**Food Labeling: Analysis, Understanding, and Perception**

Editors

**Daniela Martini Davide Menozzi**

MDPI ' Basel ' Beijing ' Wuhan ' Barcelona ' Belgrade ' Manchester ' Tokyo ' Cluj ' Tianjin

*Editors* Daniela Martini Dept. Food, Environmental and Nutritional Sciences (DeFENS) University of Milan Milan Italy

Davide Menozzi Dept. Food and Drug University of Parma Parma Italy

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Nutrients* (ISSN 2072-6643) (available at: www.mdpi.com/journal/nutrients/special issues/Food Labeling).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-1255-6 (Hbk) ISBN 978-3-0365-1254-9 (PDF)**

© 2022 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

### **Contents**




### **About the Editors**

#### **Daniela Martini**

Daniela Martini is an assistant professor of human nutrition at the University of Milan, Italy. She is also member of the Board of Directors for the Italian Society of Human Nutrition (SINU) and coordinates the SINU Young Working Group. Her expertise includes food regulation (i.e., food labelling, nutrition, and health claims), the analysis of antioxidant compounds in foods, and the evaluation of the role of foods and dietary patterns in the modulation of markers of health.

#### **Davide Menozzi**

Davide Menozzi is an associate professor of agricultural economics and rural appraisal at the University of Parma, Italy. His main research interests include analyses of consumer behavior and preferences, the economics of food quality schemes, and the evaluation of the socioeconomic sustainability of dietary behaviors. He teaches graduate and undergraduate agricultural economics courses at the University of Parma. He is an active member of the Italian Association of Agricultural and Applied Economists (AIEAA) and the European Association of Agricultural Economists (EAAE).

### *Editorial* **Food Labeling: Analysis, Understanding, and Perception**

**Daniela Martini 1,\* and Davide Menozzi 2,\***


Food labels are the first informative tool found by the customers during shopping, and are informative in terms of ingredients, nutrient content, and the presence of allergens of the selected product. However, food labeling also represents a marketing tool and may influence perception of the food quality and, in turn, the dietary choice of consumers. For this reason, there is growing research in the food labeling field and in the evaluation of its effects on consumers, food operators, and the whole market [1,2]. This is supported by a wide range of manuscripts published in recent years, for instance, with the specific purpose to better investigate how specific information on the food packaging may influence food purchases and consumption and, in general, dietary behavior [3–8]

The Special Issue "*Food Labeling: Analysis, Understanding, and Perception*" was conceived with the intention to further explore current efforts in food labeling research and welcomed original studies, as well as reviews of the literature, focusing on: (i) the analysis of the nutrient profile of products with different characteristics reported on the food labels, i.e., nutrition and health claims (NHCs), organic, gluten-free (GF); (ii) the nutrient profile underlying front-of-pack (FOP) nutrition labels and their graphical design in different countries; (iii) the consumers' perception, knowledge, and understanding of the information provided on food labeling; (iv) the impact of information on food labeling (e.g., FOP information, serving size) on consumers' willingness to pay and food choice; (v) the attitudes, beliefs, and perceptions and behavioral and socioeconomic determinants regarding the use of food labels.

This Special Issue provides a series of 25 contributions, with 20 original papers, four narrative reviews, and one commentary. This last article is a consensus by eminent exponents of the International Carbohydrate Quality Consortium (ICQC) [9], which underlines the importance of dietary fiber, which is not always mandatory on food labeling (e.g., Reg. (EU) No 1169/2011) [10]. The authors supported the need for including fiber values in food labeling by distinguishing between intrinsic and added fiber, which may also help to achieve the recommended intake by consumers. The need to consider other information on food labels has also been discussed by Marinangeli and colleagues [11], who reviewed the regulatory frameworks and examples of associated non-mandatory food labeling claims currently employed to highlight healthy carbohydrate foods to consumers. Among the information, the authors considered NHCs related to dietary fiber, glycemic index, and glycemic response, and the presence of whole carbohydrate foods and ingredients that are intact or reconstituted (e.g., whole grains).

Some studies focused on the analysis of the nutritional quality of specific food groups and/or specific nutrients. Three studies were performed within the Food Labeling of Italian Products (FLIP) [12–14], a project aiming to evaluate the nutritional quality of packaged products currently sold in the online shops of several retailers in Italy [15–17]. Specifically, two studies focused on the analysis of the food labeling of breakfast cereals [12] and pasta [14]. The first study reported an elevated inter-product variability among breakfast cereals currently sold in Italy, with only limited differences when products with NHCs and GF declarations were compared with products not carrying this information [12].

**Citation:** Martini, D.; Menozzi, D. Food Labeling: Analysis, Understanding, and Perception. *Nutrients* **2021**, *13*, 268. https:// doi.org/10.3390/nu13010268

Received: 13 January 2021 Accepted: 15 January 2021 Published: 19 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Similarly, the study performed on pasta revealed that pasta types currently on the Italian market largely vary in terms of nutrition profile, with stuffed pasta characterized by a high salt content [14]. This last aspect supports the importance of providing nutrition facts of product to consumers to help them in making informed food choices. The last study performed within the FLIP project compared the nutritional quality of organic and conventional food products, highlighting that, with just a few exceptions, prepacked organic products are not of a superior nutritional quality than conventional ones, based on the mandatory information present on their packaging [13].

Yusta-Bojo et al. [18] focused on the sugar content in the most-consumed processed foods in Spain and compared the sugar values declared on the label (LVs) with laboratory analysis values (AVs). The study findings evidenced a high adequacy of LVs with the EU labeling tolerance requirements, with only cured ham presenting significant differences between the median AVs and LVs. Lastly, Azzopardi et al. investigated the energy density (ED) of food products targeted at children sold in Australia, finding a high proportion of products with a high ED (i.e., >950 kJ/100 g) among the 548 food items considered [19]. The same study observed that the health star rating (HSR) system, one of the FOP systems introduced in Australia in 2014, did not consistently discriminate between ED levels, particularly for high-ED foods.

The HSR was also studied in another study focused on consumers' perception [20] and performed with fifteen Australian grocery shoppers. Intriguingly, the findings from this study showed that the HSR was perceived as a simple, easy-to-understand, and useful tool, despite a certain grade of skepticism concerning its conception. The consumers' perception and responses to FOP labels was also considered in another two papers published by Egnell et al. [21] and Talati et al. [22], showing results from the Netherlands and across another 12 countries, respectively, supporting that this represents a widely explored field of research. In detail, the first study [21] compared the perception and understanding of five FOP labels (HSR, Nutri-Score, multiple traffic lights (MTLs), reference intake, and warning symbols) among 1032 Dutch participants, finding a favorable perception, with Nutri-Score showing the highest performance in helping consumers to rank the products according to their nutritional quality. Conversely, in a similar study performed with over 12,000 participants across 12 countries, MTLs obtained the most favorable ratings, with mixed or neutral perceptions of the other FOP labels [22]. A third study by Breen et al. [23] compared NHCs, the HSR, and the price of snack foods sold in health food (HF) stores and aisles with the ones sold in regular areas (RAs) of supermarkets. The results showed that snack foods of HF stores displayed a significantly higher number of product claims compared to RA foods, together with a higher HSR and cost.

Botelho et al. [24] analyzed the FOP of food items shown in specific sections of the circulars of two Brazilian supermarket chains during a 10-week period, classifying them by their "unprocessed/minimally processed" versus "ultraprocessed" (UP) items and the presence and type of claims on the FOP. The NOVA systems represent another way of classifying foods that has receiving growing interest and which is based on the degree of food processing [25]. In this Special Issue, authors found that more than 50% of the items sold in the health and wellness section were UP and reported a high presence of reduced and increased nutrient content claims, suggesting that supermarkets' circulars often promote the sale of UP foods.

Besides the study of consumers' perception of FOP, it is worth investigating the predictors of consumer interest in FOP and back-of-pack labels. This was the object of a Polish study [26] which found that self-rated knowledge about nutrition healthiness is the only significant predictor in over 1000 Polish consumers, while neither demographic nor socioeconomic variables were significant predictors of interest towards food labels. Plasek and coworkers [27] focused on six categories of actors that seem to influence the perceived healthiness of foods: (i) the communication information (such as FOPs and NHCs), (ii) the product category, (iii) the shape and color of the product packaging, (iv) the ingredients of the product, (v) the organic origin of the product, and (vi) the sensory characteristics of food. Bryla [28] also found that FOP label reading is one of the predictors of the importance linked to salt content in over 1000 Polish consumers, in addition to other predictors such as the importance and attention to NHCs and the respondent's age.

Two studies applied hypothetical discrete choice experiments to analyzing consumers' choices and willingness to pay (WTP) for, respectively, fish products [29] and pork sausages [30]. Menozzi and colleagues interviewed 2500 fish consumers in five European countries to assess the relative importance and WTP for different fish species and labeled attributes (i.e., sustainability label, NHCs, product presentation, production system, and price). The findings showed positive premiums for sustainability label, NHCs, and wild-caught alternatives, with high heterogeneity across countries and species [29]. Czine et al. [30] investigated whether product characteristics indicated on food labels of sausage made from traditional Hungarian mangalica pork might influence consumers' choices. The authors found respondents' preference for the label of origin indicating meat from registered animals, and purchasing from the farmers' market is preferred over the butcher and hyper-/supermarket.

Country-of-origin (COO) labeling effects were analyzed by Bimbo et al. [31]. The authors tested the price differential associated with the COO information for extra-virgin olive oil (EVOO) in Italy, employing a hedonic price model on the purchase of EVOO products collected from 982 consumers at the supermarket checkouts. Although the mandatory COO labeling regulation for EVOO can be an effective tool for consumers to identify the origin of the product and for producers to differentiate products, the results evidenced a significant share of consumers unable to correctly identify the origin of the EVOO purchased, mostly among consumers who reported having purchased Italian EVOO.

Two experiments were conducted to analyze the effects of visual aids and color nutrition information (CNI) on sugar-sweetened beverages [32] and sweet food consumption [33]. Merillat et al. [32] assessed the effects of visual aids on judgments of sugar quantity in popular drinks and the choices of 261 individuals recruited in the USA. In the experimental condition, participants viewed beverages along with test tubes filled with the total amount of sugar in each drink and this led to a lower intention to consume any of the beverages, suggesting that this simple visual aid intervention affected judgments and choices towards curtailing sugar intake. Using an eye-tracking technique, Potthoff et al. [33] evaluated the effect of CNI based on a traffic light system adopted in Austria; participants in this study viewed images depicting sweets preceded by a colored circle informing about the sugar content of the food, with and without nutrition information. The results showed that the intervention had the opposite of the intended effect and the authors questioned whether CNI is helpful to influence initial cue reactivity toward sweet foods.

A quasi-experimental online trial on the choice of sugar foods was performed by Chen et al. [34] in Taiwan. The authors analyzed how mothers' choices of low-sugar food were affected by theory-driven nutrition interventions, finding that, after the intervention, they exhibited enhanced sugar and nutrition label knowledge, perceived behavioral control, behavioral intentions, and behavior.

Another experiment was conducted by Modlinska et al. [35] with 99 Polish individuals to assess the influence of food labeling (insect content) and appearance (traces of insect-like ingredients) on the participants' perception. The results showed that products labeled as containing insects are consumed with reluctance and in lower quantities despite their appearance, regardless of the form in which the insects are served. The authors provided recommendations for labeling strategies to help to reduce the effect of disgust.

