, of which 49 (90.7%) presented with pneumothorax, and 14.8% and 3.7% due to pneumomediastinum and pneumopericardium respectively. Of those that developed barotrauma, 52 (96.3%) patients were in acute respiratory distress syndrome (ARDS). Biochemically, the white blood cells (p = 0.001), neutrophil percentage (p = 0.012), lymphocyte percentage (p = 0.014), neutrophil: lymphocyte ratio (NLR) (p=<0.001) and lactate dehydrogenase (LDH) (p = 0.002) were found to be significantly different in patients that developed barotrauma. Intubation due to low level of consciousness (p = 0.007), a high admission COVID-GRAM score (p = 0.042), and a positive-end expiratory pressure (PEEP) higher than the control group (p = 0.016) were identified as potential risk factors for the development of barotrauma.
CONCLUSION: Patients infected with COVID-19 have a significant risk of developing barotrauma when receiving invasive mechanical ventilation. This poses a substantial impact on the hospital course of the patients and clinical outcome, correlating to a higher mortality rate in this cohort of patients.

Introduction

The novel coronavirus (COVID-19) has become a global pandemic [1]. This highly transmissible and infectious disease is found to affect multiple systems, particularly the respiratory tract. The prevalence of pneumothorax among COVID-19 patients in the intensive care unit (ICU) has been reported to be 2% [1,2]. More recently, studies have found that barotrauma-related complications due to invasive mechanical ventilation is increasingly reported, as incidence in COVID-19 patients was reported to be as high as 15% [3].

The association of acute respiratory distress syndrome (ARDS) and the development of secondary pneumothorax in mechanically ventilated patients is well documented as an independent risk factor of mortality [[4], [5], [6]]. The majority of patients that contract COVID-19 experience symptoms of a mild upper respiratory tract infection [[7], [8], [9]], however a small proportion of patients are found to develop severe pneumonia and sepsis with the potential development of ARDS and multi-system organ failure [[9], [10], [11], [12]]. The development of ARDS and its associated complications, which include septic shock, thrombotic complications, acute kidney injury (AKI), derangement of liver enzymes, cardiac injury and barotrauma are associated with poor clinical outcomes in COVID-19 patients [[9], [10], [11]]. There are a limited number of studies that have focused on the epidemiology and the potential risk factors associated with developing barotrauma in COVID-19 patients [[11], [12], [13]].

In our single-center study we aimed to identify clinical features and risk factors associated with the potential development of barotrauma in mechanically ventilated COVID-19 patients. The study also aims to investigate the impact in terms of clinical course and prognosis of COVID-19 when barotrauma related complications occur.

Methods

Study design and data collection

The present retrospective study included the first 343 confirmed COVID-19 patients transferred and/or admitted to our institution between February 26, 2020 and June 20, 2020. Our institution, located in the State of Kuwait, was dedicated solely to COVID-19 patients, predominantly for patients who were deemed critical and may require further support in the form of extracorporeal membrane oxygenation (ECMO) in the ICU.

Inclusion criteria were patients of all ages diagnosed with COVID-19 using PCR testing, in accordance with the World Health Organization (WHO) interim guidance. COVID-GRAM predictive risk score developed and internally validated in China was used to predict the risk of critical illness in hospitalized COVID-19 patients [14]. Ten variables at admission were used to predict critical illness. These include age, unconsciousness, hemoptysis, dyspnea, number of comorbidities, cancer history, chest X-ray abnormality, neutrophil to lymphocyte ratio, lactate dehydrogenase and direct bilirubin. The risk of developing critical illness was further categorized into three categories: low, medium and high.

All related patient information and clinical data were retrieved from the hospitals electronic medical record system. These included sociodemographic factors (age, gender, nationality), clinical indicators (unconsciousness on admission i.e. Glasgow coma score of less than 8, hemoptysis and shortness of breath) and biochemical inflammatory markers (white blood count, neutrophil percentage, lymphocyte percentage, neutrophil: lymphocyte ratio, lactate dehydrogenase, CRP, ferritin, direct bilirubin and d-dimer), presence of co-morbidities (diabetes, hypertension, asthma, coronary artery disease) and COVID-GRAM score.

