Clinical Endocrinology (2016) 84, 680–686 doi: 10.1111/cen.

12989

ORIGINAL ARTICLE

The persistence of maternal vitamin D deficiency and

insufficiency during pregnancy and lactation irrespective of
season and supplementation
Caroline K. Kramer*,†, Chang Ye*, Balakumar Swaminathan*, Anthony J. Hanley*,†,‡, Philip W. Connelly†,§,¶,
Mathew Sermer**, Bernard Zinman*,†,†† and Ravi Retnakaran*,†,††

*Leadership Sinai Centre for Diabetes, Mount Sinai Hospital, †Division of Endocrinology, University of Toronto, ‡Department of
Nutritional Sciences, University of Toronto, §Keenan Research Centre for Biomedical Science of St. Michael’s Hospital, ¶Department
of Laboratory Medicine and Pathobiology, University of Toronto, **Division of Obstetrics and Gynecology, University of Toronto,
and ††Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada

Conclusions The persistence of vitamin D deficiency/insuffi-

Summary ciency during pregnancy and lactation, irrespective of season
and supplementation, supports the emerging concept that cur-
Background Pregnancy and lactation comprise a critical win- rent vitamin D supplementation in antenatal care is likely inade-
dow spanning all seasons during which maternal vitamin D sta- quate.
tus potentially may influence the long-term health of the
newborn. Women typically receive calcium/vitamin D supple- (Received 27 October 2015; returned for revision 25 November
mentation through antenatal vitamins, but there has been lim- 2015; finally revised 25 November 2015; accepted 26 November
ited serial evaluation of maternal vitamin D status across this 2015)
critical window.
Design/Patients/Measurements In this prospective observa-
tional cohort study, 467 women in Toronto, Canada, underwent
measurement of serum 25-hydroxy vitamin D (25-OH-D) at Introduction
mean 29
7  2
9 weeks’ gestation, 3 months postpartum and Vitamin D is best known for its roles in the maintenance of cal-
12 months postpartum, enabling serial assessment across 3 sea- cium homoeostasis and the regulation of bone metabolism.
sons. At each assessment, vitamin D status was classified as defi- Besides these established effects, a growing body of evidence in
ciency (25-OH-D<50 nmol/l), insufficiency (25-OH-D≥50 nmol/l the past decade has implicated insufficient vitamin D as a poten-
and <75 nmol/l) or sufficiency (25-OH-D≥75 nmol/l). tial factor in a variety of nonskeletal conditions, including can-
Results The prevalence rates of vitamin D deficiency and insuf- cers, autoimmune diseases, diabetes and cardiovascular disease.1–3
ficiency were 31
5% and 35
1% in pregnancy, 33
4% and 35
3% It thus emerges that, although causality remains uncertain, vita-
at 3 months, and 35
6% and 33
8% at 12 months postpartum, min D status may hold wide-ranging implications for the health
respectively. These high rates remained stable over time of the individual.
(P = 0
49) despite declining usage of antenatal calcium/vitamin The dominant source of vitamin D in humans is endogenous
D supplementation from pregnancy to 3 months to 12 months synthesis in the epidermal layer of the skin through a process
postpartum (P < 0
001). Indeed, on mixed model analyses, vita- that is dependent upon exposure to sunlight, specifically ultravi-
min D deficiency and insufficiency in pregnancy were indepen- olet B rays. As such, populations with limited exposure to sun-
dently associated with decrements in average 25-OH-D over light are at risk for vitamin D deficiency and insufficiency, as
time of 49
6 nmol/l and 26
4 nmol/l, respectively (both reflected by low circulating concentration of 25-hydroxy vitamin
P < 0
001). In contrast, season of baseline assessment and use of D (25-0H-D). Owing to the northern latitude of their residence,
calcium/vitamin D supplements were independently associated Canadians are thus at risk for vitamin D deficiency/insufficiency,
with changes in 25-OH-D in the range of 3–5 nmol/l (both especially during the winter months.2 While this risk pertains to
P < 0
008). the entire Canadian population, it may be particularly important
in women of child-bearing age. Specifically, pregnancy and lacta-
tion comprise a unique critical window (i) that spans all seasons
Correspondence: Ravi Retnakaran, Leadership Sinai Centre for Diabetes,
(and therefore includes winter) and (ii) during which maternal
Mount Sinai Hospital, 60 Murray Street, Suite L5-025, Mailbox-21,
Toronto, Ontario, Canada M5T 3L9. Tel.: (416) 586-4800 ext. 3941; vitamin D status may affect the health of both mother and
Fax: (416) 586-8853; E-mail: [email protected] child.4–6 Indeed, in observational studies, low maternal vitamin

