ORIGINAL ARTICLE

Feto-maternal vitamin D status and infant whole-bodybone mineral content in the first weeks of lifeDK Dror1, JC King2, DJ Durand2, EB Fung2 and LH Allen1

BACKGROUND/OBJECTIVES: Compromised vitamin D status is common in pregnancy and may have adverse impacts on fetaldevelopment. The purpose of this study was to investigate the association of infant whole-body bone mineral content (WBBMC)at 8–21 days of age with feto-maternal vitamin D status in a multiethnic population in Oakland, California.SUBJECTS/METHODS: This was a cross-sectional study of 120 women and their newborn infants. Maternal and cord bloodwere collected at delivery. WBBMC was measured by dual-energy X-ray absorptiometry in term-born infants 8–21days post birth.RESULTS: No significant association was observed between unadjusted or size-adjusted WBBMC and feto-maternal vitamin Dstatus analyzed continuously or categorically. In multivariate modeling, unadjusted WBBMC was predicted by bone area(Po0.0001), weight-for-age (Po0.0001) and weight-for-length (P¼ 0.0005) Z-scores, but not by feto-maternal vitamin D status.Anthropometric predictors but not vitamin D remained significant in the multivariate model after adjustment of WBBMC for weight,bone area (bone mineral density) or logarithmically derived exponents of the denominators.CONCLUSIONS: Results of the present study do not support an association between feto-maternal vitamin D status and early infantWBBMC, raw or adjusted for inter-individual differences in size, in a multiethnic population in Northern California.

European Journal of Clinical Nutrition advance online publication, 11 July 2012; doi:10.1038/ejcn.2012.79

Keywords: vitamin D; infant; bone mineral content

INTRODUCTIONFetal bone mineralization is determined by placental mineraltransfer and fetal bone turnover, with maternal nutrient status,disease and conditions affecting placental transfer having possibleimpacts on fetal bone.1 Bone mineralization and growth duringfetal development and early infancy may have implications forchildhood growth and development, peak bone mass and laterrisk for osteoporosis.2,3

Vitamin D, which in its biologically active form 1,25(OH)2D actsas a potent genetic regulator, may influence fetal bone accretionby upregulating expression of genes encoding rate-limitingplacental calcium transport proteins (plasma membrane calciumATPase1–4).4 Maternal vitamin D status, especially during the thirdtrimester of gestation, has been associated with neonatal bonegrowth or mineralization in some studies5–7 but not in others.8

In adults, the relationship between serum 25(OH)D and bonemineral density (BMD) varies among Caucasians, African Americansand Hispanics.9 A single study with limited sample size (n¼ 50) haspreviously investigated feto-maternal vitamin D status and infantwhole-body bone mineral content (WBBMC) in the first 2 weeks oflife.10 Oakland, California hosts a racially diverse population, withskin color and genetic variance affecting maternal vitamin D statusand possibly fetal bone mineralization. The purpose of the presentstudy was to investigate the relation between feto-maternal vitaminD status and WBBMC at 8–21 days of age in a multiethnicpopulation in Oakland, CA (381N).

SUBJECTS AND METHODSThe study was approved by the institutional review boards of theUniversity of California, Davis, Children’s Hospital & Research Center,

Oakland and Alta Bates Medical Center, Berkeley. Recruitment took placebetween December 2006 and January 2008. Study subjects were mothersobtaining perinatal care at East Bay Perinatal Medical Associates inOakland, CA and their infants. Women who were 18–45 years of age andcarrying a singleton fetus were eligible for recruitment and infants born atterm (37–42 weeks of gestation) with dual-energy X-ray absorptiometry(DXA) scans between 8 and 21 days were included in analysis. Informedconsent was obtained from all the mothers upon study enrollment duringweeks 34–40 of gestation.

An interviewer-administered questionnaire was used to evaluate intakeand lifestyle factors related to vitamin D status as described previously.11

Medical records were reviewed to ascertain pre- or early-pregnancy weight,length of gestation, pregnancy and delivery events, and infant birth weight.Pre-gestational maternal body mass index (BMI) was calculated fromreported height and pre- or early-pregnancy weight recorded in themedical file.

Maternal venous blood was collected upon admission to the Alta BatesMedical Center Labor and Delivery Unit, and cord blood immediately postdelivery. Blood samples were kept at 4 1C until centrifuged and serumstored at � 80 1C until analysis. Batched samples of serum 25(OH)D wereassayed monthly at ARUP Laboratories (Salt Lake City, UT, USA) using theDiaSorin radioimmunoassay (DiaSorin Inc., Stillwater, MN, USA). An internalstandard that had been assayed in duplicate in the laboratory of Dr BruceHollis (Medical University of South Carolina, Charleston, SC, USA) wasincluded with each batch and results were adjusted accordingly. Medicalrecords were reviewed to ascertain the length of gestation and infant birthweight.

