Dietary sources and sociodemographic and lifestyle factors affecting vitamin D and calcium intakes in European adolescents: the Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Study

Abstract

Objective

To investigate dietary sources of Ca and vitamin D (VitD) intakes, and the associated sociodemographic and lifestyle factors, among European adolescents.

Design

Linear regression mixed models were used to examine sex-specific associations of Ca and VitD intakes with parental education, family affluence (FAS), physical activity and television (TV) watching while controlling for age, Tanner stage, energy intake and diet quality.

Setting

The Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA)Cross-Sectional Study.

Subjects

Adolescents aged 12·5–17·5 years (n 1804).

Results

Milk and cheese were the main sources of Ca (23 and 19 % contribution to overall Ca intake, respectively). Fish products were the main VitD source (30 % contribution to overall VitD intake). Ca intake was positively associated with maternal education (β=56·41; 95 % CI 1·98, 110·82) and negatively associated with TV viewing in boys (β=–0·43; 95 % CI −0·79, −0·07); however, the significance of these associations disappeared when adjusting for diet quality. In girls, Ca intake was positively associated with mother’s (β=73·08; 95 % CI 34·41, 111·74) and father’s education (β=43·29; 95 % CI 5·44, 81·14) and FAS (β=37·45; 95 % CI 2·25, 72·65). This association between Ca intake and mother’s education remained significant after further adjustment for diet quality (β=41·66; 95 % CI 0·94, 82·38). Girls with high-educated mothers had higher Ca intake.

Conclusions

Low-educated families with poor diet quality may be targeted when strategizing health promotion programmes to enhance dietary Ca.

Bone mineral content at older age is determined both by peak bone mass attained in young adulthood and by adult bone loss^{( 1 )}. While the clinical consequences of adverse bone health are seen predominantly in older age, accumulating evidence indicates that many predisposing factors arise in childhood and adolescence^{( 1 )}. Therefore, interventions designed to maximize bone health should start at an early age and continue through adolescence.

While peak bone mass is mainly genetically determined^{( 2 )}, lifestyle – specifically dietary interventions promoting Ca and vitamin D (VitD) intakes during childhood and adolescence – play an important role, especially when sun exposure is not frequent^{( 1 , 3 , 4 )}.

Although there are some areas of debate, there is general agreement that prevention of VitD and Ca deficiencies is a public health priority. Health benefits from adequate VitD and Ca intakes include contribution to the maintenance of normal bones and teeth, growth and development of bone, absorption or utilization of Ca and P, blood Ca concentrations and maintenance of normal muscle function^{( 5 )}. For instance, the authors of a recent systematic review and meta-analyses about nutrient and food intakes on fracture risk concluded that adverse bone health events such as fractures may occur in children and adolescents not meeting Ca dietary recommendations^{( 6 )}.

Unfortunately, previous findings on Ca and VitD intakes in European and US children and adolescents suggest very low compliance with Ca and VitD recommendations: approximately 95 % of adolescents do not meet the recommendations^{( 7 , 8 )}. Natural VitD sources most notably are oil-rich fish and eggs, although small amounts provided by meat can be important in some countries^{( 9 )}. Identifying both the food sources contributing most to daily Ca and VitD intakes and the families with children and adolescents with low Ca and VitD intakes will support the formation of targeted public health interventions when aiming to increase Ca and VitD consumption in European adolescents.

Previous regional studies with smaller sample sizes in children and adolescents have found associations of sociodemographic status and lifestyle factors with Ca and VitD intakes. A higher prevalence of inadequate Ca intake was reported in low-income Greek and Flemish children and pre-school children, as compared with their higher socio-economic counterparts^{( 7 , 10 )}. Similarly, lower VitD and Ca intakes were related to family affluence in Brazilian adolescents^{( 11 )}.

