N U TR IT ION RE S E ARCH XX ( 2 0 15 ) X XX– X XX
Available online at www.sciencedirect.com
ScienceDirect www.nrjournal.com
Review
Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review and meta-analysis Na Yang a , Linlin Wang a , Zhixia Li b , Sen Chen a , Nan Li a , Rongwei Ye a,⁎ a
Institute of Reproductive and Child Health/Ministry of Health Key Laboratory of Reproductive Health, School of Public Health, Peking University, Beijing, PR China b Department of Epidemiology and Biostatistics, School of Public Health, Peking University, Beijing, PR China
ARTI CLE I NFO
A BS TRACT
Article history:
We conducted a meta-analysis to review the effects of vitamin D supplementation during
Received 5 December 2014
pregnancy on neonatal 25-hydroxyvitamin D (25(OH)D) and calcium concentrations.
Revised 8 April 2015
Randomized controlled trials that supplemented subjects with vitamin D2 or D3 during
Accepted 14 April 2015
pregnancy and reported cord blood 25(OH)D or calcium concentrations were included. A random-effect model was used to pool the data. Subgroup analyses were performed to explore
Keywords:
the sources of heterogeneity. We searched PubMed, Web of Science, and Cochrane Library for
Vitamin D
relevant publications. Among 1768 publications identified by our search strategy, 13 studies met
Calcium
our inclusion criteria. Cord blood 25(OH)D concentration was significantly increased by maternal
Pregnancy
vitamin D supplementation (mean difference, 22.48 nmol/L; 95% confidence interval, 15.90-29.06
Neonate
nmol/L) with high heterogeneity (I2 = 98.8%, P < .0001). No effects on cord blood calcium
Meta-analysis
concentration was reported (mean difference, 0.05 mmol/L; 95% confidence interval, −0.04-0.13 mmol/L). Supplementation regimens and the different control groups may be the major sources of heterogeneity. Vitamin D supplementation during pregnancy can improve cord blood 25(OH)D concentration in women with low 25(OH)D concentration, but does not affect cord blood calcium concentration. Future researches are needed to evaluate the effect of maternal vitamin D supplementation in women with a normal 25(OH)D concentration and explore the combined effects of vitamin D, calcium, and multivitamins. © 2015 Elsevier Inc. All rights reserved.
1.
Introduction
Vitamin D is a fat-soluble vitamin that can be synthesized in the skin when exposed to ultraviolet or obtained from dietary
intake. Then it is transported to the liver and hydroxylated to form 25-hydroxyvitamin D (25(OH)D), which is the main circulating form of vitamin D. In the kidney, 1,25-dihydroxyvitamin D (1,25(OH)2D), the active form of vitamin D, is formed.
Abbreviations: 25(OH)D, 25-hydroxyvitamin D; 1,25(OH)2D, 1,25-dihydroxyvitamin D; CI, confidence interval; MD, mean difference. ⁎ Corresponding author at: Institute of Reproductive and Child Health, Peking University Health Science Center, Xueyuan Rd 38#, Haidian District, Beijing 100191, PR China. Tel.: +86 010 82801172; fax: +86 010 82801141. E-mail address: [email protected] (R. Ye). http://dx.doi.org/10.1016/j.nutres.2015.04.010 0271-5317/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
2
N UTR IT ION RE S EA RCH XX ( 2 01 5 ) X XX– X XX
1,25(OH)2D plays a vital role in the maintenance of a normal serum calcium concentration by increasing calcium absorption from the gut and inhibiting parathyroid hormone secretion from the parathyroid glands. During pregnancy, about 30 g of calcium that is needed for fetal bone mineral accretion is transferred from the mothers to their fetus [1]. The pregnant women adapt to this requirement mainly by increasing intestinal calcium absorption [2]. 25(OH)D crossing the placenta during pregnancy furnishes the vitamin D requirement of the fetus, and the circulating serum 25(OH)D concentrations at birth are determined by maternal 25(OH)D concentrations [3,4]. After birth, intestinal calcium absorption is dependent on 1,25(OH)2D. Although recent studies have found that human breast milk can transfer vitamin D more efficiently than we expected, the breastfed newborns cannot obtain enough vitamin D from their mothers because the present recommendation for nursing mothers is too low to improve the 25(OH)D concentrations in the milk [5,6]. Thus, the stored vitamin D acquired during intrauterine life is of great importance to the breastfed newborns [7]. 