Diet quality refers to the assessment of an individual’s overall dietary intake, measured against the foods and nutrients recommended by dietary guidelines, such as national food guides(Reference Gicevic, Mou and Bromage1,2) . A healthy diet is characterised by the consumption of fresh and minimally processed foods, including whole grains, fruits and vegetables. These foods are nutrient-dense, providing essential protective components such as fibre, vitamins and minerals, while containing lower levels of potentially harmful elements including saturated fats, sugars and salt(Reference Trijsburg, Talsma and de Vries3–5).
Global shifts in food systems have not only contributed to the weakening adherence to healthy dietary patterns in line with dietary recommendations but also led to an increase in the consumption of ultra-processed foods (UPF)(Reference Popkin6–Reference Louzada, Cruz and Silva8). According to the Nova classification system, UPF are industrially formulated products that are ready-to-eat and manufactured from the combination of substances extracted or derived from foods along with sensory additives to enhance the taste and palatability of the products(Reference Monteiro, Cannon and Levy9).
A systematic review demonstrated that diets predominantly composed of UPF have low nutritional quality. This is mainly due to the high densities of harmful nutrients such as saturated fat, sugar and Na and energy found in such foods(Reference Martini, Godos and Bonaccio10). Therefore, the consumption of UPF has also been related to many negative health outcomes, such as weight gain, as well as the occurrence of several non-communicable diseases(Reference Hall, Ayuketah and Brychta11–Reference Lane, Gamage and Du14). Considering these aspects, studies evaluating the consumption of UPF are useful for monitoring these dietary risk factors, supporting food and nutrition educational strategies and promoting public policies that encourage healthy eating(Reference Monteiro, Cannon and Lawrence15).
International agencies such as the FAO of the UN, the Pan American Health Organization and the International Network for Food and Obesity/Non-communicable Diseases Research, Monitoring and Action Support have suggested the use of energy contribution of UPF (expressed as % of total energy) as an indicator to monitor the nutritional quality of diets(16–Reference Vandevijvere, Monteiro and Krebs-Smith18). Studies conducted in Australia, Korea and Belgium examining the relationship between the energy contribution of UPF and the nutritional quality of diets have shown age-related differences, with findings indicating higher consumption of UPF and a decline in diet quality particularly among children and adolescents(Reference Marchese, Livingstone and Woods19–Reference Vandevijvere, De Ridder and Fiolet21).
In Brazil, Louzada et al. demonstrated that UPF consumption has a detrimental effect on overall diet quality(Reference Louzada, Ricardo and Steele22). The 2017–2018 Household Budget Survey (HBS) further revealed that younger individuals have a higher energy intake from UPF(23). However, no studies have yet examined diet quality specifically in relation to UPF consumption across different age groups in Brazil. While population-based research such as the ISA-Capital survey highlighted that adolescents tend to have the poorest diet quality compared with adults and older adults(Reference Andrade, Previdelli and Cesar24), there is still a gap in understanding how UPF consumption affects dietary quality across different age groups. The present study addresses this gap by conducting a stratified analysis by age group, investigating the relationship between UPF consumption and diet quality in adolescents, adults and older adults using data representative of the Brazilian population.
Methods
Background
The 2017–2018 HBS is the sixth conducted by the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística – IBGE, in Portuguese), which assesses the structures of consumption, expenditures, income and part of the asset variation of the households, providing a profile of the life conditions of the population based on the analysis of the household budgets(25). IBGE is the official Brazilian Population Statistics Agency and the main provider of data and information about the country. The last two HBS included the Brazilian National Dietary Survey (NDS), which assessed individual food consumption in a subsample of the households surveyed in the HBS, with data from all individuals aged 10 years and over living in the households. The assessment, compilation and structuring of this study counted on the contribution of technicians from the Ministry of Health, which added the vision of Nutrition specialists, of acknowledged experience and skill, who contributed both to the design of the project and the production of the publication, as well as to the data validation up to the final step of the analysis(23).
All these surveys used a secondary database that is publicly and freely available. The microdata, which includes data, documentation, questionnaires, table translators, reading software and calculation memory, is freely available on the IBGE website: https://www.ibge.gov.br/en/statistics/experimental-investigations/experimental-statistics/25610-pof-2017–2018-pof-en.html?edicao=28652&t=microdados. Microdata ensures confidentiality by omitting identifiable information, such as household addresses, telephone numbers and census tract numbers. Brazilian census data are protected by law (Law No. 13,709/2018 – General Data Protection Law). Supplementary Law No. 105/2001), which ensures that confidential information is not made available to the public. The IBGE was conducted in the 2017–2018 HBS and NDS following the ethical principles of the Declaration of Helsinki.
