Hostname: page-component-78c5997874-xbtfd Total loading time: 0 Render date: 2024-11-13T14:55:37.519Z Has data issue: true hasContentIssue false

Adherence to the EAT-Lancet sustainable diet and ultra-processed food consumption: findings from a nationwide population-based study in Brazil

Published online by Cambridge University Press:  04 October 2024

Leandro Teixeira Cacau*
Affiliation:
Department of Nutrition, School of Public Health, University of São Paulo, São Paulo, Brazil Center for Epidemiological Research in Nutrition and Health, University of São Paulo, São Paulo, Brazil
Thays Nascimento Souza
Affiliation:
Department of Nutrition, School of Public Health, University of São Paulo, São Paulo, Brazil Center for Epidemiological Research in Nutrition and Health, University of São Paulo, São Paulo, Brazil
Maria Laura da Costa Louzada
Affiliation:
Department of Nutrition, School of Public Health, University of São Paulo, São Paulo, Brazil Center for Epidemiological Research in Nutrition and Health, University of São Paulo, São Paulo, Brazil
Dirce Maria Lobo Marchioni
Affiliation:
Department of Nutrition, School of Public Health, University of São Paulo, São Paulo, Brazil
*
*Corresponding author: Email lcacau@usp.br
Rights & Permissions [Opens in a new window]

Abstract

Objective:

To evaluate the association between ultra-processed food consumption and adherence to the EAT-Lancet diet in a representative sample of the Brazilian population.

Design:

The study used data from the Brazilian National Dietary Survey 2017–2018 and employed linear regression models to evaluate the association between ultra-processed food consumption and adherence to the EAT-Lancet diet, as measured by the Nova food system and Planetary Health Diet Index (PHDI), respectively.

Setting:

Nationally representative sample of the Brazilian population.

Participants:

The study included 46 164 Brazilians ≥ 10 years old.

Results:

The average PHDI total score was 45·9 points (95 % CI 45·6, 46·1). The ultra-processed food consumption was, with dose-response, inversely associated with the adherence to the EAT-Lancet diet. The PHDI total score was 5·38 points lower (95 % CI –6·01, –4·75) in individuals in the highest quintile of consumption of ultra-processed foods, as compared to those in the first quintile. The PHDI score was also inversely associated with the share of processed culinary ingredients and processed foods and positively associated with the share of unprocessed or minimally processed foods.

Conclusions:

Our study showed an inverse relationship between the consumption of ultra-processed foods and the adherence to a healthy and sustainable diet.

Type
Research Paper
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2024. Published by Cambridge University Press on behalf of The Nutrition Society

Diets are intrinsically linked to human and planetary health(Reference Tilman and Clark1). Unbalanced and nutritionally poor diets, usually characterised by low amounts of fruits, vegetables, nuts and whole grain cereals and high in red and processed meat, are responsible for a major part of the global burden of diseases(Reference Afshin, Sur and Fay2), which is explained due to the food transition from traditional diets to globalised diets. Alongside the diet-disease scenario, the global food system is nowadays one of the primary drivers of adverse environmental impacts. Agriculture is responsible for about 30 % of all greenhouse gas emission (GHGE), uses 70 % of the freshwater resource and occupies about 40 % of the Earth’s surface, leading to an adverse environment and crossing the planetary boundaries(Reference Springmann, Wiebe and Mason-D’Croz3,Reference Springmann, Spajic and Clark4) .

The Food and Agriculture Organization of the United Nations (FAO) defines sustainable healthy diets as those that enhance all aspects of individual health, exert minimal environmental pressure and are accessible, affordable, safe, equitable and culturally acceptable on a global scale(5). Adherence to sustainable dietary patterns has been related to reduce diet-related mortality rates and environmental impacts, such as GHGE, water and land use(Reference Springmann, Wiebe and Mason-D’Croz3,Reference Springmann, Spajic and Clark4,Reference Laine, Huybrechts and Gunter6,Reference Wang, Li and Afshin7) . Achieving a worldwide shift from unhealthy, umbalanced diets to healthy, sustainable ones is one of the most significant challenges of this century(Reference Tilman and Clark1).

To guide countries, researchers and policy makers, the EAT-Lancet Commission proposed a global sustainable reference diet that promotes human and planetary health. This model diet, called ‘planetary health diet’, primarily includes vegetables, fruits, whole grains, legumes, nuts and unsaturated oils, while suggesting only low–to-moderate amounts of seafood and poultry, and minimal to no red meat, animal fats, added sugar, refined grains and starchy vegetables(Reference Willett, Rockström and Loken8). On a global adoption scenario, this reference diet was related to reduce diet-related deaths by 11·6 million annually (i.e. 23·6 % deaths per year) and decrease GHGE by 42 %, freshwater use by 10 % and nitrogen application by 15 %(Reference Wang, Li and Afshin7,Reference Willett, Rockström and Loken8) .

Although the EAT-Lancet report has been a landmark in the discussion of sustainable food systems for proposing a healthy and sustainable model diet, it has received some criticism for overlooking the negative effects of ultra-processed foods on human and planetary health(Reference Seferidi, Scrinis and Huybrechts9,Reference Lawrence, Baker and Pulker10) . Ultra-processed foods, as defined by the Nova food classification, are industrial formulations typically ready-to-eat, often derived from high-yield crops, such as sugars and syrups, refined starches, oils and fats, as well as protein isolates and low-commercial-value animal tissues. These formulations are designed to be visually appealing and have very intense flavours, achieved through combinations of flavourings, colourings, emulsifiers, sweeteners, thickeners and other cosmetic additives(Reference Monteiro, Cannon and Levy11). Several systematic reviews and meta-analyses showed the association between ultra-processed foods and poor-diet quality(Reference da Louzada, Ricardo and Steele12Reference Marchese, Livingstone and Woods14), higher risk of non-communicable diseases and all-cause mortality(Reference Pagliai, Dinu and Madarena15Reference Taneri, Wehrli and Roa-Díaz19), biodiversity loss(Reference Leite, Khandpur and Andrade20), as well as the temporal relationship between increasing ultra-processed food consumption and increasing GHGE, water and land use(Reference Silva, Garzillo and Rauber21).

Despite the evidence that ultra-processed foods have negative impacts on human and planetary health, the EAT-Lancet Commission report overlooks this effect and does not use or cite food processing as something to be avoided. Therefore, we wonder if these two diet quality metrics (i.e. the share of ultra-processed foods and the EAT-Lancet diet), increasingly recognised in the literature, are associated, even though they originate from different theoretical assumptions. Currently, there is no evidence on this field. Hence, the aim of this study was to address this question in a nationally representative sample of the Brazilian population.