As already mentioned, food labeling does not include only nutritional information, and this is why a series of papers focused on other aspects is included. For instance, Ontiveros and colleagues [36] focused on allergens, by evaluating the characteristics of food allergen labeling and precautionary allergen labeling (PAL) in over 10,000 products sold in six Latin American countries. The authors found a high (>87.4%) compliance with local regulations, but countries without specific regulations for allergen labeling had two-fold more products containing allergens in their ingredients lists but no food

allergen labeling, compared to countries with regulations. These results suggest that the lack of regulations for the characteristics of allergen labeling increases the risk of accidental exposure to allergens of interest.

Another interesting topic was reviewed by Van der Horst and coworkers [37], who investigated how healthy adults perceive and interpret serving size information on food packages and its influence on product perception and consumption. In their systematic review, the authors observed an overall poor conception of serving size, while the few included studies showed that labeled serving size affects portion size selection and consumption.

Finally, Rincón-Gallardo Patiño et al. [38] investigated restaurant menu labeling policies and their effects on menu reformulation. The authors found three voluntary and eight mandatory menu labeling policies primarily for energy disclosures, developed in uppermiddle- and high-income countries, whereas none was found in low- or middle-income countries. The subsequent analysis conducted by the authors showed reductions in energy for newly introduced menu items only in the US. Implications for policy, practice, and research are also provided.

Overall, the studies included in the Special Issue provide new insights in this field of research, with relevant recommendations for policy makers, business operators, and researchers for developing more effective labeling strategies, allowing consumers to make informed dietary choices. At the same time, many authors reported the need for performing further investigations to confirm and expand current findings.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Article* **Nutritional Quality of Pasta Sold on the Italian Market: The Food Labelling of Italian Products (FLIP) Study**

**Marika Dello Russo 1,†, Carmela Spagnuolo 1,† , Stefania Moccia 1,†, Donato Angelino <sup>2</sup> , Nicoletta Pellegrini 3,\* , Daniela Martini <sup>4</sup> and on behalf of the Italian Society of Human Nutrition (SINU) Young Working Group ‡**


**Abstract:** Pasta represents a staple food in many populations and, in recent years, an increasing number of pasta items has been placed on the market to satisfy needs and trends. The aims of this work were: (i) to investigate the nutritional composition of the different types of pasta currently sold in Italy by collecting the nutrition facts on their packaging; (ii) to compare energy, nutrient and salt content per 100 g and serving in fresh and dried pasta; (iii) to compare the nutrition declaration in pairs of products with and without different declarations (i.e., gluten free (GF), organic, and nutrition claims (NC)). A total of 756 items, made available by 13 retailers present on the Italian market, were included in the analysis. Data showed a wide difference between dried and fresh pasta, with high inter-type variability. A negligible amount of salt was observed in all types of pasta, except for stuffed products, which had a median high quantity of salt (>1 g/100 g and ~1.5 g/serving). Organic pasta had higher fibre and lower protein contents compared to conventional pasta. GF products were higher in carbohydrate and fat but lower in fibre and protein than not-GF products, while only a higher fibre content was found in pasta with NC compared to products not boasting claims. Overall, the results show high variability in terms of nutrition composition among the pasta items currently on the market, supporting the importance of reading and understanding food labels for making informed food choices.

**Keywords:** pasta; food labelling; nutrition declaration; nutritional composition; gluten free; nutrition claims

#### **1. Introduction**

Pasta is one of the most widespread staple foods, known at least since the time of the Etruscans, who learned how to work the wheat by grinding it, mixing it with water, levelling it in thin doughs, and cooking it on a red-hot stone. According to Italian law [1], "dried pasta" must be made with water and durum wheat semolina, while "fresh pasta" can be made with soft wheat and has a higher moisture content than "dried pasta". There are also laws regarding the preparation of "special pasta", which contains other ingredients than wheat and water: "egg pasta", manufactured with durum wheat and hen's eggs, and "stuffed pasta", which includes "fresh pasta" filled with different ingredients, as in the case of ravioli or lasagne. In this scenario, the manufacturing process of pasta is continuously updated over the years to face food needs and trends. For example, a wide range of pasta

**Citation:** Dello Russo, M.; Spagnuolo, C.; Moccia, S.; Angelino, D.; Pellegrini, N.; Martini, D.; Nutritional Quality of Pasta Sold on the Italian Market: The Food Labelling of Italian Products (FLIP) Study. *Nutrients* **2021**, *13*, 171. https://doi.org/10.3390/nu13010171

Received: 28 October 2020 Accepted: 4 January 2021 Published: 8 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

containing different ingredients, such as vegetable extracts, i.e., spinach and tomato, is currently available on the market.

A survey carried out by the International Pasta Organisation in 2014 reported that about 14.3 million tons of pasta are annually produced worldwide, mainly in Italy, the United States, Brazil, Turkey and Russia. Italians are the main pasta consumers, with 25.3 kg per capita per year, followed by Tunisians (16 kg), Venezuelans (12.2 kg) and Greeks (11.5 kg) [2]. Pasta consumption has faced a decrease in the last years, probably because of, among other reasons, the myth about dodging carbohydrate-rich foods as a "strategy" for losing weight [3]. However, it is worth noting that pasta is a key component of many healthy eating patterns, above all the Mediterranean Diet, and its consumption has been positively associated with a low body mass index and prevention of overweight and obesity risk conditions [4–8]. Among the possible reasons explaining the positive effects it has on health, pasta generally has a good nutritional quality, due to low amounts of fat and available carbohydrates. Moreover, pasta can be also a suitable vehicle for the incorporation of beneficial components, such as fibre or probiotics [9,10], considering its low cost, long shelf life, and wide range of acceptability in many consumers groups [11]. However, pasta is a very heterogeneous category, including several types of products which often differ not only in shape but most importantly in ingredients, and thus, in nutrition composition. Nevertheless, dietary recommendations do not take into account this variability for suggestions in portion sizes and frequency of consumption. Moreover, it should be carefully considered that pasta is often consumed in association with other ingredients, i.e., oil and/or grated cheese, or elaborated sauces, which can largely contribute to the energy value of the entire serving. Regardless of this aspect, the knowledge of the nutritional quality of pasta itself can be useful for evaluating the nutritional characteristics of the dish and helping consumers in their purchase. However, an overview of the nutritional quality of all the products named as "pasta" on the label is still missing in the literature.

In Europe, the nutrition information of pasta, and generally of pre-packed foods, is available to the consumer on the food label in accordance with Council Regulation No 1169/2011 [12], together with other mandatory information (i.e., list of ingredients, net amount, and name of producer). Moreover, other voluntary information can be reported on the pack, including nutrition claims (NC) and health claims (HC), the reduced presence or absence of gluten, or organic certification [12–14]. Research has shown that reading and understanding the nutrition facts and the claims reported on the label can help consumers in making healthy food choices [15]. However, it has been evidenced that the presence of NC or HC or the absence of gluten on the label can be misperceived as a guarantee of a better nutritional quality of the product [16–19]. Thus, consumers should be guided towards more informed and conscious food choices, which may lead to better dietary behaviours. In this context, the Food Labelling of Italian Products (FLIP) Study was conceived to systematically investigate the overall quality of the pre-packed foods of the most important food groups sold on the Italian market by collecting the nutrition declaration on their packaging [20,21]. The present study specifically focuses on pasta: there is, indeed, a wide variety of different pasta products sold on the shelves, and many of them boast nutrition claims and other information. A comparison of energy, nutrient, and salt content in pairs of products with and without the different declarations considered in the study (i.e., gluten free (GF), organic and their counterparts, as well as NC) was performed. Moreover, even if pasta is commonly consumed accompanied by other ingredients and sauces, the nutritional characteristics of pasta itself might be different among the various types, which may influence the nutritional quality of the entire meal. Based on these premises, it is important to investigate the nutritional composition of pasta sold in the market, also considering that the serving size can be widely different among the different types of pasta. To the best of our knowledge, no study has been carried out yet to comprehensively and systematically investigate the nutritional composition of the different varieties of pasta sold in Italy.

#### **2. Materials and Methods**

#### *2.1. Food Product Selection*

In the present study, information about pasta products was taken from the online shopping website of the main retailers present on the Italian market, as reported in a previous paper [21]. The online research was performed from July 2018 until March 2019.

All the prepacked pasta items with mandatory food information on the package, as requested by the Council Regulation (EC) no. 1169/2011 [12], were included. Conversely, the following products were excluded: not pre-packed, not available online during the collection data phase, with partial package images and/or unclear nutrition declaration, and/or an incomplete list of the ingredients. Pasta items were divided in two categories (fresh and dry), and for both, four types of pasta were selected and analysed: semolina, egg, stuffed, and special pasta.

#### *2.2. Data Extraction and Analysis*

For all the selected products, data from the complete images of the package were collected. The quali-quantitative data reported on the label of all products were recorded, including: company name, brand name, descriptive name, energy (kcal/100 g), total fat (g/100 g), saturated fatty acids (SFA, g/100 g), total carbohydrates (g/100 g), sugars (g/100 g), protein (g/100 g), and salt (g/100 g). In addition to the mandatory nutrition information indicated in the Council Regulation (EC) 1169/2011 [12], fibre content (g/100 g) was also collected. Descriptive name was used to classify the retrieved items in the two categories (fresh and dried pasta), each one including four types (semolina, egg, special, and stuffed pasta).

Once these values were retrieved, data of the energy and nutrient contents were also presented per standard serving by using the Italian suggested serving sizes for pasta [22]. The considered standard serving was 80 g for all types of dried pasta, 100 g for fresh semolina and egg pasta, and 125 g for fresh stuffed pasta.

Finally, information on the presence or absence of organic certification, GF declaration, and NC, was collected. Taking into account the disparity in the number of products with and without organic, GF, and NC declarations, the comparison of energy, nutrient, and salt content per 100 g in products with and without declarations was performed on only pairs of products, with three independent selections for each declaration. For each of the three declarations (organic, GF, NC), a similar item without declaration from the same brand was selected for each product by considering the category of pasta (i.e., fresh or dried) and type (i.e., semolina, egg, special, stuffed). For instance, for each fresh egg pasta GF, a corresponding fresh egg pasta item not GF from the same brand was chosen. When no items from the same brand but without the declaration were available, a similar item from the same category and type but another brand was randomly selected. Regarding GF products, the selection and the choice of pairs were limited to cereal-based products, thus excluding legume-based pasta, due to the high heterogeneity in terms of nutrition contents between these two types of products.

The precision of the extracted data was independently double-checked by two researchers (M.D.R. and S.M.), and inaccuracies were solved through secondary extractions by a third researcher (C.S.). After data collection, a dataset was created, grouping products into the four categories of interest: semolina pasta, egg pasta, stuffed pasta, and special pasta.

For the analysis of salt content, products were classified as "very low salt content" if they had <0.12 g of salt/100 g and "low salt content" if products had <0.3 g of salt/100 g, following the indications by Regulation (EC) No 1924/2006 [13]. The remaining products were instead classified as "medium salt content" (>0.3 but < 1 g of salt/100 g) and "high salt content" (≥ 1 g of salt/100 g) as reported by the Italian Society of Human Nutrition (www.sinu.it) in the dissemination materials produced for the "World salt awareness week" performed in the framework of the World Action on Salt and Health (http://www. worldactiononsalt.com/).

