Brazil, similar to other developing countries, experiences a dual scenario characterised by diseases related to nutritional deficiency and to overweight( Reference Araujo, Bezerra and Barbosa 1 ), commonly found in countries undergoing nutritional transition. In this context, micronutrient intake may be affected( Reference Araujo, Bezerra and Barbosa 1 – Reference Fisberg, Marchioni and Castro 6 ), highlighting the importance of micronutrient intake monitoring. The practice of nutritional monitoring is useful to guide public policies focused at preventing deficiency diseases and chronic non-communicable diseases (NCD)( Reference Araujo, Bezerra and Barbosa 1 , Reference Verly Junior, Cesar and Fisberg 2 ).
The role of minerals for human health is undeniable. However, there is a paucity of knowledge regarding the adequacy of mineral intake within the population. Methods that compare habitual intake with dietary requirements have been used to identify individuals at potential risk( Reference Fisberg, Marchioni and Castro 6 ). It is also important to identify factors that influence mineral intake, as such knowledge can assist in creating targeted public policies focusing on minerals, whose intake among the population is typically low. Vlismas et al.( Reference Vlismas, Stavrinos and Panagiotakos 7 ) suggest that socio-economic status may influence dietary habits, with education level being a consistent predictor of health, more so than household income or occupation. Other studies suggest that anthropometric status may influence mineral intake adequacy( Reference Verly Junior, Cesar and Fisberg 2 , Reference Kaidar-Person, Person and Szomstein 8 ), which might be explained by the adoption of unhealthier diets, especially by people of low socio-economic status( Reference Mayen, Marques-Vidal and Paccaud 9 ).
Given this background, our study aimed to estimate the prevalence of dietary mineral inadequacies in a representative sample of residents from urban areas of Sao Paulo, to identify foods that contribute towards mineral intake and to verify the possible association between socio-demographic and lifestyle factors and mineral intake.
Methods
This survey was approved by the Ethics Committee on Research of the Faculty of Public Health, University of Sao Paulo (Certificate of Presentation for Ethical Appreciation (CAAE) no. 26800414·1·0000·5421). Written informed consent was obtained from all individuals, or from their proxies, before commencement of the study.
Study population
This study examined data from the 2008 Health Survey of Sao Paulo (‘Inquérito de Saúde de São Paulo’; ISA-Capital), a cross-sectional, population-based survey that used a complex, stratified, multistage probability sample design to create a representative sample of non-institutionalised residents from urban areas of Sao Paulo, south-eastern Brazil.
Sample recruitment was performed in two stages. First, seventy census tracts were randomly selected from all urban census tracts in Sao Paulo. Second, 16 607 households within these census tracts were randomly selected and stratified by the following domains: (1) infants (<1 year; both sexes), (2) children (1–11 years; both sexes), (3) male adolescents (12–19 years), (4) female adolescents (12–19 years), (5) male adults (20–59 years), (6) female adults (20–59 years), (7) male elderly (≥60 years) and (8) female elderly (≥60 years). Different sampling fractions were used – based on the proportion of each age/sex group in the population of the urban areas sampled – to preserve the representativeness of each domain. The sample size was calculated on the basis of a prevalence of 0·5 with a sample error of 0·07 at a 5 % significance level and design effect of 1·5.
A total of 3271 individuals participated. Of these, 2691 individuals were aged 12 years and older, and were invited to have a dietary assessment performed. Pregnant or lactating women were excluded. Of those who had their diets assessed, 1511 individuals aged 14 years and older were considered. The cut-off point of 14 years was established to maintain the life stage (age) groups defined in the Dietary Reference Intakes: 9–13 years: puberty; 14–18 years: adolescence; 19–30 years: young adulthood; 30–50 years: middle age; 51–70 years: adulthood; and >70 years: older adults( 10 ). These age groups, reflecting the life stages, were used for the analysis of data in the present study. The survey response rate was 77 %. The findings presented here are based on the complete case analysis. The low percentage of missing values was likely missing at random or missing completely at random and will likely have minimal impact on associations. No additional information is available on those who refused to respond.
Data entry and databases
Trained interviewers visited the selected households to conduct individual interviews based on a structured questionnaire, assessing the following: socio-demographic (sex, age, education level, income) and lifestyle (smoking status, leisure-time physical activity, BMI, dietary habits) factors.
