Abstract:

This study sought to investigate the effect of race on measures of body fat (body mass index - BMI, waist circumference - WC and waist-hip ratio - WHR), as well as its relationship with diabetes, among elderly individuals living in urban areas in seven places in Brazil, according to gender. This is a cross-sectional study carried out with a probabilistic sample comprising 2,566 individuals with 65 years of age or more who participated in the FIBRA Study (Frailty in Elderly Brazilians). We used several self-reported sociodemographic variables (gender, age, race, schooling and family income), anthropometric measures of general (BMI) and abdominal obesity (WC and WHR) and self-reported diabetes. Adjusting for schooling and income, white race was associated with higher WC values (p = 0.001) and WHR (p > 0.001) for male gender, regardless of diabetes status. However, when we considered only diabetic individuals, black race became associated with general (BMI) (p = 0.007) and central obesity (CC) (p > 0.001), only among women.

Keywords:
Aged; Diabetes Mellitus; Obesity; Race or Ethnic Group Distribution

Introduction

Race and ethnicity are concepts used in health services and the scientific literature to identify socioeconomic disparities and, though they have different definition, the two are often confused ¹1. Alves C, Fortuna CMM, Toralles MBP. A aplicação e o conceito de raça em saúde pública: definições, controvérsias e sugestões para uniformizar sua utilização nas pesquisas biomédicas e na prática clínica. Gazeta Médica da Bahia 2005; 75:92-115.. Race is described as a group of individuals who share morphological or phenotypical characteristics ¹1. Alves C, Fortuna CMM, Toralles MBP. A aplicação e o conceito de raça em saúde pública: definições, controvérsias e sugestões para uniformizar sua utilização nas pesquisas biomédicas e na prática clínica. Gazeta Médica da Bahia 2005; 75:92-115.. It is represented by self-reported skin color in the main Brazilian censuses, carried out by the Brazilian Institute of Geography and Statistics (IBGE) ²2. Instituto Brasileiro de Geografia e Estatística. Características étnico-raciais da população. Um estudo das categorias de classificação de cor ou raça, 2008. http://www.ibge.gov.br/home/estatistica/populacao/caracteristicas_raciais/PCERP2008.pdf (acessado em 25/Ago/2013).
http://www.ibge.gov.br/home/estatistica/... . Ethnicity goes beyond physical characteristics. It also involves cultural, social, linguistic, religious, territorial and diet variables ¹1. Alves C, Fortuna CMM, Toralles MBP. A aplicação e o conceito de raça em saúde pública: definições, controvérsias e sugestões para uniformizar sua utilização nas pesquisas biomédicas e na prática clínica. Gazeta Médica da Bahia 2005; 75:92-115..

Despite its relevance for identifying individuals exposed to health risks, some raise questions regarding the use of race and ethnicity in research, considering the heterogeneity and methodological complexity of these variables and the subsequent absence of a consensus as to what they measure, how data should be collected and what is the most appropriate categorization of study populations ³3. Laguardia J. O uso da variável "raça" na pesquisa em saúde. Physis (Rio J.) 2004; 14:197-234..

Obesity, defined as harmful excess of body fat, is now considered a global epidemic, with growing prevalences in the population, even among the elderly ⁴4. Mathus-Vlieglen EM. Prevalence, pathophysiology, health consequences and treatment options of obesity in the elderly: a guideline. Obes Facts 2012; 5:460-83.. According to data from the National Health and Nutrition Examination Survey (NHANES) ⁵5. Fakhouri TH, Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity among older adults in the United States, 2007-2010. NCHS Data Brief 2012; (106):1-8., one third of U.S. elderly individuals (65 years or more) are obese, with a body mass index (BMI) equal to or higher than 30kg/m². In Brazil, 57.8% and 19.8% of individuals aged 65 years or more are overweight and obese, respectively, according to recent data from the Brazilian Ministry of Health ⁶6. Secretaria de Vigilância em Saúde/Secretaria de Gestão Estratégica e Participativa, Ministério da Saúde. VIGITEL Brasil 2014: vigilância de fatores de risco e proteção para doenças crônicas por inquérito telefônico. http://www.abeso.org.br/uploads/downloads/72/553a243c4b9f3.pdf (acessado em 10/Set/2015).
http://www.abeso.org.br/uploads/download... . Silva et al. ⁷7. Silva VS, Souza I, Petroski EL, Silva DAS. Prevalence and factors associated with overweight in Brazilian elderly. Rev Bras Ativ Fís Saúde 2011; 16:289-94. found obesity prevalences of 13.7% (60 to 69 years), 11.5% (70 to 79 years) and 8.3% (80 years or more) in a representative sample of elderly Brazilians.