25(OH)D concentrations are used to assess vitamin D status. The optimal serum concentrations of 25(OH)D are under debate. In 2011, the Institute of Medicine, based on assuring bone health, determined that serum 25(OH)D concentration at least 50 nmol/L meets the requirements of at least 97.5% of the population younger than 1 year [8]. For pregnant women, the estimated average requirement and recommended dietary allowance were defined as 400 and 600 IU/d, respectively, and the tolerable upper intake limit was 4000 IU/d [8]. Global reports found that the prevalence of vitamin D deficiency during pregnancy ranged from 18% to 80% [3,9–11]; the rate was especially high in Asia [12,13]. Maternal vitamin D status influences fetal growth and birth outcomes. Vitamin D deficiency rickets in early infancy is prevalent in infants whose mothers had poor vitamin D stores [14]. The association between low maternal 25(OH)D concentrations and maternal or neonatal outcomes was mixed. Several observational studies have reported that low concentrations of 25(OH)D or calcium during pregnancy may result in preterm delivery [15], low birth weight, and a higher risk of small-forgestational age [16,17]. However, some other studies have found no significant association between maternal vitamin D status and neonatal birth weight [18], head circumference [19], bone mass [20], and preterm birth [21]. Since Brooke et al [22] conducted the first controlled trial of vitamin D supplementation in Asian pregnant women in 1980, trials have been conducted to evaluate the effects of maternal vitamin D supplementation on pregnancy outcomes [23], fetal growth, or even childhood health [22,24]. Most systematic reviews of the effects of vitamin D supplementation on pregnant women and their offsprings were based on observational studies, and the results were inconsistent [25–28]. Nonetheless, the effects of vitamin D supplementation during pregnancy on neonatal 25(OH)D and calcium concentrations are not well established, and these surrogate end points can be used to detect biologically significant effects. We aimed to address these questions by systematically reviewing all randomized controlled trials of cholecalciferol (vitamin D3) or ergocalciferol (vitamin D2) supplementation that reported neonatal 25(OH)D and calcium concentrations.
2.
Methods and materials
2.1.
Search strategy
We reported this systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline and Cochrane guideline [29,30]. We searched PubMed, Web of Science, and Cochrane Library from inception to April 3, 2014, using both Mesh terms and text word. The terms included vitamin D: “vitamin D,” or “cholecalciferol,” or “ergocalciferol,” or “multivitamin,” and pregnancy: “infant,” or “newborn,” or “pregnan*,” or “prenatal,” or “infan*,” or “neona*,” or “fetus,” or “fetal,” together with trial: “randomized study,” or “randomized trial,” or “controlled clinical trial,” or “randomized controlled trial,” or “multicenter study,” or “meta-analysis,” or “review.” Articles were restricted to English and humans. In addition, we searched the databases in March 15, 2015, for newly published relevant articles.
2.2.
Study selection and data extraction
Our primary outcomes were cord blood 25(OH)D and calcium concentrations at delivery. We included randomized trials that explored the efficacy of vitamin D supplementation during pregnancy and reported the outcomes of interest, no matter whether it was the primary outcome of the original studies. Only published articles were included in our analysis. The intervention must be a preparation of vitamin D2 or D3, rather than a supplementation from food supplement. If there were other interventions (eg, calcium and multivitamin), they would have to be the same in all groups. Two authors searched the databases independently with the same terms. The search results were combined and duplicates were removed. Eligibility confirmation and data extraction were performed independently by the same two authors. A standardized spreadsheet was used to extract data from included studies. Extracted data included authors, locations of the study, population characteristics, sample size, doses and duration of vitamin D supplementation, initiation of supplementation, co-interventions, maternal vitamin D status at baseline and at delivery, and cord blood 25(OH)D and calcium concentrations at delivery.
2.3.
Assessments of methodological quality
We assessed the methodological quality of trials according to the Cochrane handbook. The features of interest in a standard “risk of bias” table of a Cochrane review are sequence generation, allocation sequence concealment, blinding, incomplete outcome data, selective outcome reporting, and other potential sources of bias [30].
2.4.