Sex was classified as male or female. Age at the date of interview was considered in complete years, and was used to group individuals according to the life stages of Dietary Reference Intakes. Educational level of the household head was defined as the number of years of education completed by the head of the family, categorised as <10 years (elementary school or less) or ≥10 years (high school or more). Categories of family income per capita were calculated using the arithmetic mean of the sum of all family members’ total monetary income – that is, each family member’s income averaged out over the number of family members. The prevailing minimum wage established by the Brazilian government at the time of data collection was US$ 258·39 (R$ 415·00), considering the exchange rate on 1 July 2008. Family income was categorised as ≤1 minimum wage per capita or >1 minimum wage per capita. Smoking behaviour was classified as ‘never’, ‘former’ or ‘current’ smoker, according to the answers provided in the structured questionnaire. Leisure-time physical activity was assessed using relevant questions from the long International Physical Activity Questionnaire( Reference Craig, Marshall and Sjostrom 11 ), and was classified as ‘insufficient’ or ‘sufficient’, according to interpretations by Haskell et al.( Reference Haskell, Lee and Pate 12 ) and Nelson et al. ( Reference Nelson, Rejeski and Blair 13 ). Sufficient physical activity comprised at least 30 min of moderate-intensity exercise daily, 5 d/week, or at least 20 min of vigorous exercise daily, 3 d/week; otherwise, it was considered as insufficient physical activity. Using self-reported height and weight, BMI was calculated as the quotient of weight (kg) over height squared (m2) and classified according to their life stage( Reference de Onis, Onyango and Borghi 14 – Reference Lebrão and Duarte 16 ).
Data from two non-consecutive, 24-h dietary recalls (24HR) were collected by trained interviewers: one during the interview conducted in the selected households and the other by telephone. The dietary interviews were conducted using the United States Department of Agriculture’s (USDA) Multiple-Pass Method and the USDA’s Automated Multiple-Pass Method, respectively. These methods guide the individual through five steps: (1) quick list – the interviewee lists, without interruption, all foods and beverages consumed the previous day; (2) forgotten foods list – the interviewer repeats the list of foods and beverages mentioned by the interviewee to identify foods that may have been forgotten; (3) time and occasion – the interviewee elaborates on the time he/she consumed foods and on what he/she considers to be a meal; (4) detail cycle – the interviewee provides descriptions and amounts of each food reported, and the interviewer reviews each occasion and the interval between occasions; and (5) final review probe – the interviewer repeats all information with the intention of collecting data on additional foods not remembered earlier( Reference Raper, Perloff and Ingwersen 17 ). The 24HR were conducted to include all days of the week and all seasons of the year.
Assessment of dietary mineral intake
The following minerals were considered in this study: K, Na, Ca, P, Mg, Zn, Mn, Cu and Se. Nutrition Data System for Research software, version 2007 (Nutrition Coordinating Centre, University of Minnesota), was used to compute data obtained by the 24HR. After entering data of 24HR, the energy intake values were checked to identify possible mistakes in data collection and entry. Subsequently, foods’ mineral values were reviewed using local information from Brazilian food tables( 18 , Reference Ferreira, Gomes and Bellato 19 ), and, when necessary, the mineral content was corrected to approximate the local reality using the data correction routine of Stata software (version 13.0; StataCorp LP).
Usual dietary intake of minerals was estimated using statistical modelling techniques incorporated in Multiple Source Method programme, a web-based tool used for estimating usual dietary intakes of nutrients and foods consumed by populations and individuals. The usual dietary intake was estimated in a three-step procedure. First, the probability of nutrient intake for each individual was estimated. Second, the usual amount of food intake in days of consumption was estimated. Finally, the above values were multiplied by each other to estimate the usual daily intake for each individual. The programme makes use of covariate information to improve the modelling of consumption probability and intake amount to estimate the distribution of usual intake( 20 ). Sex and age were the covariates used to adjust the modelling.
The prevalence of individual inadequate dietary intakes of Ca, P, Mg, Zn, Cu and Se was determined using the estimated average requirement (EAR) cut-off point approach, a simplification of the probability method, which estimates the proportion of individuals with usual intakes below the EAR (median requirement). Dietary K and Mn intakes were assessed for comparison with the adequate intake (AI), and the proportion of individuals with usual intakes equal to or greater than the AI value was determined. Na intake was assessed by risk of adverse effects using the tolerable upper intake level, which estimates the proportion of individuals potentially at risk for adverse effects( 10 ). These processes were conducted for each sex and age group.