Methods

Study design

The 2017–2018 National Dietary Survey (NDS) was used for this analysis. The NDS was integrated into the Household Budget Survey (HBS), a nationally representative survey conducted by the Brazilian Institute of Geography and Statistics to measure household consumption, expenditures and income, thus providing a comprehensive profile of the living conditions of the Brazilian population(22,23) .

Briefly, the HBS employed a two-stage cluster sampling methods, randomly selecting census sector and households with socioeconomic and geographical stratifications of the primary sampling units based on the 2000 Demographic Census. Data collection occurred from July 2017 to July 2018. Overall, the 2017–2018 HBS included 57 920 households, of which a subsample of 20 112 participated in the NDS, totalling 46 164 individuals aged ≥10 years with data on individual food consumption. Further details on sampling and data collection are available in previous publications by the Brazilian Institute of Geography and Statistics(22,23) .

Individual food consumption assessment

Dietary data were gathered using 24-hour dietary recalls conducted on two non-consecutive days at the respondent’s home by trained research agents. These interviews followed a structured approach based on the Automated Multiple-Pass Method, utilising specialised software on a tablet(23).

Initially, participants were asked to list all foods and beverages consumed the previous day, creating a ‘quick list’. Following this, detailed information was recorded about the foods, beverages and recipes, including portion sizes in household measures, preparation and cooking techniques, added items, eating occasions and meal locations. The Table of Reference Measures for Food Consumed in Brazil was then used to convert the food and beverage quantities into grams and millilitres(24).

To determine the energy and nutritional composition, the Brazilian Food Composition Table (TBCA-USP) version 7·0 was employed (https://tbca.net.br/indexen.html). The TBCA is a web-based resource comprising two databases: (1) the Nutrient Intake Evaluation Database, which offers the complete nutrient profile of Brazil’s most-consumed foods, and (2) the Biodiversity and Regional Foods Database, which centralises analytical data on the biodiversity of Brazilian foods and typical regional dishes. Information for the TBCA was obtained through direct analytical analyses of foods and by compiling data from previous publications, analytical reports from the food industry and other food composition tables. Detailed information regarding the TBCA can be found in other sources(Reference Coelho, Giuntini and Grande25,Reference Giuntini, Coelho and Grande26) . For this study, only the first 24-hour dietary recalls was used.

Adherence to EAT-Lancet dietary recommendations

The Planetary Health Diet Index (PHDI) was used to evaluate adherence to the EAT-Lancet diet, which is a 16-component dietary index that considers all food groups proposed in the EAT-Lancet diet with a continuous score for each component. The PHDI total score was previously validated and performed satisfactorily in terms of validity and reliability criteria and was associated with lower GHGE and higher overall dietary quality(Reference Cacau, De Carli and de Carvalho27).

The PHDI total score was calculated through the procedure described by Cacau et al. (Reference Cacau, De Carli and de Carvalho27) and Marchioni et al. (Reference Marchioni, Cacau and De Carli28). Briefly, we identified foods classified as unprocessed or minimally processed (such as plain cooked fruits and vegetables) and culinary preparations composed of multiple ingredients (like lasagne, stroganoff and cakes), which required disaggregating into ingredient components to classify them into PHDI categories. Culinary preparations were disaggregated based on standard homemade recipes found in national literature. For ultra-processed foods with specific main ingredients (e.g. savoury chips made from maize starch), we estimated energy distribution using added sugars and total fat levels from the nutrient database. For example, in savoury chips, the percentage of energy from total fat was assumed to represent the contribution of the vegetable oil fraction in the food. After subtracting the total fat from the savoury chips, the contribution of the refined grain group (such as corn starch) was assumed to represent the remaining energy value of the food. Processed meats, like sausage, ham and salami, were not disaggregated into individual ingredients but rather categorised based on their primary ingredient origin or commonly marketed formulation within the respective groups of red meats or chicken and substitutes. More information can be found in the original paper elsewhere(Reference Cacau, De Carli and de Carvalho27). After the disaggregation process, the ingredients were allocated in their respective PHDI component. Examples of foods and ingredients included in the PHDI components are described in Supplementary Table S1.

The 16 PHDI components are divided into four categories: (1) adequacy components (nuts and peanuts, fruits, vegetables, vegetables and whole grains), (2) optimal components (eggs, dairy, fish and seafood, tubers and potatoes, and vegetable oils), (3) ratio components (dark green vegetables/total vegetables and orange-red vegetables/total vegetables) and (4) moderation components (red meat, poultry and substitutes, animal fats and added sugars). Table 1 presents the PHDI components with their respective cut-off points and scoring criteria.

Table 1 The Planetary Health Diet Index (PHDI) components and criteria for scoring system (recommended and lower or upper limit recommended values*)

EI: energy intake.

* All values expressed as caloric ratios between each PHDI component and total daily energy intake.

Legumes: including soy.

Dairy: excluding dairy fats.

§ Vegetable oils: including palm oil.

|| Dark green vegetables (DGV)/total vegetables: ratio between the energy intake of dark green vegetables (numerator) and the total vegetable energy intake (denominator) multiplied by 100.

Red and orange vegetables (ReV)/total vegetables: ratio between the energy intake of red and orange vegetables (numerator) and the total vegetable energy intake (denominator) multiplied by 100.

** Red meat: including beef, lamb and pork.

*† Animal fats: lard, tallow and dairy fats.

For the adequacy components, the maximum score (i.e. 10 points) was assigned if consumption met or exceeded the recommendation (i.e. the criteria for maximum score). Otherwise, the score was proportional, being calculated as follows:

$${\rm{\;}}{{\rm{current\;intake}}\over{\rm{recommende\;value}}} {\times {\rm{maximum\;score}}}$$

For example, subjects who consumed 3 % of energy from fruits from the total daily energy intake received 6 points = [(3 ÷ 5) × 10 = 6].

For the optimum components, the maximum score (i.e. 10 points) was assigned when reaching the recommended value. As consumption exceeds the recommended values, the score becomes inversely proportional to consumption, being calculated as follows:

$$\scriptsize{\left( {{\rm{\;}}1{\rm{\;}} - \left| {{\rm{\;}}1 - {{{\rm{current\;intake}}}}\over{{{\rm{lower\;or\;upper\;limit\;of\;recommended\;value}}}}{\rm{\;\;}}} \right|} \right) \times {\rm{maximum\;score}}}$$

For example, subjects who consumed 0·4 % of energy from eggs of the total daily energy intake received 5 points [(1 – | 1 – 0·4 ÷ 0·8 |) × 10 = 5 points. For ratio components, the same logic as for optimum components was assigned.

