Higher Prevalence of Inadequate Vitamin D Intake in Toddlers Not Consuming a Vitamin D Supplement in Vancouver, Canada

Abstract

Background

Vitamin D is required for growth and development. However, little is known about dietary intake and status during early childhood.

Objectives

To determine vitamin D intake and status in a group of Canadian toddlers.

Methods

This is a secondary analysis of data from toddlers enrolled in a double-blind randomized controlled trial of fatty acid supplementation, between ages 1 y and 2 y in Vancouver (latitude 49°N), Canada. Dietary information was collected using a 3-d food record, and supplement use by questionnaire. Plasma 25-hydroxyvitamin D₃ (25OHD) concentrations were quantified by liquid chromatography-tandem mass spectrometry. Blood sampling was dichotomized by season as winter (September to February) and summer (March to August).

Results

At age 1 y (n = 110) and 2 y (n = 86), the top food sources of vitamin D were dairy (73‒81%), fish (6.8‒14%), and eggs (4.1‒5.5%). Vitamin D-containing supplements were consumed by 43% of the toddlers at age 1 y and 48% at 2 y. Daily median (interquartile range) total vitamin D intakes were 7.9 (4.9‒13) μg at age 1 y and 7.2 (4.8‒14) μg at age 2 y. There was a higher prevalence of inadequate intake (<10 μg/d) in toddlers not using vitamin D supplements at age 1 y (92% compared with 19%) and 2 y (96% compared with 34%). Overall, mean (standard deviation) 25OHD concentrations were 59.9 (17.6) nmol/L at age 1 y (n = 124) and 60.9 (13.6) nmol/L at 2 y (n = 104). The prevalence of low vitamin D status (<50 nmol/L) was 27% at age 1 y and 22% at 2 y. At age 1 y, the prevalence of low vitamin D status was higher in winter (adjusted prevalence 43% compared with 13%; P < 0.001). Total vitamin D intake correlated with 25OHD at both ages (r ∼0.2, P < 0.05).

Conclusions

These findings suggest a high prevalence of inadequate vitamin D intakes in toddlers who do not consume a vitamin D supplement. A national-level surveillance is needed for this important life stage.

Introduction

Vitamin D is a key regulator of calcium and phosphorous metabolism and is essential for skeletal development and optimal bone health [1]. A failure to achieve adequate vitamin D status during childhood can impair growth [2]. There is mixed evidence that inadequate vitamin D status is associated with an increased risk of diseases beyond its role in bone health, including immune and metabolic diseases, such as asthma and type 1 diabetes [3].

Vitamin D can be obtained from endogenous synthesis through sunlight exposure (UV B radiation), from food (including natural sources and fortified foods), and supplements [4]. However, endogenous synthesis is not a reliable source of vitamin D as it varies depending on latitude, skin pigmentation, and sun exposure [5].

Vitamin D requirements are based on musculoskeletal outcomes and related deficiency thresholds for the vitamin D biomarker, 25-hydroxyvitamin D₃ (25OHD), with the assumption of minimal sunlight exposure [6,7]. Dietary reference intakes for vitamin D intake recommendations, for the United States and Canada, were revised in 2011, and the estimated average requirement (EAR) for people aged 1 y and older was set to 10 μg (400 IU) [7]. For infants receiving exclusively human milk and mixed feeds (human milk and infant formula), such as in the United States and Canada, a daily vitamin D supplement of 10 μg is recommended [8,9]. Natural dietary vitamin D sources are limited (e.g., fish, egg, and mushrooms) [10,11], and for many, the main sources are fortified foods (e.g., cow milk and infant formula). As such, the toddler stage of development (age 1‒3 y) represents a time when young children may be at a higher risk of vitamin D deficiency because they no longer consume infant formula or take vitamin D supplements.

Data on the prevalence of vitamin D deficiency in toddlers are limited and vary worldwide. One contributing factor is that endogenous synthesis is affected by geographical location and seasonal fluctuation. Moreover, there is a lack of consensus on cut-offs for optimal 25OHD concentrations [12,13]. In Ireland (Cork, latitude 51°N), where sunlight exposure is limited during the winter, it was reported that toddlers (n = 741, age 2 y) had lower 25OHD concentrations during winter compared to summer (55 compared with 71 nmol/L, respectively; P < 0.001) [14]. Recent reports of vitamin D deficiency in toddlers range from 3.1% in Colombia (<30 nmol/L; n = 2104, age 1‒2 y; latitude 4°N) [15] to 30% in Jordan (<50 nmol/L; n = 139, age 1‒3 y; latitude 32°N) [16].