Mother–infant pairs visited the Children’s Hospital & Research CenterOakland’s Clinical and Translational Science Institute (CTSI) ClinicalResearch Center 8–21 days post birth. The length, weight and headcircumference of the infants were measured by trained research staff usingthe World Health Organization’s standardized protocol.12 Weight wasmeasured to the nearest gram using an infant digital scale (Seca 334, SecaCorp., Hamburg, Germany), length to the nearest 0.1 cm on an infantlength board (Shorr Infant Polylength Measuring Board, Shorr Productions,

1Allen Lab, USDA, ARS Western Human Nutrition Research Center, Davis, CA, USA and 2Children’s Hospital & Research Center, Oakland, CA, USA. Correspondence: Dr DK Dror,Allen Lab, USDA, ARS Western Human Nutrition Research Center, 430 West Health Sciences Drive, Davis, CA 95616, USA.E-mail: daphnadror@gmail.comReceived 22 November 2011; revised 21 May 2012; accepted 7 June 2012

European Journal of Clinical Nutrition (2012), 1–4& 2012 Macmillan Publishers Limited All rights reserved 0954-3007/12

www.nature.com/ejcn

Olney, MD, USA), and head circumference to the nearest 0.1 cm (ShorrProductions).

Bone mineral content, bone area and mass were determined by DXA induplicate using the Hologic Discovery A whole body infant softwarepackage (version 12.6.1, Hologic Inc., Waltham, MA, USA). The instrumentwas calibrated daily using a manufacturer-provided spine phantom. Beforeimaging, each infant was dressed only in a clean study-provided diaperand swaddled in a cotton receiving blanket to restrict movement. Infantswere breast- or formula-fed and pacified immediately before the scan, andthe scan with least movement artifact was selected for inclusion in analysis.A single trained technician evaluated the acceptability of all scans andcompartmentalized the scan into two regions (head and remainder ofbody). As the International Society of Clinical Densitometry (ISCD) has nostandard for DXA measurements in children under the age 5 years, WBBMCwas analyzed raw and adjusted for bone area (BMD) or weight inaccordance with the ISCD standards for pediatrics.13

Size for gestational age at birth was calculated based on fetal growthreference curves with o10th percentile defined as small for gestationalage and 490th percentile defined as large for gestational age.14

Z-scores for infant weight and length were calculated using CDCreference data sets.15 Seasonal boundaries were based on sine curveanalysis of maternal serum 25(OH)D concentrations and were defined as21 January–20 April, 21 April–20 July, 21 July–20 October, and 21 October–20 January as described previously.11

Statistical analysisStatistical analysis was performed using SAS software (version 9.1, SASInstitute Inc., Cary, NC, USA). Data are presented as means±s.d. 25(OH)Dwas analyzed both as a continuous and categorical variable. Normality ofstudy variables was tested using the Shapiro–Wilk statistic. Regressionmodels including the quadratic term of the variable of interest were usedto test for threshold effects. Path analysis was conducted to identifycontributors to the effect of anthropometrics on WBBMC. Differences invariable means between groups were evaluated by analysis of varianceusing Tukey’s test to correct for multiple comparison. Analysis ofcovariance was used to construct multivariate models utilizing thebackwards elimination method.

RESULTSOf 143 infants born at term who underwent DXA scans at 8–21days, 11 were excluded due to poor scan quality and 12 due tomissing maternal or cord serum 25(OH)D. The 120 infants includedin the analysis (Table 1) did not differ significantly from those whounderwent DXA scans but were not included (n¼ 23), or whowere eligible but did not return for the DXA visit (n¼ 32), ingender, maternal age, parity, or maternal or cord vitamin D status.

Bone mineral content of the infant whole body and body lessthe head were highly correlated (r¼ 0.88, Po0.0001), thereforewhole-body variables were used in the remainder of the analyses.Infant WBBMC was not significantly correlated with maternal orcord 25(OH)D (Figure 1), nor did WBBMC differ significantly bymaternal or cord vitamin D status categorized by cutpoints(Table 2) or tertiles (data not shown). Furthermore, no thresholdeffect was found between maternal or cord 25(OH)D and infantWBBMC. Excluding morbidly obese women (n¼ 18), women withgestational diabetes (n¼ 16) or large for gestational age infants(n¼ 13) did not alter the results. Other infant characteristics,including length of gestation, size for gestational age, sex, birthweight and weight or length at the DXA visit, were not associatedwith maternal or cord serum 25(OH)D.

