In recent years, public health organizations have paid close attention to ultra-processed foods (UPFs), which are industrially produced foods that contain additives and typically have high energy density and are high in saturated fats, sugar, and sodium (HFSS) (1, 2). Numerous studies have shown that consumption of UPFs is linked to increased obesity and associated health outcomes worldwide (3–15). Apart from the low-quality nutrient profile of UPFs (high in added sugars, unhealthy fats, and energy), it has been shown that ultra-processed diets are associated with the replacement of healthy dietary components, such as fiber and micronutrients (16–19). Further, some evidence suggests that UPFs are ultra-palatable and generate a reward stimulus, which has been likened to an addiction (20, 21).

Despite these health concerns, UPFs are increasingly prevalent around the globe, accounting for up to almost 60% of the average per capita daily energy intake in high-income countries (4, 5, 22–24), with rapid increases in many low- and middle-income countries (25, 26). However, significant differences remain between countries in terms of consumption, even among countries in more advanced stages of the nutrition transition. For example, recent estimates in the U.S. suggest that between 57 and 67% of total calories come from UPFs, compared to around 30% in Brazil, Chile, and Mexico (3, 27–32).

A major unresolved question relates to the prevalence of UPFs in the packaged food supply, which accounts for a sizeable proportion of diets. Understanding this is important for several reasons. First, given the flow of migrants from one country to another, understanding the prevalence of UPFs in the packaged food supply can help provide context on how individuals may be differentially exposed to UPFs over time (33–43). Second, an increasing number of countries in Latin America have implemented policies to reduce the consumption of foods HFSS. Although UPFs and HFSS are different constructs, there is a high degree of overlap between them, making it useful to understand whether countries with HFSS policies have a lower proportion of UPFs in their food supply. Lastly, alongside concerns about UPFs in general, there have been growing concerns about specific additives, including non-nutritive sweeteners (NNS), colorings, flavorings, and emulsifiers or preservatives, and their association with cancer, metabolic disturbances, and cognitive and gut issues. Despite these concerns regarding UPFs, to our knowledge, no studies have characterized the food supply in Latin American countries in terms of the prevalence of UPFs, HFSS, or additives.

To address this critical gap in the field, we leveraged a unique data resource to generate estimates of the: (1) prevalence of UPFs in the food supply of 11 countries in Latin and North America (United States, Argentina, Brazil, Chile, Colombia, Costa Rica, Ecuador, Mexico, Peru, Puerto Rico, and Venezuela) between 2018 and 2023; (2) prevalence, and average number, of additives, which include preservatives, colorants, flavor enhancers, and non-nutritive sweeteners (NNS), in the packaged food supply in the U.S. compared to Latin American countries; and (3) prevalence of foods and beverages high in calories, sugar, sodium, or saturated fat in UPFs and non-UPFs, by country.

The final sample included 207,363, of which 80.6% were foods and 19.4% were beverages. The U.S. had the most newly launched products in the period between 2018 and 2023, with 81,693 products (39.4%), followed by Brazil (42,435 products, 20.5%) and Mexico (25,169, 12.1%). The country with the least newly launched products was Venezuela, with 3,030 (1.5%).

Table 1 provides information about the number of newly launched products by country. Thus, Table 1 also lists the products for which analyses of foods and beverages “high in” were limited.

Figure 1 displays the percentage of newly launched products classified as ultra-processed by country, which ranged between 69 and 85% of the products in Venezuela and Costa Rica, respectively (Supplementary Table 4). Despite percentages being statistically different from the U.S., where the prevalence of UPFs was 81%, the magnitude of these differences was modest. An exception was Venezuela, where less than 70% of products were UPFs.