680 © 2015 John Wiley & Sons Ltd

 Vitamin D status during and after pregnancy 681

D during this window has been associated with subsequent Statistical analyses
long-term outcomes in the offspring including increased fat
All analyses were conducted using SAS 9
4 (SAS Institute, Cary,
mass at 6 years of age,7 decreased bone mineral content at
NC, USA). Continuous variables were tested for normality of
9 years of age,8 decreased peak bone mass in early adulthood9
distribution, and natural log transformations of skewed variables
and a potentially higher incidence of immune-related illnesses
were used, where necessary, in subsequent analyses. The study
such as asthma and type 1 diabetes.10–14 Despite its potential
population was first stratified into three groups according to
importance, however, there has been limited serial evaluation of
vitamin D status in pregnancy. Univariate differences across
vitamin D status in women across this critical time period.
these three vitamin D groups were assessed at recruitment in
Thus, our objective in this study was to evaluate the longitudinal
pregnancy, at 3 months and at 12 months postpartum using
course of vitamin D status in Canadian women from late preg-
analysis of variance or Kruskal–Wallis test for continuous vari-
nancy across the first year postpartum.
ables, and v2 test or Fisher exact test for categorical variables
(Table 1).
Subjects and methods Weighted least squares method was used to test whether the
respective prevalence rates of vitamin D deficiency, insufficiency
In this prospective observational cohort study, 467 women
and sufficiency change from pregnancy to 3 months postpartum
underwent assessment on three occasions: at the time of recruit-
to 12 months postpartum (Fig. 1a). The method is useful for
ment in late pregnancy, at 3 months postpartum and at
analysing repeated measures of a categorical outcome, and test-
12 months postpartum. The study protocol has been previously
ing the hypothesis of marginal homogeneity (i.e. the marginal
described in detail.15,16 In brief, pregnant women at our institu-
probability that vitamin D status group is the same over time).
tion in Toronto, Canada (latitude 43°420 N), were recruited at
We also used generalized linear mixed model to evaluate the
the time of clinical screening for gestational diabetes in late 2nd
trend of usage of calcium/vitamin D supplements from preg-
trimester. At 3 and 12 months postpartum, participants then
nancy to 3 months postpartum to 12 months postpartum within
returned to the clinical investigation unit to undergo repeat
each strata of vitamin D status (Fig. 1b). In this model, the
evaluation. Based on the timing of the 3 study visits (i.e.
usage of calcium/vitamin D supplements is a repeated-measures
~6 months between 1st and 2nd visit, followed by ~9 months
binary outcome, and duration of follow-up and vitamin D status
between 2nd and 3rd visit), these assessments reflect 3 different
group are time-dependent covariates.
seasons. The study protocol has been approved by the Mount
A mixed model was constructed to evaluate the change in
Sinai Hospital Research Ethics Board, and all participants pro-
average 25-OH-D over the duration of follow-up and to evaluate
vided written informed consent.
the potential determinants associated with the change in average
25-OH-D (Table 2). We performed the analysis with repeated-
Evaluation of women in pregnancy, at 3 months measures 25-OH-D (dependent variable), adjusting for the fol-
postpartum and at 12 months postpartum lowing covariates: age at recruitment in pregnancy, ethnicity,
duration of follow-up, duration of breastfeeding, BMI, total
At each study visit, weight was measured and interviewer-admi-
physical activity score, use of calcium/vitamin D supplements,
nistered questionnaires were completed pertaining to medical,
season of assessment in pregnancy and vitamin D status in preg-
obstetrical and family history, as previously described.15,16 Physi-
nancy. Duration of follow-up, BMI, total physical activity score
cal activity was assessed using the Baecke questionnaire,17,18 a
and use of calcium/vitamin D supplements are time-dependent
validated instrument that measures total physical activity and its
variables. Furthermore, multiple linear regression analysis was
three component domains: occupation-associated activity (work
performed for each visit to evaluate adjusted mean levels of 25-
index), sport-related physical activity (sport index) and nonsport
OH-D over time by vitamin D status group in pregnancy
leisure-time activity (leisure-time index). Work index was not
(Fig. 2a) and by season of assessment in pregnancy (Fig. 2b),
measured at 3 months postpartum, as most women would not
after adjustment for the same covariates as in the mixed model.
yet have returned to their usual occupation at that time.19
At each visit, serum parathyroid hormone (PTH) was mea-
sured using an electrochemiluminescence immunoassay on the
Results
Roche Modular E170 Analyzer (Catalogue number 11972103122;
Laval, Canada), which has a detection range from 0
6 to
Characteristics of Study Population in Pregnancy and
530 pmol/l. Vitamin D status was assessed with measurement of
across First Year Postpartum
serum 25-OH-D by competitive electrochemiluminescent
immunoassay on the Roche Modular E170 (Catalogue number The 467 women comprising the study population were stratified
05894913190; Laval, Canada). This assay has a lower reporting according to their vitamin D status in pregnancy, revealing 147
limit of 8 nmol/l. Vitamin D status was classified as per Endo- (31
5%) with vitamin D deficiency (mean 25-OH-D
crine Society guidelines20 as (i) vitamin D deficient (25-OH- 26
3  9
4 nmol/l), 164 (35
1%) with insufficiency (mean 25-
D<50 nmol/l), (ii) vitamin D insufficient (25-OH-D≥50 nmol/l OH-D 61
8  7
1 nmol/l) and 156 (33
4%) with vitamin D
and <75 nmol/l) or (iii) vitamin D sufficient (25-OH- sufficiency (mean 25-OH-D 91
7  12
9 nmol/l). Table 1 shows
D ≥75 nmol/l). the characteristics of these 3 groups at the assessments in

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686
682 C. K. Kramer et al.