Study design and sampling
This is a cross-sectional study using data from the Brazilian NDS and the HBS, both conducted by the IBGE in 2017–2018. The NDS was conducted simultaneously in a random subsample of 34·7 % of the HBS. The HBS used the IBGE Integrated Household Survey System, which selects a ‘Master Sample’, defined as a set of census tracts that cover the entire national territory. The complex sample was selected in two stages. The Primary Sampling Units (PSU) were derived from the master sample, which consists of a set of census sectors. The PSU were selected by sampling with probability proportional to the number of households in each sector. The census sectors were previously obtained by government administrative divisions, urban or rural settings and income levels based on the Demographic Census of the year 2010. In the second stage, the households were selected by simple random sampling. The sectors were distributed over the four-quarters of the year in which the research was carried out. For the assessment of individual food consumption, a subsample of households was selected from the sample by simple random sampling. The 2017–2018 HBS consisted of 5504 selected sectors and 57 920 investigated households, with 20 112 households randomly selected of the 57 920 households surveyed in the HBS. All individuals aged 10 years and over in the households surveyed in the 2017–2018 HBS were invited to participate in the NDS(25). The final sample included 46 164 individuals aged 10 years or older (Fig. 1). Additional information on the sampling process and the definition of the master sample can be obtained from official publications of the IBGE(23,25) .
Sample flow chart in 2017–2018 Household Budget Survey and Brazilian National Dietary Survey. 1Households randomly selected from the predefined stratification system. An average loss of 15 % was estimated due to possible refusals to answer the survey, and the same proportion was added to the final number of households to minimise possible losses. 2Households randomly selected from the predefined stratification system.
Data collection
Data collection was carried out over 12 months, from July 2017 to July 2018. The information collected in the 2017–2018 HBS was based on seven modules. For this study, only the modules 1 and 7 were used: the former assessed sociodemographic characteristics and the latter assessed individual food consumption. Further detailed information about data collection can be found in a previous publication by IBGE(23,25) .
Sociodemographic characteristics
The sociodemographic characteristics considered in this study were sex (male and female), age groups (10–19 years; 20–59 years and 60 years and over), years of education (0–4 years, 5–8 years, 9–11 years and 12 years and above), household per capita income (in quartiles), household arrangement (single-person, couple, couple with children, single parent with children or mixed) and macro-regions (North, Northeast, Southeast, South and Midwest). All these characteristics were answered by the reference person in the household.
Age groups were based on the 2017–2018 NDS and the WHO classification(23,26) . Per capita income was estimated by dividing the total family income (estimated from the sum of monthly monetary and non-monetary income of all family members) by the number of individuals in the family. The conversion of the Brazilian currency to US dollars was conducted based on the exchange rate as of 31 January 2018(27). Per capita income was then divided into quartiles including a per capita monthly family income of less than US$ 226 (1st quarter), from US$ 226 to less than US$ 391 (2nd quarter), from US$ 391 to less than US$ 678 (3rd quarter) and equal to or greater than US$ 678 (4th quarter).
The variable ‘household arrangement’ was constructed based on the residents’ relationship, that is, their degree of kinship or nature of the existing subordination with the reference person (who was responsible for the household). The spouse was the resident who lived conjointly with the reference person. The child/children were those considered as legitimate, adopted or brought up by the reference person and/or spouse. The arrangement ‘single parent with children’ was composed of the reference person in the household of both sexes and with at least one child. Mixed households were composed of other members with or without any degree of kinship with the reference person or spouse (e.g. son-in-law, daughter-in-law, parents, grandparents, grandchildren, siblings, domestic workers and other relatives). Regarding the macro-regions of the country, the twenty-six states were divided into five regions as proposed by IBGE (South, South East, North, Northeast and Midwest)(25).
Individual food consumption
Individual food consumption data were collected using two 24-h recalls on non-consecutive days during the week when a trained IBGE agent was present in the household. This collection process was uniform across the entire sample to ensure representativeness in each quarter of the year and to account for seasonal variations in food consumption.
