Precision Food Composition Data as a Tool to Decipher the Riddle of Ultra-Processed Foods and Nutritional Quality

BACKGROUND

Epidemiology supports a link between ultra-processed foods (UPFs) and health, mediated mainly through the clustering of foods with suboptimal nutrient profiles within UPFs. However, successful NOVA categorization requires access to a food's ingredient list, which we hypothesized can impact both UPF identification and the link between processing and composition.

METHODS

Foods (n = 4851) in the HelTH branded food composition database were classified as NOVA1-4, with or without using the ingredient lists (generic and branded approach, respectively), to identify differences in NOVA classification (chi-square test) and the estimated average nutritional composition of each NOVA group (Kruskal-Willis U test).

RESULTS

Using the ingredients list increased UPF identification by 30%. More than 30% of foods commonly assumed to be minimally processed (NOVA1-plain dairy, frozen vegetables, etc.) were reclassified as UPFs when using ingredient lists. These reclassified foods, however, had nutritional compositions comparable to NOVA1 foods and better than UPFs for energy, fat, sugars, and sodium (p < 0.001). In fact, UPFs did not show a uniform nutritional composition covering foods from Nutri-Score A (~10%) to Nutri-Score E (~20%).

CONCLUSIONS

The assumption that all UPFs have the same unfavorable nutritional composition is challenged when NOVA is applied using the appropriate branded food composition database.

How Nutritious Are French Beans (Phaseolus vulgaris L.) from the Citizen Science Experiment?

French beans are tender, immature, edible pods that are harvested early in the plant's growth cycle and are usually eaten cooked. The growth habits of French beans were studied for the first time in a Citizen Science experiment, and 19 pod samples were collected for further nutritional analysis. Various macronutrients (e.g., protein, ash, fat, carbohydrates, amino acids) and multi-element profiles were determined. A survey of their growing habits revealed that beans are usually planted once or twice a year in May and June at a length of 5-10 m, with a predominance of dwarf beans cultivation over climbing varieties, and pest resistance and stringless pods are the most important characteristics when deciding on a bean. Homogenised freeze-dried pod samples contained 16.1-23.1% protein, 4.5-8.2% ash, 0.1-1.1% fat, and 62.0-70.6% carbohydrates and had a caloric value of 337-363 kcal/100 g. Of the 17 free amino acids identified, 8 were essential (histidine, threonine, methionine, valine, lysine, isoleucine, leucine, phenylalanine) and 9 were non-essential (cysteine, aspartic acid, serine, glutamic acid, glycine, arginine, alanine, proline, tyrosine); meanwhile, of the 12 elements, 5 were macroelements and 7 were microelements. The predominant free amino acids were aspartic acid, glutamic acid, and serine. In the multiple comparisons (Box and Whisker plot), the parameters caloric value and iron showed the strongest response. A very strong positive significant Pearson correlation (≥0.95) was found for five pairs of variables within the free amino acids. Comparison of the nutrient data obtained in the pods showed near-perfect or high complementarity (85.2-103.4%) with the food composition databases for half of the parameters, suggesting that the home-grown French beans from the Citizen Science experiment are a highly nutritious vegetable.

Determining classes of food items for health requirements and nutrition guidelines using Gaussian mixture models

Introduction

The identification of classes of nutritionally similar food items is important for creating food exchange lists to meet health requirements and for informing nutrition guidelines and campaigns. Cluster analysis methods can assign food items into classes based on the similarity in their nutrient contents. Finite mixture models use probabilistic classification with the advantage of taking into account the uncertainty of class thresholds.

Methods

This paper uses univariate Gaussian mixture models to determine the probabilistic classification of food items in the South African Food Composition Database (SAFCDB) based on nutrient content.

Results

Classifying food items by animal protein, fatty acid, available carbohydrate, total fibre, sodium, iron, vitamin A, thiamin and riboflavin contents produced data-driven classes with differing means and estimates of variability and could be clearly ranked on a low to high nutrient contents scale. Classifying food items by their sodium content resulted in five classes with the class means ranging from 1.57 to 706.27 mg per 100 g. Four classes were identified based on available carbohydrate content with the highest carbohydrate class having a mean content of 59.15 g per 100 g. Food items clustered into two classes when examining their fatty acid content. Foods with a high iron content had a mean of 1.46 mg per 100 g and was one of three classes identified for iron. Classes containing nutrient-rich food items that exhibited extreme nutrient values were also identified for several vitamins and minerals.

Discussion

The overlap between classes was evident and supports the use of probabilistic classification methods. Food items in each of the identified classes were comparable to allowed food lists developed for therapeutic diets. This data-driven ranking of nutritionally similar classes could be considered for diet planning for medical conditions and individuals with dietary restrictions.

Accelerating the Classification of NOVA Food Processing Levels Using a Fine-Tuned Language Model: A Multi-Country Study

The consumption and availability of ultra-processed foods (UPFs), which are associated with an increased risk of noncommunicable diseases, have increased in most countries. While many countries have or are planning to incorporate UPF recommendations in their national dietary guidelines, the classification of food processing levels relies on expertise-based manual categorization, which is labor-intensive and time-consuming. Our study utilized transformer-based language models to automate the classification of food processing levels according to the NOVA classification system in the Canada, Argentina, and US national food databases. We showed that fine-tuned language models using the ingredient list text found on food labels as inputs achieved a high overall accuracy (F1 score of 0.979) in predicting the food processing levels of Canadian food products, outperforming traditional machine learning models using structured nutrient data and bag-of-words. Most of the food categories reached a prediction accuracy of 0.98 using a fined-tuned language model, especially for predicting processed foods and ultra-processed foods. Our automation strategy was also effective and generalizable for classifying food products in the Argentina and US databases, providing a cost-effective approach for policymakers to monitor and regulate the UPFs in the global food supply.