Finally, the moderation components received a minimum score (i.e. 0 points) when consumption reached or exceeded the upper limit of recommended value (i.e. the criteria for minimum score). When consumption was lower than the consumption limit, the calculation was as follows:

$$\small{{\left( 1 -{ {{\rm{current\;intake}}}\over{{{\rm{lower\;or\;upper\;limit\;of\;recommend\;value}}}}} \right){\rm{\;}} \times {\rm{\;maximum\;score}}}}$$

For example, subjects who consumed 1·5 % of energy from red meat of the total daily energy intake received 3·75 points [(1 – 1·5 ÷ 2·4) × 10 = 3·75].

Further detail on the PHDI development, scoring criteria, cut-off points and validity and reliability can be found elsewhere(Reference Cacau, De Carli and de Carvalho27,Reference Cacau and Marchioni29) .

Ultra-processed food classification – Nova food system

Foods and beverages reported in the NDS 2017–2018 were also classified according to the Nova food system(Reference Monteiro, Cannon and Levy11).

Briefly, this classification was conducted by researchers who were properly trained and familiar with the Nova food system. The process involved listing all reported foods and beverages to identify single-ingredient and multi-ingredient items. Single-ingredient items consist of unprocessed or minimally processed foods (e.g. raw apple, milk and unsalted nuts) or processed culinary ingredients (e.g. salt and olive oil) and were directly included in the food list. Multi-ingredient items (e.g. pies, cakes, pasta dishes and sauces) were first categorised as culinary preparations or industrially manufactured products. Items identified as industrially manufactured (e.g. ready-made pie and frozen lasagne) were included in the food list as such. Those identified as culinary preparations (e.g. homemade pie and rice with broccoli) were disaggregated into their constituent ingredients using standard food recipes, with each ingredient listed separately. Subsequently, each item in the list was classified into one of the four Nova groups(23,Reference Martinez-Steele, Khandpur and Batis30) .

The Nova food system includes four groups, namely (1) unprocessed or minimally processed foods, which are those consumed as obtained from the nature or that underwent industrial processes such as drying, boiling, freezing or others that do not add substances such as salt, sugar and/or oils or fats to the original food (e.g. fruits, vegetables, eggs and fresh meat); (2) processed culinary ingredients, which are substances obtained directly from nature or group 1 foods, which are used in the preparation, seasoning and cooking of foods (e.g. oils and fats, table sugar and salt); (3) processed foods, which are industrial products made by combining substances from group 2 with group 1 foods (e.g. cheese, fresh breads, vegetables in brine and salt-cured meat or fish); and (4) ultra-processed foods, defined as formulations of ingredients, mostly of exclusive industrial use, that result from several industrial processes and frequently are added by colours, flavours, emulsifiers and other cosmetic or sensory intensifying additives to make the final product palatable or more appealing (e.g. confectionary, salty snacks, sugary breakfast-cereal, packaged bread, carbonated soft drinks, flavoured yogurts, cookies and sausages). Thereafter, the share of contribution of each group to the total energy intake was estimated, i.e. the % energy intake of each Nova food group.

Data analyses

The mean and the corresponding 95 % CI for PHDI total score and the % energy intake from each Nova food group were presented for the entire Brazilian population. Additionally, these data were stratified by sex (male, female), age group (<19, 19–30, 31–45, 46–59, or ≥60), per capita income (quartiles), self-reported race (white, black, brown, and Asian or Indigenous) and area of residence (urban or rural).

The association between the % energy intake of each Nova food group and the PHDI scores was first assessed using linear splines graphics. Subsequently, the population was stratified into quintiles of the % energy intake of ultra-processed foods, with the lowest consumers belonging to the first quintile and the highest consumers to the fifth.

Linear regression models were fitted to evaluate if the higher % energy intake of ultra-processed foods is associated with lower PHDI total score. In these models, the PHDI total score was the outcome variable, and quintiles of share of ultra-processed food group to total diet were the explanatory variable. The model was adjusted for sex (male and female), age group (<19, 19–30, 31–45, 46–59 or ≥60), per capita income (quartiles), self-reported race (white, black, brown, yellow/Asiatic, or indigenous) and residence area (urban or rural). Analyses of association were repeated considering the quintiles of the % energy intake of the non-ultra-processed food groups (i.e. unprocessed or minimally processed foods; culinary processed ingredients; and processed foods) with the same adjustments of covariables.

As a sensitivity analysis, linear regression models were fitted using the Nova food groups expressed as energy intake (kcal), rather than % energy intake. The models were adjusted for the same covariates as in the main analyses.

Data analysis was carried out on Stata® (Statistical Software for Professionals, College Station, Texas, USA), version 14·2 and using the survey module to consider the effect of complex sampling procedures adopted in the NDS 2017–2018 and in order to allow extrapolation of results for the Brazilian population. Statistically significant differences were considered when the 95 % CI did not overlap.

Results

In the entire Brazilian population, the total PHDI had an average of 45·9 points (95 % CI 45·6, 46·1) and had a range of 1·65–95·0 points (i.e. minimum and maximum values). Higher PHDI scores were observed in women, in those with 31 years or older (31–45, 46–56 and ≥60 years), in individuals with higher per capita income and in those living in the urban area (Table 2).

Table 2 Distribution of the Planetary Health Diet Index (PHDI) and the caloric share (%) of Nova food groups for the Brazilian population and across sociodemographic characteristics (n 46 164). National Dietary Survey, 2017–2018

Values are means with their respective 95 % CI.

For the entire Brazilian population, around 20 % of the daily energy intake came from ultra-processed foods (19·7; 95 % CI 19·3, 20·1), while more than half came from unprocessed or minimally processed foods (53·4; 95 % CI 53·0, 53·8). Processed culinary ingredients corresponded to around 16 % (15·6; 95 % CI 15·4, 15·8) and processed foods to nearly 11 % (11·3; 95 % CI 11·1, 11·5). The % energy intake of the subgroups of the four Nova food groups is described Supplementary Table S2.

Ultra-processed food consumption was slightly higher among females and adolescents (those with <19 years), but tended to decrease progressively with age, and was lower in those with ≥60 years. Furthermore, ultra-processed food consumption was higher in individuals with higher per capita income and in those living in urban areas. On the other hand, unprocessed or minimally processed food consumption was slightly higher among males and in individuals with ≥60 years. Regarding self-reported race/skin colour, individuals who self-reported as black and brown had higher unprocessed or minimally processed foods intake (Table 2).