In Canada, the prevalence of vitamin D deficiency has been reported at the national level in children aged 3‒8 y to be 2.1% (n = 3525; 2012‒2019 Canadian health measures survey), based on a definition of circulating 25OHD <30 nmol/L [17]. Other smaller studies in Canada report a higher prevalence of deficiency with 32% in toddlers aged 2‒2.5 y in Toronto (n = 91; latitude 43°N), defined as circulating 25OHD <50 nmol/L [18]; and 12.5% in children aged 3‒5 y in Nunavut (n = 388; latitude 70°N), defined as circulating 25OHD <25 nmol/L [19]. Given the paucity of data on vitamin D status in children below age 3 y in Canada, we aimed to determine vitamin D intake and status in toddlers at age 1 y and 2 y from Metro Vancouver (latitude 49°N), Canada to estimate the prevalence of inadequate intakes and supplement use, vitamin D status, and whether status changes between 1 y and 2 y.

Methods

Participants

This study is a secondary analysis of data collected from toddlers between ages 1 y and 2 y enrolled in a longitudinal double-blind randomized controlled trial of arachidonic acid and DHA (200 mg/d each) supplementation. Details of the trial have been published [20]; trial registry: NCT01263912. For the present analysis, the convenience sample included toddlers with complete dietary information and plasma samples available to assess 25OHD. This research was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the University of British Columbia Clinical Research Ethics Board and the Children’s and Women’s Health Centre of British Columbia Research Ethics Board (H09-02028 and H20-02547). Written informed consent was obtained from all parents or legal guardians before participation.

Infants, born at term with no known health problems, were recruited at age 12‒14 mo (1 y, n = 133) between January 2010 and September 2014 in Metro Vancouver (latitude 49°N), British Columbia, Canada. The inclusion criteria in the trial included: appropriate weight for gestational age at birth (2500‒4000 g); singleton birth; maternal age 20‒40 y at delivery; no known food allergies, metabolic, neurological, genetic, or immune disorders; growth failure; or hospitalization for any reason; English as the primary language in the home; non-smoking home environment; human milk intake ≤2 times per day or fed ≤236 mL/d infant formula containing arachidonic acid and DHA; the primary source of milk for the toddler was cow milk or other milk substitutes not containing arachidonic acid and DHA.

During the first clinical visit, sociodemographic information was collected from the parents through a questionnaire. Anthropometric measures were assessed at each study visit, including body weight and length/height; weight status was categorized using the WHO child growth charts z-scores stratified by age and sex [21].

Intake assessment

Dietary information was collected using a 3-d food record, including 2 weekdays and 1 weekend day. Parents were instructed to record all foods and beverages that toddlers have consumed, including amount, brand, recipe, and preparation method. The dietary information was checked for clarity and completeness by the study staff. Participants with missing or incomplete dietary data were not included. Daily dietary intakes of energy, macronutrients, and vitamin D were estimated using the Food Processor Nutrition Analysis Software (ESHA Research) and the Canadian Nutrient File [11]. Food group sources of vitamin D were identified and expressed as a percentage (%) contribution to intakes. In addition, dietary sources of participants’ vitamin D intakes were categorized as natural (i.e., fish and eggs) and fortified (i.e., fortified dairy and other products).

The use of supplements containing vitamin D was assessed by questionnaire. Information, including type, brand, and amount of supplement consumed, was reported by the parents. The composition of each supplement was obtained from product labels or the manufacturer’s product information. Information on the type of supplement and the reason for consumption was also collected. Total vitamin D intakes were calculated as the sum of vitamin D intakes from foods and supplements. Daily total vitamin D intakes were compared to the current EAR (10 μg or 400 IU) for North America [7]. Vitamin D intakes, from fortified foods and supplements, above the tolerable upper intake levels for 1‒3 y-old children (European 50 μg or 2000 IU; North American 62.5 μg or 2500 IU) [6,7] are considered at risk of excess intake.

Biochemical assessments

Plasma 25OHD was used as a biomarker for vitamin D status; 91% of total 25OHD is found in D₃ form [22]. A non-fasting venous blood sample was collected in EDTA-coated tubes, and plasma was separated by centrifugation (2000 × g; 15 min; 4°C) within 20 min, aliquots were stored at −70°C until later analyses. Plasma 25OHD concentrations were quantified by stable isotope dilution liquid chromatography (LC)–tandem mass spectrometry (MS) (Waters) as described previously [23,24]. The coefficients of variability were <5%. Biochemical analyses were performed at the Analytical Core for Metabolomics and Nutrition, BC Children’s Hospital Research Institute, Vancouver, Canada.

