The added sugar intake in relation to energy is, in general, higher among children than in other population groups and commonly above the limit of 10 % of energy intake(Reference Ruxton, Garceau and Cottrell1, Reference Lambert, Agostoni, Elmadfa, Hulshof, Krause, Livingstone, Socha, Pannemans and Samartin2). Studies among children have shown that some micronutrients tend to correlate inversely with sugar consumption(Reference Gibson3–Reference Øverby, Lillegaard, Johansson and Andersen6). Even though evidence suggests that a broad range of sugar intake has no detrimental effects on micronutrient intakes in most populations, it is acknowledged that the true risk of deficiencies can only be assessed using biological indicators(Reference Ruxton, Garceau and Cottrell1). Foods simultaneously fortified with vitamins and minerals and sweetened with added sugars could mask nutrient dilution(Reference Alexy, Sichert-Hellert and Kersting7). The association of added sugar intake with fat intake has turned out to be inverse (sugar–fat see-saw phenomenon) in children, indicating that added sugar intake at least partially replaces fat intake(Reference Gibson3–Reference Alexy, Sichert-Hellert and Kersting5). The majority of epidemiological studies have not demonstrated correlations between sugar consumption and obesity(Reference Ruxton, Garceau and Cottrell1). However, consumption of sugar-sweetened drinks could lead to obesity, due to imprecise and incomplete compensation for energy consumed in liquid form(Reference Ludwig, Peterson and Gortmaker8–Reference Malik, Schulze and Hu10). Dietary data on Finnish children are dispersed and sparse. The sugar intake has been presented only as total sugars or total sucrose(Reference Ylönen, Virtanen, Ala-Venna and Räsänen11–Reference Ruottinen, Karjalainen, Pienihäkkinen, Lagström, Niinikoski, Salminen, Rönnemaa and Simell13). Food preferences, which could become determined already by the age of 2–3 years, are strongly influenced by social, demographic and lifestyle factors related to the family, particularly to the mother (Reference Birch and Fisher14–Reference Patrick and Nicklas16). However, sociodemographic factors determining the consumption of added sugars are not well known among children.
In the present study the objective was to identify the most important sources of naturally occurring and added sucrose among 3-year-old children, and the most important determinants of high intake. Furthermore, we evaluated whether a high proportion of energy from added sucrose has an impact on the intake of foods and nutrients. In the present study, the term ‘added sucrose’ is used in reference to added sucrose eaten separately at the table or used as an ingredient in processed or prepared foods. It does not include naturally occurring sucrose which is calculated separately.
Subjects and methods
The subjects in the present study belong to the Type 1 Diabetes Prediction and Prevention Study (DIPP) cohort(Reference Kupila, Muona and Simell17) (http://research.utu.fi/dipp/index.php). Written consent for screening of genetic susceptibility to type 1 diabetes of their newborn infant was obtained from the parents. The subjects carrying the high- or moderate-risk human leucocyte antigen (HLA) class II genotypes were observed for diet, growth, viral infections and type 1 diabetes-associated auto-antibodies at 3- to 12-month intervals. The study was approved by the ethics committees of the participating hospitals. The present series comprises the at-risk children born in 2001 in Tampere and Oulu University Hospitals.
A total of 471 families (56 % of those 846 invited, and 61 % of those 773 who started in the DIPP Nutrition study) returned the structured dietary questionnaire and the 3 d food record at the age of 3 years of the child. For seven children (1·5 % of all) food records were kept for only 1 d and for twenty children (4 %) for 2 d. The background information was obtained from a questionnaire completed at 3 months after delivery. Data on gestation and delivery were obtained from the Medical Birth Registries of the Oulu and Tampere University Hospitals and from the National Research and Development Centre for Welfare and Health.
Dietary data
Structured dietary questionnaire at age 3 years and 3 d food records
Data on the child's food consumption were obtained by a 3 d food record completed close to the child's third birthday. The 3 d food record comprised two consecutive weekdays and one weekend day. Additional questions about special diets were attached. A separate food record was given to day-care personnel if the child was cared for outside the home during the recording days. The families and day-care personnel received written instructions to record (type, brand and preparation method) with household measures all the foods the child consumed and vitamin and mineral supplements used. Trained study nurses checked the questionnaire and food record during the respective visits.
