Body mass index and its association with socioeconomic and behavorial variables among socioeconomically heterogeneous populations of Andhra Pradesh, India

Reddy, B Nirmala


Abstract The nature and extent of relationship between socioeconomic and behavioral variables and body mass index [BMI = weight (kg)/height (m)2] is studied in a sample of 1119 individuals (456 males and 663 females), aged 18 to 75 years, drawn from socioeconomically diverse populations from southern Andhra Pradesh, India. These populations are categorized into four groups, with graded lifestyles toward urbanization. The BMI of the participants ranged from 12.6 to 35.1 in males and from 12.3 to 34.2 in females. There is an increasing trend in mean BMI until about 50 years, followed by a decline, indicating nonlinear nature of age effects. Mean BMI also increases with better socioeconomic status of the constituent groups and with decreased physical activity level. A trend of a decrease in the proportion of individuals with chronic energy deficiency and an increase in the proportion of obese individuals is also seen from the traditional Yerukala tribe (group I) to the urbanized group 4, from the lower to the higher income categories, and from heavy to light physical activity types. Smokers show a greater proportion of obese cases compared with nonsmokers. The prevalence of obesity (BMI-25) is 6.6% in males and 10% in females. The results of the analysis of variance suggest that three of the four socioeconomic and behavioral variables (except smoking) show significant effects on age-adjusted BMI, and the R2 suggests that these variables explain 27.4% of variation in males and 17% in females. Although income explains the largest amount of variation (24%) in males, followed by physical activity and group affiliation, in females group affiliation (12.5%), followed closely by income, accounted for most of the variation. Inclusion of age in the model improved explanatory power by 5-7%. The positive association between socioeconomic status and BMI observed in the present study is qualitatively different from the negative association that characterizes contemporary Western populations.

Obesity is considered a complex phenotype that is determined by genes, environmental exposure, and interactions between genes and environment (Bouchard 1989). Body mass index (BMI) is a proxy measure for obesity and is the most commonly studied marker for it. BMI is a measure of heaviness that reflects both lean and fat tissues (Bouchard and Perusse 1988; Garn et al. 1986). Using family and adoption data, many researchers have reported low to moderate polygenic effects along with larger environmental influences on BMI (Annest et al. 1983; Bouchard et al. 1988); however, twin studies have reported larger polygenic rather than environmental influences (Austin et al. 1987; Selby et al. 1990). Evidence for a major gene effect has also been detected, based on Indian family data [see Feitosa et al. (1997), Province et al. (1990), Price et al. (1990), and Borecki et al. (1993)].

BMI is a well-documented risk factor for cardiovascular disease. Its influence on blood pressure was recently demonstrated in a socioeconomically stratified sample from India (Reddy 1998). BMI is influenced by various factors, including age, sex, socioeconomic conditions, diet, exercise, and metabolic functions. Body mass and the prevalence of obesity have been shown to be inversely associated with socioeconomic status in the United States and other industrialized countries (Van Itallie 1985; Garn 1986; Sonne-Holm and Sorensen 1986; Sobol and Stunkard 1989; Khan et al. 1991; Gortmaker et al. 1993; Randrianjohany et al. 1993). Although the association between socioeconomic factors and body mass or adiposity in males has been weaker and more versatile (Flegal et al. 1988; Croft et al. 1992), Shah et al. (1989) reported that in a high-normotensive, affluent, and largely obese Western population about 20% of variance in BMI was due to socioeconomic and behavioral variables in both males and females; Shah found that occupation and income were the most important determinants of BMI in males, whereas in females the important correlates were alcohol intake, caffeinated drink intake, and race.