Food Composition Databases (FCDBs): A Bibliometric Analysis

Food composition databases (FCDBs) are important tools that provide information on the nutritional content of foods. Previously, it was largely unclear what nutritional contents and which FCDBs were involved in highly cited papers. The bibliometric study aimed to identify the most productive authors, institutions, and journals. The chemicals/chemical compounds with high averaged citations and FCDBs used by highly cited papers were identified. In July 2023, the online database Web of Science Core Collection (WoSCC) was queried to identify papers related to FCDBs. A total of 803 papers were identified and analyzed. The first paper indexed in WoSCC was published in 1992 by Pennington, which described the usefulness of FCDB for researchers to identify core foods for their own studies. In that paper, the FCDB described was the USDA 1987-88 NFCS (the United States Department of Agriculture 1987-88 Nationwide Food Consumption Survey). The most productive author was Dr. Paul M. Finglas, the Head of the Food Databanks National Capability at the Quadram Institute (Norwich, UK) and the Managing Director of EuroFIR. His most cited paper among this dataset was about the development of an online Irish food composition database together with EuroFIR. The most productive institutions were the USDA and the World Health Organization (WHO) instead of universities. Flavonoid was the most recurring chemical class among the highly cited ones. The anti-oxidative properties and protective effects against heart disease and cancer of flavonoids might be some of the reasons for their popularity in research. Among the highly cited papers, the most heavily used FCDBs were the USDA database for the flavonoid content of selected foods, Fineli, the USDA National Nutrient Database for Standard Reference (USNDB), EuroFIR eBASIS-Bioactive Substances in Food Information Systems, and Phenol-Explorer. High-quality national and international FCDBs should be promoted and made more accessible to the research and public communities to promote better nutrition and public health on a global scale.

Systematic Review of the Literature to Inform the Development of a South African Dietary Polyphenol Composition Database

Comprehensively compiled dietary polyphenol data is required to compare polyphenol content between foods, calculate polyphenol intake and study its association with health and disease. The purpose of this review was to identify data on the presence and content of polyphenolic components in South African foods, with the aim of compiling the data into a database. An electronic literature search was conducted up until January 2020 using multiple databases. Additional literature was sourced from South African university repositories. A total of 7051 potentially eligible references were identified, of which 384 met the inclusion criteria. These studies provided information on food item name, geographical distribution, polyphenol type, quantity, and quantification method. Data for 1070 foods were identified, amounting to 4994 polyphenols. Spectrophotometry was the main method used for quantification of gross phenolic content in various assays such as total phenolic content (Folin-Ciocalteu assay), total flavonoid content (AlCl₃ assay) and condensed tannin content (vanillin-HCl assay). Phenolic acids and flavonoids were the main polyphenol classes identified. This review highlights that South Africa has abundant information on the polyphenol content of foods, which could be utilised within a food composition database for the estimation of polyphenol intake for South Africa.

Natural language processing and machine learning approaches for food categorization and nutrition quality prediction compared with traditional methods

BACKGROUND

Food categorization and nutrient profiling are labor intensive, time consuming, and costly tasks, given the number of products and labels in large food composition databases and the dynamic food supply.

OBJECTIVES

This study used a pretrained language model and supervised machine learning to automate food category classification and nutrition quality score prediction based on manually coded and validated data, and compared prediction results with models using bag-of-words and structured nutrition facts as inputs for predictions.

METHODS

Food product information from University of Toronto Food Label Information and Price Database 2017 (n = 17,448) and University of Toronto Food Label Information and Price Database 2020 (n = 74,445) databases were used. Health Canada's Table of Reference Amounts (TRA) (24 categories and 172 subcategories) was used for food categorization and the Food Standards of Australia and New Zealand (FSANZ) nutrient profiling system was used for nutrition quality score evaluation. TRA categories and FSANZ scores were manually coded and validated by trained nutrition researchers. A modified pretrained sentence-Bidirectional Encoder Representations from Transformers model was used to encode unstructured text from food labels into lower-dimensional vector representations, followed by supervised machine learning algorithms (i.e., elastic net, k-Nearest Neighbors, and XGBoost) for multiclass classification and regression tasks.

RESULTS

Pretrained language model representations utilized by the XGBoost multiclass classification algorithm reached overall accuracy scores of 0.98 and 0.96 in predicting food TRA major and subcategories, outperforming bag-of-words methods. For FSANZ score prediction, our proposed method reached a similar prediction accuracy (R²: 0.87 and MSE: 14.4) compared with bag-of-words methods (R²: 0.72-0.84; MSE: 30.3-17.6), whereas structured nutrition facts machine learning model performed the best (R²: 0.98; MSE: 2.5). The pretrained language model had a higher generalizable ability on the external test datasets than bag-of-words methods.

CONCLUSIONS

Our automation achieved high accuracy in classifying food categories and predicting nutrition quality scores using text information found on food labels. This approach is effective and generalizable in a dynamic food environment, where large amounts of food label data can be obtained from websites.

Food composition databases in the era of Big Data: Vegetable oils as a case study

Understanding the population's dietary patterns and their impacts on health requires many different sources of information. The development of reliable food composition databases is a key step in this pursuit. With them, nutrition and health care professionals can provide better public health advice and guide society toward achieving a better and healthier life. Unfortunately, these databases are full of caveats. Focusing on the specific case of vegetable oils, we analyzed the possible obsolescence of the information and the differences or inconsistencies among databases. We show that in many cases, the information is limited, incompletely documented, old or unreliable. More importantly, despite the many efforts carried out in the last decades, there is still much work to be done. As such, institutions should develop long-standing programs that can ensure the quality of the information on what we eat in the long term. In the face of climate change and complex societal challenges in an interconnected world, the full diversity of the food system needs to be recognized and more efforts should be put toward achieving a data-driven food system.

The Update of the Italian Food Composition Database of Gluten-Free Products and Its Application in Food-Based Dietary Guidelines Menus

Complete food composition databases (FCDBs) on gluten-free (GF) foods are needed to assess the nutrient intakes of celiac disease patients. The aim of the present work was to update the previously developed version of the Italian GF-FCDB and to apply it to a theoretical GF diet. The updated GF-FCDB includes the composition of 108 GF cereal-based foods, as sold, in terms of energy and macro- and micro-nutrients, imputed using food label information combined with the standard recipe approach. Three scenarios (i.e., refined, mixed, and wholegrain cereals) of the weekly guideline menu for the general Italian population were analyzed for energy and nutrient content in a theoretical dietary assessment using traditional gluten-containing (GC) foods and the corresponding GF substitutes. All GF menus were higher than the corresponding GC menus in polyunsaturated fatty acids, linoleic acid, and vitamin E. Zinc was lower in GF than in GC menus only in the wholegrain-cereal scenario. Thanks to the application of the updated GF-FCDB including a comprehensive list of micronutrients, we observed that it is possible for celiac disease patients to meet nutrient requirements by simply substituting GC with GF cereal-based products following recommendations for the general population.