Data extraction and synthesis
Both cord blood 25(OH)D and calcium concentrations at delivery were continuous variables, and the outcome was expressed as the mean difference (MD) and 95% confidence
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
3
N U TR IT ION RE S E ARCH XX ( 2 0 15 ) X XX– X XX
interval (CI). The inverse variance method was used to calculate the weighting of the studies in the meta-analysis. We assessed the heterogeneity in the results of individual studies by using the Cochrane's Q test and the I2 statistic. I2 < 50% was considered homogeneous, and a fixed-effect model was used to estimate the pooled effect; otherwise, substantial heterogeneity occurred and a random-effect model was used [30]. For the studies that had two vitamin D supplementation groups and one placebo group, we processed each dose group and the placebo group as one independent trial in the meta-analysis. Also, for the studies that had three different doses of vitamin D supplementation, we selected the minimum dose of vitamin D as the control group, and two trials were obtained from the study by comparing the other two dose groups with the control group, respectively. We prespecified subgroup analyses to explore the potential sources of heterogeneity, such as doses, the timing of initiation of supplementation, or supplementation regimens. Egger's test was used to test publication bias [31]. Sensitivity analysis was performed by excluding several low-quality trials. All analyses were performed using open-source R software, version 2.13.0.
3.
Results
3.1.
Literature search
We identified 1768 publications by our search strategy, of which 101 publications appeared to be potentially relevant after the titles and abstracts were screened. We assessed the full text of the 101 publications, and 87 publications were excluded because they were unrelated publications (32), reviews (35), observational studies (6), study protocols (2), article written in French (1), different articles based on same trials (3), trials that studied the combined effects of calcium and vitamin D (3) or the effect of food supplement (1), or trials that had neither data for cord blood 25(OH)D nor calcium concentrations (4). In addition, another 2 publications were identified by searching the databases in March 15, 2015 [32,33]. We found that the study by Harrington et al [32], which was newly identified from the databases, was based on the same trial with the study by Roth et al [34]. Because the study by Harrington et al provided both the data regarding to cord blood 25(OH)D and calcium concentrations, we included this study and excluded the study by Roth et al from our meta-analysis. In total, 15 eligible studies were identified. However, two studies were excluded from our meta-analysis for the following reasons: the 25(OH)D levels reported in the study by Brook et al [22] were extraordinarily high, which were far off from other studies performed providing a similar amount of vitamin D supplement. As for the study conducted by Marya et al [21] in 1988, we could not find the full text, although we contacted the authors to request for it. Thus, we included 13 studies in our meta-analysis. The details are shown in Fig. 1.
3.2.
Characteristics of included studies
Of the 13 studies, two studies compared two different doses of vitamin D groups and the placebo group [35,36], and another three studies compared three different doses of vitamin D [37–39], so a total of 18 randomized trials were listed in Table.
Publications identified through PubMed, Web of science and Cochrane (n = 1785)
Publications after duplicates removed (n = 1768) Publications potentially relevant by the titles and abstracts (n = 101) Excluded (n = 67) Not relevant (n = 32) Reviews ( n = 35) Publications retrieved for detailed evaluation (n = 34) Excluded (n = 20) Observational studies (n = 6) Study protocols (n = 2) Publication written in French (n = 1) Different articles based on the same trial (n = 3) Studies of the combined effects of calcium or multivitamins and vitamin D (n = 4) No data of interest (n = 4) Eligible studies identified in the first search (n = 14)
Eligible studies identified in the second search (n = 2) Excluded (n = 3) Studies based on the same trial (n = 1) Full-text not available (n = 1) Results are not credible (n = 1)
Publications included in the meta-analysis (n = 13)
Fig. 1 – Flowchart of study selection.
Cord blood 25(OH)D concentration was available in 14 randomized trials, whereas 10 trials reported cord blood calcium concentration. Among all these studies, five studies were placebo controlled [32,35,36,40,41], two studies did not give any vitamin D supplements to the control group [33,42], three studies compared two different doses of vitamin D [43–45], and another three studies compared three different doses of vitamin D [37–39]. For seven studies, the supplementation of vitamin D was initiated during the third trimester [32,33,35,41– 43,45], and for the other studies, it was initiated during the second trimester [36–40,44]. Six studies used multiple nutrients for each group, such as a standard multivitamin, elemental Ca, iron, or folic acid, and only vitamin D nutrient was different among the groups [32,33,36,38,43,45]. Supplementation regimens were various among the studies. In six studies, pills were taken daily [33,38–41,44], whereas in three studies, pill-taking was conducted weekly or monthly [32,37,43]. In another four studies, the participants in the intervention group were given a single dose of vitamin D at some point during the pregnancy [35,36,42,45].
