Intake and sources of dietary fibre and dietary fibre fractions in Finnish children · Salo TEI, Niinistö S, Korhonen TE et al.

Methods

The participants of the current study were sampled from the Type 1 Diabetes Prediction and Prevention (DIPP) Nutrition Study. The DIPP Nutrition Study was conducted as part of the main DIPP Study, which is a population-based birth cohort of Finnish children with HLA-conferred susceptibility to type 1 diabetes assessed from cord blood samples. All children with either medium or high susceptibility to type 1 diabetes born in Oulu University Hospital between September 1996 and September 2004, or Tampere University Hospital between 1997 and 2004 were invited to take part in the DIPP and DIPP Nutrition Studies. Participants of the DIPP Nutrition Study had information on dietary intake collected with 3-day food records until the age of 6 years in addition to the standard processes of the main DIPP Study. Information on sociodemographic factors was collected after the delivery with a structured questionnaire filled out by the parents.

This study assessed children’s dietary fibre intake at the ages of 6 months, 1, 3 and 6 years. All children in the DIPP Nutrition Study with at least one complete 3-day food record at any of the studied age points were included in the sample. 6-month-olds and 1-year-olds with missing growth information necessary for breast milk intake estimation were excluded at these age points. Children who developed autoimmunity were included.

Assessment of dietary intake

Three-day food records were collected at all the studied age points. The process of collecting the food records has been described in detail earlier(19). Open-ended food records with detailed instructions and pre-set dates of three consecutive days, of which 2 weekdays and 1 weekend day were provided, with separate copies for the caretakers and for day care personnel. Household measures, package sizes and weights when available were used for portion size estimation. All food records were checked upon return by trained research nurses. Dietary data were handled by trained research nutritionists. The food record data were entered into Finessi, an in-house dietary calculation software of the Finnish Institute for Health and Welfare. Finessi utilises data from the Finnish National Food Composition Database Fineli, which is maintained and updated annually by the Finnish Institute for Health and Welfare.

The database Fineli includes both food ingredients, e.g. rolled oats, as well as dishes compiled from ingredients via individual recipes, e.g. oatmeal, which includes the ingredients rolled oats, iodised salt, and 1·5 % fat milk with added vitamin D. The ingredient information for dishes is commonly based on ingredient and nutrition values declared by the manufacturer for commercial foods, and information from commonly used cookery books in Finland for non-commercial foods. Nutrient values for dishes are further based on the comprehensive nutrient values of the individual ingredients. The database also takes into account the physicochemical effects the cooking process has on the nutritional values of the dishes. There are roughly 2000 ingredients and over 3000 compiled dishes in Fineli, depending on the version. Different versions of the database are used simultaneously for cohort data such as this study, since processing methods, ingredients of commercial foods, food fortification levels and products available on the market are time dependent, and the dietary data are collected over several years. Each food record was processed with a version of the database matching the time of collection.

Because of the structure of the database, and the fact that the dietary calculation software Finessi in linked with Fineli, all dietary data inserted into Finessi can be broken down into food ingredients. Furthermore, all food ingredients belong to an ingredient class, for example spinach and lettuce into the ingredient class ‘leaf vegetables’. Breaking all foods down into ingredients and summarising them according to ingredient classes (for example, combining all ingredients from ingredient classes ‘leaf vegetables’, ‘root vegetables’ and ‘fruit vegetables’ as well as other desired vegetable classes into the food source classification ‘vegetables’) enables carrying out source analysis, in which the intake of a nutrient deriving from different food sources is assessed.

In our study, we assessed the food sources of total dietary fibre as well as dietary fibre fractions according to the following food source classifications: WheatRyeOatOther cereal products (barley, rice and other cereals)Fruits and berries (includes juices)Vegetables (includes mushrooms)PotatoesLegumes, nuts and seedsFormulaOthers (all ingredients not included above: meats, dairy products, fats, condiments, sweets, etc. May include some fibre sources, e.g. chocolate, ketchup.)

In this study, breast-feeding status was determined in 6-month-olds and 1-year-olds based on whether they had received any breast milk on the recorded 3 days. Caretakers primarily recorded the frequency of breast-feeding, not the amount of breast milk consumed. The amount of breast milk was calculated retrospectively using estimated energy requirements based on child weight and the growth of the child, and the energy intake from other foods based on the food records(20). Energy and dietary fibre from breast milk were included in energy and dietary fibre intake for 6-month-olds and 1-year-olds. After the age of 1 year breast milk intake was not considered due to very infrequent and little use by older children. We did not include dietary fibre supplements.