Lifestyle-related factors including physical activity (PA) and sedentary behaviours contribute in the variation in peak bone mass; however, its relationship with dietary patterns in children and adolescents has been scarcely studied^{( 11 , 12 )}. Television (TV) viewing, for example, especially during high school, might have long-term effects on eating choices and contribute to poor eating habits in young adulthood^{( 13 )}. However, studies investigating the relationship between TV viewing and Ca and VitD intakes specifically in European adolescents are still lacking.

Unfortunately, the aforementioned studies cannot be extrapolated to the whole European adolescent population and therefore it seems timely to investigate dietary sources of Ca and VitD intakes, and their associated sociodemographic and lifestyle factors, among European adolescents as a basis for future targeted public health strategies.

Methods

Study design

The Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Cross-Sectional Study is a population-based multi-centre investigation of the nutritional and lifestyle status of adolescents, carried out in ten European cities (Vienna in Austria, Ghent in Belgium, Lille in France, Dortmund in Germany, Athens and Heraklion in Greece, Pécs in Hungary, Rome in Italy, Zaragoza in Spain and Stockholm in Sweden) between October 2005 and December 2008. A detailed description of its design and sampling procedures is published elsewhere^{( 14 )}. The study population comprised adolescents aged 12·5–17·5 years. Adolescents were excluded from the database a posteriori if they met any of the following exclusion criteria: no measurement of weight and/or height; participated simultaneously in another clinical trial; or suffered from an acute infection the week prior to the examination. The total HELENA population consisted of 3528 eligible adolescents (52·3 % females). Participants from Heraklion and Pécs were excluded from the current analyses, as no nutrient intake information was available due to logistical problems. For the purpose of the current study, only adolescents who provided data on two non-consecutive 24 h dietary recalls were included (n 2330). Energy under-reporters (calculated using Goldberg’s cut-offs) were further excluded (n 526)^{( 15 )} because the results of descriptive analyses showed that Ca and VitD intakes were significantly higher (P<0·05) in under-reporters compared with normal reporters. After exclusion, the final sample was 1804 participants (949 females).

The study was performed following the ethical guidelines of the Declaration of Helsinki 1964, the Good Clinical Practice rules and the legislation about clinical research in humans in each of the participating countries. All study participants and their parents provided a signed informed consent form. The protocol was approved by the Human Research Review Committees of the institutions involved.

Physical examination

Body weight (in kilograms) was measured using an electronic scale (type SECA 861; UK), precision 100 g and range 0–150 kg. Standing height (to the nearest centimetre) was measured on a stadiometer (type Seca 225; UK), precision 0·1 cm and range 70–200 cm. The triceps and subscapular skinfold thickness was measured on the left side of the body. A Holtain calliper (Crymych, UK) was used to measure skinfold thickness to the nearest 0·2 mm, and a non-elastic tape was used to measure circumference to the nearest 0·1 cm. Adolescents were classified on sexual maturation using Tanner’s five-stage scale following physical examination performed by a physician (stages I, II, III, IV and V)^{( 16 )}.

Dietary assessment

Dietary intake was assessed by two non-consecutive computerized 24 h recalls, including weekdays and weekend days, and the mean of these two days (adjusted for within-person variability) was included as daily intake. The 24 h recalls were collected using the HELENA-Dietary Intake Assessment Tool (HELENA-DIAT) ⁽¹⁷⁾ . Trained dietitians assisted the adolescents to complete the 24 h recalls when needed. Adolescents autonomously selected all consumed foods and beverages from a food list in the HELENA-DIAT.

Foods and recipes recorded by the adolescents were converted into nutrients with a food composition database (German Nutrient Database, BLS) that contains a large number of nutrients and food items (approximately 12 000 coded foods). It has been demonstrated to be good at estimating nutrient intakes in European adolescents^{( 18 )}. The contribution of each of forty-four food groups to Ca and VitD intake was computed as the sum of the amount of Ca or VitD provided by the food divided by the total intake of that nutrient from all foods for the entire study population. Food and beverage consumption was expressed as grams and millilitres per day, respectively.