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
4
N UTR IT ION RE S EA RCH XX ( 2 01 5 ) X XX– X XX
Table – Characteristics of randomized controlled trials assessing the effects of maternal vitamin D supplementation on neonatal 25(OH)D or calcium concentrations Trial
N
Country
Baseline 25(OH)D (nmol/L)
25(OH)D at Initiation of delivery supplementation (nmol/L)
Mean SD
Mean SD
Cockburn et al [40] Marya et al [42]
714 England
NR
NR
42.82
95
India (Hindus)
NR
NR
NR
Mallet et al [35]
56
France (white women)
NR
NR
25.3 a
Mallet et al [35] Delvin et al [41] Hollis et al [38]
50
France (white women) 30 France (white women) 333 United States (67% AA and H women) 338 United States (67% AA and H women) 91 India
NR
NR
26.0 a
NR
NR
64.9 a
58.3
22.3 98.3 a
58.2
21.8 111.0 a 40.4 12-16 wk gestation
31.7 (14.0-57.) b
26.2 a (17.7-57.7) b
From the second trimester
Kalra et al [36]
92
32.0 (14.5-45.7) b
58.7 a (38.4-89.4) b
From the second trimester
Dawodu et al [39] Dawodu et al [39] Roth et al [45]
129 United Arab Emirates (Arab women) 127 United Arab Emirates (Arab women) 28 Bangladesh
20.5
11.9 65.4 a
3.2
12-16 wk gestation
19.6
7.7
89.6 a
4.0
12-16 wk gestation
98 a (89.0-109.0) b
27-30 wk gestation
Shakiba and Iranmanesh [37] Shakiba and Iranmanesh [37]
34
Iran (25(OH)D level
Available online at www.sciencedirect.com
ScienceDirect www.nrjournal.com
Review
Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review and meta-analysis Na Yang a , Linlin Wang a , Zhixia Li b , Sen Chen a , Nan Li a , Rongwei Ye a,⁎ a
Institute of Reproductive and Child Health/Ministry of Health Key Laboratory of Reproductive Health, School of Public Health, Peking University, Beijing, PR China b Department of Epidemiology and Biostatistics, School of Public Health, Peking University, Beijing, PR China
ARTI CLE I NFO
A BS TRACT
Article history:
We conducted a meta-analysis to review the effects of vitamin D supplementation during
Received 5 December 2014
pregnancy on neonatal 25-hydroxyvitamin D (25(OH)D) and calcium concentrations.
Revised 8 April 2015
Randomized controlled trials that supplemented subjects with vitamin D2 or D3 during
Accepted 14 April 2015
pregnancy and reported cord blood 25(OH)D or calcium concentrations were included. A random-effect model was used to pool the data. Subgroup analyses were performed to explore
Keywords:
the sources of heterogeneity. We searched PubMed, Web of Science, and Cochrane Library for
Vitamin D
relevant publications. Among 1768 publications identified by our search strategy, 13 studies met
Calcium
our inclusion criteria. Cord blood 25(OH)D concentration was significantly increased by maternal
Pregnancy
vitamin D supplementation (mean difference, 22.48 nmol/L; 95% confidence interval, 15.90-29.06
Neonate
nmol/L) with high heterogeneity (I2 = 98.8%, P < .0001). No effects on cord blood calcium
Meta-analysis
concentration was reported (mean difference, 0.05 mmol/L; 95% confidence interval, −0.04-0.13 mmol/L). Supplementation regimens and the different control groups may be the major sources of heterogeneity. Vitamin D supplementation during pregnancy can improve cord blood 25(OH)D concentration in women with low 25(OH)D concentration, but does not affect cord blood calcium concentration. Future researches are needed to evaluate the effect of maternal vitamin D supplementation in women with a normal 25(OH)D concentration and explore the combined effects of vitamin D, calcium, and multivitamins. © 2015 Elsevier Inc. All rights reserved.
1.
