Vitamin D3 for Life

Live Enlightened

Physiology

Vitamin D in Pregnancy: Maternal and Fetal Outcomes: Full Evidence Review

Previous related pageVitamin D in Pregnancy: Maternal and Fetal Outcomes: Study SummaryNext related pageHow to Read Vitamin D Studies During Pregnancy
Study at a glance

What was the question?

Does vitamin D status or supplementation during pregnancy relate to better maternal or infant health? The review separates changes in blood measurements from effects on specific health outcomes.

Who was studied?

Five core investigations examined women before or during pregnancy, and their babies. They used different starting concentrations, supplement comparisons and measurement times. The review also considers an infant-growth report that was later corrected.

What did the studies find?

Supplements raised blood status in several trials. Two trials did not show a clear overall benefit in the amount of mineral measured in babies’ bones. One found a possible benefit among winter births in an additional analysis. The other had a small imaging sample. A trial assigning supplements by chance did not show clear prevention of preeclampsia, a pregnancy complication. Associations based on observed blood status did not prove a treatment effect. The growth trial did not detect an infant-growth benefit.

What are the limits?

Not finding a clear difference does not prove treatments have identical effects. Missing measurements and selected groups limit how widely the findings apply. A recent diabetes report was excluded because of conflicting results within its report. These findings do not establish a personal dose, ideal blood target or conclusion about every pregnancy outcome.

Read the full article below ↓

Vitamin D status during pregnancy can be associated with later health outcomes without showing that supplementation prevents them. Trials also need to distinguish a higher blood concentration from an improvement in maternal or infant health. This review compares assessed evidence on neonatal bone, preeclampsia, biochemical response and infant growth, retaining both positive signals and important null findings. For the shorter account, read Vitamin D in Pregnancy: Maternal and Fetal Outcomes: Study Summary.

Start with the outcome and the comparison

The usual circulating status marker is 25-hydroxyvitamin D, abbreviated 25(OH)D. It is different from vitamin D3 given in a supplement and from the active hormone, calcitriol. Vitamin D Testing: 25-OH-D vs 1,25-OH₂-D explains those distinctions. Raising 25(OH)D establishes a biochemical response; it does not by itself establish that a pregnancy complication was prevented. This broader interpretive issue is addressed in Vitamin D Status vs Vitamin D Effect.

A randomized comparison concerns assigned treatment groups. An analysis of the concentrations participants eventually achieved concerns a different exposure: the achieved result also reflects adherence, starting status, metabolism and other circumstances. Even within a randomized trial, grouping participants by their later concentration can reintroduce confounding. A trial’s randomized result and its achieved-status association therefore should not be presented as equivalent tests of causation.

The outcome itself also matters. Neonatal bone mineral content measures mineral in the scanned region. It is not a fracture count or a complete assessment of skeletal health. Preeclampsia is a maternal complication, while infant growth is measured after birth. These endpoints cannot be combined into a general claim that vitamin D either benefits or fails to benefit every aspect of pregnancy. Bone Mineralisation, Turnover and Fractures Measure Different Outcomes explains the distinctions among skeletal measurements.

MAVIDOS: no overall neonatal bone benefit detected, with a qualified seasonal signal

The MAVIDOS trial randomized 1,134 women in the United Kingdom to vitamin D3 or placebo from around 14 weeks. The assigned supplement was 1,000 international units daily, with usual-care supplementation up to 400 units permitted. These amounts describe the trial, not a personal recommendation. Women with screening 25(OH)D below 25 or above 100 nmol/L were excluded, limiting conclusions about markedly deficient or higher-status pregnancies.

The main neonatal whole-body bone-mineral-content analysis included 665 infants, rather than every randomized pregnancy or all 737 infants with a usable scan at any site. Mean whole-body mineral content was 60.5 g in the placebo group and 61.6 g in the supplemented group. The overall difference was not statistically significant. The trial therefore did not demonstrate an overall benefit for its primary neonatal bone endpoint.

That finding does not prove that the two strategies have exactly identical effects. Missing scans and the uncertainty around an estimate remain relevant. Nor does it establish that vitamin D has no biological role in bone. The experiment tested an added supplementation strategy in a selected population. General mineral physiology, introduced in Vitamin D and Calcium Physiology and Vitamin D and Phosphate Regulation, does not by itself predict the size of an incremental trial benefit.