Based on the 24 h recalls, a validated diet quality index (DQI) was designed. This is composed of the five pillars guidelines presented in percentages: dietary quality, dietary diversity, dietary equilibrium, a meal index and a PA score^{( 19 )}. For the purpose of the current analysis, the mean of these components was calculated in 1804 adolescents; as such, the DQI ranged from −33 to 100 %, with higher scores reflecting a higher diet quality. The score was calculated for each day and a mean of the daily scores was taken as global index score of the individual.

Adolescents were asked about taking any micronutrient supplements. Nevertheless, we did not find any difference in either Ca or VitD intake between consumers of supplements and non-consumers (data not shown).

Sociodemographic factors (socio-economic status and educational level)

A self-reported questionnaire was used to collect data on living conditions, family structure, employment status, and parental occupation and education level^{( 20 )}. The Family Affluence Scale (FAS) was used as an indicator of the adolescents’ material affluence (reflecting family expenditure and consumption). The scale ranged from 0 to 8; it was re-coded and dichotomized as ‘low familial wealth’ (score=0–4) and ‘high familial wealth’ (score=5–8). Parental educational level was also recorded (primary education, lower secondary education, higher secondary education or university degree) and later dichotomized into ‘lower education’ (primary and lower secondary education) and ‘higher education’ (higher secondary education and university degree).

Lifestyle factors (physical activity and television viewing)

PA records were obtained by means of two different tools: an adolescent-adapted version of the International Physical Activity Questionnaire (IPAQ-A) and accelerometers.

The original valid and reliable questionnaire, which was developed for adults aged 18–65 years, assessed PA of the last 7 d in four different domains (work; transport; house and garden; leisure time)^{( 21 )}. To adapt the questionnaire to an adolescent study population, questions about PA at work were replaced by questions about PA at school, namely physical education, walking and moderate and vigorous PA at school. As in previous studies, moderate and vigorous PA were summed and combined into a single moderate-to-vigorous PA (MVPA) variable (expressed in min/week)^{( 22 )}.

Uniaxial accelerometers (Actigraph MTI, model GT1M; Manufacturing Technology Inc., Fort Walton Beach, FL, USA) were used to objectively measure PA. Adolescents were asked to wear the accelerometer for seven consecutive days during all waking hours, except for water-based activities. At least three days of recording, with a minimum of 8 h registration per day, was set as an inclusion criterion. The time sampling interval was set at 15 s and bouts of ≥20 min of consecutive zero counts were deleted from the data sets. Total PA was expressed as total counts recorded divided by total daily registered time (counts/min).

A standardized self-reported sedentary behaviour questionnaire was administered to assess TV watching^{( 23 )}. Adolescents reported the frequency of specified sedentary behaviours using predefined response categories separately for weekdays and weekends. The reliability (one-week test–retest) of the questionnaire has been studied previously in a similar group of adolescents, with the majority of the variables showing an almost perfect agreement (Cohen’s weighted κ value of 0·81 and 0·96 for weekdays and weekend days, respectively)^{( 23 , 24 )}.

Statistical analysis

All analyses were carried out with the statistical software package IBM SPSS Statistics Version 20.0. Kolmogorov–Smirnov tests were carried out to test the normality of the studied variables. The Multiple Source Method was used to estimate usual intakes of nutrients and foods. This statistical modelling technique takes into account within-person and between-person variability and calculates usual intakes corrected for age, sex and study centre^{( 25 )}. The population proportion formula was used to determine the percentage contribution of each food group to Ca and VitD intakes. As mentioned before, this includes summing the amount of the nutrient provided by the food and then dividing by the total intake of that nutrient from all foods for the entire study population. Dietary sources were equally distributed for boys and girls, and consequently they are presented together.

Sex-specific mixed-model linear regression analyses were then used to verify associations between Ca and VitD intakes and parental education, FAS, PA and TV watching. Variables were included separately in the model. Study centre was included as the random intercept. Age, Tanner stage and energy intakes were entered as covariates in a first model (model 1) and the DQI entered afterwards in a second model (model 2).