To reduce multicollinearities and control for infant age inmultivariate linear models, weight-for-height and height-for-ageZ-scores were substituted for raw anthropometrics. Infant WBBMCwas significantly predicted by bone area (Po0.0001) and length-for-age (Po0.0001) and weight-for-length (P¼ 0.0005) Z-scores atthe time of the DXA scan, with an overall model R2¼ 0.84. Factorsnot remaining in the final predictive model were maternal andcord 25(OH)D; maternal height, pre-pregnancy BMI or gestationaldiabetes; and infant age, length of gestation, size for gestationalage and feeding practice. In path analysis, infant age and maternalpre-pregnancy BMI were found to mediate the influence ofanthropometrics at the DXA visit on infant WBBMC.

Table 1. Characteristics of maternal and infant study subjects (n¼ 120mother–infant pairs)

Mean±s.d. Range

Infant birth weight (g) 3420±542 2290–4532Length of gestation (weeks) 39.6±1.3 37–42Maternal 25(OH)Da (nmol/l) 75.5±32.3 20.8–172.8Cord 25(OH)D (nmol/l) 44.5±22.1 13.0–157.5Infant age at DXA visit (days) 13.3±3.5 8–21Infant weight at DXA (g) 3655±544 2413–5448Weight-for-age Z-scoreb � 0.3±0.9 � 2.2–2.3Infant length at DXA (cm) 50.9±2.4 43.5–58.2Length-for-age Z-scoreb � 0.4±0.9 � 3.5–1.9Weight-for-length Z-scoreb 0.1±0.7 � 2.3–2.6Infant WBBMC (full body, g) 62.1±12.7 35.4–95.9Infant body BMC (head subtracted, g) 33.5±7.9 17.3–58.5Infant whole-body BMD (g/cm)b 0.20±0.02 0.15–0.25

Abbreviations: BMC, bone mineral content; BMD, bone mineral density;DXA, dual-energy X-ray absorptiometry. aMaternal 25(OH)D drawn at thetime of delivery. bZ-scores calculated per Centers for Disease Control andPrevention (CDC) reference data set.

20

30

40

50

60

70

80

90

100

100

Cord 25(OH)D (nmol/L)

WB

BM

C (

g)

African American Hispanic Caucasian Asian Mixed

20

30

40

50

60

70

80

90

100

180

Maternal 25(OH)D (nmol/L)

WB

BM

C (

g)

0 20 40 60 80 100 120 140 160

0 20 40 60 80

Figure 1. Distribution of WBBMC by maternal (a) and cord (b) serum25(OH)D.

Vitamin D status and neonatal bone mineral contentDK Dror et al

2

European Journal of Clinical Nutrition (2012) 1 – 4 & 2012 Macmillan Publishers Limited

As WBBMC was most strongly influenced by infant size,adjustments to the independent variable were attempted tocreate a size-neutral index and further consider a potentialinfluence of vitamin D status. Simple adjustment of WBBMC forinfant weight (kg) or bone area (as determined by DXA; BMD)over- or under-adjusted for size differences per analysis ofresiduals. To best adjust for inter-individual differences in size asdetermined through logarithmic analysis, WBBMC was divided byweight1.3 or bone area1.4. No significant correlations wereobserved between size-adjusted infant WBBMC and maternal orcord 25(OH)D. In multivariate modeling of size-adjusted infantWBBMC, anthropometric indicators were significant predictorswhereas maternal and cord 25(OH)D were not.

DISCUSSIONIn this cohort of multiethnic mother–infant pairs in Oakland, CA(381N), WBBMC in full-term infants at 8–21 days of age was notrelated to maternal or cord vitamin D status, season of birth,feeding habits or sex but was most strongly influenced by variousindices of infant size (bone area, height-for-age and weight-for-age Z-scores) and factors determined to be in the pathway for

predicting size (maternal pre-pregnancy BMI and infant age at thetime of the scan).

Because of the role that vitamin D has in bone mineralizationand the association of mineral accretion and growth early in lifewith future bone integrity, several investigators have sought tomeasure the relationship between vitamin D status and in utero orearly postnatal bone mineralization.10,16 Despite the range inserum 25(OH)D demonstrated in the present study, no significantassociation was found between maternal or cord 25(OH)Dconcentration and infant WBBMC. This finding contrasts withthat of Weiler et al.,10 in which both maternal and cord plasma25(OH)D remained significant in a multivariate model predictingWBBMC in term-born Canadian infants o15 days of age, albeitwith negative and positive coefficients, respectively. However, inthe same study infants who were vitamin D deficient (plasma25(OH)Do27.5 nmol/l) or born to vitamin D-deficient mothers(plasma 25(OHD)o37.5 nmol/l) were heavier and longer thanthose who were not. Therefore, it is possible that the difference inresults was mediated by an association between vitamin D statusand size, which was not found in the present study.