Both 24-h periods were collected through face-to-face interviews, with an 84 % response rate for both periods. These 24 h were collected using the USDA Automated Multiple-Pass Method (AMPM), which is a structured, computer-assisted, multiple-pass method developed by the US Department of Agriculture (USDA). The AMPM has five steps. Each step guides the respondent through the recall process, prompting them to provide more detailed information and reducing the likelihood of omissions(Reference Moshfegh, Rhodes and Baer28).
While the individual reported the data uninterruptedly, the IBGE agent made a list on paper. Another household member could help if one person cannot complete the 24 h. The IBGE developed software to register the time, occasion and place, and to enter all food and beverages consumed in the previous day, including recipes, ingredients, food preparation methods, portion sizes in household measures and items added to foods(23). The software issued an alert if the individual reported fewer than five foods or had not reported any foods for more than 3 h. In these cases, the IBGE agent had to confirm that all foods consumed during the day had been recorded and investigate any omissions or incomplete data. In the final stage, the information was checked for missing data and recording errors. All records with an energy intake of less than 300 kcal or more than 10 000 kcal were checked for consistency(23).
The choice of 24 h was based on the fact that it is almost universally used in population-based research and is considered the method with the lowest possibility of systematic error. Another reason was the results of the instrument validation study conducted in the 2008–2009 NDS, which showed better performance of 24 h than food records using double-labelled water as the gold standard method. Furthermore, IBGE pretested and validated the collection instruments, performed quality control procedures during data collection and deleted inconsistent records and replaced them with imputed values in order to minimise the biases inherent in the use of dietary surveys(23,Reference Rodrigues, Carli and Araújo29) .
Nutritional composition of foods
The food and beverage intake databases contained 1832 items(23). To convert these items into amounts consumed in grams or millilitres, the IBGE used the Table of Reference Measures for Food Consumed in Brazil, developed in the 2008–2009 NDS-HBS, reviewed and updated in the 2017–2018 NDS-HBS(30). The energy and nutrient content of each food item reported in 24 h was obtained from the Brazilian Food Composition Table, version 7.0(31), available at http://www.fcf.usp.br/tbca, in accordance with standards and guidelines for the generation, compilation and use of food composition data of FAO/INFOODS (Food and Agriculture Organization/International Network of Food Data Systems), at https://www.fao.org/infoods/infoods/tables-and-databases/faoinfoods-databases/en/).
Classification of ultra-processed foods
The Nova food classification system, developed by a team at the University of São Paulo, in Brazil, classifies all foods and food products into four groups according to the nature, extent and purpose of the industrial processing they undergo. It considers all physical, biological and chemical methods used during the food manufacturing process, including the use of additives(Reference Monteiro, Cannon and Levy9,Reference Monteiro, Cannon and Lawrence15) . Definitions and lists of examples for each of the four Nova groups can be found at https://www.fsp.usp.br/nupens/en/food-classification-nova/ and in the supplementary material (online Supplementary Table 1).
Sociodemographic characteristics of the study population: National Dietary Survey, Brazil, 2017–2018 (n 46 164)
As the focus of this study was on UPF, all foods and beverages reported were classified as ultra-processed or non-ultra-processed according to the NOVA food classification(Reference Monteiro, Cannon and Levy9,Reference Monteiro, Cannon and Lawrence15) . All UPF were further divided into subgroups (sweet biscuits, cakes, and pies; packaged salty snacks; bread; confectionaries; soft drinks and refreshments; milk-based drinks; pizzas, hamburgers, sandwiches, and savouries; frozen or instant ready-made dishes; reconstituted meat products; and others).
All food groupings were based not only on the Nova classification system, but also on the food composition, the list of ingredients, nutritional information, resolutions on norms and standards, food processing technology, and published studies(Reference Louzada, Ricardo and Steele22,Reference Godos, Giampieri and Al-Qahtani32,Reference Juul, Parek and Martinez-Steele33) . This process was carried out by two independent researchers. When discrepancies in classification occurred, they discussed these differences until reaching a consensus. In these cases, they resolved the issues by adopting the most conservative classification, that is, the lowest degree of processing. However, there were some exceptions to this approach, including items such as bread, ready-to-eat cereals and salty snacks.