Results

Table 1 shows descriptive characteristics of the participants. Mean Ca and VitD intakes were 833·9 (sd 372·2) mg/d and 1·8 (sd 0·9) µg/d, respectively. Boys had higher VitD and Ca intakes than girls. In addition, boys were more active than girls independently of the method used to measure PA (all P<0·05).

Table 1.

Characteristics of the study participants; adolescents (n 1804) aged 12·5–17·5 years from eight European cities, Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Study

		All		Boys (n 855)		Girls (n 949)
	n	Mean or n	sd or %	Mean or n	sd or %	Mean or n	sd or %	P
Age (years)	1804	14·7	1·2	14·8	1·3	14·7	1·2	0·311
Tanner stage	1752							0·000
Tanner I		9	0·5	9	1·1	0	0
Tanner II		130	7·4	85	10·3	45	4·9
Tanner III		432	24·7	200	24·2	232	25·1
Tanner IV		732	41·8	329	39·8	403	43·6
Tanner V		449	25·6	204	24·7	245	26·5
Ca intake (mg/d)	1804	833·9	372·2	833·9	949·1	734·9	305·2	0·000
Vitamin D intake (µg/d)	1804	1·8	0·9	2·0	1·1	1·7	0·8	0·000
Maternal education	1724							0·287
Lower education		551	32·0	268	32·7	283	31·3
Higher education		1173	68·0	552	67·3	621	68·7
Paternal education	1674							0·073
Lower education		576	34·4	260	32·6	316	36·1
Higher education		1098	65·6	538	67·4	560	63·9
Family Affluence Scale	1797							0·121
Low familial wealth		782	43·5	358	42·0	424	44·9
High familial wealth		1015	56·5	494	58·0	521	51·3
MVPA (IPAQ-A) (min/d)	1693	101·0	80·7	115·7	86·7	87·9	73·4	0·000
MVPA (Acc)* (min/d)	1222	58·7	23·6	67·5	25·0	50·7	18·9	0·000
TV viewing (min/d)	1730	110·0	66·6	111·2	66·7	109·1	66·6	0·980

MVPA, moderate-to-vigorous physical activity; IPAQ-A, adolescent-adapted International Physical Activity Questionnaire; Acc, accelerometer; TV, television.

Data are presented as means and standard deviations for continuous variables, or as numbers and percentages for categorical variables.

Significant differences between genders are indicated in bold font (P<0·05).

Main Ca and VitD sources were the same for boys and girls. Hence, Table 2 shows the main Ca and VitD sources according to different food groups for boys and girls combined. Milk and cheese were the main Ca sources (23 and 19 %, respectively). Fish products were the main VitD sources (30 %). Cakes, pies and biscuits were the second main source of VitD in our population (16 %); these products are made mainly with butter and eggs, important VitD sources, and are consumed in high amounts.

Table 2.

Dietary sources of calcium and vitamin D^* among adolescents (n 1804) aged 12·5–17·5 years from eight European cities, Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Study

	Order	%
Ca
White milk and buttermilk	1	23·40
Cheese	2	18·69
Water	3	5·51
Milk and yoghurt beverages	4	5·49
Yoghurt and quarck	5	5·14
Vitamin D
Fish products	1	30·21
Cakes, pies, biscuits	2	15·62
Eggs	3	9·76
Butter, animal fats and margarine	4	7·73
Yoghurt and quarck	5	7·46

^*Population proportion formula: sum of the amount of the nutrient provided by the food divided by the total intake of that nutrient from all foods for the entire study population (n 1804).

Table 3 presents the correlations of Ca and VitD intakes with sociodemographic and lifestyle factors. For boys, significant positive correlations were found between Ca intake and father’s education. In contrast, significant negative correlations were found between VitD intake and MVPA and TV viewing (all P<0·05). For girls, significant positive correlations were found between Ca intake and parental education (P<0·05; Table 3).

