P0ST-HOC COMPARISON OF VITAMIN D STATUS AT THREE TIMEPOINTS DURING PREGNANCY DEMONSTRATES LOWER RISK OF PRETERM BIRTH WITH HIGHER VITAMIN D CLOSER TO DELIVERY

摘要

There have been observational reports that maternal vitamin D status at baseline and not closest to delivery is a better predictor of pregnancy outcomes, suggesting that a cascade of events is set into motion that is not modifiable by vitamin D supplementation during later pregnancy. To address this issue, in this exploratory post-hoc analysis using correlation and logistic regression, we sought to measure the strength of the association between serum 25(OH)D concentrations at 3 timepoints during pregnancy: baseline, 1^st trimester (<16 wks); 2^nd trimester (16-26 wks); and 3^rd trimester (>≥27 wks) and preterm birth. It was hypothesized that the 25(OH)D value closest to delivery would be most significantly associated with preterm birth. To accomplish this objective, the datasets from NICHD (n=333) and Thrasher Research Fund (n=154) vitamin D supplementation pregnancy studies were combined.The results of this analysis were that 25(OH)D values closer to delivery were more strongly correlated with gestational age at delivery than earlier values: 1^st trimester: r=0.11 (p=0.02); 2^nd trimester: r=0.08 (p=0.09); and 3^rd trimester: r=0.15 (p=0.001). When logistic regression was performed with preterm birth (<37 weeks) as the outcome and 25(OH)D quartiles as the predictor variable, adjusting for study and participant race/ethnicity, as with the correlation analysis, the measurements closer to delivery were more significantly associated and had a higher magnitude of effect. That is, at baseline, those who had serum concentrations <50 nmol/L (20 ng/mL) had 3.3 times of odds of a preterm birth compared to those with serum concentrations ≥100 nmol/L (40 ng/mL; p=0.27). At 2^nd trimester, the odds were 2.0 fold (p=0.21) and at the end of pregnancy, the odds were 3.8 fold (p=0.01). The major findings from this exploratory analysis were: (1) maternal vitamin D status closest to delivery date was more significantly associated with preterm birth, suggesting that later intervention as a rescue treatment may positively impact the risk of preterm delivery, and (2) a serum concentration of 100 nmol/L (40 ng/mL) in the 3^rd trimester was associated with a 47% reduction in preterm births.