Table 3.
Multivariable Modelsa Assessing the Association of 25(OH)D Levels on Outcomes.
| Characteristic | Length of stay (d) | Oxygen duration (d) | ICU admission | ICU ventilation | Death |
|---|---|---|---|---|---|
| Odds ratio | Odds ratio | Odds ratio | |||
| Age >65 years | 9.3 (.78, 111) | ||||
| Dexamethasone | 4.9 (1.8, 14) | ||||
| Remdesivir | 1.83 (.72, 2.95) | 2.4 (1.1, 3.6) | |||
| 25(OH)D (ng/mL, β) | .011 (−.02, .04) | .017 (−.018, .053) | |||
| Serum sample hospital day (d, β) | .42 (−.02, .86) | .10 (−.35, .55) | −.002 (−.20, .20) | .46 (.16, .76) | .09 (−.31, .49) |
| 25(OH)D < 20 ng/mL | .71 (.21, 2.4) | 2.8 (.58, 13) | 1.6 (.10, 27) | ||
| Hypertension | −1.4 (−2.5, −.32) | ||||
| Pulmonary fibrosis | 12.5 (2.0, 78) | ||||
| Assisted ventilation | 2.0 (.10, 4.0) | ||||
| Vasopressors utilized | 3.6 (1.0, 6.2) | 5.5 (3.3, 7.7) |
a
Multivariable linear regression was performed for outcomes with continuous variables; multivariable logistic regression was used for binary outcomes. Serum 25(OH)D values <20 ng/mL and age >65 years were included in the logistic regression models for ease of interpretation.