Linear splines showed an inverse association between the % energy intake of ultra-processed foods and PHDI total score, as reported on the linear coefficients and their respective 95 % CI (–0·13; –0·14, –0·12) in Fig. 1(a). Associations in the same direction were observed for the processed culinary ingredients (Fig. 1(c)) and processed foods (Fig. 1(d)), while the % energy intake of unprocessed or minimally processed foods was directly associated with the PHDI total score (Fig. 1(b)).

Fig. 1 Crude linear splines of the association between the Planetary Health Diet Index (PHDI) score and the caloric share (%) of each Nova food group continuously (n 46 164). National Dietary Survey, 2017–2018. β coefficients and their respective 95 % CI. Legend: (a): linear spline of the association between the PHDI score and the caloric share (%) of ultra-processed foods. (b): Linear spline of the association between the PHDI score and the caloric share (%) of unprocessed or minimally processed foods. (c): Linear spline of the association between the PHDI score and the caloric share (%) of processed culinary ingredients. (d): Linear splines of the association between the PHDI score and the caloric share (%) of processed foods

Crude and adjusted linear regression models showed a dose-response association between quintiles of ultra-processed food consumption and the PHDI total score (Figs. 2 and 3, respectively). The mean PHDI total score was 5·38 points lower (95 % CI –6·01, –4·75) in individuals in the highest quintile of consumption of ultra-processed foods, as compared to those in the first quintile (Fig. 3). Similarly, the mean PHDI total score was 1·31 points lower (95 % CI –1·98, –0·65) in individuals in the highest quintile of consumption of processed foods and 1·32 points lower (95 % CI –1·88, –0·76) in individuals in the highest quintile of consumption of culinary ingredients, as compared to those in the first quintile.

Fig. 2 Crude linear coefficients and 95 % CI of the association between Planetary Health Diet Index (PHDI) scores and quintiles of the caloric share (%) of ultra-processed foods (n 46 164). National Dietary Survey, 2017–2018. Legend: a 1st quintile: mean (min – max): 0·43 (0 – 3·40); 2nd quintile 6·83 (3·40 – 9·92); 3rd quintile 13·3 (9·92 – 17·3); 4th quintile 22·8 (17·3 – 29·6); 5th quintile 45·1 (29·6 – 100). Values are expressed as linear coefficients and 95 % CI

Fig. 3 Adjusted linear coefficients and 95 % CI of association between Planetary Health Diet Index (PHDI) scores and quintiles of the caloric share (%) of ultra-processed foods (n 46 164). National Dietary Survey, 2017–2018. Legend: a 1st quintile: mean (min – max): 0·43 (0 – 3·40); 2nd quintile 6·83 (3·40 – 9·92); 3rd quintile 13·3 (9·92 – 17·3); 4th quintile 22·8 (17·3 – 29·6); 5th quintile 45·1 (29·6 – 100). Model adjusted by sex, age, per capita income, self-reported race and residence area. Values are expressed as linear coefficients and 95 % CI

On the other hand, the mean PHDI total score was 8·34 points higher (95 % CI 7·76, 8·93) in individuals in the highest quintile of consumption of unprocessed or minimally processed foods, as compared to those in the first quintile. (Table 3).

Table 3 Linear regression coefficients and their 95 % CI for the association between the caloric share (%) of non-ultra-processed food groups (i.e. unprocessed or minimally processed foods, processed culinary ingredients and processed foods) and the Planetary Health Diet Index (PHDI) total score (n 46 164). National Dietary Survey 2017–2018

* 1st quintile: mean (min – max): 29.1 (0 – 39.7); 2nd quintile 46.0 (39.7 – 51.5); 3rd quintile 56.4 (51.5 – 61.3); 4th quintile 66.6 (61.3 – 72.5); 5th quintile 81.7 (72.5 – 100).

1st quintile: 4.12 (0 – 6.95); 2nd quintile 9.36 (6.96 – 11.7); 3rd quintile 14.0 (11.7 – 16.4); 4th quintile 19.5 (16.4 – 23.0); 5th quintile 30.8 (23.0 – 100).

1st quintile: 0 (0 – 0); 2nd quintile 0.10 (0.02 – 0.16); 3rd quintile 6.29 (0.16 – 11.5); 4th quintile 15.9 (11.5 – 21.0); 5th quintile 31.8 (21.0 – 100).

Model 1: unadjusted.

Model 2: adjusted by sex, age, per capita income, self-reported race and residence area. Values are expressed as linear coefficients and 95 % CI.

In the sensitivity analyses, the results from regression models employing Nova food groups in kcal showed consistent trends and directions similar to those obtained using % energy (see online supplementary material, Supplementary Table S3).

Discussion

In this analysis of a nationally representative sample of the Brazilian population, we found that the consumption of ultra-processed foods was inversely associated with the adherence to a healthy and sustainable diet – proposed by the EAT-Lancet Commission. Similar associations were observed for the share of processed culinary ingredients and processed foods, while the consumption of unprocessed and minimally processed foods was directly associated with the PHDI scores.

Our results suggest that even though ultra-processed foods are not direct targets of the EAT-Lancet Commission recommendations, individuals with high adherence to its proposed diet also eat less of ultra-processed foods. This happens mainly because the PHDI adequacy components (fruits, vegetables, legumes, nuts and peanuts, and whole cereals) are usually eaten in the form of unprocessed or minimally processed foods (i.e. fresh or dry, unsugared and unsalted), especially in Brazil, where these foods comprise over 50 % of the typical diet(23), leaving limited space for incorporating ultra-processed foods. In a recent meta-analysis with population-based studies from 13 countries (Australia, Brazil, Canada, Chile, Colombia, France, Italy, Korea, Mexico, Portugal, Taiwan, UK and US), Martini et al. (Reference Martini, Godos and Bonaccio31) observed inverse linear relationships between the consumption of ultra-processed and unprocessed or minimally processed foods, suggesting that they may constitute opposite dietary patterns.

