Obesity Research & Clinical Practice (2009) 3, 141—148
ORIGINAL ARTICLE
Body mass index and body fat among adult Bengalee male slum dwellers in West Bengal, India Raja Chakraborty a,b, Kaushik Bose a,∗, Romendro Khongsdier c, Samiran Bisai a,c a
Department of Anthropology, Vidyasagar University, Midnapore 721102, West Bengal, India Department of Anthropology, Dinabandhu Mahavidyalaya, Bongaon, North 24 Paraganas, West Bengal, India c Department of Anthropology, North-Eastern Hill University, Shillong 793022, India b
Received 20 November 2008 ; received in revised form 5 March 2009; accepted 18 March 2009
KEYWORDS Bengalee; Slum; Body mass index; Percent body fat; Receiver operating characteristic; Obesity
∗
Summary Objective: The objective of the study is to explore the relationship between body mass index (BMI) and percent body fat (PBF) in relation to hypertension among adult Bengalee males of low socio-economic status living in a slum area of West Bengal, India. Methods: A cross-sectional survey was carried on 436 males aged 18—60 years in a slum area called Bidhan Colony, which is approximately 15 km from Kolkata city. Data on anthropometric measurements and blood pressure were collected, following standard techniques. Logistic regression and receiver operating characteristic (ROC) curve analysis were used for testing the relationship between BMI and PBF relative to hypertension. Results: About 4.25% and 50% of the normal (BMI 18.5—22.9 kg/m2 ) and overweight (BMI 23.0—24.9 kg/m2 ) subjects, respectively, were obese according to the PBF cutoff point of >25%. The ROC curve analysis indicated that the BMI cut-off ≥23 kg/m2 was appropriate for detecting obesity relative to hypertension. It was observed that the prevalence of hypertension increased significantly with age (r = 0.226, p < 0.001). Adjusting for age, the subjects with BMI 23—24.99 kg/m2 had about 3.2 times (95% CI: 1.61—6.27) greater risk of hypertension than those with BMI < 23 kg/m2 , and the risk for those with BMI ≥ 25 kg/m2 was about 4.5 times (95% CI: 2.06—9.57). As for PBF, the risk of hypertension was about 2.6 times (95% CI: 1.38—4.80) for the subjects with PBF > 25% compared to those with PBF ≤ 25%.
Corresponding author. Tel.: +91 09433403815. E-mail address: [email protected] (K. Bose).
1871-403X/$ — see front matter © 2009 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.orcp.2009.03.003
142
R. Chakraborty et al. Conclusions: Our study validated the BMI cut-off point proposed by the WHO for AsiaPacific populations for screening the individuals who are likely at risk of overweight. However, such data should be substantiated by independent risks of adverse health outcomes that need for public health intervention. © 2009 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
Introduction Accumulation of percent body fat (PBF) of >25% in males and >35% in females, corresponding to a body mass index (BMI) of ≥30 kg/m2 in young adult Caucasians, is internationally accepted and extensively used to define obesity or excess adiposity [1,2]. It is recognised as one of the major health problems in both developed and developing countries [3]. However, BMI is generally used as a measure of obesity rather than PBF because of its simplicity and high correlation with body fat [4]. The World Health Organisation (WHO) has recommended the BMI cutoffs of 25.0 kg/m2 and 30.0 kg/m2 for overweight and obesity, respectively [5]. But there is considerable evidence that these cut-off values are not applicable across ethnic groups, especially among Asian populations. It has been reported that Asian Indians, for example, have higher PBF, waist-to-hip ratio (WHR) and abdominal fat at a lower level of BMI compared with the Caucasian populations [6—8]. In Asian subjects, the risk of association with diabetes and CVD occurs at lower levels of BMI compared with the Caucasians [9—11]. Accordingly, The WHO Regional Office for Western Pacific Region, along with the International Association for the Study of Obesity (IASO) and the International Obesity Task Force (IOTF), has recommended new BMI cut-off points of 23.0 kg/m2 and 25.0 kg/m2 for defining overweight and obesity, respectively, in Asian populations [12]. Although there are several techniques to measure percent body fat, the use of skinfolds is the most preferred method because it is non-invasive, less expensive and suitable for large scale population surveys. There is increasing evidence that the relationship between BMI and PBF depends upon age, sex and ethnicity [13—16]. It has been also documented that the same ethnic group residing in different geographical locations could have a different pattern of BMI—PBF relationship. Asians including Asian Indians refer to a vast and bio-culturally diversified populations living in different levels of urbanisation, socio-economic conditions and nutrition transitions. Each of them may have a different body composition [4]. It is, therefore, recommended that further researches be undertaken with a view to
understanding the relationship between BMI and body fat in relation to risk factors and health outcomes, especially among Asian populations [4,12]. In India and other developing countries, slum dwellers are of particular interest, because they are mainly the poor people who migrated from rural areas to settle down in towns and cities, thereby getting exposed to adverse urban lifestyles and obesogenic environments. Unfortunately, there is hardly any study among the slum dwellers in India to explore the relationship between BMI and PBF [17] in relation to risk factors like hypertension, although some studies reported the relationship between BMI and risk factors for cardiovascular and metabolic disorders in non-slum areas [18—21]. The main purpose of the present study is to explore the relationship between BMI and PBF in relation to hypertension among adult Bengalee males of low socio-economic status residing in a slum area of West Bengal, India, since there is lack of information on obesity and hypertension among this group.
Materials and methods Study area and sample The present study was conducted as a part of a research project jointly undertaken by the first two authors in a slum area known as Bidhan Colony of Dum Dum, approximately 15 km from the centre of Kolkata city. Kolkata (formerly known as Calcutta) is the capital of the state of West Bengal in India. Kolkata is situated on the eastern bank of the river Ganges (also known as Hooghly River), about 120 km from the Bay of Bengal. Dum Dum, being one of the urban centres of the district is about 10 km to the north of Kolkata. The subjects of the study were adult men belonging to the Bengalee Hindu castes. The slum is situated at the right side of the railway tracks between Dum Dum Junction and the Dum Dum Cantonment Railway Stations. It is the terminal part of an urban settlement, called Purba (East) Sinthee, nearby the Dum Dum Junction Railway Station, under the South Dum Dum Municipality, North 24 Parganas of West Bengal. The
Body mass index and body fat among Bengalee males Table 1
Characteristics of the sample.
Variables
Mean
S.D.
Age (years) Height (cm) Weight (kg) Biceps skinfolds (mm) Triceps skinfolds (mm) Sub-scapular skinfolds (mm) Supra-iliac skinfolds (mm) Sum of four skinfolds (mm) Body mass index (kg/m2 ) Percent body fat (%) Systolic blood pressure Diastolic blood pressure
34.75 161.73 53.61 4.48 7.23 13.83 13.29 38.83 20.47 16.07 120.12 79.55
11.12 6.11 9.34 2.42 3.60 7.67 8.86 21.10 3.26 6.95 13.55 9.59
other side of the railway track is under the jurisdiction of Kolkata Municipal Corporation. Most of the subjects belonged to a low socio-economic status, mostly being factory workers, rickshaw-pullers and day-labourers. Ethical approval and prior permission was obtained from Vidyasagar University Ethics Committee and the institution of the first author. The municipal authorities and local community leaders were informed before the commencement of the study. Each subject was interviewed and measured at his respective household. In some cases, depending upon logistic circumstances, they were taken to a common place for examination. However, all the participants had their residence within the administrative boundary of the area under study. Overall response rate was found to be around 80%. Informed consent was also obtained from each participant. A total of 436 adult men aged 18—60 years were included in this study. For missing blood pressure data, three subjects were excluded in the analyses involving blood pressure and hypertension. Therefore, in those cases the sample size remained 433.