Vitamin D status classification, using plasma 25OHD concentrations, were categorized using the following definitions: Health Canada, deficient <30 nmol/L and sufficient ≥50 nmol/L [7]; the Endocrine Society, deficient <50 nmol/L and sufficient ≥75 nmol/L [25]; and the Canadian Pediatric Society, deficient <25 nmol/L and sufficient ≥75‒225 nmol/L [26]. A 25OHD concentration above 125 nmol/L has been associated with an increased risk of adverse events [7].

Statistical analysis

All statistical analyses were performed using SPSS Statistics (version 29; SPSS Inc.). A P value ≤0.05 was considered statistically significant. Plasma 25OHD concentrations were normally distributed, and vitamin D intakes were log-transformed before further analyses. Model assumptions were assessed by examining residuals from the fitted models. No differences by treatment group were found in vitamin D intakes or plasma 25OHD concentrations. Differences in the proportion of toddlers meeting current EAR (10 μg or 400 IU) [7] by supplement use were determined by the χ² test. The odds ratios for inadequate vitamin D intake (less than EAR) by supplement use were estimated using logistic regression at each age.

Blood sampling was dichotomized by season into winter (September to February) and summer (March to August). Differences in plasma 25OHD concentrations by potential confounders described in the literature [27], including vitamin D supplement use, season, ethnicity, and weight category, were determined by an independent samples t-test. Differences in the vitamin D status classification by season were determined using the χ² test or Fisher’s test, as appropriate. Associations between predictors and low plasma 25OHD concentrations (<50 nmol/L) were examined using generalized linear models with a Poisson distribution and log link, applying a robust variance estimator to obtain prevalence ratios and 95% confidence intervals (CIs). Separate models were fitted for toddlers aged 1 y and 2 y. Predictors included sex (male/female), vitamin D supplement use (yes/no), season (summer/winter), ethnicity (European/non-European), weight category (normal/overweight or obesity), and total vitamin D intake (more than EAR/less than EAR). Adjusted prevalence estimates for each predictor level were derived from model-based predicted probabilities (estimated marginal means) on the original scale, holding other covariates at their mean values. Relationships between total vitamin D intakes and plasma 25OHD concentrations were determined by Pearson correlation and by analysis of variance adjusted for season and total energy intake [28], where the effect size is reported as partial eta squared (ηp2).

Results

Subject characteristics

A total of 133 toddlers were enrolled in the original study, and details of the participant characteristics are presented in Supplemental Table 1. The mean (SD) age of the toddlers at baseline was 1.12 (0.07) y; 61% were male; 92% were classified in the normal weight category (mean weight-for-age z-score 0.43, stratified by sex); 56% attended the study visit during winter months; 61% were of European ethnicity (parent-reported); 68% of the mothers had a university degree. At the 2-y follow-up, 110 toddlers attended the endline study visit. Complete dietary data were available from 83% (n = 110) of the participants at age 1 y and 78% (n = 86) at age 2 y. Blood samples were available from 93% (n = 124) of the participants at age 1 y and 95% (n = 104) at age 2 y.

Vitamin D intake

Daily dietary intake of energy, macronutrients, and vitamin D is presented in Table 1. Vitamin D intakes from foods were [median (IQR)] 5.2 (3.3‒7.3) μg at age 1 y (n = 110) and 5.4 (3.9‒7.2) μg at age 2 y (n = 86). The main source of dietary vitamin D was fortified foods at age 1 (84%) and 71% at age 2 y, with a smaller contribution from natural sources. At both ages, the main foods contributing to dietary vitamin D intake were dairy products (81% and 73%, at age 1 y and 2 y, respectively), followed by fish (6.8% and 14%), and eggs (4.1% and 5.5%), as presented in Figure 1. In those that consumed these foods, daily intakes were estimated as ∼400 mL of cow milk at both ages; 20 g and 30 g of fatty fish, and 15 g and 20 g of eggs at ages 1 y and 2 y, respectively.

TABLE 1.