The group of ultra-processed foods has a poor nutrient profile characterised by lower amounts of protein, fibre and micronutrients, and higher contents of added sugar, total, saturated and trans-fatty acids(Reference Martini, Godos and Bonaccio31Reference da Louzada, Martins and Canella33) (as compared to non-ultra-processed foods). Moreover, its consumption has been linked to diminished overall dietary quality, according to findings from nationally representative cross-sectional studies conducted in Brazil(Reference da Louzada, Ricardo and Steele12), United States(Reference Martínez Steele, Popkin and Swinburn13) and Australia(Reference Marchese, Livingstone and Woods14). Conversely, previous results showed that higher scores in the PHDI (i.e. higher adherence to EAT-Lancet diet) were positively associated with higher intakes of vegetable proteins, polyunsaturated fats, fibres and micronutrients commonly present in fruits and vegetables, while negatively associated with animal protein, total fat, saturated fat, cholesterol and monounsaturated fat(Reference Cacau, De Carli and de Carvalho27,Reference Cacau, Hanley-Cook and Huybrechts34) . Alongside, the PHDI – that is, a higher adherence to EAT-Lancet sustainable diet – was also related with higher overall dietary quality(Reference Cacau, De Carli and de Carvalho27) as evaluated throughout the Revised-Brazilian Healthy Eating Index, a diet quality index adapted to the Brazilian context(Reference Previdelli, de Andrade and Pires35).

The inverse association between adherence to the EAT-Lancet diet and ultra-processed foods can also be discussed from the perspective of environmental impacts. While greater adherence to the EAT-Lancet diet was associated with lower diet-related environmental impacts (e.g. carbon, water and land footprint) in modelling studies(Reference Springmann, Wiebe and Mason-D’Croz3,Reference Wang, Li and Afshin7) , the consumption of ultra-processed foods has been also related to environmental impacts(Reference Silva, Garzillo and Rauber21,Reference Garzillo, Poli and Leite36,Reference Kesse-Guyot, Allès and Brunin37) , although still less explored. Garzillo et al. (Reference Garzillo, Poli and Leite36) using data from the 2007–2008 NDS in Brazil observed that ultra-processed foods were associated with higher carbon and water footprints, but this association remains only for water footprint, after controlling for age, sex, education, income and country regions. Silva et al. (Reference Silva, Garzillo and Rauber21) evaluated the temporal trends of GHGE, water and ecological footprint of food purchases according to Nova food system between 1987 and 2018 using the Brazilian HBS and found that GHGE increased by 21 %, water footprint by 22 % and ecological footprint by 17 %, followed by the decrease in the share of unprocessed or minimally processed foods and an increase in the food purchase of processed and ultra-processed foods. In a national cross-sectional study with a representative French sample, Kesse-Guyot et al. (Reference Kesse-Guyot, Allès and Brunin37) observed that those with the highest dietary share of ultra-processed foods also presented the highest diet-related environmental impacts indicators (e.g. GHGE, land use, use of fossil resources and ecological footprint); however, the association was fully mediated by the higher total energy intake associated with the consumption of ultra-processed foods.

Some studies have investigated the relationship of the EAT-Lancet diet with health outcomes, and the higher adherence to its recommendations was associated with lower cardiometabolic risk and lower odds for overweight and obesity in Brazilians(Reference Cacau, Benseñor and Goulart38,Reference Cacau, Benseñor and Goulart39) , lower risk of total mortality in Swedish population(Reference Stubbendorff, Sonestedt and Ramne40) and lower risk of diabetes in UK citizens(Reference Xu, Cao and Yang41) and in Mexican women(Reference López, Batis and González42). In parallel to the benefits of the EAT-Lancet diet for human health, several other studies have addressed the impact of ultra-processed foods on the risk of noncommunicable diseases worldwide. Systematic reviews and meta-analysis of observational studies reported that ultra-processed foods were associated with increased risk of type 2 diabetes(Reference Delpino, Figueiredo and Bielemann43), hypertension(Reference Wang, Du and Huang44), overweight/obesity(Reference Askari, Heshmati and Shahinfar45) and all-cause mortality(Reference Taneri, Wehrli and Roa-Díaz19). Nilson et al.,(Reference Nilson, Ferrari and da Louzada46) in a modelling study, reported that the consumption of ultra-processed foods was associated with 22 % of premature deaths from cardiovascular diseases and 34 % of mortality attributed to all causes. Although the PHDI and the % energy intake from ultra-processed foods differ in their focus of interest (i.e. with PHDI evaluating adherence to a healthy and sustainable diet, while the Nova food system focus on classifying foods based on the degree of food processing and their impact on health), both systems have yielded significant results regarding diet-disease and diet-environmental impact associations.

Inverse associations between the adherence to EAT-Lancet diet and processed culinary ingredients and processed foods were also observed. According to Nova food system classification, these groups are, respectively, composed of oils, butter, salt, and sugar and industrialised products composed by adding a processed culinary ingredient to unprocessed or minimally processed foods. The EAT-Lancet recommends that food with high contents of sugar, sodium and fats should be consumed in moderation, due to the impact on human and planetary health(Reference Willett, Rockström and Loken8). According to the Dietary Guidelines for the Brazilian Population(47), culinary ingredients and processed foods can be consumed moderately and be part of a healthy diet, especially when used in culinary preparations or recipes.

Our findings showed that ultra-processed foods comprise about 20 % of the calories of the Brazilian diet in 2017–2018, while unprocessed or minimally processed foods correspond to almost half of it. Despite that the Brazilian population is far from meeting the EAT-Lancet sustainable guidelines. Marchioni et al. (Reference Marchioni, Cacau and De Carli28) evaluated the adherence to EAT-Lancet diet throughout the PHDI in the 2017–2018 NDS and observed that the Brazilian population reached only 30 % of the possible points in the PHDI, indicating the poor adherence to the EAT-Lancet recommendations. Similar result was observed in a Brazilian cohort study, where adults reached only 40 % of the possible points in the PHDI(Reference Cacau, De Carli and de Carvalho27), even consuming more than 60 % of total calories from unprocessed or minimally processed foods and culinary ingredients(Reference Canhada, Luft and Giatti48). A possible explanation to this may be the fact that EAT-Lancet strongly recommends moderation in the consumption of animal sources, even if they are unprocessed or minimally processed foods. As an example, the EAT-Lancet recommendation for red meat is 14 g/d, ranging from 0 to 28 g, and it is well known that Brazilians have a much higher consumption of red meat(Reference de Carvalho, de Selem and Miranda49). The 2017–2018 Brazilian HBS shows that the availability of both ultra-processed foods and red meat increased between 2002–2003 and 2017–2018, while the opposite was observed for plant-based unprocessed or minimally processed foods(Reference Levy, Andrade and da Cruz50).

