TABLE 2.
Unadjusted and multivariable-adjusted associations of maternal and specimen-related factors with the maternal serum 25-hydroxyvitamin D response to prenatal vitamin D-3 supplementation1
| Model A (unadjusted) | Model B [adjusted for initial 25(OH)D and calculated vitamin D dose/wk] | Model C [adjusted for initial 25(OH)D, calculated vitamin D dose/wk, and a limited set of additional covariates2] | Model D [adjusted for initial 25(OH)D, calculated vitamin D dose/wk, and all other listed covariates] | |||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI)3 | P | β (95% CI) | P | β (95% CI) | P | β (95% CI) | P | |
| Maternal enrollment serum 25(OH)D,4 nmol/L | 0.50 (0.29, 0.71) | <0.001 | — | — | 0.39 (0.26, 0.52) | <0.001 | 0.36 (0.23, 0.50) | <0.001 |
| Maternal age, y | 0.02 (–0.72, 0.76) | 0.95 | –0.22 (–0.62, 0.19) | 0.30 | — | — | 0.07 (–0.52, 0.66) | 0.81 |
| Maternal height, cm | –0.35 (–0.93, 0.23) | 0.23 | –0.20 (–0.51, 0.10) | 0.19 | — | — | –0.13 (–0.46, 0.21) | 0.46 |
| Maternal BMI at enrollment,5 kg/m2 | –0.74 (–1.56, 0.09) | 0.08 | — | — | — | — | — | — |
| At 0 IU/wk | — | — | 0.32 (–0.66, 1.30) | 0.52 | 0.18 (–0.88, 1.23) | 0.74 | 0.08 (–1.12, 1.28) | 0.90 |
| At 4200 IU/wk | — | — | –0.52 (–1.26, 0.23) | 0.18 | –0.23 (–1.11, 0.65) | 0.61 | –0.31 (–1.29, 0.68) | 0.54 |
| At 16,800 IU/wk | — | — | –0.46 (–1.50, 0.58) | 0.39 | –0.61 (–1.74, 0.52) | 0.29 | –0.69 (–1.91, 0.53) | 0.27 |
| At 28,000 IU/wk | — | — | –1.76 (–2.60, –0.92) | <0.001 | –1.85 (–2.96, –0.74) | 0.001 | –1.73 (–2.83, –0.62) | 0.002 |
| Parity5,6 | ||||||||
| Primiparous | ref. | |||||||
| Multiparous | 1.15 (–5.16, 7.47) | 0.72 | — | — | — | — | — | — |
| At 0 IU/wk | — | — | 3.26 (–0.69, 7.21) | 0.11 | 4.70 (–0.06, 9.46) | 0.05 | 3.30 (–2.54, 9.13) | 0.27 |
| At 4200 IU/wk | — | — | –2.76 (–9.38, 3.86) | 0.41 | –3.65 (–11.78, 4.50) | 0.38 | –4.14 (–13.21, 4.96) | 0.37 |
| At 16,800 IU/wk | — | — | 3.81 (–4.18, 11.80) | 0.35 | 5.16 (–4.03, 14.35) | 0.27 | 3.70 (–6.02, 13.43) | 0.45 |
| At 28,000 IU/wk | — | — | –4.92 (–11.15, 1.30) | 0.12 | –2.92 (–10.30, 4.46) | 0.44 | –2.73 (–10.53, 5.07) | 0.49 |
| Maternal education | ||||||||
| Secondary; complete or higher | ref. | ref. | ref. | |||||
| Secondary; incomplete (grades 6–9) | 1.89 (–6.42, 10.20) | 0.66 | 0.30 (–4.10, 4.70) | 0.89 | — | — | –0.39 (–5.85, 5.07) | 0.89 |
| Primary; complete (grade 5) | 4.21 (–6.53, 14.96) | 0.44 | 2.51 (–3.30, 8.31) | 0.40 | — | — | –1.53 (–8.30, 5.24) | 0.66 |
| Primary; incomplete (grades 1–4) | –1.54 (–11.17, 8.09) | 0.75 | 0.33 (–4.41, 5.07) | 0.89 | — | — | –0.98 (–6.87, 4.91) | 0.74 |
| No education | –0.43 (–14.87, 14.01) | 0.95 | –7.54 (–17.44, 2.36) | 0.14 | — | — | –6.85 (–18.47, 4.76) | 0.25 |
| Asset index7 | 1.49 (–0.41, 3.40) | 0.13 | 0.32 (–0.66, 1.30) | 0.52 | — | — | 0.21 (–1.13, 1.55) | 0.76 |
| Paan use | ||||||||
| No | ref. | ref. | ref. | |||||
