Vitamin D status and associated factors of deficiency among Jordanian children of preschool age

Abstract

BACKGROUND/OBJECTIVES:

Vitamin D deficiency in children remains a global concern. Although literature exists on the vitamin D status and its risk factors among children in the Middle East, findings have yielded mixed results, and large, representative community studies are lacking.

SUBJECTS/METHODS:

In a nationally representative survey of 1077 Jordanian children of preschool age (12–59 months) in Spring 2010, we measured 25(OH)D₃ concentrations by liquid chromatography–tandem mass spectrometry and calculated prevalence ratios for deficiency associated with various factors.

RESULTS:

Results showed 19.8% (95% confidence interval (CI): 16.4–23.3%) deficiency (<12 ng/ml) and 56.5% (95% CI: 52.0–61.0%) insufficiency (<20 ng/ml). In adjusted models, prevalence of deficiency was higher for females compared with males (prevalence ratio (PR) = 1.74, 95% CI: 1.22–2.47, P = 0.002) and lower for children 24–35 months of age (PR = 0.64, 95% CI: 0.44–0.92, P = 0.018) compared with children 12–23 months of age. In rural areas, there was no difference in prevalence of vitamin D deficiency between those whose mothers had/did not have vitamin D deficiency (P = 0.312); however, in urban areas, prevalence of vitamin D deficiency was 3.18 times greater among those whose mothers were vitamin D deficient compared with those whose mothers were not deficient (P = 0.000).

CONCLUSIONS:

Vitamin D deficiency and insufficiency pose significant public health problems in Jordanian children with female children disproportionately affected. Strong associations between vitamin D status in children and urban residency and maternal vitamin D status suggest that the behaviors related to sun exposure in urban mothers likely also affect the sun exposure and thus vitamin D status of their children.

INTRODUCTION

The global burden of vitamin D deficiency in children¹ remains a concern because of the important role vitamin D plays in skeletal health, including bone mineral content and density, calcium absorption, and the prevention of rickets.² As ultraviolet B radiation in sunlight is the primary source of vitamin D,³ those at highest risk for deficiency are individuals with limited sun exposure,⁴ including people with dark skin pigmentation that inhibits ultraviolet B radiation absorption⁵ and urban residents where air pollution decreases levels of ultraviolet B photons.^6–8 Despite the high potential for sun exposure in the Middle East, cultural practices that limit sun exposure, breastfeeding in the absence of supplementation⁹ and a lack of weaning foods containing vitamin D increase risk for vitamin D deficiency in children.

Several studies have demonstrated high prevalence of rickets and hypovitaminosis D among children in this region. In a study of children in Turkey, 6% of children admitted to an outpatient clinic or health-care center were rachitic.¹⁰ And in rural north Yemen, 27% of children under 5 years attending an immunization clinic were diagnosed with rickets.¹¹ Among children <6 years in a large study across Saudi Arabia, approximately 23% had a serum concentration of 25-hydroxyvitamin D [25(OH)D] of 5–10 ng/ml, indicating deficiency;¹² in a separate study in urban Saudi Arabia, 9.0% of randomly selected children aged 12–72 months had serum 25(OH)D concentrations <10 ng/ml.¹³ In urban Jordan, 10.6% of children aged 3 months to 2 years of age who were admitted to hospitals for acute illnesses were diagnosed with nutritional rickets.¹⁴ In a convenience sample of children aged 4–5 years from urban/surburban areas in north Jordan, 39% of children had vitamin D insufficiency (25(OH)D<20 ng/ml).¹⁵

An extensive literature review by Thacher et al.¹⁶ describes the connection between low serum vitamin D concentration and rickets in the Middle East. The connection was further demonstrated in a study of children 6–48 months by Baroncelli et al.,⁹ who found that 71% of Egyptian patients with rickets had vitamin D deficiency (25(OH)D<15 ng/ml) compared with 42% of controls, while 86% of Turkish patients with rickets had vitamin D deficiency compared with 27% of controls. Similarly, in a clinic-based case comparison study among children 5–35 months of age in the United Arab Emirates, Dawodu et al.¹⁷ showed presence of vitamin D deficiency in over 90% of rachitic children and their mothers compared with only 22 and 51% of nonrachitic children and their mothers, respectively (P = 0.001).

