During the past two decades, various and complex economic, technological and social changes have occurred in the living conditions of children and adolescents. Several of these developments are likely to have a significant impact on behavioural changes. From the perspective of prevention and health promotion, health behaviour and its change over the years during childhood and adolescence are of special interest. Besides physical activity and social integration, health behaviour includes also dietary behaviour, which has an essential impact on development and well-being in general, as well as on development of overweight and obesity in children and adolescents. In the long term, nutrition in childhood and adolescence has a potential impact on adult morbidity and mortality(Reference Law1, Reference Smith2).
Changes which may have affected the dietary habits of children and adolescents are, for example, enlargement of the variety of available foods (e.g. fortified foods, energy-dense fast foods and soft drinks), eating and cooking habits at home, time spent on personal computers, video games and Internet use, commercial advertising and consumer information. For about two decades it could only be assumed that changes observed in other industrial nations were similar for Germany, because the most recent representative German National Food Consumption Study (Nationale Verzehrsstudie, NVS), which included children and adolescents, was conducted from 1985 until 1988. The NVS II conducted in 2006 did not include children and adolescents younger than 14 years(3). After the NVS study, only data from small local studies could give some impression of how food consumption patterns, energy and nutrient intake had changed over time. The Dortmund Nutritional and Anthropometric Longitudinally Designed (DONALD) study is the only study that has been providing long-term food and nutrient intake data of children and adolescents based on 3 d weighed dietary records in Germany since 1985. However, the sample of this study has not been intended to be representative for Germany(Reference Kroke, Manz, Kersting, Remer, Sichert-Hellert, Alexy and Lentze4). Further food consumption data come from household budget surveys (Einkommens- und Verbrauchsstichprobe, EVS). These studies have obtained information from household booklets for expenses on foods and beverages. The approximations of individual information from household data have limited validity, especially for children and adolescents(Reference Gedrich, Wagner and Karg5). The nationally representative Eating study as a KiGGS Module (EsKiMo) gives up-to-date insight into the food intake of children and adolescents and provides the opportunity to observe changes which have occurred during the last 20 years.
The current paper presents the findings of NVS and EsKiMo, revealing improvements in dietary habits and nutrient intakes and information for preventive measures.
Subjects and methods
The first German National Food Consumption Study (NVS) was conducted between 1985 and 1988 in the Federal Republic of Germany and West Berlin. It was funded by the Federal Ministry for Research and Technology. Sampling of private households was performed stratified and disproportionately by means of a defined random procedure. All persons living permanently in the selected households were asked to participate. The NVS sample included data from 1066 children aged 6–11 years and 1199 adolescents aged 12–17 years (1111 boys and 1154 girls). The food intake was assessed with 7 d dietary records. Amounts of foods were in general obtained by weighing and, if this was not possible outside the home, portion sizes were estimated using household measurements (e.g. cup, teaspoon). The total participation rate of NVS was 74 %; the participation rate of children and adolescents is not reported(Reference Kübler, Balzter, Grimm, Schek and Schneider6, Reference Adolf, Schneider and Eberhardt7).