The food consumption data were analysed using a software program developed at the National Public Health Institute (www.ktl.fi). The food composition database of the National Public Health Institute, called Fineli(18) (www.fineli.fi), is continuously updated and is the most comprehensive database in Finland. The analytical nutrient values in the database are mostly based on Finnish studies. In addition, complementary data are obtained from the Finnish food industry and international food composition tables. The system is able to accommodate the creation or modification of specific recipes, and personal recipes were used whenever possible.
For food consumption and nutrient intake, an estimate of average daily intake was calculated. Accordingly, the data on nutrient intake were analysed by food-use groups. The food-use groups were reclassified according to quantity and quality of sucrose (see the classification in Table 1). Sucrose use per child per d was calculated producing four different variables: naturally occurring sucrose, sucrose added by manufacturers, sucrose added by the consumer, and total sucrose (the sum of the aforementioned). In most of the analyses, added sucrose from manufacturers and sucrose from consumer sources were combined and named as added sucrose. The term ‘naturally occurring sucrose’ is used in reference to intrinsic sucrose that is enclosed in the cell in whole fruits and berries and naturally occurring sucrose in the products made out of fruits and berries (for example, fruit juices).
The average contribution of food groups to the intake of different sucrose classes among the children (n 471)
* The food groups are based on quantity and quality of sucrose. Other foods include all other foods which are not mentioned in the Table, such as meat, fish, etc.
† Three main contributors.
‡ Berry- or fruit-based drinks that contain naturally occurring and added sucrose.
Statistical analysis
The dietary variables were log-transformed when necessary to meet the assumptions of normal distribution. Adjustment for total energy intake was made using Willett's(Reference Willett19) residual method. Differences between means by sociodemographic characteristics were tested with the independent-samples t test (for two groups of cases) or ANOVA (for more than two groups of cases). The paired-samples t test was used to compare the means of sucrose intake on weekdays v. holidays. Pearson's correlation coefficients were used to study associations between fat and different types of sucrose (in % of energy intake). Children were divided into quartiles of proportion of energy from added sucrose. Differences between nutrient quartiles were assessed by one-way ANOVA. Differences in food intake between the quartiles were tested with the non-parametric Mann–Whitney U test (lowest and highest quartiles) and the Kruskal–Wallis test (all quartiles). A linear regression analysis was applied to study the intake of the proportion of energy from added sucrose in relation to selected sociodemographic characteristics. Sociodemographic characteristics were categorised as shown in Table 2. A P value of less than 0·05 was considered statistically significant. The SPSS 12.0 statistical package for Windows (SPSS Inc., Chicago, IL, USA) was used for the statistical analyses.
(Mean values and standard deviations)
* Differences between means by sociodemographic characteristics were tested with the independent-samples t test (for two groups of cases) or ANOVA (for more than two groups of cases).
† At the time of the birth of the child.
Results
The mean daily energy intake was 5353 (sd 988) kJ for boys and 5066 (sd 981) kJ for girls (P = 0·002). Protein accounted for 16 (sd 3) %, fat for 31 (sd 6) % and total carbohydrates for 53 (sd 6) % of the total energy intake. The energy-adjusted food consumption in boys and girls differed from each other mainly with regard to less frequently used food groups (tea, pulses, and snacks) and that of nutrient intake with regard to Fe and fluoride, which were higher among boys. No significant difference was seen in energy intake between the quartiles of added sucrose as a % of energy intake (Table 3).
(Mean values and standard deviations)
* Sum of sucrose added by manufacturers and consumers. For the lowest quartile, energy intake from added sucrose was < 7·85 % and for the highest quartile, energy intake from added sucrose was >14·5 %.
† Differences between nutrient quartiles were assessed by one-way ANOVA.
‡ Finnish Nutrition Recommendations, which are based on Nordic Nutrition Recommendations.
§ Total sugar = sucrose, fructose, lactose, maltose, galactose.