It is possible that the comparatively better vitamin D status ofthe population included in the present study may have obscuredan association between vitamin D status and infant size and bonemineral content. A higher prevalence of maternal and cord serum25(OH)D concentrations o50 nmol/l has been measured instudies conducted at more northerly latitudes (42% and 35%,respectively, at 461N and 531N in Canada17 and 60–77% and 52%,respectively, at 601N in Finland7).

In adults, variance in body size is accounted for by adjusting bonemineral content for 2-dimensional bone area (areal BMD). Inchildren, BMD is a less valid index due to the dynamic process ofbone turnover and rapid growth leading to asynchronies betweenincrease in bone area and bone mineral.18 Nevertheless, to accountfor individual size differences among infants, investigators haveadjusted WBBMC for bone area or infant weight.7,8,10 In the presentanalyses, such adjustments were made and the power of thedenominator further optimized through logarithmic analysis.Despite adjustments, vitamin D status was unassociated with bonemeasurements. That anthropometric indices remained significantpredictors in the multivariate models of adjusted WBBMC indicatesthat adjustment of the independent variable for weight or bone areaalone did not sufficiently control for inter-individual variances in size.

Limitations of the present study include restricted sample sizeand inherent difficulty in collecting DXA data in infants. Type II errorand the fact that a large percentage of mothers in the study hadadequate vitamin D status may have precluded an ability to detectdifferences in infant WBBMC. Sample size was further limited by thedifficulty in restricting movement in young infants during the DXAscan. To minimize movement artifact, infants were tightly swaddled,fed and pacified immediately before the scan, and the better of twoscans was included in data analysis. Nonetheless, conducting DXAscans on a larger sample of infants is warranted to rule out athreshold effect of vitamin D status on bone mineralization.

In conclusion, no association was found between early infantWBBMC, raw or adjusted for inter-individual differences in size,and maternal or cord vitamin D status in a multiethnic populationin Northern California.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSWe wish to acknowledge the financial support for this study from the USDA, ARSWestern Human Nutrition Research Center and the East Bay NeonatologyFoundation. We extend a special thanks to Janet M Peerson at UC Davis for herassistance with statistical analyses.

Table 2. Mean WBBMC by category (analysis of variance)

N (%) Mean±s.d.BMC (g)

P-value

Maternal 25(OH)D 0.58o27.5 nmol/l 6 (5.0) 67.2±5.227.5–79.99 nmol/l 67 (55.8) 62.2±1.6X80nmol/l 47 (39.2) 61.4±1.9

Cord 25(OH)D 0.44o27.5 nmol/l 26 (21.7) 64.9±2.527.5–79.99 nmol/l 86 (71.7) 61.3±1.4X80nmol/l 8 (6.6) 62.7±4.5

Infant sex 0.39Male 58 (48.3) 61.1±1.7Female 62 (51.7) 63.1±1.6

Season of birtha 0.3421 Jan–20 April 21 (17.5) 61.4±2.721 April–20 July 30 (25.0) 61.0±2.821 July–20 Oct 28 (23.3) 58.9±2.421 Oct–20 Jan 41 (34.2) 63.2±2.0

Feeding behavior 0.34Breastfed 50 (41.7) 60.2±1.8Formula-fed 26 (21.7) 64.1±2.5Mixed 44 (36.7) 63.2±1.9

Maternal pre-pregnancyBMIb (kg/m2)

0.04

19–24.99 (normal) 18 (16.5) 56.9±3.025–29.99 (overweight) 39 (35.8) 62.7±2.030–40 (obese) 34 (31.2) 62.7±2.1440 (morbidly obese) 18 (16.5) 69.0±3.0

Size for GA o0.0001Small for GA(o10th percentile)

19 (15.8) 48.0±2.3

Appropriate for GA 88 (73.4) 62.7±1.1Large for GA(490th percentile)

13 (10.8) 78.9±2.8

Abbreviations: BMC, bone mineral content; BMI, body mass index; GA,gestational age; WBBMC, whole-body bone mineral content. aSeasonalboundaries based on sine curve analysis of maternal serum 25(OH)Dconcentrations. bMissing values for pre-pregnancy BMI in 11 participants.