The winter-birth subgroup showed a larger mineral-content difference: about 5.5 g, with a 95% confidence interval of 1.8 to 9.1 g. The season-by-treatment interaction had a reported P value of .04. This was a secondary finding among several subgroup comparisons. It is a potentially meaningful signal, but it should not replace the primary overall result or become a universal seasonal prescribing rule.

The trial also raised maternal 25(OH)D. That biochemical separation supports the conclusion that the intervention changed exposure, even though it did not demonstrate the primary benefit overall. A reported postpartum-hemorrhage difference was vulnerable to classification problems and is not treated here as established prevention evidence. Each endpoint requires its own quality assessment rather than borrowing credibility from the trial’s main design.

Preg-D: biochemical improvement with an uncertain neonatal result

The 2026 preg-D report describes a separate Lebanese randomized trial of 330 pregnant adults. Participants entered before 17.5 weeks after screening in a specified low-status range. The comparison was between two active vitamin D3 regimens: 20,000 IU weekly versus 10,000 IU every two weeks, continued through delivery. These are research regimens, not personal dosing instructions. The study excluded severe deficiency under its screening criteria and allowed limited additional prenatal vitamin D.

Standardized biochemical measurements were available for 192 women. The higher-dose group more often reached the trial’s maternal concentration threshold at delivery, and maternal and cord concentrations were higher. These results support a biochemical effect under the tested conditions. The threshold is a trial outcome definition, not proof of a universally appropriate target.

The neonatal bone result was much less precise. Although 67 infants had acceptable scans, the table for subtotal whole-body mineral content analyzed 65. Reported means were approximately 46.0 g in the higher-dose group and 42.2 g in the lower-dose group, with P=.059; the trial used a stricter .025 significance level for its two primary endpoints. A numerical difference in a small sample does not establish a benefit, and a nonsignificant result does not establish equivalence.

Selection is central to interpretation. Only about 58% of randomized women contributed the complete standardized biochemical data, and about one fifth contributed usable neonatal scans. The supplementary comparison showed that excluded participants differed in age, education and veiling. Balance between the analyzed treatment groups does not remove every consequence of missing outcomes. Scan refusal, movement and selected stored samples further restrict generalization.

There are also reporting discrepancies in participant counts and sample-size explanations. These are retained as limitations rather than silently repaired. The main supported contrast remains clear: biochemical improvement was demonstrated, while a neonatal bone benefit was not established in the small selected imaging sample. The distinction is related to the broader explanation in Hormone Concentration, Receptor Response and Clinical Benefit: a change in exposure or hormone concentration does not automatically establish clinical benefit.

Preeclampsia: the randomized and observational VDAART results differ

Mirzakhani and colleagues examined preeclampsia in the VDAART trial. The trial assigned total vitamin D3 intakes of 4,400 versus 400 IU daily beginning at 10 to 18 weeks in families selected for an allergic-disease history. These trial amounts are not personal treatment recommendations. Among 816 pregnancies with outcome information, there were 33 preeclampsia cases in the higher-exposure group and 34 in the comparison group. Each group contained 408 participants.

The randomized risk ratio was 0.97, with a 95% confidence interval of 0.61 to 1.53. That result did not demonstrate prevention. The interval includes both a potentially lower and a potentially higher risk, so it should not be described as proof that the treatment has no possible effect. Preeclampsia was a secondary outcome, and the selection criteria limit transfer to every pregnancy population.

The same report found an association between higher achieved status at relevant time points and lower preeclampsia odds. This analysis grouped women by measured status rather than preserving only their randomized assignment. It therefore cannot establish that deliberately raising a particular woman’s concentration to that level prevents the condition. Covariate adjustment helps address measured differences but cannot turn the achieved-status comparison into a new randomized trial.

The adjusted odds-ratio interval differs between parts of the report, and odds ratios should not be substituted for the randomized risk ratio. Those reporting details are additional reasons to keep the main randomized finding and the observational association clearly separated. Detailed genomic analyses are not used here because their separate supplementary files were not assessed. The clinical result does not require an unverified molecular explanation.

Before conception: EAGeR shows why adjustment matters

Alkhalaf and colleagues’ 2026 analysis used data from EAGeR, a trial originally concerned with aspirin in women with previous pregnancy loss. Vitamin D exposure was not randomized. The investigators related preconception and early-pregnancy status to later preeclampsia. The vitamin D analysis was therefore observational, even though its participants came from a randomized trial of a different intervention.