In 2006, the dietary behaviour of 6- to 17-year-olds was assessed in a module of the German Health Interview and Examination Survey for Children and Adolescents (KiGGS)(Reference Mensink, Bauch, Vohmann, Stahl, Six, Kohler, Fischer and Heseker8). KiGGS was conducted by the Robert Koch Institute with the aim to collect comprehensive and nationwide data on the health status of children and adolescents from May 2003 to May 2006. The sample was drawn with a two-stage clustered and stratified sampling procedure. In the first stage, 167 sample points representative for German communities were selected with regard to community size and federal state. In the second stage, for every age, almost the same number of participants was randomly selected from the population registries. The overall response rate was 67 %. A total of 17 641 children and adolescents aged 0 to 17 years were examined(Reference Kurth, Kamtsiuris and Hoelling9). The Eating study as a KiGGS Module (EsKiMo) was performed by the Robert Koch Institute and the University of Paderborn, funded by the Federal Ministry of Food, Agriculture and Consumer Protection. For EsKiMo, an age-stratified random sub-sample of KiGGS participants was comprehensively asked about their food intake from January to December 2006. Data of 1234 children aged 6–11 years and 1272 adolescents aged 12–17 years were obtained (1248 boys and 1258 girls). The participation rate of EsKiMo was 63 %. As children and adolescents differ in their abilities, willingness to cooperate and personal circumstances, two different tools for data collection were used. Parents of participants younger than 12 years were asked to complete dietary records on three given consecutive days. Participants 12 years of age and older were personally interviewed about their food intake during the last four weeks following a modified diet history method with the use of the DISHES (Dietary Interview Software for Health Examination Studies) software(Reference Mensink, Heseker, Richter, Stahl and Vohmann10). Portion sizes were identified in EsKiMo using tableware models and a German picture book adapted from the EPIC-SOFT Picture Book(Reference van Kappel, Amoyel, Slimani, Vozar and Riboli11), packaging size and household measurements, or sometimes by weighing. Furthermore, all participants were asked about their sociodemographic background, leisure-time activities, lunch at school, body weight and height. As an incentive, all participants received €5 per recorded day and €10 per interview(Reference Mensink, Heseker, Richter, Stahl and Vohmann10).
Food composition database
All foods were aggregated as consumed to food groups according to the classification system used in the NVS. Similar groups were summarised (e.g. confectionery/jam/sugar, alcoholic beverages). Food intake data were converted to average daily energy and nutrient intakes using the German Nutrient Database (Bundeslebensmittelschlüssel, BLS; Max Rubner Institut, Karlsruhe, Germany). The BLS essentially covers average nutritional values of approximately 10 000 food codes available on the German market (fresh foods, processed foods and dishes). The nutrient calculation in the NVS dataset was performed with the BLS version II.2(Reference Gedrich, Wagner and Karg5) and the EsKiMo data with the BLS version II.3, which may be assumed to reflect the food composition at the time of the surveys best. In these two versions of the BLS, branded foods and fortified foods are not included. Therefore, the food compositions of about 1200 foods, mainly sweets, cereals and fast foods, were collected from producer information, Internet queries, labels and information from additional food composition databases from other countries as completion of the EsKiMo nutrient database. Furthermore a database of supplements was appended(Reference Mensink, Heseker, Richter, Stahl and Vohmann10), which was developed for the NVS II(3). Thus, in EsKiMo nutrient intake includes intakes from fortified foods and dietary supplements. For the calculation of total folate, amounts of folic acid in fortified foods and supplements were multiplied by 1·7 because of the higher bioavailability of synthetic folic acid compared with natural folate(12).
Statistical analyses
All analyses were performed with the SPSS statistical software package version 14·0 (SPSS Inc., Chicago, IL, USA). Descriptive statistics (percentiles 25, 50 and 75) summarise the data. The Mann–Whitney U test was used to test differences in the distributions of food and nutrient intakes between the years 1985–8 and 2006, taking statistical significance as P < 0·05. The analyses were done separately for children, adolescents, boys and girls.
In EsKiMo, all cases were weighed by a weighting factor which corrects for the disproportionally higher number of participants from the eastern parts of Germany, differences in the age, sex and nationality distribution of the final sample and the general population, and takes the day of the week of the records into account(Reference Kurth, Kamtsiuris and Hoelling9, Reference Mensink, Heseker, Richter, Stahl and Vohmann10).
Energy density was calculated by dividing the energy content of the diet by the weight of the diet (kJ/g). Energy density of foods (without beverages) could be calculated exactly only with the data of EsKiMo. For NVS, the sugar content of beverages (except water, tea and coffee) was estimated at 10 % (40 kcal (167 kJ)/100 g). Nutrient densities were calculated as nutrient intake (g, mg or μg) per unit of energy (MJ).