Estimated daily intakes of energy, macronutrients, and vitamin D in toddlers¹

Intakes	Age 1 y		Age 2 y
	n	Median (IQR)	n	Median (IQR)
Dietary intakes
Total energy, kcal	110	968 (766‒1105)	86	1083 (919‒1220)
Protein, g	110	43 (34‒50)	86	47 (41‒56)
Carbohydrates, g	110	114 (92‒135)	86	135 (115‒156)
Fat, g	110	36 (29‒47)	86	42 (32‒51)
Vitamin D, μg	110	5.2 (3.3‒7.3)	86	5.4 (3.9‒7.2)
Natural sources, μg	110	0.3 (0.1‒0.7)	86	0.5 (0.2‒1.4)
Fortified foods, μg	110	4.6 (2.7‒6.5)	86	4.3 (2.7‒5.6)
Supplemental intakes
Vitamin D, μg	47	10 (10‒10)	41	7.1 (3.3‒10)
Total intakes
Vitamin D, μg	110	7.9 (4.9‒13)	86	7.2 (4.8‒14)
Vitamin D supplement
Non-users, μg	63	5.4 (3.8‒7.5)	45	5.1 (3.8‒6.5)
Users, μg	47	13 (11‒17)	41	14 (8.0‒18)

Abbreviation: IQR, interquartile range.

¹Data are presented as median (IQR). Dietary intakes were assessed using the average of a 3-d food record, supplement use was assessed by questionnaire, and total intake is the sum of dietary and supplemental intakes. Vitamin D intake IU = vitamin D intake μg × 40.

FIGURE 1.

Top sources of dietary vitamin D intake in toddlers at age 1 y and 2 y. Data are presented as mean percentage of the main food groups contributing to vitamin D for dietary (natural plus fortified), natural, and fortified sources at age 1 y (n = 110, lighter shade) and 2 y (n = 86, darker shade). Dietary intakes were assessed using the average of a 3-d food record.

We further assessed vitamin D supplement use and found that 43% and 48% of toddlers at age 1 y and age 2 y, respectively, were consuming a vitamin D-containing supplement. Daily supplemental vitamin D intakes were 10 μg at age 1 y and 7.1 (3.3‒10) μg [median (IQR)] at age 2 y (Table 1). Most supplements were vitamin D drops (79% and 54%, age 1 y and 2 y, respectively), followed by multivitamins in liquid format (13% and 7.3%), gummies (6.4% and 29%), chewable tablets (0.0% and 2.4%), and >1 format (2.1% and 7.3%). Reported reasons for supplement use included recommendations from a health practitioner (51% and 37%), general health benefits (26% and 24%), other recommendations (11% and 15%), and intake of older siblings (2% and 20%).

Daily total vitamin D intakes (dietary plus supplemental intakes) were [median (IQR)] 7.9 (4.9‒13) μg at age 1 y and 7.2 (4.8‒14) μg at age 2 y (Figure 2). Median total vitamin D intakes were below the EAR (10 μg) at both ages, with 61% of toddlers having inadequate vitamin D intake (less than EAR) at age 1 y and 66% at age 2 y. Supplements contributed to 63% and 53% of total vitamin D intake in toddlers at age 1 y and 2 y, respectively. No toddler consumed above the vitamin D tolerable upper intake levels.

FIGURE 2.

Total vitamin D intake distribution in toddlers at age 1 y and 2 y. Violin plots show the total vitamin D intake (foods plus supplements) as individual data points, median (solid line), and 25th and 75th percentiles (dashed line). Data are presented for all toddlers at age 1 y (n = 110; lighter shades) and 2 y (n = 86; darker shades) and stratified by vitamin D supplement use at each age (age 1 y, supplement users n = 47 and non-users n = 63; age 2 y, supplement users n = 41 and non-users n = 45).

Plasma 25OHD concentrations

At age 1 y (n = 124), plasma 25OHD mean (SD) concentrations were 59.9 (17.6) nmol/L, and 60.9 (13.6) nmol/L at age 2 y (n = 104). At ages 1 y and 2 y, 3.2% and 0% of toddlers had plasma 25OHD concentrations below 25 nmol/L, 6.5% and 1.0% below 30 nmol/L, and 27% and 22% below 50 nmol/L, respectively (Figure 3 [7,25,26]). Few toddlers had plasma 25OHD concentrations above 75 nmol/L, 14% at age 1 y and 13% at age 2 y. No toddler had plasma 25OHD concentrations above 125 nmol/L.

FIGURE 3.

Classification of vitamin D status using different cut-off points in toddlers at age 1 y and 2 y. Data are presented as percentage of participants categorized using plasma 25-hydroxyvitamin D₃ (25OHD) concentrations for all toddlers at age 1 y (n = 124; lighter shades) and 2 y (n = 104; darker shades) and stratified by season at each age (age 1 y, summer n = 57 and winter n = 67; age 2 y, summer n = 52 and winter n = 52). Plasma 25OHD concentrations were quantified by liquid chromatography-tandem mass spectrometry. Plasma 25OHD cut-off points: The Canadian Pediatric Society, deficient <25 nmol/L and sufficient >75 nmol/L [26]. The Institute of Medicine, used by Health Canada, deficient <30 nmol/L and sufficient >50 nmol/L [7]. The Endocrine Society, deficient <50 nmol/L and sufficient >75 nmol/L [25]. Winter months, September to February; Summer months, March to August.

