Original Paper

Body Fat Percent by Bioelectrical Impedance Analysis and Risk of Coronary Artery Disease Among Urban Men with Low Rates of Obesity: The Indian Paradox
Ram B. Singh, MD, Mohammad Arif Niaz, PhD, Raheena Beegom, PhD, Gurpreet S Wander, MD, Amer S Thakur, PhD, Harbansh S Rissam, MD Centre of Nutrition, Medical Hospital and Research Centre, Moradabad, INDIA Key words: body fat, bioelectric impedance, cholesterol, triglycerides, obesity.
Objective: To determine the association between body fat percent and prevalence of coronary artery disease (CAD) and coronary risk factors in subjects with low rates of obesity. Subjects and Methods: We randomly selected 850 men, aged 25 to 64 years. The survey methods were questionnaire and bioelectrical impedance analysis for body composition. Subjects were divided into high-fat (n 357), over-fat (n 230), normal-fat (n 200) and under-fat (n 63) based on criteria of body-fat percent analysis. Results: The prevalence of CAD and the coronary-risk factors hypercholesterolemia, hypertension, diabetes, mellitus and sedentary lifestyle were significantly associated with high and moderate body fat percent despite low body-mass index (23.6 4.1 kg/m2). Mean total cholesterol, triglycerides and blood pressure were significantly associated with high and moderate body fat percent. The prevalence of smoking was weakly but inversely associated with high body-fat percent. Mean HDL cholesterol was positively associated with high body-fat percent. Body mass index was positively associated with high body-fat percent. Conclusions: High and moderate body-fat-percent subjects were associated with high prevalence of CAD and the coronary-risk factors hypertension, diabetes mellitus, higher body-mass index and sedentary lifestyle.

INTRODUCTION
Research indicates that those who suffer from excessive body weight and body fat have an increased prevalence of cardiovascular disease, diabetes and cancer [13]. In south Asian immigrants to Britain, central obesity in association with insulin resistance is an important risk factor of diabetes and cardiovascular disease [4 8]. In contrast and paradoxically, Indians have a low average body-mass index (20 30 kg/m2) and low rates of obesity (10 15%) in association with higher prevalence of cardiovascular disease and diabetes [79]. Mean waist-hip girth ratios were higher and trunk skinfolds thicker in south-Asian than in European men and women in the absence of corresponding ethnic differences in body-mass index or thigh skin folds [5,6]. In Indians, wherever they are living, glucose intolerance and insulin resistance are associated with obesity and especially with a pattern of obesity in which a high proportion of body fat is deposited on the trunk and in the

abdomen [4 9]. However, no study has examined the body composition and body-fat percent among south Asians by measuring bioelectrical impedance using a body fat analyzer [18]. The association of body-fat percent measured by bioelectrical methods with risk of coronary artery disease (CAD) is not known, although body-fat percent had consistent association in Americans with body mass index and dietary fat consumption [11]. In the present study, we examine, by bioelectrical impedance analysis, possibly for the first time, the association of body-fat percent with risk of CAD in an Indian urban population with low rates of obesity (body mass index 27 kg/m2, 13%).

SUBJECTS AND METHODS

The sampling frame for this study [12] consisted of the final population total of Moradabad city, which is 430,000 based on

Address reprint requests to: Dr. R. B. Singh, Hon. Prof. Preventive Cardiol, MHRC, Civil Lines, Moradabad-10 (UP) 244001, INDIA.

Journal of the American College of Nutrition, Vol. 18, No. 3, 268 273 (1999) Published by the American College of Nutrition 268