Results
Food consumption
Table 1 and Table 2 show the food intake by main food categories among children and adolescents. There are distinct differences for total reported food intake among adolescents between EsKiMo and NVS, with larger quantities in EsKiMo. A slightly higher food intake in EsKiMo than in NVS can be noticed among 6- to 11-year-old girls, but not among boys. Beverage intake, however, was notably higher in all age and sex groups in 2006 than 20 years ago. Differences between food intake in the 1980s and 2006 exist for all food groups, although they are not significant in all age and sex groups. Both children and adolescents ate less bread/pastries, meat, meat products/sausages (NS among 12–17-year-old boys), eggs (NS among 12–17-year-old boys), butter, potatoes/potato products and tea/coffee in 2006. In contrast, the consumed amounts of pasta/rice/cereals, cheese (NS among 6–11-year-old boys), vegetables/pulses, fruits/nuts, confectionery/jam/sugar, condiments/sauces and non-alcoholic beverages were higher than 20 years ago. Decreased intakes in the group of children but increased intakes among adolescents occurred for fish/fish products (NS among 6–11-year-old boys), milk/dairy products (NS among 12–17-year-old boys), fats/oils (NS among 12–17-year-old girls) and alcoholic beverages.
Table 1 Distribution of average daily food intakes among 6- to 11-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
Table 2 Distribution of average daily food intakes among 12- to 17-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
Energy and nutrient intakes
Table 3 and Table 4 show the energy and nutrient intakes of children and adolescents in 1985–8 and 2006. Notable differences between the two surveys exist with respect to energy and most macro- and micronutrients, which are in tendency similar for children and adolescents. Fat intake decreased significantly (NS among 12–17-year-old boys) due to declines of both saturated and unsaturated fatty acids among children and due to reduction of only unsaturated fatty acids among adolescents. Protein intake decreased among children but increased among adolescents (NS among 12–17-year-old girls). Carbohydrate intake increased in all age and sex groups except 6- to 11-year-old boys due to higher polysaccharide intake and additionally, among adolescents, because of increased mono- and disaccharide intake. Only in adolescents was consumption of dietary fibre significantly higher in 2006 than in 1985–8. The lower fat intake and higher carbohydrate intake resulted in a lower percentage of energy from fat and a higher percentage of energy from carbohydrate in the diet. Protein as a proportion of total energy intake did not change significantly among children, but decreased among adolescents. Total energy intake was lower among children, but higher among adolescents in 2006 than in the previous years. Water intake increased in all age and sex groups. Intake of vitamin D was lower in all age and sex groups, while intake of vitamin E was higher among adolescents, but not among children, in 2006 than in 1985–8. Vitamin A intake fell among children but increased among adolescents, whereas β-carotene intake increased among both children and adolescents. In 2006, the intakes of all water-soluble vitamins apart from vitamin B12 were higher compared with the 1980s, although not always significant among children. The intakes of the minerals Ca and Mg increased in all groups; the intakes of K, P, Fe and Zn increased among adolescents but decreased or did not change among children.
Table 3 Distribution of average daily energy and nutrient intakes among 6- to 11-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
Table 4 Distribution of average daily energy and nutrient intakes among 12- to 17-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
In 2006, energy density was significantly lower in all age and sex groups than 20 years ago, but nutrient density of the diet increased for both dietary fibre and most micronutrients (Table 5 and Table 6). The densities of vitamin B1, vitamin B2, vitamin B6, folate, niacin and vitamin C increased whereas the densities of vitamin B12 (NS among 6–11-year-old boys) and vitamin D decreased among both children and adolescents. β-Carotene density became higher in all age and sex groups, whereas total vitamin A density increased only among adolescents but decreased among children. Nutrient densities of K, Ca, Mg, P (NS among 12–17-year-old boys) and Fe increased, but density of Zn did not change.
Table 5 Distribution of average daily nutrient density of the diet among 6- to 11-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
Table 6 Distribution of average daily nutrient density of the diet among 12- to 17-year-old German boys and girls in 1985–8 and 2006
NVS, German National Food Consumption Study; EsKiMo, Eating study as a KiGGS Module; KiGGS, German Health Interview and Examination Survey for Children and Adolescents; P25, 25th percentile; P50, 50th percentile; P75, 75th percentile.
