Table 2.
Multivariable logistic regression analysis of the association between homocysteine and any placenta-mediated complication a (759 events b), n = 7587
| Variable | Odds ratio (95% CI) | p-value c | |
|---|---|---|---|
| Homocysteine (linear) | 0.0007 | ||
| 5 μmol/L increase | 1.629 | (1.227, 2.161) | |
| Age (restricted cubic spline, three knots) | 0.0031 | ||
| 34 versus 27 years | 1.187 | (1.063, 1.325) | |
| Race | 0.0002 | ||
| Caucasian versus others | 0.644 | (0.509, 0.814) | |
| Education | 0.0056 | ||
| College/University completed versus less than completed | 0.763 | (0.630, 0.924) | |
| Nulliparous | < 0.0001 | ||
| Yes versus no | 1.941 | (1.636, 2.303) | |
| Smoking | < 0.0001 | ||
| No | Reference | ||
| Second-hand | 0705 | (0.392, 1.268) | |
| Med/light smoker (< 10 cigarettes per day) | 1.631 | (1.228, 2.166) | |
| Heavy smoker (> 10 cigarettes per day) | 1.921 | (1.348, 2.737) | |
| Diabetes | 0.0336 | ||
| Yes versus no | 1.687 | (1.041, 2.733) | |
| BMI (restricted cubic spline, four knots) | 0.0499 | ||
| 27.3 versus 21.1 kg/m2 | 1.057 | (0.883, 1.265) | |
| Hormonal birth control prior to conception | 0.3692 | ||
| No | Reference | ||
| Oral | 0.927 | (0.784, 1.096) | |
| Injection or IUD | 0.732 | (0.436, 1.227) | |
| Chronic hypertension | < 0.0001 | ||
| Yes versus no | 2.750 | (1.687, 4.483) | |
| History of PMC (Preeclampsia, placental abruption, IUGR, stillbirth, loss) | 0.0110 | ||
| Yes versus no | 1.359 | (1.073, 1.722) | |
| Folic acid supplementation | 0.7328 | ||
| Yes versus no supplementation | 0.943 | (0.674, 1.320) | |
| Serum folate (linear) | 0.5326 | ||
| 45.1 versus 30.6 nmol/L | 1.025 | (0.949, 1.106) | |
| Gestational age at blood work (restricted cubic spline, four knots) | 0.0004 | ||
| 13.7 versus 12.4 weeks | 0.939 | (0.804, 1.095) | |
aAny placenta-mediated complication- composite of small for gestational age (SGA) <10th percentile, preeclampsia, placental abruption, and pregnancy loss
b79 missing outcome values imputed
cWald test of most meaningful hypotheses, pooled across multiple imputation datasets