#### *2.3. Statistical Analysis*

The Statistical Package for Social Sciences software (IBM SPSS Statistics, Version 24.0, IBM corp., Chicago, IL) was used to perform the statistical analysis, with a significance level set at *p* < 0.05. The normality of data distribution was firstly verified through the Kolmogorov-Smirnov test and rejected. Therefore, variables were expressed as median and interquartile range. Differences in terms of energy, macronutrients, and salt content among different types of pasta were explored using the Kruskal-Wallis test for independent samples with multiple pairwise comparisons. Analysis per 100 g was performed separately within and not between the two pasta categories (fresh and dried pasta), due to the difference in their moisture content. The analysis per serving was instead performed among all different types of pasta. The Mann-Whitney non-parametric test for two independent samples was applied for the comparisons of organic, GF, or NC pasta with their relative counterparts.

#### **3. Results**

#### *3.1. Number and Types of Products*

A total of 756 different items were analysed, categorised in fresh pasta (*n* = 269) and dried pasta (*n* = 487) based on their legal name.

Moreover, according to the nutritional characteristics of the different types of pasta, products within each category were further grouped into four pasta types (i.e., semolina, egg, stuffed, and special pasta that include other ingredients, such as rice, quinoa, amaranth, and legumes). Among fresh pasta, stuffed pasta had the largest number of items (*n* = 208), followed by egg pasta (*n* = 45) and semolina pasta (*n* = 16), and only one special pasta item, which was, thus, not considered in the following analysis. Conversely, egg pasta prevailed among dried pasta (*n* = 206), followed by semolina (*n* = 157) and special pasta (*n* = 119), while only five stuffed pasta items were found.

To compare the nutritional composition between products with and without specific declarations, the analysis was carried out on 49 pairs of organic/conventional items, 90 GF/gluten-containing products. Finally, 45 products with at least one NC declaration and 45 without NC were considered.

#### *3.2. Nutritional Composition of Pasta per 100 g*

As reported in Table 1, there was a high variability of energy and nutrient contents among fresh and dried pasta (*p* < 0.05) when different pasta types, within each category, were analysed. Considering fresh pasta, egg pasta had total median energy of 293 (288–308) kcal/100 g, which was slightly but significantly higher than stuffed pasta, which had total median energy of 274 (248–293) kcal/100 g, and semolina pasta, which had total median energy of 272 (261–273) kcal/100 g. Different results were found for energy median values of dried pasta types, where no differences were observed between egg and stuffed pasta (369 (365–374) kcal/100 g and 394 (393–403) kcal/100 g, respectively), even though they had significantly higher energy values than special and semolina pasta (351 (347–358) kcal/100 g and 354 (351–357) kcal/100 g, respectively). Overall, carbohydrates were the most abundant macronutrients, ranging from 54% of energy in stuffed fresh pasta to 71% of energy in semolina dried pasta.

In the fresh pasta category, stuffed pasta showed a significantly higher content of total fat, SFA, sugar, and salt, but a significantly lower amount of carbohydrate, compared to the other pasta types. Among the dried pasta, semolina and special pasta had a significantly lower total fat, SFA, and protein content and a greater amount of carbohydrates compared to egg- and stuffed-pasta (Table 1).

Regarding salt content, only stuffed pasta, both fresh and dried, had a median high quantity of salt (>1 g/100 g). Considering all the items (i.e., fresh and dried), low content of salt (<0.3 g/100 g) was reported by 68.1% of the products, while a medium (>0.3 g/100 g but <1 g/100 g) and high (≥1 g/100 g) category of salt content was reported for 10.9% and 21.1%, respectively. As shown in Figure 1, the type of pasta with the highest salt content was stuffed

pasta. Indeed, 98.6% of stuffed products had a medium or high salt content. Conversely, no special pasta items and only 0.6% of semolina products had a high salt content.


**Table 1.** Energy and nutritional composition across categories of pasta.

Values are expressed as median (25th–75th percentile). SFA: saturated fatty acid; Carb: carbohydrate. For each category, different superscript lowercase letters in the same row indicate significant differences among types (Kruskal–Wallis test for independent samples with multiple pairwise comparisons). # Number of items reporting fibre: Fresh pasta (All *n* = 198; Semolina *n* = 15; Egg *n* = 34; Stuffed *n* = 149), Dried pasta (All *n* = 449; Semolina *n* = 135; Egg *n* = 193; Stuffed *n* = 5; Special *n* = 116).

**Figure 1.** Salt content of the pasta products. Coloured dots refer to the classification for salt content by Council Regulation (EC) No 1924/2006 [13]: light green = very low salt content (<0.12 g/100 g); green = low salt content (<0.3 g/100 g); yellow = medium salt content (<1 g/100 g); red = high salt content (≥1 g/100 g).

Table 2 shows the nutritional content of organic, GF, and NC pasta and their respective counterparts. No differences were observed when considering organic declaration, except

for a lower protein content (12.0 (11.0–14.0) vs. 13.7 (13.0–14.6) g/100 g) and higher fibre content (3.0 (2.8–3.4) vs. 2.8 (2.7–3.0) g/100 g) in organic pasta compared to conventional pasta. GF products showed a significantly higher content of total carbohydrate and fat and a lower content of sugar, fibre and protein compared to the non-containing gluten counterpart (Table 2). No differences were identified when products with NC were compared to their counterpart, except for fibre content, which was significantly higher in pasta with NC.

**Table 2.** Energy and nutrition facts in products with and without specific declarations, on selected pairs of products (as reported in Materials and Methods section).


Values are expressed as median (25th–75th percentile). SFA: saturated fatty acid; GF: gluten free; NC: nutrition claim. For each category, asterisks indicate significant differences between groups (Mann-Whitney non-parametric test for two independent samples), *<sup>p</sup>* < 0.05. # Number of items reporting fibre: organic = 45/43; GF = 82/74; NC = 45/43 (yes/no).

#### *3.3. Nutritional Composition per Serving Size*

In order to better investigate the nutrition content of all pasta types, a further evaluation of nutrition facts (energy, nutrients, and salt) per standard serving was performed (Figure 2). This analysis was performed because the different types of pasta can be considered as alternatives as indicated in the Reference Intakes of nutrients and energy for the Italian population [22]. The analysis was carried out without dried stuffed pasta, because only five items were present. Regarding the energy content per standard serving, for fresh stuffed pasta (342 (310–366) kcal/serving) it was significantly higher than for the other pasta types, while no statistical differences were observed among fresh semolina (272 (258–273) kcal/serving), dried semolina (283 (281–286) kcal/serving), and dried special pasta (281 (278–286) kcal/serving), and between fresh egg (293 (288–308) kcal/serving) and dried egg pasta (295 (292–299) kcal/serving). The same trend was observed for total fats. Regarding the carbohydrate content per standard serving, fresh stuffed pasta and dried special pasta showed the greatest variability. In fact, the median carbohydrate content per standard serving of fresh stuffed pasta was significantly lower than other pasta types, except for fresh semolina pasta; moreover, dried semolina pasta had a median content similar to dried special pasta but significantly higher compared to all the other pasta types. Fibre content of dried special pasta had great variability and its median content did not show significant differences from fresh and dried semolina pasta. The median quantity of protein per serving was significantly higher in stuffed pasta compared to the others, except for dried egg pasta. Considering the salt content per standard serving, only fresh stuffed pasta had a median high quantity of salt (1.5 g/serving).

**Figure 2.** Box plot for energy and nutrition facts per standard serving across categories of pasta. Standard serving was 80 g for all types of dried pasta, 100 g for fresh semolina and egg pasta, and 125 g for fresh stuffed pasta. For each category, different letters indicate significant differences among pasta types (Kruskal–Wallis test for independent samples with multiple pairwise comparisons). fresh semolina (*n* = 16), fresh egg (*n* = 45), fresh stuffed (*n* = 208), dried special (*n* = 119), dried semolina (*n* = 157), dried egg (*n* = 206).

#### **4. Discussion**

To the best of our knowledge, the present study evaluated, for the first time, the nutritional quality of the different pasta products sold on the Italian market, taking into account the mandatory and some voluntary nutrition information printed on the packaging.

The first intriguing finding is related to the high number of items retrieved on the market. On the one hand, this number confirms Italians as the number one producers and consumers of pasta. It is worth remembering that pasta is indeed a widely common staple

food and it is a key product in the Mediterranean dietary pattern [23]—to such an extent that the Italian Food Dietary Guidelines suggests the consumption of one serving of pasta per day (or rice or other cereals) [24]. On the other hand, the several different types of pasta found on the market confirm the great interest of food companies in satisfying the emerging needs and trends of the customers, not only in terms of format but mostly for the many ingredients that can be used for pasta-making.

Data of the nutrition facts evidenced wide differences in terms of energy and nutrients across the two different pasta categories under study, i.e., fresh and dried pasta. Differences were also found among the types of pasta, particularly for stuffed pasta, showing a lower carbohydrate and a higher fat, sugar, and energy content with respect to the other pasta types. Such a difference observed for stuffed pasta is probably mainly due to the filling, of which the weight usually represents half of the total weight. For the same reason, this type of pasta was characterised by the highest median content of salt, with almost all stuffed products (98.6%) having a medium or high salt content. Conversely, the results indicated a negligible amount of salt, the least amount of all pasta types. This is why semolina pasta and other types of pasta are generally prepared at home as well as in the restaurant, canteens, etc. by adding salt in the boiling water; however, the final salt content is likely lower than those reported on the food label of stuffed pasta [25,26].

These results are even more evident by analysing the nutrition declaration per serving size instead of 100 g, mainly because the reference serving for stuffed pasta is generally higher than the ones for dried and fresh semolina and egg pasta [22]. This led to an increased differentiation between the nutritional quality of the different types of pasta, for instance with fresh stuffed pasta providing the highest energy per portion despite no differences were found for 100 g.

Regarding salt, values per serving confirmed the ones obtained per 100 g, highlighting that a portion of stuffed pasta contributes to a high extent to the daily intake of salt, on average equivalent to about 33% of the 5 g/day indicated as the goal by the World Health Organisation (WHO) [27] and over 2 g/serving size for 10% of the stuffed items. It is worth remembering that an excessive consumption of salt in a diet increases blood pressure and consequently the risk of adverse effects on cardiovascular health [28,29]. These results highlight, on the one hand, the importance of nutritional education and increasing knowledge in the population in taking into consideration the serving size, which can deeply influence the nutrient intake; on the other hand, the results suggest that not only food per se, but also the preparation of foods (e.g., adding salt to boiling water), has a key role in the daily intake of nutrients. For example, the high salt content in stuffed pasta suggests that the addition of salt to boiling water should be avoided, even though this is not usually reported on the pack; in some cases, it is even suggested to add salt.

To investigate the nutrition quality of the different categories and types of pasta, we also compared the nutrition facts of products with and without three different declarations, i.e., GF, organic, and NC. This aspect was taken into consideration because consumers' perception may be influenced by several types of declarations, with the so-called "halo effect" [30]. It has, indeed, been evidenced that consumers perceive foods with claims (e.g., NC) or specific front-of-pack labelling [31], as well as GF products [32] and organic foods [33], healthier than their counterparts. Thus, this misperception may influence food habits, which, in turn, may in some cases lead to an overconsumption [34]. To investigate whether the presence of declaration may affect the nutrition quality of the pasta, we selected and compared the nutrition declaration in an equal number of products with and without each of these declarations on the pack, similarly to what already done in a previous study aimed at comparing the nutritional quality of organic vs. conventional products [20]. This choice was due to the gap of product numbers with and without these declarations currently on the market. The first comparison was made between GF and gluten-containing pasta. Our results suggest that GF products had higher carbohydrates and fat contents and lower fibre, sugar, and protein contents compared to the gluten-containing products. It is worth noting that the use of legume flours/ingredients is increasing in order to

enhance the nutritional profile of GF products, resulting in significantly higher fibre and protein contents and a lower amount of carbohydrates compared to conventional semolina pasta [35,36]. However, these types of pasta have not been taken into consideration, as there were no gluten-containing counterparts for the comparison. Our results confirm the findings of an Italian survey considering GF pasta sold on the Italian market, although the authors also found a higher energy, SFA, and salt amount in GF pasta compared to the regular ones [37]. Moreover, our results are partially in agreement with the ones found by a Spanish research considering a total of 53 pasta items, 15 of which were GF [38]. The authors confirmed higher total fats and lower sugars for GF pasta, but they also found higher protein and SFA than gluten-containing pasta. Contrasting results in terms of lipid contents have been found in a UK study focusing on both GF and regular whole grain and white pasta [39]. In fact, UK white GF pasta (111 items) showed a lower content of total fat and SFA than the counterparts (96 items), whereas these data were not confirmed for the whole grain items [39].