Table 1. Characteristics of study population stratified into the following three groups based on vitamin D status in pregnancy: (i) deficient, (ii)
insufficient and (iii) sufficient

Vitamin D Status in Pregnancy

Deficient Insufficient Sufficient

(25-OH-D <50 nmol/l) (50 ≤25-OH-D <75 nmol/l) (25-OH-D ≥75 nmol/l)
n = 147 n = 164 n = 156 P

In Pregnancy
Weeks gestation at assessment (weeks) 29
0 (28
0–31
0) 30
0 (28
0–31
0) 30
0 (29
0–32
0) 0
03
Age (years) 34
2  4
3 34
5  4
2 34
3  4
3 0
80
Ethnicity: <0
001
White (%) 63
3 68
9 82
7
Asian (%) 11
6 14
7 9
0
Other (%) 25
2 16
5 8
3
Current smoking (%) 2
1 1
8 1
3 0
90
Season of blood sample collection: 0
006
Winter (%) 31
3 23
8 16
0
Spring (%) 29
9 29
9 29
5
Summer (%) 17
0 27
4 35
9
Fall (%) 21
8 18
9 18
6
25-OH-D (nmol/l) 26
3  9
4 61
8  7
1 91
7  12
9 <0
001
PTH (pmol/l) 3
5 (2
9–4
7) 2
9 (2
4–3
9) 2
7 (2
2–3
3) <0
001
Prepregnancy BMI (kg/m2) 25
6 (22
3–30
2) 23
1 (21
2–26
7) 22
8 (20
1–25
4) <0
001
Gestational weight gain up to OGTT (kg) 10
0 (6
5–13
7) 10
1 (7
9–13
9) 10
5 (8
8–13
9) 0
16
Gestational diabetes (%) 25
5 29
6 24
0 0
50
Use of calcium/vitamin D supplements (%) 74
2 87
2 89
1 <0
001
At 3 months postpartum
Total physical activity: 4
8  0
9 4
9  1
0 5
1  1
0 0
06
Sport index 1
75 (1
5–2
3) 2
00 (1
5–2
3) 2
00 (1
5–2
8) 0
08
Leisure-time index 2
85  0
6 2
92  0
6 2
95  0
54 0
24
25-OH-D (nmol/l) 41
3  16
8 61
5  17
2 83
9  16
9 <0
001
PTH (pmol/l) 3
8 (2
8–4
8) 3
1 (2
2–4
1) 2
9 (2
3–3
6) <0
001
BMI (kg/m2) 27
6 (24
2–31
8) 25
7 (22
6–29
1) 24
4 (22
5–27
8) <0
001
Use of calcium/vitamin D supplements (%) 42
9 55
4 69
2 <0
001
At 12 months postpartum
Total physical activity: 8
1  1
3 8
2  1
3 8
6  1
3 0
008
Sport index 2
3 (1
8–2
8) 2
0 (1
8–2
8) 2
5 (1
8–3
0) 0
02
Leisure-time index 3
0  0
6 3
1  0
6 3
1  0
5 0
04
Work index 3
0 (2
4–3
4) 3
0 (2
6–3
4) 3
1 (2
6–3
4) 0
47
Duration of breastfeeding (months) 8
3 (3
0–12
0) 11
0 (6
0–12
0) 9
0 (5
0–12
0) 0
11
25-OH-D (nmol/l) 41
6  20
0 58
5  18
7 83
3  21
9 <0
001
PTH (pmol/l) 4
2 (3
2–5
1) 3
4 (2
8–4
4) 3
4 (2
6–4
2) <0
001
BMI (kg/m2) 26
6 (23
4–31
7) 24
4 (21
6–27
9) 23
4 (21
3–26
3) <0
001
Use of calcium/vitamin D supplements (%) 27
9 34
8 45
5 0
006

Continuous variables presented as mean  SD (if normally distributed) or median followed by interquartile range in parentheses (if skewed). Categori-
cal variables presented as proportions.