We are familiar with the EAT-Lancet criticisms related to its dietary recommendations and the lack of discussion on ultra-processed foods(Reference Seferidi, Scrinis and Huybrechts9,Reference Lawrence, Baker and Pulker10) . Furthermore, although the report reinforces cultural adaptations to local realities can be proposed(Reference Willett, Rockström and Loken8), a global diet model may not be the best way to achieve sustainable food systems. According to Semba et al. (Reference Semba, de Pee and Kim51), global adherence to the EAT-Lancet recommendations would have a 23 % reduction in GHGE; however, in low- and middle-income countries there would be an increase of up to 283 %. Accordingly, the adoption of the EAT-Lancet diet on a global scale would reduce about 12 % of the water footprint, but would increase the water footprint in 54 low- and middle-income countries(Reference Tuninetti, Ridolfi and Laio52). In addition, Hirvonen et al. (Reference Hirvonen, Bai and Headey53) analysis showed that the EAT-Lancet diet is not affordable for more than 40 % of the world’s population living in low- and middle-income countries. However, we recognise the role that such recommendations have played in recent years, strengthening and drawing attention to the debate on food systems and sustainable diets for human health within planetary boundaries(Reference Lawrence, Baker and Pulker10). In addition to proposing a global reference sustainable diet, the EAT-Lancet Commission also calls for more sustainable food production, by incorporating sustainable agriculture techniques and agrobiodiversity, such as soil regeneration, carbon sequestration, drip irrigation and soil water harvesting. Additionally, it advocates for reducing the consumption of animal foods, particularly beef, due to the burden it represents in terms of GHGE, as well as deforestation, and suggests shifting consumption towards plant-based foods with lower environmental impact, which can contribute to soil regeneration actions(Reference Willett, Rockström and Loken8).

Finally, our study has several strengths. We used the most recent nationally representative data on dietary intake, assessed throughout the 24-h dietary recall method, with the aid of a software developed exclusively for the survey, with data collected over a one-year period, which enabled capturing seasonal variations in dietary intake. Moreover, we used a previous validated diet index to assess the adherence to the EAT-Lancet diet(Reference Cacau, De Carli and de Carvalho27) and the most used system to classify ultra-processed foods(Reference Monteiro, Cannon and Levy11). However, our study also has limitations. Although the 24-hour dietary recall is often considered more accurate than the FFQ(Reference Thompson, Subar, Coulston and Boushey54), as it allows for the reporting of a wide variety of foods and beverages consumed the day before the interview, it still possesses certain limitations inherent to all methods for collecting dietary intake data, such as the memory bias(Reference Thompson, Subar, Coulston and Boushey54Reference Willett56). Nonetheless, compared to the FFQ, the 24-hour dietary recall requires short-term memory, which implies less memory bias than the FFQ(Reference Thompson, Subar, Coulston and Boushey54Reference Willett56). Moreover, in the NDS 2017–2018 study, the 24-hour dietary recall was applied by trained research interviewers who followed strict procedures to collect dietary intake, include using a software developed for this purpose and following the steps of the Automated Multiple-Pass Method(Reference Moshfegh, Rhodes and Baer57). Nonetheless, we used only the first 24-hour dietary recall, which means that our results are not of usual intake, although that one day is considered appropriate to estimate population average dietary intakes(Reference Dodd, Guenther and Freedman58).

Our study demonstrated the inverse relationship between the consumption of ultra-processed foods, processed foods and culinary processed ingredients with the adherence to a healthy and sustainable diet, while the share of unprocessed or minimally processed food in the total diet was directly associated with the adherence to sustainable dietary recommendations.

Financial support

L.T.C. and T.N.S. received a doctoral scholarship (grant number 2019/13424-0 and 2021/14782-7, respectively) from the Sao Paulo Research Foundation (FAPESP). No funding agency had a role in the study design, data collection, analysis, decision to publish or preparation of the article.

Conflict of interest

The authors declare no conflict of interest.

Authorship

L.T.C. designed the research; T.N.S. supported on the data curation; L.T.C. carried out the data analyses; M.L.C.L. and D.M.M. advised on statistical analyses; L.T.C. wrote the first draft; T.N.S., M.L.C.L. and D.M.M. critically reviewed and revised the manuscript. All authors read and approved the final version.

Ethics of human subject participation

This study used data from the IBGE, which follows the Brazilian Government law (number 73 177 from November 20, 1973) by respecting the ethical aspects of the individuals; the database information is confidential, with no subject identification, address or telephone number available. Therefore, ethical review and approval were waived for this study due to the data being secondary and publicly available.

Data availability statement

The raw data used in this study are publicly available and can be downloaded from the following electronic address (https://www.ibge.gov.br/estatisticas/sociais/saude/24786-pesquisa-de-orcamentos-familiares-2.html?edition=28523&t=microdata). The analytic code of the PHDI computation will be made available upon request pending to the corresponding author.

Supplementary material

For supplementary material accompanying this paper visit https://doi.org/10.1017/S1368980024001678

