Hostname: page-component-7c8c6479df-r7xzm Total loading time: 0 Render date: 2024-03-28T05:17:06.980Z Has data issue: false hasContentIssue false

Life-course socio-economic factors, skin colour and abdominal obesity in adulthood in a Brazilian birth cohort

Published online by Cambridge University Press:  06 August 2009

David A González*
Affiliation:
Post-graduate Program in Epidemiology, Federal University of Pelotas, Brazil, Rua Marechal Deodoro 1160, Pelotas, RS, CEP 96020-220, Brazil
Aydin Nazmi
Affiliation:
Department of Food Science and Nutrition, California Polytechnic State University, San Luis Obispo, CA, USA
John S Yudkin
Affiliation:
University College London, London, UK
Cesar G Victora
Affiliation:
Post-graduate Program in Epidemiology, Federal University of Pelotas, Brazil, Rua Marechal Deodoro 1160, Pelotas, RS, CEP 96020-220, Brazil
*
*Corresponding author: Email david.epidemio@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

Objective

Obesity is an increasingly prevalent nutritional disorder throughout the world. In particular, abdominal obesity is associated with cardiovascular and metabolic risk. The present study aimed to evaluate the effects of skin colour and life-course socio-economic indicators on waist circumference (WC), hip circumference (HC) and waist:hip ratio (WHR) in young adults.

Design

Population-based birth cohort study. Individuals born in 1982 in Pelotas (southern Brazil) were visited on a number of occasions from birth to age 23–24 years. A sample of the cohort was sought in 2006 and 972 individuals were located. The analysis was restricted to individuals with complete data available (442 males, 414 females).

Results

In men, family income at birth and in 2004–5 were positively associated with WC and HC, but not with WHR. Regardless of current income, men born to wealthier families had larger WC and HC as adults. Skin colour was not associated with any of the outcomes. In women, early poverty was associated with smaller HC, and current poverty with larger WC. Poverty at any age thus led to higher WHR. Black women had larger WC and HC than white women, but there were no differences in WHR. All the associations were partially mediated by education and behavioural variables.

Conclusions

The effects of early socio-economic position on WC and HC persist even after adjustment for adult socio-economic position, highlighting the importance of interventions during the first years of life.

Type
Research Paper
Copyright
Copyright © The Authors 2009

Obesity is a key risk factor for the incidence of and mortality associated with CVD, diabetes, hypertension and hypercholesterolaemia(Reference de Koning, Merchant, Pogue and Anand1Reference Yusuf, Hawken and Ounpuu5). Over the past three decades, its prevalence has increased rapidly throughout the world in both sexes(Reference Jorgensen, Borch-Johnsen and Bjerregaard6Reference Monteiro, Moura, Conde and Popkin8), but men and women are differentially affected according to their socio-economic position (SEP). In middle-income countries, obesity is more prevalent in richer men and poorer women(Reference Monteiro, Conde and Popkin7, Reference Monteiro, Moura, Conde and Popkin8).

Obesity in adults is usually evaluated by BMI(3). However, BMI does not provide information on fat distribution, which is a better indicator of CVD risk(Reference Iacobellis and Sharma9). There have been calls to use waist:hip ratio (WHR) instead of other anthropometric measures for assessing CVD risk(Reference de Koning, Merchant, Pogue and Anand1, Reference Lanas, Avezum, Bautista, Diaz, Luna, Islam and Yusuf2, Reference Yusuf, Hawken and Ounpuu5). A meta-regression published in 2007 showed that WHR was more strongly associated with CVD than waist circumference (WC) among 4355 incident cases(Reference de Koning, Merchant, Pogue and Anand1). The INTERHEART Study evaluated 12 461 cases of acute myocardial infarction in fifty-two countries, confirming these findings even among individuals with normal or low BMI (BMI < 25 kg/m2)(Reference Yusuf, Hawken and Ounpuu5). It is important to study WC and hip circumference (HC) separately because the adverse effects of WHR on CVD could be due to either a relatively large WC or a small HC(Reference Seidell, Perusse, Despres and Bouchard10, Reference Snijder, Dekker, Visser, Bouter, Stehouwer, Yudkin, Heine, Nijpels and Seidell11). Furthermore, both measures represent two distinct time periods of growth and may be determined by different factors(Reference Sinclair and Dangerfield12Reference Taylor, Grant, Goulding and Williams14).

Nearly all studies evaluating socio-economic risk factors for obesity in adults are cross-sectional(Reference Monteiro, Moura, Conde and Popkin8, Reference McLaren15, Reference Sobal and Stunkard16). However, there is evidence that early poverty affects nutritional status in adult life(Reference Barros, Victora, Horta, Goncalves, Lima and Lynch17Reference Galobardes, Lynch and Davey Smith19). Studies evaluating the long-term effects of early SEP on abdominal obesity in adults are rare and principally limited to data from high-income countries; these report inverse associations in women and variable patterns for men(Reference Dekkers, Podolsky, Treiber, Barbeau, Gutin and Snieder18, Reference Kivimaki, Smith and Juonala20Reference Schooling, Jiang, Lam, Zhang, Cheng and Leung27). Research in this area is particularly relevant for low- and middle-income settings to better understand the effects of rapid nutritional transition(Reference Monteiro, Moura, Conde and Popkin8).

Studies examining the effects of ethnicity or skin colour have also been exclusive to high-income countries and have shown conflicting results(Reference Dekkers, Podolsky, Treiber, Barbeau, Gutin and Snieder18, Reference Cossrow and Falkner28Reference Hill, Sidney, Lewis, Tolan, Scherzinger and Stamm30). A recent literature review indicated that variations in body composition between blacks and whites are the result of a complex multifactorial entanglement of lifestyle, environmental and genetic differences(Reference Wagner and Heyward31), but more studies are required to clarify the mechanisms involved.

Considering the limited literature on the long-terms effects of SEP and ethnicity on body composition, the current study evaluated the influence of family income and skin colour on WC, HC, and WHR at age 23–24 years in a population-based birth cohort. Special attention was given to socio-economic trajectories from birth to adulthood and the possible mediating roles of behavioural variables.

Methods

Pelotas is a southern Brazilian city with a population of 340 000. In 2002 the city had a per capita Gross Domestic Product of $US 1958 (national mean $US 2604)(32). The main ethnic groups include European and African descendents, with a small indigenous Brazilian population. The 1982 Pelotas birth cohort study included 5914 live-born babies (99·2 % of all births) to mothers who lived in the urban area of the city. Mothers were weighed and interviewed in the hospital and babies were examined after birth and subsequently followed up on numerous occasions from infancy until the age of 23–24 years. Details regarding follow-up visits are shown in Fig. 1 and described elsewhere(Reference Victora and Barros33).

Fig. 1 Follow-up visits and variables used from the 1982 Pelotas birth cohort study

In 1997, seventy of the city’s 265 census tracts were randomly selected and all households in each sector were visited, leading to 1076 cohort members being located (follow-up rate 72 %). Between October 2004 and August 2005, the entire cohort was sought through a census of all 98000 urban households (follow-up rate 77 %). Information on socio-economic and health variables was collected. Between January and April 2006, the 1076 cohort members examined in 1997 were sought. Anthropometric, behavioural and health variables were collected on 972 subjects (90·3 % of those studied in 1997). Assessment of the three dependent variables (WC, HC and WHR) was carried out in 856 subjects (88·1 %; see Table 1). For the measurements, interviewers were trained and standardized. WC and HC were taken with individuals in standing position using a flexible 160 cm (precision: 1 mm) fibreglass measuring tape. The waist was considered the narrowest part of the trunk and was taken directly on the skin. The hip was measured at the widest point of the buttocks while wearing a gown over underwear, in order to protect privacy. A linear regression equation was used to correct the gowned measures (hipcorrected = −0·978 + 1·002 × gowned measure) based on a validation study undertaken in 105 volunteers (fifty-seven men, forty-eight women). Quality control measures included repeating 10 % of interviews during field work by a supervisor, showing over 95 % repeatability.

Table 1 Comparison of demographic and socio-economic characteristics at birth between cohort members measured and not measured in 2006 (n 856): 1982 Pelotas birth cohort study, Brazil

*All of the variables included missing data.

†Number measured for waist and hip in 2006 as a percentage of those studied at each follow-up (allowing for selection of census tracts in 1997).

χ 2test for heterogeneity.

The independent variables were collected at the different follow-up visits (Fig. 1). Self-reported skin colour was collected in 2004 in five groups according to the classification of the Brazilian Institute of Geography and Statistics (white, black, brown, yellow and indigenous)(Reference Travassos and Williams34). In Brazil, self-reported skin colour is widely used as a proxy for ethnic background. This variable was categorized as white, black/brown or other. The last group included twenty-five individuals who described themselves as ‘yellow’ or ‘indigenous’. Due to the small number of individuals in this group resulting in imprecise estimates, they are not presented as a separate category in the analyses, but they were retained in adjusted analyses to avoid reducing the study power.

The main SEP indicator used in the analyses was family income collected in 1982 and 2004. This variable reflects access to essential resources, including food, quality education and health care(Reference Sobal35). In 1982, 21·9 % of all families earned up to one minimum wage ($US 50/month) which places them well below the poverty line. In 2004, the incomes of all family members were summed, including the cohort member if currently employed, and only 5·7 % of all families had an income up to one minimum wage ($US 180/month). To ensure comparability between both periods the variable was divided into tertiles. Details on how the income variables were collected and categorized are available elsewhere(Reference Barros, Victora, Horta, Goncalves, Lima and Lynch17). Family income change was classified as: (i) always poor (bottom tertile of family income at birth and at age 23–24 years); (ii) never poor (top two tertiles in both periods); (iii) poor–not poor (bottom tertile at birth and top two tertiles at age 23–24 years); and (iv) not poor–poor (top two tertiles at birth and bottom tertile at age 23–24 years). All analyses were stratified by sex.

Own education, behavioural variables and parity (women) were collected in 2004. Dichotomous variables included current smoking (≥1 cigarette/d), sedentary behaviour (moderate physical activity <150 min/week) and low fibre intake. Ordinal variables were used to describe fat intake (very low, low, American diet, high, very high), alcohol consumption (non-drinker, up to 1 unit/d, >1 unit/d), attained education of the individual (≤7, 8–11, ≥12 years) and parity (0, 1, 2, ≥3). Details on the collection and classification of these variables are available elsewhere(Reference Nazmi, Oliveira and Victora36).

The adjusted analyses took into account the different levels of determination. The most distal determinants were skin colour and family income at birth, which were adjusted for one another (model 1). The next level included family income in 2004, that was adjusted for skin colour and income at birth (model 2). The family income change variable was adjusted solely for skin colour. To investigate possible mediating effects, all of the above variables were adjusted for current behavioural variables and attained education of the individual (model 3). Finally, these analyses were repeated with additional adjustment for concurrent BMI (model 4) to assess the effect of each explanatory variable on WC, HC and WHR independently of overall fatness. Variables were dropped from the model when their P value was ≥0·20.

ANOVA was used in crude analyses and multiple linear regression in adjusted analyses. Regression coefficients (β) and standard errors of the differences relative to the reference category (sed) are expressed in centimetres for WC and HC and as a percentage for WHR (β or sed× 100). Tests for linear trend were used for ordinal variables. The STATA statistical software package version 9·0 (Statacorp, College Station, TX, USA) was used for analysis.

A posteriori calculations of statistical power showed that the family income change variable had the least power. In men, a difference of means between any two categories equal to or greater than 4·0 cm in WC, 3·4 cm for HC and 2·0 % for WHR was detectable with 80 % power. Corresponding values for women were 3·8 cm, 3·9 cm and 2·0 %. For all other explanatory variables, detectable differences were smaller than these.

The Federal University of Pelotas Ethical Committee approved all phases of the 1982 Pelotas birth cohort study. Verbal informed consent was obtained in 1982 and written informed consent was collected in more recent follow-ups.

Results

In 2006, 856 individuals had their waist and hip measurements taken (58·4 % of the original cohort; 79·6 % of those interviewed in 1997). The sample had higher family income at birth compared with the original cohort, but there were no differences with respect to sex, skin colour, maternal education or birth weight (Table 1). Pregnant and postpartum women at the time of the measurements were excluded (n 24).

Among men, mean values were 23·8 kg/m2 for BMI, 82·0 cm for WC, 97·6 cm for HC and 0·84 for WHR. Corresponding values for women were 23·4 kg/m2, 74·2 cm, 98·7 cm and 0·75. All variables were symmetrically distributed. Elevated WC (≥94 cm and ≥80 cm, respectively) was found in 12·0 % of men and in 8·2 % of women(3).

Table 2 shows the crude associations between the explanatory variables and WC, HC and WHR. All associations varied according to sex (P for interaction ≤0·05 for all). In men, there were no significant differences in WC or HC according to skin colour. Family income at birth and in adulthood were directly associated with WC and HC. Both measures were highest among men who were never poor and lowest among those who were always poor. None of the explanatory variables were associated with WHR.

Table 2 Descriptive analysis of waist circumference (WC), hip circumference (HC), waist:hip ratio (WHR) and BMI at 23 years according to independent variables and stratified by sex: 1982 Pelotas birth cohort study, Brazil

*Excludes pregnant women (n 24).

†Test for heterogeneity.

‡Test for trend.

In women, WC and WHR were higher in black/brown compared with whites, but HC did not vary by skin colour. Family income in 1982 and 2004 were inversely associated with WC and WHR, but not with HC. Women in the always poor and in the not poor–poor groups had greater WC compared with the other groups (P of the overall difference = 0·06). Women who were poor at birth had lower HC, especially in the poor–not poor group (P = 0·05). Women who were always poor had higher WHR than those who were never poor.

Table 3 shows the adjusted analyses for men. Skin colour was not associated with any of the outcomes. The direct associations between family income at birth and both circumferences persisted when adjusted for skin colour (model 1). Further adjustment for family income in 2004 reduced the differences but the associations remained significant (model 2). The coefficients were little affected when adjusted for behavioural variables (model 3). Further adjustment for BMI (model 4) led to a reduction in the coefficients and their standard errors, but the association between early income and both circumferences remained highly significant. WHR was not associated with family income at birth in any level.

Table 3 Unadjusted and adjusted analyses for waist circumference (WC), hip circumference (HC) and waist:hip ratio (WHR) at 23 years in men according to past and concurrent socio-economic variables: 1982 Pelotas birth cohort study, Brazil

β, regression coefficient; sed, standard error of the difference.

Adjusted for: afamily income at birth; bskin colour; cfamily income in adulthood; down education, physical inactivity, alcohol consumption, smoking, fibre and fat intake; ecurrent BMI.

*Wald test for heterogeneity.

†Wald test for trend.

Family income in 2004 remained associated with an increase of approximately 4 cm in both circumferences when adjusted for skin colour and family income at birth (model 2). When behavioural variables were included, these differences fell for HC but not for WC, and both associations remained significant (model 3). The association was made null with further adjustment for BMI (model 4). For WHR, there was no dose–response association with family income in 2004; only model 3 was significant, but the relevance of this association is unclear.

For family income change both circumferences were about 5 cm larger for men who were never poor compared with those always poor, even after adjustment for skin colour and behavioural variables. There was a substantial reduction after adjustment for BMI, but the association remained significant. WHR was slightly higher among men who were never poor compared with the poor–not poor, but only after adjustment for behavioural variables (model 3); the relevance of this finding is unclear.

Summing up the results for men, both past and concurrent family income were positively associated with WC and HC. These associations were partially mediated by behavioural variables. BMI partially explained the effects of family income at birth and fully explained the effect of the concurrent family income. Whereas both circumferences increased proportionally to income, WHR remained essentially constant. Skin colour showed no associations with the outcome variables.

Table 4 shows the multivariate analyses for women. After adjustment for family income at birth (model 1), WC and HC were higher in black women compared with white women. The associations remained significant after adjustment for concurrent family income (model 2) and behavioural variables (model 3), but not when current BMI was included (model 4). The association between skin colour and WHR disappeared with adjustment for family income at birth (model 1).

Table 4 Unadjusted and adjusted analyses for waist circumference (WC), hip circumference (HC) and waist:hip ratio (WHR) at 23 years in women according to past and concurrent socio-economic variables: 1982 Pelotas birth cohort study, Brazil

β, regression coefficient; sed, standard error of the difference.

Adjusted for: afamily income at birth; bskin colour; cfamily income in adulthood; down education, physical inactivity, alcohol consumption, smoking, fibre and fat intake; ecurrent BMI.

*Wald test for heterogeneity.

†Wald test for trend.

The inverse association between family income at birth and WC lost significance after adjustment for skin colour (model 1). In all four models, HC was larger in the top two tertiles of family income at birth than in the lowest tertile. For WHR, the inverse association with family income at birth remained significant when adjusted for skin colour (model 1); and the association became progressively weaker when adjusted for mediating variables (models 2 to 4).

The inverse association between family income in 2004 and WC showed little change in models 2 and 3, but disappeared after adjustment for BMI (model 4). The association with HC was not significant in any of the models. Given its effect on WC, family income in 2004 was inversely associated with WHR in all analyses.

In terms of family income trajectory, WC tended to be smaller among women who were not poor as adults, but the differences were not quite significant (models 2 and 3). This effect disappeared after adjustment for BMI (model 4). HC was approximately 2–3 cm larger among women who were not poor at birth (never poor and not poor–poor groups) as shown in models 2 to 4. As a consequence, women who were never poor had lower WHR than other groups in all the models.

All of the analyses in women were repeated including parity as a possible mediator, but there were no noteworthy changes in any of the associations. Additional analyses restricted to nulliparous women also showed similar results.

Summarizing the results for women, early poverty was associated with smaller hips and current poverty with larger waists. Poverty at any age, therefore, led to higher WHR. Black women had higher WC and HC than white women, and consequently there were no differences in WHR. Behavioural variables played relatively weak mediating roles, but current BMI seemed to mediate most of the effects of current income.

Maternal (1982) and own education (2004) were also evaluated as independent variables, but they were not associated with the outcomes when adjusted for income and skin colour (results not shown).

Discussion

The present study aimed to investigate the influence of early and concurrent family income, socio-economic trajectory and skin colour on WC, HC and WHR in young adults. There were three differences between the sexes with respect to SEP. First, family income at birth was directly associated with WC in men, but not in women. Second, WC and HC were larger in men with high current family income, but the opposite was true in women, especially for WC. Finally, family income change had a clear cumulative effect in men (both circumferences were larger among the never poor group) but not in women (HC was directly associated with family income at birth whereas WC was inversely associated with current family income). The only consistent finding in both sexes was the direct association between family income at birth and HC. For skin colour, black/brown women had larger WC and HC than whites, but no differences in WHR were found. In men none of the measurements were related to skin colour.

The long-term effects of early SEP on abdominal obesity may be better understood by considering tissue growth. WC is affected principally by changes in subcutaneous and visceral adipose tissue. On the other hand, HC can vary according to change in fat, bone and muscle. All these tissues represent distinct periods of growth. Adipogenesis begins in utero and the number of adipocytes is determined during this period(Reference Tanner13, Reference Fonseca-Alaniz, Takada, Alonso-Vale and Lima37). After birth there are three periods with relatively fast physiological fat mass increases: (i) the first sixth months of life; (ii) during the adiposity rebound (5–6 years); and (iii) puberty(Reference Sinclair and Dangerfield12Reference Taylor, Grant, Goulding and Williams14). Hip bone formation is determined in the fetal period via hormonal, nutritional and mechanical factors, the latter two exerting effects until the first year of life(Reference Tanner13, Reference Cooper, Westlake, Harvey, Javaid, Dennison and Hanson38, Reference Sizínio and Xavier39). During puberty, hormonal factors (oestrogen) become influential again, determining morphology in both sexes(Reference Tanner13, Reference Sizínio and Xavier39). Muscle tissue is also formed early in life but changes in muscle volume begin at puberty due to hormonal and nutritional factors and physical activity(Reference Sinclair and Dangerfield12Reference Taylor, Grant, Goulding and Williams14). In adulthood fat tissue in the abdomen accumulates progressively, and after 60 years of age a decrease of muscle mass in the hip is found(Reference Taylor, Grant, Goulding and Williams14, Reference Dobbelsteyn, Joffres, MacLean and Flowerdew40Reference Zafon42). Consequently, there are critical periods for tissues growth – fetal life, early childhood and puberty – that are susceptible to adverse environmental conditions(Reference Power and Parsons43).

In men, our findings are consistent with the observation that the wealthier are more likely to be overweight in middle-income countries(Reference Monteiro, Moura, Conde and Popkin8, Reference McLaren15). A possible explanation for the association between family income at birth and WC is that wealth affects early-life fat deposition, with an independent and cumulative effect of adult poverty on abdominal adiposity. The same was observed for HC, which is partly reflective of fat tissue. HC in men was affected by current income independently of early childhood income, suggesting that socio-economic trajectories can also influence hip growth in late childhood and adolescence. Two Brazilian studies found the same direct relationships between SEP in childhood and BMI in 18- to 19-year-old men(Reference Barros, Victora, Horta, Goncalves, Lima and Lynch17, Reference Goldani, Haeffner, Agranonik, Barbieri, Bettiol and Silva44), but no other studies from low- or middle-income countries were located. Six studies evaluated the association between early SEP and WHR in adults in high-income countries, five showing inverse associations for both sexes(Reference Kivimaki, Smith and Juonala20, Reference Kuh, Hardy, Chaturvedi and Wadsworth21, Reference Laitinen, Power and Jarvelin23Reference Poulton, Caspi, Milne, Thomson, Taylor, Sears and Moffitt25) and one finding an inverse relationship in women and no association in men(Reference Laitinen, Pietilainen, Wadsworth, Sovio and Jarvelin22). Only one study was located from the low- and middle-income countries (China), which reported a direct association in males and an inverse one in females(Reference Schooling, Jiang, Lam, Zhang, Cheng and Leung27).

In women, even in middle-income countries, there is an inverse association between concurrent SEP and obesity in adulthood(Reference Monteiro, Conde and Popkin7, Reference Monteiro, Moura, Conde and Popkin8, Reference McLaren15). Our findings add to the evidence suggesting that concurrent poverty – rather than early SEP – determines central obesity. Poverty in early life, however, is associated with lower HC independent of current SEP, probably due to early developmental effects on the bony or muscle structure of the hip. Adverse conditions in the perinatal period related to poverty (obstetric complications, low maternal height, maternal and/or fetal undernutrition) can affect the anatomy of the hip, as well as cause reductions in bone mass that persist until adulthood(Reference Cooper, Westlake, Harvey, Javaid, Dennison and Hanson38, Reference Sizínio and Xavier39). Other osseous components can also be affected; for example, early-life poverty was shown to be associated with lower height at 18–19 years in the same cohort(Reference Barros, Victora, Horta, Goncalves, Lima and Lynch17). These findings agree with a recent review on the long-term effect of maternal/infant nutrition(Reference Victora, Adair, Fall, Hallal, Martorell, Richter and Sachdev45).

Associations between SEP and the outcomes tested were partially explained by adult behavioural variables such as alcohol consumption, physical inactivity, poor dietary habits (all directly associated with adiposity) and smoking (inversely associated)(Reference Koh-Banerjee, Chu, Spiegelman, Rosner, Colditz, Willett and Rimm46, Reference Lahmann, Lissner, Gullberg and Berglund47). In Pelotas, alcohol consumption is most common among richer men but there is no pattern among women(Reference Castanheira, Olinto and Gigante48). Smoking, on the other hand, is more common among poorer men and women. Moreover, there is a strong inverse association between overall physical activity and income in men(Reference Hallal, Victora, Wells and Lima49). In women, levels of activity were slightly higher in the poorest, but due to low-intensity domestic activities. Several authors have suggested that poor women in middle- and high-income countries have nutrient-poor, energy-dense diets(3, Reference Monteiro, Conde and Popkin7, Reference Monteiro, Moura, Conde and Popkin8, Reference McLaren15) and may not be as concerned with the aesthetic aspects of thinness. These differences partly explain the higher frequency of abdominal obesity in richer men and the less intense mediating effects of behavioural variables in women.

Current BMI is closely associated to the outcomes under study, and BMI-adjusted analyses investigate possible specific effects beyond an association with overall fatness. SEP at birth remained associated with WC (in men) and HC (in both sexes) after adjustment for current BMI, suggesting that early-life influences on the growth of bone and muscle, and on fat distribution, lead to permanent effects on body shape. On the other hand, concurrent SEP, while strongly associated with BMI (directly in men and inversely in women), was no longer associated with the outcomes after controlling for BMI. This suggests that – unlike early-life poverty – the effects of adult wealth or poverty are not specifically related to either waist or hip dimensions.

Parity has shown a direct association with BMI in the cohort(Reference Gigante, Rasmussen and Victora50) as well as with WC but not HC. Nevertheless, in the present analysis, adjustment for parity did not remove the social differences.

The associations with skin colour are likely related to socio-economic factors, as whites compose 67 % of the lowest tertile of income and 87 % of the highest. In women, several risk factors for CVD – obesity, hypertension, diabetes and insulin resistance – are more common in blacks than in whites(Reference Yusuf, Reddy, Ounpuu and Anand4, Reference Cossrow and Falkner28). This association has also been described with respect to WHR(Reference Dekkers, Podolsky, Treiber, Barbeau, Gutin and Snieder18, Reference Croft, Freedman, Keenan, Sheridan, Macera and Wheeler26, Reference Harris, Stevens, Thomas, Schreiner and Folsom29, Reference Hill, Sidney, Lewis, Tolan, Scherzinger and Stamm30). Croft et al.(Reference Croft, Freedman, Keenan, Sheridan, Macera and Wheeler26) showed that the difference in WHR between US blacks and whites was reduced but remained significant after adjustment for education and behavioural variables. Our results in women show that racial/ethnic differences in WHR disappeared after adjustment for SEP, but that the effects of the adjustment were different for waist and hip measures. The positive confounding for WC suggests that black women have larger waist due to being poorer and adopting the behavioural habits of this social group. For HC, there was negative confounding, suggesting that the biological growth potential of black women’s hips is not being attained due to nutritional restrictions associated with poverty in early life. In men, in accordance with the literature, there were no differences in waist or hip measures according to skin colour(Reference Yusuf, Reddy, Ounpuu and Anand4, Reference Dekkers, Podolsky, Treiber, Barbeau, Gutin and Snieder18, Reference Wagner and Heyward31).

Finally, associations between WHR and SEP result from combined effects of socio-economic trajectories on WC and HC. Men with higher family income at birth, higher concurrent income and who were never poor had both greater WC and HC, which may explain the lack of association with WHR. In women, WHR was inversely associated with income at birth (due to smaller HC among the poor) and in adult life (due to the inverse association between current income and WC). There were no differences in WHR according to skin colour – black women had both larger waists and hips than white women, whereas there were no differences in either indicator among men.

Our analytical models were possible owing to the longitudinal nature of the study. One methodological limitation is that ∼40 % of the original cohort members were not followed up. It is difficult to speculate how more losses among poorer individuals could have affected the analyses described here, but it is unlikely that they led to bias. Another possible limitation is the relatively low frequency of individuals with higher levels of abdominal obesity at 23–24 years. Nevertheless, this did not affect the results and even at this age differences of 3–5 cm in WC and HC were found between SEP groups. Finally, the ratio of white to black/brown subjects was lower compared with that in most other parts of Brazil. However, our results of higher levels of abdominal obesity among black women compared with whites and on the lack of differences in men are consistent with the literature(Reference Yusuf, Reddy, Ounpuu and Anand4, Reference Dekkers, Podolsky, Treiber, Barbeau, Gutin and Snieder18, Reference Croft, Freedman, Keenan, Sheridan, Macera and Wheeler26, Reference Cossrow and Falkner28Reference Hill, Sidney, Lewis, Tolan, Scherzinger and Stamm30).

In conclusion, we found strong sex-specific associations between socio-economic trajectories and the anthropometric measures evaluated. The early effects of poverty may be irreversible and therefore prevention depends on interventions aimed at the first years of life.

Acknowledgements

Sources of funding: The study was partially funded by The Wellcome Trust (Major Awards for Latin America on Health Consequences of Population Change). The initial phases of the cohort study were made possible by support from the National Program for Centers of Excellence (PRONEX), the Brazilian Ministry of Health (National Health Foundation), the International Development Research Center of Canada, and the United Nations Development Fund for Women (UK). Conflict of interest declaration: No conflicts of interests exist. No portion of the work has been or is currently under consideration for publication elsewhere nor has any other portion been published or posted on the Internet. Being accepted for Public Health Nutrition the paper will not be published elsewhere in the same form, in English or in any other language, without the written consent of the Nutrition Society. Authors’ contributions: D.A.G. was responsible for data collection, analysis and drafted the results of the study. A.N. participated in data collection and writing of the article. J.S.Y. contributed to the analyses and writing. C.G.V. supervised the study, participated in the analyses and writing. All of the authors read and approved the final manuscript. Acknowledgements: The authors would like to thank the researchers at the Center for Epidemiologic Research at the Federal University of Pelotas. Special thanks to Dr Denise Petrucci Gigante and Dr Anaclaudia Gastal Fassa for contributing data from the 1982 Pelotas Cohort and also to Jeovany Martínez Mesa, María Angélica Barbosa and María del Pilar Velez for their contributions to the data collection.

References

1.de Koning, L, Merchant, AT, Pogue, J & Anand, SS (2007) Waist circumference and waist-to-hip ratio as predictors of cardiovascular events: meta-regression analysis of prospective studies. Eur Heart J 28, 850856.CrossRefGoogle ScholarPubMed
2.Lanas, F, Avezum, A, Bautista, LE, Diaz, R, Luna, M, Islam, S & Yusuf, S (2007) Risk factors for acute myocardial infarction in Latin America: the INTERHEART Latin American study. Circulation 115, 10671074.CrossRefGoogle ScholarPubMed
3.World Health Organization (2000) Obesity: Preventing and Managing the Global Epidemic. Report of a WHO Consultation. WHO Technical Report Series no. 894. Geneva: WHO.Google Scholar
4.Yusuf, S, Reddy, S, Ounpuu, S & Anand, S (2001) Global burden of cardiovascular diseases: Part II: variations in cardiovascular disease by specific ethnic groups and geographic regions and prevention strategies. Circulation 104, 28552864.Google Scholar
5.Yusuf, S, Hawken, S, Ounpuu, S et al. (2005) Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case–control study. Lancet 366, 16401649.CrossRefGoogle Scholar
6.Jorgensen, ME, Borch-Johnsen, K & Bjerregaard, P (2006) Lifestyle modifies obesity-associated risk of cardiovascular disease in a genetically homogeneous population. Am J Clin Nutr 84, 2936.Google Scholar
7.Monteiro, CA, Conde, WL & Popkin, BM (2007) Income-specific trends in obesity in Brazil: 1975–2003. Am J Public Health 97, 18081812.CrossRefGoogle ScholarPubMed
8.Monteiro, CA, Moura, EC, Conde, WL & Popkin, BM (2004) Socioeconomic status and obesity in adult populations of developing countries: a review. Bull World Health Organ 82, 940946.Google ScholarPubMed
9.Iacobellis, G & Sharma, AM (2007) Obesity and the heart: redefinition of the relationship. Obes Rev 8, 3539.CrossRefGoogle ScholarPubMed
10.Seidell, JC, Perusse, L, Despres, JP & Bouchard, C (2001) Waist and hip circumferences have independent and opposite effects on cardiovascular disease risk factors: the Quebec Family Study. Am J Clin Nutr 74, 315321.CrossRefGoogle ScholarPubMed
11.Snijder, MB, Dekker, JM, Visser, M, Bouter, LM, Stehouwer, CD, Yudkin, JS, Heine, RJ, Nijpels, G & Seidell, JC (2004) Trunk fat and leg fat have independent and opposite associations with fasting and postload glucose levels: the Hoorn study. Diabetes Care 27, 372377.Google Scholar
12.Sinclair, D & Dangerfield, P (1998) Human Growth after Birth, 6th ed. New York: Oxford University Press.CrossRefGoogle Scholar
13.Tanner, JM (1989) Foetus into Man, 2nd ed. Ware: Castlemead Publications.Google Scholar
14.Taylor, RW, Grant, AM, Goulding, A & Williams, SM (2005) Early adiposity rebound: review of papers linking this to subsequent obesity in children and adults. Curr Opin Clin Nutr Metab Care 8, 607612.CrossRefGoogle ScholarPubMed
15.McLaren, L (2007) Socioeconomic status and obesity. Epidemiol Rev 29, 2948.CrossRefGoogle ScholarPubMed
16.Sobal, J & Stunkard, AJ (1989) Socioeconomic status and obesity: a review of the literature. Psychol Bull 105, 260275.CrossRefGoogle ScholarPubMed
17.Barros, AJ, Victora, CG, Horta, BL, Goncalves, HD, Lima, RC & Lynch, J (2006) Effects of socioeconomic change from birth to early adulthood on height and overweight. Int J Epidemiol 35, 12331238.CrossRefGoogle ScholarPubMed
18.Dekkers, JC, Podolsky, RH, Treiber, FA, Barbeau, P, Gutin, B & Snieder, H (2004) Development of general and central obesity from childhood into early adulthood in African American and European American males and females with a family history of cardiovascular disease. Am J Clin Nutr 79, 661668.CrossRefGoogle ScholarPubMed
19.Galobardes, B, Lynch, JW & Davey Smith, G (2004) Childhood socioeconomic circumstances and cause-specific mortality in adulthood: systematic review and interpretation. Epidemiol Rev 26, 721.CrossRefGoogle Scholar
20.Kivimaki, M, Smith, GD, Juonala, M et al. (2006) Socioeconomic position in childhood and adult cardiovascular risk factors, vascular structure, and function: Cardiovascular Risk in Young Finns Study. Heart 92, 474480.Google Scholar
21.Kuh, D, Hardy, R, Chaturvedi, N & Wadsworth, ME (2002) Birth weight, childhood growth and abdominal obesity in adult life. Int J Obes Relat Metab Disord 26, 4047.Google Scholar
22.Laitinen, J, Pietilainen, K, Wadsworth, M, Sovio, U & Jarvelin, MR (2004) Predictors of abdominal obesity among 31-y-old men and women born in Northern Finland in 1966. Eur J Clin Nutr 58, 180190.CrossRefGoogle ScholarPubMed
23.Laitinen, J, Power, C & Jarvelin, MR (2001) Family social class, maternal body mass index, childhood body mass index, and age at menarche as predictors of adult obesity. Am J Clin Nutr 74, 287294.CrossRefGoogle ScholarPubMed
24.Langenberg, C, Hardy, R, Kuh, D, Brunner, E & Wadsworth, M (2003) Central and total obesity in middle aged men and women in relation to lifetime socioeconomic status: evidence from a national birth cohort. J Epidemiol Community Health 57, 816822.CrossRefGoogle ScholarPubMed
25.Poulton, R, Caspi, A, Milne, BJ, Thomson, WM, Taylor, A, Sears, MR & Moffitt, TE (2002) Association between children’s experience of socioeconomic disadvantage and adult health: a life-course study. Lancet 360, 16401645.Google Scholar
26.Croft, JB, Freedman, DS, Keenan, NL, Sheridan, DP, Macera, CA & Wheeler, FC (1996) Education, health behaviors, and the black–white difference in waist-to-hip ratio. Obes Res 4, 505512.Google Scholar
27.Schooling, CM, Jiang, CQ, Lam, TH, Zhang, WS, Cheng, KK & Leung, GM (2008) Life-course origins of social inequalities in metabolic risk in the population of a developing country. Am J Epidemiol 167, 419428.CrossRefGoogle ScholarPubMed
28.Cossrow, N & Falkner, B (2004) Race/ethnic issues in obesity and obesity-related comorbidities. J Clin Endocrinol Metab 89, 25902594.CrossRefGoogle ScholarPubMed
29.Harris, MM, Stevens, J, Thomas, N, Schreiner, P & Folsom, AR (2000) Associations of fat distribution and obesity with hypertension in a bi-ethnic population: the ARIC study. Atherosclerosis Risk in Communities Study. Obes Res 8, 516524.Google Scholar
30.Hill, JO, Sidney, S, Lewis, CE, Tolan, K, Scherzinger, AL & Stamm, ER (1999) Racial differences in amounts of visceral adipose tissue in young adults: the CARDIA (Coronary Artery Risk Development in Young Adults) study. Am J Clin Nutr 69, 381387.CrossRefGoogle ScholarPubMed
31.Wagner, DR & Heyward, VH (2000) Measures of body composition in blacks and whites: a comparative review. Am J Clin Nutr 71, 13921402.Google Scholar
32.Instituto Brasileiro de Geografia e Estatística (2002) Produto Interno Bruto a preço de mercado e Produto Interno Bruto per capita, segundo as grandes regiões, unidades da federação e municípios – 1999–2002. Brazil: IBGE; available at http://www.ibge.gov.br/home/estatistica/economia/pibmunicipios/2002/tab01.pdfGoogle Scholar
33.Victora, CG & Barros, FC (2006) Cohort profile: the 1982 Pelotas (Brazil) birth cohort study. Int J Epidemiol 35, 237242.Google Scholar
34.Travassos, C & Williams, DR (2004) The concept and measurement of race and their relationship to public health: a review focused on Brazil and the United States. Cad Saude Publica 20, 660678.CrossRefGoogle ScholarPubMed
35.Sobal, J (1991) Obesity and socioeconomic status: a framework for examining relationships between physical and social variables. Med Anthropol 13, 231247.Google Scholar
36.Nazmi, A, Oliveira, IO & Victora, CG (2008) Correlates of C-reactive protein levels in young adults: a population-based cohort study of 3827 subjects in Brazil. Braz J Med Biol Res 41, 357367.Google Scholar
37.Fonseca-Alaniz, MH, Takada, J, Alonso-Vale, MI & Lima, FB (2006) The adipose tissue as a regulatory center of the metabolism. Arq Bras Endocrinol Metabol 50, 216229.CrossRefGoogle ScholarPubMed
38.Cooper, C, Westlake, S, Harvey, N, Javaid, K, Dennison, E & Hanson, M (2006) Review: developmental origins of osteoporotic fracture. Osteoporos Int 17, 337347.Google Scholar
39.Sizínio, S & Xavier, R (2003) Ortopedia e Traumatologia: Princípios e Prâticas, 3rd ed. Porto Alegre: Artmed Editora.Google Scholar
40.Dobbelsteyn, CJ, Joffres, MR, MacLean, DR & Flowerdew, G (2001) A comparative evaluation of waist circumference, waist-to-hip ratio and body mass index as indicators of cardiovascular risk factors. The Canadian Heart Health Surveys. Int J Obes Relat Metab Disord 25, 652661.CrossRefGoogle ScholarPubMed
41.Heitmann, BL (1992) Obesity and body fat distribution among adult Danes aged 35–65 years. An epidemiological study. Ugeskr Laeger 154, 12521257.Google Scholar
42.Zafon, C (2007) Oscillations in total body fat content through life: an evolutionary perspective. Obes Rev 8, 525530.Google Scholar
43.Power, C & Parsons, T (2000) Nutritional and other influences in childhood as predictors of adult obesity. Proc Nutr Soc 59, 267272.CrossRefGoogle ScholarPubMed
44.Goldani, MZ, Haeffner, LS, Agranonik, M, Barbieri, MA, Bettiol, H & Silva, AA (2007) Do early life factors influence body mass index in adolescents? Braz J Med Biol Res 40, 12311236.Google Scholar
45.Victora, CG, Adair, L, Fall, C, Hallal, PC, Martorell, R, Richter, L & Sachdev, HS (2008) Maternal and child undernutrition: consequences for adult health and human capital. Lancet 371, 340357.CrossRefGoogle ScholarPubMed
46.Koh-Banerjee, P, Chu, NF, Spiegelman, D, Rosner, B, Colditz, G, Willett, W & Rimm, E (2003) Prospective study of the association of changes in dietary intake, physical activity, alcohol consumption, and smoking with 9-y gain in waist circumference among 16 587 US men. Am J Clin Nutr 78, 719727.Google Scholar
47.Lahmann, PH, Lissner, L, Gullberg, B & Berglund, G (2000) Sociodemographic factors associated with long-term weight gain, current body fatness and central adiposity in Swedish women. Int J Obes Relat Metab Disord 24, 685694.CrossRefGoogle ScholarPubMed
48.Castanheira, M, Olinto, MT & Gigante, DP (2003) Socio-demographic and lifestyle factors associated with abdominal fat distribution in adults: a population-based survey in Southern Brazil. Cad Saude Publica 19, Suppl. 1, S55S65.Google Scholar
49.Hallal, PC, Victora, CG, Wells, JC & Lima, RC (2003) Physical inactivity: prevalence and associated variables in Brazilian adults. Med Sci Sports Exerc 35, 18941900.Google Scholar
50.Gigante, DP, Rasmussen, KM & Victora, CG (2005) Pregnancy increases BMI in adolescents of a population-based birth cohort. J Nutr 135, 7480.Google Scholar
Figure 0

Fig. 1 Follow-up visits and variables used from the 1982 Pelotas birth cohort study

Figure 1

Table 1 Comparison of demographic and socio-economic characteristics at birth between cohort members measured and not measured in 2006 (n 856): 1982 Pelotas birth cohort study, Brazil

Figure 2

Table 2 Descriptive analysis of waist circumference (WC), hip circumference (HC), waist:hip ratio (WHR) and BMI at 23 years according to independent variables and stratified by sex: 1982 Pelotas birth cohort study, Brazil

Figure 3

Table 3 Unadjusted and adjusted analyses for waist circumference (WC), hip circumference (HC) and waist:hip ratio (WHR) at 23 years in men according to past and concurrent socio-economic variables: 1982 Pelotas birth cohort study, Brazil

Figure 4

Table 4 Unadjusted and adjusted analyses for waist circumference (WC), hip circumference (HC) and waist:hip ratio (WHR) at 23 years in women according to past and concurrent socio-economic variables: 1982 Pelotas birth cohort study, Brazil