pregnancy, 3 months postpartum and 12 months postpartum. (mean 41
3  16
8 nmol/l), followed in turn by those in the
As expected, the women comprising the antepartum vitamin D antepartum insufficiency group (mean 61
5  17
2 nmol/l), and
deficient group had the highest prevalence of non-white ethnic- those with vitamin D sufficiency in pregnancy (mean
ity and were most likely to have been assessed during the winter. 83
9  16
9 nmol/l) (P < 0
001). Similarly, there was an accompa-
The use of calcium/vitamin D supplementation progressively nying gradient in the usage of calcium/vitamin D supplements
increased from the deficient group (74
2%) to the insufficient (42
9% vs 55
4% vs 69
2%, P < 0
001), while current BMI and PTH
group (87
2%) to the sufficient group (89
1%) ( P < 0
001). both progressively decreased across these groups (both P < 0
001).
Similarly, prepregnancy BMI and serum PTH in pregnancy both Women with vitamin D sufficiency in pregnancy were the most
showed a stepwise decline across these groups (both P < 0
001). physically active at 3 months postpartum, with total physical activ-
At 3 months postpartum, the gradients observed in pregnancy ity differing across the groups at borderline significance (P = 0
06).
persisted across these 3 groups. Indeed, women with vitamin D All of these patterns remained at 12 months postpartum, with
deficiency in pregnancy continued to have the lowest 25-OH-D the between-group difference in total physical activity now

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686
 Vitamin D status during and after pregnancy 683

(a) Vitamin D sufficiency Table 2. Mean adjusted change in average 25-OH-D over the duration
Vitamin D insufficiency of follow-up from pregnancy across the first year postpartum, per unit
Vitamin D deficiency change in each of the indicated determinants

100 Adjusted Change in 25-OH-D

P = 0·49
and postpartum period (%)
status group in pregnancy
Prevalence of vitamin D

(nmol/L)
80
Estimate (95% CI) P
60
Age (years) 0
25 (0
02 to 0
48) 0
03
40 Ethnicity (reference = white):
Asian 4
34 (7
47 to 1
20) 0
007
Other 2
17 (4
93 to 0
59) 0
12
20
Duration of 0
05 (0
18 to 0
08) 0
43
follow-up (months)
0 Duration of 0
04 (0
17 to 0
24) 0
74
Pregnancy 3-months 12-months
postpartum postpartum breastfeeding (months)
BMI (kg/m2) 0
27 (0
46 to 0
08) 0
005
(b)
Prevalence of use of vitamin D/calcium

Total physical activity score 0
16 (0
21 to 0
54) 0
4

P < 0·001 Use of vitamin 3
43 (1
77 to 5
09) <0
001
100
D/calcium supplements
supplements in pregnancy
and postpartum period (%)

Season of assessment in pregnancy (ref = summer)

80 Winter 5
39 (2
62 to 8
16) <0
001
Spring 3
54 (0
94 to 6
14) 0
008
60 Autumn 0
73 (2
14 to 3
60) 0
62
Vitamin D status in pregnancy (reference = sufficient):
40 Vitamin D deficient 49
6 (52
2 to 47
1) <0
001
Vitamin D insufficient 26
4 (28
7 to 24
0) <0
001
20
The model shows the mean adjusted change in average 25-OH-D per
unit change in each of the indicated determinants, after adjustment for
0
Pregnancy 3-months 12-months all of the other variables. For example, use of vitamin D/calcium supple-
postpartum postpartum ments was associated with an increment in 25-OH-D of 3
43 nmol/l over
the follow-up. Similarly, baseline assessment in winter (as compared to
Fig. 1 (Panel a) Prevalence of vitamin D sufficiency (white bar), summer) was associated with an increment in 25-OH-D of 5
39 nmol/l
insufficiency (grey bar) and deficiency (black bar) at each of pregnancy, over the follow-up (because the 2nd and 3rd visits would likely then
have been in summer and spring).
3 months postpartum and 12 months postpartum, respectively.
BMI, total physical activity score and use of vitamin D/calcium supple-
(Panel b) Prevalence of use of calcium/vitamin D supplements in
ments are time-dependent variables.
women with vitamin D sufficiency, insufficiency and deficiency at each
of pregnancy, 3 months postpartum and 12 months postpartum,
the overall usage of supplements declined from pregnancy to
respectively. P-values refer to the changes in proportions over time.
3 months to 12 months postpartum (P < 0
001). It thus
reaching statistical significance (P = 0
008). It thus emerges that emerges that, during the critical window of pregnancy and the
vitamin D status in pregnancy identified gradients in 25-OH-D, first year postpartum, the rates of vitamin D deficiency and
PTH, BMI, use of calcium/vitamin D supplementation and insufficiency are stable in the presence of declining usage of cal-
physical activity that persisted at 3 months and 12 months post- cium/vitamin D supplements over time (Fig. 1a and b).
partum. Furthermore, despite being measured 9 months apart
(i.e. ~3 seasons apart), 25-OH-D concentrations in the three ges-
Determinants of vitamin D status over time
tational groups were nearly unchanged between 3 and
12 months postpartum. To evaluate the relative contributions of potential determinants
As shown in Fig. 1a, the respective prevalence rates of vitamin of vitamin D status over time in this study population, we per-
D sufficiency, insufficiency and deficiency were largely stable in formed mixed model analyses of (dependent variable) 25-OH-D
the study population from pregnancy to 3 months postpartum with the following covariates: age, ethnicity, duration of follow-
to 12 months postpartum. Indeed, at each point in time, up, duration of breastfeeding, BMI, total physical activity, use of
approximately one-third of the women had vitamin D suffi- calcium/vitamin D supplements, season of baseline assessment
ciency, one-third were insufficient, and one-third were deficient, and baseline vitamin D status in pregnancy. As shown in Fig. 2a,
with no significant change in these proportions over time after adjustment for covariates, the 3 vitamin D groups in preg-
(P = 0
49). At each timepoint, there was a stepwise decrease in nancy exhibited distinct trajectories of mean adjusted 25-OH-D
the prevalence of usage of calcium/vitamin D supplements from over time, with persistent incremental differences of >20 nmol/l
the vitamin D sufficient to insufficient to deficient group at each of the 3 assessments (pregnancy, 3 months postpartum
(Fig. 1b). Fig. 1b also shows that, regardless of vitamin D status, and 12 months postpartum). Accordingly, mean adjusted 25-