Compiling of dietary fibre values

The Finnish Food Authority and the Finnish Institute for Health and Welfare were responsible for the process of updating dietary fibre values of Fineli. As mentioned previously, Fineli consists of both food ingredients and dishes. Since dishes are compiled from food ingredients, it was only necessary to produce updated dietary fibre values for food ingredients. Updating the values for food ingredients would allow recalculation of the values for dishes as well.

The Finnish Food Authority conducted novel dietary fibre analyses for 110 food ingredients with the AOAC 2011·25 method, which can separately analyse IDF, SDFP, and SDFS (TDF = IDF + SDFP + SDFS)(21,22). Values for additional ninety-six food ingredients were extracted from previous literature. Of these literature-based values, eighteen were based on AOAC 2009·01 (which separates HMWDF (IDF combined with SDFP), and SDFS) or AOAC 2011·twenty-five analyses, thirty-seven on AOAC 991·forty-three analyses (HMWDF, underestimates resistant starch, inulin and resistant maltodextrins) and forty one on other methods. After the analyses and literature review, values for 134 food ingredients were imputed from related foods. Values for additional thirty-nine food ingredients were derived from previous analyses carried out by the Finnish Food Authority. Finally, values for 1839 food ingredients could be calculated based on all the above and the recipes of the individual food ingredients. Not only dishes include recipes in Fineli, but food ingredients too, e.g. apple jam, which is at the same time a food ingredient, and has a recipe consisting of sugar, ‘water, tap water’ and ‘apple, average, with skin’.

Dietary fibre content of breast milk was based on previous literature on human milk oligosaccharides (HMO)(23). Although HMO secretion into breast milk varies between individuals as well as over the period of lactation, we estimated 100 grams of breast milk to contain 1·3 grams of HMO, which is a value previously reported for mature milk (120+ days of lactation)(23). Dietary fibre in breast milk was inserted into Fineli as SDFS since HMO are NDO.

Statistical methods

Dietary fibre intake was analysed both in grams per day as well as energy-adjusted for grams per megajoule. Medians and interquartile ranges were calculated to describe the intake of dietary fibre in different subgroups of the sample. Pearson’s chi-square test was used to test for differences in the frequencies of background variables between the age groups in order to assess differences in drop-out rate.

Age differences in absolute and energy-adjusted intake of total dietary fibre as well as dietary fibre fractions were tested separately for girls and boys. Sex differences in dietary fibre intake were tested separately according to age group and possible breastfeeding status subgroup within the age group. Differences in dietary fibre intake associated with breast-feeding status were tested separately according to age group and sex.

Age differences were assessed using univariate analysis of variance. Tukey’s honest significance difference post hoc test was used to determine which age groups differed in the age group tests. Sex and breastfeeding status differences were assessed with independent samples t test and the non-parametric Mann–Whitney U-Test as appropriate for data distribution. Data was normally distributed for all dietary fibre fractions in 3-year-olds and 6-year-olds but skewed for all fractions in 6-month-olds and SDFS in 1-year-olds.

Differences in energy-adjusted intake of total dietary fibre according to socio-demographic factors in 1-, 3-, and 6-year-olds were assessed using independent samples t test for binary variables, and univariate analysis of variance for non-binary variables. Breastfed 1-year-olds were excluded from the analysis to reduce within-group variance. Food sources of total dietary fibre and dietary fibre fractions were determined by source analysis, utilising the ingredient level dietary data as described previously. Statistical significance was predetermined as P-values ≤ 0·05. Multiple testing was controlled for using the false discovery rate (FDR) method(24). Statistical analyses were carried out in IBM SPSS Statistics 27.

Ethical considerations

Informed consent is required for the genetic screening of the DIPP study. Parents of the DIPP study children have had high participation incentive. The ethics committees of all the universities and university hospitals participating in the DIPP study have approved the study.

The current study was observational and utilised already existing data from the DIPP Nutrition Study. Carrying out the study did not cause additional participation burden for the study participants but instead allowed the high-quality data that was produced with the combined effort of the participants and research staff to be analysed as exhaustively as possible.