Diet quality assessment
The energy densities of proteins, carbohydrates, free sugar, total and saturated fats (expressed as % of energy), fibre (g/1000 kcal), potassium (mg/1000 kcal) and Na (mg/1000 kcal) were calculated. Based on nutritional recommendations for healthy diets proposed by the WHO, the inadequate consumption of free sugars and saturated fats (≥ 10 % of total energy), total fats (≥ 30 % of total energy), fibres (≤ 12·5 g/1000 kcal), potassium (< 1755 mg/1000 kcal) and Na (≥ 1 g/1000 kcal) was also assessed(4,34,35) . Adjustments were made using the Multiple Source Method (MSM) developed by the European Prospective Investigation Cancer and Nutrition (EPIC)(Reference Haubrock, Nöthlings and Volatier36). This programme estimates the usual intake of food and nutrients, eliminating intrapersonal variance of consumption(Reference Barbosa, Sichieri and Junger37,Reference Harttig, Haubrock and Knüppel38) . The MSM was used to estimate usual nutrient intake, free sugar, dietary energy and usual UPF energy intake.
Data analysis
Sociodemographic characteristics of the study population were described using absolute (n) and relative (%) frequencies. The relationship between diet quality and UPF consumption by age group, including adolescents, adults and older adults, was assessed in three ways. First, mean and standard error (se), relative contribution (%) to the total energy and quintiles of energy contribution of UPF and their subgroups were described. Second, average dietary indicators were evaluated according to quintiles (%) of the total energy from UPF foods. Linear regressions were used to examine if there was a trend in average dietary indicators according to quintiles (%) of total energy from UPF, adjusted for sociodemographic characteristics, identified in the literature as possible confounders(Reference Jager, Zoccali and MacLeod39). This approach has been used in most studies examining the effect of UPF on diet quality(Reference Martini, Godos and Bonaccio10). Third, the prevalence of inadequate nutrient consumption was calculated according to quintiles of energy contribution of UPF. These associations were investigated using a multiple Poisson regression, with robust variance, adjusted for sociodemographic characteristics, identified in the literature as possible confounders(Reference Jager, Zoccali and MacLeod39), providing the prevalence ratio (PR) and corresponding 95 % CI. The Poisson regression was used due to the high prevalence of UFP consumption among the Brazilian population. In this scenario, the PR produces more robust and narrower CI estimates(Reference Barros and Hirakata40). The assumptions of the adjusted regression models were tested using the omnibus test (P value < 0·05) and P-value for deviance (P value > 0·05) to determine the significance and goodness of fit of the models.
All analyses were performed using Stata/SE statistical package version 16 (Stata Corp.), considering the expansion factors, the complexity of the sample design and a significance level of 5 %.
Results
Out of the 46 164 participants, 52·1 % were women, 64·6 % were adults, 41·7 % of the reference person attended school for at least 12 years, 29·7 % were located in the highest per capita income, 44·9 % were living in a household arrangement composed of a couple with children and 42·7 % were located in the southeast region of Brazil (Table 1).
Ultra-processed food consumption
The results revealed a significant variation in the energy contribution of UPF foods across age groups. Adolescents exhibited the highest consumption, with their intake ranging from 6·4 % (lowest quintile) to 59·2 % (highest quintile) of their total daily energy intake. Adults followed a similar pattern, albeit with a slightly lower range of 4·6 % to 50·5 %. Older adults demonstrated the lowest consumption, with a range of 3·7 % to 42·9 % of their energy intake derived from UPF. All age groups displayed an upward trend in UPF consumption across the quintiles (Table 2).
Average of absolute and relative consumption and contribution (%) of the total energy of ultra-processed foods and their subgroups according to age groups: National Dietary Survey, Brazil, 2017–2018 (n 46 164)
Q, Quintiles of energy proportion from total UPF intake in the total energy intake; UPF, ultra-processed foods.
a Bread and toast.
b Candies, lollipops, desserts, ice cream and chocolates.
c Pasta dishes, instant noodles, powdered mixtures for soups and mash.
d Sausages, nuggets, poultry and fish finger.
e Soya products, distilled alcoholic beverages, ready-made sauces, margarine, cheeses and artificial sweeteners.
*P value < 0·05.
**P value < 0·001 for linear trend adjusted for sex, age (years), income, education (years), household arrangement and macro-region.
† 1st Q: 0–9·7; 2nd Q: 9·8–17·3; 3rd Q: 17·4–27·7; 4th Q: 27·8–42·8; 5th Q: 42·9–100.