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686
684 C. K. Kramer et al.

(a) 100 ficiency in pregnancy were associated with decrements in average

Vit D deficient
Vit D insufficient
25-OH-D of 49
6 nmol/l and 26
4 nmol/l, respectively. In con-
Vitamin D (25-OH-D) (nmol/l)

Vit D sufficient trast, season of baseline assessment, Asian ethnicity and use of
80 calcium/vitamin D supplements were associated with changes in
25-OH-D in the range of 3 to 5 nmol/l while unit changes in
age and BMI had even lesser impact (change in 25-OH-
60 D<1 nmol/l).

Discussion
40
In this cohort of Canadian women evaluated across the critical
window of pregnancy and lactation, the respective prevalence
20 rates of vitamin D deficiency, insufficiency and sufficiency were
Pregnancy 3-months 12-months
largely stable at approximately 30–35% apiece at each of early
3rd trimester, 3 months postpartum and 12 months postpartum.
(b) 100
Winter These stable rates were observed across 3 seasons and in the
Spring presence of declining usage of calcium/vitamin D supplements
Vitamin D (25-OH-D) (nmol/l)

Summer
over time. Indeed, vitamin D status in pregnancy was the domi-
80 Fall
nant independent determinant of average 25-OH-D across the
follow-up, exerting far greater impact than other factors such as
season of assessment, use of calcium/vitamin D supplements,
60
ethnicity, age and BMI. It thus emerges that vitamin D defi-
ciency/insufficiency persists across the critical window of preg-
nancy and lactation, irrespective of traditional risk factors and
40
current supplementation practices.
During pregnancy and lactation, maternal physiologic adapta-
tions serve to protect calcium homoeostasis for the foetus such
20
Pregnancy 3-months 12-months that, even in the presence of maternal vitamin D deficiency, the
newborn can have normal serum calcium and seemingly normal
Fig. 2 Adjusted mean concentrations of serum 25-OH-D at each of skeletal health.6,21 However, although causality remains unclear,
pregnancy, 3 months postpartum and 12 months postpartum in study
observational studies have linked low maternal vitamin D con-
population stratified according to (Panel a) vitamin D status in
centration during this period with adverse effects in the off-
pregnancy and (Panel b) the season of blood collection in pregnancy,
respectively. Models are adjusted for age, ethnicity, duration of follow- spring including increased fat mass at age 4 years,7 decreased
up, duration of breastfeeding, BMI, total physical activity, use of bone mineral content at age 9 years,8 decreased peak bone mass
calcium/vitamin D supplements, season of baseline assessment and in early adulthood,9 and a potentially higher incidence of
baseline vitamin D status in pregnancy. immune-related illnesses such as asthma and type 1 diabetes.10–14
Furthermore, higher vitamin D intake has been associated with a
OH-D over the duration of follow-up differed significantly decrease in pregnancy complications and improved birth out-
between the 3 groups (P < 0
0001). Specifically, post hoc tests comes.22–24 As such, there is currently considerable interest in
using the Bonferroni correction revealed that women with vita- maternal vitamin D status during and after pregnancy.
min D deficiency in pregnancy had the lowest mean adjusted Although not all using the same 25-OH-D diagnostic thresh-
25-OH-D over the duration of follow-up, followed in turn by olds, previous studies have documented suboptimal vitamin D
those with vitamin D insufficiency and sufficiency in pregnancy, status in the Canadian population,25,26 including the Canadian
respectively (P < 0
0001 for each pairwise comparison). In con- Health Measures Survey which reported that 63% of nonpreg-
trast, the 4 groups defined by the season of the baseline assess- nant women aged 20–39 years had 25-OH-D <75 nmol/l.27 In a
ment in pregnancy showed comparatively modest variation in study of 336 pregnant women, 65% had 25-OH-D <75 nmol/l
mean adjusted 25-OH-D across the 3 points in time (i.e. reflect- on cross-sectional assessment between 20 and 35 weeks’ gesta-
ing 3 different seasons) (Fig. 2b). tion.28 A study involving 17 urban obstetric hospitals across
On mixed model analysis (Table 2), the independent determi- Canada reported a 39% prevalence of 25-OH-D <50 nmol/l in
nants of average 25-OH-D over time were baseline vitamin D pregnancy.29 High rates of vitamin D deficiency/insufficiency in
deficiency and insufficiency in pregnancy, baseline assessment in pregnancy have been similarly reported in countries around the
winter and spring, use of calcium/vitamin D supplements (time- world,4,5 including those that bear resemblance to Canada in
dependent), BMI (time-dependent), age and Asian ethnicity. healthcare, ethnic composition and lifestyle, such as the United
Importantly, vitamin D deficiency and insufficiency in pregnancy States and European nations.30–32 However, to date, there have
had far greater impact on average 25-OH-D over time, than did been few longitudinal studies reporting serial evaluation of vita-
any of the other factors. Indeed, vitamin D deficiency and insuf- min D status in women beginning in pregnancy and continuing