Body Fatness and Coronary Disease in India

a 1991 census. We excluded 16 suburban streets and randomly selected 20 out of the remaining 180 streets considered urban. We randomly selected 20 to 50 adult men by blindly selecting a card out of 50 to 150 cards from each street as described earlier [12]. Each street contained 6 to 12 blocks, and any block containing 100 to 300 subjects was designated a cluster. All subjects were informed and a verbal consent taken by means of pamphlets distributed to inform subjects of the surveys utility in the prevention of diseases. They were also given their individual results along with free advice and referred for evaluation and treatment if needed. We contacted 999 adult men aged 25 to 64 years out of which 95 (9.5%) were unable to give time for detailed histories and blood tests. The remaining 904 men were invited for this study; however, bioelectrical impedance analysis was done in 850 randomly selected subjects. A pretested and validated questionnaire [8], prepared according to guidelines of the World Health Organization [13] and other Indian studies [14], was used by the dietitian and a physician for collection of data from all the subjects. The questionnaire included information on age, socioeconomic status, physical activity, ethnicity, occupation, past history of diabetes, hypertension, chest pain (Rose questionnaire) and smoking. All subjects were graded in social classes one through five based on attributes of education, occupation, housing, consumer durables and per capita income [16]. Body composition was determined by a body-fat analyzer (BF 905, Maltron, UK) [10,11] in 850 subjects. The body-fat analyzer measures the flow of electrical signals as they pass through fat and lean areas and water in the body. When the amount of fat and lean matter or water changes, so do the signals, giving a reliable and accurate measure of the amount of each of these components that make up the total weight of the person. Body-fat-percent categories for our analysis were as follows: 10, under-body-fat, 10 to 20 normal-fat, 20 to 25 over-fat and 25 higher-body-fat. These categories corresponded approximately to percentiles 0 to 25, 26 to 54, 55 to 64 and 65 of the body-mass index. These values were also in accordance with height-weight tables of average values for Indian men advised by the Life Insurance Corporation of India. Physical activity was assessed by ascertaining occupational and spare-time physical activities, and subjects were then graded into sedentary, mild, moderate and heavy based on scores of activities as described earlier [12]. Height, weight and blood pressure were recorded in all the subjects [13]. Body weights were measured by the dietitian independently in light underclothes to the nearest of 0.5 kg. Height was measured in standing position. Waist and hip girths were measured in a standing position. Waist was measured as the smallest horizontal girth between the costal margin and the iliac crests and hip as the greatest circumference at the level of the greater trochanters. Blood pressure (systolic and diastolic phase V of Korotkoff) was measured in the right arm after five minutes rest with the cuff fitted while the subject was sitting. A standard mercury sphygmomanometer was used by the same physician for bloodpressure measurement in all the subjects. A final reading in lying position was recorded in all those subjects having high blood pressure 140/90 mmHg in the lying position after five minutes rest in accordance with World Health Organization Guidelines [13]. In all subjects a 12-lead electro-cardiogram was recorded for the diagnosis of CAD.

Diagnostic Criteria
The diagnosis of high blood pressure was made when the systolic blood pressure was 140 mmHg or more and the diastolic 90 mmHg or more in accordance with the guidelines of the WHO/ISH hypertension liaison committee [15]. Diabetes mellitus was diagnosed if the fasting blood glucose was more than 7.7 mmol/L ( 140 mg/dL) and postprandial blood glucose more than 11.1 mmol/L ( 200 mg/dL) two hours after intake of 75 grams of glucose orally. Serum cholesterol was considered higher if it was more than 5.18 mmol/L ( 200 mg/dL) and triglycerides higher if they were more than 2.08 mmol/L ( 185 mg/dL). Low high-density lipoprotein cholesterol was considered when it was less than 0.9 mmol/L ( 35 mg/dL) [14]. Body-mass index was calculated and obesity defined [1,14] as a body mass index of 25 kg/m2 and overweight as a body mass index of 25 kg/m2. Sedentary lifestyle was considered in cases of office workers or shopkeepers if they walked fewer than two kilometers a day and climbed fewer than 20 flights of stairs a week without any vigorous activity at least five days of the week based on Indian criteria of activities [16]. In India, tobacco is consumed by more than one means among them cigarettes, beedies, Indian pipes, raw tobacco and chewing tobacco. Users of any form of tobacco were categorized as smokers as has been done in other studies [17]. Alcohol intake is not common in India. The criteria [13] for the diagnosis of CAD were (a) history of angina or infarction and previously diagnosed disease, (b) affirmative response to a Rose questionnaire and (c) electrocardiographic findings, namely the Minnesota codes 1-1, 4-1, 5-9, 5-2 or 9-2. Presence of any of these criteria was taken as confirmation of the diagnosis of CAD. Individual clinical criteria such as known CAD, affirmative response to a Rose questionnaire and electrocardiographic changes (Q wave changes codes 1-1 and 1-2), ST segment depression or elevation codes 4-1, 4-2 and 9-2 and T wave inversions, codes 5-1 and 5-2, were also recorded.