Discussion
Between the years 1985–8 and 2006, food consumption and nutrient intakes of children and adolescents changed considerably. On the one hand, positive changes were, for example, increases of vegetables/pulses, fruits/nuts and non-alcoholic beverage consumption and decreased consumption of meat products/sausages, butter and fats/oils. On the other hand, the decreased consumption of potatoes/potato products and bread/pastries was undesirable. The macronutrient composition of the diet improved, the micronutrient content of the diet in relation to energy intake increased and the energy density became lower. However, some micronutrients were not supplied sufficiently, fibre intake was relatively low, and fatty acid and carbohydrate compositions were unfavourable in both the 1980s and 2006.
Before discussing the implications of the results, some limitations of the analyses have to be considered. Areas and populations differ between the two nationwide surveys NVS (1985–8) and EsKiMo (2006). NVS was conducted before the German reunification and covered only the western part of Germany. EsKiMo, however, was conducted in both the western and the eastern part of Germany. Analyses of the National Health Surveys conducted in 1991 and in 1998 indicated that differences in food intake between adults from eastern and western parts of Germany existed in 1998, although these differences were smaller than those observed one year after the reunification(Reference Mensink and Beitz13). It might be suspected that the results of the present analyses were affected by including children and adolescents from the newly formed German states. However, including only children from the western parts of Germany in the analyses (implies both same area and similar survey technique) resulted in analogous findings. The only differences were that the increases in cheese intake and nutrient density of Zn became significant among boys.
Another difference between the two surveys, which should be taken into account, is the selection criteria. The NVS sample consists of households with German citizenship householders, but in the EsKiMo study, a representative sample including migrants with both German and non-German nationality was aspired. In both surveys, sufficient German language skills were necessary.
An additional point is that several methods for collecting data on food consumption are available. Crucial to the particular decision for a method are theoretical and practical considerations. General aims are to maximise motivation and participation among the sample and to minimise burden and errors of data collection with the given resources. Beyond doubt, conditions among the population on the one hand and scientific and technical resources on the other hand have changed a lot during the past two decades. Hence, it is perspicuous that NVS and EsKiMo used different methods for data collection. An advantage, but also a limitation of EsKiMo is the fact that the special needs of children and adolescents were met. The estimated 3 d record, which was conducted by the children and their parents in EsKiMo, was methodologically comparatively similar to the weighed 7 d record of the NVS. Although dietary records are often regarded as the gold standard(Reference Rockett and Colditz14), they may interfere with eating habits and are prone to under-reporting, especially in obese subjects and adolescents(Reference Livingstone, Robson and Wallace15). In contrast, assessing the food intake of adolescents with the DISHES software might be associated with difficulties in remembering and misreporting. DISHES has been validated for adults aged 19–59 years, but not yet for adolescents(Reference Mensink, Haftenberger and Thamm16). A high ratio of compliance in EsKiMo was observed including few discarded reports and interviews because of doubtful validity and relatively low number of low energy reporters. Low energy reporters were not excluded because that would reduce the representativeness of the total samples. Cut-offs for the identification of high energy reporters were not applied. A separate validation survey for this age group was not performed either in NVS or in EsKiMo. In addition, different periods of data collection (7 d v. 3 d and 4 weeks) should also be considered when interpreting the results, especially for seldom consumed foods like fish.
The data sets of NVS and EsKiMo were based upon different versions of the German Nutrient Database. Principally, it is difficult to differentiate between real changes in nutrient intake because of altered dietary behaviour and nutrient contents of food and feigned changes because of different databases when regarding changes over time. Within the GISELA Study, a study with German senior citizens, the effects of the BLS versions II.2 and II.3 on energy and nutrient calculation were analysed. The results indicated that nearly all variables differed marginally but significantly between the two databases. For example, calculated energy intake with BLS II.3 was about 1 % lower than when calculated with BLS II.2. Small positive deviations were calculated for carbohydrate and protein intakes (about +1 % difference), whereas the deviation was negative for fat intake (about −6 % difference). The amounts of most micronutrients were slightly different when calculated with the more recent version, except vitamin D, which differed notably by about −19 % among women and −36 % among men(Reference Lührmann, Kegreis, Hartmann, Dehne and Neuhäuser-Berthold17). Thus, the changes in macro- and micronutrient intake observed herein can be attributed only to a very small extent to the different nutrient database versions. The decrease of vitamin D intake, however, might be to some extent artificial.
