Abstract
OBJECTIVES: To assess mortality and different clinical factors derived from the development of atraumatic pneumothorax (PNX) and/or pneumomediastinum (PNMD) in critically ill patients as a consequence of COVID-19-associated lung weakness (CALW). DESIGN: Systematic review with meta-analysis. SETTING: Intensive Care Unit (ICU). PARTICIPANTS: Original research evaluating patients, with or without the need for protective invasive mechanical ventilation (IMV), with a diagnosis of COVID-19, who developed atraumatic PNX or PNMD on admission or during hospital stay. INTERVENTIONS: Data of interest were obtained from each article and analyzed and assessed by the Newcastle-Ottawa Scale. The risk of the variables of interest was assessed with data derived from studies including patients who developed atraumatic PNX or PNMD. MAIN VARIABLES OF INTEREST: Mortality, mean ICU stay and mean PaO2/FiO2 at diagnosis. RESULTS: Information was collected from 12 longitudinal studies. Data from a total of 4901 patients were included in the meta-analysis. A total of 1629 patients had an episode of atraumatic PNX and 253 patients had an episode of atraumatic PNMD. Despite the finding of significantly strong associations, the great heterogeneity between studies implies that the interpretation of results should be made with caution. CONCLUSIONS: Mortality among COVID-19 patients was higher in those who developed atraumatic PNX and/or PNMD compared to those who did not. The mean PaO2/FiO2 index was lower in patients who developed atraumatic PNX and/or PNMD. We propose grouping these cases under the term COVID-19-associated lung weakness (CALW).