On average, the children consumed 41 (sd 18, range 6–109) g sucrose/d, accounting for 13·3 (sd 5) % of energy intake. Total sucrose intake was higher during the weekend days; the mean sucrose intake during the weekend days was 45 v. 39 g/d on weekdays (P < 0·001). The average intake of added sucrose by manufacturers and consumers was 35 (sd 17, range 1·5–102) g, accounting for 11·3 (sd 4·8, range 0·5–32) % of the energy intake (Table 2). Quartile points of added sucrose intake were 23, 34, and 44 g/d; and 7·8, 11·0 and 14·6 % of the energy intake respectively. More than half of the children (59 %; n 279) had a mean % of the energy intake of added sucrose below the limit of 10 % of energy intake. In 5 % (n 23) of the children, the proportion exceeded 20 % of energy intake; in three children, the proportion exceeded 25 % of energy intake. The average intake of naturally occurring sucrose was 6 (sd 3·8, range 0·3–22) g/d, accounting for 2 (sd 1·2) % of the energy intake (Table 2).
The major contributor to the total sucrose intake was sucrose added by manufacturers, with 82 % of the total sucrose among the children Tables 1 and 2. Naturally occurring sucrose contributed 15 % and sucrose added by consumers 3 % of total sucrose in all children studied. The food groups entitled ‘juice drinks’, ‘yoghurt and cultured milks’ and ‘chocolate and confectionery’ were the main contributors to added sucrose intake followed by ‘ice cream and milk desserts’. In all children, fresh fruits and berries contributed most to naturally occurring sucrose (Table 1).
Food and nutrient intake by quartiles of sucrose intake as percentage of energy intake
The consumption of rye bread, porridge, fresh vegetables, cooked potatoes, skimmed milk, hard cheeses and margarine and fat spread decreased across the quartiles of added sucrose as a % of energy intake (Table 4). In contrast, the consumption of sweet bakery, biscuits, breakfast cereals, fruit and berry soups, ice cream, milk desserts, yoghurt, juice drinks, confectionery and chocolate increased across the quartiles. However, the differences close to the mean (quartiles II and III) were not always clear. The difference in food consumption between the lowest and highest quartile of added sucrose ranged from 28 % for breakfast cereals to 381 % for juice drinks.
(Mean values and standard deviations)
* Sum of sucrose added by manufacturers and consumers. For the lowest quartile, energy intake from added sucrose was < 7·85 % and for the highest quartile, energy intake from added sucrose was >14·5 %.
† Percentage difference between Q1 and Q4 in micronutrient intake ((Q1 − Q4)/Q1).
‡ Differences between the quartiles were tested with the non-parametric Kruskal–Wallis test.
§ The dietary variables were log-transformed and adjusted for total energy intake using Willett's residual method(Reference Willett19).
Fat intake as a % of energy intake was inversely associated with sucrose intake as a % of energy intake (for total sucrose as a % of energy intake, r − 0·32; for naturally occurring sucrose as a % of energy intake, r − 0·29; for added sucrose as a % of energy intake, r − 0·26; P < 0·01 for all). Intake of protein and fat was highest in the lowest quartile of added sucrose as a % of energy intake and carbohydrate intake in the highest quartile (Table 3). Across the quartiles of added sucrose as a % of energy intake from the lowest to the highest, there was a decrease in intake of all analysed nutrients, except for vitamin A, pyridoxine, vitamin C and Cu (Table 5). The difference in micronutrient intake from food between the lowest and highest quartile of added sucrose was on average 16 %, ranging from 8 % for pyridoxine to 26 % for vitamin D. Among the children, 222 (47 %) used vitamin D supplements during the recording days. There was no difference in the proportion of supplement users between the quartiles. However, intake of vitamin D from supplements was highest in the lowest quartile of sucrose as a % of energy intake, and significantly so when only supplement users were compared (Table 5).
(Mean values and standard deviations)
* Sum of sucrose added by manufacturers and consumers. For the lowest quartile, energy intake from added sucrose was < 7·85 % and for the highest quartile, energy intake from added sucrose was >14·5 %.
† Differences between nutrient quartiles were assessed by one-way ANOVA.
‡ Percentage difference between Q1 and Q4 in micronutrient intake ((Q1 − Q4)/Q1).
§ Finnish Nutrition Recommendations, which are based on Nordic Nutrition Recommendations.
∥ Number of users: 60, 53, 52 and 57 in quartile groups I, II, III and IV respectively.
¶ Children < 3 years, 5–6 μg/d all year round or 10 μg/d if on milk-restricted diet; 3-year-olds, 5–6 μg/d during October–March.
