Table 1.
Logistic regression models: results from population data
| Marker | OR per-SD | Lower 95%CI | Upper 95%CI | p-value | Brier score diff. | Increase in AUC | R 2 difference |
|---|---|---|---|---|---|---|---|
| MR-proADM | 1.18 | 1.08 | 1.30 | 0.00035 | -0.00048 | 0.00232 | 0.00349 |
| Nt-proBNP | 1.15 | 1.05 | 1.25 | 0.0017 | -0.00033 | 0.00151 | 0.00272 |
| hs-CRP | 1.08 | 1.00 | 1.17 | 0.044 | -0.00024 | 0.00084 | 0.00111 |
| CT-proAVP | 1.04 | 0.96 | 1.12 | 0.36 | 0.00000 | 0.00014 | 0.00022 |
| MR-proANP | 0.99 | 0.91 | 1.07 | 0.75 | 0.00000 | -0.00004 | 0.00003 |
All biomarkers were log-transformed. The models were adjusted for sex, age and bmi. The basic model had Brier score of 0.163, AUC of 0.759 and R 2 of 0.227