Written informed consent was obtained from the patient for publication of this study. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.

Ventilatory variables were recorded daily. These included maximum and minimum peak end-expiratory pressure (PEEP) adjusted for mechanical ventilation, fraction of inspiratory oxygen (FiO2), and partial pressure of oxygen (PaO2). Peak inspiratory pressure (PINS), pressure support (PS) and PEEP were recorded 24 h prior to the development of barotrauma. In accordance with Berlin Classification, PaO2/FIO2 ratio was used to define acute respiratory distress syndrome (ARDS) based on degree of hypoxia [17]: mild: 200 mm Hg < PaO2/FIO2 ≤ 300 mm Hg, moderate: 100 mm Hg < PaO2/FIO2 ≤ 200 mm Hg, and severe PaO2/FIO2 ≤ 100 mm Hg. In the present study, ARDS was defined among patients with a PaO2/FIO2 less than 300 mm Hg, the presence of bilateral opacities on imaging, a gradual onset (7–10 days), and no existing cardiac/pulmonary pathology prior to admission [15]. Hemodynamic instability, desaturation, and increasing ventilatory parameters were the main reason(s) behind pursuing radiological imaging leading to a diagnosis of barotrauma related complications in 15.4% of our patient population.

At our institution we utilize lung-protective ventilatory strategies, with a tidal volume of 4–6 ml/kg targeting a driving pressure of 15–17 cmH2Oo and a plateau pressure not exceeding 30 cmH2O. The Fio2 was adjusted to maintain the SpO2 between 88 and 92%. In unresponsive cases, prone positioning and inhaled Nitric oxide were utilized.

Additional recorded outcomes included the length of hospital, length of ICU stay, the number of days a patient was placed on a mechanical ventilator, the date patients developed barotrauma associated pneumothorax, and the incidence of death.

All 343 patients included in the study were admitted to the ICU. Patients that were mechanically ventilated or receiving non-invasive airway support were included in the study. Identified iatrogenic pneumothorax cases and patients diagnosed with a pre-existing pneumothorax on admission were excluded.

Outcome

The primary outcome of the study was to investigate risk factors that may have led to the development of a radiologically confirmed pneumothorax/barotrauma, and the clinical prognostic implications of developing a pneumothorax in COVID-19 patients receiving care within the intensive care unit.

Whether patients received tube thoracostomy or conservative management was also documented alongside the number of days the intercostal tube remained in-situ. Invasive procedures in the thorax in the 24 h prior to the development of the pneumothorax were recorded, including central lines, bronchoscopies, removal of intercostal tubes and endotracheal intubations. The intercostal tube was removed if there was documentation of a radiographically resolved pneumothorax, the absence of an air leak, and after the paucity of negative wall suction for at least 24 h prior.

Registration

This study was registered at https://www.researchregistry.com/(unique identifying number: researchregistry6324) and work has been reported in line with the STROCSS criteria [16].

Statistical analysis

Data were analyzed using SPSS (IBM, version 25). Descriptive statistics were used to calculate mean and standard deviations for continuous data and frequency statistics were used to calculate numbers and percentages for categorical variables. Patient characteristics that developed or did not develop pneumothorax were analyses using independent t-test for continuous variables and chi-squared test for categorical variables. Statistical significance was set at p value less than 5%.