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686
 Vitamin D status during and after pregnancy 685

across the first year postpartum. While consistent with the cur- partum in this population. Of note, Hollis et al. have recently
rent findings, these studies have been limited by modest sample demonstrated the safety of high-dose maternal vitamin D sup-
sizes (including 141 women in Denmark,33 75 women in United plementation during lactation and reported that ~6000 IU daily
Arab Emirates,34 and 30 women in Ireland35) and variable supplies the infant with adequate vitamin D.38
degrees of adjustment for potential confounding factors. In summary, two-thirds of the women in this cohort had vita-
In this context, the current study extends the literature by min D deficiency/insufficiency at each of 3 assessments traversing
providing longitudinal evaluation of 467 women from pregnancy pregnancy and the first year postpartum. These stable rates were
across the first year postpartum, with adjustment for determi- observed across 3 different seasons and in the presence of declin-
nants of vitamin D status (including season of assessment, eth- ing usage of calcium/vitamin D supplements over time. It thus
nicity, use of calcium/vitamin D supplements, BMI and physical emerges that vitamin D deficiency/insufficiency is both highly
activity). In doing so, we provide robust evidence of a high prevalent and persistent across the critical window of pregnancy
prevalence of maternal vitamin D deficiency/insufficiency that and lactation, irrespective of traditional risk factors and current
was stable across the entire duration of follow-up despite mea- supplementation practices. As such, clinical trials evaluating the
surement in 3 different seasons and declining rates of supple- effects of increased maternal vitamin D supplementation during
mentation. Furthermore, in quantifying the relative magnitude this key juncture in the life cycle may be particularly important for
of impact of each factor after adjustment for the others the health of women and ultimately that of their children.
(Table 2), we show that the baseline vitamin D status in preg-
nancy was far and away the dominant determinant of 25-OH-D Funding
concentrations across this critical window. Accordingly, these
data suggest that vitamin D deficiency/insufficiency that is This study was supported by operating grants MOP-84206 and
detected at any time during this window may be expected to MHC-115442 from the Canadian Institutes of Health Research
persist over time (e.g. regardless of season) unless it is treated. (CIHR) and OG-3-11-3300-RR from the Canadian Diabetes
It is noteworthy that (i) high rates of antepartum vitamin D Association (CDA). AJH holds a Tier-II Canada Research Chair
deficiency and insufficiency were observed despite the prevalent in Diabetes Epidemiology. BZ holds the Sam and Judy Pencer
usage of supplements in pregnancy and that (ii) these rates were Family Chair in Diabetes Research at Mount Sinai Hospital and
then largely unchanged when supplement usage declined at both University of Toronto. RR holds a Heart and Stroke Foundation
3 and 12 months postpartum (Fig. 1a and b). Taken together, of Ontario (HSFO) Mid-Career Investigator Award.
these data suggest that the doses of supplementation that were
used in the clinical care of these women did not appreciably Author contributions
raise 25-OH-D (as further evidenced by an estimated indepen-
dent effect size of only 3
4 nmol/l in the adjusted analysis in CK, AJH, PWC, MS, BZ and RR designed the research and con-
Table 2). A limitation of this study is that the precise formula- ducted the study. CK and RR designed the analysis plan and
tions and doses of supplements were not characterized; however, contributed to the analyses. CY and BS performed the statistical
for the vast majority of the women, supplementation was in the analyses. RR wrote the first draft and supervised the analysis and
form of antenatal vitamins, which typically contain 400 IU of manuscript. All authors contributed to interpretation of the data
vitamin D. Accordingly, our findings are supportive of a grow- and critical revision of the manuscript. All authors read and
ing body of literature suggesting that the typical supplementa- approved the manuscript. RR had full access to all of the data in
tion dose in current obstetrical practice (400 IU/day) is the study and takes responsibility for the integrity of the data
inadequate for preventing vitamin D deficiency/insufficiency in and the accuracy of the data analysis.
pregnancy.5 Indeed, in clinical trials, higher doses of vitamin D
(2000 and 4000 IU daily) have been shown to increase maternal Disclosure statement
serum 25-OH-D without adverse consequences.5,36 Ultimately,
the case for increasing supplementation in pregnancy will The authors have nothing to disclose.
depend upon whether a beneficial effect on clinical outcomes
can be achieved and conclusively demonstrated.37
References
A limitation of the current study is its observational design,
which precludes commentary on the potential beneficial effects 1 Holick, M.F. (2007) Vitamin D deficiency. New England Journal
that could be obtained by raising maternal serum 25-OH-D. In of Medicine, 357, 266–281.
2 Hanley, D.A., Cranney, A., Jones, G. et al. (2010) Vitamin D in
this regard, clinical trials of vitamin D supplementation during
adult health and disease: a review and guideline statement from
pregnancy and lactation will be critical. The importance of such
Osteoporosis Canada. CMAJ, 182, E610–E618.
trials (some of which are ongoing) is underscored by the wide-
3 McGreevy, C. & Williams, D. (2011) New insights about vitamin
spread prevalence of vitamin D deficiency/insufficiency and the D and cardiovascular disease: a narrative review. Annals of Inter-
anticipated ease of implementing the intervention (supplementa- nal Medicine, 155, 820–826.
tion). The findings of the current study highlight the relevance 4 Mulligan, M.L., Felton, S.K., Riek, A.E. et al. (2010) Implications
of these trials for pregnant women and the potential particular of vitamin D deficiency in pregnancy and lactation. American
importance of continued supplementation in the first year post- Journal of Obstetrics and Gynecology, 202, 429. e1-9.