Measurements The field investigation including anthropometric measurements was carried out by the first author (R.C.). Information on ethnicity, age and some socio-economic were collected from each partici-
Table 2
143
Figure 1 Prevalence of obesity according to BMI categories in relation to PBF ≤ 25% and >25%.
pant with the help of a pre-tested questionnaire. All the anthropometric measurements were taken following the standard techniques [22]. Height and weight were measured to the nearest 0.1 cm and 0.5 kg, respectively, using standard anthropometer, and weight scale, respectively. Four skinfolds namely, biceps (BSF), triceps (TSF), sub-scapular (SSF) and supra-iliac (SISF), were measured to the nearest 0.2 mm using a skinfold calliper (Holtain Ltd., UK). Single instruments were used for each type of measurements to avoid inter-instrumental errors. Technical errors of measurements were found to be within the acceptable limits [23]. Blood pressure was measured by R.C. on the right arm of each subject, using standard stethoscope and a digital blood pressure monitor (Home Health, Switzerland) following the prescribed protocol. Resting systolic and diastolic blood pressures (in mmHg) was measured with the subject in a sitting position for at least 15 min prior to measurement and again at least 10 min after the first reading. The mean values of two measures were used in analyses. BMI was computed as weight (in kg) divided by height (in meter squared). Hypertension was defined as a systolic BP ≥ 140 mmHg and/or diastolic BP ≥ 90 mmHg, whereas obesity was defined as PBF > 25%. PBF was calculated using the sum of four
Mean (SD) age and prevalence (%) of obesity (PBF > 25) according to BMI categories.
BMI categories (kg/m2 )
Number
Age (mean ± S.D.)
25% 9 (2.59) 26 (50.00) 28 (75.68)
144
R. Chakraborty et al.
skinfolds according to the equations of Siri [24] and Durnin and Womersley [25] as generally followed by other studies in Indian populations [21,26]. The equations are expressed as follows: PBF =
4.95 − 4.5 × 100 density
where density = 1.1765 − 0.0744 ×log10 (BSF + TSF + SSF + SISF)
Statistical analyses Data were analysed using SPSS package for windows (SPSS Inc., Chicago, IL, USA). Two-tailed test was used by setting the significance level at 5%. Mean and standard deviation (S.D.) values for age, anthropometric variables and blood pressure were computed. The distributions of the anthropometric variables were not significantly skewed. Pearson correlation coefficient (r) was used to test the correlation between two variables. Receiver operating characteristics (ROC) curve analysis was used to determine the best BMI cut-off point against two categories of PBF i.e., non-obese (coded as 0) and obese (coded as 1). The odds ratio (OR) with 95% CI relative to the prevalence of hypertension was derived from the coefficient of logistic regression after adjusting for age.
Results Table 1 presents the means and standard deviations of age, blood pressure and anthropometric variables. Mean age of the subjects was
34.7 ± 11.13 years, and the mean values of BMI and PBF were 20.47 ± 3.26 kg/m2 and 16.07 ± 6.95%, respectively. The mean values of diastolic and systolic blood pressure were 79.55 ± 9.59 mmHg and 120.12 ± 13.55 mmHg, respectively, which seemed to be in normal condition. Table 2 shows the classification of the subjects according to BMI categories in relation to their PBF. It was found that BMI increased significantly with age of the individuals (r = 0.115, p < 0.016) and was positively correlated with PBF (r = 0.817, p < 0.0001). Following the recent recommendation of BMI classification for Asian populations [12], the prevalence of overweight was 11.93% and that of obesity 8.49% out of 436 subjects. It is, however, observed that about 50% of these overweight subjects would be classified as obese on the basis of their PBF. Similarly, about 24% of the obese subjects according to BMI would be classified as non-obese if their PBF was taken into consideration. This sort of misclassification is expected to be high especially if the BMI cut-off points of 25—29.9 kg/m2 and 30.0 kg/m2 are taken into consideration for classifying the subjects into overweight and obesity categories [5]. The present findings, therefore, seem to support the recommendation for reducing the BMI cut-off points for Asian populations, although a misclassification of the subjects is likely to persist irrespective of such reduced BMI categories. Fig. 1 shows the prevalence of obesity according to BMI categories in relation to PBF. About 4.25% and 50% of the normal (18.5—22.9 kg/m2 ) and overweight (23.0—24.9 kg/m2 ) subjects, respectively, were obese according to the PBF cut-off point of 25. Using ROC curve analysis, Table 3 shows sensitivity and specificity values according to selected threshold values of BMI against the reference PBF > 25%. The area under ROC curve (AUC) (Fig. 2) was 0.941 ± 0.021 with a CI of
Table 3 Sensitivity and specificity of different BMI thresholds in detection of obesity against the PBF reference value of >25%. BMI
Sensitivity (95% CI)
Specificity (95%CI)
Positive predicted value
Negative predicted value
21.50 22.00 22.50 22.90 23.00 23.10 23.50 24.00 24.50 25.00
92.1 (82.4—97.3) 92.1 (82.4—97.3) 87.3 (76.5—94.3) 85.7 (74.6—93.2) 85.7 (74.6—93.2) 84.1 (72.7—92.1 63.5 (50.4—75.3) 57.1 (44.0—69.5) 52.4 (39.4—65.1) 44.4 (31.9—57.5)
76.94 (72.3—81.1) 83.11 (78.9—86.8) 88.20 (84.5—91.3) 89.81 (86.3—92.7) 91.15 (87.8—93.8) 91.42 (88.1—94.1) 93.57 (90.6—95.8) 96.78 (94.4—98.3) 97.05 (94.8—98.5) 97.86 (95.8—99.1)
40.3 47.9 55.6 58.7 62.1 62.4 62.5 75.0 75.0 77.8
98.3 98.4 97.6 97.4 97.4 97.2 93.8 93.0 92.3 91.2
Body mass index and body fat among Bengalee males
Figure 2 ROC curve of sensitivity and specificity by BMI against reference PBF.