A list of foods contributing to mineral intake was obtained, taking into account the sampling design, as suggested by Block et al.( Reference Block, Dresser and Hartman 21 ). In brief, all foods consumed in the first 24HR were classified into food groups according to frequency of consumption and similarity of mineral content. If the food was consumed fewer than five times, it was grouped with other food items according to the similarity of mineral content – for example, smoothies of papaya, banana, apple, avocado and/or strawberry. If consumed at least five times, the food was kept without grouping – for example banana. Once classified, the percentage of minerals contributed was calculated using sample weighting, and the foods were sequenced in rank order of contribution.
Statistical analysis
Statistical analyses were performed using the survey data commands available in Stata software, which consider the sampling weights (probability weights), clustering and stratification of the survey design( 22 ). A 5 % significance level was considered.
Socio-demographic and health-risk lifestyle factors were expressed as relative frequencies. Mineral intakes are presented as means, standard errors and percentiles stratified by life stage and sex. In order to reduce the errors associated with dietary measurements, the intake of minerals was adjusted according to the total energy intake using the residual method( Reference Willett, Howe and Kushi 23 ). Multiple linear regression methods were used to verify the relationships between mineral intakes (dependent variables) and socio-demographic and lifestyle factors (independent variables). All regression models were adjusted for BMI and energy intake.
Results
In total, 1511 individuals were assessed. Of these, 58·1 % were women, 63·6 % of household heads completed <10 years of education, 50·7 % earned more than one minimum wage per capita, 66·6 % self-reported as never smokers (18·1 % were former smokers), 86·9 % engaged in insufficient leisure-time physical activity and 51·9 % had healthy weight (42·5 % were overweight or obese; Table 1). Their energy intakes are presented in Table 1.
Baseline characteristics of urban residents in Sao Paulo: Inquérito de Saúde de São Paulo; ISA-Capital study, 2008 (Numbers, percentages, mean values and standard deviations)
* Family income per capita is the sum of all family members’ monetary income averaged out over the number of family members. Minimum wage=US$ 258·39 (R$ 415·00).
† Sufficient physical activity: moderate-intensity exercise for at least 30 min daily on 5 d/week or vigorous-intensity exercise for at least 20 min daily on 3 d/week; otherwise, it was considered as insufficient physical activity( Reference Haskell, Lee and Pate 12 , Reference Nelson, Rejeski and Blair 13 ).
‡ BMI classifications: 14–19 years: underweight, BMI<3rd percentile; healthy weight, BMI>3rd percentile and <85th percentile; overweight, BMI>85th percentile and <97th percentile; obese, BMI>97th percentile( Reference de Onis, Onyango and Borghi 14 ); 20–59 years: underweight, BMI<18·5 kg/m2; healthy weight, BMI 18·5–24·9 kg/m2; overweight, BMI 25·0–29·9 kg/m2; obese, BMI≥30 kg/m2 ( 15 ); and ≥60 years: underweight, BMI<23 kg/m2; healthy weight, BMI 23·0–27·9 kg/m2; overweight, BMI 28–29·9 kg/m2; obese, BMI≥30 kg/m2 ( Reference Lebrão and Duarte 16 ).
Among the minerals assessed, the prevalence of inadequate intake was highest for Na, Ca and Mg; this was consistent across all life stages (age groups). The prevalence of inadequate intake varied from 50·0 to 95·4 % for Na, 77·9 to 99·8 % for Ca and 82·9 to 99·5 % for Mg. Adolescents had a considerable prevalence of low dietary P intake, especially female adolescents (Tables 2 and 3). Almost none of the individuals had a K intake equal to or above the AI (Table 4).
Dietary intakes of minerals, their distribution and probability of inadequacy in men residing in urban areas of Sao Paulo, according to life stage: Inquérito de Saúde de São Paulo (ISA-Capital) study, 2008 (Numbers, means values and standard deviations)
EAR, estimated average requirement; UL, upper limit level.
* 14–18 years: adolescence; 19–30 years: young adulthood; 30–50 years: middle age; 51–70 years: adulthood; >70 years: older adults( 10 ).
† For Na, the results are the proportion of individuals with risk of adverse effects by Na intakes, calculated using the UL.
Dietary intakes of minerals, their distribution and probability of inadequacy in women residing in urban areas of Sao Paulo, according to life stage: Inquérito de Saúde de São Paulo (ISA-Capital) study, 2008 (Numbers, mean values and standard deviations)
EAR, estimated average requirement; UL, upper limit level.
