Vitamin D requirements have become one of the most highly debated and controversial topics in nutrition. Recommendations for vitamin D intake during pregnancy are a central part of this discussion. The publication of a controlled trial of vitamin D supplementation in pregnant women by Hollis and coworkers in this issue of JBMR provides important data about the effects of different doses of vitamin D3 supplementation on serum 25-hydroxyvitamin D [25(OH)D] levels.1
Hollis and coworkers report that providing daily supplements of 2000 or 4000 IU of vitamin D3 during the second and third trimesters of pregnancy led to higher serum 25(OH)D concentrations in women during pregnancy and in their newborn infants than supplements of 400 IU/d.1 As suggested by a footnote, further results from this controlled trial may be provided in the future in the peer-reviewed literature. For now, an evaluation of these results and how they might guide vitamin D recommendations during pregnancy is appropriate.
In this report,1 increases in both vitamin D supplementation and serum 25(OH)D level were associated with higher serum 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] levels. This outcome was the primary one for which this study was designed and sample size determined. However, there was no evidence of any clinical outcome differences based on supplement dose, and such outcomes were not part of the original study design. Specifically, there were no significant differences based on supplement dose in the 350 subjects who completed the study in C-section rate, gestational age, birth weight, or neonatal intensive-care unit (NICU) admission rate. Rates of preeclampsia and preterm birth based on supplement dose were not provided. Although NICU admission is not a complete assessment measure for neonatal outcome, these findings provide evidence that different levels of vitamin D supplementation had no significant positive or negative effect on the need for advanced neonatal care. Of note is that although an intention-to-treat analysis was indicated in the methods, outcome data are not available for the over 100 subjects who discontinued the study or were withdrawn from it without a known pregnancy loss.
Outcomes such as neonatal or maternal bone mineral content, postneonatal infant health outcomes, and long-term bone outcomes for mother and infant are not provided in this report.1 There is conflicting scientific evidence regarding the long-term effects of maternal vitamin D supplementation on childhood health outcomes2, 3 and no evidence about long- or short-term outcomes for maternal bone health from supplementation at doses of 4000 IU/d during pregnancy.
In evaluating their data, the authors indicated, “These findings suggest that the current vitamin D EAR and RDA for pregnant women issued in 2010 by the Institute of Medicine should be raised to 4000 IU vitamin D per day so that all women regardless of race attain optimal nutritional and hormonal vitamin D status throughout pregnancy.”1
This statement is considerably different from recent recommendations from several policy-oriented groups, including the Institute of Medicine (IOM),4 the American College of Obstetricians and Gynecologists (ACOG),5 and the Endocrine Society.6 Therefore, careful consideration should be made as to whether these data,1 in the absence of evidence for differences in clinical outcomes according to vitamin D dose or serum 25(OH)D level, warrant reconsideration of those recommendations and routine supplementation of all pregnant women with 4000 IU/d of vitamin D.
Reconsideration of the recently released IOM vitamin D dietary recommendations4 would be enhanced by an increased understanding of the terms EAR and RDA. The estimated average requirement (EAR) and the recommended dietary allowance (RDA) have different meanings for individual diets and for public health policy.4 The EAR for vitamin D of 400 IU/d was derived by the IOM based on a target 25(OH)D level of 40 nmol/L. It represents a recommendation most relevant to population groups and is a value at which about half of young adults, including pregnant women, meet their requirements. The RDA of 600 IU/d was set based on a target 25(OH)D level of 50 nmol/L by the IOM. It is an intake recommendation that is more useful for evaluating an individual's diet and reflects intakes that meet the needs of at least 97.5% of the population.
Once one can agree on the indicators for an “optimal” serum 25(OH)D value, dose-response and population-distribution data can be used to derive estimates of the EAR and subsequently the RDA based on this value.4 Specifically, in the report by Hollis and colleagues,1 about half of mothers achieved a 25(OH)D level of 80 nmol/L, a value above both the EAR- and RDA-targeted 25(OH)D values with a supplement of 400 IU/d.