References

Tilman, D & Clark, M (2014) Global diets link environmental sustainability and human health. Nat 515, 518522.Google Scholar
Afshin, A, Sur, PJ, Fay, KA et al. (2019) Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the global burden of disease study 2017. Lancet 393, 19581972.Google Scholar
Springmann, M, Wiebe, K, Mason-D’Croz, D et al. (2018) Health and nutritional aspects of sustainable diet strategies and their association with environmental impacts: a global modelling analysis with country-level detail. Lancet Planet Health 2, e451e461.Google Scholar
Springmann, M, Spajic, L, Clark, MA et al. (2020) The healthiness and sustainability of national and global food based dietary guidelines: modelling study. BMJ 370, m2322.Google Scholar
Food and Agriculture Organization of the United Nations and World Health Organization (2019) Sustainable Healthy Diets: Guiding Principles. Rome: Food and Agriculture Organization.Google Scholar
Laine, JE, Huybrechts, I, Gunter, MJ et al. (2021) Co-benefits from sustainable dietary shifts for population and environmental health: an assessment from a large European cohort study. Lancet Planet Health 5, e786e796.Google Scholar
Wang, DD, Li, Y, Afshin, A et al. (2019) Global improvement in dietary quality could lead to substantial reduction in premature death. J Nutr 149, 10651074.Google Scholar
Willett, W, Rockström, J, Loken, B et al. (2019) Food in the anthropocene: the EAT–lancet commission on healthy diets from sustainable food systems. Lancet 393, 447492.Google Scholar
Seferidi, P, Scrinis, G, Huybrechts, I et al. (2020) The neglected environmental impacts of ultra-processed foods. Lancet Planet Health 4, e437e438.Google Scholar
Lawrence, MA, Baker, PI, Pulker, CE et al. (2019) Sustainable, resilient food systems for healthy diets: the transformation agenda. Public Health Nutr 22, 29162920.Google Scholar
Monteiro, CA, Cannon, G, Levy, RB et al. (2019) Ultra-processed foods: what they are and how to identify them. Public Health Nutr 22, 936941.Google Scholar
da Louzada, MLC, Ricardo, CZ, Steele, EM et al. (2018) The share of ultra-processed foods determines the overall nutritional quality of diets in Brazil. Public Health Nutr 21, 94102.Google Scholar
Martínez Steele, E, Popkin, BM, Swinburn, B et al. (2017) The share of ultra-processed foods and the overall nutritional quality of diets in the US: evidence from a nationally representative cross-sectional study. Popul Health Metr 15, 6.Google Scholar
Marchese, L, Livingstone, KM, Woods, JL et al. (2022) Ultra-processed food consumption, socio-demographics and diet quality in Australian adults. Public Health Nutr 25, 94104.Google Scholar
Pagliai, G, Dinu, M, Madarena, MP et al. (2021) Consumption of ultra-processed foods and health status: a systematic review and meta-analysis. Br J Nutr 125, 308318.Google Scholar
Elizabeth, L, Machado, P, Zinöcker, M et al. (2020) Ultra-processed foods and health outcomes: a narrative review. Nutrients 12, 1955.Google Scholar
Chen, X, Zhang, Z, Yang, H et al. (2020) Consumption of ultra-processed foods and health outcomes: a systematic review of epidemiological studies. Nutr J 19, 86.Google Scholar
Lane, MM, Davis, JA, Beattie, S et al. (2021) Ultraprocessed food and chronic noncommunicable diseases: a systematic review and meta-analysis of 43 observational studies. Obes Rev 22, e13146.Google Scholar
Taneri, PE, Wehrli, F, Roa-Díaz, ZM et al. (2022) Association between ultra-processed food intake and all-cause mortality: a systematic review and meta-analysis. Am J Epidemiol 191, 13231335.Google Scholar
Leite, FHM, Khandpur, N, Andrade, GC et al. (2022) Ultra-processed foods should be central to global food systems dialogue and action on biodiversity. BMJ Glob Health 7, e008269.Google Scholar
Silva, JT, Garzillo, JMF, Rauber, F et al. (2021) Greenhouse gas emissions, water footprint, and ecological footprint of food purchases according to their degree of processing in Brazilian metropolitan areas: a time-series study from 1987 to 2018. Lancet Planet Health 5, e775e785.Google Scholar
Instituto Brasileiro de Geografia e Estatística (2020) Pesquisa de Orçamentos Familiares 2017–2018: Avaliação Nutricional da Disponibiliade Domiciliar de Alimentos no Brasil. Rio de Janeiro: IBGE.Google Scholar
Instituto Brasileiro de Geografia e Estatística (2020) Pesquisa de Orçamentos Familiares (POF) 2017–2018: Análise do Consumo Alimentar Pessoal no Brasil. Rio de Janeiro: IBGE.Google Scholar
Instituto Brasileiro de Geografia e Estatística (IBGE) (2011) Pesquisa de Orçamentos Familiares 2008–2009: Tabela de Medidas Referidas Para os Alimentos Consumidos no Brasil. Rio de Janeiro: IBGE.Google Scholar
Coelho, KS, Giuntini, EB, Grande, F et al. (2019) 12th IFDC 2017 special issue – Brazilian food composition table (TBCA): development and functionalities of the online version. J Food Compost Anal 84, 103287.Google Scholar
Giuntini, EB, Coelho, KS, Grande, F et al. (2019) 12th IFDC 2017 special issue – Brazilian nutrient intake evaluation database: an essential tool for estimating nutrient intake data. J Food Compost Anal 83, 103286.Google Scholar
Cacau, LT, De Carli, E, de Carvalho, AM et al. (2021) Development and validation of an index based on EAT-lancet recommendations: the planetary health diet index. Nutrients 13, 1698.Google Scholar
Marchioni, DM, Cacau, LT, De Carli, E et al. (2022) Low adherence to the EAT-lancet sustainable reference diet in the Brazilian population: findings from the national dietary survey 2017–2018. Nutrients 14, 1187.Google Scholar
Cacau, LT & Marchioni, DM (2022) The planetary health diet index scores proportionally and considers the intermediate values of the EAT-lancet reference diet. Am J Clin Nutr 115, 1237.Google Scholar
Martinez-Steele, E, Khandpur, N, Batis, C et al. (2023) Best practices for applying the Nova food classification system. Nat Food 4, 445448.Google Scholar
Martini, D, Godos, J, Bonaccio, M et al. (2021) Ultra-processed foods and nutritional dietary profile: a meta-analysis of nationally representative samples. Nutrients 13, 3390.Google Scholar
da Louzada, MLC, Martins, APB, Canella, DS et al. (2015) Impact of ultra-processed foods on micronutrient content in the Brazilian diet. Rev Saude Publica 49, 18.Google Scholar
da Louzada, MLC, Martins, APB, Canella, DS et al. (2015) Ultra-processed foods and the nutritional dietary profile in Brazil. Rev Saude Publica 49, 38.Google Scholar
Cacau, LT, Hanley-Cook, GT, Huybrechts, I et al. (2023) Relative validity of the planetary health diet index by comparison with usual nutrient intakes, plasma food consumption biomarkers, and adherence to the Mediterranean diet among European adolescents: the HELENA study. Eur J Nutr 62, 25272539.Google Scholar
Previdelli, ÁN, de Andrade, SC, Pires, MM et al. (2011) A revised version of the Healthy Eating Index for the Brazilian population. Rev Saude Publica 45, 794798.Google Scholar
Garzillo, JMF, Poli, VFS, Leite, FHM et al. (2022) Ultra-processed food intake and diet carbon and water footprints: a national study in Brazil. Rev Saude Publica 56, 6.Google Scholar
Kesse-Guyot, E, Allès, B, Brunin, J et al. (2022) Environmental impacts along the value chain from the consumption of ultra-processed foods. Nat Sustain 6, 192202.Google Scholar
Cacau, LT, Benseñor, IM, Goulart, AC et al. (2023) Adherence to the EAT-lancet sustainable reference diet and cardiometabolic risk profile: cross-sectional results from the ELSA-Brasil cohort study. Eur J Nutr 62, 807817.Google Scholar
Cacau, LT, Benseñor, IM, Goulart, AC et al. (2021) Adherence to the planetary health diet index and obesity indicators in the Brazilian longitudinal study of adult health (ELSA-Brasil). Nutrients 13, 3691.Google Scholar
Stubbendorff, A, Sonestedt, E, Ramne, S et al. (2022) Development of an EAT-lancet index and its relation to mortality in a Swedish population. Am J Clin Nutr 115, 705716.Google Scholar
Xu, C, Cao, Z, Yang, H et al. (2022) Association between the EAT-lancet diet pattern and risk of type 2 diabetes: a prospective cohort study. Front Nutr 8, 784018.Google Scholar
López, GE, Batis, C, González, C et al. (2022) EAT-lancet healthy reference diet score and diabetes incidence in a cohort of Mexican women. Eur J Clin Nutr 77, 348355.Google Scholar
Delpino, FM, Figueiredo, LM, Bielemann, RM et al. (2022) Ultra-processed food and risk of type 2 diabetes: a systematic review and meta-analysis of longitudinal studies. Int J Epidemiol 51, 11201141.Google Scholar
Wang, M, Du, X, Huang, W et al. (2022) Ultra-processed foods consumption increases the risk of hypertension in adults: a systematic review and meta-analysis. Am J Hypertens 35, 892901.Google Scholar
Askari, M, Heshmati, J, Shahinfar, H et al. (2020) Ultra-processed food and the risk of overweight and obesity: a systematic review and meta-analysis of observational studies. Int J Obes 44, 20802091.Google Scholar
Nilson, EAF, Ferrari, G, da Louzada, MLC et al. (2022) The estimated burden of ultra-processed foods on cardiovascular disease outcomes in Brazil: a modeling study. Front Nutr 9, 1043620.Google Scholar
Ministério da Saúde (2014) Guia Alimentar Para a População Brasileira, 2nd ed. Brasília: Ministério da Saúde.Google Scholar
Canhada, SL, Luft, VC, Giatti, L et al. (2020) Ultra-processed foods, incident overweight and obesity, and longitudinal changes in weight and waist circumference: the Brazilian longitudinal study of adult health (ELSA-Brasil). Public Health Nutr 23, 10761086.Google Scholar
de Carvalho, AM, de Selem, SSC, Miranda, AM et al. (2016) Excessive red and processed meat intake: relations with health and environment in Brazil. Br J Nutr 115, 20112016.Google Scholar
Levy, RB, Andrade, GC, da Cruz, GL et al. (2022) Three decades of household food availability according to NOVA - Brazil, 1987–2018. Rev Saude Publica 56, 75.Google Scholar
Semba, RD, de Pee, S, Kim, B et al. (2020) Adoption of the ‘planetary health diet’ has different impacts on countries’ greenhouse gas emissions. Nat Food 1, 481484.Google Scholar
Tuninetti, M, Ridolfi, L & Laio, F (2022) Compliance with EAT–lancet dietary guidelines would reduce global water footprint but increase it for 40 % of the world population. Nat Food 3, 143151.Google Scholar
Hirvonen, K, Bai, Y, Headey, D et al. (2020) Affordability of the EAT–lancet reference diet: a global analysis. Lancet Glob Health 8, e59e66.Google Scholar
Thompson, FE & Subar, AF (2013) Dietary assessment methodology. In Nutrition in the Prevention and Treatment of Disease, 3rd ed. pp. 546 [Coulston, A, Boushey, C, editors]. Cambridge, MA: Academic Press.Google Scholar
Shim, J-S, Oh, K & Kim, HC (2014) Dietary assessment methods in epidemiologic studies. Epidemiol Health 36, e2014009.Google Scholar
Willett, W (2012) Nutritional Epidemiology, 3rd ed. Oxford: Oxford University Press.Google Scholar
Moshfegh, AJ, Rhodes, DG, Baer, DJ et al. (2008) The US department of agriculture automated multiple-pass method reduces bias in the collection of energy intakes. Am J Clin Nutr 88, 324332.Google Scholar
Dodd, KW, Guenther, PM, Freedman, LS et al. (2006) Statistical methods for estimating usual intake of nutrients and foods: a review of the theory. J Am Diet Assoc 106, 16401650.Google Scholar
Figure 0