Table 3.

Crude Spearman rank correlations of calcium (mg/d) and vitamin D (µg/d) intakes with sociodemographic and lifestyle factors among adolescents (n 1804) aged 12·5–17·5 years from eight European cities, Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Study

	Ca intake		Vitamin D intake
	Boys	Girls	Boys	Girls
Maternal education	0·06	0·15	0·03	0·02
Paternal education	0·09	0·08	0·04	−0·02
FAS	0·05	−0·01	0·06	−0·05
MVPA (IPAQ-A)	−0·04	−0·04	−0·08	0·02
MVPA (Acc)	−0·05	−0·04	−0·09	0·03
TV viewing	−0·07	−0·02	−0·07	0·01

FAS, Family Affluence Scale; MVPA, moderate-to-vigorous physical activity; IPAQ-A, adolescent-adapted International Physical Activity Questionnaire; Acc, accelerometer.

Significant values are indicated in bold font (P<0·05).

Table 4 presents the results of the mixed-model linear regression analyses for Ca and VitD intakes and sociodemographic and lifestyle factors. The models were first adjusted for age, Tanner stage and energy intake (model 1) and further for diet quality (model 2). Ca intake was positively associated with maternal education (β=56·41; 95 % CI 1·98, 110·82) and negatively associated with TV viewing in boys (β=–0·43; 95 % CI −0·79, −0·07); however, these associations lost significance upon further adjustment for diet quality. In girls, Ca intake was positively associated with mother’s (β=73·08; 95 % CI 34·41, 111·74) and father’s education (β=43·29; 95 % CI 5·44, 81·14) and FAS (β=37·45; 95 % CI 2·25, 72·65). The association between Ca intake and mother’s education remained significant after further adjustment for diet quality (β=41·66; 95 % CI 0·94, 82·38).

Table 4.

Mixed-model analyses between calcium (mg/d) and vitamin D (µg/d) intakes and sociodemographic and lifestyle factors among adolescents (n 1804) aged 12·5–17·5 years from eight European cities, Healthy Lifestyle in Europe by Nutrition in Adolescence (HELENA) Study

	Boys						Girls
	Model 1*			Model 2*			Model 1*			Model 2*
	β	95 % CI	P	β	95 % CI	P	β	95 % CI	P	β	95 % CI	P
Ca intake
Maternal education	56·41	1·98, 110·82	0·042	−14·32	−76·18, 47·54	0·650	73·08	34·41, 111·74	0·000	41·66	0·94, 82·38	0·045
Paternal education	17·08	−38·42, 72·59	0·546	−14·26	−76·62, 48·11	0·654	43·29	5·44, 81·14	0·025	15·66	−23·87, 55·20	0·437
FAS	44·35	−2·86, 91·57	0·066	13·71	−39·23, 66·67	0·611	37·45	2·25, 72·65	0·037	−6·63	−42·66, 29·40	0·718
MVPA (IPAQ-A) (min/d)	0·21	−0·78, 0·50	0·153	0·38	−0·70, −0·07	0·016	0·10	−0·15, 0·36	0·426	−0·25	−0·51, 0·01	0·055
MVPA (Acc) (min/d)	−0·04	−1·15, 1·06	0·942	−0·49	−1·74, 0·75	0·439	0·40	−0·69, 1·50	0·469	0·06	−1·05, 1·16	0·921
TV viewing (min/d)	−0·43	−0·79, −0·07	0·019	−0·11	−0·51, 0·29	0·585	−0·14	−0·40, 0·12	0·293	0·10	−0·17, 0·37	0·455
Vitamin D intake
Maternal education	0·03	−0·14, 0·19	0·748	−0·06	−0·22, 0·09	0·417	−0·01	−0·14, 0·11	0·829	−0·08	−0·20, 0·04	0·199
Paternal education	0·04	−0·13, 0·21	0·657	−0·03	−0·19, 0·12	0·687	0·04	−0·07, 0·17	0·441	−0·00	−0·12, 0·11	0·970
FAS	0·13	−0·02, 0·28	0·083	0·12	−0·02, 0·26	0·089	−0·03	−0·14, 0·09	0·625	−0·04	−0·15, 0·06	0·417
MVPA (IPAQ-A) (min/d)	−0·00	−0·00, 0·00	0·317	0·00	−0·00, 0·00	0·095	0·00	−0·00, 0·00	0·834	0·00	−0·00, 0·00	0·997
MVPA (Acc) (min/d)	−0·00	−0·01, 0·00	0·197	−0·00	−0·01, 0·00	0·261	0·00	−0·00, 0·00	0·814	0·00	−0·00, 0·00	0·845
TV viewing (min/d)	−0·00	−0·00, −0·00	0·034	−0·00	−0·00, 0·00	0·216	−0·00	−0·00, 0·00	0·373	−0·00	−0·00, 0·00	0·924