An important issue is whether it is necessary to target a 25(OH)D level of at least 80 nmol/L during pregnancy or in newborn infants rather than the 50 nmol/L value for the RDA target set by the IOM. In this report, data are provided showing an increase in the urinary calcium/creatinine (Ca/Cr) ratio with higher 25(OH)D levels that according to the authors “appear” to “normalize” at about 75 nmol/L. However, no statistical analysis of this conclusion is provided. Although a low Ca/Cr ratio has been associated with preeclampsia,7 there is no evidence from controlled trials for a decrease in preeclampsia with supplemental vitamin D to increase the Ca/Cr ratio and no established definition of a normal urine Ca/Cr ratio related to biologic outcomes in the second and third trimesters of pregnancy. A spot Ca/Cr ratio in the unfasted state may not identify risks or benefits of either calcium or vitamin D supplementation in any population.8 A recent study found no significant relationship between early-pregnancy 25(OH)D levels and preeclampsia.9
Additionally, there is no evidence for any targeted value of 1,25(OH)2D3 related to clinical maternal or neonatal outcomes or evidence that maximizing 1,25(OH)2D has specific health benefits. It would be important to evaluate results in this cohort among potentially clinically relevant biochemical outcomes indices achieved by supplementation and neonatal outcomes (eg, birth weight, gestational age, preeclampsia risk).
Hollis and colleagues comment on the optimization of calcium absorption at a level of 25(OH)D of 80 nmol/L.1 This value, however, is not established in pregnancy, and calcium absorption efficiency was optimized at much lower 25(OH)D levels in large studies involving older children and adolescents.10, 11 Whether 80 nmol/L is a valid target for optimizing calcium absorption in nonpregnant adults remains highly controversial. During pregnancy, calcium absorption efficiency is very high and has not been shown to be related to high-dose vitamin D supplementation.12 Finally, no data exist relating a range of neonatal 25(OH)D levels to calcium absorption efficiency. Clinical evidence of vitamin D deficiency has not been shown for serum 25(OH)D values above approximately 25 to 30 nmol/L in neonates and young infants.1
From a safety perspective, it is reassuring that no complications were seen in the women who completed this study.1 However, this finding needs to be interpreted with caution. As noted by the authors, they evaluated supplementation only during the second and third trimesters of pregnancy. Therefore, recommendations and safety considerations for the entirety of pregnancy are premature. The randomization as required by the institutional review board prevented those with high baseline values from receiving the highest supplement dose. Furthermore, several women with high serum 25 (OH)D levels were taken off supplementation.
The small sample size and high dropout and withdrawal rates are also not reassuring proof of safety for the entire population of pregnant women in the United States. The proportion of patients who could be analyzed was well below the 95% target after randomization for most intention-to-treat analyses.13. A sample size of about 350 women would be unlikely to find a complication occurring, for example, in 0.3% of pregnant women in the United States each year. In summary, the small sample size and protocol constraints could have failed to demonstrate significant increases in complications affecting many pregnant women each year.
It is likely, however, that a pregnancy supplement of 4000 IU/d of vitamin is safe for most women and that supplement dose is the amount set by the IOM as its tolerable upper intake level for total daily vitamin D intake.4 However, concerns about the long-term safety of high-dose vitamin D supplementation in pregnancy remain. These concerns, including relevant animal research data, were described in a recent report by Roth.14 The key conclusion is that long-term safety data are needed in a large population cohort before being certain of the safety of high doses of vitamin D during pregnancy.
Faced with a myriad of recommendations, both pregnant women and their caregivers must decide how to proceed. The RDA of 600 IU/d total vitamin D intake remains well supported by current data. Alternately, the not dissimilar view of the ACOG that vitamin D intakes of 1000 to 2000 IU/d are safe may be considered as a guide, although the ACOG does not specifically recommend these intakes or routine screening of serum 25(OH)D levels during pregnancy.5
Pregnant women can readily achieve a vitamin D intake of 600 IU/d with a typical diet including daily supplementation with a prenatal vitamin containing 400 IU of vitamin D. Achieving an intake of vitamin D at or above 1000 IU/d may require a vitamin D supplement in addition to a typical prenatal multivitamin supplement. Costs and benefits related to additional supplements and to routine screening of pregnant women for serum 25(OH)D should be evaluated by independent groups such as the US Preventative Services Task Force before implementation.
The study by Hollis and colleagues1 provides novel data about vitamin D in pregnancy. However, currently published research data fall short of providing evidence for adopting a routine vitamin D supplementation strategy of 4000 IU/d or even 2000 IU/d for all pregnant women. The choice to provide such a supplement should be an individual one. We should insist on adequately sized controlled trials with biologically meaningful endpoints determined before the trial begins related to maternal and infant health, as well as long-term safety data, in order to best advise pregnant women and their caregivers related to vitamin D.