Coronary heart disease mortality, cardiovascular

International Journal of Cardiology 217 (2016) 64–68

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Coronary heart disease mortality, cardiovascular disease mortality and all-cause mortality attributable to dietary intake over 20 years in Brazil Leandro Fórnias Machado de Rezende a,⁎, Catarina Machado Azeredo b, Daniela Silva Canella c, Olinda do Carmo Luiz a, Renata Bertazzi Levy a, Jose Eluf-Neto a a b c

Faculdade de Medicina da Universidade de São Paulo, Departamento de Medicina Preventiva, Av. Dr. Arnaldo 455, 2° andar, Sao Paulo, Sao Paulo 01246-903, Brazil Faculdade de Medicina da Universidade Federal de Uberlândia, Curso de Nutrição, Av. Pará, 1720, Bloco 2U, Sala 20, Campus Umuarama, Uberlândia, Minas Gerais 38.405-320, Brazil Instituto de Nutrição da Universidade do Estado do Rio de Janeiro, Departamento de Nutrição Aplicada, Rua São Francisco Xavier, 524, Rio de Janeiro, RJ20559-900, Brazil

a r t i c l e

i n f o

Article history: Received 1 March 2016 Received in revised form 28 April 2016 Accepted 30 April 2016 Available online 03 May 2016 Keywords: Diet Mortality Cardiovascular diseases Epidemiology

a b s t r a c t Background/objectives: In the last two decades, in Brazil, there has been a decreasing trend of consumption of in natura or minimally processed food, while intake of ultra-processed food has markedly increased. We estimated the contribution of dietary intake in trends from coronary heart disease mortality (CHDM), cardiovascular disease mortality (CVDM), and all-cause mortality (ACM) over 20 years in Brazil. Methods: We used a representative sample of Brazilian households located in metropolitan areas to estimate dietary intake in 1987/88 and 2008/09. For both periods, we estimated fractions of CHDM, CVDM, and ACM attributable to healthy (fruits and vegetables) and unhealthy food items (sugar-sweetened beverages, processed and red meat). We also estimated the number of prevented or postponed deaths attributable to these food items. Results: The fraction of CHDM attributable to all food items increased from 28.6% in 1987/88 to 38.7% in 2008/09. CVDM attributable to food items increased from 13.7% in 1974 to 19.3% in 2008/09. ACM attributable to all food items increased from 20.1% in 1987/88 to 27.3% in 2008/09. Without the decrease in healthy food item consumption, and the increase in unhealthy food items, 3195 deaths from coronary heart disease, 5340 from cardiovascular disease, and 16,970 from all causes could have been prevented or postponed. Conclusions: The burden of cardiovascular diseases and mortality attributable to dietary intake has increased over the last 20 years in Brazil. These findings suggest a need for a population prevention approach, focused on dietary intake to reduce the burden of disease. © 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Cardiovascular disease mortality (CVDM), mostly coronary heart disease mortality (CHDM) and stroke, accounted for almost two thirds of all deaths in 2013 [1]. Despite the high burden of CVDM, agestandardized death rates have fallen since 1990, especially in highincome countries [1]. In many middle-income countries, including Brazil, CVDM rates also declined in the last decade [2]. These CVDM rate trends might be partially explained by changes in modifiable risk factors such as smoking, alcohol consumption, physical inactivity, and dietary consumption. Many studies have shown the association between different food items (i.e. fruits and vegetables, sugar-sweetened beverages, processed and red meat) with CVDM [3–7]. In the last two decades, there has been slightly greater consumption of healthy food worldwide, although, by ⁎ Corresponding author. E-mail addresses: [email protected] (L.F.M. Rezende), [email protected] (C.M. Azeredo), [email protected] (D.S. Canella), [email protected] (O.C. Luiz), [email protected] (R.B. Levy), [email protected] (J. Eluf-Neto).

http://dx.doi.org/10.1016/j.ijcard.2016.04.176 0167-5273/© 2016 Elsevier Ireland Ltd. All rights reserved.

