Time dependency and unique etiology of barotrauma in COVID-19: A retrospective cohort study with landmark analysis and pathological approach.

BACKGROUND: Barotrauma frequently occurs in coronavirus disease 2019. Previous studies have reported barotrauma to be a mortality-risk factor; however, its time-dependent nature and pathophysiology are not elucidated. To investigate the time-dependent characteristics and the etiology of coronavirus disease 2019-related-barotrauma. METHODS AND FINDINGS: We retrospectively reviewed intubated patients with coronavirus disease 2019 from March 2020 to May 2021. We compared the 90-day survival between the barotrauma and non-barotrauma groups and performed landmark analyses on days 7, 14, 21, and 28. Barotrauma within seven days before the landmark was defined as the exposure. Additionally, we evaluated surgically treated cases of coronavirus disease 2019-related pneumothorax. We included 192 patients. Barotrauma developed in 44 patients (22.9%). The barotrauma group's 90-day survival rate was significantly worse (47.7% vs. 82.4%, p < 0.001). In the 7-day landmark analysis, there was no significant difference (75.0% vs. 75.7%, p = 0.79). Contrastingly, in the 14-, 21-, and 28-day landmark analyses, the barotrauma group's survival rates were significantly worse (14-day: 41.7% vs. 69.1%, p = 0.044; 21-day: 16.7% vs. 62.5%, p = 0.014; 28-day: 20.0% vs. 66.7%, p = 0.018). Pathological examination revealed a subpleural hematoma and pulmonary cyst with heterogenous lung inflammation. CONCLUSIONS: Barotrauma was a poor prognostic factor for coronavirus disease 2019, especially in the late phase. Heterogenous inflammation may be a key finding in its mechanism. Barotrauma is a potentially important sign of lung destruction.

Mechano-inflammatory sensitivity of ACE2: Implications for the regional distribution of SARS-CoV-2 injury in the lung.

The coronavirus disease (COVID-19) caused by SARS-CoV-2 can result in severe injury to the lung. Computed tomography images have revealed that the virus preferentially affects the base of the lung, which experiences larger tidal stretches than the apex. We hypothesize that the expression of both the angiotensin converting enzyme-2 (ACE2) receptor for SARS-CoV-2 and the transmembrane serine protease 2 (TMPRSS2) are sensitive to regional cell stretch in the lung. To test this hypothesis, we stretched precision cut lung slices (PCLS) for 12 h with one of the following protocols: 1) unstretched (US); 2) low-stretch (LS), 5% peak-to-peak area strain mimicking the lung base; or 3) high-stretch (HS), the same peak-to-peak area strain superimposed on 10% static area stretch mimicking the lung apex. PCLS were additionally stretched in cigarette smoke extract (CSE) to mimic an acute inflammatory exposure. The expression of ACE2 was higher whereas that of TMPRSS2 was lower in the control samples following LS than HS. CSE-induced inflammation substantially altered the expression of ACE2 with higher levels following HS than LS. These results suggest that ACE2 and TMPRSS2 expression in lung cells is mechanosensitive, which could have implications for the spatial distribution of COVID-19-mediated lung injury and the increased risk for more severe disease in active smokers and patients with COPD.

Modeling lung perfusion abnormalities to explain early COVID-19 hypoxemia.

Early stages of the novel coronavirus disease (COVID-19) are associated with silent hypoxia and poor oxygenation despite relatively minor parenchymal involvement. Although speculated that such paradoxical findings may be explained by impaired hypoxic pulmonary vasoconstriction in infected lung regions, no studies have determined whether such extreme degrees of perfusion redistribution are physiologically plausible, and increasing attention is directed towards thrombotic microembolism as the underlying cause of hypoxemia. Herein, a mathematical model demonstrates that the large amount of pulmonary venous admixture observed in patients with early COVID-19 can be reasonably explained by a combination of pulmonary embolism, ventilation-perfusion mismatching in the noninjured lung, and normal perfusion of the relatively small fraction of injured lung. Although underlying perfusion heterogeneity exacerbates existing shunt and ventilation-perfusion mismatch in the model, the reported hypoxemia severity in early COVID-19 patients is not replicated without either extensive perfusion defects, severe ventilation-perfusion mismatch, or hyperperfusion of nonoxygenated regions.

Can Hyperperfusion of Nonaerated Lung Explain COVID-19 Hypoxia?

Early stages of the novel coronavirus disease (COVID-19) have been associated with 'silent hypoxia' and poor oxygenation despite relatively small fractions of afflicted lung. Although it has been speculated that such paradoxical findings may be explained by impairment of hypoxic pulmonary vasoconstriction in infected lungs regions, no studies have confirmed this hypothesis nor determined whether such extreme degrees of perfusion redistribution are physiologically plausible. Here, we present a mathematical model which provides evidence that the extreme amount of pulmonary shunt observed in patients with early COVID-19 is not plausible without hyperperfusion of the relatively small fraction of injured lung, with three-fold increases in regional perfusion to afflicted regions. Although underlying perfusion heterogeneity (e.g., due to gravity or pulmonary emboli) exacerbated existing shunt in the model, the reported severity of hypoxia in early COVID-19 patients could not be replicated without considerable reduction of vascular resistance in nonoxygenated regions.