0.914—0.961 (p < 0.0001). The results indicated that a BMI ≥ 23 kg/m2 would be most appropriate for detecting obesity among the male slum dwellers of the present study. It is evident from Table 3 that the ROC curve for the BMI cut-off point of ≥25 kg/m2 against the reference PBF > 25% resulted in about 44% sensitivity (95% Cl: 31.9—57.5) and 98% specificity (95% Cl: 95.8—99.1). However, if the proposed cut-off for screening obesity is lowered to BMI ≥ 23 kg/m2 , the amount of sensitivity increased substantially from 44% to about 86% (95% Cl: 74.6—93.2), while that of specificity decreased marginally from 98% to about 91% (95% CI: 87.8—93.8). Similarly, the positive predicted
Table 4
145 value increased from 62% to 79% and the negative predicted value decreased from about 97% to 91% (highlighted in bold in Table 3). It may be worthwhile to mention that the prevalence of obesity in the present study increased from 8.5% for the BMI cut-off ≥25 kg/m2 to about 20.41% for the BMI cutoff ≥23 kg/m2 , resulting in an additional increase of about 12%. On the other hand, when the BMI cut-off ≥30 kg/m2 was taken into consideration [5], such an additional increase was about 19% (from 5 to 89 out of 436 individuals). The relationship between BMI and PBF was further tested taking into account the prevalence of hypertension (Table 4). The overall prevalence of hypertension was 17.6%. It was observed that the prevalence of hypertension increased significantly with age (r = 0.226, p < 0.001). The OR with 95% CI relative to the prevalence of hypertension derived from the coefficient of logistic regression indicated that the subjects aged 35—44 and ≥45 years were, respectively, about 3.6 times (95% CI: 1.38—9.28) and 7.5 times (95% CI: 2.95—18.89) more likely to suffer from hypertension as compared to those in the age group ≤ 24 years (p < 0.001). Adjusting for age, the subjects with BMI 23—24.9 kg/m2 were likely to have about 3.2 times (95% CI: 1.61—6.27) greater risk of hypertension than those with BMI < 23 kg/m2 , and the risk for those with BMI ≥ 25 kg/m2 was about 4.5 times (95% CI: 2.06—9.57). There was no significant difference between these two categories of BMI with respect to the prevalence of hypertension, although the subjects with BMI ≥ 25 kg/m2 had about 1.4 times (95% CI: 0.57—3.43) greater risk of hypertension than those with BMI 23—24.9 kg/m2 . As for PBF, the risk of hypertension was about 2.6 times (95% CI: 1.38—4.80) for the subjects with PBF > 25% as compared to those with PBF ≤ 25%. Thus, the
Summary of the logistic regression analysis of hypertension on age, BMI and PBF.
Parameters
N
Prevalence (%)
Odds ratio* (95% CI)
p-Level
Age groups (years) ≤24 25—34 35—44 ≥45
97 129 110 97
6 (6.19) 17 (13.18) 21(19.09) 32 (32.99)
— 2.30 (0.87—6.08) 3.58 (1.38—9.28) 7.47 (2.95—18.89)
— 0.092 0.009 0.001
BMI categories 25
371 62
56 (15.09) 20 (32.26)
— 2.57 (1.38—4.80)a
— 0.003
a
Adjusted for age.
146 present findings revealed that the proposed BMI cut-off point of ≥23 kg/m2 based on PBF > 25% was also associated with an increased risk of hypertension.
Discussion It is evident from the present analyses that there is considerable misclassification of the individuals as obese and non-obese on the basis of their BMI alone. The validity of BMI as a measure of obesity has, of course, been questioned by many studies in both developed and developing countries [1,2,13,17]. The major concern is that BMI is simply a crude measure of body weight relative to height, which tells nothing about the relative proportion of body fat and other components of body composition [26]. On the other hand, it is the degree of body fatness that should be considered a risk factor from the clinical and physiological points of view. In addition, the relationship between BMI and body fat is compounded by age, sex and ethnicity [13—16,27]. These factors make the diagnosis of obesity more complicated on the basis of BMI alone. Several studies in various Asian countries like China [28,29] Taiwan [30], Hong Kong [31] and Japan [32] have reported an association between a BMI > 22.3 kg/m2 and increased atherogenic risk factors. The risk of co-morbidities of diabetes, dyslipidemia, and hypertension was found to increase significantly with a BMI of >22.0 kg/m2 [33]. In short, there is considerable evidence that Asians have a greater percentage of total body fat at the same BMI values than the Europeans and the risk of chronic diseases in these populations increased significantly at a much lower BMI compared with the Europeans [34,35]. Our data on hypertension seem to be consistent with those earlier findings. Therefore, the present findings support the recommendation for reducing the BMI cut-off points for Asian populations, although a misclassification of the subjects is likely to persist irrespective of such reduced BMI categories depending upon the risk factors for a specific population. Our findings are also consistent with a study among urban males of southern India, where a BMI of >23 kg/m2 was significantly associated with an increased risk of diabetes [18]. Other studies have, however, suggested different cut-off points varying between 21.5 kg/m2 and 24 kg/m2 [17,19,20,36]. A recent study among the Bengalee males of Kolkata found that a BMI of 24 kg/m2 was the best cut-off for defining obesity [21]. The ROC curve analysis of our data indicated that the BMI cut-off point of ≥23 kg/m2 was appro-
R. Chakraborty et al. priate for detecting obesity (PBF > 25) relative to hypertension among the slum dwellers. Therefore, it is likely that the relationship between BMI and body fat relative to risk factors in India varies from one region to another, or from population to population depending on socio-economic and environmental conditions. There is, however, considerable evidence that the proportion of Indian subjects with a high risk of type 2 diabetes mellitus, hypertension and cardiovascular diseases is substantial at BMI values lower than the WHO cutoff point ≥25 kg/m2 for overweight [5]. The new proposed cut-off point ≥23 kg/m2 for overweight in Asia-Pacific populations [12] seems to be consistent with the present sample, although a large proportion of the overweight individuals at risk (23—24.9 kg/m2 ) may fall in the categories of obesity (25—29.9 kg/m2 for grade 1, and ≥30 kg/m2 for grade 2) relative to PBF and/or risk factors in Indian populations. This should not be confused in a population study in which the main purpose is to screen the individuals at risk of obesity and co-morbidities. However, the major concern may still prevail, if a large proportion of the Indian subjects are at risk of co-morbidities with BMI < 23 kg/m2 [20,36]. On the basis of available evidence including the present findings, the new BMI cut-off of ≥23 kg/m2 for AsiaPacific populations may be more appropriate than the conventional cut-off ≥25 kg/m2 for overweight at least for Indian populations. More studies are needed to carry out among different Indian populations, taking into consideration the relationship between BMI and PBF relative to risk factors. There are limitations of our study. Our study has taken PBF of >25% as a reference for determining the reliability of BMI as a diagnostic tool for assessing obesity. PBF was estimated from skinfold measurements, using prediction equations derived from European populations as generally followed by other studies [17,19,20,36]. In addition to technical errors of measurements, these prediction equations may have large prediction errors [37,38]. It is, therefore, difficult at present to suggest that the prediction methods of PBF based on skinfolds and bioelectrical impedance should be preferred over BMI. Although direct measurements of body fat like in vivo neutron activation analysis would be a better measure of obesity [4], such types of method would not be easily available especially for researchers in developing countries. The use of BMI is preferable because it is easier, speedier and less expensive; although its reliability as a measure of fatness can be questioned especially at the individual level. Considering our findings and other studies, it is likely that a BMI cut-off point relative to risk factors
Body mass index and body fat among Bengalee males would vary across Indian populations living in different ecological and socio-environmental conditions. The BMI cut-off point of ≥23 kg/m2 as recommended by the WHO for Asia-Pacific region [12] may be used for screening the proportion of people who are likely at risk of overweight. However, such data should be substantiated by independent and interactive risks of adverse health outcomes that need for public health intervention, taking into consideration the possibility of under- and over-estimation that may result in an unnecessary burden on the part of public health expenditure. The use of PBF in relation to BMI and risk factors is likely to be more informative about the nutritional and health status of the study population for further clinical investigation and intervention at the individual level. Lastly, it must be mentioned here that the widely used terms like Asian or Indian characterise a vast and diverse population. Diversity in Asian populations based on ethnicity, culture, degrees of urbanisation, socio-economic condition and nutrition transition [4] and therefore, our results may not applicable to all of them. Similar studies are needed among the various ethnic groups of diverse economic backgrounds to arrive at a more panIndian consensus.