Concentrations did not differ by sex, supplement use, ethnicity, or weight category (all P > 0.05; Table 2) [21]. At age 1 y, 2.3% of vitamin D supplement users compared to 3.7% of non-users had plasma 25OHD concentrations below 25 nmol/L; 4.7% and 7.4%, respectively, had plasma 25OHD concentrations below 30 nmol/L; 35% and 22%, respectively, had plasma 25OHD concentrations below 50 nmol/L; and 21% and 11%, respectively, had plasma 25OHD concentrations above 75 nmol/L. At age 2 y, no subjects had plasma 25OHD concentrations below 25 nmol/L; no vitamin D supplement users and 1.6% of non-users had 25OHD concentrations below 30 nmol/L; 20% and 23%, respectively, had 25OHD concentrations below 50 nmol/L; and 12% and 14%, respectively, had 25OH concentrations above 75 nmol/L.

TABLE 2.

Plasma 25-hydroxyvitamin D concentrations at age 1 y and 2 y, by demographic and lifestyle variables in toddlers¹

Concentrations (nmol/L)	Age 1 y		Age 2 y
	n	Mean (SD)	n	Mean (SD)
Sex
Male	76	59.8 (18.0)	65	61.7 (13.4)
Female	48	60.0 (17.2)	39	59.7 (14.1)
Vitamin D supplement
Non-users	81	59.4 (16.4)	64	59.7 (12.6)
Users	43	60.6 (19.8)	40	62.9 (15.2)
Season2
Winter	67	55.6 (18.7)∗	52	60.5 (14.5)
Summer	57	64.9 (14.9)	52	61.4 (12.8)
Ethnicity
European	75	59.3 (16.5)	71	62.3 (14.1)
Non-European	38	59.7 (19.3)	31	58.3 (12.4)
Weight category3
Normal	113	60.6 (17.4)	93	61.1 (12.7)
Overweight and obesity	10	50.7 (18.7)	8	65.9 (16.3)

Abbreviations: SD, standard deviation; WHO, World Health Organization; 25OHD, 25-hydroxyvitamin D; LC-tandem MS, liquid chromatography-tandem mass spectrometry.

¹Data are presented as mean (SD).

²Winter, September to February; Summer, March to August. Plasma 25OHD concentrations were quantified by LC-tandem MS. ³Weight status was categorized using the WHO child growth charts z-scores stratified by age and sex [21]. Differences by vitamin D supplement use, season, ethnicity, and weight category were determined using an independent samples t-test at each age.

^∗P < 0.01.

At age 1 y, vitamin D concentrations were 14% lower during winter compared to summer (55.6 nmol/L compared with 64.9 nmol/L; P = 0.03); this was not observed at age 2 y (60.5 nmol/L compared with 61.9 nmol/L) (Table 2). At age 1 y, there was a higher prevalence of low vitamin D status (plasma 25OHD <50 nmol/L) in toddlers during the winter months compared to summer (adjusted prevalence 43% compared with 14%; P < 0.007). This was not observed at age 2 y (adjusted prevalence 26% compared with 19%; P = 0.506). The prevalence ratio of being classified with low vitamin D status during the winter months compared to summer was 0.32 (95% CI: 0.14, 0.73), but not at age 2 y (0.73; 95% CI: 0.29, 1.83) (Table 3).

TABLE 3.

Determinants of low plasma 25-hydroxyvitamin D concentrations (<50 nmol/L) in toddlers¹

	Age 1 year			Age 2 years
	Adjusted prevalence, %	Adjusted prevalence ratios (95% CI)	P value	Adjusted prevalence, %	Adjusted prevalence ratios (95% CI)	P value
Sex
Female	22	0.81 (0.44, 1.51)	0.521	29	1.74 (0.75, 4.06)	0.197
Male	27			17
Ethnicity
Non-European	21	0.75 (0.39, 1.45)	0.385	27	1.45 (0.53, 3.9)3	0.479
European	29			18
Weight category
Overweight or obese	29	1.42 (0.67, 3.01)	0.364	25	1.23 (0,21, 7.36)	0.821
Normal	21			20
Season
Summer	14	0.32 (0.14, 0.73)	0.007	19	0.73 (0.29, 1.83)	0.506
Winter	43			26
Vitamin D supplement Use
Yes	35	1.98 (0.96, 4.10)	0..065	23	1.09 (0.41, 2.90)	0.860
No	18			21
Total vitamin D intake
>EAR (10 μg/d)	21	0.32 (0.14, 1.41)	0.393	17	0.59 (0.21, 1.68)	0.319
<EAR (10 μg/d)	30			29

Abbreviations: CI, confidence interval; EAR, estimated average requirement.