Regarding the nutritional profile of products boasting NC, only fibre content was significantly higher than in products without claims. This is plausibly due to the fact the almost all NCs found in these products were related to fibre (i.e., 38 items "source or rich in fibre"), while only 7 items boasted a NC claim related to protein. Intriguingly, a survey conducted on 87 pasta and rice items sold on the Irish market found that 31% of the products considered boasted a NC or a HC and that most of the NC referred to fat (including saturated fats) and carbohydrates, followed by sugars and protein [40]. Overall, data from the present study support the hypothesis that NC should not be considered as marker of the overall quality of food products, as already indicated in previous studies on different types of products [21]. This suggests that more effort should be made in nutrition education to avoid misperceptions, which lead to inappropriate food choices and possibly overconsumption [41].

Finally, we also compared the nutritional quality of organic and conventional pasta. In agreement with studies comparing conventional and organic durum wheat products [42–44], our data only showed a significantly lower protein content and a higher fibre amount in organic pasta with respect to the conventional counterpart. Even though no other significant difference between organic and conventional products was found, these variations were not due to the co-existence of other characteristics, such as a different number of wholemeal in the organic vs. the conventional counterparts, as none of the 49 paired items were wholemeal. These results are in line with a previous publication, where it was highlighted that the organic certification cannot be intended as a marker of the general nutritional quality of the products [20].

Our work showed strengths and limitations, mainly attributable to the methodology used for product selection. On the one hand, we analysed for the first time the nutritional composition of a high number of different pasta products retrieved from the major retailers present on the Italian market that have a home-shopping website, thus including the majority of pasta sold in Italy. On the other hand, the exclusion of products sold by local groceries and discounts as well as shops dedicated to special foods, i.e., GF items, might have limited the product analysis. Another limitation of the study concerns the different origin of the nutritional data on the label, which could be based on laboratory analysis or calculation from the ingredients used or generally established and accepted data, creating a putative bias in data origin. Moreover, the high variability of filling characteristics found among stuffed pasta items made tricky the intra- and inter-type comparisons of the nutritional quality. Finally, the comparison of the nutritional quality between the pairs of products with and without declarations could be considered a limitation. However, including all the 756 items in the analysis would have affected the findings because of the large difference in the number of products of the same type, with and without declarations such as GF, organic, or NC. Conversely, the comparison of items of the same brand was a way to avoid that the brand name can act as a possible cause of bias.

#### **5. Conclusions**

To the best of our knowledge, this is the first study which comprehensively analysed the nutritional composition of a wide range of fresh and dried pasta products sold in the Italian market. Data showed that pasta types currently on the market are very different in terms of nutrition profile, and not really comparable. Particularly stuffed pasta was characterised by a high salt content, representing a large proportion of the maximum of 5 g/day indicated by the WHO. This last aspect particularly highlights the need to clarify as much as possible the nutrition facts of product to the consumer, as pasta is usually eaten by adding sauces and/or toppings which might further increase the energy, macronutrient, and salt intake. Linked to this, we also advised that salt, mainly the discretionary one, should be carefully reduced or avoided in cooking this type of pasta. It is, indeed, particularly crucial to increase consumer awareness about the choice of both adequate pasta type and dressing and their contribution to the nutritional quality of the entire dishes.

Overall, findings from the present study are particularly of interest and should be taken into account in dietary recommendations, which currently provide only information regarding the serving size of the different type of pasta, but not about their frequency of consumption. Stuffed pasta probably should not be regularly consumed as an alternative to semolina pasta. Thus, the awareness of the consumers about the nutrition profile of the different types of pasta could be the topic of targeted nutrition education interventions aimed to improve their knowledge and, in turn, their food habits. Finally, with this study focusing on pasta products, we confirm that organic or other declarations, i.e., NC, cannot be an overall marker of the nutritional quality of the product and, thus, this topic should also be the object of future nutrition education targeted to consumers.

**Author Contributions:** M.D.R., C.S., and S.M. were involved in the protocol design, data analyses, interpretation of results, and drafting of the manuscript; D.A. participated in the protocol design, data analysis, and drafting of the manuscript; N.P. participated in the protocol design and critically reviewed the manuscript; D.M. conceived the study, supervised the data collection, and had primary responsibility for the final content. Other members of the Italian Society of Human Nutrition (SINU) Young Working Group were involved in the protocol design. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors wish to thank all students who participated to the development of the dataset.

#### **SINU Young Working Group**


**Conflicts of Interest:** The present publication has been conceived within the Italian Society of Human Nutrition (SINU) Young Group, and it has been made without any funding from food industries or other entities. The authors declare no conflict of interest.

#### **References**


### *Article* **Supermarket Circulars Promoting the Sales of 'Healthy' Foods: Analysis Based on Degree of Processing**

**Alyne Michelle Botelho 1, Anice Milbratz de Camargo <sup>1</sup> , Kharla Janinny Medeiros 1, Gabriella Beatriz Irmão 1, Moira Dean <sup>2</sup> and Giovanna Medeiros Rataichesck Fiates 1,\***


Received: 27 July 2020; Accepted: 27 August 2020; Published: 21 September 2020

**Abstract:** The health and wellness food sector grew 98% from 2009 to 2014 in Brazil, the world's fourth-biggest market. The trend has reached supermarket circulars, which recently started to feature whole sections advertising health and wellness-enhancing foods. This study identified food items advertised in circulars' specific sections of two Brazilian supermarket chains (one regional, one national) during a 10-week period. Foods were classified according to degree of food processing and presence/type of claims on their front-of-pack (FoP) labels. Comparison between groups of Unprocessed/Minimally Processed foods vs. Ultra-processed foods and presence/type of claims employed Pearson chi-square test. From the 434 alleged health and wellness-enhancing foods advertised, around half (51.4%) were classified as Ultra-processed. Presence of reduced and increased nutrient-content claims was significantly higher in labels of Ultra-processed foods. Most frequent claims addressed sugar and fibre content. Brazilian supermarket circulars were found to be promoting the sale of Ultra-processed foods in their health and wellness sections, leading to a situation that can mislead the consumer and bring negative health outcomes.

**Keywords:** supermarket circulars; ultra-processed; food label; health claims

#### **1. Introduction**

Healthy eating is essential for health promotion and protection, and as a determinant factor in preventing chronic non-transmissible diseases [1]. Nevertheless, the access to a healthy and adequate diet is proving to be a challenge for modern societies, with the eating practices of Brazilians in different stages of life and across all socioeconomic strata being far from what is considered desirable [2]. Consumption of processed and ultra-processed foods has been growing exponentially in the Brazilian population and is considered a contributing factor for the increased prevalence of obesity and non-communicable diseases in the country [3,4]. Consequently, a new version of the Dietary Guidelines for the Brazilian Population was published, instructing individuals to limit the consumption of processed foods, avoid consumption of ultra-processed foods, and choose fresh and minimally processed foods as the core of their diets. According to the Guidelines, a healthy diet is based on the consumption of natural or minimally processed foods; and of dishes and meals containing such foods [1].

Published in 2015, the Dietary Guidelines for the Brazilian Population established specific eating directives based on degree of food processing, employing a classification system later improved and published under the name of NOVA (a name, not an acronym) [5]. According to NOVA, industrial formulations containing little or no fresh ingredients and food additives to add colour, flavour, texture, and additional sensory properties to unprocessed foods and preparations containing them are classified as Ultra-processed foods [1,5]. The nature of the processes and ingredients used in their manufacture, and their displacement of unprocessed or minimally processed foods and freshly prepared dishes and meals, make ultra-processed foods intrinsically unhealthy. In spite of this, as foods typically energy-dense, rich in sugar, fat, and salt, they are hyper-palatable and cheap, which contributes to their high consumption. Ultra-processed foods are poor in dietary fibre, protein, vitamins, and minerals; and additives contained in their formulation increase shelf-life without increasing their cost [1,5,6].

On the other hand, dietary patterns based on dishes and meals made from a variety of unprocessed or minimally processed plant foods, prepared, seasoned, and cooked with processed culinary ingredients and complemented with processed foods are the healthier ones [6]. Examples of ultra-processed foods include but are not limited to cookies, fizzy drinks, confectionery items, cereal bars, bottled sauces, instant noodles, and sweetened milk-based beverages. Their intake is discouraged, while that of unprocessed or minimally processed foods is encouraged to constitute the core items of the population's diet [1]. Unprocessed (or natural) foods are edible parts of plants (seeds, fruits, leaves, stems, roots) or of animals (muscle, offal, eggs, milk), and also fungi, algae and water, after separation from nature [5]. Minimally processed foods, that together with unprocessed foods make up NOVA group 1 are unprocessed foods altered by industrial processes such as removal of inedible or unwanted parts, drying, crushing, grinding, fractionating, roasting, boiling, pasteurisation, refrigeration, freezing, placing in containers, vacuum packaging or non-alcoholic fermentation. None of these processes add salt, sugar, oils or fats, or other food substances to the original food. Their main aim is to extend the life of grains (cereals), legumes (pulses), vegetables, fruits, nuts, milk, meat and other foods, enabling their storage for longer use, and often to make their preparation easier or more diverse [6].

Despite of what is recommended by official guidelines, people have complex and diversified interpretations about the concept of healthy eating, which reflect their personal, social, cultural, and environmental experiences [7]. The concept of healthy eating is frequently unclear for individuals and is not understood and interpreted identically by all [8]. This can lead to the adoption of different practices in the name of healthy eating [9].

A definitive and universally accepted concept of healthy eating does not exist, but its association with better health and disease prevention is largely recognised [10]. During the second half of the last century, the increased availability and diversity of (un)healthy foods considerably modified the concept of what constitutes a healthy diet [11]. As the focus changed from a nutrient-based approach to a food-based one, food classification systems based on degree of processing were proposed [12]. In this context, the higher the processing degree to which a food has been submitted, the lower is the frequency in which it should be ingested as part of a healthy diet [5].

As ultra-processed foods tend to be energy-dense and low-cost, low energy cost could be one mechanism linking ultra-processed foods with high consumption and consequent negative health outcomes [13]. In Brazil however, the total cost of diets based on natural or minimally processed foods is still lower than the cost of diets based on ultra-processed foods. Relatively expensive perishable foods such as some vegetables, fruits, and fish are and should be consumed with other natural or minimally processed foods that have lower prices, such as rice, beans, potatoes, cassava, and other staple traditional Brazilian foods. Calculations based on Brazilian household budget surveys show that diets based on fresh and minimally processed foods, and dishes and meals made with these foods and culinary ingredients, are cheaper than diets made of ultra-processed foods, as well as being healthier [1].

