Covid-19 interstitial pneumonia can lead to acute respiratory failure (ARF). Some Authors have described a discrepancy between the initial relatively well-preserved lung mechanics, and the severity of hypoxemia, with a higher-than-expected level of pulmonary shunt. The aims were to non-invasively investigate shunt fraction in patients with Covid-19-ARF compared with patients with other causes of ARF.

Observational study of hospitalized patients with Covid-19-ARF and other causes of ARF at PG23 Hospital in Bergamo (Italy) between Jun 2020 and Nov 2021. Shunt fraction was estimated by a non-invasive system, during spontaneous breathing (Beacon® Caresystem) and chest CT scan data were processed by a medical imaging automated Artificial Intelligence (AI) software (Aview: Lung Textures® Coreline Soft).

We enrolled 43 adult patients (8 female), mean age (±SD) 67±12 years and mean BMI 26,8±4,5 Kg/m2. Covid-19-ARF patients were 67% (29/43). CAP was the most common cause of other ARF (10/14). No differences in terms of age and BMI were described between the two groups. Pulmonary gas exchange impairment was similar, median PaO2/FIO2 ratio was 254 [IQR 162-297] in Covid-19-ARF and 269 [IQR 201-296] in other ARF (p=0.41). No significant differences in radiological pattern or degree of involvement on chest CT scan were identified by AI software analysis between the two groups. Nevertheless, mean shunt fraction resulted significantly increased in Covid-19-ARF (18±7%) than in other causes of ARF (12±9%; p=0.04). 

The shunt fraction appears to be increased in Covid-19-ARF compared with other causes of ARF, without any difference in quantitative involvement and radiological pattern at chest CT scan, however further investigations are needed to validate this non-invasive technique.