Factors associated with high sucrose intake
Intake of added sucrose as a % of energy intake differed significantly when the following were taken into consideration: number of siblings, paternal professional education, and type of day-care (Table 2). The children who followed a special diet because of milk allergy or lactose intolerance (n 58; 12 % of all) had higher added sucrose intake as a % of energy intake compared with others (12·6 v. 11·2 % energy intake; P = 0·032). When all relevant characteristics are included in a linear regression model, type of day-care (home v. pre-school or kindergarten) and paternal professional education (vocational school v. academic education) were the strongest determinants of energy intake from added sucrose (Table 6).
(β Coefficients and 95 % confidence intervals)
* The model included all covariates presented in the Table.
Discussion
In the present study, sucrose added by manufacturers accounted for 82 %, naturally occurring sucrose for 15 % and sucrose added by consumers for 3 % of total sucrose in the diet of the children studied. Juice drinks, yoghurt and cultured milks, and chocolate and confectionery were the key contributors to the intake of added sucrose; fresh fruits and berries and fruit and berry juices were the key contributors to the intake of naturally occurring sucrose. A high proportion of added sucrose in the diet had an unfavourable impact on the intake of recommended foods and key nutrients in 3-year-old Finnish children. The type of day-care, paternal professional education and a milk-restricted diet were the strongest determinants of high added sucrose intake.
The majority of added sugars in the diet comprise of sucrose, but fructose, glucose as well as different syrups are also found in this group. Added sucrose calculated in the present study is, therefore, an underestimate of the total amount of added sugars, and is not directly comparable with studies providing estimates for added sugars. However, the consumption of added sucrose reliably predicts the total consumption of added sugars. In a German study(Reference Linseisen, Gedrich, Karg and Wolfram20) the mean contribution of sucrose to total energy intake in children aged 4–6 years was 14 %, of which 15–25 % was assumed to be accounted for by naturally occurring sucrose. In the present study, naturally occurring sucrose contributed 15 % to the total sucrose intake. About half of the children maintained their added sucrose intake below the recommended limit of 10 % energy intake(21). Since the recommendation applies to all ‘refined sugars’ (not only sucrose), the real proportion of children meeting the recommended limit is smaller.
The average total sucrose intake in the present study was remarkably higher than in the Finnish STRIP-intervention study in the 1990s(Reference Karjalainen, Söderling, Sewón, Lapinleimu and Simell12, Reference Ruottinen, Karjalainen, Pienihäkkinen, Lagström, Niinikoski, Salminen, Rönnemaa and Simell13) but lower than in Finnish 1–3-year-olds in a survey from the late 1980s(Reference Ylönen, Virtanen, Ala-Venna and Räsänen11). Between-country comparisons regarding the actual intake of added sugars are difficult and partly inaccurate due to discrepancies in the calculation or mode of expression. The proportion of added sugars from total energy in the diet of children was above 10 % of energy intake in studies in Denmark(Reference Lyhne and Ovesen22), Norway(Reference Øverby, Lillegaard, Johansson and Andersen6), Germany(Reference Alexy, Sichert-Hellert and Kersting5), UK(Reference Gibson3) and the USA(Reference Farris, Nicklas, Myers and Berenson4, Reference Krebs-Smith23). The consequence of this high intake on diet quality is still unclear. The lack of measurement and reported intake precision complicates further analyses on the health effects of high added sugar intake. There might only be a long-term effect, which is difficult to measure, that only appears later on in life when the habit of a high-sucrose diet is established.
As observed by others in cross-sectional surveys(Reference Gibson3–Reference Alexy, Sichert-Hellert and Kersting5), the effect of added sugar intake on fat intake was inverse, supporting the ‘sugar–fat see-saw phenomenon’. Energy intake did not differ with added sucrose intake, which supports earlier notions that the intake of added sucrose, at least partially, replaces fat intake in children's diets(Reference Øverby, Lillegaard, Johansson and Andersen6). Most of the micronutrients were inversely related to added sucrose intake as has also been observed in earlier studies(Reference Gibson3, Reference Farris, Nicklas, Myers and Berenson4, Reference Øverby, Lillegaard, Johansson and Andersen6, Reference Alexy, Sichert-Hellert and Kersting7). Among the children in the lowest quartile of added sucrose as a % of energy intake, the mean intake of most nutrients came closest to the recommended levels. However, intakes of most nutrients were adequate regardless of the level of added sucrose intake.