147
[8]
[9]
[10]
[11]
[12] [13]
[14]
[15]
[16]
Conflict of interest statement [17]
This manuscript does not have any conflict of interest. [18]
References [1] Deurenberg P, Yap M, van Stavaren WA. Body mass index and percent body fat: a meta analysis among different ethnic groups. Int J Obes 1998;22:1164—71. [2] Deurenberg P, Deurenberg Yap M, Wang J, Lin FP, Schmidt G. The impact of body build on the relationship between body mass index and percent body fat. Int J Obes 1999;23:537—42. [3] Zimmet P, Alberti KG, Shaw J. Global and societal implications of the diabetes epidemic. Nature 2001;414:782—7. [4] WHO expert consultation. Appropriate body mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157—63. [5] WHO. Physical status: the use and interpretation of anthropometry. Technical Report Series No. 854. Geneva: World Health Organization; 1995. [6] Ramachandran A, Snehalatha C, Viswanathan V, Viswanathan M, Haffner SM. Risk of non-insulin dependent diabetes mellitus conferred by obesity and central adiposity in different ethnic groups: a comparative analysis between Asian Indians, Mexican Americans and Whites. Diabetes Res Clin Pract 1997;36:121—5. [7] Deurenberg-Yap M, Schmidt G, van Stavaren WA, Deurenberg P. The paradox of low body mass index and high body
[19]
[20]
[21]
[22]
[23]
[24]
[25]
fat percentage among Chinese, Malaya and Indians in Singapore. Int J Obes 2000;24:1011—7. Deurenberg P, Deurenberg-Yap M, Guricci S. Asians are different from Caucasians and from each other in their body mass index/body fat per cent relationship. Obes Rev 2002;3:141—6. McKeigue PM, Shah B. Relationship of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991;337:382—6. Banerji MA, Faridi N, Atluri R, Chaiken RL, Lebovitz HE. Body composition, visceral fat, leptin and insulin resistance in Asian Indian men. J Clin Endocrinol Metab 1999;84:137— 44. Chandalia M, Abate N, Garg A, Stray-Gundersen J, Gundy SM. Relationship between generalised and upper body obesity to insulin resistance in Asian Indian men. J Clin Endocrinol Metab 1999;84:2329—35. WHO. The Asia Pacific perspective: redefining obesity and its treatment. Geneva: World Health Organization; 2000. Norgan NG. Body mass index and body energy stores in developing countries. Eur J Clin Nutr 1990;44(S1):79— 84. Gallagher D, Visser M, Sepulveda D, Pierson RN, Harris T, Heymsfield S. How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am J Epidemiol 1996;143(3):228—39. Taylor RW, Jones IE, Williams SM, Goulding A. Evaluation of waist circumference, waist-to-hip ratio and the Conicity index as screening tools for high trunk fat mass, as measured by dual energy X-ray absorptiometry, in children aged 3—19 years. Am J Clin Nutr 2000;72:490—5. Bose K. The interaction of waist-hip ratio and Conicity index with subcutaneous adiposity in two ethnic groups: native British and migrant Pakistani men. Anthropol Anz 2001;59:275—82. Misra A, Pandey RM, Sinha S, Guleria R, Sridhar V, Dudeja V. Receiver operating characteristics curve analysis of body fat and body mass index in dyslipidemic Asian Indians. Indian J Med Res 2003;117:170—9. Snehalatha C, Viswanathan V, Ramachandran A. Cut off values for normal anthropometric variables in Asian Indian adults. Diabetes Care 2003;26:1380—4. Vikram NK, Misra A, Pandey RM, Dudeja V, Sinha S, Ramadevi J, et al. Anthropometry and body composition in Northern Asian Indians patients with type 2 diabetes: receiver operating characteristic (ROC) curve analysis of body mass index with percentage body fat as standard. Diabetes Nutr Metab 2003;16(1):32—40. Singh KD, Dhillon JK, Arora A, Gill BS. Receiver operating characteristic curve analysis of BMI and percentage body fat in type 2 diabetics of Punjab. Indian J Physiol Pharmacol 2004;48(1):73—80. Ghosh JR, Bandyopadhyay AR. Waist circumference as a measure of total body fat in Asian Indians. Int J Body Comp Res 2007;5(3):89—93. Lohman TG, Roche AF, Martorell R. Anthropometric standardization reference manual. Chicago: Human Kinetics Books; 1988. Ulijaszek SJ, Kerr DA. Anthropometric measurement error and the assessment of nutritional status. Br J Nutr 1999;82:165—77. Siri WE. Body composition from fluid spaces and density: analysis of methods. In: Brozek J, Hensche A, editors. Techniques for measuring body composition. Washington, DC: National Academy of Sciences; 1961. p. 223—44. Durnin JVGA, Womersley J. Body fat assessed from total body density and its estimation from skin fold thickness:
148
[26]
[27]
[28]
[29]
[30]
[31]
[32]
R. Chakraborty et al. measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77—97. Khongsdier R. BMI and morbidity in relation to body composition: a cross-sectional study of a rural community in North-East India. Br J Nutr 2005;93:101—7. Bose K, Mascie-Taylor CGN. Conicity index and waist-to-hip ratio and their relationship with total cholesterol and blood pressure in middle aged Europeans and migrant Pakistani men. Ann Hum Biol 1998;25:11—6. Zhou B, Wu Y, Yang J, Li Y, Zhang H, Zhao L. Overweight is an independent factor for cardiovascular disease in Chinese population. Obes Rev 2002;3:157—76. Jia WP, Xiang K, Chen L, Lu JX, Wu YM. Epidemiological study on obesity and its comorbidities in urban Chinese older than 20 years of age in Shanghai, China. Obes Rev 2002;3:157—76. Chang C, Wu C, Chang C, Yao WJ, Yang YC, Wu JS, Lu FH. Low body mass index but high percent body fat in Taiwanese subjects: implications of obesity cut off. Int J Obes 2003;27:253—9. Lee Z, Ctitchley J, Ko GT, Anderson PJ, Thomas GN, Young RP. Obesity and cardiovascular risk factors in Hong Kong Chinese. Obes Rev 2002;3:173—82. Ito H, Nakasuga K, Chen L, Nakasaga K, Oshima A, Maruyama T, et al. Detection of cardiovascular risk fac-
[33]
[34]
[35]
[36]
[37]
[38]
tors by indices of obesity obtained from anthropometry and dual-energy absorptiometry in Japanese individuals. Int J Obes 2003;27:232—7. Ko GTC, Chan JCN, Cockram CS, Woo J. Prediction of hypertension, diabetes, dyslipidemia or albuminuria using simple anthropometric indexes in Hong Kong Chinese. Int J Obes 1999;29:1136—42. Wang J, Thornton JC, Russel M, Burastero S, Heymsfield SB, Pierson RN. Asians have lower BMI but higher percent body fat than do Whites: comparisons of anthropometric measurements. Am J Clin Nutr 1994;60:23—8. Kim Y, Suh YK, Choi H. BMI and metabolic disorders in South Korean adults: 1998 Korea National Health and Nutrition Survey. Obes Res 2004;12(3):445—53. Dudeja V, Mishra A, Pandey RM, Devina G, Kumar G, Vikram NK. BMI does not accurately predict overweight in Asian Indians in Northern India. Br J Nutr 2001;86:105—12. Davies PSW. Anthropometry and body composition. In: Ulijaszek SJ, Mascie-Taylor CGN, editors. Anthropometry: the individual and the population. Cambridge: Cambridge University Press; 1994. p. 130—40. Peterson MJ, Czerwinski SA, Siervogel RM. Development and validation of skinfold-thickness prediction equations with a 4-compartment model. Am J Clin Nutr 2003;77:1186— 91.