Evaluation of the New Individual Fatty Acid Dataset for UK Biobank: Analysis of Intakes and Sources in 207,997 Participants

The Oxford WebQ is an online 24 h dietary assessment tool used by several large prospective studies. This study describes the creation of the new individual fatty acid (FA) dataset for the Oxford WebQ and reports intakes and sources of dietary individual FAs in the UK Biobank. Participants who completed ≥1 (maximum of five) 24 h dietary assessments were included (n = 207,997). Nutrient intakes were obtained from the average of all completed 24 h dietary assessments. Nutrient data from the UK McCance and Widdowson's The Composition of Foods and the US Department of Agriculture food composition tables were used to calculate intakes of 21 individual FAs. The individual FA dataset included 10 saturated fatty acids (SFAs), 4 monounsaturated fatty acids (MUFAs), and 7 polyunsaturated fatty acids (PUFAs; including alpha-linolenic (18:3), eicosapentaenoic (20:5), and docosahexaenoic (22:6) acids). Palmitic (16:0; mean ± standard deviation (SD): 13.5 ± 5.7 g/d) and stearic (18:0; 5.2 ± 2.5) acids were the main contributors to SFAs, and the main sources of these were cereals and cereal products (mostly desserts/cakes/pastries), milk and milk products (mostly cheese and milk), and meat and meat products. Oleic acid (18:1; 24.2 ± 9.8) was the main MUFA, derived mainly from cereals and cereal products, and meat and meat products. Linoleic acid (18:2; 9.7 ± 4.3) was the main PUFA, derived mostly from cereals and cereal products, and vegetables (including potatoes) and vegetable dishes. The individual FA dataset for the Oxford WebQ will allow future investigations on individual FAs and disease risk.

Development of a Canadian Food Composition Database of Gluten-Free Products

Country-specific food composition data are needed for gluten-free (GF) food products to assess nutritional adequacy and diet quality. This research aimed to develop a comprehensive GF food composition database for key GF foods consumed in Canada. Average nutrient data from 167 products were estimated from Nutrition Fact Panel labels and the commercial ingredient list, using an iterative and systematic approach. The database reports mean values for energy and 29 nutrients per 100 g for 33 GF commercial grain-based foods. Nutrient values were evaluated with Health Canada’s nutrient content claims per standard reference serving. On average, GF products were, at minimum, a source of thiamin (73%), riboflavin (70%), niacin (58%), iron (58%), fibre (55%), magnesium (48%), folate (36%), zinc (19%), and calcium (15%). Most GF products were low in saturated fat (85%) and cholesterol (64%) but only 15% were low in total fat and 6% were free of sugar. Micronutrient enrichment and the use of nutrient-dense whole grain flours, legume flours, oil seed husks, and functional fibre ingredients varied within and between categories and brands but appeared to contribute to nutrient content. This database provides a new tool to enhance GF diet assessment in individuals or populations in Canada.

Statistical Methods for the Analysis of Food Composition Databases: A Review

Evidence-based knowledge of the relationship between foods and nutrients is needed to inform dietary-based guidelines and policy. Proper and tailored statistical methods to analyse food composition databases (FCDBs) could assist in this regard. This review aims to collate the existing literature that used any statistical method to analyse FCDBs, to identify key trends and research gaps. The search strategy yielded 4238 references from electronic databases of which 24 fulfilled our inclusion criteria. Information on the objectives, statistical methods, and results was extracted. Statistical methods were mostly applied to group similar food items (37.5%). Other aims and objectives included determining associations between the nutrient content and known food characteristics (25.0%), determining nutrient co-occurrence (20.8%), evaluating nutrient changes over time (16.7%), and addressing the accuracy and completeness of databases (16.7%). Standard statistical tests (33.3%) were the most utilised followed by clustering (29.1%), other methods (16.7%), regression methods (12.5%), and dimension reduction techniques (8.3%). Nutrient data has unique characteristics such as correlated components, natural groupings, and a compositional nature. Statistical methods used for analysis need to account for this data structure. Our summary of the literature provides a reference for researchers looking to expand into this area.

Review of dietary assessment studies conducted among Khmer populations living in Cambodia

Background

Despite economic growth, Cambodia continues to have high rates of malnutrition, anaemia and nutrition‐related deficiencies. Government policies promote nutrition strategies, although dietary intake data is limited. A detailed synthesis of existing intake data is needed to inform nutrition policy and practice change. This review aims to characterise and assess quality of dietary assessment methods and outcomes from individual‐level ‘whole diet’ studies of Khmer people living in Cambodia.

Methods

Searches were conducted using PRISMA‐ScR guidelines. Included papers reported dietary intake at an individual level for ‘whole diet’. Studies using secondary data or lacking dietary assessment details were excluded. Extracted data included dietary assessment features, nutrient/food group intakes and database.

Results

Nineteen publications (15 studies) were included, with nine carried out among children under 5 years and six among women. Eleven studies reported intake by food groups and four by nutrients, prominently energy, protein, vitamin A, iron, calcium and zinc. Inconsistent intakes, food groupings and reporting of study characteristics limited data synthesis. All but one study used 24‐h recalls. Trained local fieldworkers used traditional interview‐administered data collection and varied portion estimation tools. Food composition databases for analysis were not tailored to the Cambodian diet. Overall quality was rated as ‘good’.

Conclusions

We recommend the development of a best‐practice protocol for conducting dietary assessment, a Cambodia‐specific food composition database and a competent trained workforce of nutrition professionals, with global support of expertise and funding for future dietary assessment studies conducted in Cambodia.

Fifteen studies with highly variable intake data included in the review.

The food composition databases used were not specific to Cambodian diet.

Minimum reporting standards and best practice protocols recommended, including in‐country nutrition training.

Lack of whole population dietary intake data indicates the need for a national survey.