contrast, intake of unhealthy food items increased to a greater extent [8]. In Brazil, a worse trend was observed. In the last two decades, there has been a decreasing trend in the consumption of in natura and minimally processed food in total caloric intake from 44.0 to 38.9%. In addition, consumption of ultra-processed food, such as ready-to-eat or drink formulations with a high content of sugar, salt, fat and additives, increased in total caloric intake from 18.7 to 29.6% [9]. In this study, we aimed to estimate the contribution of dietary intake trends for coronary heart disease mortality (CHDM), CVDM, and allcause mortality (ACM) over 20 years in Brazil. 2. Methods 2.1. Estimation of dietary intake We obtained data on the average intake of healthy (fruits, vegetables) and unhealthy food (sugar-sweetened beverages, processed and red meat) from the National Household Budget Survey (Pesquisa de Orçamentos Familiares — POF) carried out in two periods: March 1987 to February 1988 [10] and May 2008 to May 2009 [11]. These food

L.F.M. Rezende et al. / International Journal of Cardiology 217 (2016) 64–68

items were selected based on the evidence of association with CHDM, CVDM, and ACM [3–7]. POF 1987/88 and POF 2008/09 were representative samples of Brazilian households located in 11 metropolitan areas distributed around the five regions of the country (Belem in the North; Fortaleza, Recife and Salvador in the Northeast; Belo Horizonte, Rio de Janeiro and Sao Paulo in the Southeast; Curitiba and Porto Alegre in the South; and the Federal District and the municipality of Goiania in the Mid-west). The total sample of households in these areas was 13,611 in 1987/88 and 15,399 in 2008/09 [10,11]. These 11 metropolitan regions together represent around a third of the total Brazilian household population [10–12], but present higher socioeconomic status, urbanization, and income inequality when compared to the whole country [24]. Despite the national coverage of POF 2008/09, which included 55,970 households, we decided to analyze only the data from the 11 metropolitan areas (15,399 households) in order to follow the same sampling criteria as the POF 1987/88 survey. We also chose to analyze these metropolitan areas because mortality data are better registered. Both surveys used complex multi-stage, clustered sample design, involving geographic and socioeconomic stratification among all the census tracts of the country, followed by random sampling of sectors (first stage) and households (second stage). Data collection was performed over 12 months, distributed uniformly among tracts, in order to assure representation in the four quarters of the year. Further details regarding the sampling procedures are available elsewhere [10,11]. For both surveys, we extracted the purchase data for each food item, and quantities were converted and expressed in grams per capita per day (g/day). We considered all types of fruits, vegetables (excluding tubers) and sugar-sweetened beverages (cola, guaraná, orange, lemon, apple, grape, non-specified, other sodas). We defined red meat as beef, pork, and lamb as fresh, chilled or frozen. We defined processed meat as salted or cured meat and sausages (e.g. sausage, ham, salami, and bologna). 2.2. Effects of dietary intake on coronary heart disease mortality, cardiovascular disease mortality and all-cause mortality

65

We used the following levels of consumption to determining risk: consumption of red meat above the recommended 70 g/day; any consumption of processed meat and sugar-sweetened beverages, since only no consumption was considered to be risk free; and consumption lower than 400 g for fruit and vegetable intake (of which 2/3 should be from vegetables (240 g) and 1/3 from fruits (160 g/day)) [15]. Finally, the joint impact of all food item consumption was estimated based on the following equation: n