These studies and others have noted various associated factors of low vitamin D levels and rickets in children in the Middle East, including female sex,^13,18 younger age,¹³ breastfeeding without vitamin D supplementation,^9,14,19,20 presence of anemia,¹⁴ higher maternal education level⁹ and low levels of maternal serum vitamin D.^9,20,21 However, although literature exists on vitamin D status and its risk factors among children in the Middle East, findings have yielded mixed results, and large, representative population studies are lacking. Furthermore, laboratory testing has often used methods which are considered sub-optimal (for example, radioimmunoassay, enzyme immunoassay and competitive protein-binding assay).²²

This paper presents findings from the first nationally representative survey of vitamin D status among women (15–49 years) and children (12–59 months) in Jordan.²³ The survey was conducted, in part, as a baseline assessment for monitoring the fortification of flour with vitamin D in Jordan. We used liquid chromatography-–tandem mass spectrometry, a gold-standard laboratory method,²² to measure vitamin D status and examine factors associated with vitamin D deficiency among preschool-aged children and their mothers in Jordan. A previous publication provides results of vitamin D status and associated factors of deficiency in women.²⁴ Briefly, 60.3% (95% confidence interval (CI): 57.1–63.4%) of women were deficient (<12 ng/ml), and prevalence of deficiency was higher among women who covered their skin, lived in urban areas or had completed at least secondary education. This paper presents the findings of vitamin D status and associated factors of deficiency among children.

SUBJECTS AND METHODS

Survey population and sampling design

Survey staff collected data during a national household-based micronutrient survey of women (15–49 years) and children (12–59 months) in Jordan during March–April 2010, when historical UV Index averages range between 7 and 10 (high to very high UV exposure levels).²⁵ Adults/caregivers provided informed consent verbally for their participation and for the participation of their minor children. The Al Basheer Hospital Human Resources Committee provided ethical approval; the CDC Institutional Review Board deemed the survey public health practice and therefore did not review the protocol. Survey staff gathered a sample of blood from each participant and applied a questionnaire to mothers or caregivers regarding themselves and/or their minor child. Survey staff also measured the height/length and weight of all participating children.

We calculated sample size estimates for the primary survey objectives. Assuming 0.6 children (12–59 months) and 1.4 women (15–49 years) in each household, a design effect of 2.0, and an 80% participation rate, we calculated that 166 clusters with 12 households in each cluster (1992 households) were needed. This sample size was sufficient to estimate child vitamin D deficiency with 5% precision.

The survey team used multistage stratified cluster sampling methods to select households from the sampling frame developed from Jordan’s 2004 Population and Housing Census. The sampling frame did not include those residing in remote areas (mostly nomads), individuals living in collective dwellings (for example, hotels, work camps, hospitals and prisons) or non-Jordanian households. The census defined households as Jordanian if the head of household stated that s/he was Jordanian. Using probability proportionate to size sampling, the survey team selected 166 clusters across 30 strata. In each selected cluster, the team then enumerated and mapped Jordanian households and systematically selected 12 households for participation. From participating households, the survey team recruited all women (15–49 years) and children (12–59 months) to participate. Detailed methods are described elsewhere.²³

Biochemical testing

A trained phlebotomist collected venous blood from the forearm of each participant using EDTA tubes for hematological analysis and plain plastic colorless vacuum tubes for analysis of serum 25(OH)D₃. The field team transported the samples (4–10 °C in a cold box with frozen gel packs) to Jordan’s Central Public Health Laboratory (CPHL) for the measurement of hemoglobin as part of a complete blood count (Beckman Coulter Inc., 2003), and centrifugation, aliquoting and freezing of serum until further analysis of vitamin D. The CPHL processed all samples within 24 h of collection, and the staff maintained a cold chain from collection to processing. The Jordan University of Science and Technology (JUST) measured serum 25(OH)D₃ concentrations with liquid chromatography–-tandem mass spectrometry method,²⁶ a gold-standard of measurement.²² The interassay coefficient of variation was 2%, and the limit of detection was 1.0 ng/ml. Results from participation in the CDC Vitamin A Laboratory – External Quality Assurance (VITAL-EQA) program, an external quality assurance program for a number of micronutrients including 25(OH)D,²⁷ showed excellent precision and minimal bias.

Cutoffs of 25(OH)D used for the analysis were selected based on existing literature and the Institute of Medicine’s report on vitamin D:² severely deficient (<5 ng/ml), deficient (<12 ng/ml) and insufficient (<20 ng/ml). Anemia was defined as hemoglobin (Hb) <11 g/dl.²⁸

Covariates associated with deficiency

In addition to anemia and maternal vitamin D status, we also investigated the association of several other covariates with vitamin D deficiency in children, including sex, residence (urban or rural), reported age of child in months (12–23, 24–35 and 36–59), age of the child when first given something other than breast milk to consume (< 3, 3 to<6 and ⩾ 6 months), maternal education (less than secondary versus secondary or higher) and weight status (underweight/normal weight and overweight/obese) based on body mass index (BMI) for age Z-score.