Branded Foods Databases as a Tool to Support Nutrition Research and Monitoring of the Food Supply: Insights From the Slovenian Composition and Labeling Information System

Branded foods databases are becoming very valuable not only in nutrition research but also for clinical practice, policymakers, businesses, and general population. In contrast to generic foods, branded foods are marked by rapid changes in the food supply because of reformulations, the introduction of new foods, and the removal of existing ones from the market. Also, different branded foods are available in different countries. This not only complicates the compilation of branded foods datasets but also causes such datasets to become out of date quickly. In this review, we present different approaches to the compilation of branded foods datasets, describe the history and progress of building and updating such datasets in Slovenia, and present data to support nutrition research and monitoring of the food supply. Manufacturers are key sources of information for the compilation of branded foods databases, most commonly through food labels. In Slovenia, the branded food dataset is compiled using standard food monitoring studies conducted at all major retailers. Cross-sectional studies are conducted every few years, in which the food labels of all available branded foods are photographed. Studies are conducted using the Composition and Labeling Information System (CLAS) infrastructure, composed of a smartphone application for data collection and online data extraction and management tool. We reviewed various uses of branded foods datasets. Datasets can be used to assess the nutritional composition of food in the food supply (i.e., salt, sugar content), the use of specific ingredients, for example, food additives, for nutrient profiling, and assessment of marketing techniques on food labels. Such datasets are also valuable for other studies, for example, assessing nutrient intakes in dietary surveys. Additional approaches are also being tested to keep datasets updated between food monitoring studies. A promising approach is the exploitation of crowdsourcing through the mobile application VešKajJeš, which was launched in Slovenia to support consumers in making healthier dietary choices.

Development of an Unified Food Composition Database for the European Project "Stance4Health"

The European Commission funded project Stance4Health (S4H) aims to develop a complete personalised nutrition service. In order to succeed, sources of information on nutritional composition and other characteristics of foods need to be as comprehensive as possible. Food composition tables or databases (FCT/FCDB) are the most commonly used tools for this purpose. The aim of this study is to describe the harmonisation efforts carried out to obtain the Stance4Health FCDB. A total of 10 FCT/FCDB were selected from different countries and organizations. Data were classified using FoodEx2 and INFOODS tagnames to harmonise the information. Hazard analysis and critical control points analysis was applied as the quality control method. Data were processed by spreadsheets and MySQL. S4H’s FCDB is composed of 880 elements, including nutrients and bioactive compounds. A total of 2648 unified foods were used to complete the missing values of the national FCDB used. Recipes and dishes were estimated following EuroFIR standards via linked tables. S4H’s FCDB will be part of the smartphone app developed in the framework of the Stance4Health European project, which will be used in different personalized nutrition intervention studies. S4H FCDB has great perspectives, being one of the most complete in terms of number of harmonized foods, nutrients and bioactive compounds included.

A Community-Based Approach to Integrating Socio, Cultural and Environmental Contexts in the Development of a Food Database for Indigenous and Rural Populations: The Case of the Batwa and Bakiga in South-Western Uganda

Comprehensive food lists and databases are a critical input for programs aiming to alleviate undernutrition. However, standard methods for developing them may produce databases that are irrelevant for marginalised groups where nutritional needs are highest. Our study provides a method for identifying critical contextual information required to build relevant food lists for Indigenous populations. For our study, we used mixed-methods study design with a community-based approach. Between July and October 2019, we interviewed 74 participants among Batwa and Bakiga communities in south-western Uganda. We conducted focus groups discussions (FGDs), individual dietary surveys and markets and shops assessment. Locally validated information on foods consumed among Indigenous populations can provide results that differ from foods listed in the national food composition tables; in fact, the construction of food lists is influenced by multiple factors such as food culture and meaning of food, environmental changes, dietary transition, and social context. Without using a community-based approach to understanding socio-environmental contexts, we would have missed 33 commonly consumed recipes and foods, and we would not have known the variety of ingredients’ quantity in each recipe, and traditional foraged foods. The food culture, food systems and nutrition of Indigenous and vulnerable communities are unique, and need to be considered when developing food lists.

Identifying Nutrient Patterns in South African Foods to Support National Nutrition Guidelines and Policies

Food composition databases (FCDBs) provide the nutritional content of foods and are essential for developing nutrition guidance and effective intervention programs to improve nutrition of a population. In public and nutritional health research studies, FCDBs are used in the estimation of nutrient intake profiles at the population levels. However, such studies investigating nutrient co-occurrence and profile patterns within the African context are very rare. This study aimed to identify nutrient co-occurrence patterns within the South African FCDB (SAFCDB). A principal component analysis (PCA) was applied to 28 nutrients and 971 foods in the South African FCDB to determine compositionally similar food items. A second principal component analysis was applied to the food items for validation. Eight nutrient patterns (NPs) explaining 73.4% of the nutrient variation among foods were identified: (1) high magnesium and manganese; (2) high copper and vitamin B₁₂; (3) high animal protein, niacin, and vitamin B₆; (4) high fatty acids and vitamin E; (5) high calcium, phosphorous and sodium; (6) low moisture and high available carbohydrate; (7) high cholesterol and vitamin D; and (8) low zinc and high vitamin C. Similar food patterns (FPs) were identified from a PCA on food items, yielding subgroups such as dark-green, leafy vegetables and, orange-coloured fruit and vegetables. One food pattern was associated with high sodium levels and contained bread, processed meat and seafood, canned vegetables, and sauces. The data-driven nutrient and food patterns found in this study were consistent with and support the South African food-based dietary guidelines and the national salt regulations.

Updated Database and Trends of Declared Low- and No-Calorie Sweeteners From Foods and Beverages Marketed in Spain

Background: The past few years have witnessed an increase in the availability of food products containing one or more low- and no-calorie sweeteners (LNCS) in the Spanish market, mostly due to the new massive reformulation plan. However, these are not included in food composition tables or databases, and, therefore, assessment of their intake among the population is complex. This study aims to update a database including commercialized foods and beverages.

Method: A systematic search of ingredients information from the different food and beverage categories was undertaken during 2019 by recording the availability and type of LNCS declared in the information of the product from labels and online shopping platforms of retailers from Spain to update a previous food composition database compiled in 2017.

