The amount of subpleural lung infiltrates in patients affected by SARS-CoV2 is proportional to the number of days of spontaneous breathing

Introduction: Specific features differentiate SARS-CoV2-induced acute respiratory failure from the classical definition of acute respiratory distress syndrome (ARDS) [1] By lung computed tomography (CT) it is possible to characterize extension and severity of lung damage and infer regional lungs mechanics during ARDS [2] The role of lung CT for SARS-CoV2 related lung engagement remains to be confirmed Objectives: A primarily subpleural distribution of lung infiltrates induced by SARS-CoV2 has been hypothesized The correlation between the extent of lung infiltrated and days of spontaneous breathing has never been tested before Methods: This study has been performed on CT images collected from 20 patients affected by a severe form of SARS-CoV2 infection in need of intensive care Chest CT, acquired during static conditions and without contrast agent were selected For each CT, twenty equally spaced images between apex and diaphragmatic dome were analyzed Each CT image was divided into three subpleural regions of interest (ROIs - between 0 and 1 cm;1 and 2 cm;2 and 3 cm) concentric to the visceral pleura Four quadrants (external non-dependent,internal nondependent, internal dependent, external dependent) were delineated for each lung, depending on figures centroids (Figure) Quadrants were used to differentiate between atelectasis, affecting dependent regions, and SARS-CoV2 specific lung infiltrates in the external and non-dependent regions Four lung Hounsfield units (HU) compartments were defined [3]: hyper (- 1,000 to - 800 HU), normally (- 800 to - 500 HU), poorly (- 500 to - 100 HU) and not inflated (- 100 to + 100 HU) Friedman's test (α-value < 0 05) was used for statistics Days of spontaneous breathing with and without ventilatory support were defined for each patient Correlations between HU compartments and days of spontaneous breathing were performed (Spearman correlation, rs, α-value < 0 05) Results: Twenty-three chest CT were analyzed Hypo-inflated and not-inflated compartments were significantly more represented in the subpleural regions of the lung (0-1 cm) compared to more central ones (2-3 cm), both in dependent and not dependent quadrants Consistently, the amount of normo-inflated lung was more represented in deeper regions of the lung The hyper-inflated compartment was significantly more represented in non-dependent subpleural lung regions (0-1 cm) A positive correlation was found between days of spontaneous breathing and hypo-/not- inflated lung compartments, in not-dependent (rs = 0 54, p = 0 017) and external regions (rs = 0 49, p = 0 032) Conclusion: The primarily subpleural distribution of SARS-CoV2 specific lung infiltrates has been confirmed using objective lung-CT analysis Further, it subsequently extends towards more central regions of the lung A direct correlation between extent of SARS-CoV2-induced lung infiltrated and days of spontaneous breathing was shown (Table Presented)