Wintertime Vitamin D status and its related risk factors among children living in Auckland, New Zealand

Vitamin D is a secosteroid and an essential nutrient, which has a crucial role in the absorption of calcium from the intestine, regulation of serum calcium and bone health.1 Observational and clinical studies suggest there is a relationship between vitamin D deficiency and insufficiency and skeletal health problems (eg, rickets, metabolic bone disease and hypocalcaemia) during childhood.2,3 Therefore an adequate 25(OH)D concentration is considered important for ensuring bone health during childhood and later in life.

In New Zealand, a small quantity of vitamin D comes from a limited number of foods that naturally contain vitamin D (salmon, herring, tuna and mackerel). There is a limited range of foods including milk and yoghurt, which are sometimes fortified with vitamin D,4 although there is no mandatory fortification in New Zealand.4 Therefore, it would be hard to reach acceptable blood levels of vitamin D through diet alone. However, endogenous synthesis through sunlight exposure is the major source of vitamin D for most people living in New Zealand since diet alone is not adequate to meet recommended 25(OH)D concentrations.4 Nevertheless, the efficacy of cutaneous synthesis of vitamin D can be influenced by various factors including geographic latitude, season, time of day,5 ethnicity,6 obesity,7 waist circumference,8 age9 and gender.10 Furthermore, some other conditions such as skin pigmentation, use of sunscreens, exposed body surface and exposure duration,11 and restricted sunlight exposure habits (eg, clothing)12 can affect the cutaneous vitamin D synthesis.

Unfortunately, there are limited data available regarding the vitamin D status and risk factors for vitamin D deficiency in New Zealand children.6,13,14 Therefore, the aims of the present study were to assess wintertime vitamin D status in New Zealand children living in Auckland and to identify related risk factors for deficiency.

Methods

Study participants

In this cross-sectional study, we recruited children aged 8–11 years from six Auckland, New Zealand primary schools (one from north, two from east, two from south, and one from central) (in August 2016 and 2017—late winter in southern hemisphere). We originally approached schools through a collaboration of primary school science teachers and asked for expressions of interest. We then endeavored to recruit schools specifically to include a range of sociodemographic levels and ethnicities. The study protocol was approved by the Human Ethics Committee of Massey University (Southern A; approval no. MUHECN 16/42). All children and their parents provided written informed consent prior to participating in the study. The study was run in agreement with the Declaration of Helsinki.15 Children who were apparently healthy were sought. Children were ineligible when they met the following exclusion criteria 1) a history of any disease affecting vitamin D metabolism (eg, cardiac, kidney or liver disease) or 2) a history of any long-term medication use (eg, steroids) 3) having had any surgical implants, metal screws or similar, or 4) having a cast. All data collection from the children took place at their schools on one occasion.

Data collection

Sample size

Children were stratified by gender (two groups), ethnicity (six categories), and skin colour (four groups) and logistic regression used to determine contribution of risk factors for vitamin D deficiency. A sample of 10–15 per factor per group is the standard requirement for regression analysis. Therefore: 2*6*4*10–15 = 480–720 participants.16

Anthropometric measurements

Information about participants’ weight and height were collected. Children were asked to remove their shoes to measure their height (to an accuracy of 0.1cm) using a Seca 213 portable stadiometer. Body weight was recorded in minimal clothing to the nearest 0.1kg through InBody 230 (Biospace Co. Ltd., Seoul, Korea). Body mass index (BMI) was derived by dividing weight (kg) by the square of height (m2). We calculated BMI adjusted-for-age as described by Cole et al (2000 and 2007).17,18 Percentage of body fat (%BF) was measured by bioelectrical impedance analysis (BIA), the InBody 230 (Biospace Co. Ltd., Seoul, Korea) according to the standard procedure provided by the manufacturer. The Inbody 230 has been validated in the same population (the abstract in proceedings).

