Integration of sera metabolites and coagulation indicators
yielded higher predictive performance
To assess the potential of metabolites and coagulation indicators as
biomarkers for risk stratification of AMN-SARS-CoV-2, logistic
regression was employed. AUC values of 0.8393, 0.7029, and 0.7653 were
obtained for PCT, Fbg, and FDP, respectively (Figure 5A). Upon inclusion
of the three coagulation indicators in the logistic regression model, an
improved AUC value of 0.8878 was observed. The incorporation of
ornithine, citrulline, L-proline, and ADMA further enhanced the AUC to
0.9082 (Figure 5B). Finally, when both the coagulation indicators and
metabolites were integrated into the logistic regression model, the
highest AUC (0.964) was obtained, indicating better accuracy in
discriminating between AMN-SARS-CoV-2 and SARS-CoV-2-no AMN cases
(Figure 5C).