Results: A total of 1,238 products were identified. The major groups were sugar and sweets (24%), non-alcoholic beverages (21%), cereals and grains (19%), and milk and dairy products (14%) accounting for >70% of total products. The mainly declared LNCS were sorbitol (19.5%), sucralose (19.5%), and acesulfame K (19.2%).

Conclusion: There is a wide variety of products that include LNCS as a main ingredient with higher availability than when compared with the results of database of 2017, consequently, it might be expected that LNCS are commonly consumed at present in the Spanish diet.

The Design and Development of a Food Composition Database for an Electronic Tool to Assess Food Intake in New Caledonian Families

The food environment in New Caledonia is undergoing a transition, with movement away from traditional diets towards processed and discretionary foods and beverages. This study aimed to develop an up-to-date food composition database that could be used to analyze food and nutritional intake data of New Caledonian children and adults. Development of this database occurred in three phases: Phase 1, updating and expanding the number of food items to represent current food supply; Phase 2, refining the database items and naming and assigning portion size images for food items; Phase 3, ensuring comprehensive nutrient values for all foods, including saturated fat and total sugar. The final New Caledonian database comprised a total of 972 food items, with 40 associated food categories and 25 nutrient values and 615 items with portion size images. To improve the searchability of the database, the names of 593 food items were shortened and synonyms or alternate spelling were included for 462 foods. Once integrated into a mobile app-based multiple-pass 24-h recall tool, named iRecall.24, this country-specific food composition database would support the assessment of food and nutritional intakes of families in New Caledonia, in a cross-sectional and longitudinal manner, and with translational opportunities for use across the wider Pacific region.

A New Food Composition Database of Lactose-Free Products Commercialized in Spain: Differences in Nutritional Composition as Compared to Traditional Products

We developed a new database to evaluate the nutritional composition of lactose-free products from Spain. The database includes dairy products and other products, all of which show the "lactose-free" declaration on their label, accounting for 327 products in total. Of these, 123 are dairy products, 16 are non-dairy products which include a dairy ingredient (5%) and 188 items (57% of the sample) are non-dairy products that do not contain any dairy ingredient. The main subgroups are yogurt (25%), milk (24%), and cheese (17%). Nineteen percent of the compiled products included nutritional claims on their labels. Most lactose-free products did not contain either added sugars or low- or no-calorie sweeteners (58%), while 34% included added sugars and only 6%, sweeteners or a combination of both (2%). We found that 19.5%, mainly within the milk subgroup, were fortified with vitamins A, D, E, K, B₉, and B₁₂, P, and Ca. There were no significant differences in the nutritional composition between lactose-free products and traditional products. According to the NOVA classification, 55% of compiled lactose-free products were ultra-processed, and 20% processed. The array of lactose-free products marketed in Spain proves that there are enough, both in quantity and quality, to satisfy the dairy needs of lactose intolerants.

Retrospectively Estimating Energy Intake and Misreporting From a Qualitative Food Frequency Questionnaire: An Example Using Australian Cohort and National Survey Data

Qualitative food frequency questionnaires (Q-FFQ) omit portion size information from dietary assessment. This restricts researchers to consumption frequency data, limiting investigations of dietary composition (i.e., energy-adjusted intakes) and misreporting. To support such researchers, we provide an instructive example of Q-FFQ energy intake estimation that derives typical portion size information from a reference survey population and evaluates misreporting. A sample of 1,919 Childhood Determinants of Adult Health Study (CDAH) participants aged 26-36 years completed a 127-item Q-FFQ. We assumed sex-specific portion sizes for Q-FFQ items using 24-h dietary recall data from the 2011-2012 Australian National Nutrition and Physical Activity Survey (NNPAS) and compiled energy density values primarily using the Australian Food Composition Database. Total energy intake estimation was daily equivalent frequency × portion size (g) × energy density (kJ/g) for each Q-FFQ item, summed. We benchmarked energy intake estimates against a weighted sample of age-matched NNPAS respondents (n = 1,383). Median (interquartile range) energy intake was 9,400 (7,580-11,969) kJ/day in CDAH and 9,055 (6,916-11,825) kJ/day in weighted NNPAS. Median energy intake to basal metabolic rate ratios were 1.43 (1.15-1.78) in CDAH and 1.35 (1.03-1.74) in weighted NNPAS, indicating notable underreporting in both samples, with increased levels of underreporting among the overweight and obese. Using the Goldberg and predicted total energy expenditure methods for classifying misreporting, 65 and 41% of CDAH participants had acceptable/plausible energy intake estimates, respectively. Excluding suspected CDAH misreporters improved the plausibility of energy intake estimates, concordant with expected body weight associations. This process can assist researchers wanting an estimate of energy intake from a Q-FFQ and to evaluate misreporting, broadening the scope of diet-disease investigations that depend on consumption frequency data.

Evaluation of the Ability of Diet-Tracking Mobile Applications to Estimate Energy and Nutrient Intake in Japan

We evaluated the energy and nutrient intake estimates of popular Japanese diet-tracking mobile applications (apps). We identified five diet-tracking apps in the iTunes store during August 2020. A researcher entered the dietary data from a one-day paper-based dietary record (DR) previously obtained from apparently healthy free-living adults (15 males and 15 females; 22-65 years) into each app. The energy and nutrient intakes estimated by the apps were compared with those calculated using the Standard Tables of Food Composition in Japan based on the paper-based DR (reference method). The number of dietary variables available ranged from one (energy in Mogutan) to 17 (FiNC). Compared to the DR-based estimates, the median energy intake was significantly overestimated by MyFitnessPal, Asken, Calomiru, and Mogutan. Moreover, the intakes of many nutrients were overestimated by Asken and Calomiru and underestimated by MyFitnessPal. For energy intake, the Spearman correlation coefficient between the DR and the apps was lowest for Mogutan (0.76) and highest for FiNC (0.96). The median correlation coefficient for nutrient intakes was lower in MyFitnessPal (0.50) than in the other three apps (0.80 in Asken, 0.87 in FiNC, and 0.88 in Calomiru). These results suggest that intake calculations differ among apps. Further evaluation is needed in free-living settings, where users input their own food intake.