Introduction
Vitamin D is a fat-soluble vitamin that can be synthesized in the skin when exposed to ultraviolet or obtained from dietary
intake. Then it is transported to the liver and hydroxylated to form 25-hydroxyvitamin D (25(OH)D), which is the main circulating form of vitamin D. In the kidney, 1,25-dihydroxyvitamin D (1,25(OH)2D), the active form of vitamin D, is formed.
Abbreviations: 25(OH)D, 25-hydroxyvitamin D; 1,25(OH)2D, 1,25-dihydroxyvitamin D; CI, confidence interval; MD, mean difference. ⁎ Corresponding author at: Institute of Reproductive and Child Health, Peking University Health Science Center, Xueyuan Rd 38#, Haidian District, Beijing 100191, PR China. Tel.: +86 010 82801172; fax: +86 010 82801141. E-mail address: [email protected] (R. Ye). http://dx.doi.org/10.1016/j.nutres.2015.04.010 0271-5317/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
2
N UTR IT ION RE S EA RCH XX ( 2 01 5 ) X XX– X XX
1,25(OH)2D plays a vital role in the maintenance of a normal serum calcium concentration by increasing calcium absorption from the gut and inhibiting parathyroid hormone secretion from the parathyroid glands. During pregnancy, about 30 g of calcium that is needed for fetal bone mineral accretion is transferred from the mothers to their fetus [1]. The pregnant women adapt to this requirement mainly by increasing intestinal calcium absorption [2]. 25(OH)D crossing the placenta during pregnancy furnishes the vitamin D requirement of the fetus, and the circulating serum 25(OH)D concentrations at birth are determined by maternal 25(OH)D concentrations [3,4]. After birth, intestinal calcium absorption is dependent on 1,25(OH)2D. Although recent studies have found that human breast milk can transfer vitamin D more efficiently than we expected, the breastfed newborns cannot obtain enough vitamin D from their mothers because the present recommendation for nursing mothers is too low to improve the 25(OH)D concentrations in the milk [5,6]. Thus, the stored vitamin D acquired during intrauterine life is of great importance to the breastfed newborns [7]. 25(OH)D concentrations are used to assess vitamin D status. The optimal serum concentrations of 25(OH)D are under debate. In 2011, the Institute of Medicine, based on assuring bone health, determined that serum 25(OH)D concentration at least 50 nmol/L meets the requirements of at least 97.5% of the population younger than 1 year [8]. For pregnant women, the estimated average requirement and recommended dietary allowance were defined as 400 and 600 IU/d, respectively, and the tolerable upper intake limit was 4000 IU/d [8]. Global reports found that the prevalence of vitamin D deficiency during pregnancy ranged from 18% to 80% [3,9–11]; the rate was especially high in Asia [12,13]. Maternal vitamin D status influences fetal growth and birth outcomes. Vitamin D deficiency rickets in early infancy is prevalent in infants whose mothers had poor vitamin D stores [14]. The association between low maternal 25(OH)D concentrations and maternal or neonatal outcomes was mixed. Several observational studies have reported that low concentrations of 25(OH)D or calcium during pregnancy may result in preterm delivery [15], low birth weight, and a higher risk of small-forgestational age [16,17]. However, some other studies have found no significant association between maternal vitamin D status and neonatal birth weight [18], head circumference [19], bone mass [20], and preterm birth [21]. Since Brooke et al [22] conducted the first controlled trial of vitamin D supplementation in Asian pregnant women in 1980, trials have been conducted to evaluate the effects of maternal vitamin D supplementation on pregnancy outcomes [23], fetal growth, or even childhood health [22,24]. Most systematic reviews of the effects of vitamin D supplementation on pregnant women and their offsprings were based on observational studies, and the results were inconsistent [25–28]. Nonetheless, the effects of vitamin D supplementation during pregnancy on neonatal 25(OH)D and calcium concentrations are not well established, and these surrogate end points can be used to detect biologically significant effects. We aimed to address these questions by systematically reviewing all randomized controlled trials of cholecalciferol (vitamin D3) or ergocalciferol (vitamin D2) supplementation that reported neonatal 25(OH)D and calcium concentrations.
2.
Methods and materials
2.1.