Obesity plays a crucial etiological role in a series of chronic conditions, chief among them diabetes mellitus ⁴4. Mathus-Vlieglen EM. Prevalence, pathophysiology, health consequences and treatment options of obesity in the elderly: a guideline. Obes Facts 2012; 5:460-83.^{), (8}8. Edson EJ, Sierra-Johnson J, Curtis B. Diabetes and obesity in older adults: a call to action. Rev Clin Gerontol 2009; 19:135-47.^{), (9}9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28., which was responsible for 5.1 million deaths worldwide in 2013 ¹⁰10. International Diabetes Federation. IDF Diabetes atlas. http://www.idf.org/diabetesatlas (acessado em 10/Jun/2014).
http://www.idf.org/diabetesatlas... . Particularly among the elderly, obesity, especially abdominal obesity, and diabetes have similar characteristics, which include chronic inflammation (characterized by higher inflammation markers) and insulin resistance ¹¹11. Hanley AJG, Wagenknecht LE, Norris JM, Bryer-Ash M, Chen YI, Anderson AM, et al. Insulin resistance, beta cell dysfunction and visceral adiposity as predictors of incident diabetes: the Insulin Resistance Atherosclerosis Study (IRAS) Family Study. Diabetologia 2009; 52:2079-86., and which lead to common comorbidities, such as metabolic syndrome, cardiovascular and renal diseases, bone fragility, dementia and sleep disorders. These conditions lead to unfavorable health outcomes, such as disabilities and reduction in quality of life ⁸8. Edson EJ, Sierra-Johnson J, Curtis B. Diabetes and obesity in older adults: a call to action. Rev Clin Gerontol 2009; 19:135-47..

Socioeconomic disparities in obesity, as with other health conditions, are well-established in the literature ¹²12. Anselmo O MT, Costa JSD, Kac G, Pattussi MP. Epidemiologia da obesidade abdominal em mulheres adultas residentes no Sul do Brasil. Arch Latinoam Nutr 2007; 57:349-56.^{), (13}13. Barros MBA, Francisco PMSB, Zanchetta LM, César CLG. Trends in social and demographic inequalities in the prevalence of chronic diseases in Brazil. PNAD: 2003-2008. Ciênc Saúde Coletiva 2011; 16:3755-68.^{), (14}14. Lima-Costa MF, De Oliveira C, Macinko J, Marmot M. Socioeconomic inequalities in health in older adults in Brazil and England. Am J Public Health 2012; 102:1535-41.^{), (15}15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.^{), (16}16. Louie GH, Ward MM. Socioeconomic and ethnic differences in disease burden and disparities in physical function in older adults. Am J Public Health 2011; 101:1322-9.. Race and ethnicity, demographic characteristics which are also widely correlated with socioeconomic status ¹⁷17. Araújo ED, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. The use of the variable race/color within public health: possibilities and limits. Interface Comunic Saúde Educ 2009; 13:383-94.^{), (18}18. Perreira KM, Telles EE. The color of health: skin color, ethnoracial classification, and discrimination in the health of Latin Americans. Soc Sci Med 2014; 116:241-50.^{), (19}19. Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30:1438-52.^{), (20}20. Chor D. Desigualdades em saúde no Brasil: é preciso ter raça. Cad Saúde Pública 2013; 29:1272-5., are associated with obesity, evaluated by anthropometric indicators of general (BMI) ⁷7. Silva VS, Souza I, Petroski EL, Silva DAS. Prevalence and factors associated with overweight in Brazilian elderly. Rev Bras Ativ Fís Saúde 2011; 16:289-94.^{), (9}9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (21}21. Kelley EA, Bowie JV, Griffith DM, Bruce M, Hill S, Thorpe RJ Jr. Geography, race/ethnicity and obesity among men in the United States. Am J Mens Health 2016; 10:228-36.^{), (22}22. Zhang H, Rodriguez-Monguio R. Racial disparities in the risk of developing obesity-related diseases: a cross-sectional study. Ethn Dis 2012; 22:308-16.^{), (23}23. Ng JH, Bierman AS, Elliot MN, Wilson RL, Xia C, Scholle SH. Beyond black and white: race/ethnicity and health status among older adults. Am J Manag Care 2014; 20:239-48. and central fat (waist circumference - WC and waist-hip ratio - WHR) ⁹9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (24}24. Ferreira MG, Valente JG, Silva RMVG, Sichieri R. Waist circumference and waist-to-hip ratio as indicators of fat location in black, white, and mulatto Brazilian men. Ethn Dis 2007; 17:256-61.^{), (25}25. Castanheira M, Olinto MTA, Gigante DP. Associação de variáveis sócio-demográficas e comportamentais com a gordura abdominal em adultos: estudo de base populacional no Sul do Brasil. Cad Saúde Pública 2003; 19 Suppl 1:S55-65..