* 14–18 years: adolescence; 19–30 years: young adulthood; 30–50 years: middle age; 51–70 years: adulthood; >70 years: older adults( 10 ).
† For Na, the results are the proportion of individuals with risk of adverse effects by Na intakes, calculated using the UL.
Dietary intakes of minerals with adequate intake (AI), their distribution and probability of adequacy in men and women residing in urban areas of Sao Paulo, according to life stage: Inquérito de Saúde de São Paulo (ISA-Capital) study, 2008 (Numbers, mean values and standard deviations)
* 14–18 years: adolescence; 19–30 years: young adulthood; 30–50 years: middle age; 51–70 years: adulthood; >70 years: older adults( 10 ).
In some age-/sex-stratified groups, >20 % of individuals had inadequate intakes of Zn, Cu and Se. Men aged >19 years were more susceptible to inadequate Zn intakes than women, and this inadequacy worsened with age. Unlike Zn, Cu intake inadequacy did not worsen with age, and the highest prevalence of inadequate Cu intake was among women. Female adolescents and adults (≥31 years) had lower Se intakes than men, which became more pronounced with age (Tables 2 and 3). A considerable proportion of individuals had Mn intakes equal to or above the AI: 64·3–81·8 % for men and 77·2–90·3 % for women (Table 4).
The ten foods listed in Table 5 (online Supplementary Tables S5·1 and S5·2) contributed to about 60 % of the total mineral intake. Overall, white rice, brown beans and French bread were the major foods contributing to dietary mineral intake in the sample assessed. Brown beans was among the top five mineral contributors; it was not listed in the top 10 only for Se (general: 132th position; men: 105th position; women: 148th position). White rice and French bread emerged as contributors to all minerals assessed: white rice stayed in the top three foods for most minerals, and French bread was in the top six foods for all minerals. Beef, chicken and milk provided a share of some minerals. Milk was the main contributor of Ca. Meat (beef, chicken, pork and fish) contributed to all minerals except Mn. However, Mn was the only mineral that had among its top 10 food contributors vegetables (lettuce, 8th position (men: 9th; women: 9th)) and whole-grain foods (whole-wheat bread, 5th position (general, men and women)). In addition, banana contributed to Mn (6th (men: 6th; women: 6th)), K (9th (men: 10th; women: 6th)) and Mg (9th (men: 9th; women: 7th)) intakes, and raw tomato contributed to K (8th (men: 8th; women: 7th)) and Cu (10th (men: 14th; women: 12th)), as well as, only for men, Ca (9th) intakes.
Main food contributors to dietary intake of minerals in residents of urban areas of Sao Paulo: Inquérito de Saúde de São Paulo (ISA-Capital) study, 2008 (Percentages of contribution)
* Varied salted flavours.
† Mixed meats.
‡ Varied flavours.
In the multiple linear regression model tested (Table 6), significant positive associations were identified between intake of Ca and household head education level ≥10 years and age ≥51 years; intakes of K, P and Mg and all life stages (age groups); intakes of Na and Mn and life stage 31–50 years; intakes of Na, Ca, Mn and Cu and life stage 51–70 years; intakes of Ca and Mn and life stage >70 years; intake of Se and family income more than one minimum wage; and intakes of K, Ca and Mg and sufficient physical activity.
Multiple linear regression analysis of dietary intake of mineralsFootnote * with socio-demographic and lifestyle factors in residents from urban areas of Sao Paulo: Inquérito de Saúde de São Paulo (ISA-Capital) study, 2008 (β Non-standard regression coefficients and 95 % confidence intervals)
Ref. reference.
* Mineral intakes were adjusted according to the individual energy intake( Reference Willett, Howe and Kushi 23 ) and intra-individual variability( 20 ).
† Cu intakes (mg/d) were log transformed to obtain normal distribution.
‡ Family income per capita is the sum of all family members’ monetary income averaged out over the number of family members. Minimum wage=US$ 258·39 (R$ 415·00).
§ Sufficient physical activity: moderate-intensity exercise for at least 30 min daily on 5 d/week or vigorous-intensity exercise for at least 20 min daily on 3 d/week; otherwise, it was considered as insufficient physical activity( Reference Haskell, Lee and Pate 12 , Reference Nelson, Rejeski and Blair 13 ).