All groups had a mean of vitamin D (from food) and Fe intake that did not meet the Nordic Nutrition Recommendations(21), supporting earlier studies that found a shortfall of these nutrients among Finnish children(Reference Lamberg-Allardt and Viljakainen24–Reference Piirainen, Laitinen and Isolauri26). The nutritional consequences of an increased added sucrose intake are, therefore, most critical for decreased intakes of vitamin D and Fe. The differences in vitamin D and Fe intakes between the four quartiles of added sucrose as a % of energy intake are partly explained by a decreased consumption of fortified liquid milk products and margarines (vitamin D) and rye (Fe).
In the present study, the main food groups contributing to sucrose intake were ‘juice drinks’, ‘yoghurt and cultured milks’, ‘chocolate and confectionery’ and ‘ice cream and milk desserts’. These are the same as have been found in other child populations(Reference Alexy, Sichert-Hellert and Kersting5, Reference Øverby, Lillegaard, Johansson and Andersen6, Reference Ruottinen, Karjalainen, Pienihäkkinen, Lagström, Niinikoski, Salminen, Rönnemaa and Simell13, Reference Lyhne and Ovesen22). For vegetables there was a 42 % decrease in intake from the low to the high quartiles of added sucrose as a % of energy intake. Øverby et al. (Reference Øverby, Lillegaard, Johansson and Andersen6) found a very similar 45 % decrease among Norwegian 4-year-old children. The consequence of decreasing vegetable intake is critical, as Finnish children are known to consume remarkably low amounts of vegetables and fruits(Reference Talvia, Räsänen, Lagström, Pahkala, Viikari, Rönnemaa, Arffman and Simell27).
Having a milk-restricted diet seems to endow a Finnish 3-year-old child with greater odds for a high added sucrose intake. A substantial proportion of changes in the added sucrose intake by children can be attributed to changes in their beverage consumption patterns. Inverse associations between added sucrose intake and riboflavin and Ca point to a reduction in milk consumption, as shown earlier as well(Reference Alexy, Sichert-Hellert and Kersting7). It is an enormous challenge to replace milk in a child's diet with a drink providing similar nutrients without any increase in added sucrose intake.
It is valuable to speculate why the added sucrose intake was higher in children cared for at home. A snack-dominated meal pattern might be higher when cared for at home. Snacks appear to be associated with higher sucrose intake and lower intake of micronutrients(Reference Ovaskainen, Reinivuo, Tapanainen, Hannila, Korhonen and Pakkala28). The observed increased sucrose intake during weekends is another facet of the same issue. Compared with the present results, the difference in sucrose intake between the weekends and weekdays was even higher in 3–5-year-old Swedish children(Reference Sepp, Lennernäs, Petterson and Abrahamsson29). The difference may be due to differences in meal habits at home compared with day-care, or to different meal habits during weekends as compared with weekdays. In addition, there is an established tradition of ‘Saturday sweets’ for children in both Finnish and Swedish families.
Interestingly, we found an association between paternal, but not maternal, education and the intake of both naturally occurring and added sucrose by 3-year-old children. Paternal, but also maternal, education determined the sucrose intake in an earlier Finnish study(Reference Ruottinen, Karjalainen, Pienihäkkinen, Lagström, Niinikoski, Salminen, Rönnemaa and Simell13). An association between mothers', but not fathers', education and 4-year-olds' intake of sugar was found in an earlier Norwegian study(Reference Øverby, Lillegaard, Johansson and Andersen6). Overall, the impact of the paternal education on the eating pattern and dietary quality in the offspring is not as widely studied as is maternal education. It is important to note that naturally occurring and added sucrose related differently to the background variables. An overall, high intake of naturally occurring sucrose seems to be more related to better overall dietary quality than a high intake of added sucrose, which is clearly explained by their main dietary sources. Including sugars which occur naturally in fruit juices within the group of added sugars, as in the WHO statement(30), is questionable when estimating an overall effect of added sucrose on the dietary quality.