The relationship between general and central fat with race and ethnicity is not yet clear ⁷7. Silva VS, Souza I, Petroski EL, Silva DAS. Prevalence and factors associated with overweight in Brazilian elderly. Rev Bras Ativ Fís Saúde 2011; 16:289-94.^{), (9}9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (24}24. Ferreira MG, Valente JG, Silva RMVG, Sichieri R. Waist circumference and waist-to-hip ratio as indicators of fat location in black, white, and mulatto Brazilian men. Ethn Dis 2007; 17:256-61.^{), (25}25. Castanheira M, Olinto MTA, Gigante DP. Associação de variáveis sócio-demográficas e comportamentais com a gordura abdominal em adultos: estudo de base populacional no Sul do Brasil. Cad Saúde Pública 2003; 19 Suppl 1:S55-65.. Some studies show associations with high BMI and WC values in white men ²¹21. Kelley EA, Bowie JV, Griffith DM, Bruce M, Hill S, Thorpe RJ Jr. Geography, race/ethnicity and obesity among men in the United States. Am J Mens Health 2016; 10:228-36.^{), (24}24. Ferreira MG, Valente JG, Silva RMVG, Sichieri R. Waist circumference and waist-to-hip ratio as indicators of fat location in black, white, and mulatto Brazilian men. Ethn Dis 2007; 17:256-61.^{), (25}25. Castanheira M, Olinto MTA, Gigante DP. Associação de variáveis sócio-demográficas e comportamentais com a gordura abdominal em adultos: estudo de base populacional no Sul do Brasil. Cad Saúde Pública 2003; 19 Suppl 1:S55-65. while in other studies, obesity is associated with women or individuals of both genders and black ⁹9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (15}15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.^{), (22}22. Zhang H, Rodriguez-Monguio R. Racial disparities in the risk of developing obesity-related diseases: a cross-sectional study. Ethn Dis 2012; 22:308-16.^{), (23}23. Ng JH, Bierman AS, Elliot MN, Wilson RL, Xia C, Scholle SH. Beyond black and white: race/ethnicity and health status among older adults. Am J Manag Care 2014; 20:239-48. and Hispanic ²²22. Zhang H, Rodriguez-Monguio R. Racial disparities in the risk of developing obesity-related diseases: a cross-sectional study. Ethn Dis 2012; 22:308-16.^{), (23}23. Ng JH, Bierman AS, Elliot MN, Wilson RL, Xia C, Scholle SH. Beyond black and white: race/ethnicity and health status among older adults. Am J Manag Care 2014; 20:239-48. race/ethnicity, in comparison with white individuals. Meanwhile, national and international findings have shown associations between diabetes and minority races/ethnicities ²⁶26. World Health Organization. Waist circumference and waist-hip ratio - report of a WHO Expert Consultation. http://whqlibdoc.who.int/publications/2011/9789241501491_eng.pdf (acessado em 29/Ago/2011).
http://whqlibdoc.who.int/publications/20... ^{), (27}27. Noble JM, Manly JJ, Schupf N, Tang MX, Luchsinger JA. Type 2 diabetes and ethnic disparities in cognitive impairment. Ethn Dis 2012; 22:38-44.^{), (28}28. Whitson HE, Hastings SN, Landerman LR, Fillenbaum GG, Cohen JH, Johnson KS. Black-white disparity in disability: the role of medical conditions. J Am Geriatr Soc 2011; 59:844-50.^{), (29}29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7., which is explained by a set of social, economic, biological and environmental factors ²⁹29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7..

Given the scarcity of descriptive data, especially in Brazil, regarding the association between race and obesity among the elderly, including the presence of chronic diseases, among which diabetes, this study sought to: investigate the effect of race on measures of general (BMI) and central fat (WHR and WC) according to gender in a sample of elderly individuals living in urban areas in seven places in Brazil, adjusting for schooling and family income variables; and subsequently evaluate the variation of this effect on the presence of self-reported diabetes.

Materials and methods

Participants

The study sample comprised 2,566 elderly community dwellers aged 65 years or more who lived in the following places in Brazil: Campinas (São Paulo State), Belém (Pará State), Parnaíba (Piauí State), Campina Grande (Paraíba State), Poços de Caldas (Minas Gerais State), Ivoti (Rio Grande do Sul State) and Ermelino Matarazzo district (São Paulo State). This is a cross-sectional study that used secondary data from the multi-center, multidisciplinary population study FIBRA (Frailty in Elderly Brazilians) - Unicamp group. The FIBRA Study sought to evaluate frailty of urban elderly community dwellers (65 years of more) according to socioeconomic, psychological and biological aspects.

The FIBRA sample was selected through a simple random sample of the census sectors of the urban zone of the seven places where data was collected. For each of these places, there were pre-established quotas of the census sectors (90 in Campinas, 93 in Belém, 75 in Poços de Caldas, 62 in Ermelino Matarazzo, 60 in Campina Grande, 60 in Parnaíba and 27 in Ivoti) to be visited by recruiters. The samples should include quotas for men and women and for the age groups 65-69 years, 70-74 years, 75-79 years and 80 years or more, proportionally to the distribution of these segments in the elderly population residing in each urban area. In all places, for each census sector, researchers planned the recruitment of an oversample of 25% of intended elderly participants, respecting the gender and age criteria, to compensate possible losses due to participants not showing up or dropping out at the time of the data collection.

In order to calculate sample size for each city, researchers estimated the sample size needed to get a population proportion of 50% of a given characteristic being studied - the value for which the sample size is the greatest possible (p = 0.50; q = 0.50). The number of census sectors selected through a draw equaled the ratio between the intended number of elderly individuals and the number of urban census sectors present in each city. Researchers established a sample plan that estimated a minimum size of 601 elderly individuals (for a 4% sampling error) for places with more than one million inhabitants (Campinas and Belém) and 384 elderly individuals (for a 5% sampling error) for the other places, with less than one million inhabitants. Ivoti, which had a universe of elderly individuals of 646, was the exception, with a sample estimated at 235 elderly individuals (for a 5% sampling error).