Vitamin D status and neonatal bone mineral contentDK Dror et al

3

& 2012 Macmillan Publishers Limited European Journal of Clinical Nutrition (2012) 1 – 4

REFERENCES1 Namgung R, Tsang RC. Factors affecting newborn bone mineral content:

in utero effects on newborn bone mineralization. Proc Nutr Soc 2000; 59:55–63.

2 Sayers A, Tobias JH. Estimated maternal ultraviolet B exposure levels in pregnancyinfluence skeletal development of the child. J Clin Endocrinol Metab 2009;94: 765–771.

3 Winsloe C, Earl S, Dennison EM, Cooper C, Harvey NC. Early life factors in thepathogenesis of osteoporosis. Curr Osteoporos Rep 2009; 7: 140–144.

4 Martin R, Harvey NC, Crozier SR, Poole JR, Javaid MK, Dennison EM et al. Placentalcalcium transporter (PMCA3) gene expression predicts intrauterine bone mineralaccrual. Bone 2007; 40: 1203–1208.

5 Morley R, Carlin JB, Pasco JA, Wark JD. Maternal 25-hydroxyvitamin D and para-thyroid hormone concentrations and offspring birth size. J Clin Endocrinol Metab2006; 91: 906–912.

6 Namgung R, Tsang RC, Lee C, Han DG, Ho ML, Sierra RI. Low total bodybone mineral content and high bone resorption in Korean winter-born versussummer-born newborn infants. J Pediatr 1998; 132: 421–425.

7 Viljakainen HT, Saarnio E, Hytinantti T, Miettinen M, Surcel H, Makitie O et al.Maternal vitamin D status determines bone variables in the newborn.J Clin Endocrinol Metab 2010; 95: 1749–1757.

8 Akcakus M, Koklu E, Budak N, Kula M, Kurtoglu S, Koklu S. The relationshipbetween birthweight, 25-hydroxyvitamin D concentrations and bone mineralstatus in neonates. Ann Trop Paediatr 2006; 26: 267–275.

9 Hannan MT, Litman HJ, Araujo AB, McLennan CE, McLean RR, McKinlay JB et al.Serum 25-hydroxyvitamin D and bone mineral density in a racially and ethnicallydiverse group of men. J Clin Endocrinol Metab 2008; 93: 40–46.

10 Weiler H, Fitzpatrick-Wong S, Veitch R, Kovacs H, Schellenberg J, McCloy U et al.Vitamin D deficiency and whole-body and femur bone mass relative to weight inhealthy newborns. CMAJ 2005; 172: 757–761.

11 Dror DK, King JC, Durand DJ, Allen LH. Association of modifiable andnonmodifiable factors with vitamin D status in pregnant women and neonates inOakland, CA. J Am Diet Assoc 2011; 111: 111–116.

12 de Onis M, Onyango AW, Van den Broeck J, Chumlea WC, Martorell R. Measure-ment and standardization protocols for anthropometry used in the constructionof a new international growth reference. Food Nutr Bull 2004; 25: S27–S36.

13 Gordon CM, Bachrach LK, Carpenter TO, Crabtree N, El-Hajj Fuleihan G,Kutilek S et al. Dual energy X-ray absorptiometry interpretation and reportingin children and adolescents: the 2007 ISCD Pediatric Official Positions. J ClinDensitom 2008; 11: 43–58.

14 Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M et al.A new and improved population-based Canadian reference for birth weight forgestational age. Pediatrics 2001; 108: E35.

15 Z-score data files (database on the Internet). Centers for Disease Controland Prevention, National Center for Health Statistics. (cited 1 May 2011).Available from http://www.cdc.gov/growthcharts/zscore.htm.

16 Mahon P, Harvey N, Crozier S, Inskip H, Robinson S, Arden N et al. Low maternalvitamin D status and fetal bone development: cohort study. J Bone Miner Res2010; 25: 14–19.

17 Newhook LA, Sloka S, Grant M, Randell E, Kovacs CS, Twells LK. Vitamin Dinsufficiency common in newborns, children and pregnant women living inNewfoundland and Labrador, Canada. Matern Child Nutr 2009; 5: 186–191.

18 Molgaard C, Thomsen BL, Michaelsen KF. Whole body bone mineral accretion inhealthy children and adolescents. Arch Dis Child 1999; 81: 10–15.

Vitamin D status and neonatal bone mineral contentDK Dror et al

4

European Journal of Clinical Nutrition (2012) 1 – 4 & 2012 Macmillan Publishers Limited