Non-invasive positive pressure ventilation (NPPV) emerged as an efficient tool for treatment of Coronavirus disease-19 (COVID-19) pneumonia. The factors influencing NPPV failure are still elusive. The aim of the study was to investigate the relationships between semiquantitative chest computed tomography (CT) scoring and NPPV failure and mortality in patients with COVID-19
The semi-quantitative CT-score was calculated at hospital admission. Subgroups were identified according to the ventilation strategy used (oxygen delivered by Venturi mask n=53; NPPV-responder n=38; NPPV-failure n=21). The study's primary endpoint was the use of NPPV. The secondary endpoints were NPPV failure and in-hospital death, respectively. CT-score progressively increased among groups (6 vs 9 vs 14, p<0.05 between all). CT-score was an independent predictor of all study endpoints (primary endpoint: 1.25 [95% CI 1.1-1.4], p=0.001; NPPV failure: 1.41 [95% CI 1.18-1.69], p<0.001; in-hospital mortality: 1.21 [95% CI 1.07-1.38], p=0.003). According to receiver operator characteristics curve analysis, CT-score was the most accurate variable for prediction of NPPV failure (area under the curve 0.862 with p<0.001; p<0.05 vs other variables).