Overall, the percentage of beverages considered ultra-processed was slightly higher than the percentage of foods for all countries (Table 2; Supplementary Tables 5, 6). As expected, the fruits and vegetables group consistently had the lowest percentage of UPFs, ranging between 35% (Puerto Rico) and 64% (Peru). Similarly, between 25% (Venezuela) and 72% (Puerto Rico) of pasta, rice, and other starchy dishes were ultra-processed. On the other hand, candy, gum, chocolate, desserts, and ice cream had the highest prevalence of UPFs, ranging from 97% (Peru) and 99% (Chile and Colombia), and between 94% (Brazil and Ecuador) and 100% (Chile and Peru), respectively. Among beverages, ultra-processing was high for all categories, with more than 80% of beverages being UPFs across all countries. The highest percentage of UPFs observed for carbonated soft drinks and sports and energy drinks, in which almost 100% of products were UPFs.

We next examined the percentage of foods and beverages containing additives by country (Figure 2). Flavor additives were the most prevalent class in the food supply, ranging from 60% of all products in Venezuela to 78% in Costa Rica. The other additives (e.g., emulsifiers and thickeners) class was less prevalent, ranging from 49% of products in Venezuela to 70% in Costa Rica. The percentage of foods and beverages containing preservatives ranged from 50% (Venezuela) to 64% (Costa Rica). Between 23% (Venezuela) and 44% (Argentina, Mexico, and Puerto Rico) of products contained coloring. NNS were the least prevalent across all countries, ranging from 7% in Puerto Rico to 20% in Chile.

Countries varied in terms of the mean number of additives among products (Table 3). Among products that contained additives, the mean number varied between 3.9 additives for Venezuela and 7.1 for Peru. Across all countries, Venezuela had the lowest means for all categories of additives except NNS. The highest means were observed for flavor additives, which ranged from 2.8 (Venezuela) to 4.8 (U.S.), and for other additives, ranging from 2.3 (Venezuela) to 4.0 (Peru).

We compared the country-specific percentage of foods “high in” for UPFs and non-UPFs (Figure 3; Supplementary Table 7). As expected, a large percentage of UPFs were considered high in energy, sugar, and saturated fats. Interestingly, non-UPFs products had a similar or higher prevalence of high sodium than UPFs in most countries, except Colombia, Ecuador, Peru, and Puerto Rico.

Overall, smaller percentages of beverages were “high in” than foods (22 to 67% of beverages vs. 51–89% for foods, Figure 3; Supplementary Table 8). In Venezuela, none of the newly launched, non-UPFs beverages were classified as high in energy, sodium, or saturated fats. A higher percentage of ultra-processed beverages, compared to non-ultra-processed (17–52% vs. 0–24%), was classified as high in energy, except in Ecuador and Mexico. In general, a larger percentage of UPFs beverages were high in sugar, except in the U.S. (25–52% vs. 5–33%), and high in sodium, except in Brazil, Ecuador, Mexico, and Peru, compared to non-UPFs beverages (11–37% vs. 0–33%). In most countries, a larger prevalence of non-UPFs beverages high in saturated fats was observed, compared to UPFs, except for the U.S., Argentina, Puerto Rico, and Venezuela (0–38% vs. 4–17%).

In this study, the prevalence of newly launched products between 2018 and 2023 in the U.S. and Latin America that were considered ultra-processed was high across all countries, with 81% in the U.S. and ranging from 69% (Venezuela) to 85% (Costa Rica). However, it is important to note that the availability of packaged foods in those countries is still different. For instance, in the U.S., over 80,000 products were introduced or modified during that period, whereas in other countries with a smaller population, such as Chile, Costa Rica, Ecuador, Peru, Puerto Rico, and Venezuela, that number was much lower, ranging between 3,000 and 6,000. The prevalence of packaged and ultra-processed foods in the diet of these countries’ populations still varies, and a high prevalence of UPFs among newly launched products does not necessarily translate into a high percentage of consumption since intake so depends on the prevalence of UPFs in existing products as well as consumption of packaged and non-packaged foods in general. Still, given the association between UPFs and adverse health outcomes (14, 59–62), the overall high prevalence of UPFs among newly launched products across countries represents an important public health concern.

