Ultra-processed foods consumption and diet quality among preschool children and women of reproductive age from Argentina

We utilised dietary data collected during the 2005 National Survey of Health and Nutrition (ENNyS) of Argentina, a nationally representative and cross-sectional survey of maternal and child populations of urban areas carried out by the Ministry of Health⁽³²⁾. Using a probabilistic multi-stage sample involving localities, census blocks and three independent samples: 32 474 boys and girls between 6 months and 5 years, 8307 women aged 11–49 years, and 1941 pregnant women. We analysed data from 7022 children (2–5 years of age), 2165 adolescent females (10–19 years of age) and 4414 adult women (20–49 years of age).

Food consumption was collected at the interviewee’s home by one 24-h dietary recall interview conducted by trained nutritionists, for adolescents and adults or to the person responsible for feeding young children in the case of 2–5-year-old participants. The institution’s managers gave information about the food consumed in day-care centres and schools by children and adolescents. It was registered the consumption of everything eaten by individuals the day before the survey, including food, beverages (except drinking water and infusions), and mineral and vitamin supplements of particular interest. The interviewers used visual food models with colour photographs of portions of different food sizes and references of raw and cooked weight amounts. The composition of some preparations was previously standardised using essential and regional recipes. Details on the methodology of ENNyS can be found elsewhere⁽³²⁾.

Additional information for the analysis was obtained from a general questionnaire with details of household characteristics and assets and sociodemographic characteristics of the head of households and participants. Aspects of the study were provided, and participants were asked to sign informed consent before their inclusion in the study.

The food items were sorted into mutually exclusive food subgroups within: (1) unprocessed or minimally processed foods (eleven subgroups: e.g. fresh meat, roots, and tubers, cereals, vegetables, legumes, fruits); (2) processed culinary ingredients (four subgroups: e.g. plant oils, table sugar, animal fats); (3) processed foods (five subgroups: e.g. unpackaged fresh bread, cheese, ham, and salted meat, vegetables and fruits preserved in brine or sugar syrup); and (4) UPF (seventeen subgroups: e.g. carbonated soft drinks, sweet or savoury snacks, confectionery, industrial desserts, reconstituted meat products, shelf-stable or frozen meals, industrial packaged bread), according to the grade of processing by following the NOVA system classification⁽⁶⁾. Details of categorisation can be found in Table 1↗ in Supplemental Material.

To obtain the nutritional profile of diets, we estimated the energy and nutrients provided by each food item. Given that a complete local database for the nutrient compositions of foods and beverages is not available for Argentina, we combined different sources of nutritional composition databases. Although the Argenfoods database⁽³³⁾ was the primary source, we also utilised the USDA database⁽³⁴⁾, the Germany database⁽³⁵⁾ and the National Nutritional Institute of Salta University database⁽³⁶⁾. Also, we have used the information from food labels for some products that are not reported in any of the databases mentioned above and some data obtained in the CESNI laboratory (Centro de Estudios Sobre Nutrición Infantil, a Non-Government Organization of Argentina).

Thus, the diet’s nutrient composition was assessed by the daily intake of energy, total proteins, total carbohydrates, available carbohydrates, free sugars, fibre, total fats, saturated fats, Na and K. It was calculated considering the edible part of each food item and their energy and nutrient composition. Finally, the contribution of every NOVA food group was estimated.

The intake was reported (mean and standard error) in absolute terms (grams or milligrams) and considering the energy intake relative to the total daily intake of the specific participant.

The dietary pattern of children, adolescent females and adult females was described by distributing the total energy intake according to the four NOVA food groups, and within these groups, according to selected subgroups.

After that, we analysed the average energy contribution of UPF and the impact of the consumption of these products on specific food groups and critical nutrients related to NCD by considering the recommendations of the WHO⁽³⁷⁾. Firstly, the 421 food items reported in the 24-h dietary recall were classified into five food groups, fresh vegetables, fresh fruits, legumes, nuts and seeds, and wholegrains (Table 2↗ in Supplemental Material) and were estimated the consumption in 2000 kcal/d. Secondly, we assess the daily intake of protein, free sugar, fibre, total fats, saturated fat, trans-fat, fibre, Na and K^(37–41). Fibre, Na and K intake were expressed per 2000 kcal, while the other nutrients were expressed as a percentage of total energy intake. The energy density of solid fraction was calculated by dividing the sum of energy from the intake of solid foods by the amount in grams of these foods. The recommendations used for this indicator were those proposed by the World Cancer Research Fund⁽⁴²⁾.