Results

Demographic, clinical and biochemical characteristics are presented in Table 1. Of the 343 patients included in the study, 285 (83.1%) were male and the mean age was 55.9 (13.5) years. Fifty-four (15.4%) patients developed barotrauma, of which 49 out of 54 patients presented with pneumothorax, and 14.8% and 3.7% was due to pneumomediastinum and pneumopericardium respectively. Of the 54 patients that developed barotrauma 52 (96.3%) patients were in ARDS (Table 2), and 5 (9.3%) received ECMO. The mean duration of developing barotrauma from admission was 10.9 (10.8) days, and mean duration of intubation prior to barotrauma development was 9.7 (11.1) days (Table 1).

Table 1

Demographic, Clinical and Biochemical characteristics according patients with or without Barotrauma.

	Barotraumaa n = 54	No Barotraumaa n = 289	p valueb
Demographic
Age, years	55.3 (15.0)	56.0 (13.3)	0.76
Gender
Male, n (%)	42 (77.8)	243 (84.1)	0.32
Female, n (%)	12 (22.2)	46 (15.9)
Nationality
Kuwaiti, n (%)	18 (33.3)	92 (31.8)	0.78
Non-Kuwaiti, n (%)	36 (66.7)	197 (68.2)
Anthropometric
Weight, kgs	81.1 (21.0)	85.4 (20.2)	0.18
Height, m	166.0 (7.8)	167.9 (8.2)	0.17
BMI, kg/m2	29.5 (5.3)	30.1 (6.8)	0.62
Symptoms
Shortness of breath, n (%)	26 (48.1)	145 (50.2)	0.88
Hemoptysis, n (%)	54 (100)	289 (100)	–
Unconsciousness on admission	36 (66.7)	113 (46.0)	0.007
Comorbidity
Diabetes, n (%)	20 (37.0)	123 (42.6)	0.55
hypertension, n (%)	24 (55.6)	131 (45.3)	1.00
CHD/IHD, n (%)	4 (7.4)	43 (14.9)	0.20
COPD, n (%)	1 (1.9)	4 (1.4)	0.58
Asthma, n (%)	6 (11.1)	27 (9.3)	0.62
CVD, n (%)	2 (3.7)	12 (4.2)	1.00
Hepatitis, n (%)	2 (3.7)	5 (1.7)	0.30
Cancer, n (%)	4 (7.4)	8 (2.8)	0.10
CKD, n (%)	4 (7.4)	18 (6.2)	0.76
Immunodeficiency, n (%)	1 (1.9)	2 (0.7)	0.40
History of cancer, n (%)	4 (7.4)	8 (2.8)	0.10
Outcome
Gram score	194.1 (159.1)	159.9 (101.3)	0.042
Risk c
Medium, n (%)	39 (11.4)	136 (47.1)	0.006
High, n (%)	14 (4.1)	152 (52.6)
Death, n (%)	38 (70.4)	153 (52.9)	0.025
ETT, n (%)	54 (100)	225 (77.9)	<0.001
ECMO	5 (9.3)	19 (6.6)	0.56
Biochemical
WBC	12.1 (7.7)	9.5 (4.8)	0.001
Neutrophils	83.8 (12.4)	79.0 (12.9)	0.012
Lymphocytes	10.0 (10.5)	13.8 (10.1)	0.014
NLR	17.5 (15.9)	10.0 (9.2)	<0.001
CRP	134.0 (111.4)	142 .4 (119.6)	0.71
Procalcitonin	4.3 (12.3)	22.2 (242.7)	0.63
Direct bilirubin	8.2 (10.3)	10.1 (29.8)	0.63
D-dimer assay	2216.2 (3918.1)	2565.6 (3191.5)	0.58
LDH	1118.9 (3513.6)	408.0 (270.2)	0.002
Ferritin	1413.8 (1966.4)	1082.8 (1490.1)	0.22
Length of Stay	16.4 (8.5)	20.1 (14.0)	0.11
Length of stay (ICU)	14.9 (7.8)	13.9 (12.1)	0.60
Pressure
PEEP minimum	7.7 (2.6)	6.7 (3.8)	0.06
PEEP Maximum	14.5 (2.7)	12.5 (5.7)	0.016

Data are presented as means (standard deviations) for continuous variables, unless indicated. Data presented as number (percentage) for categorical variables.