Healthy and demographic questionnaires

Parents were asked to identify their child’s ethnic group. Ethnicity was categorised as 1) European New Zealanders, 2) Māori, 3) Pacific, 4) South Asian and 5) Chinese/Korean/Southeast Asian, and 6) other ethnicity. If participants reported more than one ethnic group, then based on the priority system,19,20 the child was assigned to one ethnic groups. The following ethnicity prioritisation was used for analytical purposes: Māori>Pacific>South Asian> Chinese/Korean/Southeast Asian>European New Zealanders>Other. For instance, if Māori was one of the reported groups, then the child was allocated to Māori group. Parents were asked to classify their child’s skin colour. Skin colour was categorised as ‘fair’, ‘medium’, ‘olive’ or ‘dark/brown’.

Information about the sunlight exposure, applying sunscreen, and the frequency of using sunscreen was obtained from a self-reported parents’ questionnaire. Parents were asked to identify which parts of the body were usually exposed to sunlight, and what the usual sunscreen usage was.

Physical activity was evaluated with the short version of the “International Physical Activity Questionnaire (IPAQ)”.21 Three types of activity were assessed: walking, moderate-intensity and vigorous-intensity activities. Data collected using IPAQ were categorised into three levels: low, moderate and high.21

Dried blood spot sampling and vitamin D assay

The dried blood spot (DBS) method was used, as it is a minimally invasive and convenient technique for children. A trained researcher used a single-use safety lancet to collect sufficient capillary blood from the child’s finger. Blood spots were collected onto a Whatman filter paper card within a pre-marked area. The sample cards were air dried at room temperature before being placed into a sealed plastic bag. All the DBS cards were refrigerated and stored at 4°C until sent to the Canterbury Health Laboratories for further analysis.

Measurement of 25(OH)D in dried blood spots was performed by a newly developed method utilising two-dimensional liquid chromatography tandem mass spectrometry (2D LC-MS/MS) following methanol and hexane extraction of the dried blood spots. Capillary calculated serum concentrations of 25(OH)D were determined by assuming a hematocrit of 0.39. The method has been thoroughly validated for research applications.22

Statistical analysis

The computer software statistical package program SPSS version 24 (IBM Corporation, New York, NY, USA) and R statistical package, version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/) were used to analyse the data. The variables were tested for normality using the Kolmogorov-Smirnov test and Shapiro-Wilk (S-W) tests and for homogeneity using the Levene’s test.

The data were normally distributed so parametric tests were used. For baseline characteristics and measurements, continuous variables are expressed as mean±standard deviation (SD) (Table 1) and categorical as number and percentages (n(%)) (Table 2). Pearson’s correlation coefficients were used to test relationships between continuous variables. The independent T-test, ANOVA test and Chi-squared test were used to compare 25(OH)D levels between groups.

Table 1: Characteristics of population group (n=507) and Pearson correlation between 25(OH)D and continuous variables.

‍

SD standard deviation.

The association between 25(OH)D status (<50nmol/L) and potential determinants were tested using univariate analysis. The following independent variables (all with a P-value of<0.20) were included in the model: age, %BF, gender, physical activity and ethnicity. Age and %BF were entered into the model as continuous variables. All other variables were treated as categorical variables. Reference categories were being male, having normal physical activity and being of New Zealand European ethnicity. Body Mass Index, body part sun exposure and sunscreen user did not meet the screening criteria (P-value≥0.20) so were not included as they were unlikely to contribute to a model contacting other potential determinants of 25(OH)D<50nmol/L. Interaction between ethnicity and skin colour did not pass the initial screen, therefore only ethnicity entered to the model. There was not any collinearity between variables. Forward stepwise multiple logistic regression analysis with the entry criterion set at P-value<0.05 was used to determine which variables to include in the final model.

Results

A total of 507 children (237 (47%) boys) participated. The main characteristics of the study population group are presented in Table 1. The distribution of ethnic groups was: 191 (38%) New Zealand European, 64 (13%) Māori, 108 (21%) Pacific, 47 (9%) South Asian, 65 (13%) Chinese/Korean/Southeast Asian and 32(6%) of others not specified. The mean age was 10±1 years (range 8–11 years). 25(OH)D concentrations were negatively associated with age (P<0.05), weight, BMI and %BF (P<0.01). Most participants were normal (86%) with a mean BMI of 19 kg/m2. High physical activity levels were reported in 36% of participants. The most common skin colour was ‘medium’ (40%) with ‘olive’ (29%) ranking second. Our results showed 90% of children exposed only legs; arms and legs; face and arms and legs to the sunlight. The majority used sunscreen only in summer time (87%) and 75% applied sunscreen on their only legs; arms and legs; face and arms and legs.