To calculate BMI, field staff measured barefoot height or length to the nearest 0.1 cm using a portable wooden Shorr stadiometer (Olney, MD, USA). The age of the child was calculated based on the difference between the child’s birth date and the date of the measurement. Field staff measured standing height for children older than 24 months of age and recumbent length for children ⩽ 24 months of age. Field staff measured body weight of the children to the nearest 0.1 kg using UNICEF Seca Uniscales (Birmingham, UK). The weight of children who could not stand on their own was assessed using the mother–child tare function on the scale. We calculated BMI-for-age Z-scores as an indicator of overweight using the most recent World Health Organization (WHO) international growth reference/standard.²⁶ For analysis, BMI-for-age Z-score was dichotomized as underweight/normal (⩽ 2) or overweight/obese (> 2).²⁷

Statistical analysis

We reviewed biochemical and measurement data for physiologic plausibility. Values were considered missing if they were outside of the following cutoffs: 25(OH)D₃ (>76.8 ng/ml) (based on the maximum concentration measured in 2.5 years of U.S National Health and Nutrition Examination Survey data); Hb (4.0–18.0 g/dl);²⁹ BMI-for-age Z-score (−5.0 to 5.0).³⁰ The analysis team calculated each cluster’s response rate to account for nonresponse and sample weights for stratification. Except where noted, all analyses presented take into account the complex sample design with sample weights. For vitamin D status of children, we calculated the percent severely deficient (<5 ng/ml), deficient (<12 ng/ml) and insufficient (<20 ng/ml), and the 95% CI.

We examined covariates associated with vitamin D deficiency in Jordanian children of preschool age by calculating prevalence, prevalence ratios (PR) and 95% CI for covariates. We followed a five-step process using binomial regression:³¹

1. we identified covariates with an independent association with vitamin D deficiency and retained covariates with a moderately significant univariate relationship with vitamin D deficiency (P<0.25);

2. we assessed all two-way interactions among covariates retained in step 1 (P<0.05);

3. we created a multivariable model with interaction terms identified in step 2 and their main effects and all covariates retained in step 1 and then used a backward elimination method to remove covariates that were nonsignificant (P>0.10) and not confounders of other effect measures in the model (<15% change to other effect measures);

4. in the final model, we used a standard statistical significance level of P<0.05 to determine significance of interaction terms and covariates; and

5. we calculated Cramer’s V for categorical covariates and phi coefficients for dichotomous covariates to assess collinearity.

Observed interactions were further explored by restricting the analysis to relevant subgroups defined by the interaction terms. We used Stata 12.0 s.e. (StataCorp, College Station, TX, USA, 2011) for statistical analysis.

RESULTS

Survey staff invited a total of 1992 heads of household to participate. Of these, 1741 (87.4%) heads of households agreed to participate, 157 (7.9%) refused and 94 (4.7%) were not available at the time of the visit. Of the households that agreed to participate, survey staff invited, via the caregiver, 1090 eligible children to participate. Caregivers agreed to complete a questionnaire for 1077 (98.8%) children. Among these children, 947 (87.9%) caregivers also agreed to allow the survey team to obtain a sample of the child’s blood. Complete data for assessing vitamin D status were available for 915 children and for assessing covariates associated with deficiency for 734 children (see Figure 1). A blood sample for assessing vitamin D status of mothers was available for 84.4% (n = 772) of mothers.

Figure 1.

Flow chart of participation in the 2010 Jordan National Micronutrient Survey.

All children for whom participation was refused (n = 13) lived in an urban area. Table 1 compares characteristics of participants who provided a blood sample with characteristics of those whose caregiver refused for their child to provide a blood sample but completed a questionnaire.

Table 1.