Statistical Significance set at p < 0.05.

Barotrauma associated with pneumothorax, pneumomediastinum and pneumopericardium.

Independent t-test was carried for continuous variables. Chi-squared test was carried out for categorical variables.

Only for 53/54 patients and 288/289 patients.

Table 2

Pressure and Volume variables specific to patients that development pneumothorax.

Barotrauma	Data
Pneumothorax, n (%)	49 (90.7)
Pneumomediastinum, n (%)	8 (14.8)
Pneumopericardium, n (%)	2 (3.7)
PEEP 24 h prior to development of barotrauma	11.7 (3.5)
Duration of intubation prior to development of barotrauma	9.7 (11.1)
Duration of barotrauma development from admissionb	10.9 (10.8)
PaOb 24 h pre-trauma	92.9 (40.8)
FiOb 24 h pre trauma	73.4 (21.8)
PaOb/FiOb ratio 24 h	142. 0 (80.8)
ARDS, n (%)	52 (96.3)
PINS 24 h priora	28.2 (8.3)
ICD duration	8.2 (9.9)
PSa	16.0 (6.9)
Procedure 24 h prior, n (%)	6 (11.1)

Data presented as mean (standard deviation) for continuous variables, unless indicated for categorical variables.

Data available for 34/54 patients.

Data available for 50/54 patients.

Several inflammatory markers were found to be statistically different between patients that developed barotrauma compared to those that did not develop barotrauma. White blood cell count (WBC) (12.1 vs. 9.5; p = 0.001), neutrophil percentage (83.8 vs. 79.0; p = 0.012), lymphocyte percentage (10 vs 13.8; p = 0.014), neutrophil: lymphocyte ratio (NLR) (17.5 vs. 10; p=<0.001) and lactate dehydrogenase (LDH) (1118.9 vs. 408.0; p = 0.002). No statistically significant differences were found between CRP, procalcitonin, direct bilirubin, d-dimer assay and ferritin.

Upon admission unconsciousness was significantly higher among patients that developed and those that did not develop barotrauma (66.7% vs. 46.05; p = 0.007). All 54 patients with barotrauma were intubated and mechanically ventilated compared to 77.9% those without pneumothorax (p < 0.001), and mortality was significantly greater among those with barotrauma (70.4% vs. 52.9%; p = 0.025). No significant difference was found between barotrauma and non-barotrauma patients in receiving ECMO.

COVID-GRAM score was significantly higher among patients with barotrauma compared to patients that did not develop barotrauma (194.1 vs. 159.9; p = 0.042), PEEP maximum was greater among patients with barotrauma compared to those without (p = 0.016), and the incidence of death was significantly higher among patients that developed barotrauma compared to patients that did not develop barotrauma (70.4% vs. 52.9; p = 0.025).

There were no significant differences in demographic variables, presence of comorbidities, BMI, and length of hospital/ICU stay between patients with or without barotrauma.

Discussion

The present study focused on COVID-19 patients, analyzing the associated risk factors of barotrauma and the clinical implications with respect to their clinical course. Barotrauma complications are evident among COVID-19 patients and has been widely documented in the literature [17,18]. In this single center study, the incidence of barotrauma related complications in COVID-19 patients was 15%, which is comparable to the incidence reported by McGuinness G et al. [18].

The main indication for ICU admission in our study was patients inability to maintain oxygen saturations above 90% on high flow nasal cannula at a rate of 40 L per minute with a fraction of inspired oxygen of 100% (Fi02 100%), or the patient's inability to protect his/her airway. Several other studies have highlighted the significance of previously existing lung disease and the development of barotrauma during the delivery of invasive ventilation. In particular, conditions such as pneumonia, chronic obstructive pulmonary disease (COPD), and lung cancer are found to predispose patients to develop barotrauma [18]. Similarly, in the current study, all patients that developed barotrauma (n = 54) had underlying COVID-19 pneumonia, 11.1% were asthmatic and 1.9% had known case of COPD, In our study, 279 patients received invasive mechanical ventilation, of which 19.4% (n = 54) developed barotrauma that was radiologically confirmed consisting of either a pneumothorax, pneumomediastinum or surgical emphysema. The incidence of barotrauma in COVID-19 patients was higher to previous published studies [18].