| Yes | –2.02 (–13.32, 9.27) | 0.73 | 4.66 (–1.17, 10.49) | 0.12 | — | — | 5.23 (–1.74, 12.21) | 0.14 |
| Prenatal calcium intake,8 g/d | 6.68 (–5.04, 18.4) | 0.26 | –1.14 (–7.92, 5.64) | 0.74 | — | — | –3.07 (–11.42, 5.27) | 0.44 |
| Maternal enrollment plasma CRP, ln(μg/mL) | –0.68 (–3.40, 2.03) | 0.62 | –0.55 (–2.00, 0.90) | 0.46 | — | — | 0.13 (–1.72, 1.98) | 0.89 |
| Maternal delivery serum retinol, μg/dL | 0.03 (–0.25, 0.32) | 0.81 | — | — | — | — | — | — |
| At 0 IU/wk | — | — | 0.02 (–0.22, 0.26) | 0.88 | — | — | 0.10 (–0.17, 0.37) | 0.45 |
| At 4200 IU/wk | — | — | 0.001 (–0.24, 0.24) | 0.99 | — | — | 0.04 (–0.22, 0.30) | 0.76 |
| At 16,800 IU/wk | — | — | 0.22 (–0.20, 0.63) | 0.31 | — | — | 0.24 (–0.14, 0.63) | 0.21 |
| At 28,000 IU/wk | — | — | –0.39 (–0.68, –0.09) | 0.010 | — | — | –0.23 (–0.52, 0.05) | 0.11 |
| Maternal delivery serum folate, ng/mL | 0.67 (–0.12, 1.45) | 0.10 | 0.49 (0.05, 0.93) | 0.029 | 0.66 (0.08, 1.24) | 0.026 | 0.60 (–0.02, 1.22) | 0.06 |
| Maternal delivery serum ferritin, ln(ng/mL) | –3.66 (–7.57, 0.24) | 0.07 | — | — | — | — | — | — |
| At 0 IU/wk | — | — | –1.62 (–3.75, 0.51) | 0.16 | –1.99 (–5.59, 1.60) | 0.28 | –1.57 (–5.45, 2.32) | 0.43 |
| At 4200 IU/wk | — | — | –5.45 (–10.36, –0.54) | 0.030 | –4.33 (–10.49, 1.82) | 0.17 | –4.03 (–10.48, 2.43) | 0.22 |
| At 16,800 IU/wk | — | — | –6.69 (–12.31, –1.08) | 0.019 | –5.74 (–11.75, 0.26) | 0.06 | –6.80 (–12.42, –1.19) | 0.018 |
| At 28,000 IU/wk | — | — | 0.50 (–3.06, 4.07) | 0.78 | 0.66 (–3.43, 4.75) | 0.75 | 1.38 (–2.93, 5.69) | 0.53 |
| Duration of supplementation, wk | 0.95 (–0.36, 2.27) | 0.16 | 0.27 (–0.48, 1.03) | 0.48 | — | — | –0.06 (–1.00, 0.88) | 0.90 |
| Season of blood collection | ||||||||
| Dec–Feb | ref. | ref. | ref. | ref. | ||||
| Mar–May | –2.81 (–12.97, 7.35) | 0.59 | 4.04 (–1.57, 9.66) | 0.16 | 0.99 (–5.27, 7.24) | 0.76 | –0.60 (–7.37, 6.17) | 0.86 |
| Jun–Aug | 1.11 (–7.87, 10.08) | 0.81 | 8.45 (3.63, 13.26) | 0.001 | 8.84 (2.40, 15.28) | 0.007 | 6.78 (–0.06, 13.62) | 0.052 |
| Sep–Nov | –2.30 (–10.18, 5.57) | 0.57 | 1.59 (–2.65, 5.83) | 0.46 | 0.80 (–3.78, 5.37) | 0.73 | –0.48 (–5.18, 4.21) | 0.84 |
| Time of day of blood collection9 | ||||||||
| Morning | ref. | ref. | ref. | |||||
| Afternoon | 5.32 (–2.31, 12.95) | 0.17 | 2.53 (–1.62, 6.68) | 0.23 | — | — | –0.13 (–5.13, 4.86) | 0.96 |
| Evening | 4.65 (–2.87, 12.18) | 0.23 | 1.39 (–2.55, 5.33) | 0.49 | — | — | 1.75 (–2.72, 6.22) | 0.44 |
| Days since last dose10 | –1.32 (–2.38, –0.25) | 0.016 | –0.31 (–0.91, 0.29) | 0.31 | –0.32 (–0.96, 0.33) | 0.34 | –0.29 (–0.96, 0.38) | 0.39 |
| Time between blood collection and 25(OH)D analysis,11 mo | 0.36 (–0.05, 0.77) | 0.08 | 0.23 (–0.01, 0.47) | 0.06 | 0.21 (–0.07, 0.50) | 0.14 | 0.65 (0.24, 1.07) | 0.002 |
| Assay drift,12 mo | 0.02 (–0.37, 0.40) | 0.93 | –0.08 (–0.30, 0.14) | 0.47 | — | — | –0.59 (–0.98, –0.21) | 0.002 |