According to Euromonitor International, Brazilian population's interest in healthy foods has increased between 2009 and 2014 [14]. The health and wellness food sector accounts for a US \$35 billion market each year and is expected to grow on average 5% per year until 2021. The 'free from' food category presents the largest growth, stimulated by the increased (but unrelated to dietetic intolerance) consumption of gluten and lactose-free foods [14].

Nutrient-content claims are regulated in Brazil as 'Complementary Nutrition Information' (CNI), defined as 'representations which affirm, suggest or imply that a product has particular nutritional properties especially, but not solely restricted to its energy, protein, fat, carbohydrate and fibre content, and also vitamin and mineral content'. CNIs may refer to absolute or relative/comparative nutrient content of food products using terms as: 'without', 'no', 'absence', 'low content', 'does not contain'; and 'presence', 'contains', 'high content', 'rich', 'source of'. Regulation on parameters for the voluntary display of CNI on front-of-pack (FOP) labels of packaged products exist since 2012 [15], but claims are allowed without consideration for foods' whole nutrient composition or degree of processing. Therefore, an ultra-processed food product containing high levels of sugar and/or sodium may display a nutrient-content claim of 'low fat' or 'vitamin rich' on its label.

Health claims are regulated by a resolution published in 1999 and amended in 2004, which determines that a 'health property claim' is one that affirms, suggests or implies a relationship between the food/ingredient and diseases or health-related conditions [16]. It may also describe a physiological role which assists normal growth, development and functions, and contributes to health maintenance and reduced risk of diet-related disease [16].

Nutrition labelling is designed to help consumers make healthier food choices, provided they understand the vocabulary or layout used to display nutritional information. For this reason, the highlight of positive characteristics in food products by means of nutrient or health claims is regarded as a marketing strategy to promote sales [17], as many are found on unhealthy food items [18]. Claims have the potential to both inform and mislead consumers, depending on the information that is highlighted and the kind of product displaying this information [19]. The highlight of positive characteristics as nutrient claims on front-of-pack labels can generate a 'health halo' effect, when consumers' assessment of a single positive characteristic of the food affects their judgment about the quality of the food as a whole [20]. Health halos can be conferred by claims concerning just one nutrient, because consumers often make generalisations about the overall health of a product based on one piece of information found on labels [21].

Notably, the 'free-from' consumption trend has influenced the content of supermarket circulars. Together with images of products' front-of-pack labels, circulars present products' prices to aid consumers to plan their shopping, but also significantly influencing their shopping decisions. National and regional supermarket chains have started to dedicate whole sections of circulars to the promotion of foods designated by them (possibly together with manufacturers), as health and wellness-enhancing [14].

'Wellness' refers to the positive, subjective state that is opposite to illness [22], an evolving process toward achieving one's full potential [23]. Wellness is positive/affirming and holistic, and encompasses lifestyle, spiritual, and environment wellbeing domains; it also accounts for the physical, mental, and social domains implied in health, and thus health is dependent on sufficient wellness [23]. As consumers today are more health conscious than ever before, and the food and beverage industry is driven by consumer demand and popular trends, the health and wellness trend is increasingly prevalent. Health and wellness foods such as energy bars, gluten and dairy-free products, products containing organic/prebiotic ingredients, and fortified/functional foods are significantly more expensive than the regular offering, which makes them an expensive luxury in many emerging markets [14].

Research on the content of supermarket circulars found that most advertised foods are unhealthy and not conducive to the adoption of a diet in line with official recommendations [24–31]. Studies reported that circular offers did not contribute towards an environment that supports healthy eating behaviour, but most only assessed the products advertised on the front pages and not in the entire circulars. Additionally, they were mostly conducted in European countries, North America,

and Australia. Only one study was identified reporting the analysis of Latin American (Brazilian) supermarkets' entire circulars [31].

The situation can mislead consumers and negatively impact their health, as supermarket circulars are reportedly used by consumers as planning tools [32,33], and can predict subsequent memory of the advertised product or brand [34]. Foods advertised in circulars are also further promoted in-store and positioned in strategic places such as end-of-aisle or islands located in places of great circulation, in order to significantly influence shopping decisions [35].

The present study seeks to extend this stream of research focusing on circulars not only advertising novel products or promotions, but dedicating sections specifically to the promotion of as health and wellness-enhancing foods.

To our knowledge, no papers about the quality of products advertised in health and wellnessenhancing sections of supermarket circulars have been published. The aim of this study was to analyse, according to degree of food processing, the quality of foods advertised in the health and wellness-enhancing sections of supermarket circulars from Brazilian supermarkets, and identify presence and type claims on the front-of-pack labels of such products.

#### **2. Materials and Methods**

This cross-sectional study was undertaken in the capital city of Santa Catarina state (Florianópolis), southern Brazil. The capital was chosen out of convenience (near the university where the research team works), and also because it is the state's second largest city. The supermarket chains for circulars' collection were defined according to the frequency of circular distribution (fortnightly) and the presence of a specific section advertising health and wellness-enhancing foods. Two supermarket chains (one national and one regional) with stores in the capital distributed circulars with the aforementioned characteristics. Both pertain to the group of 50 companies (national chain: 14th position; regional chain: 44th position) with the highest gross sales (R \$2,711,219,166.00 and R \$928,708,550.00, national and regional chain, respectively) in 2018 according to the ranking of the Brazilian Association of Supermarkets [36]. The national chain has a total of 29 stores in Brazil, of which six are located in Florianópolis; the regional chain has a total of 22 stores in the state and nine are located in Florianópolis.

To collect the circulars, one store from each chain was conveniently chosen, both located in a residential middle-class neighbourhood near two university campuses. A total of 20 circulars (10 from each chain) were collected in situ or downloaded from the supermarket website, between October 2018 and April 2019, at 15-day intervals. Nineteen printed circulars were retrieved; one circular from the regional chain was only available online. Collection was paused between December and February to avoid the influence of seasonal offers in circulars (Christmas and New Years' holiday season and summer vacation).

All images of products advertised in the health and wellness sections from retrieved circulars were analysed. Different package sizes and shapes of the same product from the same brand were counted as one (e.g., spaghetti and penne pasta). Different flavours of the same food item (e.g., grape juice and orange juice from the same brand) were counted as different items. The same happened to similar products by different brands.

Manufacturers' websites and supermarket stores were then visited to retrieve the ingredient lists of all products (except for unprocessed foods items) in order to proceed with the categorisation according to degree of processing. Foods were categorised into one of four groups as (a) unprocessed and minimally processed foods (U/MP); (b) processed culinary ingredients (PCI); (c) processed foods (P); and (d) ultra-processed foods (UP) (Table S1) [5,6]. A decision flowchart specifically developed to guide the categorisation was used [37]. Whenever the flowchart was not applicable, a conservative criterion [38] was applied (i.e., product categorised in the lower degree of processing).

Products' images on circulars' pages (examples in Figures S1 and S2) were analysed to identify the claims on the 'front-of-pack' (FoP) food labels. Illegible content was further investigated on manufacturers' websites, or in situ at the supermarket stores. Identified claims were further classified as 'Complementary Nutritional Information—CNI' or 'Additional Claims'. CNIs referred to 'reduced amount or absence' and 'increased amount or presence' of determined nutrients and energy value [39]. The terms identified and types of nutrient-content claims are presented in Table 1.

**Table 1.** Types of claims, terms, and content/nutrients identified as regulated Complementary Nutritional Information—CNI.


Reference: elaborated by the authors based on Brazilian Legislation [39].

Claims on FoP labels that did not meet the criteria to be considered CNIs (i.e., 'whole', 'organic', 'healthy', 'natural' [40], were categorised as Additional Claims—AC. In Brazil, legislation determines that the presence of lactose and gluten in foods must be reported in the package labelling section containing the product's ingredient list (usually on the back of the package) [41,42], to alert consumers who are allergic or intolerant. Therefore, the presence of such claims on FoP labels was characterised in the present study as AC.

Due to the large number of different CNIs and ACs identified on FoP labels, they were grouped into similar themes defined by two of the researchers and independently checked by a third researcher. Inconsistencies were discussed and resolved (Table S2).

Information on the description of products was organised in a Microsoft Excel® spreadsheet. Descriptive statistic was used to present data as absolute and relative frequencies, means, standard deviations (SD) or median, interquartile range (IQR) (depending on normality of distribution, assessed with Shapiro-Wilk test). Pearson chi-square test was used to compare the presence of increased and reduced nutrient-content CNIs and ACs in Unprocessed/Minimally Processed vs. Ultra-processed foods. Significance was established at *p* < 0.05. Stata version 13.0 (StataCorp, College Station, TX, USA) was used for data analysis. A post-hoc effect size analysis (w) of Pearson's chi-square tests was conducted with G\*power 3.1.9.2 considering an alpha of 0.05 [43].

Review by a Research Ethics Committee was not required for the study, as it did not involve human subjects.

#### **3. Results**

#### *3.1. Health and Wellness Food Sections' Characteristics and Degree of Processing of Advertised Food Items*

Analysis of the 20 circulars obtained from the two supermarket chains (10 for each chain) led to the identification of 434 food items, with an average number of 21.7 (5.88 SD) foods per health and wellness section. Just over half (51.4%, *n* = 223) of the foods advertised were categorised as ultra-processed, followed by unprocessed/minimally Processed foods at 32.5% (*n* = 141), P foods at 8.7% (*n* = 38), and PCI at 7.4% (*n* = 32).

The three most frequently advertised ultra-processed foods were biscuits (21.1%, *n* = 47), processed cheese (10.8%, *n* = 24), and flavoured yoghurts (8.5%, *n* = 19), followed by vegetable-based beverages, granola, popcorn, ready-to-drink tea, and breads. The three most featured unprocessed/minimally processed food products were fruit juice (15%, *n* = 21), milk (13.6%, *n* = 19), and fish (14%, *n* = 10), followed by tapioca, fresh fruits, and ground coffee.

#### *3.2. Complementary Nutritional Information—Reduced Amount*/*Absence*

A total of 48 different CNIs were identified on the products' FoP labels. From all the foods classified as unprocessed/minimally processed, and ultra-processed, 155 (43%) presented at least one CNI of reduced content or absence. Presence of this type of CNI was significantly higher (χ<sup>2</sup> = 28.67,

*p* < 0.001, effect size = 0.35) in the ultra-processed food group (77%, *n* = 120) when compared with the unprocessed/minimally processed group (23%, *n* = 35). The most frequent CNI of reduced content/absence was about sugar. Only foods from the ultra-processed group presented claims of 'reduced content or absence of saturated fat', 'reduced energy content', and 'light' (Figure 1).

**Figure 1.** Distribution of Complementary Nutritional Information about 'reduced content' or 'absence' of different nutrients in the analysed products. Legend: U/MP = Unprocessed/Minimally Processed foods. UP = Ultra-processed foods.

#### *3.3. Complementary Nutritional Information—Increased Content*/*Presence*

From the foods classified as unprocessed/minimally processed and ultra-processed, 79 (22%) products presented at least one CNI of increased content or presence of a nutrient. This type of CNI was significantly more common (χ<sup>2</sup> = 18.78, *p* < 0.001, effect size = 0.60) in the ultra-processed food group (82%, *n* = 65) when compared to the unprocessed/minimally processed group (18%, *n* = 14). The most frequent CNI was related to fibre content, which corresponded to more than 50% of the CNIs in both groups (Figure 2).

**Figure 2.** Distribution of Complementary Nutritional Information about 'increased content' or 'presence' of different nutrients in the analysed products. Legend: U/MP = Unprocessed/Minimally Processed foods. UP = Ultra-processed foods.