Capillary (calculated) serum concentrations of 25(OH)D ranged from 9–123nmol/L. Mean±SD 25(OH)D concentration was 64±21nmol/L. The prevalence of 25(OH)D<50nmol/L, 50–75nmol/L, and ≥75nmol/L was 28, 41 and 31%, respectively (Figure 1).

Figure 1: Prevalence of 25(OH)D status (<50nmol/L, 50–75nmol/L and ≥75nmol/L).

The relationship between 25(OH)D levels and their potential determinants are summarised in Table 2.

Mean 25(OH)D did not differ significantly between boys and girls (66±20 vs 62±21nmol/L; P>0.05), or by levels of physical activity. However, mean 25(OH)D was significantly higher in normal compared with overweight/obese children (65±21 vs 55±18nmol/L; P<0.005). New Zealand European children had the highest mean 25(OH)D concentrations of 75±20nmol/L and South Asian children had the lowest 49±20nmol/L. Participants reporting ‘dark/brown’ skin colour had lower mean 25(OH)D levels (51±18nmol/L) compared to the other skin colours. Body parts exposed to sunlight, use of sunscreen, frequency of use of sunscreen and body parts where sunscreen applied were significantly associated with 25(OH)D (Table 2).

Table 2: 25(OH)D levels and its related determinants.

‍

‍

1Independent T-test (for two categories) and ANOVA test (more than two categories), 2Chi-squared test, 3Cole et al (2000) and Cole et al (2007).
SD standard deviation.

Logistic regression analysis results on 25(OH)D and its determinants are shown in Table 3. There was a significant inverse association between %BF and an increased odds ratio (OR) of 25(OH)D<50nmol/L (OR 0.96, CI 95%: 0.94–0.99, P<0.001). All other ethnic groups compared to reference group (New Zealand European) were at higher risk of 25(OH)D<50nmol/L (OR from 0.07 to 0.28, P<0.0001).

Table 3: Results of stepwise linear regression identifying determinants of 25(OH)D<50 nmol/L.

‍

Following independent variables were included in the model: age, body fat percentage, gender, physical activity and ethnicity.
1Change in odds of 25(OH)D<50nmol/L occurring for each unit change in determinant variable. If >1.0, as determinant variable increases, odds of 25(OH)D<50nmol/L increase. If <1.0, as determinant variable increases, odds of 25(OH)D<50nmol/L decrease.

R2 =0.22 (Hosmer and Lemeshow), 0.12 (Cox and Snell), and 0.17 (Nagelkerke). Model X2=65.49.

CI confidence interval; OR odds ratio; NZ New Zealand.

Discussion

There are limited data available regarding wintertime vitamin D status and its related determinants in New Zealand children. Approximately one-third of the study population had 25(OH)D≥75nmol/L and about one-third had 25(OH)D<50nmol/L. Body fat percentage and ethnicity were the strongest predictors of 25(OH)D<50nmol/L, while many established contributors to vitamin D status (eg, gender, physical activity) were not associated with 25(OH)D<50nmol/L in this population.

There is not a worldwide consensus about acceptable vitamin D concentrations. Different cutoff points have been set for populations based on the relationship between vitamin D status and various criteria such as parathyroid hormone (PTH) levels, intestinal calcium absorption and bone mineralisation. It has been shown that 25(OH)D levels less than 50nmol/L may cause hypo-calcemia and secondary hyperparathyroidism in children.23 Based on New Zealand consensus, aiming for 25(OH)D≥50nmol/L seems prudent.4 In our study, nearly one-third (31%) of participants had 25(OH)D≥75nmol/L, 41% had 25(OH)D=50–75nmol/L and 25(OH)D<50-nmol/L was detected in approximately one-third (28%) of children. Our data are comparable to previous studies in children and adolescents.6–8,24–27 For instance, Alemzadeh et al,24 using the same cutoff points for vitamin D categories, showed 32% and 41.7% of 6–18 year olds (49 Caucasian, 39 Hispanic and 39 African American) had 25(OH)D<50nmol/L and between 50 and 75nmol/L, respectively. There are a few studies that have investigated the status of 25(OH)D among the New Zealand paediatric population. Rockell et al6 used 37.5nmol/L 25(OH)D as the cut-off value and found a high prevalence of 25(OH)D insufficiency 41%, 59% and 25% among Māori, Pacific and New Zealand European school-age children aged 5–14 years, respectively. Recently, Cairncross13 collected capillary blood spots late-winter to early spring and measured 25(OH)D in 1,329 preschool children (2–4 years old). Their results showed (86) 7% and (642) 48% of children had vitamin D deficiency (<25nmol/L) and insufficiency (<50nmol/L) respectively. Therefore, further demonstrating that wintertime 25(OH)D concentrations are of concern for some New Zealand children.