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686
686 C. K. Kramer et al.

5 Wagner, C.L., Taylor, S.N., Johnson, D.D. et al. (2012) The role 22 Hollis, B.W. & Wagner, C.L. (2013) Vitamin D and pregnancy:
of vitamin D in pregnancy and lactation: emerging concepts. skeletal effects, nonskeletal effects, and birth outcomes. Calcified
Women’s Health, 8, 323–340. Tissue International, 92, 128–139.
6 Kovacs, C.S. (2013) Maternal vitamin D deficiency: fetal and 23 Sablok, A., Batra, A., Thariani, K. et al. (2015) Supplementation
neonatal implications. Seminars in Fetal & Neonatal Medicine, of vitamin D in pregnancy and its correlation with feto-maternal
18, 129–135. outcome. Clinical Endocrinology, 83, 536–541.
7 Crozier, S.R., Harvey, N.C., Inskip, H.M. et al. (2012) Maternal 24 Wagner, C.L., Baggerly, C., McDonnell, S. et al. (2016) Post-hoc
vitamin D status in pregnancy is associated with adiposity in the analysis of vitamin D status and reduced risk of preterm birth in
offspring: findings from the Southampton Women’s Survey. two vitamin D pregnancy cohorts compared with South Carolina
American Journal of Clinical Nutrition, 96, 57–63. march of dimes 2009-2011 rates. Journal of Steroid Biochemistry
8 Javaid, M.K., Crozier, S.R., Harvey, N.C. et al. (2006) Maternal and Molecular Biology, 155(Pt B), 245–251.
vitamin D status during pregnancy and childhood bone mass at 25 Greene-Finestone, L.S., Berger, C., de Groh, M. et al. (2011)
age 9 years: a longitudinal study. Lancet, 367, 36–43. 25-Hydroxyvitamin D in Canadian adults: biological, environ-
9 Zhu, K., Whitehouse, A.J., Hart, P.H. et al. (2014) Maternal vita- mental, and behavioral correlates. Osteoporosis International, 22,
min D status during pregnancy and bone mass in offspring at 1389–1399.
20 years of age: a prospective cohort study. Journal of Bone and 26 Gagnon, C., Baillargeon, J.P., Desmarais, G. et al. (2010) Preva-
Mineral Research, 29, 1088–1095. lence and predictors of vitamin D insufficiency in women of
10 Camargo, C.A. Jr, Rifas-Shiman, S.L., Litonjua, A.A. et al. (2007) reproductive age living in northern latitude. European Journal of
Maternal intake of vitamin D during pregnancy and risk of Endocrinology, 163, 819–824.
recurrent wheeze in children at 3 y of age. American Journal of 27 Langlois, K., Greene-Finestone, L., Little, J. et al. (2010) Vitamin
Clinical Nutrition, 85, 788–795. D status of Canadians as measured in the 2007 to 2009 Canadian
11 Anderson, L.N., Chen, Y., Omand, J.A. et al. (2015) Vitamin D Health Measures Survey. Health Reports, 21, 47–55.
exposure during pregnancy, but not early childhood, is associ- 28 Li, W., Green, T.J., Innis, S.M. et al. (2011) Suboptimal vitamin
ated with risk of childhood wheezing. Journal of Developmental D levels in pregnant women despite supplement use. Canadian
Origins of Health and Disease, 6, 308–316. Journal of Public Health, 102, 308–312.
12 Beckhaus, A.A., Garcia-Marcos, L., Forno, E. et al. (2015) Mater- 29 Wei, S.Q., Audibert, F., Hidiroglou, N. et al. (2012) Longitudinal
nal nutrition during pregnancy and risk of asthma, wheeze and vitamin D status in pregnancy and the risk of pre-eclampsia.
atopic diseases during childhood: a systematic review and meta- BJOG, 119, 832–839.
analysis. Allergy, 70, 1588–1604. 30 Bodnar, L.M., Simhan, H.N., Powers, R.W. et al. (2007) High
13 Hypponen, E., Laara, E., Reunanen, A. et al. (2001) Intake of prevalence of vitamin D insufficiency in black and white preg-
