Metabolic Syndrome (MetS) is characterized by the coexistence of metabolic risk factors for type 2 diabetes mellitus and cardiovascular diseases, including insulin resistance, dyslipidemia, central obesity, and hypertension [1,2]. Furthermore, it has shown an increasing prevalence and affects about a quarter of the world’s population [3]. In Brazil, data from the 2013 Pesquisa Nacional de Saúde (PNS, National Health Survey) revealed a prevalence of 38.4% in the adult population [4].

Weight gain is one of the major risk factors for MetS. Data from the National Health and Nutrition Examination Survey in the United States show that MetS was present in 5% of those with normal weight, 22% of those overweight, and 60% of those with obesity [5]. In this sense, lifestyle modifications, especially related to dietary changes and physical activity, are the main recommended interventions for the management of MetS [1,6].

It has been shown that legume consumption exerts beneficial effects on the prevention and as an adjunct in the treatment of metabolic disorders. In a study with Canadian adults, regular consumption of legumes such as peas, chickpeas, beans, and lentils promoted beneficial changes in risk factors for MetS, such as waist circumference, insulin resistance, glycated hemoglobin, HDL-cholesterol, peptide-C, and fasting insulin [7].

Beans are widely consumed in Brazil [8] and can be highlighted for their nutritional profile, representing an important source of nutrients, especially proteins, dietary fiber, complex carbohydrates, and micronutrients such as vitamins, minerals, and phytochemicals [9,10].

Due to its nutritional characteristics, bean consumption has been attributed to various health benefits, including reduced cardiovascular risk [11], serum cholesterol [12], blood glucose [13,14], serum triglycerides [15], weight [16,17], and waist circumference [18]. Despite this, studies relating to bean consumption and MetS are still scarce. In view of the above, the present study was conducted with the objective of verifying the association between bean consumption and components of MetS, as well as with the presence of diagnosed MetS in adults.

The data for this study were obtained from the Inquérito de Saúde Domiciliar no Piauí (ISAD-PI, Household Health Survey in Piauí) from 2018-2019 [19], a population-based and household-based cross-sectional study, referring to a representative sample of the municipalities of Teresina and Picos, to analyze the living conditions and health status of the urban population, through visits to private households.

For this investigation, adult individuals of both sexes, residing in the municipality of Teresina in the state of Piauí, with collected data on food consumption and biochemical analyzes were included.

The study sampling plan was carried out by a cluster sampling process, in two stages: Primary Sampling Units (PSU), composed of census sectors, and in the second stage, households, based on the Instituto Brasileiro de Geografia e Estatística (IBGE, The Brazilian Institute of Geography and Statistics) census data for the year 2010 [20].

For the sample size calculation, the population size and the number of private households in Teresina (767,557 inhabitants; 210,093 households) were considered. From this data, the average number of individuals of both sexes per household was calculated in each of the following age groups: children under 2 years old; children aged 3 to 4 years; children aged 5 to 9 years; adolescents aged 10 to 14 years; adolescents aged 15 to 19 years; adults aged 20 to 59 years and the elderly over 60 years. A total of 578 households was then estimated. The final sample size for this study was adjusted using n=n₀/0.90, assuming a response rate of 90%, resulting in n≅642 households. All residents of the selected households were invited to participate upon signing the Free and Informed Consent Form.

The overall sampling fraction used in this study was f = aMi ∑Mi x M ’ i where: f = overall sampling fraction; a = total number of PSUs to be selected in the first stage; M ‘ i = number of households in the PSU “i”; = number of households to be drawn from each selected PSU. In this way, all regions of the urban area of the municipality were covered, so that it was possible to have an actual picture of the city of Teresina, considering that all data represent the socioeconomic and health situation of the municipality.

Following the same sampling plan, 50% of the households were selected, forming a subsample for the collection of food consumption data, through the application of 24-hour dietary recalls, as well as for blood analysis. Thus, the number of individuals in the final sample was considered adequate for public health studies, allowing adequate precision, in general, with a Coefficient of Variation (CV) for proportions below 20% [19]. Further details about the sample size and sampling plan of the ISAD-PI have been published in the study by Rodrigues et al. [19].

After the conclusion of the research, the final sample included 497 households in Teresina, and the subsample consisted of 248 households. In the end, data on food consumption and blood tests were obtained from 365 and 242 adult individuals, respectively. Of these, a total of 192 individuals were included in this study because they had both data collected (Figure 1). In the present study, most estimates had a CV below 20% and, in more than 90% of them, the design effect was less than 1.5, indicating precision in the estimated parameters and that the sampling design was successful [9].

Demographic data (age, sex, skin color, and marital status), socioeconomic (education and income), lifestyle (alcohol consumption, smoking status, and physical activity) and use of medications were collected through the application of structured questionnaires, applied by trained interviewers, using mobile devices, using the Epicollect 5® application (Imperial College London) [21]. Anthropometric data (weight, height, and waist circumference), blood pressure, and food consumption data (24-hour dietary recall) were also measured. In addition, blood was collected for biochemical analysis (fasting glucose, triglycerides, and HDL-c - High-Density Lipoprotein).