Significant negative associations were identified between female sex and intakes of K, Na, P, Mg, Zn and Mn; household head education level ≥10 years and intake of Na; former smoker status and intake of Ca; and between current smoker status and intakes of K, Ca, P and Cu.
Discussion
Of the minerals assessed in the present study, the only one with probably no likely risk of inadequacy in urban residents of Sao Paulo was Mn. Mn is an essential mineral that functions as an antioxidant and plays an important role in amino acid, cholesterol and carbohydrate metabolism( 24 ). However, a few studies have assessed levels of dietary Mn intake. Hence, the RDA of Mn has not yet been established, making it difficult to assert whether there is, in fact, adequacy of intake( 24 ).
Similar to Mn, the RDA for K and Na could not be derived( 25 ). High levels of dietary Na and lower levels of dietary K were observed among both sexes and all life stages (age groups) assessed. Inadequate K and suboptimal Na intake has also been observed globally( Reference Araujo, Bezerra and Barbosa 1 , Reference Brown, Tzoulaki and Candeias 26 – Reference Powles, Fahimi and Micha 29 ). This combination is considered harmful to health because of the high rates of morbidity and mortality related to CVD. On the other hand, high sodium chloride intake results in increased urinary Na excretion, which leads to loss of other minerals, and thus predisposes the body to changes in homoeostasis( 30 ). According to Powles et al.( Reference Powles, Fahimi and Micha 29 ), Brazilian adults (≥20 years) have the highest dietary Na intakes (4110 mg/d) in Latin America, and higher than those observed in the USA (3610 mg/d) and Canada (3720 mg/d). These authors also observed higher dietary Na intake in men. A possible explanation is that men generally have higher energy requirements than women. Other reasons include an increased concern with appearance and greater health consciousness among women. As evidenced by Souza et al.( Reference Souza, Bezerra and Pereira 31 ), the present study showed that Na intake was provided by traditional Brazilian foods such as rice and beans and other foods such as bread, meats, pizza and salted margarine, as well as the discretionary addition of salt (especially in the form of sodium chloride) during cooking for seasoning. These foods contributed significantly as a source of Na intake. An increased intake of more K-rich foods such as fruits, leafy green vegetables and root vegetables – combined with a reduced intake of high-salt-content products – especially the addition of salt – could help achieve the physiological requirements for K and Na and could decrease the burden of NCD( 25 ).
To improve this scenario, the World Health Organization( 32 ) set a global target of reducing salt intake by 30 % by 2025, aiming for a salt intake of <5g/d. Several countries have implemented salt reduction strategies. These strategies have included engaging with the food industry to re-formulate products, establish the Na content targets for foods, provide consumer education and ensure front-of-pack labelling; imposing taxes on high-salt foods; and introducing interventions in public institutions( Reference Trieu, Neal and Hawkes 33 ). In Brazil, the strategies adopted by the Ministry of Health are promoting healthy eating; providing education and information for healthcare and food handling and manufacturing professionals; and re-formulating processed foods( Reference Nilson, Jaime and Resende 34 ).
In Brazil, NCD are the leading cause of all deaths (72 %) and the largest contributor to disease burden. Diabetes and hypertension have become increasingly prevalent, rising in parallel with overweight( Reference Schmidt, Duncan and Azevedo e Silva 35 ). In addition to the association between dietary Na and NCD, a causal link between dietary inadequacy of Ca and Mg and NCD has been suggested( Reference Sales and Pedrosa 36 – Reference Choi and Bae 41 ). In the present study, the inadequacy of Ca and Mg intake was more pronounced than that of Na; this could increase NCD. One of the reasons for Mg inadequacy is the low consumption of foods rich in Mg, with most of it being derived from foods that are believed to be intermediate or low sources of this micronutrient. The simple change from white rice (12 mg/100 g of food) to whole rice (43 mg/100 g of food) could have a positive effect on Mg intake, especially if the other cereals were also consumed in their whole form.
Dairy products contributed towards Ca intake in the sample assessed. However, its consumption was insufficient to supply physiological needs. Other foods, which are not particularly rich in Ca, were listed because they were consumed in large quantities. Studies have suggested that consumption of dairy products is associated with a decreased risk of CVD, type 2 diabetes and overweight( Reference Soedamah-Muthu, Ding and Al-Delaimy 42 – Reference Schwingshackl, Hoffmann and Schwedhelm 44 ).