FAS, Family Affluence Scale; MVPA, moderate-to-vigorous physical activity; IPAQ-A, adolescent-adapted International Physical Activity Questionnaire; Acc, accelerometer.

^*Model 1: adjusted for age, Tanner stage and energy intake; model 2: adjusted for model 1 and diet quality index.

Significant values are indicated in bold font (P<0·05).

Discussion

Main results

In this large European, population-based, cross-sectional study in adolescents, milk and cheese were the main food sources of Ca, and fish, cakes, pies and biscuits the main food sources of VitD. Ca intake was positively associated with parental education and FAS in girls, although only maternal education remained significant after adjusting for the DQI; and positively associated with mother’s education and negatively associated with TV viewing in boys, although both associations disappeared after adjusting for the DQI. To the best of our knowledge, the present study is the largest that has assessed dietary sources of Ca and VitD intakes and their determinants in European adolescents.

Similarly to our study, milk and cheese have been shown to be the main Ca sources in Flemish pre-school children^{( 7 )}, younger Spanish children^{( 26 )} and US children^{( 27 )}; and fish the main VitD source in Spanish children^{( 26 )}. Cakes, pies and biscuits were the second main source of VitD in our population; these products are made mainly with butter and eggs, important VitD sources, and are consumed in high amounts. In the USA and Canada, milk and ready-to eat cereals are usually the main VitD sources^{( 27 , 28 )}; this may be due to either differences between European and American fortification policies or the fact that fish consumption is higher in Europe than in the USA or Canada^{( 29 )}. In adults, in the USA, VitD-fortified milk makes the highest contribution to VitD intake and in Scandinavian countries oil-rich fish consumption is relatively high, although both fortification and supplementation policies have also been implemented^{( 30 )}. In the HELENA study, we use the German food composition database and therefore VitD contents in foods are representative of Germany, where the emphasis is on encouraging more outdoor sun exposure rather than on supplementation or fortification.

We observed associations between Ca intake and parental education in boys and girls, but they remained significant only for female adolescents when adjusting for diet quality. These results are in agreement with previous studies of Brazilian^{( 31 )} and Mexican^{( 32 )} adolescents in whom Ca intake was shown to be inversely associated with socio-economic status. Female European adolescents usually present healthier and varied dietary patterns^{( 33 )} compared with their male counterparts, which could explain differences by sex.

We did not find any association between TV watching and Ca or VitD intake in girls. TV viewing was a modifiable factor negatively associated with Ca intake only in boys, but this association disappeared when adjusting for diet quality. We failed to find an association between PA and VitD intake although previous studies have suggested that playing more outside and watching less TV can lead to higher levels of VitD in adolescents^{( 34 )}; however, VitD status and intakes are not positively associated in the HELENA study sample^{( 4 )}.