With respect to biochemical findings, our study found that white blood count (WBC), neutrophil count, neutrophil lymphocyte ratio (NLR) and lactate dehydrogenase (LDH) were statistically greater among patients that developed barotrauma. These findings are similar to a study by Shioe et al. that reported higher levels of LDH and neutrophils among patients infected with SARS-CoV virus, which were also found to be associated with a higher severity of lung injury. However, the study did not compare their findings to the patients that did not develop barotrauma [20]. Furthermore, we documented a statistically significant lower lymphocyte count in patients whom subsequently developed barotrauma. This finding has been found to be associated with cases of increased severity as T cells play a fundamental role in the initial and continued immune response to coronaviruses, as such T cells are unnaturally depleted [21,22].

From a clinical perspective, a Glasgow Coma Score of 8 or less requiring immediate intubation was a statistically significant variable for the development of barotrauma. This is further highlighted in our results where all 54 patients in the barotrauma group were receiving invasive mechanical ventilation in comparison to the 77.9% whom did not develop barotrauma (p < 0.001).

Acute respiratory distress syndrome (ARDS) has been recognized as an independent risk factor for developing barotrauma while receiving invasive ventilation. In the literature, the incidence of ARDS has been found to vary, reaching 15% in several previous studies [[23], [24], [25], [26], [27]]. In our study, of the 54 patients that developed barotrauma, 52 (96.3%) fit the criteria of ARDS within the 24 h preceding the event.

Moreover, a higher maximum PEEP was found among patients that developed barotrauma compared to the non-barotrauma patients. Although increasing PEEP values on the ventilator is part of guideline-based ‘lung recruitment’ maneuvers, the underlying pathophysiology of COVID-19 lung damage has been postulated to be different than typical ARDS cases [28,29]. The inability to adopt recruitment measures in COVID-19 induced ARDS lungs presents a unique situation where barotrauma is highly likely [28]. Furthermore, contrary to typical ARDS, the presence of conserved lung compliance in COVID-19 induced ARDS is distinctively different [28].

In the current study, the internally validated COVID-GRAM tool was used to assess the severity of disease and care required upon admission. A higher score was found among patients that subsequently developed barotrauma related incidents (p = 0.042). To our knowledge this is the first time the COVID-GRAM has been used to predict barotrauma. We reported a higher score in the patients who developed barotrauma related incidents (P value = 0.042).

In our study, the mortality was significantly greater among patients that developed barotrauma 70.4% compared to 52.9% in the non-barotrauma group (p = 0.025). These findings were echoed in the study by De Lassence et al. [30] Additionally, it should be noted that the mortality rate was also significantly high in the patients whom did not develop barotrauma related events; potentially a representation of a morbid cohort of patients.

Conclusion

In conclusion, there is a significant risk of developing barotrauma in mechanically ventilated COVID-19 patients. In our cohort of patients, we have identified specific variables that have been associated with the development of barotrauma related complications. Selecting the optimal ventilator parameters for lung recruitment while protecting against barotrauma is a delicate balance. In our study, we have reported that the development of barotrauma carries a considerable morbidity and mortality in COVID-19 patients. Further research may be required to support and re-produce these findings which may aid in establishing lung protective protocols tailored to the COVID-19 lung.

Author contribution

All authors contributed equally to the protocol, data collection, statistical analysis, and writing. Provenance and peer review Not commissioned, externally peer-reviewed.

Declaration of competing interest

All authors declare no conflict of interest.