Table 3.
Multiple Logistic Regression Estimates Based on Backward Selection Eliminating from a Model with Known Risk Factors (Gestational Age, Birth Weight,) and Significant Risk Factors Reported in Table 2a.
| Specification | β | OR (95% CI) | P-value |
|---|---|---|---|
| Gestational age, wk | |||
| ≤ 28 | 2.57 | 13.11 (5.90–29.15) | <0.001 |
| 28–30 | 1.72 | 5.61 (3.12–10.09) | <0.001 |
| 30–32 | 1.03 | 2.80 (1.66–4.74) | <0.001 |
| >32 | 1 | [Reference] | |
| Birth weight, g | |||
| <1,000 | 2.27 | 9.65 (4.64–20.06) | <0.001 |
| 1,000–1,500 | 0.98 | 2.67 (1.640–4.34) | <0.001 |
| ≥1,500 | 1 | [Reference] | |
| Weight gain rate, g/d | |||
| ≤ 12 | 0.98 | 2.65 (1.49–4.72) | 0.001 |
| 12–18 | 0.66 | 1.94 (1.21–3.11) | 0.006 |
| >18 | 1 | [Reference] | |
| Blood transfusion | 0.71 | 2.03 (1.14–3.64) | 0.017 |
| Invasive mechanical ventilation | 0.56 | 1.74 (1.15–2.66) | 0.009 |
| NT-proBNP ≥ 25,000 ng/L | 0.41 | 1.51 (1.00–2.28)b | 0.048 |
| Constant | −4.40 | 0.01 | <0.001 |
CI, confidence interval; OR, odds ratio; PMA, postmenstrual age; ROP, retinopathy of prematurity.
a
The multivariate model started with all the significant predictors in Tables 1, 2. The model went through the stepwise selection by keeping only the statistical significant variables in the final multivariate model.
b
If three decimal places are reserved, the 95% CI is 1.004–2.282.