vitamin D and risk of type 1 diabetes: a birth-cohort study. Lan- nant women residing in the northern United States and their
cet, 358, 1500–1503. neonates. Journal of Nutrition, 137, 447–452.
14 Sørensen, I.M., Joner, G., Jenum, P.A. et al. (2012) Maternal 31 Johnson, D.D., Wagner, C.L., Hulsey, T.C. et al. (2011) Vitamin
serum levels of 25-hydroxy-vitamin D during pregnancy and risk D deficiency and insufficiency is common during pregnancy.
of type 1 diabetes in the offspring. Diabetes, 61, 175–178. American Journal of Perinatology, 28, 7–12.
15 Kramer, C.K., Swaminathan, B., Hanley, A.J. et al. (2014) 32 Karras, S.N., Anagnostis, P., Annweiler, C. et al. (2014) Maternal
Prospective associations of vitamin D status with beta-cell func- vitamin D status during pregnancy: the Mediterranean reality.
tion, insulin sensitivity and glycemia: the impact of parathyroid European Journal of Clinical Nutrition, 68, 864–869.
hormone status. Diabetes, 63, 3868–3879. 33 Milman, N., Hvas, A.M. & Bergholt, T. (2011) Vitamin D sta-
16 Kramer, C.K., Swaminathan, B., Hanley, A.J. et al. (2014) Vita- tus during normal pregnancy and postpartum. A longitudinal
min D and parathyroid hormone status in pregnancy: impact on study in 141 Danish women. Journal of Perinatal Medicine, 40,
insulin sensitivity, beta-cell function and gestational diabetes 57–61.
mellitus. Journal of Clinical Endocrinology and Metabolism, 99, 34 Narchi, H., Kochiyil, J., Zayed, R. et al. (2010) Maternal vitamin
3262–3269. D status throughout and after pregnancy. Journal of Obstetrics
17 Baecke, J.A., Burema, J. & Frijters, J.E. (1982) A short question- and Gynaecology, 30, 137–142.
naire for the measurement of habitual physical activity in epi- 35 Zhang, J.Y., Lucey, A.J., Horgan, R. et al. (2014) Impact of preg-
demiology studies. American Journal of Clinical Nutrition, 36, nancy on vitamin D status: a longitudinal study. British Journal
936–942. of Nutrition, 112, 1081–1087.
18 Pereira, M.A., Fitzgerald, S.J., Gregg, E.W. et al. (1997) A collec- 36 Hollis, B.W., Johnson, D., Hulsey, T.C. et al. (2011) Vitamin D
tion of Physical Activity Questionnaires for health-related supplementation during pregnancy: double-blind, randomized
research. Medicine and Science in Sports and Exercise, 29, S1–S205. clinical trial of safety and effectiveness. Journal of Bone and Min-
19 Kew, S., Ye, C., Hanley, A.J. et al. (2014) Cardiometabolic impli- eral Research, 26, 2341–2357.
cations of postpartum weight changes in the first year after deliv- 37 Perez-Lopez, F.R., Pasupuleti, V., Mezones-Holguin, E. et al.
ery. Diabetes Care, 37, 1998–2006. (2015) Effect of vitamin D supplementation during pregnancy
20 Holick, M.F., Binkley, N.C., Bischoff-Ferrari, H.A. et al. (2011) on maternal and neonatal outcomes: a systematic review and
Evaluation, treatment, and prevention of vitamin D deficiency: meta-analysis of randomized controlled trials. Fertility and Steril-
an Endocrine Society clinical practice guideline. Journal of Clini- ity, 103, 1278–1288. e4.
cal Endocrinology and Metabolism, 96, 1911–1930. 38 Hollis, B.W., Wagner, C.L., Howard, C.R. et al. (2015) Maternal
21 Kovacs, C.S. (2008) Vitamin D in pregnancy and lactation: versus infant vitamin D supplementation during lactation: a ran-
maternal, fetal, and neonatal outcomes from human and animal domized controlled trial. Pediatrics, 136, 625–634.
studies. American Journal of Clinical Nutrition, 88, 520S–528S.

© 2015 John Wiley & Sons Ltd

Clinical Endocrinology (2016), 84, 680–686