Weight (kilograms) and height (meters) data were collected according to the recommendations of Cameron [22] and Jelliffe, Jelliffe [23]. Waist circumference – WC (centimeters) was measured according to the World Health Organization – WHO recommendation [24]. Blood pressure was measured using an aneroid sphygmomanometer (Durashock Welch AllynTycos®, NY, USA, Model DS-44) following the recommendations of the Brazilian Society of Cardiology [25].

Food consumption data were obtained through the application of a 24-hour dietary recall, using the Multiple Pass Method [26]. The household measures reported by the interviewees were transformed into grams (g) or milliliters (mL) using the Table for Assessment of Food Intake in Household Measures, based on the study by Pinheiro et al. [27]. Energy intake was estimated based on the Tabela Brasileira de Composição de Alimentos (TACO, Brazilian Food Composition Table) [28], the Tabela de Composição Nutricional dos Alimentos Consumidos no Brasil (TBCA, Table of Nutritional Composition of Foods Consumed in Brazil) [29] and the Food Composition Table – Support to Decision Nutrition [30].

Bean consumption was categorized based on the energy value equivalent to a serving size, defined as one medium shallow ladle, approximately 80g (50% grain and 50% broth) corresponds to approximately 55 kcal [27,28]. The groups were categorized as follows: <55 kcal – less than one medium shallow ladle; 55 to 100 kcal – one to two medium shallow ladles; >110 kcal – two or more medium shallow ladles.

Blood was collected after a 12-hour overnight fast. To obtain the serum, the samples were centrifuged (3000 rpm – 5 minutes) within 4 hours after blood collection, the concentrations of glucose, serum triglycerides, and serum HDL-cholesterol (High-Density Lipoprotein) were quantified using the colorimetric-enzymatic method, using Labtest® reagent kits.

For the diagnosis of MetS, the NCEP ATP III criteria (National Cholesterol Education Program Adult Treatment Panel III) [31] were used. It was defined by the combination of at least three of the following criteria: 1) abdominal obesity, defined by waist circumference in men >102 cm and in women >88 cm; 2) serum triglycerides ≥150 mg/dL; 3) serum HDL-cholesterol (HDL-c) <40 mg/dL in men and <50 mg/dL in women; 4) Systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85mmHg; 5) Fasting glucose ≥110 mg/dL.

The study was submitted to the Ethics Committee of the Universidade Federal do Piauí (UFPI, Federal University of Piauí) and approved under opinion number 2.552.426. After clarification, all individuals participated in the research by signing theFree and Informed Consent Form.

For variables with missing values, imputation was performed using the Predictive Mean Matching method, not exceeding 20%.

Imputation for missing data was performed for the following variables analyzed in this study: income, education level, alcohol consumption, and smoking status. Predictor variables included geographic location of the household, sex, age, employment status, and marital status.

All analyzes were performed using the survey module of the Stata version 14 program, which considers the sampling design (equiprobabilistic) in the estimation of the results. The normality of the distribution of the variables was assessed using the Kolmogorov-Smirnov test. A 95% Confidence Interval (95% CI) was assumed, and a p-value <0.05 was considered significant.

The Chi-square test was used to verify the association between the frequency of bean consumption and the explanatory variables in the bivariate analysis. The multiple analysis of the association between MetS and bean consumption was performed using the Prevalence Ratio (PR), obtained through Poisson regression, with robust variance adjustment (95% CI).

The selection of confounding variables was guided by a theoretical model involving exposure, outcome, and confounding variables, designed using Directed Acyclic Graphs (DAGs) in the Dagitty software version 3.2. After the elaboration of the DAG, a minimum set of confounding variables was indicated, established by the backdoor criterion, which were age, sex, education level, income, alcohol consumption, smoking status, physical activity, and medication use [32].

Table 1 presents the characterization of the sample according to demographic, socioeconomic, lifestyle aspects, and prevalence of MetS. It was observed that most of the studied population is female (72.9%), with black, brown, or indigenous skin color (77.1%), studied up to high school or less (68.7%), has no partner (55.7%), has a family income less than or equal to two minimum salaries (75.5%) and works (58.9%). Regarding lifestyle habits, the majority does not consume alcoholic beverages (53.1%), does not smoke (90.6%), and is sufficiently active (94.3%).

The prevalence of MetS was found to be 31.2% (95% CI: 23.5-40.3). Moreover, an association with education level was observed, being more prevalent in individuals with lower education (up to high school or less) (36.4%; p=0.0211) compared to individuals with higher education (higher education or more) (20.0%).

Table 2 shows the bean consumption data according to demographic, socioeconomic, and lifestyle aspects. No associations were observed in the analyzed variables.

Table 3 shows the prevalence ratio of MetS and MetS components according to bean consumption in adult individuals. Bean consumption was not associated with MetS components; however, it was observed that individuals who consumed more than 110 kcal of beans/day had a 48% lower prevalence ratio of MetS (PR: 0.52; 95% CI: 0.29-0.91) compared to individuals who consumed less than 55 kcal of beans/day.