Intake of P, unlike the other minerals assessed, showed considerable inadequacy among adolescents, and very low to low intake levels in the other age groups. P is found in nature as inorganic phosphate. In the body, it is regulated by the kidneys and has an important structural role( 30 ). The inadequacies observed among adolescents can be explained by the higher requirements needed to support their intense period of growth( 30 ). Phosphate salts are commonly used by industries as additives, making processed foods good sources when ingested in significant amounts( 30 ). In the present study, the main contributors of P were meat, milk and Brazilian traditional foods. It is important to emphasise that, although most food sources have good P bioavailability, beans and whole-grain cereals contain a non-protoplasmic phytic acid that the human digestive system cannot hydrolyse; hence, this P is not directly available( 30 ).
In terms of Zn intake, in the present study, men were more susceptible to inadequate Zn intake than women. This difference may be reflective of the lower intake of Zn-rich foods, combined with the higher sex-specific recommended requirements( 45 ). In addition, the highest inadequacy observed was among older adults, consistent with the findings of the Brazilian National Dietary Survey( Reference Fisberg, Marchioni and Castro 6 ). This age-related inadequacy of intake may be explained by the commonly observed monotony of the diet and difficulties older people experience with chewing, which impairs the consumption of foods such as meat. Foods that contributed towards Zn intake in the present sample varied from sources that are rich in Zn (red meat) to sources that have moderate (beans and milk) or poor (bread and white rice) Zn content. Consumption of whole grains could help achieve the recommended Zn intake. Inadequacies in dietary Zn can lead to impairment of the immune system, changes in oxidative stress and other alterations that can raise the risk of developing NCD( 45 ).
Inadequacies in the intake of Cu and Se have also been implicated as factors that may contribute to the increased risk of developing NCD; the role of both minerals as antioxidants would be one reason for this( 45 , 46 ). Se, when ingested in large amounts, can also increase the risk of diabetes, thus taking on an uncertain role( 46 ). In the present study, Cu and Se, similar to Zn, showed moderate-to-low inadequacies, and women had lower intakes. These findings can probably be explained by the lower amount of foods consumed by women, which consequently reduces the intake quantity of these minerals, which have the same dietary reference intake for both sexes.
Although liver, one of the most Cu-rich foods, has been listed among the contributors of this mineral, the other foods listed are considered to have moderate levels of Cu, supporting the observed inadequacy. For Se, in addition to meat and rice, foods containing wheat contributed to the consumption of this mineral. Beans did not appear to contribute because the variety widely consumed in Sao Paulo – brown beans or ‘carioquinha’ beans – have the lowest Se concentrations (0·1 μg/100 g) compared with other varieties such as black beans, 11·9 μg/100 g; red beans, 3·2 μg/100 g; and white beans, 2·6 μg/100 g( Reference Ferreira, Gomes and Bellato 19 ).
This present study confirms that socio-demographic and lifestyle factors influence the intakes of minerals differently. Sex and age were the main factors influencing mineral intake, showing that, although it is suggested that high cost and low access to foods might explain low mineral intake – because most mineral-rich foods are usually more expensive( Reference Verly Junior, Cesar and Fisberg 2 , Reference Vlismas, Stavrinos and Panagiotakos 7 , Reference Yang, Hsu-Hage and Tian 47 , Reference Lim, Park and Lee 48 ) – other personal factors, including lifestyle, may influence mineral intake to an equal or greater extent. Being female, for example, in the present study showed one interesting aspect: decreased Na intake. The decreased Na intake among women was also observed in the Thai National Health Examination Survey IV, specifically in women ≥30 years( Reference Satheannoppakao, Kasemsup and Inthawong 49 ). On the other hand, except for Na, Zn, Cu and Se, older age was associated with increased mineral intake. This observation may reflect a better diet quality in older people as a response to minimise the diseases common of age, as described by Andrade( Reference Andrade 50 ).
In the present study, the level of education of the household head and family income per capita were associated with intakes of different minerals. This highlights that these two factors impact the intake of minerals differently and independently, as has been suggested by some authors( Reference Vlismas, Stavrinos and Panagiotakos 7 , Reference Araujo, Verly Junior and Junger 51 , Reference Si Hassen, Castetbon and Cardon 52 ). Although the education level of the household head was shown to influence Na and Ca intake, family income per capita was shown to influence only the intake of Se. Education may be a potential means of improving micronutrient intake; this information can be considered while developing public policies.