The prevalence of UPFs in soft drinks was high across all countries. This finding is expected, given that sodas and sports/energy drinks typically contain colors, flavors, or sweeteners. More unexpectedly, the proportion of UPFs in the category of plain and flavored waters was also relatively high across countries (63–91%), suggesting a high prevalence of flavored waters since bottled waters are not UPFs. This high rate may potentially reflect an artifact of the Mintel database, which generates new records every time a product is reformulated or re-marketed (e.g., changes packaging). Since bottled water cannot be reformulated and may be less subject to new marketing schemes, it may have fewer records in the database than flavored waters, which is a rapidly growing category. The global flavored water market size in 2022 was estimated to be US$16.6 bi and expected to grow to US$ 40.6 billion in 10 years (63).

Mirroring the high prevalence of UPFs, a high percentage of newly launched products contained additives, particularly flavoring (60–78%) and other additives (i.e., emulsifiers, thickeners, etc.; 49–70%). Among the additives, NNS was the category with the lowest prevalence (7–20%) for all countries. Overall, Venezuela was the country with the lowest mean number of additives across all categories except NNS, likely due to the ongoing political and economic scenario (64, 65). The total number of additives varied between 3.9 (Venezuela) and 7.1 (Peru). A study investigating additives in products purchased by U.S. households found that, between 2001 and 2019, the mean number of additives increased from 4.0 to 4.6, a lower average than what was observed among newly introduced products in the same market (56). In addition, the same study found that between 49.6 and 59.5% of products contained any additive, an estimate lower than what we observed (56). This might be due to the outcomes weighted by sales volume, while the current study only focuses on product availability in the market and not on purchases by consumers. Our results were in line with a study developed in Brazil, which concluded that almost 80% of the products contained at least one additive, with flavoring agents being the most commonly found (58.8%) (66). It is possible that that percentage was even higher, given the different classifications of additive classes adopted in the study. Another study investigating the presence of 64 most commonly used additives in the U.S. estimated that 64.9% of the foods contained at least one additive (67). In France, a study revealed that 53.8% of the food supply contained at least one additive (68).

The excess of nutrients of concern and energy is one of the top reasons UPFs have drawn the attention of public health organizations. In this study, UPFs were, in general, higher in nutrients of concern and energy than non-UPFs, supporting that view, especially among foods. However, it is important to note that, although beverages had, in general, a lower percentage of products high in energy, sugar, sodium, and saturated fats than foods, they had a higher prevalence of products considered UPFs. This is because beverages often contain flavoring and coloring but are rarely considered high in nutrients of concern other than sugar. In addition, diet beverages often contain NNS, which reduces their sugar content but increases the number of additives (69–71). This indicates that, although there is an overlap between nutrient content and the presence of markers of ultra-processing, which include additives, these two approaches are not the same. This is supported by Popkin et al., who concluded that the definition of foods and beverages HFSS did not encompass all products that should be targeted by policy as unhealthy, and that the best method to identify them is a combination between Nova and HFSS (55, 72).

This study has public health implications. We found an overall equally high prevalence of UPFs in the U.S. and Latin American countries. Studies have described the transformation of the food system in Latin America in the last decades, with increased participation of large retailers and multinational food industries (73–75). The increased availability of UPFs has contributed to dietary changes, including increased UPFs intake and, consequently, multiple negative health outcomes (76–78). However, it is still unclear whether this is due to the unique aspects of processing, which can lead to overconsumption or higher energy density and, hence, higher caloric intake (79). We also found a high prevalence of food additives. However, additive categories are broad classes representing a range of different substances with different biological impacts on the body. Although studies have demonstrated that additives are considered safe in small amounts (80), others have shown associations between some of those compounds and allergic reactions, hyperactivity, gut dysbiosis, increased risk of cancer, obesity, and other metabolic disturbances, and gastrointestinal issues (81–85). It is challenging to determine the dose–response or how much individuals usually ingest, given that nutrition facts labels do not include the amount included in foods and beverages. Furthermore, most products contain multiple additives, making it difficult to pinpoint the individual impact of each additive on health (86). However, it is important to also consider the possible synergistic and antagonistic effects between different additives. For instance, a recent prospective cohort study has found associations between commonly consumed additive mixtures and type 2 diabetes (87). Future studies should focus on better understanding the potential detrimental effects of additives in the diet to inform regulations.