Search strategy
We reported this systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline and Cochrane guideline [29,30]. We searched PubMed, Web of Science, and Cochrane Library from inception to April 3, 2014, using both Mesh terms and text word. The terms included vitamin D: “vitamin D,” or “cholecalciferol,” or “ergocalciferol,” or “multivitamin,” and pregnancy: “infant,” or “newborn,” or “pregnan*,” or “prenatal,” or “infan*,” or “neona*,” or “fetus,” or “fetal,” together with trial: “randomized study,” or “randomized trial,” or “controlled clinical trial,” or “randomized controlled trial,” or “multicenter study,” or “meta-analysis,” or “review.” Articles were restricted to English and humans. In addition, we searched the databases in March 15, 2015, for newly published relevant articles.
2.2.
Study selection and data extraction
Our primary outcomes were cord blood 25(OH)D and calcium concentrations at delivery. We included randomized trials that explored the efficacy of vitamin D supplementation during pregnancy and reported the outcomes of interest, no matter whether it was the primary outcome of the original studies. Only published articles were included in our analysis. The intervention must be a preparation of vitamin D2 or D3, rather than a supplementation from food supplement. If there were other interventions (eg, calcium and multivitamin), they would have to be the same in all groups. Two authors searched the databases independently with the same terms. The search results were combined and duplicates were removed. Eligibility confirmation and data extraction were performed independently by the same two authors. A standardized spreadsheet was used to extract data from included studies. Extracted data included authors, locations of the study, population characteristics, sample size, doses and duration of vitamin D supplementation, initiation of supplementation, co-interventions, maternal vitamin D status at baseline and at delivery, and cord blood 25(OH)D and calcium concentrations at delivery.
2.3.
Assessments of methodological quality
We assessed the methodological quality of trials according to the Cochrane handbook. The features of interest in a standard “risk of bias” table of a Cochrane review are sequence generation, allocation sequence concealment, blinding, incomplete outcome data, selective outcome reporting, and other potential sources of bias [30].
2.4.
Data extraction and synthesis
Both cord blood 25(OH)D and calcium concentrations at delivery were continuous variables, and the outcome was expressed as the mean difference (MD) and 95% confidence
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
3
N U TR IT ION RE S E ARCH XX ( 2 0 15 ) X XX– X XX
interval (CI). The inverse variance method was used to calculate the weighting of the studies in the meta-analysis. We assessed the heterogeneity in the results of individual studies by using the Cochrane's Q test and the I2 statistic. I2 < 50% was considered homogeneous, and a fixed-effect model was used to estimate the pooled effect; otherwise, substantial heterogeneity occurred and a random-effect model was used [30]. For the studies that had two vitamin D supplementation groups and one placebo group, we processed each dose group and the placebo group as one independent trial in the meta-analysis. Also, for the studies that had three different doses of vitamin D supplementation, we selected the minimum dose of vitamin D as the control group, and two trials were obtained from the study by comparing the other two dose groups with the control group, respectively. We prespecified subgroup analyses to explore the potential sources of heterogeneity, such as doses, the timing of initiation of supplementation, or supplementation regimens. Egger's test was used to test publication bias [31]. Sensitivity analysis was performed by excluding several low-quality trials. All analyses were performed using open-source R software, version 2.13.0.
3.
Results
3.1.
Literature search
We identified 1768 publications by our search strategy, of which 101 publications appeared to be potentially relevant after the titles and abstracts were screened. We assessed the full text of the 101 publications, and 87 publications were excluded because they were unrelated publications (32), reviews (35), observational studies (6), study protocols (2), article written in French (1), different articles based on same trials (3), trials that studied the combined effects of calcium and vitamin D (3) or the effect of food supplement (1), or trials that had neither data for cord blood 25(OH)D nor calcium concentrations (4). In addition, another 2 publications were identified by searching the databases in March 15, 2015 [32,33]. We found that the study by Harrington et al [32], which was newly identified from the databases, was based on the same trial with the study by Roth et al [34]. Because the study by Harrington et al provided both the data regarding to cord blood 25(OH)D and calcium concentrations, we included this study and excluded the study by Roth et al from our meta-analysis. In total, 15 eligible studies were identified. However, two studies were excluded from our meta-analysis for the following reasons: the 25(OH)D levels reported in the study by Brook et al [22] were extraordinarily high, which were far off from other studies performed providing a similar amount of vitamin D supplement. As for the study conducted by Marya et al [21] in 1988, we could not find the full text, although we contacted the authors to request for it. Thus, we included 13 studies in our meta-analysis. The details are shown in Fig. 1.