Each nutritional indicator was estimated for the overall diet, for NOVA 1-2-3 categories (the fraction of non-UPF) and NOVA 4 category (the fraction of UPF). The indicators were used to evaluate the dietary quality of the population strata corresponding to the distribution quintiles of caloric contribution from UPF to total calories. Linear regression analyses were used to identify the direction and the statistical significance of the association between the distribution quintiles of caloric contribution from UPF (as % of the energy of UPF) and nutritional indicators as the dependent variable, with and without fitting for confounding variables (age, gender in children 2–5 years, geographic region, years of schooling of the head of the family in children and adolescent aged 10–19 years, years of education of women aged 20–49 years, unsatisfaction of basic needs, household income per capita divided into quintiles).

Individuals were classified into five strata according to the caloric value that UPF contributed to the total value of their diet. These strata were related to the distribution quintiles of caloric contribution from UPF across the population.

We estimated a mean daily energy intake of 1636 ± 620 kcal among children, 1956 ± 829 kcal in adolescent females and 1695 ± 800 kcal in adult females, with differences in the participation of each NOVA food group (Fig. 1). For more details on the contribution of each food item, see Table 1.

Minimally processed food accounted for 41 % of daily energy intake in children, while in adolescent and adult females, the intake of minimally processed foods provided 31 % and 33 % of daily energy, respectively. Milk and plain yogurt were the essential food items in children (15 %), and meat and cereals contributed with most calories in adolescent and adult females (10–12 %). The relative energy intake from processed culinary ingredients was 15 % in children, 16 % in adolescent females and 20 % in adult females, being table sugar the primary source of calories in this group (about 7–10 %). Processed foods contributed 16 % of total calories in children, 21 % in adolescent females and 20 % in adult females. The main contributor was fresh bread with about 12–15 % of total energy.

Concerning UPF, they represented 27·5 % of daily energy in children and 31·5 % in adolescent females, while in adult females, the participation was 26 %. Most calories from UPF were related to cookies, pastries, cakes, packaged bread and crackers (about 11 % of total daily intake), followed by soft drinks and juices (about 5 % of total daily intake). Sweets and candies represented 5 % of total daily intake in children and adolescents and 2 % in adult women. Milk-based drinks and yogurt contributed 3 % of energy in children and 1 % in adolescents and women.

As can be observed in Table 2, the overall diet of the analysed Argentinian population presented an unhealthy profile. On the one hand, the number of fruits and vegetables consumed was very low (169–227 g/d), well below the WHO recommendations (400 g/d). In addition, the total intake of wholegrains, legumes, nuts and seeds reached a low quantity of 7–14 g/d. On the other hand, the consumption of free sugar and saturated fats exceeds the recommended levels (<10 % of the total energy of both nutrients) in the three age groups. Moreover, the consumption of fibre and potassium was deficient, three- and twofold below the optimal intake level, respectively. In addition, although the reported Na intake was relatively low, the Na:K ratio was <1:3 in all groups. However, the survey did not account for salt intake during meal preparation or at the table. Thus, Na consumption could have been higher than the figure reported here, increasing the Na:K ratio further. It is worth mentioning that the consumption of trans-fats was just below the recommended intake (<1 %).

UPF contributed significantly to the intake mentioned above of unhealthy nutrients. Compared to the fraction of the diet concerning non-UPF, the fraction referring to UPF has 1·6 to 1·8 more energy/g, 1·5 to 2 more trans-fat, 2·1 to 4·2 more added sugar, 3·3 to 5·3 more Na, depending on the age group. In addition, the UPF fraction was two times lower in fibre, proteins and K content. Non-processed and ultra-processed fractions have similar fat and saturated fat contents, especially in adolescents and women. Regarding saturated fats, its primary sources were both UPF and non-UPF because of the high consumption of animal products in Argentina, particularly red meat (Table 2).

An interesting pattern emerged when the population was classified according to their consumption of UPF and adjusted by sociodemographic variables. Table 3 presents the nutritional dietary profile indicators for the five strata of the population corresponding to increasing quintiles in energy contribution from UPF. Besides UPF being an important source of free sugar in all the analysed groups, a positive association between UPF consumption and free sugar intake was found only in children and adolescent females (r = 0·47 and 0·34, P < 0·001). While UPF did not contribute significantly to the intake of trans-fat, higher consumption of UPF was also positively associated with trans-fat intake in the three groups (r = 0·40, 0·46 and 0·51, P < 0·001). Na, total fat and saturated fat also increased with higher consumption of UPF, but the association was weaker (r < 0·30, P < 0·001; see Table 3). The same trend was observed for fibre, protein and K (r < -0·35, P < 0·001; Table 3 for more details). Also, we found a negative association between the consumption of UPF and fresh and whole foods (r < -0·30, P < 0·001), showing a potential displacement effect. This result was particularly evident in the highest quintile of UPF consumption, in which a deficient intake of fresh fruits was observed in all age groups analysed.