Previous studies have suggested a significant inverse association between vitamin D status and fat mass, due to vitamin D being a fat-soluble vitamin.28,29 Our results showed an inverse correlation between weight, BMI and %BF and vitamin D status. Children with normal BMI showed significantly higher 25(OH)D compared to overweight/obese children (65nmol/L vs 55nmol/L). Wortsman et al,29 suggesting obesity is associated with vitamin D insufficiency since vitamin D can be deposited in adipocytes, and therefore its bioavailability decreases. Also, in our logistic regression analysis, having a higher %BF increased the chance of having 25(OH)D<50nmol/L after accounting for age, gender and physical activity.

In our study New Zealand European children had higher 25(OH)D concentration (75 nmol/L) compared with all other ethnic groups. Rockell et al6 demonstrated the prevalence of vitamin D insufficiency (<37.5nmol/L) was 38, 58 and 23% in Māori, Pacific and New Zealand European children aged 5–14 years, respectively. Cairncross13 found that the prevalence of vitamin D deficiency (<25 nmol/L) was 9 and 23% in Māori and Pacific children, respectively, compared with 3% in New Zealand European children. We found 11% of New Zealand European children, 31% of Māori children and 34% of Pacific children had 25(OH)D less than 50nmol/L. The highest prevalence of 25(OH)D<50nmol/L was in South Asian children (62%) with a mean of 49nmol/L. Results from the logistic regression analysis demonstrated ethnicity is a predictor of 25(OH)D<50nmol/L and South Asian children had the highest odds of 25(OH)D<50nmol/L compared to the New Zealand European group. In a review paper, Akhtar30 reported that vitamin D deficiency (20–50nmol/L) is highly prevalent among South Asian population. A possible reason for these results can be related to differences in skin colour among different ethnic groups. In the present study, most New Zealand European children (56%) reported ‘medium’ skin colour while 36% of Māori children reported having ‘olive’ skin colour. About half (47%) of Pacific children and more than half (57%) of South Asian children had dark/brown skin colour. The negative association of vitamin D status with skin colour is well documented in previous studies.26,27 We found that children reporting a ‘dark/brown’ skin colour had significantly lower 25(OH)D concentrations than children reporting a ‘fair’ skin colour (51nmol/L vs 67nmol/L). It is suggested that the cutaneous synthesis of vitamin D is lower in individuals with a darker skin colour due to the ultraviolet radiation beta (UVβ) photons being less absorbed through melanin pigmentation.31

Contrary to previous studies,10,32–34 and in line with Vierucci et al7 and Avagyan et al,25 we did not find a significant difference in 25(OH)D concentration between boys (66nmol/L,) and girls (62nmol/L). It has been proposed that gender differences in vitamin D are related to BMI status.28,29 In the current study, there was not a significant difference in BMI between boys (mean 18kg/m2) and girls (mean 19kg/m2).

It has been established that sedentary lifestyle and less physically active people have lower 25(OH)D values.35 Less physically active people usually spend less time outdoors and therefore have limited opportunity for sun exposure, with an increased risk of obesity. In this study we did not find a significant relationship between vitamin D status and physical activity. Also, the mean of BMI across all three physical activity levels did not differ significantly.

In the literature, several factors have been postulated that affect cutaneous synthesis of vitamin D through sun exposure, such as body surface area, time of the day, season, latitude, degree of skin pigmentation and sunscreen use.36,37 In this study, we asked parents to identify which parts of the body were usually exposed to sunlight, frequency of use and where sunscreen was applied. Our data confirm the prevalence of vitamin D deficiency is lower in children who exposed more surface of their body (only legs; arms and legs; face and arms and legs) to the sunlight.