Strengths and limitations

The HELENA study has several strengths. The sampling procedure and the strict standardization of the fieldwork among the countries involved in the study avoided to a great extent the kind of confounding bias due to inconsistent protocols that makes comparing results from isolated studies difficult. In addition, the study includes important sets of confounders, i.e. age, diet quality and PA, which is crucial to analyse the association.

In contrast, our study has some of the inherited limitations of cross-sectional studies, mainly that the results cannot be interpreted in terms of cause-and-effect relationships. Assessing the diets of younger age groups is considered to be challenging, because their diets are highly variable from day to day. Thus, although adolescents are able to report, reports may be less accurate compared with adults^{( 35 )}.

Therefore, limitations in relation to the use of recalls as a method of assessment should be considered. For instance, recalled intakes are prone to under-reporting, rely on memory, require many days to capture an individual’s usual intake, are affected by within-person variation and provide imprecise estimation of servings. Although two days have been recommended as the minimum number of recalls required to correct for within-person variability or day-to-day variation, more than two measurements are still better when considering infrequently consumed foods such as fish for estimating usual intakes. Also, to compensate for the effect of within-person variation we used the Multiple Source Method^{( 25 )}. It should be noted as well that because of regular patterns of white milk consumption, Ca intakes generally have a relatively low ratio of intra- to inter-individual variance^{( 36 )}. Furthermore, under-reporters were excluded in our sample to avoid unreliable dietary intakes^{( 37 )}.

In addition, VitD intake calculated from the German Nutrient Database (BLS) could have been a rough estimate of the dietary VitD intake. The BLS is based on German, American, Swedish, Danish and Dutch food composition tables and thus on different analytical values. Meat and meat products have recently been shown to be better sources of VitD than previously thought^{( 9 )} because what is present in meat is mostly the intermediary metabolite, calcidiol, which on a molar basis is many times more potent than cholecalciferol^{( 38 )}. Two different updates of this food composition database have been released since the HELENA study was performed, in 2010 the German BLS 3·01 and in 2014 the German BLS 3·02^{( 39 )}, and therefore VitD estimation may be more appropriate now than it was before. However, it is worth noting that the ideal approach would be to compare different nutrient data of different food composition databases especially when aiming to estimate the nutrient intakes in Europe.

Food fortification policy varies throughout Europe. Mandatory VitD fortification of margarine was introduced in some countries in Europe as a way of preventing VitD deficiency in children and today many breakfast cereals are also fortified with VitD. Finland and Sweden began to fortify their milk with VitD in the 1990s but, overall, fortification of milk is not widely practised in Europe. Contributions from supplements were not taken into account due to supplementation differences between countries. Nevertheless, we did not find any difference in either Ca or VitD intake between consumers of supplements and non-consumers (data not shown).

Although these data are now 5–8 years old, to date the HELENA study is the only large-scale study among European adolescents and thus provides a valuable resource and benchmark for studying trends in Europe.

Recommendations

According to previous literature, European adolescents present insufficient serum VitD concentrations^{( 40 )} and they do not comply with VitD and Ca intake recommendations. Sun exposure is the main VitD source; unfortunately, during wintertime and especially in north European countries sun exposure is not enough to achieve optimal serum VitD levels.

As food fortification policy varies throughout Europe, the promotion of Ca and VitD dietary sources seems to be the most appropriate option. In this regard, enhancing the consumption of fish and milk could importantly contribute to higher VitD and Ca intakes, respectively, particularly among lower socio-economic classes.

Future policy development and implementation needs to target the main sources of dietary Ca and VitD in various populations, including low-educated families. In addition to the promotion of fish and milk consumption, government and manufacturers should encourage children and adolescents to eat healthier to support bone health.

Conclusions

Milk and cheese were the main food sources of Ca intake and fish was the most important source of VitD. Differences were observed by sex; female adolescents with high-educated parents had higher intakes of Ca. Low-educated families with poor diet quality should be the main target when strategizing health promotion programmes to enhance dietary Ca and VitD intakes.