Comparing Calculated Nutrient Intakes Using Different Food Composition Databases: Results from the European Prospective Investigation into Cancer and Nutrition (EPIC) Cohort

This study aimed to compare calculated nutrient intakes from two different food composition databases using data from the European prospective investigation into cancer and nutrition (EPIC) cohort. Dietary intake data of the EPIC cohort was recently matched to 150 food components from the U.S. nutrient database (USNDB). Twenty-eight of these nutrients were already included in the EPIC nutrient database (ENDB-based upon country specific food composition tables), and used for comparison. Paired sample t-tests, Pearson's correlations (r), weighted kappa's (κ) and Bland-Altman plots were used to compare the dietary intake of 28 nutrients estimated by the USNDB and the ENDB for 476,768 participants. Small but significant differences were shown between the USNDB and the ENDB for energy and macronutrient intakes. Moderate to very strong correlations (r = 0.60-1.00) were found for all macro- and micronutrients. A strong agreement (κ > 0.80) was found for energy, water, total fat, carbohydrates, sugar, alcohol, potassium and vitamin C, whereas a weak agreement (κ < 0.60) was found for starch, vitamin D and vitamin E. Dietary intakes estimated via the USNDB compare adequately with those obtained via the ENDB for most macro- and micronutrients, although the agreement was weak for starch, vitamin D and vitamin E. The USNDB will allow exposure assessments for 150 nutrients to investigate associations with disease outcomes within the EPIC cohort.

Updated Food Composition Database for Cereal-Based Gluten Free Products in Spain: Is Reformulation Moving on?

We developed a comprehensive composition database of 629 cereal-based gluten free (GF) products available in Spain. Information on ingredients and nutritional composition was retrieved from food package labels. GF products were primarily composed of rice and/or corn flour, and 90% of them included added rice starch. The most common added fat was sunflower oil (present in one third of the products), followed by palm fat, olive oil, and cocoa. Only 24.5% of the products had the nutrition claim “no added sugar”. Fifty-six percent of the GF products had sucrose in their formulation. Xanthan gum was the most frequently employed fiber, appearing in 34.2% of the GF products, followed by other commonly used such as hydroxypropyl methylcellulose (23.1%), guar gum (19.7%), and vegetable gums (19.6%). Macronutrient analysis revealed that 25.4% of the products could be labeled as a source of fiber. Many of the considered GF food products showed very high contents of energy (33.5%), fats (28.5%), saturated fatty acids (30.0%), sugars (21.6%), and salt (28.3%). There is a timid reformulation in fat composition and salt reduction, but a lesser usage of alternative flours and pseudocereals.

NIH workshop on human milk composition: summary and visions

Nationally representative data from mother-child dyads that capture human milk composition (HMC) and associated health outcomes are important for advancing the evidence to inform federal nutrition and related health programs, policies, and consumer information across the governments in the United States and Canada as well as in nongovernment sectors. In response to identified gaps in knowledge, the National Institute of Diabetes and Digestive and Kidney Diseases of the NIH sponsored the "Workshop on Human Milk Composition-Biological, Environmental, Nutritional, and Methodological Considerations" held 16-17 November 2017 in Bethesda, Maryland. Through presentations and discussions, the workshop aimed to 1) share knowledge on the scientific need for data on HMC; 2) explore the current understanding of factors affecting HMC; 3) identify methodological challenges in human milk (HM) collection, storage, and analysis; and 4) develop a vision for a research program to develop an HMC data repository and database. The 4 workshop sessions included 1) perspectives from both federal agencies and nonfederal academic experts, articulating scientific needs for data on HMC that could lead to new research findings and programmatic advances to support public health; 2) information about the factors that influence lactation and/or HMC; 3) considerations for data quality, including addressing sampling strategies and the complexities in standardizing collection, storage, and analyses of HM; and 4) insights on how existing research programs and databases can inform potential visions for HMC initiatives. The general consensus from the workshop is that the limited scope of HM research initiatives has led to a lack of robust estimates of the composition and volume of HM consumed and, consequently, missed opportunities to improve maternal and infant health.

Antioxidant Properties of Four Commonly Consumed Popular Italian Dishes

Four popular dishes belonging to Italian cuisine and widely consumed in the country were experimentally prepared in a dedicated lab-kitchen following a validated and standardized protocol. This study provides their antioxidant properties evaluating the contribution of extractable and non-extractable bioactive compounds, and identifying the assessment of interactions between their natural active compounds and the food matrix. Ferric reducing antioxidant power (FRAP) values in aqueous-organic extract ranged from the highest antioxidant activity in torta di mele (10.72 µmol/g d.m.) to that in besciamella (2.47 µmol/g d.m.); in residue, pasta alla carbonara reached the highest value (73.83 µmol/g d.m.) following by that in pasta alla amatriciana (68.64 µmol/g d.m.). Total polyphenol content (TPC) ranged in aqueous-organic extracts between 36.50 and 64.28 mg/100 g d.m. and in residue from 425.84 to 1747.35 mg/100 g d.m. Our findings may contribute to the updating of the Italian Food Composition Database, by providing for the first time a value for the antioxidant properties. This could contribute to encourage the consumption of recipes rich in key nutrients and bioactive molecules. This information is useful and important for determining the association between diet and a healthy status.

Evaluation of Reliability of the Recomputed Nutrient Intake Data in the National Heart, Lung, and Blood Institute Twin Study

The nutrient intake dataset is crucial in epidemiological studies. The latest version of the food composition database includes more types of nutrients than previous ones and can be used to obtain data on nutrient intake that could not be estimated before. Usual food consumption data were collected among 910 twins between 1969 and 1973 through dietary history interviews, and then used to calculate intake of eight types of nutrients (energy intake, carbohydrate, protein, cholesterol, total fat, and saturated, monounsaturated, and polyunsaturated fatty acids) in the National Heart, Lung, and Blood Institute Twin Study. We recalculated intakes using the food composition database updated in 2008. Several different statistical methods were used to evaluate the validity and the reliability of the recalculated intake data. Intra-class correlation coefficients between recalculated and original intake values were above 0.99 for all nutrients. R² values for regression models were above 0.90 for all nutrients except polyunsaturated fatty acids (R² = 0.63). In Bland⁻Altman plots, the percentage of scattering points that outlay the mean plus or minus two standard deviations lines was less than 5% for all nutrients. The arithmetic mean percentage of quintile agreement was 78.5% and that of the extreme quintile disagreement was 0.1% for all nutrients between the two datasets. Recalculated nutrient intake data is in strong agreement with the original one, supporting the reliability of the recalculated data. It is also implied that recalculation is a cost-efficient approach to obtain the intake of nutrients unavailable at baseline.

