Cristiane Bonaldi Cano Reliability of Food Labels from Products Marketed in the City of São

CONCLUSÕES: Foram encontrados altos índices de não conformidade dos dados nutricionais nos rótulos de alimentos destinados ao público adolescente e infantil, indicando a urgência de ações de fi scalização e de outras medidas de rotulagem nutricional.


INTRODUCTION
Greater contribution of industrialized foods, rich in sugar and fat, in the Brazilian family diet, to the detriment of basic foods that are sources of complex carbohydrates and dietary fi ber, is a relevant characteristic of food pattern changes in the last decades. 10a a Instituto Brasileiro de Geografi a e Estatística.Pesquisa de Orçamentos Familiares 2002-2003: análise da disponibilidade domiciliar de alimentos e do estado nutricional no Brasil.Rio de Janeiro; 2004.
Governmental actions in Brazil include seeking to guarantee both Nutrition and Food Safety, a in accordance with the WHO Global Strategy on Diet, Physical Activity and Health. 13Assuring the presence of useful and reliable information on food labels is a right guaranteed by the Brazilian Consumer Protection Code.b Food labeling entitles the consumer to obtain nutritional information as well as to access quality and safety food parameters.Access to this information meets the requirements of the legislation and, at the same time, encourages companies to invest in the nutritional improvement of their products, whose explicit composition may infl uence the consumer towards purchase. 8A refl ection on the historical evolution of food label legislation in Brazil, compared to other countries, has recently been published. 8According to Ferreira & Lanfer-Marquez, 8 the international food market progress and the recognition of consumer rights are among the factors that require standardization of regulations and constant improvement of nutrition label norms.
In 1999, when the Agência Nacional de Vigilância Sanitária (ANVISA -National Health Surveillance Agency) was founded, nutritional labeling of foods became mandatory in Brazil.The main pieces of legislation concerning manufactured food labeling in Brazil are the following: RDC (Resolution of the Collegiate Board of Directors) 259/02, c which defi nes and establishes measures and serving sizes, including cooking measures and its corresponding serving sizes in grams or milliliters, and details the utensils commonly used in the kitchen, their approximate dimensions and capacities; RDC 360/03, d which establishes, among other specifi cations, the mandatory inclusion of the energy value and protein, carbohydrate, total fat, saturated fat, trans fat, dietary fi ber and sodium contents on manufactured food labels.This legislation enables: a criterion for rounding numbers; a 20% variability in nutrition information; and also the gathering of nutrient data by means of physical-chemical analyses, theoretical calculations based on the product formula, data compiled from food composition tables or data provided by raw material manufacturers.
Tolerance of 20% for non-compliance (plus or minus) between the data stated on the label and the "actual" data does not meet the Consumer Protection Code.b However, it considers the inevitable variation of raw material composition as well as the changes that oc-cur due to food processing, and the gathering of food composition data from other countries, for certain foods/dishes.
Nutritional labeling must provide reliable information so that this instrument may be able to fulfi ll its purpose of helping consumers with their choices, and also supporting health professionals to offer guidance on diet composition.
The Brazilian situation concerning fulfi llment of nutrition label legislation is still discouraging.This situation is corroborated by research conducted between 1997 and 2004, which shows how often the legislation is disobeyed. 5e present study aimed to assess the reliability of nutrition information stated on food labeling of industrialized foods.

METHODS
This study deals with the assessment of nutritional value data, stated on nutritional labeling of pre-packaged foods by law.These data were compared to those obtained from physical-chemical analyses performed in laboratory.
Sampling plan was comprised by industrialized salty and sweet foods, which are described as often consumed by children and adolescents. 4,e These products were purchased in the city of São Paulo, Southeastern Brazil, between 2001 and 2005.Samples were gathered using the Health Surveillance Agency standard inspection procedure, in accordance with the current legislation, f and subsequently sent to the Instituto Adolfo Lutz (Adolfo Lutz Institute) for analysis.
A total of 153 samples of industrialized products from 84 different brands were analyzed.Among the salty products, 56 products and 34 brands were analyzed, while among sweet products, 97 products from 50 distinct brands were analyzed.Samples were divided into sweet products (cookies, wafers, milk chocolate, white chocolate, chocolates) and salty products (corn snacks, wheat snacks, potato chips, peanuts) with different fl avors and from distinct brands, disregarding the format established by the manufacturer or the product lot number (Table 1).Samples of products were collected in duplicate for the analyses, 15  the carbohydrate content was obtained by the di-• fference between a hundred and the sum of values of moisture, ashes, proteins, total fat and dietary fi ber.
To calculate the energy value, 4 kcal per gram of • carbohydrates and proteins and 9 kcal per gram of total fat were considered.
Results of nutritional composition were expressed in g/100g.Samples were approved or rejected by comparing the values obtained experimentally to the amounts of nutrients stated on labels.Thus, for each sample analyzed, the serving size (in grams) stated on the label was calculated, considering the 20% variability tolerated by the RDC 360/03.The interval of rejection considered, or percentage range of each nutrient, consists in the frequency of non-compliance variation in the samples.
Descriptive data analysis was performed and the 95% confi dence interval for the mean value of each nutrient was calculated.The Minitab software, version 13.0, was used for statistical treatment of data.