Available online at www.sciencedirect.com
ORIGINAL ARTICLE
Body mass index and body fat among adult Bengalee male slum dwellers in West Bengal, India Raja Chakraborty a,b, Kaushik Bose a,∗, Romendro Khongsdier c, Samiran Bisai a,c a
Department of Anthropology, Vidyasagar University, Midnapore 721102, West Bengal, India Department of Anthropology, Dinabandhu Mahavidyalaya, Bongaon, North 24 Paraganas, West Bengal, India c Department of Anthropology, North-Eastern Hill University, Shillong 793022, India b
Received 20 November 2008 ; received in revised form 5 March 2009; accepted 18 March 2009
KEYWORDS Bengalee; Slum; Body mass index; Percent body fat; Receiver operating characteristic; Obesity
∗
Summary Objective: The objective of the study is to explore the relationship between body mass index (BMI) and percent body fat (PBF) in relation to hypertension among adult Bengalee males of low socio-economic status living in a slum area of West Bengal, India. Methods: A cross-sectional survey was carried on 436 males aged 18—60 years in a slum area called Bidhan Colony, which is approximately 15 km from Kolkata city. Data on anthropometric measurements and blood pressure were collected, following standard techniques. Logistic regression and receiver operating characteristic (ROC) curve analysis were used for testing the relationship between BMI and PBF relative to hypertension. Results: About 4.25% and 50% of the normal (BMI 18.5—22.9 kg/m2 ) and overweight (BMI 23.0—24.9 kg/m2 ) subjects, respectively, were obese according to the PBF cutoff point of >25%. The ROC curve analysis indicated that the BMI cut-off ≥23 kg/m2 was appropriate for detecting obesity relative to hypertension. It was observed that the prevalence of hypertension increased significantly with age (r = 0.226, p < 0.001). Adjusting for age, the subjects with BMI 23—24.99 kg/m2 had about 3.2 times (95% CI: 1.61—6.27) greater risk of hypertension than those with BMI < 23 kg/m2 , and the risk for those with BMI ≥ 25 kg/m2 was about 4.5 times (95% CI: 2.06—9.57). As for PBF, the risk of hypertension was about 2.6 times (95% CI: 1.38—4.80) for the subjects with PBF > 25% compared to those with PBF ≤ 25%.
Corresponding author. Tel.: +91 09433403815. E-mail address: [email protected] (K. Bose).
1871-403X/$ — see front matter © 2009 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.orcp.2009.03.003
142
R. Chakraborty et al. Conclusions: Our study validated the BMI cut-off point proposed by the WHO for AsiaPacific populations for screening the individuals who are likely at risk of overweight. However, such data should be substantiated by independent risks of adverse health outcomes that need for public health intervention. © 2009 Asian Oceanian Association for the Study of Obesity. Published by Elsevier Ltd. All rights reserved.
Introduction Accumulation of percent body fat (PBF) of >25% in males and >35% in females, corresponding to a body mass index (BMI) of ≥30 kg/m2 in young adult Caucasians, is internationally accepted and extensively used to define obesity or excess adiposity [1,2]. It is recognised as one of the major health problems in both developed and developing countries [3]. However, BMI is generally used as a measure of obesity rather than PBF because of its simplicity and high correlation with body fat [4]. The World Health Organisation (WHO) has recommended the BMI cutoffs of 25.0 kg/m2 and 30.0 kg/m2 for overweight and obesity, respectively [5]. But there is considerable evidence that these cut-off values are not applicable across ethnic groups, especially among Asian populations. It has been reported that Asian Indians, for example, have higher PBF, waist-to-hip ratio (WHR) and abdominal fat at a lower level of BMI compared with the Caucasian populations [6—8]. In Asian subjects, the risk of association with diabetes and CVD occurs at lower levels of BMI compared with the Caucasians [9—11]. Accordingly, The WHO Regional Office for Western Pacific Region, along with the International Association for the Study of Obesity (IASO) and the International Obesity Task Force (IOTF), has recommended new BMI cut-off points of 23.0 kg/m2 and 25.0 kg/m2 for defining overweight and obesity, respectively, in Asian populations [12]. Although there are several techniques to measure percent body fat, the use of skinfolds is the most preferred method because it is non-invasive, less expensive and suitable for large scale population surveys. There is increasing evidence that the relationship between BMI and PBF depends upon age, sex and ethnicity [13—16]. It has been also documented that the same ethnic group residing in different geographical locations could have a different pattern of BMI—PBF relationship. Asians including Asian Indians refer to a vast and bio-culturally diversified populations living in different levels of urbanisation, socio-economic conditions and nutrition transitions. Each of them may have a different body composition [4]. It is, therefore, recommended that further researches be undertaken with a view to
understanding the relationship between BMI and body fat in relation to risk factors and health outcomes, especially among Asian populations [4,12]. In India and other developing countries, slum dwellers are of particular interest, because they are mainly the poor people who migrated from rural areas to settle down in towns and cities, thereby getting exposed to adverse urban lifestyles and obesogenic environments. Unfortunately, there is hardly any study among the slum dwellers in India to explore the relationship between BMI and PBF [17] in relation to risk factors like hypertension, although some studies reported the relationship between BMI and risk factors for cardiovascular and metabolic disorders in non-slum areas [18—21]. The main purpose of the present study is to explore the relationship between BMI and PBF in relation to hypertension among adult Bengalee males of low socio-economic status residing in a slum area of West Bengal, India, since there is lack of information on obesity and hypertension among this group.
Materials and methods Study area and sample The present study was conducted as a part of a research project jointly undertaken by the first two authors in a slum area known as Bidhan Colony of Dum Dum, approximately 15 km from the centre of Kolkata city. Kolkata (formerly known as Calcutta) is the capital of the state of West Bengal in India. Kolkata is situated on the eastern bank of the river Ganges (also known as Hooghly River), about 120 km from the Bay of Bengal. Dum Dum, being one of the urban centres of the district is about 10 km to the north of Kolkata. The subjects of the study were adult men belonging to the Bengalee Hindu castes. The slum is situated at the right side of the railway tracks between Dum Dum Junction and the Dum Dum Cantonment Railway Stations. It is the terminal part of an urban settlement, called Purba (East) Sinthee, nearby the Dum Dum Junction Railway Station, under the South Dum Dum Municipality, North 24 Parganas of West Bengal. The
Body mass index and body fat among Bengalee males Table 1
Characteristics of the sample.
Variables
Mean
S.D.
Age (years) Height (cm) Weight (kg) Biceps skinfolds (mm) Triceps skinfolds (mm) Sub-scapular skinfolds (mm) Supra-iliac skinfolds (mm) Sum of four skinfolds (mm) Body mass index (kg/m2 ) Percent body fat (%) Systolic blood pressure Diastolic blood pressure
34.75 161.73 53.61 4.48 7.23 13.83 13.29 38.83 20.47 16.07 120.12 79.55
11.12 6.11 9.34 2.42 3.60 7.67 8.86 21.10 3.26 6.95 13.55 9.59
other side of the railway track is under the jurisdiction of Kolkata Municipal Corporation. Most of the subjects belonged to a low socio-economic status, mostly being factory workers, rickshaw-pullers and day-labourers. Ethical approval and prior permission was obtained from Vidyasagar University Ethics Committee and the institution of the first author. The municipal authorities and local community leaders were informed before the commencement of the study. Each subject was interviewed and measured at his respective household. In some cases, depending upon logistic circumstances, they were taken to a common place for examination. However, all the participants had their residence within the administrative boundary of the area under study. Overall response rate was found to be around 80%. Informed consent was also obtained from each participant. A total of 436 adult men aged 18—60 years were included in this study. For missing blood pressure data, three subjects were excluded in the analyses involving blood pressure and hypertension. Therefore, in those cases the sample size remained 433.