#### *3.4. Additional Claims*

A total of 136 claims classified as Additional Claims were identified on the FoP labels of foods advertised in the circulars. From the foods classified as unprocessed/minimally processed and ultra-processed, 269 (74%) presented at least one AC. The presence of ACs was not statistically different (χ<sup>2</sup> = 0.29, *p* = 0.590) between unprocessed/minimally processed (38%, *n* = 102) and ultra-processed (62%, *n* = 167) groups. In the ultra-processed group, the most frequent ACs were 'whole grain and fibre' (21.7%, *n* = 36), 'free-from' or 'low in lactose' (15.1%, *n* = 25), 'highlight on the presence of ingredients' (10.1%; *n* = 17), and, 'free-from gluten' or 'wheat-free' (9.8%; *n* = 16) (Figure 3).

**Figure 3.** Distribution of Additional Claims in the analysed products, classified by themes (more information available in Supplementary Materials—Table S2). Legend: U/MP = Unprocessed/Minimally Processed foods. UP = Ultra-processed foods.

#### **4. Discussion**

This study identified and categorised food products advertised in health and wellness sections of supermarket circulars according to degree of processing, and identified claims on the front-of-pack labels of such products.

Firstly, we found that the most advertised group was of ultra-processed foods. The fact that the present research was limited to the health and wellness sections of supermarket circulars means that such materials may be negatively influencing consumers' food purchases, contributing to the divergence of their diets from the recommendations of the Dietary Guidelines for the Brazilian Population [1]. In Brazilian supermarkets, as in many countries, circulars are made available to consumers online, in-store and also posted into their letter boxes with the intention to promote new products and present special offers [44]. Price affects consumers' choices [45] and most purchases influenced by promotions are unhealthy, as reported in a systematic review [46]. Advertised products are further promoted in-store and positioned in strategic places such as end-of-aisle or islands located in places of great circulation, in order to influence shopping decisions [35]. In addition, circulars are reportedly used by consumers as planning tools [32,33] and can predict subsequent memory of the advertised product or brand [34]. One Brazilian study found that ultra-processed foods were three times more frequently advertised in supermarket circulars than unprocessed/minimally processed foods [31]. Another study analysed circulars from 12 Latin-American countries and reported that in six countries unhealthy foods were more advertised than healthy foods [24]. Our findings are supported by the aforementioned studies about supermarket circulars, but by focusing on the sections dedicated to the promotion of health and wellness-enhancing foods, we extend this stream of research to the phenomenon of increase in the population's interest for healthy foods, reflected in selling strategies [14].

Another important finding was that ultra-processed foods presented significantly more CNIs than unprocessed/minimally processed foods. The most frequent CNIs in ultra-processed food packages were the increased content/presence of fibre and the reduction/absence of sugar. It is quite concerning from a public health point of view that ultra-processed foods are regarded as providers of fibre, because despite fibre content and absence of sugar, other undesirable ingredients such as hydrogenated vegetable fat, modified starch, food additives, artificial sweeteners or other formulations exclusively for industrial use which characterise these foods may also be present. The latter are responsible for increasing the palatability and shelf-life of ultra-processed foods [5,6]. Frequently, the presence of a CNI on labels of ultra-processed foods highlights a positive quality (usually a single nutrient), while negative aspects are not so visibly disclosed. In this sense, CNIs act as advertising strategies to promote sales instead of fulfilling their objective to provide more accessible information to the consumer [17,18]. Additionally, the term 'premium' is often used by the food manufacturing industry to refer to ultra-processed foods that, compared with 'regular' products, contain less 'bad' ingredients such as trans fats, sugar and salt, and more 'good' ingredients such as vitamins, minerals or whole foods such as fruits and nuts. While some of these modifications are positive, others may be harmful, as they will not make these products healthy, but mislead the consumer to think that they are [47]. The influence of nutrient claims on food products' FoP over consumers' perception has been established [48–50]. As Brazil does not have a nutrient profile system in place to evaluate the composition of foods, CNIs that highlight positive qualities of a food (e.g., high fibre content) can mislead the consumer when placed on products whose intake should be limited, such as ultra-processed foods. Brazilian researchers have already suggested that regulations on the use of nutrient claims in products bearing marketing strategies directed to children should be revised, so that only products with appropriate nutrient profiles should be allowed to display nutrient claims [51]. The same can be argued here, where foods advertised as healthy and wellness-enhancing are actually classified as ultra-processed due to their high degree of processing. These findings are cause for concern, as the consumption of ultra-processed foods is associated with higher body mass index and greater prevalence of both excess weight and obesity, as well as other non-communicable diseases [6,52].

Lastly, more than two-thirds of ultra-processed and unprocessed/minimally processed products advertised presented at least one additional claim (AC). Regarding ultra-processed foods, the most frequent ACs identified were related to whole grain, lactose and gluten content. All these claims may induce the consumer to perceive the products as healthier than they are. In spite of the information indeed being relevant to consumers with health issues, the ever-increasing interest by the industry to highlight the absence of lactose and gluten, for example, is not due to increased prevalence of celiac or lactose-intolerant individuals. What informs this strategy is the recent trend in the health and wellness sector to associate the free-from gluten and lactose products with a healthier lifestyle [53,54]. The study by Hartmann et al., (2018) evaluated the effect of free-from labels (including gluten and lactose) on consumers' perception in four European countries, and discovered that consumers evaluated such products in a simplistic way and considered the ones labelled as free-from to be healthier [54]. In another recent study, individuals perceived products labelled as gluten-free as less energy-dense and less processed than similar ones containing gluten [55]. In respect to whole grain products, research revealed that consumers perceive whole grain foods as healthier, and more nutritionally balanced and natural than the refined ones [56] and although the innovative added-fibre to refined grain products may present a solution to increase fibre intakes [57], the mere addition of soluble or insoluble fibre to a product turns it into a ultra-processed food [6]. The ample use of health-related information on food labels and in the media can confuse the consumer [58] and trigger halo effects, in which a consumer thinks a product is healthier than it actually is [59,60] and lead to indulgence when eating them [61,62].

Educational level appears to influence the perception of free-from gluten and lactose products [63], as well as the concern with healthy eating and the practice of unhealthy weight loss strategies [21]. Up to today, no evidence of benefits for healthy individuals to avoid gluten has been obtained, but the idea is so common among certain social strata that the Brazilian Society for Food and Nutrition published a position paper stating that there is no evidence to support the fact that a gluten-free diet would be beneficial for a healthy individual [64]. The benefits attributed by healthy individuals to a lactose-free diet also remain unproven [54].

This study has some limitations. We were able to analyse circulars from only two supermarket chains in the city where the study was conducted, but this was due to the fact that only these two chains featured the health and wellness-enhancing section in their circulars. Additionally, there may be slight differences between circulars from distinct stores from the same chain, but as we collected circulars from only one store from each supermarket chain, we were not able to point those differences out. Regarding the higher presence of CNI of reduced content/absence and of increased content/presence of a nutrient in ultra-processed foods, the results were significant, but with small (0.35) and medium (0.60) effect sizes [65], respectively. Interpretation should be cautious, but amplifying data collection time would most likely not improve size effect as circulars were very similar to each other throughout the months. As we analysed claims present only on the FoP (and not on the other sides of packages) an underestimation of claims may have happened. However, it should be noticed that the FoP is also the first contact the consumer has with products.

To our knowledge, this is the first paper that evaluates the health and wellness-enhancing foods advertised in supermarket circulars. Our findings have important implications to promote healthy eating environments in Brazil. Results highlight to public policy-makers a scenario where food industry places CNIs and additional claims in ultra-processed foods' FoP, which are further promoted by supermarkets as 'healthy' in circulars. Ideally, Brazilian legislation should be revised so that only information truly informing healthy choices to the consumer can be present in FoP labels of products. This is important because using successful advertising practices to promote healthy choices has the potential to enhance the health and well-being of consumers and reduce the expanding healthcare costs [66]. Attention should also be given by health professionals when guiding the population to healthier food choices within the supermarket environment, by disclosing this type of selling strategy.

Further research can be directed at exploring how consumers perceive the advertising of ultra-processed foods in health and wellness-enhancing sections of supermarket circulars and if, how do they shop in these sections. More studies aiming to identify the influence of health claims on FoP over consumer choices are needed, as a systematic review has demonstrated that although those claims are tested in healthy foods, they are mostly used in unhealthy foods [67]. Another potentially interesting area for future research is to investigate whether supermarket circulars in other countries use a similar strategy as in Brazil to promote ultra-processed foods as 'healthy or wellness-enhancing'.

#### **5. Conclusions**

Results indicate that supermarket chains included in this study are promoting the sales of ultra-processed foods in the health and wellness-enhancing sections of their promotion circulars. This can lead consumers to inadvertently choose unhealthy foods when trying to adopt a healthy diet, which suggests a need for revision of Brazilian legislation regarding FoP labelling.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6643/12/9/2877/s1, Figure S1. Examples of circulars' health and wellness sections (page 1); Figure S2. Examples of circulars' health and wellness sections (page 2); Table S1. NOVA food groups: definition according to the extent and purpose of food processing, and examples; Table S2. Additional Claims (AC) on FoP labels of foods advertised in 'health and wellness' sections of supermarket circulars, classified by themes.

**Author Contributions:** Conceptualisation, M.D. and G.M.R.F.; data curation, G.M.R.F.; formal analysis, A.M.B., A.M.d.C. and G.B.I.; funding acquisition, M.D. and G.M.R.F.; project administration, M.D. and G.M.R.F.; supervision, K.J.M. and G.M.R.F.; writing—original draft, A.M.B., K.J.M. and G.B.I.; writing—review and editing, A.M.d.C. and M.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by Newton Mobility Grant Scheme 2015 (Award Reference NG150026 to M.D. and G.M.R.F.), the UK Academies Fellowships Research Mobility, and Young Investigator Awards for UK Researchers in Brazil FAPESC/CONFAP/FUNDO NEWTON (Call Nº 02/2017 to M.D. and G.M.R.F.). Coordination for the Improvement of Higher Education Personnel (CAPES) and the National Council for Scientific and Technological Development (CNPq) provided funding in the form of scholarships to A.M.B.; A.M.d.C.; K.J.M. and G.B.I.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Characteristics of Allergen Labelling and Precautionary Allergen Labelling in Packaged Food Products Available in Latin America**

**Noé Ontiveros <sup>1</sup> , Jesús Aristeo-López Gallardo 2, Jesús Gilberto Arámburo-Gálvez <sup>3</sup> , Carlos Eduardo Beltrán-Cárdenas <sup>2</sup> , Oscar Gerardo Figueroa-Salcido <sup>3</sup> , José Antonio Mora-Melgem 2, Diana María Granda-Restrepo <sup>4</sup> , Cecilia Ivonne Rodríguez-Bellegarrigue <sup>5</sup> , Marcela de Jesús Vergara-Jiménez 2, Feliznando Isidro Cárdenas-Torres <sup>2</sup> , Martina Hilda Gracia-Valenzuela 6,\* and Francisco Cabrera-Chávez 2,\***


Received: 1 August 2020; Accepted: 30 August 2020; Published: 4 September 2020

**Abstract:** The characteristics of food allergen labelling are relevant for avoiding accidental exposure to the allergens of interest but no Latin American country has evaluated these characteristics. Our aim was to evaluate the characteristics of food allergen labelling and precautionary allergen labelling (PAL) in six Latin American countries. All data were collected directly from the supermarkets surveyed. A total of 10,254 packaged food products were analyzed, of which 63.3% (*n* = 6494) and 33.2% (*n* = 3405) featured allergen labelling and/or PAL, respectively. Most products complied with local regulations (≥87.4% for both locally produced and imported). Thirty-three types of PAL statements were detected; the most frequent was "may contain traces of ... " (35.1%). Countries without regulations on the characteristics of allergen labelling had two-fold more products that contained allergens in their ingredients lists but no food allergen labelling. The use of PAL in countries that regulate it (38.2%) was as high as that in countries without PAL regulations (19.2%–44.7%). The findings suggest that the lack of regulations for the characteristics of allergen labeling increases the risk of accidental exposure to allergens of interest. Our findings also suggest that beyond regulations, a scientific approach is required for minimizing and standardizing the use of PAL.