Table 1 The Planetary Health Diet Index (PHDI) components and criteria for scoring system (recommended and lower or upper limit recommended values*)

Figure 1

Table 2 Distribution of the Planetary Health Diet Index (PHDI) and the caloric share (%) of Nova food groups for the Brazilian population and across sociodemographic characteristics (n 46 164). National Dietary Survey, 2017–2018

Figure 2

Fig. 1 Crude linear splines of the association between the Planetary Health Diet Index (PHDI) score and the caloric share (%) of each Nova food group continuously (n 46 164). National Dietary Survey, 2017–2018. β coefficients and their respective 95 % CI. Legend: (a): linear spline of the association between the PHDI score and the caloric share (%) of ultra-processed foods. (b): Linear spline of the association between the PHDI score and the caloric share (%) of unprocessed or minimally processed foods. (c): Linear spline of the association between the PHDI score and the caloric share (%) of processed culinary ingredients. (d): Linear splines of the association between the PHDI score and the caloric share (%) of processed foods

Figure 3

Fig. 2 Crude linear coefficients and 95 % CI of the association between Planetary Health Diet Index (PHDI) scores and quintiles of the caloric share (%) of ultra-processed foods (n 46 164). National Dietary Survey, 2017–2018. Legend: a 1st quintile: mean (min – max): 0·43 (0 – 3·40); 2nd quintile 6·83 (3·40 – 9·92); 3rd quintile 13·3 (9·92 – 17·3); 4th quintile 22·8 (17·3 – 29·6); 5th quintile 45·1 (29·6 – 100). Values are expressed as linear coefficients and 95 % CI

Figure 4

Fig. 3 Adjusted linear coefficients and 95 % CI of association between Planetary Health Diet Index (PHDI) scores and quintiles of the caloric share (%) of ultra-processed foods (n 46 164). National Dietary Survey, 2017–2018. Legend: a 1st quintile: mean (min – max): 0·43 (0 – 3·40); 2nd quintile 6·83 (3·40 – 9·92); 3rd quintile 13·3 (9·92 – 17·3); 4th quintile 22·8 (17·3 – 29·6); 5th quintile 45·1 (29·6 – 100). Model adjusted by sex, age, per capita income, self-reported race and residence area. Values are expressed as linear coefficients and 95 % CI

Figure 5

Table 3 Linear regression coefficients and their 95 % CI for the association between the caloric share (%) of non-ultra-processed food groups (i.e. unprocessed or minimally processed foods, processed culinary ingredients and processed foods) and the Planetary Health Diet Index (PHDI) total score (n 46 164). National Dietary Survey 2017–2018

Supplementary material: File

Cacau et al. supplementary material

Cacau et al. supplementary material
Download Cacau et al. supplementary material(File)
File 152.8 KB