Measurements The field investigation including anthropometric measurements was carried out by the first author (R.C.). Information on ethnicity, age and some socio-economic were collected from each partici-
Table 2
143
Figure 1 Prevalence of obesity according to BMI categories in relation to PBF ≤ 25% and >25%.
pant with the help of a pre-tested questionnaire. All the anthropometric measurements were taken following the standard techniques [22]. Height and weight were measured to the nearest 0.1 cm and 0.5 kg, respectively, using standard anthropometer, and weight scale, respectively. Four skinfolds namely, biceps (BSF), triceps (TSF), sub-scapular (SSF) and supra-iliac (SISF), were measured to the nearest 0.2 mm using a skinfold calliper (Holtain Ltd., UK). Single instruments were used for each type of measurements to avoid inter-instrumental errors. Technical errors of measurements were found to be within the acceptable limits [23]. Blood pressure was measured by R.C. on the right arm of each subject, using standard stethoscope and a digital blood pressure monitor (Home Health, Switzerland) following the prescribed protocol. Resting systolic and diastolic blood pressures (in mmHg) was measured with the subject in a sitting position for at least 15 min prior to measurement and again at least 10 min after the first reading. The mean values of two measures were used in analyses. BMI was computed as weight (in kg) divided by height (in meter squared). Hypertension was defined as a systolic BP ≥ 140 mmHg and/or diastolic BP ≥ 90 mmHg, whereas obesity was defined as PBF > 25%. PBF was calculated using the sum of four
Mean (SD) age and prevalence (%) of obesity (PBF > 25) according to BMI categories.
BMI categories (kg/m2 )
Number
Age (mean ± S.D.)
25% 9 (2.59) 26 (50.00) 28 (75.68)
144
R. Chakraborty et al.
skinfolds according to the equations of Siri [24] and Durnin and Womersley [25] as generally followed by other studies in Indian populations [21,26]. The equations are expressed as follows: PBF =
4.95 − 4.5 × 100 density
where density = 1.1765 − 0.0744 ×log10 (BSF + TSF + SSF + SISF)
Statistical analyses Data were analysed using SPSS package for windows (SPSS Inc., Chicago, IL, USA). Two-tailed test was used by setting the significance level at 5%. Mean and standard deviation (S.D.) values for age, anthropometric variables and blood pressure were computed. The distributions of the anthropometric variables were not significantly skewed. Pearson correlation coefficient (r) was used to test the correlation between two variables. Receiver operating characteristics (ROC) curve analysis was used to determine the best BMI cut-off point against two categories of PBF i.e., non-obese (coded as 0) and obese (coded as 1). The odds ratio (OR) with 95% CI relative to the prevalence of hypertension was derived from the coefficient of logistic regression after adjusting for age.
Results Table 1 presents the means and standard deviations of age, blood pressure and anthropometric variables. Mean age of the subjects was
34.7 ± 11.13 years, and the mean values of BMI and PBF were 20.47 ± 3.26 kg/m2 and 16.07 ± 6.95%, respectively. The mean values of diastolic and systolic blood pressure were 79.55 ± 9.59 mmHg and 120.12 ± 13.55 mmHg, respectively, which seemed to be in normal condition. Table 2 shows the classification of the subjects according to BMI categories in relation to their PBF. It was found that BMI increased significantly with age of the individuals (r = 0.115, p < 0.016) and was positively correlated with PBF (r = 0.817, p < 0.0001). Following the recent recommendation of BMI classification for Asian populations [12], the prevalence of overweight was 11.93% and that of obesity 8.49% out of 436 subjects. It is, however, observed that about 50% of these overweight subjects would be classified as obese on the basis of their PBF. Similarly, about 24% of the obese subjects according to BMI would be classified as non-obese if their PBF was taken into consideration. This sort of misclassification is expected to be high especially if the BMI cut-off points of 25—29.9 kg/m2 and 30.0 kg/m2 are taken into consideration for classifying the subjects into overweight and obesity categories [5]. The present findings, therefore, seem to support the recommendation for reducing the BMI cut-off points for Asian populations, although a misclassification of the subjects is likely to persist irrespective of such reduced BMI categories. Fig. 1 shows the prevalence of obesity according to BMI categories in relation to PBF. About 4.25% and 50% of the normal (18.5—22.9 kg/m2 ) and overweight (23.0—24.9 kg/m2 ) subjects, respectively, were obese according to the PBF cut-off point of 25. Using ROC curve analysis, Table 3 shows sensitivity and specificity values according to selected threshold values of BMI against the reference PBF > 25%. The area under ROC curve (AUC) (Fig. 2) was 0.941 ± 0.021 with a CI of
Table 3 Sensitivity and specificity of different BMI thresholds in detection of obesity against the PBF reference value of >25%. BMI
Sensitivity (95% CI)
Specificity (95%CI)
Positive predicted value
Negative predicted value
21.50 22.00 22.50 22.90 23.00 23.10 23.50 24.00 24.50 25.00
92.1 (82.4—97.3) 92.1 (82.4—97.3) 87.3 (76.5—94.3) 85.7 (74.6—93.2) 85.7 (74.6—93.2) 84.1 (72.7—92.1 63.5 (50.4—75.3) 57.1 (44.0—69.5) 52.4 (39.4—65.1) 44.4 (31.9—57.5)
76.94 (72.3—81.1) 83.11 (78.9—86.8) 88.20 (84.5—91.3) 89.81 (86.3—92.7) 91.15 (87.8—93.8) 91.42 (88.1—94.1) 93.57 (90.6—95.8) 96.78 (94.4—98.3) 97.05 (94.8—98.5) 97.86 (95.8—99.1)
40.3 47.9 55.6 58.7 62.1 62.4 62.5 75.0 75.0 77.8
98.3 98.4 97.6 97.4 97.4 97.2 93.8 93.0 92.3 91.2
Body mass index and body fat among Bengalee males
Figure 2 ROC curve of sensitivity and specificity by BMI against reference PBF.
