Calcification of the thoracic aorta on low-dose chest CT predicts severe COVID-19
Fig 2
Multivariate logistic regression for prediction of severe COVID-19.
Multivariate regression was modeled including the independent variables patient age, sex, an anamnestic cardiovascular comorbidity score and the magnitude of aortic wall calcification (AWC). Logistic regression was fit to a logarithmic scale for AWC to reduce the impact of outliers. Within our model, AWC was the only significant regressor to predict severe COVID-19 (p = 0.004). The 95% confidence interval is illustrated by a grey band (alpha = 0.05).