The adoption of healthier habits by participants engaging in sufficient physical activity was reflected by the increased intakes of K, Ca and Mg. Smoking, even in those categorised as former smokers, was negatively associated with mineral intake. The negative effects of smoking may lead to the development of unhealthier habits and by possible changes in gustatory capacity.
Doubtless, the global decline observed in the intake of whole grains( Reference Micha, Khatibzadeh and Shi 53 ) directly affects the inadequacy of mineral intake, as does the lower consumption of fruits and vegetables. Urgent consideration needs to be given to the question of how to approach the population in order to make individuals and the food industry aware of the necessity of adopting healthy lifestyle habits, in order to prevent diseases. From this perspective, dietary guidelines are important tools for health promotion. Such guidelines can encourage healthy eating habits that take into account changes in population eating habits, health conditions and advances in scientific knowledge( 54 ).
Globally, public policies for minerals have been more focused on Fe, I and Na( Reference Trieu, Neal and Hawkes 33 , 55 – Reference Darnton-Hill and Nalubola 57 ). Some countries have also focused on Zn and Ca, encouraging Zn and Ca fortification of foods( 56 , Reference Darnton-Hill and Nalubola 57 ). Other minerals, whose intakes are frequently inadequate, such as Mg, are not contemplated in these actions. This may be due to a paucity of data regarding the population intake of these minerals. New public policies that include the monitoring of more minerals are necessary to avoid mineral deficiencies and the consequences thereof.
A critical appraisal of the present study must to be taken into consideration. Evaluation of dietary intake is susceptible to both random and systematic errors. To minimise this bias, 24HR was conducted by a trained interviewer using standardised methods, as multiple-pass methods minimise the omission of possibly forgotten foods and standardise the level of detail for describing foods. Moreover, the interviewers were trained to communicate directly with the participants to minimise the influence of psychological determinants of misreporting( Reference Poslusna, Ruprich and de Vries 58 ). In addition, the intra-individual variation was removed, and the effect of energy was considered for the associations tested. Finally, Se content in foods, especially in plants, can vary according to the availability of this mineral in the soil where the plant is grown; this may also influence the Se content in meats, depending on the animal feedstuff( 46 ). Therefore, it is important to recognise that, despite the corrections made in the database with data analysis of Brazilian foods( Reference Ferreira, Gomes and Bellato 19 ), the values obtained for all minerals may contain inaccuracies.
In conclusion, residents from urban areas of Sao Paulo have significantly inadequate intakes of most of the minerals assessed, taking all ages/life stages and both sexes into account. The high prevalence of inadequate intake raises concerns because of the association with impairments in vital functions and development of NCD. Moreover, the major foods contributing towards dietary intake of minerals were not the most mineral-rich food sources, but reflected the Brazilian cultural dietary pattern. Socio-demographic and lifestyle factors may influence the ability to reach the dietary reference intake of minerals. These findings may contribute to identification of population subgroups that are vulnerable to inadequate intake of minerals, to better target public health policies and to assist in monitoring established policies.
Acknowledgements
The authors are grateful to the ISA team and to the members of the Food Consumption Research Group. Particular thanks go to Prof. Chester Luiz Galvão César (Faculdade de Saúde Pública, Universidade de São Paulo) and Prof. Moisés Goldbaum (Faculdade de Medicina, Universidade de São Paulo) for ISA coordination, as well as to Dr Maria Cecilia Goi Porto Alves (Instituto de Saúde, Secretaria da Saúde, Governo do Estado de São Paulo) for her sample size analysis of the study population.
This study was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq (R. M. F., grant no. 473100/2009-6, D. M. M., grant no. 481176/2008-0); by the Fundação de Amparo à Pesquisa do Estado de São Paulo – FAPESP (R. M. F. grant nos 2009/15831-0, 2012/22113-9); by the Prefeitura do Município de São Paulo (C. L. G. C., no grant number); and by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – CAPES (C. H. S., no grant number). Funders cited had no role in the design, analysis or writing of this article.
C. H. S. performed the analysis and interpretation of data and drafted the article; M. d. M. F. and D. A. S. V. critically revised the analysis of data and the article; D. M. M. participated in designing the ISA study and critically revised the article; R. M. F. participated in designing the ISA study, supervising the analysis and critically revising the article. All authors read and approved the submitted version.
None of the authors has any potential conflicts of interest to declare.
Supplementary Material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114516004633