PAF ¼ 1−∏ ð1−PAFi Þ i¼1

where PAFi is the PAF of each food marker, when applicable. 2.4. Deaths prevented or postponed attributable to food items We estimated deaths prevented or postponed attributable to food items using a similar model to previous publications [16,17]. Data on the total metropolitan area population and age distribution from 1987/88 (average) and 2008/09 (average) were obtained from the Brazilian Census Bureau. Age-standardized rates from CHDM (CID 9: 410 to 414; CID 10: I20 to I25), CVDM (CID 9: 390 to 459; CID 10: I00 to I99), and ACM were obtained from the Brazilian National Mortality Database — DATASUS [18], for 1987/88 (average) and 2008/09 (average). Multiplying the age-specific mortality rates for 1987/88 by the population in each age stratum in 2008/09, we calculated the number of deaths from the three outcomes that would have been expected in 2008/09 if the mortality rates in 1987/88 had remained unchanged. In addition, we subtracted the number of deaths observed in 2008/09 from the number expected in 2008/09, achieving the total number of deaths prevented or postponed in this period. Finally, to estimate the number of deaths prevented or postponed attributable to food items, we multiplied the number of deaths for each of the three outcomes in 1987/88 by the difference between the PAF in 2008/09 and 1987/88. 3. Results

The magnitude of association of each food item with outcomes of interest was retrieved through a literature review. We searched for recent meta-analysis on the Medline database using keywords related to exposures (vegetables, fruits, sugar-sweetened beverages, processed and red meat) and outcomes of interest (CHDM, CVDM, and ACM). When more than one meta-analysis was eligible for the exposure–outcome relationship, we retrieved relative risks (RRs) from the most recent study. The RRs retrieved from published meta-analysis are given in Table A.1. Further details of the search strategy and included meta-analysis are given in Table A.2. 2.3. Calculation of population attributable fraction We estimated the population attributable fraction (PAF) for CHDM, CVDM, and ACM associated with each food item. Our estimates were based on a theoretical minimum risk, a risk factor elimination perspective, according to international recommendations for dietary intake [13, 14]. We estimated the percentage of CHDM, CVDM, and ACM attributable to each food item for 1987/88 and 2008/09 using the following equation:

PAF ¼

R−1 R

where R = exp[In(RRunit) × x ]RRunit = relative risk for each unit increment in the exposurex = average exposure in the population.

From 1987/88 to 2008/09, the age-standardized mortality rates (per 1000) for CHDM sharply decreased from 1.23 to 0.66, CVDM from 3.69 to 1.97, and ACM from 9.13 to 6.59. The total number of deaths from CHDM, CVDM, and ACM prevented or postponed in the period was 31,611, 96,213, and 133,839, respectively (Table 1). The fraction of CHDM attributable to all food items increased from 28.6% in 1987/88 to 38.7% in 2008/09 in Brazilian metropolitan areas (Table 2). Approximately 3195 deaths from coronary heart disease (10% of the CHDM rate reduction between 1987/88 and 2008/09) were not prevented or postponed because of the decreased consumption of fruits and vegetables, and the increased consumption of processed meat and sugar-sweetened beverages, between 1987/88 and 2008/09 (Table 3). CVDM attributable to all food items in Brazilian metropolitan areas increased from 13.7% in 1987/88 to 19.3% in 2008/09 (Table 2). Approximately 5340 deaths from CVD (5.5% of the CVDM rate reduction between 1987/88 and 2008/09) were not prevented or postponed because of the decreased consumption of fruits and vegetables and the increased consumption of processed meat between 1987/88 and 2008/09 (Table 3). ACM attributable to all food items increased from 20.1% in 1987/88 to 27.3% in 2008/09 in Brazilian metropolitan areas (Table 1). Approximately 16,970 deaths from all causes (12.7% of the ACM rate reduction between 1987/88 and 2008/09) were not prevented or postponed because of the decreased consumption of fruits and vegetables and the increased consumption of processed meat between 1987/88 and 2008/09 (Table 3).

66


Table 1 Coronary heart disease, cardiovascular disease, all-cause mortality: age-standardized rates, deaths observed and expected, and total number of deaths prevented or postponed. Brazilian metropolitan areas, 1987/88 and 2008/09. Age-standardized rates (per 1000)

Coronary heart disease mortality Cardiovascular disease mortality All-cause mortality * **

Deaths observed

1987/88**

2008/09*

1987/88**

2008/09*

1.23 3.69 9.13

0.66 1.97 6.59

31,636 95,351 235,692

36,689 110,782 350,829

Deaths expected in 2008/09*

Total number of deaths prevented or postponed

68,299 206,994 484,668

31,611 96,213 133,839

Average values between 2008 and 2009. Average values between 1987 and 1988.