All participants understood instructions, were permanent residents of the household and census sector, and agreed to participate in the research, signing an Informed Consent form. Exclusion criteria were the same used in the Cardiovascular Health Study³⁰30. Ferrucci L, Guralnik JM, Studenski S, Fried LP, Cutler Jr GB, Walston JD, et al. Designing randomized, controlled trials aimed at preventing or delaying functional decline and disability in frail, older persons: a consensus report. J Am Geriatr Soc 2004; 52:625-34.. Data collection took place in the communities, in sites to which participants traveled by their own means. More details on sampling and participant recruitment processes for the FIBRA study can be found in a previously-published article ³¹31. Pinto JM, Neri AL. Factors associated with low life satisfaction in community-dwelling elderly: FIBRA Study. Cad Saúde Pública 2013; 29:2447-58..

The State University of Campinas Ethical Review Board approved this study, under the protocol n. 208/2007, according to the demands and procedures established in the National Health Council's Resolution n. 466/12, which regulates research involving humans ³²32. Conselho Nacional de Saúde. Resolução nº 466, de 12 de dezembro de 2012. http://www.conselho.saude.gov.br/resolucoes/2012/Reso466.pdf (acessado em 15/Dez/2015).
http://www.conselho.saude.gov.br/resoluc... .

Instruments and measures

Demographic and socioeconomic data regarding gender, age, race, schooling and family income were obtained through self-reports (date of birth, male vs female gender, years of schooling and family income in gross values). Age, schooling and income variables were grouped in the following categories, respectively: 65-69, 70-74, 75-79, ≥ 80 years; never attended school, 1-4 years, ≥ 5 years; 0.0-1.0, 1.1-3.0, 3.1-5.0, ≥ 5.1 minimum wage. Race is based on the criteria established by the IBGE ² and was self-reported by participants. In this study, we only included white, brown and black races because other races (indigenous and yellow) were a minority in the sample.

Anthropometric data were collected by trained examiners according to World Health Organization (WHO) recommendations ³³33. World Health Organization. Physical status: the use and interpretation of anthropometry - Report of a WHO Expert Committee. http://whqlibdoc.who.int/trs/WHO_TRS_854.pdf?ua=1 (acessado em 05/Mai/2007).
http://whqlibdoc.who.int/trs/WHO_TRS_854... . Participants were weighed using a digital scale (manufactured by G-Tech. Accumed-Glicomed, Rio de Janeiro, Brazil) and height was measured through a scale (200cm) graduated in centimeters and millimeters. BMI was calculated using the equation: BMI = weight (kg) ÷ height² (m) and classified according to criteria established by the Pan American Health Organization (PAHO) for the elderly ³⁴34. Organización Panamericana de la Salud. XXXVI Reunión del Comité Asesor de Investigaciones en Salud. Encuesta Multicéntrica - Salud Bienestar y Envejecimiento (SABE) en América Latina y el Caribe. Informe preliminar. http://www.opas.org/program/sabe.htm (acessado em 05/Jun/2007).
http://www.opas.org/program/sabe.htm... (underweight < 23; normal weight ≥ 23 and < 28; overweight ≥ 28 and < 30; obesity ≥ 30). WC and hip circumference (HC) were measured with an inelastic millimetered tape (150cm long). With these measures, we calculated WHR (WC ÷ HC), classified according to metabolic risk, according to Bray & Gray's ³⁵35. Bray GA, Gray DS. Obesity. Part I - pathogenesis. Western J Med 1988; 149:429-41. recommendations (risk for men and women, respectively: < 0.91 and < 0.76 low; 0.91-0.98 and 0.76-0.83: moderate; > 0.98 and > 0.83: high/very high). WC was classified according to values suggested by the WHO (risk for men and women, respectively: ≥ 94 and ≥ 80: increased; ≥ 102 and ≥ 88: substantially increased) ²⁶26. World Health Organization. Waist circumference and waist-hip ratio - report of a WHO Expert Consultation. http://whqlibdoc.who.int/publications/2011/9789241501491_eng.pdf (acessado em 29/Ago/2011).
http://whqlibdoc.who.int/publications/20... .

Researchers carried out the Mini-Mental State Examination (MMSE) before collecting data referring to physical health, including self-reported diabetes. Elderly individuals with a lower score than the cut-off score for their level of schooling were excluded from the study (cut-off scores established by Brucki et al. ³⁶36. Brucki SMD, Nitrini R, Caramelli P, Bertolucci PHF, Okamoto IH. Sugestões para o uso do Mini-Exame do Estado Mental no Brasil. Arq Neuropsiquiatr 2003; 61:777-81., minus one standard deviation: 17 for illiterate individuals; 22 for 1 to 4 years; 24 for 5 to 8 years; 26 for 9 or more years of schooling).

Diabetes was evaluated through the following dichotomous self-report question: "has a doctor ever told you that you have the following diseases?", with diabetes listed as one of the chronic diseases that were part of this item.

Some variables had missing data because participants failed to answer questions or it was impossible to carry out the anthropometric measurements. There was a 13.9% information loss for family income; 0.15% for schooling; 0.97% for race; 0.89% for BMI; 2.02% for WC; and 2.06% for WHR.

Statistical analysis

We analyzed the data using the IBM SPSS, version 20 (IBM Corp., Armonk, USA). In order to describe the sample profile, we carried out frequency analyses (absolute frequencies and percentages) with the categorical data and calculated averages and standard deviations for the continuous numerical variables. We carried out a chi-squared test in order to compare race with socioeconomic status, self-reported diabetes and body fat measures.