It is suggested that properly applying sunscreen with a sun protection factor (SPF) of 30 can decrease cutaneous synthesis of vitamin D by as much as 95–99%.36 In our study, similar to Cairncross et al,14 our participants who reported that they applied sunscreen and were more frequent users, had higher 25(OH)D concentrations compared to the other groups. An explanation for these findings may be because of the lack of parents’ knowledge about how to apply sunscreen. Furthermore, we did not ask parents or children how often they renewed application of sunscreen each day and the SPF of the sunscreen used. The amount of sunscreen applied38 and how often it is re-applied influences its effectiveness. We could also speculate that the children who were using sunscreen were going out in the sun, but those who were not applying sunscreen were not exposing themselves.

In this study, we observed that 28% of the study population presented with 25(OH)D values<50nmol/L. An effective strategy is needed to prevent vitamin D status less than 50nmol/L and its related health consequences among children. Increasing sun exposure, especially in high-risk children (overweight/obese, darker skin colour, specific ethnic groups) should be the first-line treatment. For most people living in New Zealand, exposure to the sun is the major source of vitamin D.4 The body is able to synthesise sufficient vitamin D with adequate skin sunlight exposure.39 However, there is no evidence regarding the safe threshold level of ultraviolet (UV) radiation exposure from the sun without increasing risk of skin cancer.4 A balance between avoiding excessive sun exposure to prevent skin cancer, and enough sun exposure to achieve adequate vitamin D levels, is required.4
The New Zealand Ministry of Health recommends a daily walk or some other form of outdoor physical activity in the early morning or late afternoon between the summer months of September and April.4 Between May and August (winter in New Zealand), except at high altitudes or near highly reflective surfaces (eg, snow or water), sun protection is generally not recommended. During this time exposure to direct sunlight, especially in the hours around noon when ultraviolet radiation beta (UVβ) levels are highest, will be enough for adequate cutaneous production of vitamin D.4 However, pediatricians should monitor children’s sun exposure and vitamin D levels in high-risk subjects, including those who are overweight/obese, have dark skin colour or who are of South Asian ethnicity. These groups may benefit from vitamin D supplementation.4

Strengths and limitations

One of the limitations of this study is that our results are not necessarily transferable to other populations since we recruited only healthy children living in Auckland, New Zealand. Also we recruited children based on their schools’ collaboration, so we cannot claim that our participants are representative of all children living in Auckland. For instance, the proportion of Pacific and Asian groups seem high while the Māori on the low side. However, we tried to include a range of sociodemographic levels and ethnicities. This study was a cross-sectional study at the end of winter, and lacked longitudinal assessment of vitamin D status, therefore we could not consider the effect of seasonal variation on vitamin D status. In addition, we collected some information regarding vitamin D risk factors (eg, body exposure area, use of sunscreen, physical activity) through self-reported questionnaires, so recall bias could have influenced our results. Although the DBS method is minimally invasive for measuring vitamin D and previous studies have used it,14 it has not been specifically validated in children. Despite these limitations, this study had some notable strengths and suggested a general pattern of vitamin D status and its related risk factors among school-age children living in Auckland, New Zealand. We recruited a large sample of healthy children from a broad range of sociodemographic and ethnic backgrounds. Also we used %BF, which is a better indicator of fatness in an individual than BMI.

Conclusion

Our data indicates an association between winter vitamin D status and its related factors in children living in Auckland, New Zealand. 25(OH)D<50nmol/L was present in about one-third of our participants. Ethnicity and %BF were the only significant predictors of 25(OH)D<50nmol/L, with no association shown between gender, physical activity and 25(OH)D in our population. Being overweight and obese, being from South Asian ethnicity, having darker skin colour and exposing less surface of the body to sunlight significantly affected vitamin D status. Poor vitamin D status during childhood can affect long-term health, so opportunities to intervene during childhood should be pursued. A strong consideration should be given to the high-risk children and production of cutaneous vitamin D through sunlight exposure.

‍