0.914—0.961 (p < 0.0001). The results indicated that a BMI ≥ 23 kg/m2 would be most appropriate for detecting obesity among the male slum dwellers of the present study. It is evident from Table 3 that the ROC curve for the BMI cut-off point of ≥25 kg/m2 against the reference PBF > 25% resulted in about 44% sensitivity (95% Cl: 31.9—57.5) and 98% specificity (95% Cl: 95.8—99.1). However, if the proposed cut-off for screening obesity is lowered to BMI ≥ 23 kg/m2 , the amount of sensitivity increased substantially from 44% to about 86% (95% Cl: 74.6—93.2), while that of specificity decreased marginally from 98% to about 91% (95% CI: 87.8—93.8). Similarly, the positive predicted
Table 4
145 value increased from 62% to 79% and the negative predicted value decreased from about 97% to 91% (highlighted in bold in Table 3). It may be worthwhile to mention that the prevalence of obesity in the present study increased from 8.5% for the BMI cut-off ≥25 kg/m2 to about 20.41% for the BMI cutoff ≥23 kg/m2 , resulting in an additional increase of about 12%. On the other hand, when the BMI cut-off ≥30 kg/m2 was taken into consideration [5], such an additional increase was about 19% (from 5 to 89 out of 436 individuals). The relationship between BMI and PBF was further tested taking into account the prevalence of hypertension (Table 4). The overall prevalence of hypertension was 17.6%. It was observed that the prevalence of hypertension increased significantly with age (r = 0.226, p < 0.001). The OR with 95% CI relative to the prevalence of hypertension derived from the coefficient of logistic regression indicated that the subjects aged 35—44 and ≥45 years were, respectively, about 3.6 times (95% CI: 1.38—9.28) and 7.5 times (95% CI: 2.95—18.89) more likely to suffer from hypertension as compared to those in the age group ≤ 24 years (p < 0.001). Adjusting for age, the subjects with BMI 23—24.9 kg/m2 were likely to have about 3.2 times (95% CI: 1.61—6.27) greater risk of hypertension than those with BMI < 23 kg/m2 , and the risk for those with BMI ≥ 25 kg/m2 was about 4.5 times (95% CI: 2.06—9.57). There was no significant difference between these two categories of BMI with respect to the prevalence of hypertension, although the subjects with BMI ≥ 25 kg/m2 had about 1.4 times (95% CI: 0.57—3.43) greater risk of hypertension than those with BMI 23—24.9 kg/m2 . As for PBF, the risk of hypertension was about 2.6 times (95% CI: 1.38—4.80) for the subjects with PBF > 25% as compared to those with PBF ≤ 25%. Thus, the
Summary of the logistic regression analysis of hypertension on age, BMI and PBF.
Parameters
N
Prevalence (%)
Odds ratio* (95% CI)
p-Level
Age groups (years) ≤24 25—34 35—44 ≥45
97 129 110 97
6 (6.19) 17 (13.18) 21(19.09) 32 (32.99)
— 2.30 (0.87—6.08) 3.58 (1.38—9.28) 7.47 (2.95—18.89)
— 0.092 0.009 0.001
BMI categories 25
371 62
56 (15.09) 20 (32.26)
— 2.57 (1.38—4.80)a
— 0.003
a
Adjusted for age.
146 present findings revealed that the proposed BMI cut-off point of ≥23 kg/m2 based on PBF > 25% was also associated with an increased risk of hypertension.
Discussion It is evident from the present analyses that there is considerable misclassification of the individuals as obese and non-obese on the basis of their BMI alone. The validity of BMI as a measure of obesity has, of course, been questioned by many studies in both developed and developing countries [1,2,13,17]. The major concern is that BMI is simply a crude measure of body weight relative to height, which tells nothing about the relative proportion of body fat and other components of body composition [26]. On the other hand, it is the degree of body fatness that should be considered a risk factor from the clinical and physiological points of view. In addition, the relationship between BMI and body fat is compounded by age, sex and ethnicity [13—16,27]. These factors make the diagnosis of obesity more complicated on the basis of BMI alone. Several studies in various Asian countries like China [28,29] Taiwan [30], Hong Kong [31] and Japan [32] have reported an association between a BMI > 22.3 kg/m2 and increased atherogenic risk factors. The risk of co-morbidities of diabetes, dyslipidemia, and hypertension was found to increase significantly with a BMI of >22.0 kg/m2 [33]. In short, there is considerable evidence that Asians have a greater percentage of total body fat at the same BMI values than the Europeans and the risk of chronic diseases in these populations increased significantly at a much lower BMI compared with the Europeans [34,35]. Our data on hypertension seem to be consistent with those earlier findings. Therefore, the present findings support the recommendation for reducing the BMI cut-off points for Asian populations, although a misclassification of the subjects is likely to persist irrespective of such reduced BMI categories depending upon the risk factors for a specific population. Our findings are also consistent with a study among urban males of southern India, where a BMI of >23 kg/m2 was significantly associated with an increased risk of diabetes [18]. Other studies have, however, suggested different cut-off points varying between 21.5 kg/m2 and 24 kg/m2 [17,19,20,36]. A recent study among the Bengalee males of Kolkata found that a BMI of 24 kg/m2 was the best cut-off for defining obesity [21]. The ROC curve analysis of our data indicated that the BMI cut-off point of ≥23 kg/m2 was appro-
R. Chakraborty et al. priate for detecting obesity (PBF > 25) relative to hypertension among the slum dwellers. Therefore, it is likely that the relationship between BMI and body fat relative to risk factors in India varies from one region to another, or from population to population depending on socio-economic and environmental conditions. There is, however, considerable evidence that the proportion of Indian subjects with a high risk of type 2 diabetes mellitus, hypertension and cardiovascular diseases is substantial at BMI values lower than the WHO cutoff point ≥25 kg/m2 for overweight [5]. The new proposed cut-off point ≥23 kg/m2 for overweight in Asia-Pacific populations [12] seems to be consistent with the present sample, although a large proportion of the overweight individuals at risk (23—24.9 kg/m2 ) may fall in the categories of obesity (25—29.9 kg/m2 for grade 1, and ≥30 kg/m2 for grade 2) relative to PBF and/or risk factors in Indian populations. This should not be confused in a population study in which the main purpose is to screen the individuals at risk of obesity and co-morbidities. However, the major concern may still prevail, if a large proportion of the Indian subjects are at risk of co-morbidities with BMI < 23 kg/m2 [20,36]. On the basis of available evidence including the present findings, the new BMI cut-off of ≥23 kg/m2 for AsiaPacific populations may be more appropriate than the conventional cut-off ≥25 kg/m2 for overweight at least for Indian populations. More studies are needed to carry out among different Indian populations, taking into consideration the relationship between BMI and PBF relative to risk factors. There are limitations of our study. Our study has taken PBF of >25% as a reference for determining the reliability of BMI as a diagnostic tool for assessing obesity. PBF was estimated from skinfold measurements, using prediction equations derived from European populations as generally followed by other studies [17,19,20,36]. In addition to technical errors of measurements, these prediction equations may have large prediction errors [37,38]. It is, therefore, difficult at present to suggest that the prediction methods of PBF based on skinfolds and bioelectrical impedance should be preferred over BMI. Although direct measurements of body fat like in vivo neutron activation analysis would be a better measure of obesity [4], such types of method would not be easily available especially for researchers in developing countries. The use of BMI is preferable because it is easier, speedier and less expensive; although its reliability as a measure of fatness can be questioned especially at the individual level. Considering our findings and other studies, it is likely that a BMI cut-off point relative to risk factors
Body mass index and body fat among Bengalee males would vary across Indian populations living in different ecological and socio-environmental conditions. The BMI cut-off point of ≥23 kg/m2 as recommended by the WHO for Asia-Pacific region [12] may be used for screening the proportion of people who are likely at risk of overweight. However, such data should be substantiated by independent and interactive risks of adverse health outcomes that need for public health intervention, taking into consideration the possibility of under- and over-estimation that may result in an unnecessary burden on the part of public health expenditure. The use of PBF in relation to BMI and risk factors is likely to be more informative about the nutritional and health status of the study population for further clinical investigation and intervention at the individual level. Lastly, it must be mentioned here that the widely used terms like Asian or Indian characterise a vast and diverse population. Diversity in Asian populations based on ethnicity, culture, degrees of urbanisation, socio-economic condition and nutrition transition [4] and therefore, our results may not applicable to all of them. Similar studies are needed among the various ethnic groups of diverse economic backgrounds to arrive at a more panIndian consensus.