Demographic characteristics of Jordanian participants aged 12–59 months by whether they provided a blood sample, 2010

Characteristic	Provided blood (%) n = 915	Refused blood (%) n = 130	Chi-square P-valuea
Sex			0.27
Male	51.6	56.2
Female	48.4	43.8
Residence			0.81
Rural	21.2	21.5
Urban	78.8	78.5
Age (months)			0.10
12–23	23.9	26.9
24–35	23.2	30.0
36–59	52.9	43.1
Age child first given something other than breast milkb			0.07
< 3 months	15.3	15.5
3 to < 6 months	53.5	43.4
6+ months	31.2	41.1
BMI-for-age Z-scorec			0.39
Underweight/normal (⩽ 2)	92.2	88.4
Overweight/obese (> 2)	7.8	11.6
Anemia (Hb < 11 g/dl)d			N/A
Normal	83.0	N/A
Anemic	17.0	N/A
Maternal level of education completede			0.23
Below secondary	49.6	57.8
Secondary and above	50.4	42.2
Maternal vitamin D deficiencyf			0.33
Normal	39.4	31.8
Deficient	60.6	68.2

Abbreviation: BMI, body mass index. NOTE: Percentages unweighted.

^aAdjusted for cluster survey design.

^bn = 914 for children who provided blood and n = 129 for children who refused blood.

^cn = 902 for children who provided blood and n = 86 for children who refused blood

^dn = 880 for children who provided blood.

^en = 885 for children who provided blood and n = 128 for children who refused blood.

^fn = 772 for children who provided blood and n = 63 for children who refused blood.

All observed 25(OH)D₃ values were within the range of physiologically plausible values, thus, we did not drop any observations from analysis. Figure 2 depicts the unweighted frequency distribution of serum 25(OH)D₃ concentration among all children (n = 915); the overall median 25(OH)D₃ concentration was 18.2 ng/ml with an inter-quartile range of 12.8–24.3 ng/ml. On the basis of weighted data, more than half of children (56.5%, 95% CI: 52.0–61.0) had 25(OH)D₃ concentrations <20.0 ng/ml; 19.8% (95% CI: 16.4–23.3) of children had 25(OH)D₃ concentrations <12.0 ng/ml; 1.5% (95% CI: 0.7–2.3) of children were severely deficient, with 25(OH)D₃ concentrations <5.0 ng/ml.

Figure 2.

Unweighted frequency distribution of serum 25(OH)D₃ concentration among Jordanian children (12–59 months old, n = 915), 2010.

Among the covariates examined for association with vitamin D deficiency in children, we recorded a missing value for one child who had an implausible Hb measure (Hb>18.0 g/dl).²⁹ For BMI-for-age Z-score, we recorded missing values for four otherwise eligible children because of extreme measures (Z-score>5.0).³⁰

A comparison of prevalence of deficiency across covariates (Table 2) showed statistically significant differences in the prevalence of vitamin D deficiency (25(OH)D₃<12 ng/ml). Prevalence was lower among males, children in rural areas, 24–35-month olds and children with non-vitamin-D-deficient mothers. We observed no collinearity between the covariates in the final regression model. In our final adjusted model, the prevalence of vitamin D deficiency was 1.74 times higher for females (95% CI: 1.22–2.47) than for males (P = 0.002). Compared with children 12–23 months of age, children 24–35 months had a lower prevalence of deficiency (PR = 0.64, 95% CI: 0.44–0.92, P = 0.018). However, there was no difference between the youngest and oldest age groups. Adjusted Wald test P = 0.046 confirmed joint significance of the categories of the age variable (Table 3).

Table 2.

Prevalence and unadjusted PR for vitamin D deficiency (25[OH]D₃<12.0 ng/ml) by select characteristics for children (n = 915) aged 12–59 months, Jordan 2010

Characteristic	N	% Vitamin D deficient (95% CI)	Unadjusted/univariate
			PR (95% CI)	P-valuea
Sex
Male	472	14.0 (10.9, 17.7)	1.00
Female	443	25.9 (16.6, 23.5) P < 0.001b	1.86 (1.37–2.51)	< 0.001
Residence
Rural	194	13.6 (9.5–19.1)	1.00
Urban	721	21.4 (17.5, 25.8) P = 0.024b	1.57 (1.05–2.35)	0.028
Age (months)
12–23	219	24.1 (18.8–30.3)	1.00
24–35	212	14.6 (10.3–20.2)	0.61 (0.42–0.87)	0.008
36–59	484	20.2 (15.9–25.3) P = 0.048b	0.84 (0.60–1.16)	0.286
Age child first given something other than breast milk
< 3 months	140	18.2 (11.6–27.3)	1.00
3 to < 6 months	489	20.1 (15.8–25.1)	1.10 (0.70–1.74)	0.667
6+ months	285	19.9 (15.4–25.4) P = 0.906b	1.10 (0.68–1.77)	0.709
BMI-for-age Z-score
Underweight/normal (⩽ 2)	832	19.7 (16.4–23.5)	1.00
Overweight/obese (> 2)	70	22.2 (14.4–32.6) P = 0.570b	1.13 (0.75–1.70)	0.565
Anemia (Hb < 11 g/dL)
Normal	730	18.5 (15.0–22.7)	1.00
Anemic	150	24.1 (18.1–31.4) P = 0.121b	1.30 (0.94–1.82)	0.117
Maternal level of education completed
Below secondary	439	22.2 (17.5–27.7)	1.21 (0.87–1.69)	0.261
Secondary and above	446	18.3 (14.1–23.4) P = 0.260b	1.00
Maternal vitamin D deficiency
Normal	304	10.2 (7.0–14.7)	1.00
Deficient	468	26.4 (21.7–31.8) P < 0.001b	2.59 (1.70–3.95)	< 0.001