4. Discussion We estimated the extent to which dietary intake contributed to CHDM, CVDM, and ACM over 20 years in Brazilian metropolitan areas. Even though CHDM, CVDM, and ACM rates fell between 1987/88 and 2008/09, we observed an increase in the burden of these outcomes attributable to dietary intake. The decrease in consumption of healthy food items, and increase in consumption of unhealthy food items resulted in failure to prevent or postpone 3195 deaths from CHD, 5340 from CVD, and 16,970 from all causes. The explanation for the decrease in age-standardized rates of CHDM, CVDM, and ACM in these two decades, even with the increase in the negative contribution of dietary intake, might be related to other risk factors that have also been causally related to cardiovascular diseases, such as smoking and physical inactivity. In Brazil, prevalence of smoking among adults halved between 1989 (29%) and 2013 (15%) [19,20]. Additionally, leisure-time physical activity has increased (annual increase of 1.9%) over the last decade in Brazil, while television-viewing time has decreased (annual decline in 5%) [21]. Further, coverage of the Family Heath Program (FHP), the largest primary care program in the world, has increased. This has also been shown to be associated with a reduction in hospitalizations and mortality in Brazil between 2000 and 2009. FHP is related to a higher number of health education activities, domiciliary visits from health professionals and physicians, which might help to explain the plausibility of the associations [22].

Table 2 Population attributable fraction for outcomes associated with food items in Brazilian metropolitan areas 1987–2008. Population attributable fraction 1987/08

2008/09

Coronary heart disease mortality Fruits Vegetables Sugar-sweetened beverages Processed meat Red meat Joint PAF

2.2 20.7 1.4 6.7 – 28.6

5.2 24.1 2.9 12.3 – 38.7

Cardiovascular disease mortality Fruits Vegetables Sugar-sweetened beverages Processed meat Red meat Joint PAF

2.0 9.1 – 4.2 0.0 13.7

4.8 10.7 – 7.7 0.0 19.3

All-cause mortality Fruits Vegetables Sugar-sweetened beverages Processed meat Red meat Joint PAF

2.5 14.7 – 3.9 0.0 20.1

5.9 16.7 – 7.1 0.0 27.3

Finally, in the last two decades, there has been a significant improvement in primary and secondary prevention of cardiovascular diseases. For instance, a previous study conducted in two European countries found that an increase in statin prescription partly explained the reduction of CHDM between 2001 and 2011. On the other hand, discontinuation of statins, which is more common in low-income areas, increases the risk of CHD events [23]. However, to the best of our knowledge, there are no studies evaluating such impact on mortality trends in Brazil. Our results highlight the importance of dietary intake in the burden of disease in Brazilian metropolitan areas, which present higher socioeconomic status, urbanization, and income inequality when compared to the whole country [24]. These urbanized regions have been shown to be the most exposed to unhealthy dietary intake in low–middle-income countries [9,25]. Nevertheless, food items analyzed in our study presented similar trends between 2002 and 2008 in these metropolitan areas and in Brazil as a whole, despite the socioeconomic differences [9]. These results suggest that the increase in the burden of the analyzed outcomes attributable to dietary intake in our study might be happening in Brazil as a whole. Globally, the contribution of dietary intake to the burden of disease has also increased. In 1990, dietary intake was considered the second risk factor, accounting for 7%, for global disability-adjusted life years (DALY) but increased to become the leading risk factor, accounting for 10%, of global DALY in 2010. Among food items, low intake of fruit, vegetables, nuts, seeds, and whole grains was the greatest contributor to the burden of disease [26]. In our study, low intake of vegetables and high intake of processed meat were the main food items related to the burden of CHDM, CVDM, and ACM over the last 20 years. The food items included in our study represent a small fraction in terms of total energy consumption estimated for the Brazilian population. National data for 2008–2009 indicate that fruits, vegetables, processed meat and sugar-sweetened beverages accounted for around 2% or less of total calories each, and red meat accounted for around 5%, totaling around 12% of total energy consumption [9]. Nonetheless, according to NOVA — a food classification based on the extent and purpose of industrial processing, these food items are part of two broad food groups: ultra-processed food products and in natura or minimally processed food [27]. Research using this classification has shown an increase in intake of the former group, which is replacing the latter [9]. Therefore, the scenario indicated in our study would probably be worse if these groups were studied as a whole. Unfortunately, there are no estimates of the magnitude of association between dietary intake and the three outcomes analyzed in our study based on this food classification, and this has prevented us from conducting a wider analysis. A limitation of our study is mainly related to household purchase of food items, which does not necessarily represent individual consumption. However, international and national studies found similar results when comparing household purchase and individual intake [28–30]. Additionally, food bought and consumed away from home was not included in the survey. In Brazil, between 2002 and 2008, there was an increase in spending on food consumed outside the home [31], which may