**Keywords:** allergen labelling; Latin America; packaged food products

#### **1. Introduction**

Food-allergic individuals must avoid the allergen of interest in their diets. Food-allergic reactions can vary from mild symptoms to life-threatening anaphylaxis. Intergovernmental members of the Codex Alimentarius have established mandatory general guidelines for the allergen labelling of packaged food products and countries can adapt these guidelines to their particular needs and regulations [1,2]. Despite this, accidental exposure to the allergen of interest often occur, which sometimes have fatal consequences [3–5]. The prevalence rates of food allergies in children and adults vary between 0.6–10.5% and 3–10%, respectively, but current data suggest that the prevalence of this condition is increasing worldwide [6,7]. This fact represents a challenge for packaged food producers regarding the declaration of allergens. In this context, for allergens used as ingredients in the production of specific foods, allergen labelling must be clear and understandable for the general population and comply with local and overseas regulations for exportation purposes. Conversely, there are no regulations in most countries for packaged food products at risk of cross-contamination with food allergens during the manufacturing process and food producers often use precautionary allergen labelling (PAL) (i.e., "may contain" statements) for filling this gap [8]. Although PAL is not intended to replace allergen risk analysis and management, most food products with PAL do not have detectable levels of the allergen(s) of interest and, therefore, do not represent a risk for the vast majority of food-allergic individuals [9,10]. This fact encourages food-allergic individuals to take the risk of consuming foods with PAL and casts doubts among healthcare professionals about the safe use of foods with PAL in allergen-restricted diets [11–13]. Furthermore, the excessive use of PAL limits the availability of packaged food products for food-allergic individuals and/or parents/caregivers, which could give rise to an additional economic burden [9,12,13]. Developed and emerging countries have evaluated the prevalence and characteristics of food products with PAL [14–16] but in Latin America, no study has addressed either the characteristics of food allergen labelling or PAL. Regulations for food allergen labelling can vary from country to country, with the characteristics of the label and the list of allergens declared on it representing the main sources of variation [2]. The Latin American region has particular characteristics in trade, including food products. There are three major trade blocs in Latin America: the Common Market of the South (MERCOSUR) (encompassing Argentina, Brazil, Paraguay, Uruguay, and Venezuela), the Pacific Alliance (encompassing Chile, Colombia, México, and Peru), and the Secretariat for Central American Economic Integration (SIECA) (encompassing Costa Rica, Guatemala, Honduras, El Salvador, Nicaragua, and Panama) but the harmonization of food regulations remains debated [8,17]. Furthermore, although most Latin American countries regulate the allergens to be declared, the characteristics of food allergen labelling are not regulated in some countries such as México, Panamá, and Colombia, and only Argentina, Brazil, and Chile have regulated the use of PAL [17–20]. Thus, the aim of the present study was to evaluate the characteristics of food allergen labelling and PAL in six Latin American countries.

#### **2. Materials and Methods**

#### *2.1. Survey and Data Collection*

Cluster sampling was carried out from January to December 2019. Supermarkets were chosen based on information available online and the suggestions of local citizens. At least one local and one multinational supermarket featuring a flow of consumers with different socioeconomic statuses were included in the survey. Three supermarkets were sampled in Culiacán and three in Mexico City (México, North America), three in San Salvador (El Salvador), four in Panama City (Panamá, Central America), four in Medellín (Colombia), three in Quito (Ecuador), and four in Buenos Aires (Argentina, South America). The sample size consisted of all the packaged food products available on the shelves at the time of the survey. For products with multiple package sizes, only one size was included in the analysis to avoid bias [14]. Duplicate products found across the supermarkets in each country were recorded only once. Information was captured via digital images that covered the brands and

name of the products as well as ingredients and details of allergen labelling, PAL (when available), and place of manufacture [15]. All images were verified twice to corroborate the match between the label captions and products. The information extracted from the images consisted of the following: product brand and name, place of manufacture, ingredients, food allergen declaration, typography (bold, italics, highlighted, colored, etc.), and PAL statements.

#### *2.2. Food Categories and Allergens*

Food products were categorized into twelve groups: baked goods, snacks, confectionaries, baby foods, condiments and sauces, jams and spreads, beverages, powders and pastes, instant foods, chilled and frozen foods, canned foods, and packaged raw foods [15]. The food allergens considered by the Codex Alimentarius guidelines were the focus of the present study (milk and its derivates, egg, soy, wheat, and other cereals containing gluten, peanuts, nuts, fish, and crustaceans) [1].

#### *2.3. Compliance with the Regulations for the Characteristics of Food Allergen Labelling*

Both local and imported packaged food products were verified for their compliance with local regulations for the characteristics of allergen labelling regulations or Codex Alimentarius guidelines for this purpose. The following characteristics were verified: for Argentina, allergen labelling statements with the legend "contains ... " and the use of special typography (capital letters, bold letters, or different colors than the label) as well as the place where the allergen labelling statement appeared on the packaging [18]; for Ecuador, allergen labelling statements with the legend "contains ... ", the use of capital letters, and the place where the allergen labelling statement appeared on the packaging [21]; for El Salvador, the allergen labelling statement with any legend, the use of any special typography to highlight the allergen labelling statement, and the place where the allergen labelling statement appeared on the packaging [22]; for Colombia, México, and Panamá, the allergen labelling statement with any allergen labelling statement and typography [1,23,24]. Regardless of whether the products featured food allergen labelling, all ingredient lists were inspected to verify that they did not include undeclared allergens, i.e., products without food allergen labelling but containing allergens as ingredients were also recorded. The analyses of compliance with local regulations for the characteristics of food allergen labelling only included products with food allergen labelling.

#### *2.4. Frequency of Food Allergen Labelling, PAL, and the Type of PAL*

All products were examined for corroborating the presence of either food-allergic labelling or PAL or both types of allergen labelling. Independently, if the food allergen was declared in the list of ingredients (or not), the products were examined for the presence of PAL. All types of PAL were included in the study.

#### *2.5. Data Analysis*

Descriptive statistics was used and the results are shown as percentages. Confidence intervals were calculated (95%) using the OpenEpi software version 3.01.

#### **3. Results**

#### *3.1. Frequency of Food Allergen Labelling in Packaged Food Products*

Twenty-four supermarkets from six countries (seven cities) were sampled. In total, 10,254 packaged food product labels were analyzed: 71.1% of the products (*n* = 7288; 95% CI, 70.19–71.95) had food allergen labelling and/or PAL. Most of the products without food allergen labelling or PAL were naturally allergen-free products such as chips, juices, rice cookies, and packaged raw food among others (67.70%; *n* = 2008/2966; 95% CI, 65.98–69.38), while 63.3% (*n* = 6494; 95% CI, 62.39–64.26) of the products featured food allergen labelling (Figure 1). Panamá and Argentina had the lowest (*n* = 922; 52.71%) and the highest (*n* = 903; 71.66%) percentages of products with food allergen labelling, respectively (Figure 1A). The most commonly declared allergens were milk, including its derivatives (*n* = 3397; 52.32%); wheat and other cereals containing gluten (*n* = 3335; 51.36%); and soybean, including its derivatives (*n* = 2696; 41.52%) (Figure 1B). The less frequently declared food allergens were fish and crustaceans (*n* = 220; 3.38% and *n* = 58; 0.90%, respectively) (Figure 1B). The most frequent statements used for food allergen labelling were "contains ... " (87.2%) followed by "this product contains ... " (4.6%), "Allergens: contains ... " (2.9%), and "contains ingredients from ... " (2.8%) (Figure 1C). The most commonly used typography for allergen labelling statements was bold letters (70.52%), in either lower case or capital letters, and capital letters without bold typography (58.3%) (Figure 1D).

**Figure 1.** Percentage of products with food allergen labelling and the characteristics of the labelling in commercially available packaged food products in six Latin American countries. (**A**): Percentage of products with food allergen labelling by country (in pie charts: green and red sections; proportions of products with or without food allergen labelling, respectively); (**B**): Percentage of food allergens declared in food allergen labelling; (**C**): Statements used for food allergen labelling; (**D**): Typography used for food allergen labelling (the summation of percentages is greater than 100% due to the use of two or more typographical characteristics in combination).

#### *3.2. Compliance with Local Regulations for the Characteristics of Food Allergen Labelling*

Most packaged foods with food allergen labelling were compliant with the local regulations for this purpose (91.0%, *n* = 5909/6494). Beverages (95.3%), snacks (93.1%), and packaged raw food (92.7%) were the food categories that featured the best compliance with local regulations (Figure 2). The category of baby foods was non-compliant with local regulations to a greater extent than the other categories (15.2%) and was one of the categories with the lowest percentage of allergen labelling (52.27%) (Figure 2). For local or imported food products, 90.67% (range: 77.44%–97.11%) and 87.44% (range: 74.39%–97.92%) of locally produced and imported products complied with local regulations,

respectively (Table 1). In total, 6494 products had food allergen labelling and 585 of them (9.0%) were in non-compliance with local regulations. Of these 585 products, 263 (45.0%) had undisclosed allergens in their food allergen labelling, 295 (50.4%) did not use the proper allergen labelling typography, and 76 products (13.0%) used an allergen labelling statement other than those enforced by local regulations (Supplementary Material Table S1). El Salvador, Ecuador, and Argentina were the countries with the highest proportions of products in non-compliance with local regulations (22.00%, 11.47%, and 8.30%, respectively), followed by México, Colombia, and Panamá (6.24%, 5.77%, and 2.16%, respectively) (Table 1). Countries with the highest proportions of products without food allergen labelling but containing allergens as ingredients were Panamá (36.15%, *n* = 299), Colombia (35.64%, *n* = 236), and México (35.10%, *n* = 231), followed by El Salvador (32.74%, *n* = 167), Argentina (22.12%, *n* = 79), and Ecuador (15.81%, *n* = 118) (Table 1). The number of products without food allergen labelling but containing allergens as ingredients was more than two-fold higher, on average, in Colombia, Panamá, and México than in Argentina, Ecuador, and El Salvador (766 vs. 364, respectively) (Table 1). On average, this means that for every 5.7 (1130 out of 6494) products with allergen labelling, there is one product without food allergen labelling but containing allergens as an ingredient in the Latin American countries included in the survey (Table 1).

**Figure 2.** Compliance and non-compliance with local regulations for the characteristics of food allergen labelling by food category. Percentages in orange bars (-): frequency of non-compliance of food products with local regulations for food allergen labelling; Percentages in blue bars (-): frequency of products in compliance with local regulations for food allergen labelling. Percentages at the tops of the bars correspond to the percentages of packaged foods with food allergen labelling.


