Table 2.
Multivariable linear regression analysis of the determinants of plasma homocysteine, with plasma homocysteine as a continuous dependent variable (n 7587)
| Variable | Effect | 95 % CI | P-value* |
|---|---|---|---|
| Demographic characteristics | |||
| Age (rcs, three knots) | 0·0013 | ||
| 34 v. 27 years | −0·027 | −0·070, 0·016 | |
| Race | <0·0001 | ||
| Caucasian | Reference | ||
| African | −0·370 | −0·561, −0·179 | |
| Middle Eastern | −0·438 | −0·594, −0·282 | |
| Asian | −0·468 | −0·588, −0·349 | |
| Other | −0·405 | −0·612, −0·197 | |
| Education | 0·3987 | ||
| College/university completed v. less than completed | −0·032 | −0·105, 0·042 | |
| Health indicators | |||
| BMI (rcs, three knots) | 0·0103 | ||
| 27·3 v. 21·1 kg/m2 | −0·070 | −0·119, −0·022 | |
| Diabetes | 0·6197 | ||
| Yes v. no | −0·056 | −0·279, 0·166 | |
| Chronic hypertension | 0·0001 | ||
| Yes v. no | 0·502 | 0·250, 0·754 | |
| Pregnancy characteristics | |||
| Nulliparous | <0·0001 | ||
| Yes v. no | 0·140 | 0·073, 0·207 | |
| Smoking | <0·0001 | ||
| No | Reference | ||
| Second-hand exposure | 0·298 | 0·097, 0·499 | |
| Med/light smoker (<10 cigarettes per day) | 0·418 | 0·297, 0·539 | |
| Heavy smoker (≥10 cigarettes per day) | 0·864 | 0·705, 1·022 | |
| Hormonal birth control prior to conception | 0·1747 | ||
| No | Reference | ||
| Oral | −0·046 | −0·108, 0·015 | |
| Injection or IUD | 0·091 | −0·088, 0·270 | |
| History of PMC (pre-eclampsia, placental abruption, SGA, stillbirth/loss) | 0·0078 | ||
| Yes v. no | −0·124 | −0·215, −0·033 | |
| Folic acid supplementation | <0·0001 | ||
| None | Reference | ||
| >0 and ≤1 mg | −0·310 | −0·442, −0·178 | |
| >1 mg | −0·329 | −0·476, −0·181 | |
| Homocysteine metabolism | |||
| Gestational age at bloodwork (rcs, five knots) | <0·0001 | ||
| 13·4 v. 12·4 weeks | 0·016 | −0·063, 0·095 | |
| Serum folate (rcs, five knots) | <0·0001 | ||
| 45·1 v. 30·6 nmol/l | 0·137 | 0·065, 0·210 | |
rcs, restricted cubic spline; PMC, placenta-mediated complication.
Wald test of most meaningful hypotheses, pooled across multiple imputation datasets.