Food consumption
Although it is necessary to discuss the data with some caution, they provide basic and relevant information.
The comparison of food consumption among children and adolescents in Germany in 1985–8 and 2006 revealed significant changes. During the past two decades, total food intake increased and food selection altered. The amounts consumed of pasta/rice/cereals, cheese, vegetables/pulses, fruits/nuts, confectionery/jam/sugar, condiments/sauces and non-alcoholic beverages increased significantly. Consumption of bread/pastries, meat, meat products/sausages, eggs, butter, potatoes/potato products and tea/coffee decreased in contrast. Intakes of both fats/oils and milk/dairy products decreased significantly among children, but fats/oils increased among adolescent boys and milk/dairy products increased among adolescent girls. Despite mainly improved food selection, more plant foods but less animal, high-fat and high-sugar foods should be consumed to meet the German recommendations for food intake(Reference Kersting, Alexy and Clausen18). In addition, beverage selection should be optimised, because energy-supplying juices and sugar-sweetened drinks (soft drinks) accounted for about half the beverages (data not shown). Because of the clear associations of soft drink consumption and nutrition and health outcomes, especially overweight and obesity(19–Reference Vartanian, Schwartz and Brownell21), consumption of these beverages should be replaced by energy-free beverages. The main results are to some extent in accordance with findings of studies from Germany and other European countries and the USA, despite different periods and methods of data collection and evaluation (food classification, age groups). The German DONALD study found significant positive trends for beverage consumption in all age and sex groups, bread/cereals among 9- to 13-year-old boys and 4- to 8-year-old girls, and negative trends for both meat/fish/eggs and fats/oils among 4- to 18-year-old boys and 4- to 13-year-old girls and for fruits/vegetables among 4- to 8-year-old boys and 9- to 13-year-old girls, between 1985 and 2000. The consumption of milk/milk products, potatoes/pasta/rice and candy/cakes did not change significantly(Reference Alexy, Sichert-Hellert and Kersting22). In addition, fortified foods became more important for child nutrition(Reference Sichert-Hellert, Kersting, Alexy and Manz23). The Bogalusa Heart Study revealed lower consumption of breads/grains, mixed meats, eggs, fats/oils, vegetables/soups, desserts and candy, but higher consumption of cheese, fruits/juices, beverages, snacks, condiments and poultry, among 10-year-olds in 1992–4 compared with 1973–4. The consumption of seafood, beef, pork and milk did not change(Reference Nicklas, Elkasabany, Srinivasan and Berenson24). Other studies from the USA indicated an increase of beverage intake among 2- to 18-year-olds(Reference Nielsen and Popkin25) and among 11- to 18-year-olds an increase of vegetable consumption and a decline of milk consumption(Reference Cavadini, Siega-Riz and Popkin26). The Nationwide Nutritional Survey of Food Behaviour of the Italian population (INN-CA) conducted in 1980–4 and 1994–6 observed increasing amounts of cakes/biscuits/pastries, ham/salami and savoury snacks among Italian children (1- to 9-year-olds) and adolescents (10- to 17-year-olds)(Reference Turrini, Saba, Perrone, Cialfa and D’Amicis27).