**Table 1.** Compliance and non-compliance with local regulations for the characteristics of food allergen labelling.

#### *3.3. Frequency and Characteristics of PAL*

The proportion of products with PAL was 33.2% (*n* = 3405/10,254; 95% CI, 32.29–34.13). This accounts for 46.7% of the total of products with food allergen labelling and/or PAL (3405/7288). PAL was identified in 58.8% (*n* = 827) of the baked goods, 55.37% (*n* = 304) of confectionaries, and 47.84% (*n* = 666) of snacks (Figure 3A). PAL was identified in only 8.5% (*n* = 45) of the canned foods category (Figure 3A). The most frequently used PAL statements were "may contain traces of ... " (*n*= 1195; 35.01%), "may contain ... " (*n* = 1004; 29.5%), and "made in plant that processes ... " (*n* = 261; 7.7%) (Figure 3B). Thirty more PAL statements were identified in 28.60% (*n* = 974) of the products evaluated (Table S1). Some products declared two different types of PAL (0.9%; *n* = 31). The most commonly declared allergens in PAL were nuts (*n* = 1549; 45.5%), soy (*n* = 1466; 43%), and milk (*n* = 1423; 41.8%) (Figure 3B). On average, PAL was identified in 23.6% (*n* = 622) of naturally allergen-free food products and in 25.5% (*n* = 2611) of products with food allergen labelling. Ecuador (19.2%; *n* = 407), Panamá (24.9%; *n* = 436), and El Salvador (31.7%; *n* = 428) had the lowest percentages of products with PAL and Argentina (38.2; *n* = 482), México (42.7%; *n* = 822), and Colombia (44.7%; *n* = 482) had the highest percentages.

**Figure 3.** Percentages of packaged foods with precautionary allergen labelling (PAL) and the most frequent types of PAL. (**A**): frequencies of products with PAL by food categories; (**B**): pie center: most frequently used types of PAL in packaged foods; pie periphery: most frequently declared food allergens in PAL (The summation of percentages is more than 100% due to a combination of two or more allergens in PAL).

#### **4. Discussion**

The correct allergen labelling of food products is essential to minimize accidental exposure to the allergens of interest among food-allergic individuals and to ameliorate the socioeconomic burden associated with this condition. In the present study, the characteristics of the food allergen labelling and PAL of products available in six Latin American countries were analyzed. The results show that, independently of the country sampled, at least one in two packaged food products available on the shelves of the main supermarkets featured food allergen labelling. This illustrates that food-allergic individuals should analyze the labels of each packaged food product that they choose to buy for the first time or analyze such labels regularly to ensure that food producers have not changed their ingredients or manufacturing processes. Consequently, food-allergic individuals will spend additional time buying foods at the supermarket because they have to identify suitable foods. This extra time can be 39% greater than the time spent by non-food-allergic individuals [9]. The most commonly declared food allergens in the six Latin American counties surveyed were the eight foods considered for allergen labelling in the Codex Alimentarius guidelines (milk, wheat, soy, egg, nuts, peanuts, fish, and crustaceans) [1]. These major food allergens declared in the allergen labelling analyzed were reported to be the main triggers of immediate-type food allergy symptoms by the parents of 11,277 Latin American children with different nationalities [25–28]. Certainly, as per the opinions of others [17], the scientific evidence is still insufficient to state that food allergens other than those considered by the Codex Alimentarius guidelines should be added to the food allergen labelling regulations in Latin American countries. The method for stating food allergens in allergen labelling is important to avoid misinterpretations and to easily identify the presence of the allergen of interest. In the present study, the most common allergen labelling statements were "contains", "this product contains", "contains ingredients from", and "allergen: contains", either in bold or capital letters or a combination of these typographies. These results are consistent with the current regulations of Argentina and Ecuador, which enforce the use of the statement "contains" in capital letters (Ecuador) or special typography (capital letters, bold letters, or different colors than the label (Argentina) [18,21]. Although previous studies highlighted that using symbols in combination with word-based allergen labelling statements can be more effective for communicating with food-allergic individuals than using word-based labelling only [29,30], no Latin American regulations for food allergen labelling enforce the use of symbols for such a purpose. The use of symbols in food allergen labelling could, perhaps, reduce the extra time that food-allergic individuals spend identifying suitable foods.

Regulations for food allergen labelling are mainly intended to protect food-allergic individuals from accidental exposure to the allergen of interest and establish the characteristics of the labels. Non-compliance with local regulations entails sanctions ranging from financial penalties to partial or total closure of the responsible company. In general, 91.0% of the food allergen labelling analyzed complied with local regulations. The food category of baked goods had the highest percentage of allergen labelling (88.8%) and was also one of the food categories with the highest and lowest percentages of compliance (91.4%) and non-compliance (8.6%) with local regulations for allergen labelling. Wheat is the major allergenic ingredient in most conventional formulations of baked goods so its declaration in food allergen labelling is mandatory across all Latin American countries [17]. Furthermore, wheat-based baked goods trigger symptoms not only in wheat-allergic individuals but also in some individuals with celiac disease or irritable bowel syndrome as well as in all those with non-celiac gluten sensitivity [31]. Conversely, baby food was one of the categories with the lowest percentage of allergen labelling (52.3%). Most of the baby foods without allergen labelling had no allergens in their ingredient lists but 6.52% of them had undisclosed allergens. This is intriguing because the target population of these foods is particularly vulnerable for developing food allergies [31]. Although food allergen labelling regulations can vary from one country to another, on average, the percentages of locally produced (90.7%) and imported (87.4%) packaged foods that complied with local food allergen labelling regulations were quite similar. This suggests that beyond differences in country to country regulations, most packaged food products comply with the local food

allergen labelling regulations where they are sold. Argentina, Ecuador, and El Salvador established the typography to be used for the food allergen labelling statement and the place where the statement has to be located on the food packaging, with differences in typography between Ecuador and the other two countries [18,21,22]. Furthermore, Argentina and Ecuador have regulated the use of "contains" as their allergen labelling statement [18,22]. Colombia and México have also established food allergen labelling regulations but the allergen labelling statement, typography of the statement, and place where the statement has to be located on the food packaging remain to be established [23,24]. Panamá has not yet established regulations for food labelling [17] but the Codex Alimentarius guidelines are enforced [32,33]. As expected, the countries with "weak" allergen labelling regulations (Colombia, México, and Panamá) had the highest percentage (92.2%, on average) of packaged food products that complied with local allergen labelling regulations. Although this compliance could facilitate the trading of foods among countries, it could also have serious consequences for food-allergic individuals. In this context, our findings show that some products without allergen labelling contained allergens as ingredients and that, on average, the number of products with this peculiarity was two-fold higher in countries that have not yet established the characteristics for their food allergen labelling statements compared with those that have already done so (766 vs. 364). Thus, the risk of choosing a food product at the supermarket without allergen labelling when that food contains allergens is two-fold higher in countries that have not yet regulated the characteristics for food allergen labelling. Others have reported undisclosed allergens in packaged food products using analytical methods [34–36]. Overall, our results suggest that regulations should consider all the characteristics of food allergen labelling to avoid putting food-allergic individuals at increased risk of accidental exposure to their allergens of interest.

PAL is used in food products at the risk of cross-contamination with food allergens during the manufacturing process but there are no regulations for its use in most countries. The present study shows that PAL is widely used in Latin America (33.2% of food products) and that even the proportion of naturally allergen-free products with PAL is considerable (23.6%). Others have reported similar percentages of PAL (28.6%–39.9%) [14,15,37] or even higher (65%) [38]. The PAL statement "may contain traces of ... " was the most commonly utilized statement, which agrees with other studies [15,38] but 32 other types of PAL were also identified in the present study. It should be noted that PAL generates anxiety and confusion among food-allergic individuals and/or parents/caregivers [39] as its excessive use could contribute to increasing the socioeconomic disease burden [12]. Furthermore, both healthcare professionals and food-allergic individuals are facing a dilemma regarding what to do with PAL products [11–13] as most food products with PAL have undetectable levels of allergens and are not risky for most food-allergic individuals [9,10]. Argentina had one of the highest percentage of products with PAL (38.2%) although this country was also the only one that has regulated the use of PAL among the countries surveyed [9,18]. These findings suggest that the regulation of PAL may not be enough to limit its use or to ameliorate the negative impact that the excessive use of PAL has on food-allergic individuals and/or their parents/caregivers. In addition to Argentina, other countries such as South Africa, Brazil, and Chile have regulated the use of PAL. Legislation in these countries allows the use of PAL as long as its use is substantiated by a documented risk assessment demonstrating adherence to good manufacturing practices [9,19,20]. Alternatively, a scientific approach called voluntary incidental trace allergen labelling (VITAL) was developed by the Australian manufacturing industry in 2007 [39]. This approach is based on the reference doses (thresholds) for specific allergens [12,40]. An allergen dose that is likely to trigger allergic reactions in 1% or 5% of the food-allergic population is called ED01 or ED05, respectively [12]. This promising proposal has not been yet endorsed by public health agencies [12] but it seems to have the approval of the majority of the scientific community. Undoubtedly, the endorsement of VITAL by public health agencies, its regulation by countries, and its implementation by food producers will strongly contribute to ameliorating the socioeconomic burden of food allergies and solving the dilemma on how to govern foods products with PAL.

The main strengths of the study are its large number of labels analyzed and its inclusion of countries with different regulations for the characteristics of food allergen labelling including a country that has regulated the use of PAL. These strengths allowed us to draw deeper conclusions than previous studies that have been carried out in only one country. The main limitations of this study include not addressing consumer understandings, perceptions, and interpretations of food allergen labelling and the lack of laboratory tests for confirming the presence or absence of food allergens in products with PAL or in products without food allergen labelling but containing allergens as ingredients. Despite these limitations, the present study provides groundwork for future studies based on immunological methods for detecting food allergens in packaged food products and increases our knowledge about the characteristics of food allergen labelling in the Latin American region.

#### **5. Conclusions**

This is the first study carried out in Latin America to evaluate the characteristics of food allergen labelling in commercially available packaged foods. The packaged food products available with food allergen labelling accounted for 63.3% and the vast majority of those products, both locally produced and imported, complied with local regulations. The countries (Colombia, México, and Panamá) that have not yet regulated the characteristics of their food allergen labelling had more than two-fold greater number of products without labelling when those products contained allergens as ingredients compared with the countries that have already done so (Argentina, Ecuador, and El Salvador), putting their food-allergic populations at increased risk of accidental exposure to their allergens of interest. The use of PAL was high in all of the countries surveyed including one country that has regulated the use of PAL. Furthermore, there were many types of PAL (33), with the most common being "may contain traces of ... ". Additional strategies for PAL regulations should be implemented to minimize and standardize the use of PAL.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6643/12/9/2698/s1, Table S1: Variations of precautionary allergen labelling (PAL).

**Author Contributions:** Conceptualization, N.O. and F.C.-C.; methodology, D.M.G.-R., C.I.R.-B., and M.H.G.-V.; formal analysis, J.A.-L.G., J.G.A.-G., C.E.B.-C., O.G.F.-S., J.A.M.-M., F.I.C.-T., and M.d.J.V.-J.; investigation, J.A.-L.G., D.M.G.-R., and C.I.R.-B.; resources, N.O., F.C.-C., D.M.G.-R., and C.I.R.-B.; data curation, J.A.-L.G., J.G.A.-G., C.E.B.-C., O.G.F.-S., J.A.M.-M., and F.I.C.-T.; writing—original draft preparation, N.O., F.C.-C., and J.A.-L.G.; writing—review and editing, M.d.J.V.-J. and M.H.G.-V.; supervision, D.M.G.-R. and C.I.R.-B.; project administration, N.O. and F.C.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors wish to thank CONACyT for the post-graduate fellowship given to J.A.L.-G. and J.G.A.-G., as well as INAPI Sinaloa (Young Scientific Talents, 2019 program) for the financial support given to C.E.B.-C. and J.A.L.-G.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