This study also has policy implications. During the study period, many Latin American countries have implemented food policies focusing on front-of-package warning labels to inform about the presence of nutrients of concern, such as sodium, added sugars, and saturated fats. Currently, Brazil, Chile, Colombia, Mexico, and Peru have such policies, with Mexico and Colombia also requiring warning labels for the presence of NNS and Mexico and Chile requiring a label for excess calories. This approach, while effective in informing decisions and contributing to a decrease in the intake of these nutrients, can have effects on the formulation of products since companies are incentivized to reduce sugar, sodium, and saturated fat to avoid the policy (88–93). However, these reductions in nutrients of concern do not necessarily translate to reductions in the prevalence of UPFs since sugar, sodium, and saturated fat content alone do not make a product UPFs, suggesting that such policies could keep the prevalence of UPFs stable or possibly even increase it, without improving the products’ healthfulness. An interesting example is beverages: the data from Chile show a substantial replacement of sugar with NNS in beverages, with a 35.4% increase in sweetness from NNS and a 14.5% decrease in sweetness from total sugars (94). This would seem to increase the prevalence of UPFs in beverages since NNS is an additive; however, most beverages with added sweeteners also often contain added color and/or flavor. Although the reformulation might be beneficial in reducing sugar, it is likely that the percentage of UPFs would remain the same. Unfortunately, we were unable to evaluate percentual changes in UPFs over time due to the pooled nature of our data. Annual-level data would be useful to examine how the food supply changed regarding both UPFs and HFSS in response to policy action. At a minimum, our results suggest that if policymakers want to reduce UPFs, they will need to use broader criteria than the classic approach of looking at HFSS only to address this issue.

This study is not without limitations, with the main one being related to the dataset. The GNPD includes only packaged, newly launched products. This can introduce bias to the analyses in a few different ways. First, a large proportion of minimally processed foods, such as fruits, vegetables, milk, eggs, and meats, are not packaged and thus were not included in the analyses, which can result in overestimates of the proportion of UPFs. Second, the dataset is restricted to newly launched products, and unless products underwent any reformulations and/or changes in packaging between 2018 and 2023, those were not included in the analyses despite potentially being often consumed. This could introduce selection bias since UPFs are more likely to undergo reformulations and changes in packaging as a result of marketing strategies (95–99) and thus more likely to be included in the dataset. However, despite these limitations, the study is informative about what types of products are being introduced in the markets and might reflect trends in consumer behavior and desirability.

Some other limitations include not considering the market share of products, instead attributing equal weight to every newly launched product. This prevents any inferences about the intake of UPFs, given that ultra-processed products could be consumed in smaller or larger frequency and quantity. Future studies should investigate the intersection between food supply and purchases or dietary intake regarding ultra-processing. In addition, many of the products did not contain information about nutrients of concern or energy, and this was differential across countries due to distinct local laws and regulations for the required information in the NFL and/or packaging, which limited our ability to apply the Chilean NPM and could have resulted in selection bias. In particular, our results might be an underestimation if manufacturers chose not to include that information on products that tend to be higher in those nutrients. Lastly, our algorithmic approach to identify additives and classify UPFs did not undergo a formal validity test.

The study also presents strengths. First, it investigates a large database of newly introduced foods and beverages in the U.S. and ten other markets across Latin America. Second, the database comprises information contained in the packaging of products, which results in objective information about the nutritional content. Third, identifying UPFs algorithmically proves beneficial because it allows for ingredient lists of individual products to be searched for additives, avoiding subjectivity and the classification of every product in the same group in a similar manner, despite the composition. Lastly, to identify UPFs, we searched for a comprehensive list containing more than 4,000 additives, increasing the likelihood of capturing all products.

Newly launched products in the U.S. and Latin America are largely ultra-processed, with a high prevalence of additives and products high in energy, sodium, sugar, and saturated fats, nutrients that contribute to the development of obesity, non-communicable diseases, and early mortality. However, since this study only reflects newly launched products, future studies should investigate the prevalence of UPFs across the entire food supply. Policies such as front-of-package labeling might be helpful in informing the general public that foods are ultra-processed.