Abbreviations: CI, confidence interval; PR, prevalence ratio. Prevalences and PR weighted and CI adjusted for sample design. Count for each characteristic may not add to 915 because of missing data.

^aP-value for the subgroup difference in prevalence ratio.

^bP-value for the subgroup difference in prevalence of vitamin D deficiency.

Table 3.

Adjusted PR for vitamin D deficiency (25[OH]D3<12.0 ng/ml) by select characteristics for children (n = 772^a) aged 12–59 months, Jordan 2010

Characteristic	Adjusted PR
	PR (95% CI)b	P-value
Sex
Male	1.00
Female	1.74 (1.22–2.47)	0.002
Age (months)
12–23	1.00
24–35	0.64 (0.44–0.94)	0.018c
36–59	0.91 (0.66–1.26)	0.564c
Residence and maternal vitamin D deficiency
Rural
Mother vitamin D normal	1.00
Mother vitamin D deficient	0.68 (0.32–1.44)	0.312
Urban
Mother vitamin D normal	0.53 (0.25–1.12)	0.095
Mother vitamin D deficient	1.68 (0.91–3.09)	0.098

Abbreviations: CI, confidence interval; PR, prevalence ratio. PR weighted and CI adjusted for sample design.

^aFinal analytic sample size.

^bPR adjusted for all characteristics shown.

^cAdjusted Wald test P = 0.046 indicating joint significance of age categories.

We observed an interaction between urban/rural residence and maternal vitamin D deficiency. To further investigate the interaction, we reran the analysis restricted first to only children in rural areas and then to only children in urban areas. Among children living in rural areas, we found no association between low maternal vitamin D status and low child vitamin D status after controlling for other variables (PR = 0.61, 95% CI: 0.23–1.61, P = 0.306). However, we did observe a significant association between low maternal vitamin D status and low child vitamin D status among children living in urban areas (PR = 3.18, 95% CI: 1.85–5.46, P = 0.000).

DISCUSSION

In our representative survey of Jordanian children, one in five were found to be deficient and nearly three in five have insufficient levels of vitamin D. As few other studies in the Middle East have used liquid chromatography–tandem mass spectrometry methodology for measuring 25(OH)D₃ status, as was used in the present survey, interassay differences must be considered when results are compared with those of other studies.^7,32 high-performance liquid chromatography methods (used by Olang et al.)¹⁸ have been shown to approximate liquid chromatography–tandem mass spectrometry results, whereas radioimmunoassay (used by Bahijri et al.),¹³ enzyme immunoassay (used by Gharaibeh and Stoeker)¹⁵ and competitive protein-binding assay (used by Sedrani et al.)¹² are known to underestimate 25(OH)D concentrations.²² Also, the majority of previous studies were not nationally representative and were conducted in hospital rather than community settings.

Insofar as findings can be related to other studies, the moderately high prevalence of deficiency and insufficiency in children in the present survey was higher than that of a convenience sample of children from community centers in urban/suburban areas in north Jordan;¹⁵ similar to that among a national sample of children in Saudi Arabia (deficiency);¹² and about 21 percentage points higher (insufficiency) than that of a national sample of 15–24-month olds in Iran.¹⁸ This observation may be attributed to the higher distribution of urban children (78.8%) included in the present survey compared to the Iran study (50.7%).

In a global review of vitamin D status, Mithal et al.¹ report that 8% of US children aged 1–5 years have insufficient levels (<20 ng/ml). This compares to 56.5% of Jordanian children in the present survey. Although direct comparisons across countries are difficult because of differences in definitions and age categories, this global review generally indicates that vitamin D deficiency is more common in all age groups in the Middle East and South Asia, compared to other regions.