Association of Free Sugar Intake Estimated Using a Newly-Developed Food Composition Database With Lifestyles and Parental Characteristics Among Japanese Children Aged 3-6 Years: DONGuRI Study

BACKGROUND

The lack of comprehensive food composition databases for sugar contents in Japanese foods has led to the lack of nutritional epidemiologic studies on sugar intake in Japanese population. This cross-sectional study aimed to investigate the association of free sugar intake estimated using a newly developed food composition database with the characteristics and lifestyles of Japanese children aged 3-6 years.

METHODS

The food composition database contained information on sugars in 2,222 commonly consumed Japanese foods. Using this database, we estimated the sugar (total, added, and free sugars) intakes derived from a 3-day weighed dietary record of 166 boys and 166 girls aged 3-6 years living in 24 prefectures in Japan.

RESULTS

The mean free sugar intake was 26.8 g/d (standard deviation [SD], 12.3 g/d), while the mean value for energy intake was 7.8% (SD, 3.2%). The prevalence of excessive free sugar intake (≥10% of energy intake) was 21.7%. Among the characteristics and lifestyles examined, screen time was most strongly associated with the prevalence of excessive free sugar intake: multivariate adjusted odds ratios for screen time <0.5, ≥0.5 to <1, and ≥1 h/d were 1.0 (reference), 3.81 (95% confidence interval, 1.04-13.98), and 4.36 (95% confidence interval, 1.16-16.35), respectively. Additionally, younger age, shorter sleep, and mothers with office work and service and sales jobs (compared with those with professional and managerial jobs) were significantly associated with a higher prevalence of excessive free sugar intake.

CONCLUSIONS

This study showed the sugar intake of Japanese children aged 3-6 years is positively associated with screen time.

Development of a harmonized food grouping system for between-country comparisons in the TEDDY Study

The Environmental Determinants of Diabetes in the Young (TEDDY) is an international study aiming to investigate associations between dietary and other environmental factors and the risk of developing islet autoimmunity and type 1 diabetes. Dietary intake was assessed using a 24-hour recall and repeated 3-day food records and analyzed using country-specific food composition databases (FCDBs) in Finland, Germany, Sweden, and the U.S. with respective in-house calculation programs. A food grouping harmonization process between four country-specific FCDBs was conducted to evaluate and achieve comparability on food group definitions and quantification of food consumption across the countries. Systematic review revealed that the majority of existing food groups of the TEDDY FCDBs were not comparable. Therefore, a completely new classification system of 15 mutually exclusive main food groups (e.g. vegetables) and 89 subgroups (e.g. root vegetables, leafy vegetables) was developed. Foods and beverages were categorized into basic foods (single ingredient) and composite dishes (multiple ingredients). Composite dishes were broken down to ingredients using food composition data available in the FCDBs or generic recipes created for the harmonization effort. The daily consumption of every food group across FCDBs was quantified consistently as either raw or prepared weight depending on the food group to achieve maximal comparability.

Testing the Capacity of a Multi-Nutrient Profiling System to Guide Food and Beverage Reformulation: Results from Five National Food Composition Databases

Nutrient profiling ranks foods based on their nutrient composition, with applications in multiple aspects of food policy. We tested the capacity of a category-specific model developed for product reformulation to improve the average nutrient content of foods, using five national food composition datasets (UK, US, China, Brazil, France). Products (n = 7183) were split into 35 categories based on the Nestlé Nutritional Profiling Systems (NNPS) and were then classified as NNPS 'Pass' if all nutrient targets were met (energy (E), total fat (TF), saturated fat (SFA), sodium (Na), added sugars (AS), protein, calcium). In a modelling scenario, all NNPS Fail products were 'reformulated' to meet NNPS standards. Overall, a third (36%) of all products achieved the NNPS standard/pass (inter-country and inter-category range: 32%-40%; 5%-72%, respectively), with most products requiring reformulation in two or more nutrients. The most common nutrients to require reformulation were SFA (22%-44%) and TF (23%-42%). Modelled compliance with NNPS standards could reduce the average content of SFA, Na and AS (10%, 8% and 6%, respectively) at the food supply level. Despite the good potential to stimulate reformulation across the five countries, the study highlights the need for better data quality and granularity of food composition databases.

Voluntary Folic Acid Fortification Levels and Nutrient Composition of Food Products from the Spanish Market: A 2011-2015 Update

INTRODUCTION

Folic acid (FA) is a synthetic compound commonly added for voluntary fortification of food products in many European countries. In our country, food composition databases (FCDB) lack comprehensive data on FA fortification practices and this is considered a priority research need when undergoing nutritional assessment of the population.

METHODS

A product inventory was collected and updated by visiting retail stores in Madrid Region, conducting online supermarket searches, and by the provision of food label information by manufacturers. Euro-FIR FCDB guidelines for data compilation and harmonization were used.

RESULTS

The FCDB, compiled between 2011 and 2015, includes FA as well as macro and micronutrient data from 338 fortified foodstuffs. As compared to previous FCDB updates (May 2010), 37 products have ceased to declare added FA in their labels, mainly yogurt and fermented milk products. The main food subgroup is 'breakfast cereals' (n = 95, 34% of total). However, the highest average FA fortification levels per recommended serving were observed in the 'milk, milk products, and milk substitutes' group at ≥35% FA Nutrient Reference Values (NRV, 200 µg, EU Regulation 1169 of 2011) (60-76.3 µg FA per 200 mL). Average contribution to the FA NRV per food group and serving ranged between 16%-35%.

CONCLUSION

Our data show a minor decrease in the number of FA fortified products, but vitamin levels added by manufacturers are stable in most food groups and subgroups. This representative product inventory comprises the main FA food source from voluntary fortification in our country. It is therefore a unique compilation tool with valuable data for the assessment of dietary intakes for the vitamin.