Energy and nutrient intakes
The overall changes of food intake were reflected in improvements of nutrient intakes. In 2006 most D-A-CH reference values(28) were met better than in the 1980s. First, water intake increased in all age and sex groups; thus the median water intake of children nearly reached the dietary recommendations and adolescents met the recommendations for water intake. The recommended intake of carbohydrates (>50 % energy) was not met in median in any group in the 1980s, but reached the median in all age and sex groups in 2006. Although the intake of polysaccharides increased, mono- and disaccharides accounted still for half of total carbohydrate intake. In contrast, the intake of dietary fibre was also relatively low (2·4 g/MJ recommended). Therefore, the intake of complex carbohydrates and dietary fibre needs increasing further. The recommended fat intake (30–35 % energy for 6- to 14-year-olds, <30 % energy for ≥15-year-olds) was exceeded by far in all groups in the 1980s. In 2006, median fat intake was higher than recommended only among 15- to 17-year-olds (data not shown). Even though the intake of SFA fell in all groups, in comparison to the D-A-CH reference value (saturated:unsaturated fatty acids = 1:2) their proportion was still rather high in 2006. Further improvement of the fatty acid composition is necessary. Median protein intake was much higher than the recommendations in all groups (between 17 g/d and 60 g/d depending on age and sex recommended), thus nearly all children and adolescents ingested more protein than adequate. Other studies indicate similar changes of macronutrient intake. Trend analyses with data of the DONALD study came to the result that overall fat intake, SFA and MUFA intakes of the subjects decreased, whereas carbohydrate intake increased between 1985 and 2000(Reference Alexy, Sichert-Hellert and Kersting22). As evaluations of the Bogalusa Heart Study indicate, trends in total fat, SFA and MUFA and percentage of energy from fat were negative among 10-year-old children between 1984–5 and 1992–4. Total carbohydrate intake and percentage of energy from carbohydrates remained similar(Reference Nicklas, Elkasabany, Srinivasan and Berenson24). Results of the Continuing Survey of Food Intake by Individuals (CSFII) showed a decline in percentage of energy from fat and protein which was compensated by an increase in carbohydrate intake among 11- to 18-year-olds(Reference Cavadini, Siega-Riz and Popkin26). In Northumbrian 11- to 12-year-olds, the percentage of energy from fat also decreased between 1980 and 2000, alongside an increase in percentage of energy from starch, but not from sugars(Reference Fletcher, Rugg-Gunn, Matthews, Hackett, Moynihan, Mathers and Adamson29, Reference Rugg-Gunn, Fletcher, Matthews, Hackett, Moynihan, Kelly, Adams, Mathers and Adamson30).
The present results indicate that energy intake was lower among children, but higher among adolescents in 2006 compared with 1985–8. On the one hand, this may be due to misreporting as discussed before. On the other hand, the observation may be explained by changed physical activity and body height and weight. Comparable studies support the finding that energy intake develops differently among children and adolescents. Among boys of the DONALD study energy intake decreased slightly, albeit significant only in the group of 14- to 18-year-olds. Among girls energy intake had a negative trend for 4- to 8-year-olds and a positive trend for 9- to 18-year-olds, however not significant(Reference Alexy, Sichert-Hellert and Kersting22). In the third National Health and Nutrition Examination Survey in the USA (NHANES III, 1988–94) energy intake of 6- to 11-year-old children remained stable, but increased for adolescents aged 12–19 years since NHANES II (1976–80)(Reference Troiano, Briefel, Carroll and Bialostosky31). The CSFII conducted in 1989–91 and 1994–6 among 11- to 18-year-olds came to the result that energy intake increased(Reference Cavadini, Siega-Riz and Popkin26). In contrast, energy intakes of 11- to 12-year-old boys and girls fell in Northumberland, England, between 1980 and 2000(Reference Fletcher, Rugg-Gunn, Matthews, Hackett, Moynihan, Mathers and Adamson29).
Besides absolute energy intake, energy density is of special interest because there is convincing evidence that energy-dense foods promote the development of overweight and obesity(19, 32). Consuming nutrient-dense foods instead of energy-rich but nutrient-poor foods helps to improve the nutrient:energy ratio(Reference Drewnowski33). From this point of view, it is favourable that energy density decreased but the nutrient density of diets increased. However, in both surveys, the energy density of food (without beverages) was relatively high compared with the cut-off point of 125 kcal (523 kJ)/100 g food of the second World Cancer Research Fund report(Reference Troiano, Briefel, Carroll and Bialostosky31). In the 1980s, median energy density of food was about 170 kcal (711 kJ)/100 g food among children, 185 kcal (774 kJ)/100 g food among adolescent boys and 179 kcal (749 kJ)/100 g food among adolescent girls. In 2006, median energy density of food in these groups was about 150 kcal (627 kJ)/100 g food, 161 kcal (673 kJ)/100 g food and 144 kcal (603kJ)/100 g food, respectively. Even the 25th percentiles of energy density were above the cut-off point in all age and sex groups in NVS and EsKiMo (data not shown).