Certain limitations of the present study should be considered when interpreting the findings. The 3 d food records give an accurate estimate of usual intake for most frequently used foods such as porridge, milk and bread spreads. However, for occasionally used foods, many more days are required. The same applies to nutrients. In the present study with 3 d of diet records, the correlation coefficients between observed and true intake of nutrients ranged from 0·16 for vitamin A to 0·98 for Fe, being 0·65 for sucrose and 0·61 for added sucrose (M Erkkola, unpublished results; for the method, see Nelson et al. (Reference Nelson, Black, Morris and Cole31)). A 7 d diet record would have been needed to achieve r ≥ 0·8 (18 d for r ≥ 0·90) between observed and true sucrose intake. The ‘Saturday sweets’ habit contributes to daily variation in sucrose intake, and, consequently, a child's average sucrose intake could differ according to the type of days recorded. The families were advised to record two consecutive weekdays and one weekend day in order to obtain the influence of day of the week in sucrose intake. However, the small number of records may provide a standard deviation that is greatly overestimated. Furthermore, measurements of associations are substantially weakened.
We did not exclude potential under-reporters from the analysis. Diet during childhood tends to be highly variable from day to day, and the identification of reliable under-reporters is difficult. Ideally, we would have included anthropometric data, so that we had something against which to check reliability of reporting. However, data on children's weight and height were not available. One fourth of boys (26 %) and girls (25 %) had their average daily energy intake below the FAO, WHO & United Nations University energy requirements computed for moderate levels of physical activity (2–3-year olds: boys 4·7 and girls 4·4 MJ/d)(32). Based on within- and between-individual variability of energy among the subjects in the present study, 5 d of diet records are needed to achieve r ≥ 0·8 (12 d for r ≥ 0·9) between observed and true energy intake. In the present study, diet was recorded for 3 d, so r = 0·72.
In a Norwegian validation study among 2-year-old children, the food items under-reported were typically sucrose-rich foods such as cake, soft drinks and sweets, while the over-reported foods were more healthy foods such as bread, fruit and potatoes(Reference Andersen, Lande, Trygg and Hay33). In the present study, this could imply that the subjects in the top quartile of the added sucrose distribution were more likely to under-report their consumption of sucrose-rich foods than subjects in other quartiles. The true association could then be obscured and the influence of the consumption of added sucrose on intake of nutrients would be stronger than that observed.
Although the present cohort carries increased HLA-conferred susceptibility to type 1 diabetes, the children are expected to be representative of the general population of young Finnish children. Almost 20 % of the Finnish population have increased HLA-conferred predisposition to type 1 diabetes, while only 3–4 % of those actually progress to clinical disease(Reference Ilonen, Reijonen, Herva, Sjöroos, Iitiä, Lövgren, Veijola, Knip and Åkerblom34). The distribution of subjects by sociodemographic characteristics was comparable with the Finnish 3-year-olds in general(Reference Vuori and Gissler35).
In conclusion, the primary rationale for the recommendation to reduce the intake of refined sugars, to ensure adequate intakes of Fe and other essential nutrients as well as dietary fibre, seems reasonable. The recommendation for increased use of vegetables and whole-grain food products should also be promoted, not only to prevent extra added sucrose intake, but as an overall base for preventing obesity and chronic diseases.
Acknowledgements
We express our gratitude to the children and parents who participated. We are grateful to the DIPP research nurses, doctors, nutritionists and laboratory staff for excellent collaboration over the years. We thank Jennifer Burt Davis for language editing. The study was supported by the Academy of Finland (grants 63 672, 79 685, 79 686, 80 846, 201988, 210632), the Finnish Diabetes Association, the Finnish Diabetes Research Foundation, the Finnish Pediatric Research Foundation, the Häme Foundation of the Finnish Culture Fund, the Juho Vainio Foundation, the Yrjö Jahnsson Foundation, Danisco, Medical Research Funds, Turku, Oulu and Tampere University Hospitals, JDRF (grants 197032, 4–1998–274, 4–1999–731, 4–2001–435), Novo Nordisk Foundation and the EU Biomed 2 Program (BMH4-CT98–3314). S. M. V. designed the DIPP Nutrition Study and is responsible for the study. S. M. V. and M. E. designed the present study. M. E. drafted the manuscript, did the statistical analysis and together with C. K.-K. and H. T. did the special analysis concerning sucrose intake. M.-L. O. is the senior investigator in the DIPP Nutrition Study and is responsible for the food composition database. H. R. is also responsible for the food composition database. M. K. is the principal investigator of the DIPP study in Oulu and Tampere, and R. V. is the senior investigator of the DIPP study in Oulu. C. K.-K., P. K., J. L., H. T. and H. R. are responsible for the management and analysis of food consumption data. All the co-authors participated in the evaluation of the results and in editing the final manuscript.