¹Prevalence ratios and 95% CIs were obtained from generalized linear models with a Poisson distribution and log link, applying a robust variance estimator. Adjusted prevalence estimates for each predictor level were derived from model-based predicted probabilities (estimated marginal means) on the original scale, holding other covariates at their mean values.

The relationship between plasma 25OHD concentrations and total vitamin D intake is illustrated in Supplemental Figures 1 and 2. A weak positive correlation was found between plasma 25OHD concentrations and total vitamin D intake, r = 0.21 (P = 0.035) at age 1 y and r = 0.23 (P = 0.036) at age 2 y (Supplemental Figure 1). At age 1 y, the correlation was stronger in winter (r = 0.32, P = 0.015) compared to summer (r = 0.03, P = 0.858) (Supplemental Figure 2). At age 1 y, this association was retained after adjusting for season and total energy intake (ηp2 = 0.050; P = 0.023) and was stronger in winter (ηp2 = 0.110; P = 0.012, adjusted model) and not present in summer (ηp2 = 0.000; P = 0.897, adjusted model). At age 2 y, the association was no longer significant after adjusting for season and total energy intake (ηp2 = 0.037; P = 0.087).

Discussion

In this study, we investigated vitamin D intakes and status in healthy toddlers at ages 1 y and 2 y living in Vancouver (latitude 49°N), Canada, and observed the following main findings. At both ages, toddlers consumed similar food sources of vitamin D, but median daily vitamin D intakes were below recommendations. Almost half of the toddlers were consuming a vitamin D supplement; non-users, compared to users, had a higher prevalence of inadequate total vitamin D intakes. Mean 25OHD concentrations were similar at both ages, but vitamin D status differed by season at age 1 y. We observed a positive relationship between plasma 25OHD concentrations and total vitamin D intakes.

The daily median vitamin D intake from foods in our study are consistent with data from other Canadian studies conducted in young children in Montreal (5.9 μg, n = 508, age 2‒5 y) [29] and Nunavik (6.8 μg, n = 217, age 1‒4.5 y) [30], and toddlers in Belgium (5.5 μg, n = 263, age 1‒3 y) [31], and in France (4.6‒5.3 μg, n = 241, age 1‒2 y) [32]. Intakes estimated in our study are lower than the mean dietary intakes reported in toddlers in the United States (7.1 μg, n = 1133, age 1‒2 y) [33]. Even lower daily intakes of vitamin D from foods have been reported in this age group in other countries [14,31,32,[34], [35], [36], [37], [38], [39]], ranging from a daily median of 1.3 μg in Denmark (n = 323, age 3 y) [34] to a mean intake of 3.5 μg in Ireland (n = 467, age 2 y) [14]. Similarly, a lower median intake of vitamin D has been reported in young children in Japan (3.7 μg, n = 235, age 1‒6 y) [40] and in older children in Canada (4.2 μg, n = 2686, age 10‒11 y) [41].

Fortified products were the main source of dietary vitamin D intakes, contributing to >70% in our study, largely from fortified cow milk, and natural sources such as fish and eggs. Comparable food sources have been reported in toddlers from the Montreal study (fortified milk ∼70% and fish ∼12%) [29], and in Ireland (fortified foods 64%, fish 8%, and eggs 5%) [14]. In contrast, a much higher contribution of natural food sources has been reported in Japan, with fish as the main contributor to dietary vitamin D intake (77%), followed by non-fortified milk (9.7%), and eggs (6.5%) [38]. This variation in dietary sources is related to the difference in vitamin D fortification policies between countries [42]. Food fortification with vitamin D is likely the best option for increasing vitamin D intake across the pediatric population, with milk and milk products consumed by toddlers. Furthermore, Canada has recently allowed an increase in the vitamin D food fortification level in milk and margarine, but this policy was not implemented at the time of data collection for this study [43].