Higher deficiency among female children compared with males is consistent with a study showing lower vitamin D concentrations among preschool-aged female children in health centers throughout Iran.¹⁸ Bahijri¹³ also observed this sex difference and noted that preschool-aged boys in urban Saudi Arabia had a sun exposure time of more than 2h per day versus 20 min for girls.

Regarding age, the lowest levels of deficiency were observed among children 24–35 months, a finding also observed by Bahijri¹³ who noted an increased exposure to sunlight and fewer diarrheal attacks compared with younger age groups. Bahijri¹³ observed no significant differences in vitamin D intake across the age groups of children 12–72 months, and the relatively lower levels of vitamin D among the older age group were attributed to social customs keeping older girls more at home.

Although literature outside of the Middle East^6–8 has cited urban residence as a factor associated with a higher prevalence of vitamin D deficiency, this observation appears to be the first among children in the Middle East—likely because of the inability to make urban/rural comparisons in previous studies. The observation that only urban children of mothers with deficiency were more likely to be deficient, compared with urban children of mothers who were not deficient, suggests that maternal behaviors, sun exposure and other factors that impact the vitamin D status of mothers in urban areas may also have a similar impact on the vitamin D status of their children.

In a study of preschool-aged children in north Jordan, Gharaibeh and Stoecker¹⁵ observed a negative correlation between family income and vitamin D levels, citing the fact that children of lower socioeconomic status are more likely to have increased sun exposure from spending more time outdoors because of poor housing conditions. However, the only available socioeconomic indicator in the present study was maternal education level, which did not show a statistically significant association with child vitamin D status.

Breastfeeding is another factor related to maternal vitamin D status, where prolonged exclusive breastfeeding without a supplemental vitamin D source can increase risk of deficiency, because human milk contains only about one-tenth (0.3–0.8 ug/l) of the infant’s need of vitamin D.³³ Also, low vitamin D concentrations in mothers can further impact the low concentration of vitamin D in breast milk.³⁴ Although the present survey focused on children older than 12 months, the combination of a high prevalence of breastfeeding in Jordan (97% of children in this study) and a prevalence of 60.3% (95% CI: 57.1–63.4%) deficiency among Jordanian women of reproductive age²⁴ heightens the concern of exclusive breastfeeding without vitamin D supplementation among infants in Jordan. The age the child when first given something other than breast milk does not show a statistically significant association with deficiency in the present survey, though information available to completely assess breastfeeding behaviors was limited. However, collective findings of the survey may point to a lack of alternative dietary sources of vitamin D in Jordan, further emphasizing the importance of improving the availability of infant and child foods and supplements containing vitamin D.

Despite research demonstrating decreased bioavailability of vitamin D in obese individuals, which might contribute to lower vitamin D concentrations,³⁵ we observed no association between BMI-for-age Z-score and vitamin D deficiency, which is consistent with findings of Cole et al.³⁶ The lack of association in our study may be explained by the decision to group overweight and obesity because of small numbers of obese children (BMI-for-age Z-score>3; n = 18).²⁷

Limitations of the present survey include a lack of information on serum 25(OH)D₂ concentrations and other factors that affect vitamin D concentration, including dietary intake and supplementary feeding, individual sun exposure and behavior related to sun exposure, medication use, medical conditions, and skin color. It would be advisable for future studies to include these factors, especially information on individual sun exposure. Numbers for the multivariate analysis were limited because of missing data for various variables, including 13 missing BMI-for-age Z-score; 35 missing Hb; 143 missing maternal vitamin D status; and 30 missing maternal education level (Figure 1). The majority of missing data is likely because of invasiveness of blood collection procedures; the large reduction in the number of children available for analysis because of missing maternal vitamin D status may be a source of bias.

This first nationally representative survey on vitamin D status among preschool-aged children in Jordan can serve as baseline data for Jordan’s wheat flour fortification program. Following the micronutrient survey, in June 2010, Jordan’s Ministry of Health added vitamin D to the panel of micronutrients included in its wheat flour fortification. Two μg (80 IU) of dry vitamin D₃ is added for each 200 g serving of flour (or 300 g serving of bread), meeting 40% of the WHO recommended nutrient intake of 200 IU daily for vitamin D³⁷ for preschool-aged children. Considering the high levels of deficiency in Jordanian children and women of reproductive age,²⁴ and the updated Institute of Medicine recommendations that call for 400–600 IU daily to meet the needs of children aged 1–18 years, fortification levels should be reviewed and the impact of fortification on population vitamin D concentrations should be evaluated.^2,38