On the other hand, de Vasconcellos (1994), using Brazilian surveys, reported a higher probability of having thin individuals at low income levels, suggesting a positive association between income and BMI. Kennedy and Garcia (1994), who examined the trends in BMI for males and females across countries, reported that BMI of males showed a more consistent relation with increasing household income than the BMI of females. Delpeuch et al. (1994) reported a large prevalence of low BMI in rural areas and high BMI in urban areas of the Congo, a central African country; they also found that BMI is positively and significantly associated with socioeconomic level and that it is particularly sensitive to economic changes over time. In India Naidu and Rao (1994), using data from studies conducted by the National Institute of Nutrition, reported that mean BMI values were lower in landless agricultural laborers and in other low-income groups compared with cultivators, artisans, and high-income groups. In addition, Bharati (1989) reported a positive association between body dimensions and socioeconomic status in southern West Bengal, and Sanjeev et al. (1991) reported a significant difference in fatness between lower and upper socioeconomic status males and females. Sanjeev et al. (1991) also reported that no lower socioeconomic status individuals were in the overweight category, whereas 12.1% of upper socioeconomic status subjects were classified as overweight or severely overweight, according to their BMI.

The nutritional situation of any community is based on certain socioeconomic conditions, such as occupation, per capita income, and sometimes even population social background. In turn, BMI has a strong relationship with socioeconomic conditions and pattern of food intake. Behavioral aspects of lifestyle, such as diet, physical activity, and smoking, may be different in different socioeconomic groups, which may be reflected in the degree and nature of their relationship to BMI. These effects may be more perceptible in the populations of developing countries such as India, which are rapidly being urbanized. Based on occupational specializations, the traditional social stratification of the Indian population has in it an implicit disparity in the privileges that the different groups are entitled to, with the more privileged sections getting better opportunities for education, mobility, etc. compared with those in the lower strata. This stratification in turn determined the degree to which different populations responded to urbanization. Although this situation provides many appropriate study designs to understand the nature and extent of the effects of socioeconomic and behavioral variables on BMI or other related measures, such as adiposity, there have been few attempts at such studies (Sanjeev et al. 1991; Naidu and Rao 1994). Hence in the present study I attempt to delineate the effect of socioeconomic and behavioral variables on BMI among the socioeconomically contrasting populations of Andhra Pradesh, India.

Materials and Methods

Sample. Anthropometric, socioeconomic, and behavioral (caste affiliation, income, habitual physical activity, and smoking) data were measured for 1119 subjects (456 males and 663 females), aged 18 to 75 years. No systematic sampling design was adhered to in drawing the samples, although care was taken not to include related subjects; the subjects were brought one by one to a village central location, managed by the panchayat (local administrative unit of a village) president, where the relevant data were obtained. In the towns the data were gathered in the respective homes of the subjects. Because I am female, there were more female volunteers, and thus the sample is somewhat biased and represents a greater proportion of females. Furthermore, the samples were deliberately drawn from heterogeneous sections of the population, representing as many as seven castes of varying social status and a traditional seminomadic tribe (the Yerukala).

The castes and the tribe were categorized into four groups, with graded lifestyles toward urbanization: the traditional seminomadic tribe (the Yerukala) (group 1), hard-working agricultural and other laborers of the Mala caste and Muslims (group 2), the land-owning agriculturists castes, Reddy and Balija (group 3), and urban, sedentary Brahmin, Vyshya, and Marwadi (group 4). Broadly, these population groups also conform to the subpopulations’ socioeconomic and ethnographic affiliations (Thurston 1907, 1909). For example, although the Brahmins, Vyshyas, and Marwadis constituting group 4 come from the uppermost strata of Hindu society (Malhotra 1984), the Balija caste (group 3) is an offshoot of the Reddy caste (Thurston 1907). It is also common knowledge that most religious converts in India are from the depressed castes such as Malas; hence I group Muslims with the Malas in group 2.

The populations in group 4 (Brahmin, Vyshyas, and Marwadis) were mostly from urban areas and were engaged mostly in government service or in business, whereas the subjects in groups 1 and 2 were sampled from rural areas. Members of group 3 were sampled from rural and semirural areas. The Yerukala tribe, constituting group 1, is quite distinct in their lifestyle compared with the studied caste groups. They still lead a largely seminomadic life, move from village to village with their herd of pigs, and live in temporary tents erected near one of the villages or in the open. During their travels, they subsist largely by begging from the villagers for cooked food and grain. Thus they live in perpetual uncertainty of both the quality and quantity of their food consumption.

Populations in group 2, who are from the lower castes and from rural areas, depend on hard physical labor. Although populations in group 2 live under much less privileged conditions compared with groups 3 and 4, they interact regularly with those groups as laborers in the agricultural fields and hence are more exposed to the modern ways of life compared with the Yerukala tribe.

Anthropometry. Height and weight were measured using standard procedures (Weiner and Lourie 1981). BMI was computed as weight in kilograms divided by height squared (height was measured in meters).

Behavioral Measures. Among the behavioral variables physical activity levels were ascertained by inquiring about the nature of work usually performed by the subject and were classified according to the guidelines of Rao (1983) into three categories: heavy, medium, and light. Farm laborers, rickshaw pullers, washermen, and other artisan groups were included in the heavy physical activity category, whereas shopkeepers, landowners not actually engaged in tilling the land, servants, and housewives personally doing household work constituted the medium physical activity category. The light physical activity category included professionals, such as doctors, lawyers, and lecturers, businessmen, managers, clerks, and other officers, who generally lead sedentary lives. Students in the habit of playing outdoor games regularly were included in the medium physical activity category, and other students fell in the light physical activity category.

Physical activity levels are obviously highly correlated with occupation. Although quantitative levels of physical activity in terms of energy expenditure are not precisely ascertained, the type of habitual physical activity levels give a fair idea of the qualitative levels for classification into one of the physical activity grades.

The subjects were also classified as current smokers and nonsmokers. Former smokers were included under nonsmokers. Smokers are users of cigarettes and beedies (beedies are made with unprocessed tobacco and are rolled with dry tendoo leaf instead of paper); cigars are not used in this region. All females, irrespective of group, were nonsmokers.

Income and Socioeconomic Status. Subjects were classified into low-, middle-, or high-income groups based on monthly income level: (1) below 500 rupees, (2) 500-1499 rupees, and (3) 1500 rupees or above (equivalent to about $20, $50, and above). These broad income categories are based on the average perception of the income level from the standard of living.

All subjects in group 4 were vegetarians, whereas all others in the remaining three groups were nonvegetarians. Although I do not have quantitative data on individual or household food consumption, the average patterns can be summarized as follows. All the populations thrive on a highcarbohydrate and low animal protein diet. The populations in group 4 do not consume meat, fish, or eggs. However, because of financial and other constraints, the populations included in groups 1 and 2 also consume little animal protein. Vegetable protein (in the form of lentils) and leafy vegetables are generally consumed by all groups but relatively much less so by the members of groups 1 and 2. Consumption of milk and dairy products is popular only among the economically better off populations in groups 3 and 4. Members of these groups, especially group 4, also consume relatively more oils and fatty foods.

Statistical Analyses. Statistical analyses of data were accomplished using BMDP and SPSSx packages. BMI distributions were initially examined in male and female samples. The distributions were nearly normal and free from outliers. The bivariate plots of BMI and age showed a nonlinear relationship for both males and females. The stepwise regression analysis of BMI on cubic polynomials of age resulted in the selection of both age and age2. Therefore BMI was adjusted by including a quadratic term for age. The adjusted BMI values were subjected to an analysis of variance (ANOVA) to study the effect of socioeconomic and behavioral variables, both individually and simultaneously in the model, assuming a high degree of collinearity among some of the independent variables. This analysis is equivalent to the analysis of covariance in which BMI (not adjusted for age) is a dependent variable, the socioeconomic and behavioral variables are the factors, and the selected age terms, age and age2, are the covariates. The differences in mean BMI between various levels of each variable was assessed using Turkey’s multiple comparison test.


Age Trends in BMI. Mean BMI shows a gradual increase until about 5060 years of age, followed by a decline, indicating the nonlinear nature of age effects (Figure 2). BMIs significantly differ among the age groups in both males (F = 5.40, p = 0.000) and females (F = 8.40, p = 0.000). Females show a slightly higher mean BMI in each of the age groups after 30 years. However, age plots of mean BMI in different social groups (Figure 3) suggest that this pattern is characteristic only of the affluent groups 3 and 4 and not of the traditional populations of groups 1 and 2; mean BMI is significantly (p

Effects of Socioeconomic and Behavioral Variables on BMI. The means and standard errors of age-adjusted BMI are presented in Table 2 for different categories of each of the explanatory variables. Similar to the trend for unadjusted BMI (Table 2), there is a consistent increase in mean BMI from group 1 to 4, from low to high income groups, and from heavy to light physical activity. However, smokers and nonsmokers have similar mean BMIs. These trends are similar but relatively less consistent in females. The F values from ANOVA suggest significant heterogeneity (p ‘ 0.05) of means among categories of three of the four independent variables (Table 2), excluding smoking (p = 0.76). The RZ values suggest that income explains by far the largest amount of variation in BMI (24.5%) in males, followed by physical activity (16.8%) and group affiliation (16.5%); in females the maximum variation is accounted for by group affiliation (12.5%), closely followed by income (10.4%).

An analysis of covariance, in which BMI instead of age-adjusted BMI is used as a dependent variable, each of the categorical variables is used as a factor, and age and age2 are covariates, was also performed to examine the effect of the addition of age terms to the model. Age accounted for 5-7% more variation in BMI than was explained by each factor after eliminating the effect of age.

The analysis of differences in mean BMI between categories of each variable suggests that in males the mean BMI of traditional and hard-working groups 1 and 2 is significantly smaller than that of the more affluent groups 3 and 4; the means of groups 1 and 2 and groups 3 and 4 are similar. The mean BMI of adults with light physical activity is significantly (p

It is intuitively known and also apparent from the data that the categorical variables, especially group membership, income, and physical activity patterns, are highly interrelated. For example, members of affluent groups 3 and 4 are richer and physically less active than those of traditional groups 1 and 2. Therefore the effects of these variables may be confounding; given some of these categorical variables, others may not significantly account for the variation in BMI. Therefore ANOVA was done using adjusted BMI as the dependent variable and four categorical variables as factors simultaneously. The classical experimental approach used for this analysis adjusts for all other factors in the model while estimating the effects of a particular factor. This analysis, however, suggests that, except for smoking, each of the other categorical variables significantly explains BMI variation independent of other factors in males and females (Table 3). None of the interaction terms were significant, however. Nevertheless, the amount of variation explained by the four categorical variables together is 27.4%, only about 3% more than that explained by income alone (24.5%) in males. The amount of variation is about 17% in females and is about only 4% more than group membership alone could account for (12.5%). Inclusion of age and age2 as covariates in the model (with unadjusted BMI as the dependent variable) explains 5% more variation in BMI.


The results of the present study are in general agreement with earlier studies investigating the etiology of obesity in India (Sanjeev et al. 1991; Naidu and Rao 1994) and other developing countries, such as Brazil (Delpeuch et al. 1994), Congo (Kennedy and Garcia 1994), and Kenya (de Vasconcellos 1994), in that BMI or obesity is positively associated with socioeconomic status and negatively associated with physical activity level. Both monthly income and group membership denoting socioeconomic status in our study show a strong positive association with BMI. This is reflected not only in the mean BMIs but also in the proportion of obese individuals in different socioeconomic categories. Although only about 2% of the adult males in the rural and hard-working manual workers of groups 1 and 2 were found to be obese, 8% of males of the land-owning agriculturist group 3 and 14% of males of the most affluent group 4 were found to be obese (BMI => 25). For females these proportions were similar for the first three groups, but 24% compared with 14% in males were found to be obese in the affluent group 4. This frequency is as high as that found for the US population (Van Itallie 1985), although the standards used for those populations (BMI=>27) are somewhat higher (Frisancho 1990; Must et al. 1991).

Chronic energy deficiency (CED) is a term used to indicate an inadequate household food supply, and one of the methods recently suggested for assessing CED is BMI (James et al. 1988; Ferro-Luzzi et al. 1992). Although about 53% of adult males and 40% of adult females suffer from some form of CED in groups 1 and 2, this incidence is about 18% and 36% in group 3 and 25% and 19% in group 4, respectively. The greater proportion of females with CED in group 3 could be partly due to the relatively greater proportion of females in the sample being from rural backgrounds. Persistence of CED in about 20% of the adult population of even the most affluent group 4 is typically characteristic of developing nations. Thus, although obesity appears to be an emerging problem in the socioeconomically affluent sections of the Indian population, CED continues to be a problem even in this stratum and remains to be tackled. The results support the inference of Naidu and Rao (1994) that in Indian rural populations CED is of primary significance rather than obesity or overweight, as is the case in Western populations.

Both obesity and CED are consequences of chronic imbalance between energy intake and expenditure (Jequier 1987); although this imbalance is positive for obesity, it should be negative for a CED state to be manifested. The high levels of energy expenditure required for performing hard physical labor are obviously not being balanced by commensurate levels of energy intake by the economically poor populations of groups 1 and 2, hence the high frequency of CED among them. This imbalance is expected to have a greater impact on males rather than females who do hard physical labor, resulting in a higher proportion of CED among males. Furthermore, BMI is a result of complex interaction between nutritional intake, health status, and physical activity patterns (Parizkova 1977).

That males are less frequently obese than females is a universal observation. This difference can be seen most conspicuously in the most urbanized group 4 of the present study, which complements de Vasconcellos’s (1994) observation in Brazil; de Vasconcellos observed that females are thinner than males in the poorest regions, whereas in the richest regions the probability of finding an obese female is higher.

Furthermore, age is more strongly associated with BMI in females than in males. This is reflected in mean BMI values at all ages after 30 years (Figure 2), but with an increasing trend of sexual dimorphism. The enhanced association between age and BMI in females is thought to be an artifact of cumulative impact of pregnancies (Noppa and Bengtsson 1980), because some of the weight gained during pregnancy may be retained. That the significant association between BMI and age is characteristic only of the two affluent groups 3 and 4 and not of the rural groups confirms de Vasconcellos’s (1994) observation among Brazilians that BMI decreases with age in rural areas and increases with age in urban areas.

The positive association between socioeconomic status and BMI observed in the present study and in many other studies from developing countries is qualitatively different from the negative association characterizing contemporary Western populations (Van Itallie 1985; Forman et al. 1986; Garn 1986; Shah et al. 1989; Khan et al. 1991; Croft et al. 1992; Gortmaker et al. 1993; Randrianjohany et al. 1993; Stunkard and Sorensen 1993). In the industrialized West, where the populations had initially experienced greater consumption of a protein-rich diet followed by a rise in cardiovascular deaths, the more literate upper strata of the population, at least, are known to eat a more balanced diet, consciously avoiding food items that may contribute to obesity, high cholesterol levels, etc. They also probably exercise during their leisure time, being more conscious of the need to check their weight. This has probably been just the opposite in developing countries, including India, where the average population is undernourished. The economically well-off upper strata of the population in developing countries tend to consume relatively more protein and fat-rich items and are also relatively much less active, a trend probably typical of the initial stages of industrialization, even in developed countries. The same explanation is plausible for the lower BMI observed among nonmanual workers compared with manual workers in the United States, against the usually observed negative association between physical activity level and BMI.

Again, the disproportionately larger incidence of obesity and the larger mean BMI observed among the constituent members of group 4, who are vegetarians, compared with the so-called nonvegetarians constituting groups 1 to 3 has to be explained in the cultural context of the present sample. Members of group 4, although vegetarians, depend mostly on sedentary urban occupations such as government service, business, etc. and traditionally consume relatively much larger quantities of saturated fats and oils besides substantial quantities of vegetable protein in the form of pulses. A high fat intake may be conducive to weight gain through its effect on metabolic rate (Flatt 1978; Achenson et al. 1984; Swaminathan et al. 1985). A positive relationship between fat intake and BMI, independent of energy intake, has also been reported (Romieu et al. 1988; Dreon et al. 1988). The so-called nonvegetarians in the southern parts of India in general and particularly the constituent populations in groups 1, 2, and 3, who hail from those areas, consume little meat or fish and do so only irregularly. Also, these populations consume relatively much lower quantities of fats and oils and milk and milk products. Overall, the energy intake among the rural manual workers of groups 1 and 2 and even among the land-owning population of group 3 is much lower compared with the most affluent group 4, whereas the levels of energy expenditure are expected to be much higher among groups 1 to 3.

Of the four socioeconomic and behavioral variables considered, three (except smoking) show a significant association with or effect on age-adjusted BMI variation, whether considered individually or together. Of these variables, monthly income is the most significant in males, explaining about 24.5% of variation in BMI, but group membership explains the maximum amount of variation (12.5%) in females, closely followed by income. Inclusion of age in the model improved explanatory power by 5-7%. Given that most of the smokers, especially in the first three groups, use beedies, which are mild compared with cigars, the lack of a smoking effect is not surprising.

A considerable variation (5-25%) has been observed in other studies in the amount of variation in BMI explained by socioeconomic and behavioral variables. This could be due to methodological variation among the studies in the form of the number and kind of the considered independent variables and to the population backgrounds. At the same time, the estimates of genetic contribution also vary vastly, from 25% to 75% (Bouchard and Tremblay 1990; Sorensen et al. 1991). This suggests that a large part of the variation in BMI remains unexplained even in the present populations.

I conclude that the results of the present study generally agree with previous findings among the populations of developing countries, both in India and elsewhere. As highlighted by Naidu and Rao (1994), CED is of significant concern among rural populations, whereas certain sections of the socioeconomically advanced urban Indians, as demonstrated by the females of group 4, are dangerously surging toward developing a high load of obesity, which is one of the risk factors for developing cardiovascular diseases. Therefore the measures for public health intervention should be bifocal, critically evaluating the socioeconomic and cultural backgrounds of the concerned populations.

Acknowledgments I am thankful to the University Grants Commission for awarding me a research scientistship and a research grant; to the Director of the Indian Statistical Institute for logistic support; to Susanta Kr. Bera for computational assistance; and to P. Bhimasankaram, B. Mohan Reddy, and T. Krishnan for discussions on statistical analyses. The data were collected during 1985-1986 as part of my M. Phil. dissertation, under the supervision of P. Chengal Reddy. Finally, I thank the two anonymous reviewers whose comments helped to improve the presentation of the results. Received 10 February 1997; revision received 15 January 1998.

Literature Cited

Achenson, K.J., Y. Schutz, T. Bessard et al. 1984. Nutritional influences on lipogenesis and thermogenesis after a carbohydrate meal. Am. J. PhysioL 246:E62-70. Annest, J.L., C.F. Singh, P. Biron et al. 1983. Familial aggregation of blood pressure and weight in adoptive families. II. Analysis of the role of shared genes and shared household environment in explaining family resemblance for height, weight, and selected weight/ height indices. Am. J. Epidemiol. 117:492-506.

Austin, M.A., M.C. King, R.D. Bawol et al. 1987. Risk factors for coronary heart disease in adult female twins: Genetic heritability and shared environmental influences. Am. J. Epidemiol. 125:308-318.

Bharati, P. 1989. Variation in adult body dimensions in relation to economic condition among the Mahishyas of Howrah district, West Bengal, India. Ann. Hum. Biol. 16:529-541. Borecki, I.B., G.E. Bonney, T. Rice et al. 1993. Influence of genotype dependent effects of covariates on the outcome of segregation analysis of the body mass index. Am. J. Hum. Genet. 53:676687.

Bouchard, C. 1989. Genetic factors in obesity. Med. Clin. N. Am. 73:67-81. Bouchard, C., and L. Perusse. 1988. Heredity and body fat. Annu. Rev. Nutr. 8:259-277. Bouchard, C., and A. Tremblay. 1990. Genetic effects in human energy expenditure. Int. J.

Obes. 14(suppl. 1):49-58.

Bouchard, C., L. Perusse, C. Leblance et al. 1988. Inheritance in the amount and distribution of human body fat. Int. J. Obes. 12:205-215.

Croft, J.B., D.S. Strogatz, S.A. James et al. 1992. Socioeconomic and behavioral correlates of body mass index in black adults: The Pitt County Study. Am. J. Public Health 82:821826.

Delpeuch, F., A. Cornu, J.P. Massamba et al. 1994. Is body mass index sensitively related to socioeconomic status and economic adjustment? A case from the Congo. Eur. J. Clin. Nutr. 48 (suppl. 3):141-147.

de Vasconcellos, M.T.L. 1994. Body mass index: Its relationship with food consumption and socioeconomic variables in Brazil. Eur. J. Clin. Nutr. 48(suppl. 3):115-123. Dreon, D.M., B. Frey-Hewitt, N. Ellsworth et al. 1988. Dietary fat: Carbohydrate ratio and obesity in middle-aged men. Am. J. Clin. Nutr. 47:995-1000. Feitosa, M., T. Rice, A.N. Reddy et al. 1997. Is the major gene evidence for body mass index due to energy balance? Unpublished.

Ferro-Luzzi, A., A. Sette, M. Franklin et al. 1992. A simplified approach of assessing adult chronic energy deficiency. Eur. J. Clin. Nutr. 46:173-186. Flatt, J.P 1978. The biochemistry of energy expenditure. In Obesity Research 11, G.A. Bray, ed. London, England: Newman, 211-228.

Flegal, K.M., W.R. Harlan, and J.R. Landis. 1988. Secular trends in body mass index and skinfold thickness with socioeconomic factors in young adult men. Am. J. Clin. Nutr. 48:544-551.

Forman, M.R., F.L. Trowbridge, E.M. Gentry et al. 1986. Overweight adults in the United States: The behavior risk factor surveys. Am. J. Clin. Nutr. 44:410-416.

Frisancho, A.R. 1990. Anthropometric Standards for the Assessment of Growth and Nutritional Status. Ann Arbor, MI: University of Michigan Press.

Garn, S.M. 1986. Family-line and socioeconomic factors in fatness and obesity. Nutr. Rev. 44:381-386.

Garn, S.M., W.R. Leonard, and V.M. Hawthorne. 1986. Three limitations of the body mass index. Am. J. Clin. Nutr. 44:996-997.

Gortmaker, S.L., A. Must, J.M. Perrin et al. 1993. Social and economic consequences of overweight in adolescence and young adulthood. New EngL J. Med. 329:1008-1012. James, W.P.T., A. Ferro-Luzzi, and J.C. Waterlow. 1988. Definition of chronic energy deficiency in adults: Report of a working party of International Dietary Energy Consultative Group. Eur. J. Clin. Nutr. 42:969-981.

Jequier, E. 1987. Energy utilization in human obesity. Ann. NYAcad. Sci. 132:594-611. Kennedy, E., and M. Garcia. 1994. BMI and economic productivity. Eur. J. Clin. Nutr. 48(suppl. 3):45-55.

Khan, H.S., D.F. Williamson, and J.A. Stevens. 1991. Race and weight change in US women: The roles of socioeconomic and marital status. Am. J. Public Health 81:319-323. Malhotra, K.C. 1984. Population structure among the Dhangar caste cluster of Maharashtra, India. In The People of South Asia, J.R. Lukacs, ed. New York: Plenum Press, 295-324. Must, A., G.E. Dallal, and W.H. Dietz. 1991. Reference data for obesity: 85th and 95th percentiles of body mass (wt/ht2) and triceps skinfold thickness. Am. J. Clin. Nutr. 53:839846.

Naidu, A.N., and N.P. Rao. 1994. Body mass index: A measure of the nutritional situation in Indian populations. Eur. J. Clin. Nutr. 48(suppl. 3):131-140. Noppa, H., and C. Bengtsson. 1980. Obesity in relation to socioeconomic status: A population study of women in Goteborg, Sweden. J. EpidemioL Commun. Health 34:139-142. Parizkova, J. 1977. Body Fat and Physical Fitness. The Hague, Netherlands: Martinus Nijhoff. Price, R.A., R. Ness, and P. Laskarzewski. 1990. Common major gene inheritance of extreme overweight. Hum. Biol. 62:747-765.

Province, M.A., P. Amqvist, J. Keller et al. 1990. Strong evidence for a major gene for obesity in the large, unselected total Community Health Study of Tecumseh. Am. J. Hum. Genet. 47(suppl.):A143.

Randrianjohany, A., B. Balkau, J. Cubeau et al. 1993. The relationship between behavioral pattern, overall, and central adiposity in a population of healthy French men. Int. J. Obes. 17:651655.

Rao, S.R. 1983. Blood pressure levels of Rellis (India) with special reference to variation with age. Hum. Hered. 33:287-290.

Reddy, B.N. 1998. Blood pressure and adiposity: A comparative evaluation among the socioeconomically diverse groups of Andhra Pradesh, India. Am. J. Hum. Biol. 10:5-21. Romieu, I., W.C. Willet, M.J. Stampfer et al. 1988. Energy intake and other correlates of relative weight. Am. J. Clin. Nutr. 47:406412.

Sanjeev, G.D., I. Indrajit, and F.E. Johnston. 1991. Skinfold thickness, body circumferences, and their relation to age, sex, and socioeconomic status in adults from northwest India. Am. J. Hum. Biol. 3:469-477.

Selby, J.V., B. Newman, C.P. Quesenberry Jr. et al. 1990. Genetic and behavioral influences on body fat distribution. Int. J. Obes. 14:593-602.

Shah, M., R.W. Jeffrey, P.J. Hannan et al. 1989. Relationship between sociodemographic and behavior variables and body mass index in a population with high normal blood pressure: Hypertension prevention trial. Eur. J. Clin. Nutr. 43:583-596. Shetty, P.S., and W.P.T. James. 1994. Body Mass Index: A Measure of Chronic Energy Deficiency in Adults. FAO Food and Nutrition Paper 56. Rome, Italy: Food and Agricultural Organization.

Sobol, J., and A.J. Stunkard. 1989. Socioeconomic status and obesity: A review of the literature. Psychol. Bull. 105:260-275.

Sonne-Holm, S., and T.I.A. Sorensen. 1986. Prospective study of attainment of social class of severely obese subjects in relation to parental social class, intelligence, and education. Br. Med. J. 292:586-589.

Sorensen, T.I.A., A.J. Stunkard, and C. Holst. 1991. Recessive inheritance of a major gene for human obesity. Int. J. Obes. 15(suppl. 1):13.

Stunkard, A.J., and T.I.A. Sorensen. 1993. Obesity and socioeconomic status: A complex re

lation. New Engl. J. Med. 329:1036-1037.

Swaminathan, R., R.F.G.J. King, J. Holmfield et al. 1985. Thermic effect of feeding carbohydrate fat and protein and mixed meal in lean and obese subjects. Am. J. Clin. Nutr. 42:171-181.

Thurston, E. 1907. Ethnographic Notes in Southern India. Madras, India: Government Press. Thurston, E. 1909. Castes and Tribes of Southern India. Madras, India: Government Press. Van Itallie, T.B. 1985. Health implications of overweight and obesity in the United States. Ann. Intern. Med. 103:983-988.

Weiner, J.S., and J.A. Lourie. 1981. Practical Human Biology. London, England: Academic Press.

‘ Anthropometry and Human Genetics Unit, Indian Statistical Institute, Calcutta 700 035, India.

Copyright Wayne State University Press Oct 1998

Provided by ProQuest Information and Learning Company. All rights Reserved

You May Also Like

Book reviews — Basic and Clinical Biostatistics by Beth Dawson-Saunders and Robert G. Trapp

Book reviews — Basic and Clinical Biostatistics by Beth Dawson-Saunders and Robert G. Trapp / Statistical First Aid by Robert P. Hirsch and Ric…

Obituary: Arthur Mourant (1904-1994)

Obituary: Arthur Mourant (1904-1994) Roberts, Derek F Arthur Mourant, who brought a new dimension to biological anthropology, died o…

La Alta Alpujarra Oriental (Southeast Spain)

Spontaneous abortion pattern in an isolated Mediterranean population: La Alta Alpujarra Oriental (Southeast Spain) Luna, F Abstract …

Lessons from sexual orientation studies

Molecular investigations into complex behavior: Lessons from sexual orientation studies Pattatucci, Angela M ANGELA M. PATTATUCCI 1 …