The prevalence of vitamin D supplement use in our study was comparable to previous Canadian studies reporting 58% in toddlers (n = 2312, age 1+ y) [44], and 28% (n = 508, age 2‒5 y) [29], and 34% in young children [17]. Current policy on vitamin D supplementation in Canada includes a daily vitamin D supplement of 10 μg for young children receiving human milk [8,9]; in our study, the intake of human milk was limited due to the inclusion criteria of the original trial (≤2 times per day). Vitamin D supplement use in toddlers in the United States has been reported at 21% (n = 1133, age 1‒2 y) [45]. In European countries, vitamin D supplement use in toddlers is much more variable [14,31,32,34,36,37,46], ranging from 7.2% in Germany (n = 164, age 2 y) [46] to 70% (n = 236, age 1‒3 y) in Belgium [31]. These differences can be related to differences in recommendations for vitamin D supplementation and adherence rates [47].

Our findings that the median total vitamin D intakes (from food sources and supplements) were below the recommendations at both ages and that the majority of the toddlers were consuming below the EAR (<10 μg), are in line with other studies from Canada (55% below EAR) [29,41], United States (78% below EAR) [45], and Europe [14,31,48]. As we had expected, supplements, rather than food, provided the majority of the vitamin D in toddlers who used supplements, which is similar to reports by others (58‒62%, age 1‒3 y) [36,49]. Our findings that supplement non-users had inadequate total vitamin D intakes compared to supplement users have also been reported by others [14,29,32] and emphasize the difficulty in meeting vitamin D recommendations from food sources alone in this age group. The toddlers in this study had normal caloric intake and diets with a normal distribution of macronutrients, but despite this, a large proportion had inadequate vitamin D intake. This data, together with previous reports, is supportive of more universal recommendations for vitamin D supplementation in toddlers, irrespective of their breastfeeding status. Furthermore, if adequate vitamin D intake relies on supplementation, this might exacerbate existing health disparities, whereas food fortification of milk and other commonly used foods might be a more equitable way to increase vitamin D intake in all population groups. Recently, Canada approved an increase in the vitamin D food fortification level in milk and margarine [43]. Therefore, future research should evaluate whether this updated fortification policy leads to a higher proportion of the population meeting the recommended vitamin D intake level.

The mean plasma 25OHD concentrations in our study are similar to reports by others in toddlers from several different countries [14,15,17,34,38,[50], [51], [52]]. For example, the rates of deficiency that we observed using either <30 nmol/L or <50 nmol/L are similar to what has been reported by others in this age group in Canada (32% below 50 nmol/L) [18] and in young children (2.1‒4.6% <30 nmol/L and 12‒26% below 50 nmol/L) [17,29,30,53]; and from European reports on toddlers (4‒6% <30 nmol/L and 19‒27% below 50 nmol/L) [14,34,46], and Japan (8.5% <37 nmol/L and 29.6% <50 nmol/L) [38]. Conversely, a lower prevalence of deficiency has been reported in the United States (0.5% <30 nmol/L and 7.1% <50 nmol/L) [54] and higher rates of deficiency in New Zealand (11% <30 nmol/L and 45‒51% <50 nmol/L) [50,51]. In countries closer to the Equator, such as Colombia and Mexico (latitude 4°N and 23°N, respectively), a high prevalence of toddlers with vitamin D concentrations <50 nmol/L has also been reported (42.5% and 24.6%, respectively) [15,55]. The lack of a universal consensus regarding the definition of vitamin D deficiency [13] makes study comparison challenging; reporting the 25OHD cut-off used to define vitamin D status is crucial.

We found higher 25OHD concentrations in summer compared to winter at age 1 y, but not at age 2 y. This difference in seasonal variation could be partially explained by differences in sun exposure by age. It is possible that at age 2 y, toddlers in our study spend more time outdoor during the winter than at age 1 y and therefore had greater endogenous vitamin D synthesis. Consistent with our findings at age 1 y, other studies have reported seasonal variation in vitamin D status in toddlers, including Ireland (Cork, latitude 51°N; winter 54.5 nmol/L compared with summer 71.2 nmol/L) [14], Denmark (Copenhagen and Frederiksberg, latitude 55°N, winter 58.9 nmol/L compared with summer 77.0 nmol/L) [34], and the Netherlands (Utrecht, latitude 52°N; winter 62.2 nmol/L compared with summer 96.2 nmol/L) [27]. A study in Japan also reported higher 25OHD concentrations in summer, 82.4 nmol/L compared with 62.4 winter nmol/L (Tokyo, latitude 36°N) [56]. These observations support the usefulness of assessing vitamin D status stratified by season during early childhood, as the risk of vitamin D deficiency during limited sunlight exposure months may be overlooked.