Since the study sample did not have all the assumptions required for an analysis of covariance (ANCOVA) (normality of dependent variables, homogeneity of variance and homogeneity of slopes), we carried out Quade's nonparametric ANCOVA (following the procedures described by Marôco ³⁷37. Marôco J. Testes não paramétricos para amostras independente. In: Marôco J, organizador. Análise estatística com o SPSS Statistics. 6ª Ed. Pêro Pinheiro: Report Number; 2014. p. 356-61.⁾⁾ in order to verify the effect of race on anthropometric measures of body fat (BMI, WC and WHR as dependent variables), adjusting for schooling and family income (covariables). We carried out these analyses for the entire sample and, later, for the sample of elderly diabetics (all stratified according to gender). For the multiple comparisons of estimated averages (± standard error) among race groups, we used Bonferroni's test. In all analyses, we used a probability of error type I (α) of 5%, that is, p < 0.05.

Results

The sample predominantly comprised women (65.51%), white and brown participants and individuals aged between 65 and 74 years (average age = 72.38 ± 5.58). A significant percentage of individuals reported never having attended school (19.5%) and half of the sample reported having 1 to 4 years of formal education, and family income between 1.1 and 3.0 times the minimum wage (Table 1). The frequency of self-reported diabetes was 19.4% among men and 21.7% among women.

When we classified participants according to race (Table 2), there was a greater proportion of illiterate individuals among black and brown participants (X²(4) = 23.78; p < 0.001), and wealthier white individuals (X²(6) = 34.29; p < 0.001). White women had higher schooling (X²(4) = 54.35; p < 0.001) and were wealthier (X²(6) = 59.32; p < 0.001) than women of other races.

There was an association with higher abdominal fat among white men (in comparison with black and brown men), represented by WC (X²(4) = 16.62; p = 0.002) and by WHR (X²(4) = 15.82; p = 0.003). Adjusting for schooling and family income, we found the same effect of white race on higher WC and WHR, in comparison with brown (p = 0.001) and black and brown individuals (p < 0.001) respectively (Table 3), only among men. However, among women, we found no variation of any anthropometric measure according to race.

For men, diabetes frequency did not vary according to race, while among women, diabetes was more frequent among brown and black women, when compared with white women (X²(2) = 15.40; p < 0.001) (Table 2).

Among diabetics, the effect of race on body fat measures was not present, but it was evident on women's BMI and WC, adjusting for socioeconomic conditions (Table 4). Black diabetic women had the highest BMI (30.30kg/m²⁾ and WC (98.67cm), in comparison with brown women (p = 0.007 and p < 0.001). These values indicate general and, especially, abdominal obesity. Brown diabetic women had the lowest abdominal fat, when compared with the other race groups.

Discussion

This study sought to describe and compare anthropometric measures of body fat according to race among elderly Brazilians, as well as to evaluate the effect of race on these indicators for individuals who self-reported diabetes.

In this sample, there was a high percentage of white men and women, followed by brown men and women. Although race, indicated by self-reported skin color, represents the individual's phenotypic characteristic, is also the result of a sociocultural construction that is dependent on the individual's context ¹⁷17. Araújo ED, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. The use of the variable race/color within public health: possibilities and limits. Interface Comunic Saúde Educ 2009; 13:383-94..

According to Penner & Sapperstein ³⁸38. Penner AM, Saperstein A. How social status shapes race. Proc Natl Acad Sci U S A 2008; 105:19628-30., individuals' perception of self-reported race is fluid and changes over time because it relates, in part, to individuals' social status. The researchers observed that, in a sample of North Americans, those unemployed, incarcerated or poor were more likely to identify as black than as white. A Brazilian study showed that men aged 40 years or older tended to self identify as brown (as opposed to white) when interacting with black interviewers (in comparison with white interviewers). Furthermore, black interviewers (in comparison with white interviewers) were less likely to evaluate black men (aged 40 years or older) as black (as opposed to white). These data suggest that an interaction between age, gender and race of research participants and interviewers may mediate the results obtained in processes for racial classification ³⁹39. Bastos JL. Does the way I see you affect the way I see myself? Associations between interviewers' and interviewees' color/race in southern Brazil. Cad Saúde Pública 2009; 25:2111-4..

However, Fuchs et al. ⁴⁰40. Fuchs SC, Guimarães SM, Sortica C, Wainberg F, Dias KO, Ughini M, et al. Reliability of race assessment based on the race of the ascendants: a cross-sectional study. BMC Public Health 2002; 2:1. considered self-reported race as a reliable and useful measure, especially in epidemiological studies, in addition to being the main measure for evaluating race in Brazilian censuses.

In this study, white race was associated with a high degree of abdominal fat (WC and WHR), when compared with brown and black race, only among men. Similar data were found in other Brazilian studies. Ferreira et al. ²⁴24. Ferreira MG, Valente JG, Silva RMVG, Sichieri R. Waist circumference and waist-to-hip ratio as indicators of fat location in black, white, and mulatto Brazilian men. Ethn Dis 2007; 17:256-61., in a sample of 1,235 also did not find an association between race and IMC, even though study participants were younger (20 to 59 years) than those in our study. However, they found higher values for WC (p < 0.01) and WHR (p = 0.05) among white men over 30 years of age (in comparison with black men), when adjusting for age, body fat percentage, alcohol consumption, smoking, physical activity, income and schooling.