The analysis included 552 complete preconception records and 530 records at eight weeks among a selected live-birth population. Before body-mass-index adjustment, the reported association comparing deficient with adequate preconception status had a risk ratio of 2.32, with a 95% confidence interval of 1.09 to 4.95. After that adjustment, it was 1.45, with an interval of 0.64 to 3.29. The early-pregnancy estimate was also imprecise.

The change after adjustment shows that body size materially affects interpretation of this association. It does not prove that body size fully explains it, establish a causal mediation pathway, or demonstrate that vitamin D treatment would alter risk. Selection for previous pregnancy loss, missing observations and conditioning on available pregnancy outcomes also affect applicability.

An exploratory concentration change point in the paper was chosen by comparing possible statistical models. It is not a treatment threshold. Sparse observations at higher concentrations and apparent reporting inconsistencies further limit precise subgroup inference. Why Vitamin D Blood Test Results Need Context explains why a status result needs information about the person, assay and timing; this pregnancy example adds the distinction between confounder adjustment and causal evidence.

Hollis: successful repletion is a different endpoint

Hollis and colleagues’ pregnancy trial provides a useful biochemical comparison alongside the clinical-outcome studies. Assigned vitamin D3 regimens beginning around 12 to 16 weeks changed measured status and attainment of a study-defined concentration threshold. The trial did not demonstrate significant group differences in gestational age, birthweight or higher-level neonatal care.

Those clinical comparisons were not a definitive test of every possible pregnancy benefit. Follow-up was incomplete, adherence was imperfect, and the trial’s main objective was biochemical status. Its abstract and body also report different enrollment counts. The findings support the narrower conclusion that an exposure can improve a laboratory endpoint without establishing a clinical advantage in the same report.

The monitoring data likewise do not establish universal safety. Trial participants, duration, starting concentrations and clinical oversight matter. No dose from this or another study is presented as a personal regimen, and no observed biochemical threshold is adopted as an individual goal. A pregnancy treatment decision requires a different kind of assessment from reading an experimental comparison.

Infant growth and the importance of corrected sources

The Bangladesh MDIG trial investigated supplementation during pregnancy and, in one strategy, after birth. Its principal growth report did not detect an infant-growth benefit. This is useful contrary evidence against assuming that a biochemical response must improve growth. It concerns a different population, timing and outcome from neonatal bone scans or maternal preeclampsia.

A 2021 correction changed coding for parity and infant sex and affected numerical material in the report and supplement. The publisher stated that the text and conclusions were unchanged. This review therefore uses only the correction-disclosed qualitative growth conclusion. It does not reproduce affected older estimates, denominators, subgroup results or adverse-event comparisons without current corrected numerical material.

That limited use is not a claim that the interventions were equivalent or that supplementation cannot benefit any infant outcome. Later reports from the same underlying trial are not independent replications simply because they examine another endpoint. Likewise, neonatal measurements and later infant growth should remain distinct from a review of long-term developmental programming.

A recent diabetes report was assessed but not used to estimate benefit

A 2026 randomized report in women with established gestational diabetes was examined in full, including its tables, model figure and supplement. It contained material inconsistencies in change calculations, glucose results, model descriptions and statistical reporting. These problems prevent reliable effect estimation from that report in this review.

The study is therefore not counted as a qualifying confirmation of benefit or harm. This exclusion is a judgment about the report’s interpretability, not an allegation of misconduct or a claim that it has been retracted. Because the participants already had gestational diabetes, the comparison also could not simply be presented as a trial preventing its initial development. Excluding it leaves uncertainty; it does not establish a diabetes null result.

What can responsibly be concluded

QuestionMain interpretation
Does supplementation raise status?Several trials demonstrate biochemical improvement under their tested conditions
Does it improve neonatal bone overall?MAVIDOS and preg-D did not establish their overall bone-mineral-content benefit; subgroup signals and imprecision remain
Does it prevent preeclampsia?The assessed VDAART randomized comparison did not demonstrate prevention; status associations are not treatment effects
Does early status establish a target?EAGeR’s observational and exploratory findings do not establish an individual treatment threshold
Does it improve infant growth?The correction-disclosed MDIG conclusion did not detect benefit in that trial

The evidence supports neither a universal benefit claim nor a universal claim of no effect. Population, baseline status, start time, comparator, adherence, missing data and endpoint all change the question being answered. These selected studies do not settle every obstetric outcome, fetal-loss question, diabetes question or severely deficient population. Biological importance and benefit from increasing a particular exposure remain separate scientific questions.