The present study showed a high prevalence of MetS in the studied population. Santos et al. [33], similarly to this study, identified a prevalence of 30.9% of MetS in adult individuals from Florianópolis, Santa Catarina. Likewise, a study conducted with individuals from Pará showed a prevalence of 29.9% in adults [34].

Individuals with lower education levels showed a higher prevalence of MetS. This result was also identified in a study conducted with data from the 2013 Pesquisa Nacional de Saúde (PNS, Brazilian National Health Survey), in which the prevalence of MetS was almost double in individuals with 0 to 8 years of education, compared to those with more years of education [4]. Similarly, in a study conducted by Ramires et al. [35], the lowest level of education was associated with a higher prevalence of MetS in women.

Education is one of the major determinant factors of an individual’s lifestyle, as people with higher levels of education are less likely to have unhealthy lifestyle habits associated with the risk of developing MetS. In addition, educational level is also associated with access to health services [36].

Furthermore, bean consumption was associated with a lower prevalence of MetS in adults. The protective effect of bean consumption on the presence of MetS was also verified by Chang et al. [37], in a 6.5-year longitudinal study conducted in Taiwan, in which it was observed that a bean-free diet was associated with an increased chance of developing MetS by up to 83% in men and 45% in women.

Several studies have shown that bean consumption improves metabolic aspects, such as reduced waist circumference [16,38,39], reduced body fat percentage [38], reduced blood pressure [7,39], and improved postprandial glycemic control [40]. In a randomized crossover study with adults, it was demonstrated that the consumption of brown beans compared to the consumption of white bread provided a significant reduction in glycemia and insulinemia [13].

In the present study, although no associations were observed between bean consumption and the components of MetS individually, an association was observed between bean consumption and the presence of these components when present simultaneously in the diagnosis of MetS. Thus, the protective effect of bean consumption was demonstrated, moreover, it should be considered that the risk of negative health outcomes related to MetS is synergistically greater than the estimate of its factors separately [4].

The protective effect of beans can be explained by their high content of dietary fiber and complex carbohydrates with slow digestion, which delay gastric emptying and, consequently, help reduce the release of postprandial glucose, favoring glycemic control. Beans are also characterized by being a low-energy-density food and having a low-fat content, which can favor weight control. In addition [41], beans are also rich in saponins, substances associated with improved insulin sensitivity and reduced cholesterol concentrations [42,43].

In a randomized crossover clinical trial, Reverri et al. [44], demonstrated that the inclusion of black beans in a typical Western-style meal promotes the attenuation of postprandial insulin and moderately improves the antioxidant capacity of adults with MetS, which may be partially explained by the fiber content and antioxidant properties of beans. A large part of the antioxidant properties of beans comes from its high content of phenolic compounds, whose antioxidant action can contribute to the prevention of oxidative events that can trigger the onset of metabolic complications [10].

Bean consumption can also lead to the production of bioactive peptides through proteolytic hydrolysis during gastrointestinal digestion. Bioactive peptides from beans can promote antioxidant, antithrombotic, and antihypertensive effects. Some of these beneficial effects of bioactive peptides from beans can be explained by the ability of bioactive peptides to inhibit certain enzymes related to the onset of chronic non-communicable diseases, such as dipeptidyl peptidase IV, alpha-amylase, and alpha-glucosidase, associated with type 2 Diabetes Mellitus, as well as the inhibition of the angiotensin-converting enzyme, related to the development of hypertension [10].

Approximately 58.6% of the Brazilian population consumes beans five or more days a week, mostly men and those with lower education levels [45]. However, research on the acquisition of food items for household consumption in Brazil points to a reduction in minimally processed or unprocessed foods, such as beans, to the detriment of increased consumption of ultra-processed foods [46,47].

Thus, a change in the population’s dietary consumption profile is perceptible, since Cruz et al. [48], when evaluating the consumption of dietary fiber in the Brazilian population and its relationship with the intake of ultra-processed foods, observed that the highest quintile of consumption of ultra-processed foods was inversely associated with insufficient consumption of fiber, unprocessed or minimally processed foods.

It is known that ultra-processed foods have some particularities in their composition, such as high amounts of sugar, sodium, saturated fat, as well as high caloric content and low or absent content of dietary fiber and micronutrients. This composition, in turn, increases the risk of mortality and development of chronic non-communicable diseases, such as obesity [49-51].

This study presents limitations, as there were losses in the sample size. However, studies evaluating the association between bean consumption in the diet and the presence of MetS are still scarce, as most studies evaluate the consumption of legumes in general or only isolated nutrients and compounds from beans or other health outcomes other than the diagnosis of MetS per se. Thus, the findings of the present study can serve as a basis for the development of more studies on the topic.

A high prevalence of MetS was observed in the adult population of Teresina, with a higher proportion in individuals with lower educational levels. Additionally, a higher participation of bean consumption in the diet was inversely associated with the prevalence of MetS, constituting a protective factor for the study population.

Given the scarcity of studies evaluating the effects of bean consumption on the diagnosis of MetS, there is still a need for more research to elucidate the topic, and the present work can serve as a basis for the development of future research.