‡ 1st Q: 0–7·2; 2nd Q: 7·3–12·3; 3rd Q: 12·4–19·9; 4th Q: 20·0–32·9; 5th Q: 33·0–100.
§ 1st Q: 0–6·2; 2nd Q: 6·3–10·5; 3rd Q: 10·6–16·5; 4th Q: 16·6–26·7; 5th Q: 26·8–100.
Dietary quality and ultra-processed food
The analysis of the nutritional quality of the diets of adolescents, adults and older adults showed a significant decreasing tendency in the average levels of protein, fibre and potassium intake and an increasing tendency in the energy densities of carbohydrates, free sugars, total fats (except for the older adults), saturated fat and Na as the energy contribution of UPF increased. It is noteworthy that the diets of the three age groups had high Na content and low potassium content in all energy contribution strata of UPF (Table 3).
Average diet indicators according to quintiles (%) of the total energy of ultra-processed foods by age groups: National Dietary Survey, Brazil, 2017–2018 (n 46 164)
Q, Quintiles of energy proportion from total UPF intake in the total energy intake; UPF, ultra-processed foods.
a–h , WHO recommended values for proteins (10–15 % of total energy), carbohydrates (55–75 % of total energy), free sugar (< 10 % of total energy), total fats (15–30 % of total energy), saturated fat (< 10 % of total energy), fibre (> 12·5 g/1000 kcal), potassium (> 1·755 mg/1000 kcal) and Na (< 1 g/1000 kcal).
*P value < 0·001 for linear trend adjusted for sex, age (years), income, education (years), household arrangement and macro-region.
† 1st Q: 0–9·7; 2nd Q: 9·8–17·3; 3rd Q: 17·4–27·7; 4th Q: 27·8–42·8; 5th Q: 42·9–100.
‡ 1st Q: 0–7·2; 2nd Q: 7·3–12·3; 3rd Q: 12·4–19·9; 4th Q: 20·0–32·9; 5th Q: 33·0–100.
§ 1st Q: 0–6·2; 2nd Q: 6·3–10·5; 3rd Q: 10·6–16·5; 4th Q: 16·6–26·7; 5th Q: 26·8–100.
When comparing adolescents in the last quintile of energy contribution of UPF with those in the first quintile, higher PR of inadequate free sugar consumption (PRadj = 2·02, 95 % CI = 1·82, 2·25), total fats (PRadj = 1·21, 95 % CI = 1·08, 1·35) and fibre (PRadj = 1·88, 95 % CI = 1·68, 2·10) were observed. Adults in the last quintile of energy contribution of UPF compared with those in the first quintile had inadequate intake of free sugar, increased by 86 % (PRadj = 1·86, 95 % CI = 1·75, 1·98), of total fats, increased by 10 % (PRadj = 1·10, 95 % CI = 1·03, 1·17), and fibres, increased by 70 % (PRadj = 1·70, 95 % CI = 1·60, 1·80). On the other hand, older adults in the highest quintile of energy contribution of UPF compared with those in the first quintile had inadequate intake of free sugar, increased by 48 % (PRadj = 1·48, 95 % CI = 1·30, 1·69), and fibres, increased by 24 % (PRadj = 1·24, 95 % CI = 1·09, 1·40) (Table 4).
Inadequate nutrient intake according to quintiles of energy contribution of ultra-processed foods by age groups: National Dietary Survey, Brazil, 2017–2018 (n 46 164)
Q, quintiles of energy proportion from total UPF intake in the total energy intake; PR adjusted, prevalence ratios adjusted for sex, age (years), income, education (years), household arrangement and macro-region; UPF, ultra-processed foods.
All models were tested for goodness of fit using the omnibus test (P value < 0·05) and P value for deviance (P value > 0·05).
a P value < 0·05.
b P value < 0·001.
* The usual consumption estimated by the Multiple Source Method.
† 1st Q: 0–11·4; 2nd Q: 11·5–18·3; 3rd Q: 18·4–25·9; 4th Q: 26·0–36·0; 5th Q: 36·1–82·5.
‡ 1st Q: 0–7·6; 2nd Q: 7·7–13·3; 3rd Q: 13·4–19·3; 4th Q: 19·4–27·9; 5th Q: 28·0–81·7.