3.2.
Characteristics of included studies
Of the 13 studies, two studies compared two different doses of vitamin D groups and the placebo group [35,36], and another three studies compared three different doses of vitamin D [37–39], so a total of 18 randomized trials were listed in Table.
Publications identified through PubMed, Web of science and Cochrane (n = 1785)
Publications after duplicates removed (n = 1768) Publications potentially relevant by the titles and abstracts (n = 101) Excluded (n = 67) Not relevant (n = 32) Reviews ( n = 35) Publications retrieved for detailed evaluation (n = 34) Excluded (n = 20) Observational studies (n = 6) Study protocols (n = 2) Publication written in French (n = 1) Different articles based on the same trial (n = 3) Studies of the combined effects of calcium or multivitamins and vitamin D (n = 4) No data of interest (n = 4) Eligible studies identified in the first search (n = 14)
Eligible studies identified in the second search (n = 2) Excluded (n = 3) Studies based on the same trial (n = 1) Full-text not available (n = 1) Results are not credible (n = 1)
Publications included in the meta-analysis (n = 13)
Fig. 1 – Flowchart of study selection.
Cord blood 25(OH)D concentration was available in 14 randomized trials, whereas 10 trials reported cord blood calcium concentration. Among all these studies, five studies were placebo controlled [32,35,36,40,41], two studies did not give any vitamin D supplements to the control group [33,42], three studies compared two different doses of vitamin D [43–45], and another three studies compared three different doses of vitamin D [37–39]. For seven studies, the supplementation of vitamin D was initiated during the third trimester [32,33,35,41– 43,45], and for the other studies, it was initiated during the second trimester [36–40,44]. Six studies used multiple nutrients for each group, such as a standard multivitamin, elemental Ca, iron, or folic acid, and only vitamin D nutrient was different among the groups [32,33,36,38,43,45]. Supplementation regimens were various among the studies. In six studies, pills were taken daily [33,38–41,44], whereas in three studies, pill-taking was conducted weekly or monthly [32,37,43]. In another four studies, the participants in the intervention group were given a single dose of vitamin D at some point during the pregnancy [35,36,42,45].
Please cite this article as: Yang N, et al, Effects of vitamin D supplementation during pregnancy on neonatal vitamin D and calcium concentrations: a systematic review..., Nutr Res (2015), http://dx.doi.org/10.1016/j.nutres.2015.04.010
4
N UTR IT ION RE S EA RCH XX ( 2 01 5 ) X XX– X XX
Table – Characteristics of randomized controlled trials assessing the effects of maternal vitamin D supplementation on neonatal 25(OH)D or calcium concentrations Trial
N
Country
Baseline 25(OH)D (nmol/L)
25(OH)D at Initiation of delivery supplementation (nmol/L)
Mean SD
Mean SD
Cockburn et al [40] Marya et al [42]
714 England
NR
NR
42.82
95
India (Hindus)
NR
NR
NR
Mallet et al [35]
56
France (white women)
NR
NR
25.3 a
Mallet et al [35] Delvin et al [41] Hollis et al [38]
50
France (white women) 30 France (white women) 333 United States (67% AA and H women) 338 United States (67% AA and H women) 91 India
NR
NR
26.0 a
NR
NR
64.9 a
58.3
22.3 98.3 a
58.2
21.8 111.0 a 40.4 12-16 wk gestation
31.7 (14.0-57.) b
26.2 a (17.7-57.7) b
From the second trimester
Kalra et al [36]
92
32.0 (14.5-45.7) b
58.7 a (38.4-89.4) b
From the second trimester
Dawodu et al [39] Dawodu et al [39] Roth et al [45]
129 United Arab Emirates (Arab women) 127 United Arab Emirates (Arab women) 28 Bangladesh
20.5
11.9 65.4 a
3.2
12-16 wk gestation
19.6
7.7
89.6 a
4.0
12-16 wk gestation
98 a (89.0-109.0) b
27-30 wk gestation
Shakiba and Iranmanesh [37] Shakiba and Iranmanesh [37]
34
Iran (25(OH)D level