In addition, the evaluation of the NVS and EsKiMo data revealed higher absolute intakes and higher densities of most vitamins and minerals in 2006 than in 1985–8. This is mainly due to changes in food intake and consumption of fortified foods. Among children of the DONALD study the intake of water-soluble vitamins and minerals from fortified foods increased significantly since the 1980s(Reference Sichert-Hellert, Kersting, Alexy and Manz23). In addition, the changes may be explained to a small extent by supplement intake. In EsKiMo, the inclusion of supplements slightly increased the median intakes and densities of some micronutrients (e.g. vitamins E and C, Ca, Mg) among adolescents, but not among children (data not shown). The D-A-CH recommendations for nutrient densities of vitamin B6, vitamin B12 and Zn were met in all age and sex groups in NVS and EsKiMo. In contrast, folate and vitamin D densities were too low compared with the recommendations in all age and sex groups in both surveys. Among girls, the densities of Fe and Ca were lower than recommended in both surveys. The recommendations for the intakes of vitamin D and folate were not met in any age and sex group in the two surveys. In addition, the intakes of vitamin E, Ca and vitamin A (only EsKiMo) were lower than recommended during childhood. Thus, the most critical nutrients observed in NVS and EsKiMo were folate, vitamin D, vitamin A, vitamin E, Ca and Fe. In the DONALD study, vitamin C density increased linearly and nutrient density of vitamin A remained constant between 1986 and 2001. Vitamin E intake increased from 1986 to 1993 and decreased afterwards(Reference Alexy, Kersting and Sichert-Hellert34). In the US CSFII, Fe intake was higher in 1994–6 than in 1989–91, while vitamin A and C, folate, Ca and fibre remained the same(Reference Cavadini, Siega-Riz and Popkin26). In Northumberland, vitamin A and C and Fe intake increased, while Ca intake decreased, among 11- to 12-year-olds between 1980 and 1990(Reference Adamson, Rugg-Gunn, Butler, Appleton and Hackett35).
Conclusion
Food and nutrient intakes of German children and adolescents changed significantly during the past 20 years. Macronutrient composition and micronutrient intake improved. However, the intake of polysaccharides and dietary fibre should be enhanced and mono- and disaccharide intake should be reduced in future. In addition, the intake of folate, vitamins D, A and E, Ca and Fe have to increase to meet the recommendations. The general trend of higher micronutrient content of the diet in relation to energy intake and lower energy density should proceed. To improve the energy and nutrient density of the diet, the intake of nutritious foods with high water content should be promoted. Further efforts will be necessary to allow well-balanced diets with more plant foods and energy-free beverages, but less animal, high-fat and high-sugar foods for all children and adolescents and to stabilise enduring favourable dietary habits. The results are of interest for all involved in health policy and responsible for nutrition campaigns.
Acknowledgements
Sources of funding: The study was supported by the German Federal Ministry of Food, Agriculture and Consumer Protection. This paper was conceived and written independently of the funding agency. Conflict of interest declaration: There are no conflicts of interest. Authorship responsibilities: G.B.M.M. and H.H. designed and managed EsKiMo; C.V., A.R. and A.S. were responsible for data collection and aggregation; A.S. conducted the presented data analysis and wrote the manuscript. All co-authors contributed critical revisions and gave final approval for submission. Acknowledgements: We would like to thank PD Dr K. Gedrich, who made the dataset of the NVS available. We are also very grateful to the families whose participation was crucial and to the colleagues who supported EsKiMo.