147
[8]
[9]
[10]
[11]
[12] [13]
[14]
[15]
[16]
Conflict of interest statement [17]
This manuscript does not have any conflict of interest. [18]
References [1] Deurenberg P, Yap M, van Stavaren WA. Body mass index and percent body fat: a meta analysis among different ethnic groups. Int J Obes 1998;22:1164—71. [2] Deurenberg P, Deurenberg Yap M, Wang J, Lin FP, Schmidt G. The impact of body build on the relationship between body mass index and percent body fat. Int J Obes 1999;23:537—42. [3] Zimmet P, Alberti KG, Shaw J. Global and societal implications of the diabetes epidemic. Nature 2001;414:782—7. [4] WHO expert consultation. Appropriate body mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004;363:157—63. [5] WHO. Physical status: the use and interpretation of anthropometry. Technical Report Series No. 854. Geneva: World Health Organization; 1995. [6] Ramachandran A, Snehalatha C, Viswanathan V, Viswanathan M, Haffner SM. Risk of non-insulin dependent diabetes mellitus conferred by obesity and central adiposity in different ethnic groups: a comparative analysis between Asian Indians, Mexican Americans and Whites. Diabetes Res Clin Pract 1997;36:121—5. [7] Deurenberg-Yap M, Schmidt G, van Stavaren WA, Deurenberg P. The paradox of low body mass index and high body
[19]
[20]
[21]
[22]
[23]
[24]
[25]
fat percentage among Chinese, Malaya and Indians in Singapore. Int J Obes 2000;24:1011—7. Deurenberg P, Deurenberg-Yap M, Guricci S. Asians are different from Caucasians and from each other in their body mass index/body fat per cent relationship. Obes Rev 2002;3:141—6. McKeigue PM, Shah B. Relationship of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991;337:382—6. Banerji MA, Faridi N, Atluri R, Chaiken RL, Lebovitz HE. Body composition, visceral fat, leptin and insulin resistance in Asian Indian men. J Clin Endocrinol Metab 1999;84:137— 44. Chandalia M, Abate N, Garg A, Stray-Gundersen J, Gundy SM. Relationship between generalised and upper body obesity to insulin resistance in Asian Indian men. J Clin Endocrinol Metab 1999;84:2329—35. WHO. The Asia Pacific perspective: redefining obesity and its treatment. Geneva: World Health Organization; 2000. Norgan NG. Body mass index and body energy stores in developing countries. Eur J Clin Nutr 1990;44(S1):79— 84. Gallagher D, Visser M, Sepulveda D, Pierson RN, Harris T, Heymsfield S. How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am J Epidemiol 1996;143(3):228—39. Taylor RW, Jones IE, Williams SM, Goulding A. Evaluation of waist circumference, waist-to-hip ratio and the Conicity index as screening tools for high trunk fat mass, as measured by dual energy X-ray absorptiometry, in children aged 3—19 years. Am J Clin Nutr 2000;72:490—5. Bose K. The interaction of waist-hip ratio and Conicity index with subcutaneous adiposity in two ethnic groups: native British and migrant Pakistani men. Anthropol Anz 2001;59:275—82. Misra A, Pandey RM, Sinha S, Guleria R, Sridhar V, Dudeja V. Receiver operating characteristics curve analysis of body fat and body mass index in dyslipidemic Asian Indians. Indian J Med Res 2003;117:170—9. Snehalatha C, Viswanathan V, Ramachandran A. Cut off values for normal anthropometric variables in Asian Indian adults. Diabetes Care 2003;26:1380—4. Vikram NK, Misra A, Pandey RM, Dudeja V, Sinha S, Ramadevi J, et al. Anthropometry and body composition in Northern Asian Indians patients with type 2 diabetes: receiver operating characteristic (ROC) curve analysis of body mass index with percentage body fat as standard. Diabetes Nutr Metab 2003;16(1):32—40. Singh KD, Dhillon JK, Arora A, Gill BS. Receiver operating characteristic curve analysis of BMI and percentage body fat in type 2 diabetics of Punjab. Indian J Physiol Pharmacol 2004;48(1):73—80. Ghosh JR, Bandyopadhyay AR. Waist circumference as a measure of total body fat in Asian Indians. Int J Body Comp Res 2007;5(3):89—93. Lohman TG, Roche AF, Martorell R. Anthropometric standardization reference manual. Chicago: Human Kinetics Books; 1988. Ulijaszek SJ, Kerr DA. Anthropometric measurement error and the assessment of nutritional status. Br J Nutr 1999;82:165—77. Siri WE. Body composition from fluid spaces and density: analysis of methods. In: Brozek J, Hensche A, editors. Techniques for measuring body composition. Washington, DC: National Academy of Sciences; 1961. p. 223—44. Durnin JVGA, Womersley J. Body fat assessed from total body density and its estimation from skin fold thickness:
148
[26]
[27]
[28]
[29]
[30]
[31]
[32]
R. Chakraborty et al. measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77—97. Khongsdier R. BMI and morbidity in relation to body composition: a cross-sectional study of a rural community in North-East India. Br J Nutr 2005;93:101—7. Bose K, Mascie-Taylor CGN. Conicity index and waist-to-hip ratio and their relationship with total cholesterol and blood pressure in middle aged Europeans and migrant Pakistani men. Ann Hum Biol 1998;25:11—6. Zhou B, Wu Y, Yang J, Li Y, Zhang H, Zhao L. Overweight is an independent factor for cardiovascular disease in Chinese population. Obes Rev 2002;3:157—76. Jia WP, Xiang K, Chen L, Lu JX, Wu YM. Epidemiological study on obesity and its comorbidities in urban Chinese older than 20 years of age in Shanghai, China. Obes Rev 2002;3:157—76. Chang C, Wu C, Chang C, Yao WJ, Yang YC, Wu JS, Lu FH. Low body mass index but high percent body fat in Taiwanese subjects: implications of obesity cut off. Int J Obes 2003;27:253—9. Lee Z, Ctitchley J, Ko GT, Anderson PJ, Thomas GN, Young RP. Obesity and cardiovascular risk factors in Hong Kong Chinese. Obes Rev 2002;3:173—82. Ito H, Nakasuga K, Chen L, Nakasaga K, Oshima A, Maruyama T, et al. Detection of cardiovascular risk fac-
[33]
[34]
[35]
[36]
[37]
[38]
tors by indices of obesity obtained from anthropometry and dual-energy absorptiometry in Japanese individuals. Int J Obes 2003;27:232—7. Ko GTC, Chan JCN, Cockram CS, Woo J. Prediction of hypertension, diabetes, dyslipidemia or albuminuria using simple anthropometric indexes in Hong Kong Chinese. Int J Obes 1999;29:1136—42. Wang J, Thornton JC, Russel M, Burastero S, Heymsfield SB, Pierson RN. Asians have lower BMI but higher percent body fat than do Whites: comparisons of anthropometric measurements. Am J Clin Nutr 1994;60:23—8. Kim Y, Suh YK, Choi H. BMI and metabolic disorders in South Korean adults: 1998 Korea National Health and Nutrition Survey. Obes Res 2004;12(3):445—53. Dudeja V, Mishra A, Pandey RM, Devina G, Kumar G, Vikram NK. BMI does not accurately predict overweight in Asian Indians in Northern India. Br J Nutr 2001;86:105—12. Davies PSW. Anthropometry and body composition. In: Ulijaszek SJ, Mascie-Taylor CGN, editors. Anthropometry: the individual and the population. Cambridge: Cambridge University Press; 1994. p. 130—40. Peterson MJ, Czerwinski SA, Siervogel RM. Development and validation of skinfold-thickness prediction equations with a 4-compartment model. Am J Clin Nutr 2003;77:1186— 91.
Available online at www.sciencedirect.com
