Table 2. Multinomial logistic regression with dependent variable severity of disease.
| Covariates | Severity of disease | |||||
|---|---|---|---|---|---|---|
| Hospitalization | ICU | |||||
| OR | (95% CI) | p | OR | (95% CI) | p | |
| Univariate models: | ||||||
| CMV-reactive | 3.4 | (2.0, 5.9) | < 0.001 | 4.5 | (2.3, 8.6) | < 0.001 |
| HSV-reactive | 6.0 | (2.6, 13.9) | < 0.001 | 10.2 | (3.0, 35.0) | < 0.001 |
| Age group | 1.4 | (1.2, 1.6) | < 0.001 | 2.1 | (1.7, 2.7) | < 0.001 |
| Male | 1.5 | (0.9, 2.5) | 0.433 | 2.8 | (1.6, 5.2) | 0.001 |
| Cardio-vascular co-morbidity | 13.0 | (5.8, 29.0) | < 0.001 | 32.3 | (13.4, 77.7) | < 0.001 |
| Respiratory co-morbidity | 2.7 | (1.0, 7.5) | 0.065 | 3.7 | (1.3, 10.8) | 0.018 |
| Nephrological co-morbidity * | 24.2 | (1.3, 433.4) | 0.031 | 31.5 | (1.7, 584.0) | 0.021 |
| Diabetes mellitus | 6.3 | (2.1, 18.6) | 0.001 | 9.1 | (3.0, 27.7) | < 0.001 |
| Any comorbidity | 8.2 | (4.5, 15.0) | < 0.001 | 29.7 | (12.8, 69.0) | < 0.001 |
| Multivariate model 1: | ||||||
| CMV-reactive | 3.1 | (1.7, 5.6) | < 0.001 | 5.0 | (2.4, 10.5) | < 0.001 |
| HSV-reactive | 3.6 | (1.5, 8.9) | 0.005 | 4.5 | (1.2, 17.6) | 0.029 |
| Age group | 1.3 | (1.1, 1.6) | 0.003 | 2.2 | (1.7, 2.8) | < 0.001 |
| Multivariate model 2: | ||||||
| CMV-reactive | 3.1 | (1.6, 5.9) | 0.001 | 5.2 | (2.3, 12.1) | < 0.001 |
| HSV-reactive | 4.8 | (1.8, 13.1) | 0.002 | 6.5 | (1.5, 28.2) | 0.012 |
| Age group | 1.0 | (0.8, 1.3) | 0.856 | 1.5 | (1.1, 2.0) | 0.005 |
| Any comorbidity | 8.1 | (4.0, 16.7) | < 0.001 | 22.5 | (8.4, 59.9) | < 0.001 |
The reference category is: Mild disease.
* Firth Penalized Likelihood correction in two separate binary logistic regression models due to quasi-complete separation of the data; Firth’s correction is not yet implemented for multinomial regression.