Objective
We undertook a systematic review to assess normative levels of vitamin D in early pregnancy and association with subsequent pregnancy outcomes.
Study Design
Medline and Embase databases and reference lists were searched. Inclusion criteria were pregnant populations, blood sample taken during the first trimester, and serum hydroxyvitamin D levels assessed.
Results
Eighteen studies reported vitamin D levels in first trimester (n = 11-3730), and 5 examined pregnancy outcomes. Mean vitamin D concentrations differed when stratified by ethnicity: white (mean [SD]: 29.4 [11.7] to 73.1 [27.1] nmol/L) and nonwhite (15.2 [12.1] to 43 [12] nmol/L). Most studies used general population cut points to define deficiency and found a large proportion of women deficient. Two articles examined risk of preeclampsia and reported differing findings, whereas 2 of 3 found low levels associated with increased risk of small-for-gestational age births.
Conclusion
There is no clear definition of vitamin D deficiency in pregnancy and insufficient evidence to suggest low vitamin D levels in early pregnancy are associated with adverse pregnancy outcomes.
There is increasing interest in the association between low vitamin D levels and adverse pregnancy outcomes, including preeclampsia, fetal growth restriction, preterm birth, and stillbirths. Most studies investigating vitamin D in pregnancy have measured and assessed levels late in pregnancy or at birth, thus at a stage too late to intervene or change the outcome of the pregnancy and often in high-risk populations more susceptible to vitamin D deficiency, such as ethnic or black women.
Dermal skin absorption of ultraviolet B (UVB) rays from sunlight is the major source of vitamin D, and this is affected by season, latitude, time of day, use of sunscreen, skin pigmentation, and clothing. It is also obtained in lesser amounts through diet, particularly oily fish. Although low levels are associated with adverse health outcomes, various definitions of normative and deficient levels of vitamin D are reported for the general population. There is currently no information on normative levels of vitamin D in pregnant populations.
Measurement of vitamin D in the first trimester provides an ideal opportunity for early identification of pregnancies at risk with ample time for monitoring, follow-up, and early prophylaxis that may potentially reduce the risk of adverse maternal and infant outcomes. The aim of this study was to undertake a systematic review of primary studies to assess recent normative levels of vitamin D in early pregnancy and determine any association between first-trimester vitamin D levels with subsequent adverse pregnancy outcomes.
Materials and Methods
An electronic search of published articles was carried out using Medline (1950 to August 2010) and Embase (1966 to August 2010), with the search terms vitamin D and pregnancy, limiting the search to studies of humans. Search terms were exploded and subheadings included, and terms subsequently combined with no limitations placed on the searches. The results were reviewed using the title and abstract, and full text articles of potentially relevant studies were obtained. Reference lists from relevant reviews and selected articles were also checked.
The selection criteria for review included those studies conducted in populations of pregnant women in which a blood sample was taken during the first trimester of pregnancy and serum hydroxyvitamin D (25[OH]D) levels assessed. Articles were excluded if they were a review, case report, comment, letter, conference report, or guideline; vitamin D was measured using body fluids other than blood; the mean gestational age of the sample at the time of serum collection was greater than 14 weeks; or the study was limited to women with specific medical conditions.
For the review of recent normative levels, studies were limited to high-income countries to reduce major discrepancies in nutritional status, maternal medical and pregnancy conditions, and maternity care. Selected articles were further limited if the vitamin D that had been measured was the active metabolite (1,25 dihydroxyvitamin D [1,25 (OH) 2 D]) because previous studies have deemed this to be an unreliable biomarker.
Information on the study design, geographical location, study participants, entry criteria, and vitamin D levels was obtained from relevant articles using a data extraction spread sheet. Two reviewers independently extracted the data from each article and findings were compared. Any discrepancies were resolved by consensus. Several original authors were also contacted for missing information (such as overall or stratified mean vitamin D levels).
We originally planned to pool the results of studies of normative values of vitamin D and the risk of adverse outcomes. However, because of inconsistent reporting of results across studies, we were unable to combine findings or undertake any metaanalyses. Thus, general findings from studies are summarized and, where possible, results using similar methods are collated and presented in summary figures. Ethics approval for this systematic review was not required.
Results
The electronic search resulted in 1643 articles, and an additional 5 articles were identified from the reference list search. After assessing titles and abstracts for the 1643 citations, 1488 articles were excluded ( Figure 1 ). Examples of excluded studies included those assessing vitamin D levels in second and third trimester and those assessing nonrelevant exposures such as other metabolites of vitamin D (1,25[OH] 2 D).
Outcomes not relevant to this review included neonatal vitamin D, calcium levels, or rickets among infants. Nonblood measures of vitamin D using biological samples such as amniotic fluid or urine were also excluded. Of the remaining 155 articles reviewed in full, a further 137 were excluded because they did not meet our prespecified criteria. Eighteen relevant articles were retained because these provided details of vitamin D levels among various groups of pregnant women in the first trimester; 5 of these articles also assessed pregnancy outcomes.
The Table presents the characteristics of the 18 studies included in the systematic review, including the 5 studies that reported pregnancy outcomes. These studies were conducted in various countries at different latitudes and included women from a range of ethnic backgrounds. All studies included prospective cohorts of women with a singleton pregnancy and without medical conditions, opportunistically recruited while attending antenatal clinics in early pregnancy.