In a study on individuals aged between 20 and 69 years, Castanheira et al. ²⁵25. Castanheira M, Olinto MTA, Gigante DP. Associação de variáveis sócio-demográficas e comportamentais com a gordura abdominal em adultos: estudo de base populacional no Sul do Brasil. Cad Saúde Pública 2003; 19 Suppl 1:S55-65. also found a higher WC among white men than among brown and black men (p < 0.001), but did not find this association among women. According data from the Brazilian Household Budget Survey (POF, in Portuguese), a BMI of ≥ 25kg/m² was associated with white race, in comparison with brown race, among elderly individuals over 60 years of age ⁷7. Silva VS, Souza I, Petroski EL, Silva DAS. Prevalence and factors associated with overweight in Brazilian elderly. Rev Bras Ativ Fís Saúde 2011; 16:289-94..

However, the international literature shows different results than those found in this study. Most show an association between black and Hispanic race and general obesity (BMI) in samples of adult and elderly individuals ⁹9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (21}21. Kelley EA, Bowie JV, Griffith DM, Bruce M, Hill S, Thorpe RJ Jr. Geography, race/ethnicity and obesity among men in the United States. Am J Mens Health 2016; 10:228-36.^{), (22}22. Zhang H, Rodriguez-Monguio R. Racial disparities in the risk of developing obesity-related diseases: a cross-sectional study. Ethn Dis 2012; 22:308-16.^{), (24}24. Ferreira MG, Valente JG, Silva RMVG, Sichieri R. Waist circumference and waist-to-hip ratio as indicators of fat location in black, white, and mulatto Brazilian men. Ethn Dis 2007; 17:256-61.. Data from NHANES show an association between general (BMI) and abdominal obesity (WC) with black race, in comparison with white and Hispanic races, in U.S. elderly women (> 60 years).

The discrepancy may result, first, from different classifications used to define race/ethnicity. This limits this type of investigation and makes data comparison and discussion difficult. Additionally, one must consider differences regarding socioeconomic profile and its implications on lifestyle and eating habits among different races/ethnicities in Brazil and in developed countries.

The literature establishes a strong relationship between a lower socioeconomic status and black or brown race ¹⁷17. Araújo ED, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. The use of the variable race/color within public health: possibilities and limits. Interface Comunic Saúde Educ 2009; 13:383-94. and with worse health conditions ¹³13. Barros MBA, Francisco PMSB, Zanchetta LM, César CLG. Trends in social and demographic inequalities in the prevalence of chronic diseases in Brazil. PNAD: 2003-2008. Ciênc Saúde Coletiva 2011; 16:3755-68., including obesity ¹²12. Anselmo O MT, Costa JSD, Kac G, Pattussi MP. Epidemiologia da obesidade abdominal em mulheres adultas residentes no Sul do Brasil. Arch Latinoam Nutr 2007; 57:349-56.^{), (15}15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.. However, though black and brown elderly individuals in this study had unfavorable schooling and income levels, it was white men who showed the worst metabolic condition, indicated by higher WC values. Monteiro et al. ⁴¹41. Monteiro CA, Conde WL, Popkin BM. Independent effect of income and education on the risk of obesity in the Brazilian adult population. J Nutr 2001; 131:881-6. found that higher income, which would lead to greater consumption of food, was a risk factor for obesity, especially among men.

Despite the fact that this study's analyses were adjusted for schooling and income, environmental factors related to health behaviors, such as physical activity levels, eating habits, smoking and alcohol consumption were not addressed, but could influence the findings regarding racial differences.

Additionally, the evaluation of socioeconomic status over an individual's life course would better explain obesity among adults, as the literature shows. Wealth or poverty during childhood could affect nutritional status and fat deposits later in life ⁴²42. González DA, Nazmi A, Yudkin JS, Victora CG. Life-course socio-economic factors, skin color, and abdominal obesity in adulthood in a Brazilian birth cohort. Public Health Nutr 2009; 12:2225-35.^{), (43}43. Taveras EM, Gillman MW, Kleinman KP, Rich-Edwards JW, Rifas-Shiman SL. Reducing racial/ethnic disparities in childhood obesity: the role of early life risk factors. JAMA Pediatr 2013; 167:731-8.. González et al. ⁴²42. González DA, Nazmi A, Yudkin JS, Victora CG. Life-course socio-economic factors, skin color, and abdominal obesity in adulthood in a Brazilian birth cohort. Public Health Nutr 2009; 12:2225-35. verified that, regardless of family income at the time of the study, men born into families with higher purchasing power had higher WC in adulthood. Thus, this study's findings must be interpreted with caution, since controlling for socioeconomic variables does not completely eliminate their effect on race.