Only a few studies have reported on the relationship between vitamin D intake and status in toddlers [14,29,34,38,55]. We found a positive relationship between 25OHD concentrations and total vitamin D intakes, which were stronger in winter than in summer at age 1 y. Similar findings have been reported in previous studies in Canada [57] and in Ireland [14]. In contrast, no association between vitamin D intakes and biomarkers was reported in Denmark [34] and in Japan [38], which could in part be related to the use of only dietary intake data (not total intakes of vitamin D). Moreover, the lack of association between vitamin D status and supplement use in our study could be due to other factors, such as parents providing vitamin D supplements less frequently than reported, time spent outdoors (proxy for sun exposure) not collected, and blood samples not collected in the fasting state.

We found the risk of vitamin D deficiency is common and may not be prevented by current dietary vitamin D intakes, particularly in winter seasons at age 1 y. However, it is important to note that estimates of inadequate vitamin D intakes are often higher than those of vitamin D deficiency [34]. Indeed, we found that 61‒66% of toddlers had vitamin D intakes below the EAR compared to plasma 25OHD concentrations of only 1‒7% below 30 nmol/L, 22‒27% below 50 nmol/L, and 86‒87% below 75 nmol/L. This disagreement may be related to the contribution of sunlight exposure to 25OHD concentrations.

The strengths of our study include that it is the first report exploring the relationship between vitamin D dietary intakes and vitamin D status (plasma 25OHD) in toddlers in Canada. We collected information between January 2010 and September 2014, which allowed us to stratify data by season. We collected dietary information and estimated intakes using the average of 3-d food recalls, which reflect usual intakes and allow for intake assessment. We collected data on the prevalence of vitamin D supplement use and daily dose, enabling the estimation of total nutrient intakes. Furthermore, we analyzed vitamin D status by quantifying plasma 25OHD by LC-MS/MS, which is considered the gold standard method. However, our findings must be interpreted with consideration of our study limitations. We used a convenience sample, which may not be representative of the Canadian population, highlighting the need for updated monitoring in nationally representative samples. The study did not collect information on sunlight exposure, skin pigmentation, or use of sunscreen, which may have influenced plasma 25OHD concentrations. Dietary information was reported by parents and/or caregivers, including limitations of self-report bias and the accuracy of the nutrient databases. Moreover, non-fasting venous samples were collected, which may influence the plasma 25OHD concentration. However, this is standard clinical practice aimed at minimizing the stress associated with fasting and venous blood draws in toddlers.

In summary, our findings suggest that in this sample of healthy Canadian toddlers, many are not meeting the dietary vitamin D intake recommendations, and some children may be at higher risk of vitamin D deficiency during the winter. Around half of the participants were consuming a vitamin D supplement. In addition, we found that total vitamin D intakes are reflected by plasma 25OHD at age 1 y, but not at age 2 y. These results highlight the need to conduct regular assessments of nutritional status in nationally representative samples across multiple age groups to monitor nutrients that may be difficult to obtain sufficiently through the diet, like vitamin D. Further research should assess whether the latest increase in the Canadian vitamin D food fortification policy will positively impact the percentage of the population meeting the vitamin D intake recommendations and improve vitamin D status.

Author contributions

The authors’ responsibilities were as follows– AMW, AMD: designed and conducted the research; AMW, VML: analyzed the dietary and supplement intake data; AMW: performed the vitamin D biomarker of status analysis, validated the database, conducted the statistical analyses, and drafted the manuscript; TJG, CP, RE, AMD: contributed to data interpretation and manuscript revision; AMD: had primary responsibility for the final content; and all authors: read and approved the final manuscript.

Data availability

Data described in the manuscript will be made available upon request, pending application and approval. Request to access the dataset should be directed to AMD (email: adevlin@bcchr.ubc.ca).

Funding

This research was supported by an Egg Farmers of Canada grant to AMD, RE, and AMW. DSM Nutritional Products sponsored the original intervention trial but had no role in the study design and data analyses. AMW was supported by a Bertram Hoffmeister Postdoctoral Fellowship from BC Children’s Hospital Research Institute and Society to Cell Clyde Hertzman Memorial Fellowship from the University of British Columbia Social Exposome Cluster and the Human Early Learning Partnership. AMD, RE, and CP are each supported by an Investigator Grant from BC Children’s Hospital Research Institute.

Conflict of interest

AMD is an editorial board member for the Journal of Nutrition and played no role in the Journal’s evaluation of the manuscript. All other authors report no conflicts of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article: