Diet in childhood is not only of great importance to the well-being and growth of the child, it is also a potential determinant of adult morbidity and mortality.
Many chronic diseases occurring in later adult life – such as CVD, osteoporosis and some cancers – have been related to the diet or to the intake of specific dietary components, with protective or detrimental influenceReference Clarke, Frost, Collins, Appleby and Peto1–Reference De Stefani, Boffetta, Ronco, Correa, Oreggia, Deneo-Pellegrini, Mendilaharsu and Leiva6. Some disease-promoting pathways are considered to be initiated early in life, in fact perhaps already during growth in utero Reference Barker, Eriksson, Forsen and Osmond7–Reference Schack-Nielsen, Molgaard, Larsen, Martyn and Michaelsen10. At the same time, childhood is the period of life when dietary habits are being formed and may persist in later life. Hence, food consumption patterns in childhood tend to be associated with subsequent risk of developing these chronic diseases in adult lifeReference Kalkwarf, Khoury and Lanphear11, Reference Nicklas12. In other words, establishing a healthy diet in early childhood may be one way of contributing to the prevention of future morbidity and mortality.
From a public health viewpoint, long-term strategies aimed at prevention are urgently needed. However, in order to plan and develop targeted health promotion actions, a good description and identification of the major problems is necessary.
To date, with the exception of certain studies conducted in communities or cities, detailed information on dietary habits of young children does not exist in Flanders. In an attempt to bridge this gap in available descriptive data, a population-based survey has recently been carried out in pre-school children in Flanders, aimed at estimating nutrient and food intakes. Here, the results on energy and nutrient intakes from this study are reported and interpreted from the the perspective of existing national and international nutrient recommendations.
Subjects and methods
Study population
The target population for this study were all pre-school children living in Flanders – the northern Dutch-speaking part of Belgium. Using a cross-sectional epidemiological design, representative samples of pre-school children aged 2·5–6·5 years were selected on the basis of random cluster sampling at the level of schools, stratified by province and age. Cluster sampling was carried out in two stages: first, schools were selected as primary sampling units from lists made available by the Ministry of Flanders for Education; and, secondly, classrooms were selected as secondary sampling units. Within every school participating in the study, one class was randomly selected for each age group (secondary sampling unit), including all the children within the selected classes as final sampling units. Children were excluded from the study when: (1) they were staying in an institution (e.g. a hospital school), where the food was provided by the institution; (2) they were not attending school during the whole period of the fieldwork; (3) they were living abroad; and (4) when neither of their parents spoke Dutch.
Out of a sample of sixty-three eligible nursery schools, thirteen schools refused to participate in this dietary survey and seven schools refused to distribute food diaries (method described below), as this was considered too much of a burden for the children and their parents. The remaining forty-three schools that participated in the study were proportionally spread over the different provinces in Flanders. Within these forty-three schools, a total of 2095 children were invited to participate in this study.
Instruments
Both a semi-quantitative FFQ and an estimated diet record (EDR) method have been used. However, only the results of the EDR, collected over three consecutive days, are reported and discussed here. In a general questionnaire, asking about socio-demographic and lifestyle characteristics of the child and its parents, the parents were also asked to report the weight and height of their child.
In the structured EDR, days were subdivided into six eating occasions, namely breakfast, morning snacks, lunch, afternoon snacks, dinner and evening snacks. Detailed information on the type (including brand names) and portion size of the foods consumed was collected using an open entry format. On a separate sheet, parents were invited to give details on recipes, ingredients, cooking methods, etc. Only good quality food diaries (EDR), including three completed record days and containing sufficiently detailed descriptions of the food products and portion sizes consumed, were included in the analysis. Two dieticians, with long-standing experience in nutritional epidemiological fieldwork, carried out the exclusion procedure of the EDR. As a cross-check, average energy intake (EI) and nutrient intakes were calculated as the mean of the three recorded days. Diaries that produced very high or very low estimates of intake for some nutrients (e.g. energy, calcium and iron) were rechecked by the dieticians. This amounted to 5 % of the diaries. In this cross-check, only diaries having extremely low estimates for some nutrient intakes, explained by an exceptional day (such as sickness of the child), have been excluded from the study. After these quality checks, the remaining diaries were coded and entered in a ‘Diet Entry & Storage’ program (BECELReference Nederlandse Unilever Bedrijven and Rotterdam13). The food list and food composition data for this program were based on the following tables: the Belgian food composition database (FCDB)14, the Dutch FCDB15, the FCDB of the Belgian Institute Paul Lambin16 and the UK McCance and Widdowson's FCDB17.
Data collection
The directors of the schools and the teachers of the classes that participated in the study were given detailed information and instructions about the study. The teachers were asked to distribute the questionnaires and diaries among the children and to assist in motivating the parents to participate. Teachers were also asked to collect the completed questionnaires and diaries, which the parents could return in a sealed envelope for the sake of confidentiality. To ensure that all the days of the week would be covered equally, the research team determined beforehand the days to be registered by each child. In the EDR, detailed instructions were elaborated for the parents and they had to sign an informed consent in order to take part in the study.
The fieldwork of this study was carried out from October 2002 to February 2003. The Ethical Committee of the Ghent University Hospital granted ethical approval for the study.
Comparison with nutritional guidelines
In this report, adequacy of macro- and micronutrient intakes was primarily evaluated against the Belgian age-specific recommendations, which are mainly based on the RDA concept of the European Union ‘Scientific Committee on Food’18, 19. This commonly used concept of RDA is based on three main dimensions: estimated average requirement (EAR), population reference intake (PRI) and lowest threshold intake (LTI) (Table 1) 18. However, experimental data on the nutrient needs of children are more scarce and in general less reliable than for adults. The current information is inadequate to give EAR or LTI values for children and is limiting itself to giving PRI that have been derived (in the absence of reliable data) by extrapolation from the PRI of young adults on the basis of energy expenditure18. For some nutrients known to be essential, but with inadequate data to make EAR, PRI or LTI values, an acceptable range (AR) of intakes is given.
Table 1 Comparison between the RDA concept, used in Belgium and the more recently developed DRI concept of the Institute of Medicine
More recently, the Food and Nutrition Board of the Institute of Medicine (IOM) in the USA created the new dietary reference intakes (DRI) concept. The RDA and DRIs concept are explained in more detail in Table 1. When the DRI system included an EAR, adequate intake (AI) or tolerable upper intake level (UL) value for a nutrient, that differed importantly from our Belgian RDA recommendations, the proportion of the population having usual intakes below the EAR and/or above the UL were also calculated. This was the case for sodium, potassium, calcium, iron and magnesium. Since the Belgian RDA for children do not include an RDA for fibre intake, the AI of the DRI concept was used as reference value20. For macronutrients, AMDR for individuals as a proportion of total EI have been used (Table 1)20.
Statistical analysis
Usual nutrient and energy intakes were computed based on the recommendation of the IOM regarding the need to determine the distribution of usual nutrient intakes for assessing diets of population groups in relation to the recommendations21. Because of day-to-day variations in individual nutrient intakes, a large number of days of intake data are typically needed to determine usual nutrient and/or energy intakes for an individual. Unfortunately, it is seldom feasible to collect these required long-term data for each person. Therefore, a statistical modelling method that accounts for within-individual variation in nutrient intakes while requiring relatively few days of intake per individual was needed. In this survey, the NUSSER method (developed at Iowa State University) was used for estimating usual intake distributions and the proportion below or above defined cut-off values, based on the 3 d estimated diet recordsReference Guenther, Kott and Carriquiry22, Reference Nusser, Carriquiry, Dodd and Fuller23. The software program used to carry out the method was the Software for Intake Distribution Estimation (C-side)24. This report presents estimates of usual energy and nutrient intakes, including the mean intakes, standard deviation (sd), median and standard error (se) for gender/age groups for which RDA/DRI have been established: children 1–3 and 4–7 years old. In addition, the percentage below or above the reference values was calculated. The percentages of individuals with intakes below the EAR are estimates of the prevalence of inadequacy. Although the percentages of individuals with intakes below the PRI and AI are also presented herein, it should be noted that unlike an EAR, a PRI and AI cannot be used to estimate the prevalence of inadequacy in a population21.
The Student's t test or Mann–Whitney U test was used to compare the means of different groups of children.
Adopted Goldberg's cut-offs have been used for the purpose of identifying under-reporters. Although Goldberg's cut-offs were developed for adults, they have been adopted for use in childrenReference Sichert-Hellert, Kersting and Schoch25, Reference Torun, Davies, Livingstone, Paolisso, Sackett and Spurr26. Sichert-Hellert and colleagues used the physical activity level values for light physical activity, estimated by Torun et al. to recalculate the cut-offs for under-reporting in 3 d dietary records, according to the formulas proposed by Goldberg and colleagues. Using a CV for EI of 24 % for 1–5 year olds and 23 % for other age groups, as given by Nelson et al. Reference Nelson, Black, Morris and Cole27, they considered records with EI/BMR ratios up to 0·97 or 1·07 depending on the subjects' age and sex, as an implausible measurement of the actual 3 d energy intakeReference Sichert-Hellert, Kersting and Schoch25.
Results
Response rate
A total of 2095 children were invited to complete the 3 d food diaries (EDR). In total, 1052 of them returned a diary (participation rate = 50 %). As mentioned before, only good quality food diaries (EDR), containing sufficiently detailed descriptions of the food products and portion sizes consumed, were used for analysis. In total, the food diaries of twenty-six children had to be excluded because of quality problems.
Of the 1026 remaining children, 696 completed 3 d diaries, 208 completed 2 d diaries and 122 only 1 d. When using C-side, at least three recorded days are required when data are collected over consecutive days. Therefore, 31 % (330 children) of the returned questionnaires had to be excluded from the analyses since only one or two record days were completed. Characteristics of the children (e.g. age, BMI category) and socio-demographic characteristics of the parents registering a 3 d diary were compared with those of parents registering less than 3 d, and no significant differences were found (data not shown). Since only good quality 3 d diaries were included in these analyses, the total number of diaries of use for this study was reduced from 1052 to 696 diaries. Because age information was missing for thirty-five children, only 661 children were included in the analysis for age groups.
Characteristics of the study population
The characteristics of the study population are included in Table 2. Boys and girls were almost equally represented. In total, 30 % of the children were younger than 4 years old, and 70 % were at least 4 years old. Fewer than 3 % of the children had to follow a medical diet (e.g. diabetes) and fewer than 2 % followed a special eating pattern (e.g. vegetarian, macrobiotic, etc.). Thirty-two percent of the children were taking dietary supplements. These supplements included mainly multivitamin supplements (Table 2). In total, 10 % of the children were overweight or obese.
Table 2 Characteristics of the children and their parents
* n 696.
The proportion of children studied in each province (30 % in Antwerp, 24 % in East-Flanders, 22 % in West-Flanders, 15 % in Flemish Brabant and 9 % in Limburg) compared well with the proportions derived from the target population, namely pre-school children in Flanders (28, 23, 18, 18 and 13 %, respectively).
In our study, the mean ratios of observed energy intake (OEI)/BMR were 1·7 (sd 0·4) for boys and girls younger than 4 years old, and 1·7 (sd 0·3) and 1·5 (sd 0·3) for boys and girls at least 4 years old, respectively. There were no children with a lower ratio than Goldberg's cut-off adapted for children in the group of children younger than 4 years old and only < 2 % for children at least 4 years old (0·5 % in boys and 1·4 % in girls). The data presented in the rest of this report have not been adjusted for under-reporting.
Comparison with nutritional guidelines
Energy
Mean OEI (6543 (sd 1062) kJ for boys; 5757 (sd 926) kJ for girls) exceeded the estimated energy requirements (EER) for boys (6040 (sd 653) kJ) and approached the EER for girls (5798 (sd 649) kJ) in children younger than 4 years old (Table 3). However, for children at least 4 years old, the OEI (6408 (sd 932) kJ for boys; 5914 (sd 928) kJ for girls) was below the EER (6995 (sd 950) kJ for boys; 6740 (sd 950) kJ for girls).
Table 3 Usual energy (kJ/d) and macronutrient intakes in g/d calculated from estimated diet records
IOM, Institute of Medicine.
* Significance level of the differences between boys and girls according to t-test or Mann–Whitney U-test.
n (boys < 4 years old) 102.
n (boys ≥ 4 years old) 236.
n (girls < 4 years old) 95.
n (girls ≥ 4 years old) 228.
Macronutrients
The mean and median macronutrient intakes in g/d have been summarised in Table 3, while the percentages of energy derived from the different macronutrients have been compared with the recommendations in Table 4. As shown in Table 4, the mean percentage of energy derived from carbohydrates was above the recommended minimum level of 50 en%. The mean percentage of energy derived from proteins reached the AMDR upper level of 15 en%, while the mean percentage of energy derived from total fat just reached the AMDR lower level of 30 en%. The mean percentage of energy derived from saturated fatty acids (SFA) was above the AMDR upper level of 12 en%, while the mean percentage derived from MUFA and PUFA was below the AMDR lower level of 12 and 8 en%, respectively. The mean fibre intake observed was much lower than the recommended level of 14 g/1000 kcal (or 3·35 g/1000 kJ). The mean water intake for children at least 4 years old was lower than the AR lower level of 75 ml/kg per day. However, the mean water intake for children younger than 4 years old reached this AR lower level (Table 4).
Table 4 Usual macronutrient intakes calculated from estimated diet records, compared with reference values and the proportion of the population with usual intakes greater than the upper intake levels (UL)
* Significance level of the differences between boys and girls according to t-test or Mann–Whitney U-test.
† Standard error not displayed when percentage is 0 or 100.
‡ Estimated energy requirement (EER).
§ Percentage of total energy intake supplied by the nutrient in question and the recommended intake expressed as acceptable macronutrient distribution ranges (AMDR).
¶ Adequate intake (AI).
‖ Acceptable range (AR).
n (boys < 4 years old) 102.
n (girls < 4 years old) 95.
n (boys ≥ 4 years old) 236.
n (girls ≥ 4 years old) 228.
In the observed intake distribution, more than two-thirds of the children did not reach their AMDR lower level for PUFA, MUFA and fibre intakes (Table 4), while almost two-thirds and a half of the children exceeded the AMDR upper level for SFA and proteins, respectively. Almost half of the children and more than half of the children did not reach the minimum recommendation for total fat (30 % of energy) and water intakes (75 ml/kg per day), respectively (Table 4).
Micronutrients
Examination of mean micronutrient intakes showed that these exceeded the RDA for most nutrients (Table 5). For mineral intakes, only the mean iron intake (7–8 mg) was below the RDA of 10 mg/d. For vitamin intakes, only the mean vitamin D intake (2 μg) was below the AR lower level of 5 μg/d. In the observed intake distribution, < 15 % of the children had vitamin D and iron intakes above the Belgian recommended level and < 70 % had calcium intakes above the recommended levels (Table 5). Although most of the children did not reach our Belgian RDA for iron intakes (Table 6), when comparing them with the iron EAR from the IOM, almost all of them reached this EAR value (Table 6). Since the AI for calcium intake in children younger than 4 years old from the IOM is much lower (500 mg) than our Belgian RDA value (800 mg), the proportion of this population with usual calcium intakes lower than the AI is only 5 % in comparison with >30 % below our Belgian RDA.
Table 5 Usual micronutrient intakes calculated from estimated diet records, compared with reference values and the proportion of the population with usual intakes greater than the upper intake levels (UL)
* Significance level of the differences between boys and girls according to t-test or Mann–Whitney U-test.
† Standard error not displayed when percentage is 0 or 100.
‡ Adequate intake (AI).
§ Population reference intake (PRI).
¶ Acceptable range (AR).
n (boys < 4 years old) 102.
n (girls < 4 years old) 95.
n (boys ≥ 4 years old) 236.
n (girls ≥ 4 years old) 228.
Table 6 Proportion of population with usual intakes of sodium, potassium, calcium, iron and magnesium below the EAR/AI and above the UL of the IOM recommendations
IOM, Institute of Medicine.
* Standard error not displayed when percentage is 0 or 100.
† Adequate intake (AI).
‡ Upper level (UL).
§ Estimated average requirement (EAR).
n (boys < 4 years old) 102.
n (boys ≥ 4 years old) 236.
n (girls < 4 years old) 95.
n (girls ≥ 4 years old) 228.
Mean sodium, potassium and magnesium intakes exceeded the Belgian AR upper level. It should be noted that sodium derived from table salt was not included in these analyses (Table 5).
When comparing sodium intakes with the UL of the IOM, more than three-quarters of the children younger than 4 years old and >40 % of the children at least 4 years old exceeded these UL (Table 6). Although almost all the children exceeded the Belgian AR upper levels for potassium, in comparison with the potassium AI from the IOM, more than three-quarters of the children did not reach this AI value (Table 6). Almost all the children reached the IOM magnesium EAR (Table 6).
Nutrient intakes from dietary supplements were not included in these analyses. However, as mentioned before (Table 2), around one-third of the children were recorded in the general questionnaire as receiving vitamin and/or mineral supplements (18 % daily; 4 % 4–6 times a week; 8 % 2–3 times a week; 2 % once a week and 2 % less than once a week).
Discussion
Main results
In our study, both sexes were almost equally represented and one-third of the children were younger than 4 years old. Although recommendations differed for pre-school children younger than 4 years old and children at least 4 years old, they did not differ between sexes. In this population-based sample of pre-school children, the mean dietary intakes of energy and most nutrients, estimated from parental records of foods and drinks consumed, were higher in boys than in girls (Tables 3 and 5). However, from Table 4, it could be concluded that the quality of the diet, as judged by the percentage of energy contributed by each of the macronutrients, was not always higher in boys than in girls.
Although mean daily EI approached the EER for children younger than 4 years old, for children between 4 and 6·5 years old, the mean OEI were below the EER in both sexes. A possible reason why the EER was higher than the recorded EI might be that children nowadays are less active than in the past, when the EER were calculated18, Reference Torun, Davies, Livingstone, Paolisso, Sackett and Spurr26, Reference Davies, Gregory and White28, Reference Goran, Carpenter and Poehlman29. The contribution from SFA was higher than the AMDR upper level of 12 en% for children, while that from MUFA and PUFA was lower than the AMDR lower levels of 12 and 8 en%, respectively. Although issues related to fat in the diet of children are controversial, different studies indicate that the quality of fat in the diet strongly influences the lipid profiles in children. Shea et al. showed that, just like in adults, the consumption of SFA is positively correlated with total and LDL-cholesterol levels in pre-school childrenReference Shea, Basch, Irigoyen, Zybert, Rips, Contento and Gutin30. In addition, there is evidence that high LDL-cholesterol levels are an important risk factor for CVD in later lifeReference Nicklas, Farris, Smoak, Frank, Srinivasan, Webber and Berenson9, Reference Nicklas, Farris, Major, Frank, Webber, Cresanta and Berenson31, Reference Nicklas, Dwyer, Feldman, Luepker, Kelder and Nader32. Recently, Zhang indicated that high intakes of PUFA may contribute to an improved performance of cognitive functioning in pre-school children, in contrast to SFA and cholesterol intakesReference Zhang, Hebert and Muldoon33. Consequently, the opposite percentages of energy derived from SFA, MUFA and PUFA, in comparison with the recommendations, might be a source of concern for the health of our Belgian pre-school children.
Only a small number of children reached the recommended daily level of fibre intake, and more than two-thirds of the children between 4 and 6·5 years old had a total daily water intake lower than the AR lower level. Gomes et al. found that intakes of dietary fibre below the minimum recommendation are an important risk factor for chronic functional constipation in childrenReference Gomes, Maranhao, Pedrosa and Morais34. In addition, a low water (fluid) intake might also be related to medical dysfunction such as constipation, kidney stones and abdominal pain, and increased risk of a urinary tract infection. Since experimental and clinical data on some of these subjects are conflictingReference Arnaud35–Reference Shah, Aurangzeb, Khan, Bhatti and Khan39, the low fibre and water intakes found in our Belgian pre-school children's diet underline the necessity for further investigations.
For most vitamins and minerals, the mean and median intakes were in excess of the RDA, implying that the chances of any of the children in this study having inadequate intakes of one of these nutrients are rather small. The exceptions to this were iron and vitamin D. The mean and median iron intakes were approximately 30 % lower than our national RDA, and about 90 % of the children had an iron intake lower than this RDA. However, in comparison with the iron EAR from the IOM, the mean iron intake of our Belgian pre-school children was above the EAR and almost all children reached this EAR value40. While no EAR values are available in Belgium for iron intakes in young children, our RDA value for children younger than 4 years old (10 mg) is much higher than the RDA value of the IOM (7 mg). Although part of the differences in Fe recommendations between countries may be explained by the different assumptions made about Fe absorption from local dietsReference Prentice, Branca, Decsi, Michaelsen, Fletcher, Guesry, Manz, Vidailhet, Pannemans and Samartin41, these low Fe intakes in comparison with our Belgian recommendation might indicate the need to reconsider our current Fe recommendation.
The mean vitamin D intake was roughly half of the AR lower level, and more than 90 % of the children had vitamin D intakes lower than the AR lower level. However, our low estimated vitamin D intakes do not necessarily imply a problem with vitamin D deficiency, as in practice most vitamin D is derived from the action of sunlight on the skin. Although no intakes of vitamin D supplements have been reported, it is noteworthy that possible contributions from multivitamin supplements have not been taken into account, which might also cause an underestimation of the true vitamin D intake. However, very young children (up to 4 years of age) in Belgium might risk having an inadequate vitamin D intake during winter months, when there is less exposure to sunlightReference Emmett, Rogers and Symes42, Reference Davies, Bates, Cole, Prentice and Clarke43.
Further, our results revealed that more than three-quarters of the children younger than 4 years old and >40 % of the children at least 4 years old exceeded the sodium UL from the IOM. Although mean potassium intakes exceeded our Belgian AR upper levels, more than three-quarters of the children did not reach the IOM potassium AI. It is important to note that some studies indicate that elevated blood pressure, which may lead to stroke, CHD and kidney disease, is associated with increased sodium and inadequate potassium intakes37. Taking into account that sodium derived from table salt was not included in the analyses, the estimated sodium intakes are probably an underestimation of the real sodium intake. Consequently, these increased sodium and inadequate potassium intakes, in comparison with the IOM recommendations, are a cause for concern for the health of our Belgian pre-school children.
Stengths and limitations of the study
While 10 % of the children were found to be overweight or obese, it should be noted that the weight and height of the children were reported by the parents. Since the ability of parents to estimate the weight and height of their pre-school child is only limited, this percentage should be interpreted with cautionReference Huybrechts, De Bacquer, Van Trimpont, De Backer and de Henauw44.
The representativeness of our study was tested by comparing our study sample with some characteristics of the target population, namely pre-school children in Flanders. First, the proportion of children attending schools in Flanders at the considered age was calculated. Data from the Ministry of Flanders for Education revealed that about 94 % of the children between 2·5 and 6·5 years old were attending schools in Flanders in 200345, 46. This percentage is a rough estimation, based on the total population of pre-school children (age 2·5–6·5 years old) in Flanders in 2003 and the number of pre-school children attending schools at that time. It is noteworthy that the majority of the children not attending schools are in the age group of 2·5–3 years old, since most children start attending schools between the age of 2·5 and 3. This might explain why only 30 % of the children studied were younger than 4 years old. When taking into account that about 6 % of Flemish pre-school children are not attending schools, it should be noted that our sampling frame ‘the lists of schools made available by the Ministry of Flanders for Education’ might introduce some selection bias, since children not attending schools might differ from pre-school children attending schools. Although our primary education in Belgium is free of charge, it is possible for instance that children from the lower socio-economic classes are less represented in the total population of pre-school children attending schools.
Secondly, the proportion of Flemish pre-school children in the different provinces was compared with the proportions in our study population, and this comparison confirmed good demographic representativeness of our study sample45.
While the response rate in our study reached 50 %, the total number of children included in the analysis was further reduced, since several children did not reach the stringent inclusion criteria specified for the analysis in C-side: an EDR, including three ‘good-quality’ record days. Although, no doubt, willingness to participate leads to some selection bias, these data represent a more general population of pre-school children in Flanders, in comparison with other food consumption surveys in children, which are mostly restricted to local areas.
It is also noteworthy that like any dietary assessment methodology, the EDR is prone to a degree of misreporting which may have influenced our classification of compliance and non-compliance with dietary reference intakes. However, the percentage of under-reporters in our study sample was rather low, and exclusion of presumed misreporters could also bias the resultsReference Black, Goldberg, Jebb, Livingstone, Cole and Prentice47. In addition, a 3 d EDR does not necessarily reflect individuals' usual intake. However, a statistical modelling method (the NUSSER method) that accounts for within-person variability was used in order to calculate valid usual nutrient and energy intakes. Since all days of the week were included in the study, it was possible to adjust our data to remove the effect of day of the week. Unfortunately, it was impossible to correct for seasonal variations, because our fieldwork was conducted only during autumn and winter time. No data were found about potential seasonal influences on nutrient intakes in this population group in Belgium. However, from our National Food Consumption Survey in 2004, it could be concluded that seasonal variations were only limited for nutrient intakes in our Belgian population of at least 15 years oldReference Devriese, Huybrechts, Moreau and Van Oyen48. These low seasonal variations in our Belgian population could be due to the widespread availability of most foods all year round.
Another potential limitation encountered was that EAR values for pre-school children are still lacking for most of the nutrients in the RDA concept, as well as in the DRI concept, which made it difficult to establish inadequacy. However, when comparing with a PRI or AI value, it is still possible to say if the intake of an individual is adequate, when intakes are above the PRI/AIReference Prentice, Branca, Decsi, Michaelsen, Fletcher, Guesry, Manz, Vidailhet, Pannemans and Samartin41.
Finally, it should be noted that food composition data, used for calculating nutrient intakes, might also introduce some bias in dietary surveys reporting nutrient intakes. In particular, our assessment of fibre intake could be distorted by the use of inadequate FCDB, since these databases have limited information on fibre contents, which could contribute to underestimates of intakes of this nutrient49.
Comparison with previous studies
The authors compared their study results with other studies concerning nutrient intake in pre-school children, namely the enKID StudyReference Aranceta, Serra-Majem, Perez-Rodrigo, Ribas-Barba and Delgado-Rubio50, the DONALD studyReference Kroke, Manz, Kersting, Remer, Sichert-Hellert, Alexy and Lentze51 and the different studies reported in the review of Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52 For the comparison with the study of Lambert et al., it is noteworthy that a variety of collection methods were used in this review and there was no consistency in the ages of the children surveyed or the age cut-off points. However, most surveys gave data for males and females separately at all ages. Just under half of the surveys were nationally representative and most of the remainder were regional. In addition it is important to note that each country used a different set of food composition data, which differ in definitions, analytical methods, units and modes of expressionReference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. The enKid study is a population-based, cross-sectional nutrition survey in Spanish children and adolescents (2–24 years old) conducted between 1998 and 2000. In the enKID study, dietary assessment was completed by means of a 24 h recall and an FFQ completed in an interview with the mother or caregiver for children under 13 years old. A second 24 h recall was completed on 25 % of the sample, allowing for adjustment of intakes for random intraindividual variationReference Aranceta, Serra-Majem, Perez-Rodrigo, Ribas-Barba and Delgado-Rubio50.
The DONALD study (which started in 1985) is a cohort study collecting detailed data on diet, metabolism, growth and development from healthy subjects between infancy and adulthood (3 months to 18 years old). Parents of the children or the older subjects themselves kept 3 d weighed dietary records, weighing and recording all foods and fluids consumed as well as leftovers using electronic food scalesReference Kroke, Manz, Kersting, Remer, Sichert-Hellert, Alexy and Lentze51.
Our results for EI fell in the ranges reported by Lambert et al. for the same age and gender groups, and our finding that the EI expressed relative to body weight decreased with increasing age was also in agreement with the findings of Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. It is noteworthy that also for the analyses in the 80 dietary surveys reported by Lambert et al. under-reporters have not been excluded. The enKid study reported slightly higher EI in children 2–5 years old. However, in the enKid study, the under-reporters (18·7 % calculated with a Goldberg cut-off of 1·14 for EI/BMR) have been excluded for these analysesReference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53. Although, in contrast to our study, the under-reporters in the DONALD study have been excluded for the analyses, the total EI of pre-school children was slightly lower in the DONALD study than in our studyReference Kersting, Sichert-Hellert, Lausen, Alexy, Manz and Schoch54.
Our results for fatty acid intakes were comparable with the studies reported by Lambert et al., except from the Southern Mediterranean countries, which had higher MUFA intakes. The cholesterol intake was comparable with the intakes from The Netherlands, the UK and Denmark reported by Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. Total fat and cholesterol intakes reported in the enKid study were much higher in comparison with our dataReference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53. When comparing our results with the DONALD study, we found that the percentage of energy derived from fat in the DONALD study was much higher than in our studyReference Alexy, Sichert-Hellert and Kersting55, Reference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56. This higher fat intake was mainly attributable to higher SFA and MUFA intakes in the DONALD studyReference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56. Also the total cholesterol intake was much higher in the DONALD study than in our studyReference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56.
Our finding that the contribution to EI from proteins exceeded the AMDR upper levels was comparable with most of the studies reported by Lambert et al., except from Sweden and France where the protein intakes were higher than 16 % of the energy. Also the protein intakes reported in the enKid study were slightly higher than those reported in our studyReference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53. The percentages of energy derived from proteins in the DONALD study were much lower than in our studyReference Alexy, Sichert-Hellert and Kersting55, Reference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56.
Although the contribution to EI from total carbohydrates was comparable with the results reported in other studiesReference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52, the contribution to EI from simple carbohydrates was high in comparison with other studiesReference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. Even though the total carbohydrate intake in Spanish children (in g/d) was comparable with our results, because of the higher fat and protein intakes, the percentage of energy derived from total carbohydrates in Spanish children was much lower than in our studyReference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53. The percentage of energy derived from carbohydrates in the DONALD study was much lower than in our studyReference Alexy, Sichert-Hellert and Kersting55, Reference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56.
For the children younger than 4 years old, our fibre intake results were slightly higher than those reported by Lambert et al. For children at least 4 years old, our results were comparable with those of Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. Fibre intakes in young children from the enKid study were lower than in our studyReference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53 and also fibre intakes in the DONALD study were much lower than in our studyReference Kersting, Sichert-Hellert, Alexy, Manz and Schoch56.
Since most food consumption studies in children do not provide information about the water or fluid intakeReference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52, we were not able to compare the low water intakes in our Belgian pre-school children with other studies in pre-school children. However, as mentioned before, the importance of fluid (water) intake in pre-school children should be further investigated.
The sufficient intakes of most micronutrients (except from iron and vitamin D intakes) found in our pre-school children were comparable with other studies also reporting sufficient intakes for most micronutrients, except for iron and/or vitamin D intakesReference Emmett, Rogers and Symes42, Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52, Reference Serra-Majem, Ribas-Barba, Perez-Rodrigo and Bartrina53. Our results for vitamin D and iron were comparable with those from France, The Netherlands and the UK, as reported by Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52. In the DONALD study, however, not only the iron intake, but also the intake of all micronutrients studied (vitamin C, thiamine, riboflavin, calcium and iron) was much lower than in our studyReference Kersting, Alexy and Sichert-Hellert57.
The high sodium intakes in our study were comparable with those reported by Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52, while the low potassium intakes estimated in our study were still somewhat higher than those reported by Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52.
The high magnesium intakes found in our study were comparable with those reported by Lambert et al., while the phosphorus intakes in our study were a little higher for children younger than 4 years old and slightly lower for those at least 4 years old in comparison with the results reported by Lambert et al. Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin52.
Finally, the authors would like to underline that methods of measuring food intake in children are not standardised across Europe and intake data are generally poor. This can make comparisons between countries difficult and inaccurate, creating uncertainties over the true nutrient intakes of children and adolescents across Europe. These uncertainties should be kept in mind in the comparison with other studies described above.
Conclusion
Although diets of pre-school children in Flanders were adequate in most nutrients during autumn and winter time, it is necessary to revise our national iron recommendations and to analyse the implications of low iron, vitamin D and fibre intakes in pre-school children. Also the effect of low water intakes in pre-school children should be investigated. Furthermore, the SFA, MUFA and PUFA intakes, which were opposite to the recommendations, and the excessive sodium intakes might be potentially modifiable risk factors for CVD in later life.
Acknowledgements
We thank Mia Bellemans and Mieke De Maeyer, the dieticians of our team, for the data input and their contribution to the conceptualisation of the FFQ. We especially acknowledge Professor D. De Bacquer, Professor H. Van Oyen, Dr C Vereecken, ir. I. Sioen and ir. C. Matthys for their critical advice. We also are extremely grateful to all the parents and teachers who participated in this project and generously volunteered their time and knowledge. Funding for this project was provided by the Belgian Nutrition Information Center.