§ 1st Q: 0–6·6; 2nd Q: 6·7–11·4; 3rd Q: 11·5–16·7; 4th Q: 16·8–24·1; 5th Q: 24·2–80·2.
Discussion
The present study found that in the 2017–2018 NDS, UPF accounted for approximately 27 % of the total energy intake for Brazilian adolescents, 20 % for adults and 16 % for older adults. As consumption of UPF increased across all age groups, there was a significant decline in the average intake of essential nutrients like protein, fibre and potassium. Conversely, there was a concerning rise in energy density, carbohydrates (particularly free sugars), total fats (except for older adults), saturated fats and Na as the energy contribution of UPF increased.
Adolescence represents a critical developmental period marked by significant behavioural changes, including the adoption of potentially unhealthy habits such as excessive screen time, increased consumption of meals away from home and decreased family meal participation(Reference Rodríguez-Barniol, Pujol-Busquets and Bach-Faig41,Reference Rauber, Martins and Azeredo42) . These emerging patterns can establish long-term lifestyle trajectories with lasting health implications. Particularly concerning is the high intake of UPF in this population, which may adversely affect adolescent growth and development while potentially establishing detrimental dietary patterns that persist into adulthood(5). A previous study by Crisóstomo et al. (Reference Crisóstomo, Rosal and de Carvalho43) investigated UPF consumption within age groups in a Northeastern Brazilian capital city and found that adolescents had the highest intake (26·4 %), while the older adults consumed the least (17·5 %), with results similar to ours. It is worth noting that younger people are the priority target of marketing and advertising strategies for these foods, and therefore, compared with adults and older adults, they are more likely to try most UPF(Reference Ozcariz, Pudla and Martins44,Reference Sadeghirad, Duhaney and Motaghipisheh45) . Other factors such as convenience, lack of culinary skills and cost may be also important promoters of consumption of these foods among younger individuals(Reference Almeida, Scagliusi and Duran46). Recent evidence also suggests that UPF may induce addictive biological and behavioural responses, consequently driving excessive eating patterns and contributing to obesity, supporting the need for future research(Reference LaFata and Gearhardt47).
Our results are also consistent with a previous national study utilising HBS data, which identified a significant inverse relationship between UPF consumption and the dietary content of essential nutrients like protein, fibre, vitamins and minerals(Reference Louzada, Ricardo and Steele22). These findings are particularly relevant as a reduction in the dietary share of UPF holds the potential to significantly improve overall dietary quality.
Our findings indicate that the types of UPF consumed were similar across all three age groups. Pizza, hamburgers, sandwiches, savoury snacks, soya products, distilled alcoholic beverages, ready-made sauces, margarine, cheeses and artificial sweeteners were the primary sources of energy. Additionally, sweet biscuits, cakes and pies were significant contributors for adolescents and adults, whereas packaged salty snacks and bread held prominence among older adults. Several of these items are consistent with dietary patterns observed in other countries. For instance, breads, fast-food snacks and salted biscuits are noted in Canada(Reference Polsky, Moubarac and Garriguet48), while breads, biscuits, cakes and industrial desserts are prevalent in Portugal(Reference Miranda, Rauber and de Moraes49). A study conducted in Brazil showed that rice, coffee, beans, salt, bread and beef are the most commonly consumed foods, with noticeable variations among teenagers. This highlights the importance of further investigating age-related differences in diet quality(Reference Rodrigues, Souza and Bezerra50).
The stratification of age groups by quintiles of energy contribution of UPF revealed that the average consumption of these foods can be up to 60 %, 50 % and 43 % of the total energy consumed in the last quintile by adolescents, adults and older adults in the last quintile, respectively. These values were substantially lower than those found in studies conducted in high-income countries(Reference Machado, Steele and Levy51,Reference Rauber, Louzada and Steele52) . In Australia, for example, the average consumption of these foods accounted for 84 %, 72 % and 65 % of the total energy consumed by adolescents, adults and older adults, respectively(Reference Machado, Steele and Levy51). The affordability, convenience and aggressive marketing may be responsible for Australian diets based on UPF(Reference Marchese, Livingstone and Woods19). Whereas in Brazil, fresh and minimally processed foods, especially traditional items of national cuisine, such as rice, beans, roots and tubers, along with dishes derived from these staples still represent the basis of the Brazilian diet(23,Reference Rodrigues, Souza and Bezerra50) .