Composition and Nutrient Information of Non-Alcoholic Beverages in the Spanish Market: An Update

The aim of this study was to draw an updated map of the nutrition facts in the different categories of non-alcoholic beverages in the Spanish market based on the information declared on the labels of these products; we expect this first step to justify the need for the coordination and harmonization of food composition tables in Spain so that there will be an updated database available to produce realistic scientific nutrient intake estimates in accordance with the actual market scenario. Materials and Methods: The nutrition facts declared on the labels of non-alcoholic beverages by manufacturers in Spain were compiled and studied. Results: The database included 211 beverages classified in 7 groups with energy/carbohydrate content per 100 mL ranging from 0–55 kcal/0–13 g for soft drinks; 2–60 kcal/0–14.5 g for energy drinks; 24–31 kcal/5.8–7.5 g for sports drinks; 1–32 kcal/0–7.3 g for drinks containing mineral salts in their composition; 14–69 kcal/2.6–17 g for fruit juice, nectar, and grape musts; 43–78 kcal/6.1–14.4 g for vegetable drinks; and 33–88 kcal/3.6–14 g for dairy drinks. Conclusion: The current non-alcoholic beverage market is a dynamic, growing, and highly innovative one, allowing consumers to choose according to their preferences, needs, or level of physical activity at any moment of the day.

Development of a New Branded UK Food Composition Database for an Online Dietary Assessment Tool

The current UK food composition tables are limited, containing ~3300 mostly generic food and drink items. To reflect the wide range of food products available to British consumers and to potentially improve accuracy of dietary assessment, a large UK specific electronic food composition database (FCDB) has been developed. A mapping exercise has been conducted that matched micronutrient data from generic food codes to “Back of Pack” data from branded food products using a semi-automated process. After cleaning and processing, version 1.0 of the new FCDB contains 40,274 generic and branded items with associated 120 macronutrient and micronutrient data and 5669 items with portion images. Over 50% of food and drink items were individually mapped to within 10% agreement with the generic food item for energy. Several quality checking procedures were applied after mapping including; identifying foods above and below the expected range for a particular nutrient within that food group and cross-checking the mapping of items such as concentrated and raw/dried products. The new electronic FCDB has substantially increased the size of the current, publically available, UK food tables. The FCDB has been incorporated into myfood24, a new fully automated online dietary assessment tool and, a smartphone application for weight loss.

Standardization of the Food Composition Database Used in the Latin American Nutrition and Health Study (ELANS)

Between-country comparisons of estimated dietary intake are particularly prone to error when different food composition tables are used. The objective of this study was to describe our procedures and rationale for the selection and adaptation of available food composition to a single database to enable cross-country nutritional intake comparisons. Latin American Study of Nutrition and Health (ELANS) is a multicenter cross-sectional study of representative samples from eight Latin American countries. A standard study protocol was designed to investigate dietary intake of 9000 participants enrolled. Two 24-h recalls using the Multiple Pass Method were applied among the individuals of all countries. Data from 24-h dietary recalls were entered into the Nutrition Data System for Research (NDS-R) program after a harmonization process between countries to include local foods and appropriately adapt the NDS-R database. A food matching standardized procedure involving nutritional equivalency of local food reported by the study participants with foods available in the NDS-R database was strictly conducted by each country. Standardization of food and nutrient assessments has the potential to minimize systematic and random errors in nutrient intake estimations in the ELANS project. This study is expected to result in a unique dataset for Latin America, enabling cross-country comparisons of energy, macro- and micro-nutrient intake within this region.

International collaborative project to compare and track the nutritional composition of fast foods

BACKGROUND

Chronic diseases are the leading cause of premature death and disability in the world with over-nutrition a primary cause of diet-related ill health. Excess quantities of energy, saturated fat, sugar and salt derived from fast foods contribute importantly to this disease burden. Our objective is to collate and compare nutrient composition data for fast foods as a means of supporting improvements in product formulation.

METHODS/DESIGN

Surveys of fast foods will be done in each participating country each year. Information on the nutrient composition for each product will be sought either through direct chemical analysis, from fast food companies, in-store materials or from company websites. Foods will be categorized into major groups for the primary analyses which will compare mean levels of saturated fat, sugar, sodium, energy and serving size at baseline and over time. Countries currently involved include Australia, New Zealand, France, UK, USA, India, Spain, China and Canada, with more anticipated to follow.

DISCUSSION

This collaborative approach to the collation and sharing of data will enable low-cost tracking of fast food composition around the world. This project represents a significant step forward in the objective and transparent monitoring of industry and government commitments to improve the quality of fast foods.

Estimation of trans fatty acid intake in Japanese adults using 16-day diet records based on a food composition database developed for the Japanese population

BACKGROUND

The Standard Tables of Food Composition in Japan do not include information on trans fatty acids. Previous studies estimating trans fatty acid intake among Japanese have limitations regarding the databases utilized and diet assessment methodologies. We developed a comprehensive database of trans fatty acid food composition, and used this database to estimate intake among a Japanese population.

METHODS

The database was developed using analytic values from the literature and nutrient analysis software encompassing foods in the US, as well as values estimated from recipes or nutrient compositions. We collected 16-day diet records from 225 adults aged 30 to 69 years living in 4 areas of Japan. Trans fatty acid intake was estimated based on the database and the 16-day diet records.

RESULTS

Mean total fat and trans fatty acid intake was 56.9 g/day (27.7% total energy) and 1.7 g/day (0.8% total energy), respectively, for women and 66.8 g/day (25.5% total energy) and 1.7 g/day (0.7% total energy) for men. Trans fatty acid intake accounted for greater than 1% of total energy intake, which is the maximum recommended according to the World Health Organization, in 24.4% of women and 5.7% of men, and was particularly high among women living in urban areas and those aged 30-49 years. The largest contributors to trans fatty acid intake were confectionaries in women and fats and oils in men.

CONCLUSIONS

Although mean trans fatty acid intake was below the maximum recommended intake of the World Health Organization, intake among subgroups was of concern. Further public health efforts to reduce trans fatty acid intake should be encouraged.