67

Table 3 Deaths from coronary heart disease, cardiovascular disease, and all-causes that were prevented or postponed as result of changes in dietary intake in the Brazilian metropolitan areas, 1987 to 2008.

Fruits Vegetables Sugar-sweetened beverages Processed meat Red meat Total food markers

Absolute level food markers (g/day)

Change in risk factor

Deaths prevented or postponed

1987

2008

Absolute change (g/day)

Relative change (%)

Coronary heart disease

Cardiovascular disease mortality

All-causes

No. deaths

% of total reduction

No. deaths


No. deaths


131.5 82.8 40.0

90.2 54.2 81.8

−41.3 −28.6 41.8

−31.4 −34.5 104.5

−949 −1076 −475

−3.0 −3.4 −1.5

−2670 −1526 –

−2.8 −1.6 –

−8014 −4714 –

−6.9 −3.5 –

9.9 61.6

18.6 47.6

8.7 −14.2

87.9 −22.7

−1772 – −3195

−5.6 – −10.1

−3337 0 −5340

−3.5 0.0 −5.5

−7542 0 −16,970

−5.6 0.0 −12.7

lead to an underestimation of PAF related to unhealthy food items, and an overestimate of PAF related to healthy food items. However, food consumption outside the home corresponded to only 16% of total consumption of calories in 2008 [32]. Finally, using household purchase did not allow us to analyze the degree to which dietary intake differed between subgroups (e.g. sex, age). Despite these limitations, the main strength in using household purchase lies in the possibility of analyzing, for the first time, the contribution of dietary intake to deaths from CHD, CVD, and all causes over 20 years in Brazilian metropolitan areas. We also made some assumptions in the use of RR. We used average RR estimates from meta-analysis, which did not consider risk differences between subgroups. On the other hand, studies included in the meta-analysis adjusted by possible mediators of the relationship between food items and outcomes analyzed, such as body mass index and metabolic profile, provided a conservative estimate of the RR and, consequently, our estimates. Additionally, these meta-analyses did not include any Brazilian studies in the summary effect, and we therefore made RR portability assumptions for: 1) exposure measure and; 2) disease latency in the cohorts are similar to Brazilian metropolitan areas; 3) absence of effect modification [33]. 5. Conclusion In conclusion, the burden of CHDM, CVDM, and ACM attributable to dietary intake increased over the last 20 years in Brazilian metropolitan areas. These findings suggest the need for a population prevention approach focused on dietary intake, such as economic incentives to consume healthful foods, labeling and information actions, multicomponent interventions in schools and workplaces, and restrictions on advertising and marketing ultra-processed food [34], among others, in order to reduce the burden of disease in Brazil.

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10] [11]

[12] [13]

[14]

[15]

Funding sources

[16]

LFMR acknowledges the PhD scholarship: grant #2014/25614-4, São Paulo Research Foundation (FAPESP).