Regarding the analysis of diet quality, most nutritional indicators found in the first quintile of energy contribution of UPF of the diet of adolescents, adults and older adults followed the values recommended by the WHO, except for Na (high) and potassium (low). A possible explanation for these results is that Na intake in the country is quite high, mainly due to the high amounts of cooking salt and salt-based condiments used in culinary preparations(Reference Ricardo, Andrade and Salvador53). We emphasise that, although the nutritional profile of diets with lower energy contribution of UPF was healthier, there is no amount considered safe for the intake of these products, as they tend to have an unbalanced nutritional content(Reference Jaime, Campello and Monteiro54).
Our findings suggest that as the proportion of UPF in the diets of all age groups increases, the nutritional quality tends to decline, with many indicators deviating significantly from WHO recommendations. This trend is consistent with both national and international research(Reference Martini, Godos and Bonaccio10). A systematic review with meta-analysis of data from representative national surveys revealed a significant increase in free sugar, total fat and saturated fat content, alongside a decrease in proteins, fibres and several micronutrients including potassium, Mg, vitamin C, vitamin D, Zn, phosphorus, vitamin B12 and niacin, as UPF consumption increases(Reference Martini, Godos and Bonaccio10). These findings align with our own and support previous studies.
In contrast to the previous national survey based on NDS 2008–2009 data, our findings revealed a direct correlation between Na consumption and the increasing presence of UPF in the diets of adolescents, adults and older adults(Reference Louzada, Martins and Canella55). This discrepancy may arise from variations in the types of UPF subgroups consumed, as documented in different editions of the NDS. Specifically, our study noted a significant rise in the intake of items associated with higher Na density across quintiles of UPF energy contribution. Similar trends of increasing Na density with greater consumption of UPF were observed in studies from Chile, Canada and the USA(Reference Cediel, Reyes and Corvalán56–Reference Steele, Popkin and Swinburn58). Another possible factor is the variation in dietary data collection methods between the two surveys. While the HBS 2008–2009 employed a food record, the HBS 2017–2018 utilised the 24 h. This change was predominantly influenced by the widespread adoption of the 24 h method in population-based research due to its perceived lower susceptibility to systematic errors. Nonetheless, prior research demonstrated that harmonisation analysis strategies enable the assessment of food consumption trends across different survey methods(Reference Rodrigues, Carli and Araújo29).
This study also demonstrated that all age groups with high consumption of UPF, when compared with their counterparts, presented statistically significant increased prevalence in the excessive consumption of free sugar, total fats, saturated fat and Na and insufficient dietary fibre and potassium. Much evidence, including systematic reviews, shows that excessive consumption of free sugar contributes to diets with higher energy density and increases the risk of obesity, type 2 diabetes and dental caries(Reference Macdonald59,60) , while high consumption of fats and Na and low consumption of fibre are positively associated with unfavourable outcomes associated with cardiometabolic disorders and mortality(Reference Lane, Gamage and Du14,Reference Kim, Je and Giovannucci61–Reference Veronese, Solmi and Caruso63) .
In the case of Brazilian older adults, the consumption of these foods may represent a potential risk to their health conditions, given the high prevalence of chronic diseases among this population(64). In addition, there is further evidence indicating that consumption of UPF can cause damage to the quality of life and health of older people, which is associated with a decrease in telomere length, a biological marker of cell ageing(Reference Alonso-Pedrero, Ojeda-Rodríguez and Martínez-González65), with increased risk of developing frailty syndrome(Reference Sandoval-Insausti, Blanco-Rojo and Graciani66) and a higher likelihood of having abdominal obesity(Reference Sandoval-Insausti, Jiménez-Onsurbe and Donat-Vargas67), dyslipidemia(Reference Donat-Vargas, Sandoval-Insausti and Rey-García68) and decline in renal function(Reference Rey-García, Donat-Vargas and Sandoval-Insausti69).
Based on the above, several countries have implemented strategies to guide adequate and healthy food choices, such as the development of dietary guidelines, standards for front-of-package labelling of foods, and food and nutrition education activities; structuring of tax policies (taxing of sweetened beverages); and adoption of regulatory advertising measures. Food Guides in several countries have adopted the NOVA classification system to establish their recommendations on healthy eating and emphasise that UPF should be avoided(5,70–73) .