Model A was an unadjusted model with 25(OH)D at delivery as the outcome and the listed predictor as the exposure (1 individual model was created for each listed predictor). Model B additionally adjusted for maternal 25(OH)D at enrollment and calculated vitamin D dose per week (1 individual model was created for each listed predictor). Model C was a single parsimonious multivariable model that included enrollment 25(OH)D, calculated vitamin D dose per week, and any predictor variables with a P value <0.1 in Models A and/or B, alongside any variables that were a priori hypothesized to influence attained 25(OH)D concentrations. Model D adjusted for all listed hypothesized predictors in a single model. Robust SEs were estimated to account for heteroskedasticity. CRP, C-reactive protein; ref., reference; 25(OH)D, 25-hydroxyvitamin D.
Covariates adjusted for included: maternal BMI at enrollment, parity, maternal delivery serum folate, maternal delivery serum ferritin, season of blood collection, days since last dose, and time between blood collection and analysis.
β represents the difference in delivery 25(OH)D concentrations for every 1-unit increase in continuous variables or compared with the reference category for categorical variables. For selected predictor variables for which we examined interactions with vitamin D dose, associations are shown at specified supplemental vitamin D doses; in these instances, coefficients are predicted margins that represent the change in 25(OH)D for a 1-unit increase in the predictor (continuous variable) or difference compared with the reference category (categorical variable), at the specified vitamin D dose level.
Effect estimates of the association of enrollment 25(OH)D with delivery 25(OH)D are not shown for Model B because enrollment 25(OH)D was a covariate for all other predictors assessed. Hence, a different effect estimate was generated for 25(OH)D corresponding to each predictor variable investigated (Model B), rather than a single unique effect estimate for 25(OH)D.
Interaction with dose in adjusted and multivariate models.
Defined based on the total number of previous live births, irrespective of previous miscarriage or abortions, and was inclusive of the current pregnancy. Hence, parity was categorized as primiparous (no previous live birth) or multiparous (≥1 live birth).
Determined by ownership of household items, using principal components analysis.
Estimated by a targeted, nonquantitative FFQ.
Categorized as morning (00:00 to 11:59), afternoon (12:00 to 16:59), or evening (17:00 to 23:59).
Defined as the number of days between blood sample collection and the last administered vitamin D (or placebo) intervention dose, whereby the maximum number of days was truncated at 14.
Used to examine potential minor degradation during freezer storage.
Estimated by the number of months between completion of the first 25(OH)D assay and the assay concerning the 25(OH)D measurement of interest.