Another important aspect is individuals' caloric output throughout life, particularly that which results from labor and displacement. Evidence shows that men with lower schooling ⁴⁴44. Sávio KEO, Costa THM, Schmitz BAS, Silva EF. Sex, income and level of education associated with physical activity level among workers. Rev Saúde Pública 2008; 42:457-63. and income ⁴⁴44. Sávio KEO, Costa THM, Schmitz BAS, Silva EF. Sex, income and level of education associated with physical activity level among workers. Rev Saúde Pública 2008; 42:457-63.^{), (45}45. Florindo AA, Hallal PC, Moura EC, Malta DC. Practice of physical activities and associated factors in adults, Brazil, 2006. Rev Saúde Pública 2009; 43 Suppl 2:S65-73. were more active in their work and displacements. These individuals are more likely to do informal work (which require greater physical effort), tend to live farther away from work, are more likely to use public transportation and spend more time with desplacement ⁴⁴44. Sávio KEO, Costa THM, Schmitz BAS, Silva EF. Sex, income and level of education associated with physical activity level among workers. Rev Saúde Pública 2008; 42:457-63.. Given that, in our study, black and brown men had lower schooling and income than white men, we may suppose that the higher abdominal fat among white men could be related to better life and work conditions, which would lead to a more sedentary lifestyle and lower energy consumption.

However, in our study, when considering only participants who had self-reported diabetes, we found higher BMI and WC values among black women, when compared with brown women, regardless of economic condition (income) and schooling.

The relationship between abdominal fat and diabetes is well-established in the literature, which shows that obesity (general and central) is a risk factor for diabetes, since it favors a state of chronic inflammation and insulin resistance, contributing to an increased prevalence of the disease, especially among the elderly ⁽⁴4. Mathus-Vlieglen EM. Prevalence, pathophysiology, health consequences and treatment options of obesity in the elderly: a guideline. Obes Facts 2012; 5:460-83.^{), (9}9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (11}11. Hanley AJG, Wagenknecht LE, Norris JM, Bryer-Ash M, Chen YI, Anderson AM, et al. Insulin resistance, beta cell dysfunction and visceral adiposity as predictors of incident diabetes: the Insulin Resistance Atherosclerosis Study (IRAS) Family Study. Diabetologia 2009; 52:2079-86..

The literature has shown an association between a higher prevalence and incidence of diabetes with minority races or ethnic groups, such as black and Hispanic ²³23. Ng JH, Bierman AS, Elliot MN, Wilson RL, Xia C, Scholle SH. Beyond black and white: race/ethnicity and health status among older adults. Am J Manag Care 2014; 20:239-48.^{), (27}27. Noble JM, Manly JJ, Schupf N, Tang MX, Luchsinger JA. Type 2 diabetes and ethnic disparities in cognitive impairment. Ethn Dis 2012; 22:38-44.^{), (28}28. Whitson HE, Hastings SN, Landerman LR, Fillenbaum GG, Cohen JH, Johnson KS. Black-white disparity in disability: the role of medical conditions. J Am Geriatr Soc 2011; 59:844-50.. In a study on 941 elderly men and women, Noble et al. ²⁷27. Noble JM, Manly JJ, Schupf N, Tang MX, Luchsinger JA. Type 2 diabetes and ethnic disparities in cognitive impairment. Ethn Dis 2012; 22:38-44. observed a higher diabetes prevalence among non-white individuals. The prevalence was 19.6% and 20.1% among Hispanic and black individuals, respectively, versus 8.2% among white individuals (p < 0.001).

Likewise, Whitson et al. ²⁸28. Whitson HE, Hastings SN, Landerman LR, Fillenbaum GG, Cohen JH, Johnson KS. Black-white disparity in disability: the role of medical conditions. J Am Geriatr Soc 2011; 59:844-50., in an analysis adjusted for gender and socioeconomic status, showed that elderly black individuals were more obese, had more disabilities and a higher diabetes prevalence, when compared with white individuals.

Racial differences associated with higher morbity ¹⁹19. Oliveira BLCA, Thomaz EBAF, Silva RA. The association between skin color/race and health indicators in elderly Brazilians: a study based on the Brazilian National Household Sample Survey (2008). Cad Saúde Pública 2014; 30:1438-52.^{), (46}46. August KJ, Sorkin DH. Racial and ethnic disparities in indicators of physical health status: do they still exist throughout late life? J Am Geriatr Soc 2010; 58:2009-15^{), (47}47. Warner DF, Brown TH. Understanding how race/ethnicity and gender define age-trajectories of disability: an intersectionality approach. Soc Sci Med 2011; 72:1234-48., obesity ⁹9. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007; 29:6-28.^{), (15}15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.^{), (22}22. Zhang H, Rodriguez-Monguio R. Racial disparities in the risk of developing obesity-related diseases: a cross-sectional study. Ethn Dis 2012; 22:308-16.^{), (29}29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7. and disability ¹⁶16. Louie GH, Ward MM. Socioeconomic and ethnic differences in disease burden and disparities in physical function in older adults. Am J Public Health 2011; 101:1322-9.^{), (28}28. Whitson HE, Hastings SN, Landerman LR, Fillenbaum GG, Cohen JH, Johnson KS. Black-white disparity in disability: the role of medical conditions. J Am Geriatr Soc 2011; 59:844-50. are largely explained by unfavorable socioeconomic conditions, with lead to an accumulation of unhealthy behaviors and lifestyle, due to lower access to information, quality education and health services. According to Chor ²⁰20. Chor D. Desigualdades em saúde no Brasil: é preciso ter raça. Cad Saúde Pública 2013; 29:1272-5., race, socioeconomic status and gender are crucial variables that must be analyzed together, given their inter-relations, which influence disparities and create vulnerabilities to health risks.