[17]

Conflict of interest

[18] [19]

None. Acknowledgments

[20]

[21]

None. [22]

References [1] GBD collaborators, Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis

[23]

for the Global Burden of Disease Study 2013, Lancet (London, England) 385 (2015) 117–171. M.I. Schmidt, B.B. Duncan, G. Azevedo e Silva, A.M. Menezes, C.A. Monteiro, S.M. Barreto, et al., Chronic non-communicable diseases in Brazil: burden and current challenges, Lancet (London, England) 377 (2011) 1949–1961. L. Dauchet, P. Amouyel, S. Hercberg, J. Dallongeville, Fruit and vegetable consumption and risk of coronary heart disease: a meta-analysis of cohort studies, J. Nutr. 136 (2006) 2588–2593. C. Huang, J. Huang, Y. Tian, X. Yang, D. Gu, Sugar sweetened beverages consumption and risk of coronary heart disease: a meta-analysis of prospective studies, Atherosclerosis 234 (2014) 11–16. R. Micha, S.K. Wallace, D. Mozaffarian, Red and processed meat consumption and risk of incident coronary heart disease, stroke, and diabetes mellitus: a systematic review and meta-analysis, Circulation 121 (2010) 2271–2283. X. Wang, Y. Ouyang, J. Liu, M. Zhu, G. Zhao, W. Bao, et al., Fruit and vegetable consumption and mortality from all causes, cardiovascular disease, and cancer: systematic review and dose–response meta-analysis of prospective cohort studies, BMJ (Clin. Res. Ed.) 349 (2014) g4490. I. Abete, D. Romaguera, A.R. Vieira, A. Lopez de Munain, T. Norat, Association between total, processed, red and white meat consumption and all-cause, CVD and IHD mortality: a meta-analysis of cohort studies, Br. J. Nutr. 112 (2014) 762–775. F. Imamura, R. Micha, S. Khatibzadeh, S. Fahimi, P. Shi, J. Powles, et al., Dietary quality among men and women in 187 countries in 1990 and 2010: a systematic assessment, Lancet Global Health 3 (2015) e132–e142. A.P. Martins, R.B. Levy, R.M. Claro, J.C. Moubarac, C.A. Monteiro, Increased contribution of ultra-processed food products in the Brazilian diet (1987–2009), Rev. Saude Publica 47 (2013) 656–665. Instituto Brasileiro de Geografia e Estatística, Pesquisa de Orçamentos Familiares, 1987–88: regiões metropolitanas, IBGE, Rio de Janeiro, 1991. Instituto Brasileiro de Geografia e Estatística. Pesquisa de Orçamentos Familiares 2008–2009: Avaliação nutricional da disponibilidade de alimentos no Brasil. Rio de Janeiro: IBGE; 2010. Instituto Brasileiro de Geografia e Estatística, SIDRA: banco de dados agregados, IBGE, Brasília, 2015. World Health Organization, Diet, nutrition and the prevention of chronic diseases, Report of the Joint WHO/FAO Expert Consultation, World Health Organization, Geneva, 2003. World Cancer Research Fund/American Institute for Cancer Research, Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective, AICR, Washington DC, 2007. World Cancer Research Fund Hong Kong, Eat Right, Live Bright, WCRF HC, Hong Kong, 2009. E.S. Ford, U.A. Ajani, J.B. Croft, J.A. Critchley, D.R. Labarthe, T.E. Kottke, et al., Explaining the decrease in U.S. deaths from coronary disease, 1980–2000, N. Engl. J. Med. 356 (2007) 2388–2398. B. Unal, J.A. Critchley, S. Capewell, Explaining the decline in coronary heart disease mortality in England and Wales between 1981 and 2000, Circulation 109 (2004) 1101–1107. Brazil. Ministry of Health, Datasus. Sistema de Informacao sobre Mortalidade, 2012. C.A. Monteiro, T.M. Cavalcante, E.C. Moura, R.M. Claro, C.L. Szwarcwald, Populationbased evidence of a strong decline in the prevalence of smokers in Brazil (1989– 2003), Bull. World Health Organ. 85 (2007) 527–534. D.C. Malta, T.P. Oliveira, M.L. Vieira, L. Almeida, C.L. Szwarcwald, Use of tobacco and exposure to tobacco smoke in Brazil: resuts from the National Health Survey 2013, Epidemiol Serv Saúde. 24 (2015) 239–248. G.I. Mielke, P.C. Hallal, D.C. Malta, I.M. Lee, Time trends of physical activity and television viewing time in Brazil: 2006–2012, Int. J. Behav. Nutr. Phys. Act. 11 (2014) 101. D. Rasella, M.O. Harhay, M.L. Pamponet, R. Aquino, M.L. Barreto, Impact of primary health care on mortality from heart and cerebrovascular diseases in Brazil: a nationwide analysis of longitudinal data, BMJ (Clin. Res. Ed.) 349 (2014) g4014. F. Vancheri, B. Wettermark, L.E. Strender, L.G. Backlund, Trends in coronary heart disease mortality and statin utilization in two European areas with different