Sources and access

Principal reports and claim-relevant material were assessed. Uninspected Hollis plots are not used for threshold claims; VDAART gene workbooks are not used for genomic conclusions. MDIG is restricted to its correction-disclosed qualitative conclusion. GDM2026 was excluded after full assessment for material reporting conflicts. Shared trial reports are not counted as independent confirmations.

Maternal gestational vitamin D supplementation and offspring bone health (MAVIDOS): a multicentre, double-blind, randomised placebo-controlled trial. Cooper C et al. Lancet Diabetes & Endocrinology. 2016;4:393 to 402.

Maternal and neonatal outcomes of vitamin D supplementation in hypovitaminosis D pregnancy: preg-D randomized trial. Chakhtoura M et al. Journal of Clinical Endocrinology & Metabolism. 2026. Supplement DOI 10.6084/m9.figshare.32314851.v1.

Early pregnancy vitamin D status and risk of preeclampsia. Mirzakhani H et al. Journal of Clinical Investigation. 2016;126:4702 to 4715.

Preconception vitamin D status and subsequent risk of preeclampsia: A secondary cohort analysis from the EAGeR trial. Alkhalaf ZM et al. Pregnancy Hypertension. 2026;43:101417.

Vitamin D Supplementation during Pregnancy: Double Blind, Randomized Clinical Trial of Safety and Effectiveness. Hollis BW et al. Journal of Bone and Mineral Research. 2011;26:2341 to 2357. Correction: Erratum: Vitamin D supplementation during pregnancy: Double-blind, randomized clinical trial of safety and effectiveness. 2011.

Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Roth DE et al. New England Journal of Medicine. 2018;379:535 to 546. 2021 correction: Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Same trial, not a separate study.

The role of vitamin D in pregnancy outcomes in gestational diabetes: a randomized controlled trial. Liu Y et al. Endocrine Connections. 2026;15:e250932. Assessed but excluded from reliable effect estimates for the reasons described above.

Related reading

Original sources

Source

Preconception vitamin D status and subsequent risk of preeclampsia: A secondary cohort analysis from the EAGeR trial

A secondary analysis of the EAGeR Trial (2006 to 2012), which was a multisite, prospective, double-blind, block-randomized, placebo-controlled clinical trial of women with regular menstrual cycles and 1 to 2 prior pregnancy losses.

Publication details

Preconception vitamin D status and subsequent risk of preeclampsia: A secondary cohort analysis from the EAGeR trial. Alkhalaf ZM et al. Pregnancy Hypertension. 2026;43:101417.

Read more →
Source

Vitamin D Supplementation during Pregnancy: Double Blind, Randomized Clinical Trial of Safety and Effectiveness

Published in Hollis BW et al. Journal of Bone and Mineral Research (2011).

Publication details

Vitamin D Supplementation during Pregnancy: Double Blind, Randomized Clinical Trial of Safety and Effectiveness. Hollis BW et al. Journal of Bone and Mineral Research. 2011;26:2341 to 2357. Correction: Erratum: Vitamin D supplementation during pregnancy: Double-blind, randomized clinical trial of safety and effectiveness. 2011.

Read more →
Source

Erratum: Vitamin D supplementation during pregnancy: Double-blind, randomized clinical trial of safety and effectiveness

Vitamin D Supplementation during Pregnancy: Double Blind, Randomized Clinical Trial of Safety and Effectiveness.

Publication details

Vitamin D Supplementation during Pregnancy: Double Blind, Randomized Clinical Trial of Safety and Effectiveness. Hollis BW et al. Journal of Bone and Mineral Research. 2011;26:2341 to 2357. Correction: Erratum: Vitamin D supplementation during pregnancy: Double-blind, randomized clinical trial of safety and effectiveness. 2011.

Read more →
Source

Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth

Published in Roth DE et al. New England Journal of Medicine (2018).

Publication details

Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Roth DE et al. New England Journal of Medicine. 2018;379:535 to 546. 2021 correction: Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Same trial, not a separate study.

Read more →
Source

Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth

Published in Roth DE et al. New England Journal of Medicine (2018).

Publication details

Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Roth DE et al. New England Journal of Medicine. 2018;379:535 to 546. 2021 correction: Vitamin D Supplementation in Pregnancy and Lactation and Infant Growth. Same trial, not a separate study.

Read more →