Brazil recently introduced new front-of-package nutritional labelling standards for UPF, highlighting high levels of critical nutrients like sugar, Na and unhealthy fats, enabling consumers to easily identify unhealthy products(74,Reference Martins and Jaime75) . Chile pioneered this approach, showing significant reductions in UPF purchases(Reference Taillie, Bercholz and Popkin76). Fiscal policies, such as Mexico’s tax on high-energy foods and sugary drinks and Barbados’ tax on sweetened beverages, have also led to decreased consumption(Reference Batis, Rivera and Popkin77,Reference Alvarado, Unwin and Sharp78) . In Brazil, voluntary agreements to reduce Na and sugar in foods have faced criticism for lack of transparency, weak goals and no penalties for non-compliance(79–Reference Scrinis and Monteiro81). Although these measures aim to reduce dietary risks, they may not guarantee healthier products and could inadvertently promote UPF consumption, undermining efforts to encourage minimally processed foods(5,Reference Scrinis and Monteiro81) . Additionally, regulating marketing and advertising, especially among children and adolescents, remains crucial, with challenges emerging in digital media regulation(82).
This study has several limitations and strengths. One primary constraint lies in the underreporting of consumed items, inherent to the 24-h method, potentially impacting dietary and nutritional estimates. However, the inclusion of the AMPM during data collection enhances standardisation and systematisation, mitigating interview bias and enhancing data accuracy. This method was also used during the data analysis stage to estimate usual food and nutrient intake, thereby eliminating the effect of intra-individual variation in day-of-week consumption. Misclassification errors could result in either underestimation or overestimation of UPF consumption. To minimise these errors, research teams cross-reviewed each other’s classifications and resolved discrepancies through discussion until consensus was reached. In cases of uncertainty, a conservative approach was adopted, favouring the lower degree of processing or assuming a homemade recipe, which may have led to an underestimation of UPF intake.
Finally, some of the survey data used in this study were collected several years ago. While these represent the most recent nationally representative data available, dietary patterns may have evolved since 2018. Analyses from a prior study, involving data from 2018 to 2022, found an increase in unprocessed or minimally processed foods and processed culinary ingredients prices (R$ 15·3/kg in 2018 to R$ 22·5/kg in 2025), and a decrease in UPF prices (R$ 22·1/kg in 2018 to R$ 20·2/kg in 2025), which may have encouraged the replacement of traditional meals for the consumption of unhealthy foods(Reference Andrade, Caldeira and Mais83).
Furthermore, the findings are only applicable to individuals aged 10 years and older and do not encompass younger Brazilian children. Nonetheless, utilising a recent database representing the Brazilian population’s food consumption is advantageous. Additionally, adjustments made to estimate usual food consumption reduce errors in estimating inadequate nutrient intake.
Moreover, this study provides valuable insights into UPF consumption at a national level and, to the best of our knowledge, offers the first assessment of age group differences in such consumption.
Conclusion
Our study corroborates previous findings indicating that UPF consumption may detrimentally affect diet quality, and we observed differences in the dietary quality of adolescents, adults and the elderly, even descriptively, highlighting the unique dynamics of each age group. Therefore, we advocate for nutrition education initiatives and limited access to UPF to promote diet quality and healthy dietary habits.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S000711452510411X
Acknowledgements
A.C.T.R was supported by a Doctoral Funding from Research Support Foundation of the State of Rio de Janeiro (Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro – FAPERJ n. 202·728/2017). The funder had no role in the study design, data collection, analysis, interpretation of data, preparation of the manuscript and decision to submit the paper.
A. C. T. R.: Conceptualisation, Data curation, Investigation, Methodology, Formal Analysis, Writing – Original Draft, Writing – Review and Editing. F. d. S. B. B.: Conceptualisation, Data curation, Investigation, Methodology, Formal analysis, Project administration, Supervision, Writing – Original Draft, Writing – Review and Editing. D. M. d. S.: Conceptualisation, Data curation, Investigation, Methodology, Project administration, Supervision, Writing – Original Draft, Writing – Review and Editing. E. S. d. C.: Writing – Original Draft, Writing – Review and Editing. C. C. C.: Writing – Review and Editing; A. R. A. A.: Data curation, Investigation, Methodology, Writing – Original Draft, Writing – Review and Editing.
The authors declare no conflict of interests.
Ethical approval for the study was provided by Coventry University (Project Reference Number: P134794).
Open access