68

[24] [25]

[26]

[27]

[28]

L.F.M. Rezende et al. / International Journal of Cardiology 217 (2016) 64–68 population risk levels: Stockholm and Sicily, Int. Cardiovasc. Forum Journal. 1 (2014) 140–146. Instituto Brasileiro de Geografia e Estatística, Pesquisa Nacional por Amostra de Domicilio (PNAD). September 17, 2015 ed. IBGE, Rio de Janeiro, 2008/2009. J.C. Moubarac, R.M. Claro, L.G. Baraldi, R.B. Levy, A.P. Martins, G. Cannon, et al., International differences in cost and consumption of ready-to-consume food and drink products: United Kingdom and Brazil, 2008–2009, Global Public Health 8 (2013) 845–856. S.S. Lim, T. Vos, A.D. Flaxman, G. Danaei, K. Shibuya, H. Adair-Rohani, et al., A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010, Lancet (London, Engl. 380 (2012) 2224–2260. J.C. Moubarac, D.C. Parra, G. Cannon, C.A. Monteiro, Food classification systems based on food processing: significance and implications for policies and actions: a systematic literature review and assessment, Curr. Obesity Reps. 3 (2014) 256–272. W. Becker, Comparability of household and individual food consumption data — evidence from Sweden, Public Health Nutr. 4 (2001) 1177–1182.

[29] A. Naska, V.G. Vasdekis, A. Trichopoulou, A preliminary assessment of the use of household budget survey data for the prediction of individual food consumption, Public Health Nutr. 4 (2001) 1159–1165. [30] R.M. Claro, P.C. Jaime, K. Lock, R.M. Fisberg, C.A. Monteiro, Discrepancies among ecological, household, and individual data on fruits and vegetables consumption in Brazil, Cadernos de Saude Publica 26 (2010) 2168–2176. [31] R.M. Claro, L.G. Baraldi, A.P. Martins, D.H. Bandoni, R.B. Levy, Trends in spending on eating away from home in Brazil, 2002–2003 to 2008–2009, Cadernos de Saude Publica 30 (2014) 1418–1426. [32] Instituto Brasileiro de Geografia e Estatística, Pesquisa de Orçamentos Familiares 2008– 2009 – Análise do consumo alimentar pessoal no Brasil, IBGE, Rio de Janeiro, 2011. [33] K. Steenland, B. Armstrong, An overview of methods for calculating the burden of disease due to specific risk factors, Epidemiol. (Cambridge, Mass) 17 (2006) 512–519. [34] D. Mozaffarian, A. Afshin, N.L. Benowitz, V. Bittner, S.R. Daniels, H.A. Franch, et al., Population approaches to improve diet, physical activity, and smoking habits: a scientific statement from the American Heart Association, Circulation 126 (2012) 1514–1563.