The higher obesity among diabetic black elderly women found in our study may have been influenced by unfavorable conditions faced by this group over their lives ¹⁵15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.^{), (47}47. Warner DF, Brown TH. Understanding how race/ethnicity and gender define age-trajectories of disability: an intersectionality approach. Soc Sci Med 2011; 72:1234-48., even in cases in which they reached higher levels of schooling ¹⁷17. Araújo ED, Costa MCN, Hogan VK, Araújo TM, Dias AB, Oliveira LOA. The use of the variable race/color within public health: possibilities and limits. Interface Comunic Saúde Educ 2009; 13:383-94., due to reduced opportunities for upward social mobility, in addition to differences and worse care in health services, a reflex of racial discrimination and social exclusion ²⁰20. Chor D. Desigualdades em saúde no Brasil: é preciso ter raça. Cad Saúde Pública 2013; 29:1272-5.^{), (47}47. Warner DF, Brown TH. Understanding how race/ethnicity and gender define age-trajectories of disability: an intersectionality approach. Soc Sci Med 2011; 72:1234-48.. Cunningham et al. ⁴⁸48. Cunningham TJ, Berkman LF, Kawachi I, Jacobs Jr. DR, Seeman TE, Kiefe CI, et al. Changes in waist circumference and body mass index in the US CARDIA cohort: fixed-effects associations with self-reported experiences of racial/ethnic discrimination. J Biosoc Sci 2013; 45:267-78. showed that, over an eight year period, there was a significant increase in WC and BMI among black women who reported greater racial discrimination (the same was not found for black and white men or for white women).

Authors further emphasize that racial differences in obesity ¹⁵15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86. and diabetes ²⁹29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7. result from a complex interaction that encompasses, beyond socioeconomic and environmental factors (lifestyle), biological/physiological conditions that include: lower resting and total energy output among black individuals (especially women) than in white individuals ²⁹29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7.; lower levels of adiponectins (associated with higher body fat and metabolic syndrome) in black individuals (versus white individuals) ¹⁵15. Agyemang P, Powell-Whiley TM. Obesity and black women: special considerations related to genesis and therapeutic approaches. Curr Cardiovasc Risk Rep 2013; 7:378-86.; in addition to a possible predisposition of ethnic minorities to insulin resistance ²⁹29. Staiano AE, Harrington DM, Johannsen NM, Newton Jr LR, Sarzynski MA, Swift DL, et al. Uncovering physiological mechanisms for health disparities in type 2 diabetes. Ethn Dis 2015; 25:31-7.^{), (49}49. Tillin T, Hughes AD, Godsland IF, Whincup P, Forouhi NG, Welsh P, et al. Insulin resistance and truncal obesity as important determinants of the greater incidence of diabetes in Indian Asians and African Caribbeans compared with Europeans. Diabetes Care 2013; 36:383-93.^{), (50}50. Shai I, Jiang R, Manson JE, Stampfer MJ, Willet WC, Colditz GA, et al. Ethnicity, obesity and risk of type 2 diabetes in women. Diabetes Care 2006; 29:1585-90.. However, further studies are necessary in order to clarify these associations.

It is worth noting that brown elderly diabetic women had lower WC values, particularly when compared with white women (who were also diabetic), which may be partly explained by lower income and schooling among brown women, which would lead them to have a higher energy output, due to greater use of public transportation or greater workload of domestic work than wealthier women ⁴⁵45. Florindo AA, Hallal PC, Moura EC, Malta DC. Practice of physical activities and associated factors in adults, Brazil, 2006. Rev Saúde Pública 2009; 43 Suppl 2:S65-73.. However, this argument is not valid when comparing black and brown women, since both groups have unfavorable socioeconomic conditions.

This study has several limitations. First, the cross-sectional design does not enable us to establish a causality relationship between the variables. Another limitation is the use of self-reports to evaluate diabetes mellitus, which may lead to an underestimation of the disease's prevalence, since some elderly individuals may be unaware of their diagnosis. In spite of this, in an investigation carried out on 10,321 individuals (average age = 63 years), use of self-reported diabetes had high validity and specificity in the identification of its prevalence and incidence, when compared with standard recommendations, based on fasting plasma glucose levels, glycosylated hemoglobin (HbA1c) and medication use ⁵¹51. Schneider ALC, Pankow JS, Heiss G, Selvin E. Validity and reliability of self-reported diabetes in the Atherosclerosis Risk in Communities Study. Am J Epidemiol 2012; 176:738-43.. The method used to classify race may also be considered a limitation, given the subjective and dynamic nature of this criterion based on self-reported skin color. However, we emphasize this study's relevance, which established a profile of relationship between these variables in a sample of elderly Brazilians, data that is still lacking in the literature.

Conclusion

In a sample of elderly Brazilians, we identified the effect of white race on higher WC and WHR values only for men, regardless of the presence of self-reported diabetes. However, when considering only diabetic individuals, black race became associated with general (BMI) and central obesity (WC), when compared with brown race, only among women. The effect of race on body fat measures was present even when adjusting for socioeconomic variables (schooling and family income).

These data show the importance of public policies that include individualized strategies that are directed at the specificities of each racial group. Public policies must seek to promote health, considering the prevention of obesity as the main risk factor for diabetes, as well as the adequate management of these conditions, with a focus on individuals' function and quality of life.

Acknowledgments

We thank CNPq (n. 555082/2006-7) and Capes (n. BEX12339/13-0)

References

Publication Dates

History