RESULTS
As regards nutritional composition (Table 2), all products analyzed showed high energy value, as well as fat and sodium contents.In contrast, dietary fi ber values were low (mean values between 1.36 and 2.87g/ 100g), except for peanut samples (6.14g/ 100g).
All samples analyzed showed some non-compliance of the nutrition information stated on the food label (Table 3).
Among salty products, 0 to 50% of the samples were rejected due to their protein content, of which samples of potato chips showed the highest percentage (50%) among the products analyzed.For carbohydrates, variation of non-compliance was between 0 and 40%, and wheat snacks were the only product that did not show samples rejected for this nutrient.None of the samples of salty products was approved in terms of their dietary fi ber, sodium and saturated fat contents.Corn snacks showed the highest frequency of samples rejected: 69% for dietary fi ber, 72% for sodium, 85% for total fat and 41% for saturated fat.
Among sweet products, 10% to 40% were rejected due to their protein content.The frequency of samples rejected, in relation to the total fat content, varied between 0 and 75% in the case of chocolates and milk chocolate.
None of the cookie samples showed non-compliance for total fat.However, the saturated fat content data obtained in laboratory was not found to be in conformity with that stated on the label in 52% of the cookie samples.For almost all the products (except for white chocolate), there was between 6% and 36% of rejection in terms of the dietary fi ber content.The frequency of rejection for dietary fi ber content was high in the case of chocolates (29%), milk chocolate (29%) and cookies (36%).Chocolates were the sweet product with the highest frequencies of samples showing noncompliance of nutrients, as 40% of the samples were rejected for their protein content, 75% for total fat, 14% for saturated fat and 29% for dietary fi ber.

DISCUSSION
To compare nutrient analytical quantifi cation data with those stated on food labels, some factors that may interfere with the sampling plan and result interpretation should be considered.Among these are the following: number of samples, raw material control, type of manufacturing process adopted, storage, quality control procedures employed, analytical methods or food composition tables used by companies to determine nutritional information of products. 15In this study, some of these factors could not be strictly controlled, as samples were randomly collected and sent to the Health Surveillance Agency for analysis.a One limitation to this study was associated with the lack of differentiation of samples whose nutrient values were either overestimated or underestimated, because only the 20% variability for non-compliance allowed by the current legislation was considered for statistical purposes.
Discrepancy between nutrient data obtained in laboratory and those stated on the label by the manufacturer, in the case of the products analyzed, could be explained by: a) analytical matters related to: extraction methods for total fat and fractions, interference of the food matrix composition with fi ber analytical determination or differences associated with ingredient composition, such as products with several fi llings; 15 or b) the nutritional value calculation from food composition tables, based on the product's ingredients or raw materials.However, these differences, regardless of their cause, must not surpass the 20% variability (plus or minus) tolerated by the current legislation (RDC 360/03 Resolution).
The food groups analyzed were selected due to children's and adolescents' preferences and also due to their high fat and sodium contents and low dietary fi ber content, which may have an impact on morbimortality indices. 3,4,14In the present study, these nutrients showed the highest proportions of non-compliance of samples.
Food labeling, by informing consumers about the quality and amount of nutritional constituents of products, must contribute to promote appropriate food choices and be used as a nutritional education tool for the population. 6,9hus, legitimacy of information is mandatory.
Another important implication of the high frequency of rejected label data is associated with epidemiological studies, as assessment of food consumption is based on nutritional data provided by food labels. 3,11,12,14Lack of reliability of nutritional information on labels may become bias when estimating consumption data and also compromise the identifi cation of associations between diet factors and physiopathological factors involved with obesity and non-communicable chronic diseases.

Table 1 .
Samples and brands of salty and sweet foods analyzed.City of São Paulo, Southeastern Brazil, 2001-2005.Spices and lemon, cream cheese, onion and parsley, natural, Parmesan cheese.