Laboratory Data
Blood samples were collected in the lying position using a tourniquet after a overnight fast. Total cholesterol and triglycerides were estimated by enzymatic methods. High density lipoprotein cholesterol was estimated after precipitation of nonhigh-density lipoprotein cholesterol with managese heparin substrate. Low-density lipoprotein cholesterol was obtained

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

269

Body Fatness and Coronary Disease in India

using Friedewalds formula: low density lipoprotein total cholesterol high density lipoprotein (triglycerides/5). All blood samples were analysed on the day of blood collection. Body-mass index, total and high-density lipoprotein cholesterol, triglycerides and systolic blood pressure also showed an increasing trend with increase in body-fat percent. Diastolic blood pressure showed no significant association with body fat percent (Table 4). There was a significant positive rank correlation of the level of body-fat percent with age, body-fat weight, body weight, body mass index, waist-hip ratio and systolic blood pressure (Table 5). No significant rank correlation of body fat percent was observed with diastolic blood pressure, total cholesterol and low-density lipoprotein cholesterol. However, high-density lipoprotein cholesterol showed significant negative and triglycerides positive rank correlation with body fat percent.

Statistical Analysis
All p values were two tailed, and significance was taken as p 0.05. A relation was determined between body-fat percent and prevalence of CAD and hypertension, hypercholesterolemia, diabetes mellitus, smoking and sedentary lifestyle, and the significance of any trend was calculated with the MentelHaenzel X2 test. Body-mass index, systolic and diastolic blood pressure, total cholesterol, HDL cholesterol and triglycerides were determined at various levels of body fat percent, and the significance of trends was checked by Kendalls test. Spearmans coefficient of rank correlation (r) was calculated for status of body-fat percent with various clinical and biochemical risk factors. We used the SPSS statistical package (SPSS Inc, Chicago) for analysis of our data.

DISCUSSION
This study shows that CAD and coronary risk factors were significantly associated with the level of body-fat percent determined by bioelectrical impedance analysis in a cohort of urban men in north India. High fat and overfat subjects were associated with a high prevalence of CAD. Hypercholesterolemia, hypertension, diabetes mellitus and sedentary lifestyle were also more prevalent among high-fat and over-fat people. On the other hand smoking was more common among underfat people. There is scant evidence on the association of bodyfat percent with coronary risk factors [11,18]. Studies in developed countries [19 21] showed that insulin resistance and high-plasma triglycerides levels were important correlates of truncal obesity, which is an important risk factor for CAD in south Asians. These investigators proposed a link between regional fat accumulation, regional variation in adipocyte morphology and lipolytic activity and the metabolic profile. Since then, results from several epidemiological studies [3 8,21] have shown that a high proportion of fat on the trunk or abdomen was associated with an increased probability of developing cardiovascular disease in association with insulin resistance [22]. South Asians have a tendency toward abdominal obesity and insulin resistance [4 9]; however, no study has examined the association of body-fat percent with CAD. In one study [18] among 453 Thai females, the prevalence of overall obesity (27.0%) and central obesity (54.1%) (WHR 0.80%) was quite high. Total and low density lipoprotein cholesterol, triglycerides, apo B and fasting blood glucose showed significant increases with increase in body-mass index or body fat measured by bioelectric impedance. In another study [11] among 205 American female adults, intake of dietary fat was significantly related to adipocity and body-fat percent with control for multiple potential confounding factors. Changes in body-fat percent may be dependent on an imbalance between energy intake and energy expenditure. As Indians become affluent, urbanized and mechanized, the demand for physical activity declines and the energy is stored as

RESULTS
The body-fat percent categories of the various age groups of the 850 men are shown in Table 1. There was positive correlation of age with body-fat percent (Spearmans r men 0.45, p 0.01). We found an overall increase in the prevalence of clinical and electrocardiographic CAD among obese and overweight subjects based on body-fat-percent criteria, and the trend was significant as shown in Table 2. Social classes 1 to 3 (higher and middle socioeconomic status) were more common among over- and high-fat subjects than normal- and low-fat subjects respectively (91.1% vs. 45.6%, p 0.01). Higher social classes 1 to 3 were mainly from families of professionals, business persons, shopkeepers and skilled workers and social classes four and five had physically demanding occupations. About half of the men were Hindus and half were Moslems and access to health care was adequate. The prevalence of coronary risk factors among people of various body-fat percent is shown in Table 3. An increasing level of body-fat percent was associated with a significant rising trend in the prevalence of hypercholesterolemia, hypertension, diabetes mellitus and sedentary lifestyle. However, smoking was greater among normal- and under-fat subjects. Table 1. Distribution [n (%)] of Body Fat Percent Categories among Men of Different Age Groups
Age groups 25 to 34 35 to 44 45 to 54 55 to 64 TOTAL No. of Under body Normal fat Over fat Subjects fat ( 10%) (1020%) (2125%) 284 275 170 121 850 26 (9.1) 27 (9.8) 5 (2.9) 5 (4.1) 63 (7.4) High fat ( 25%)

88 (30.9) 70 (24.6) 100 (35.2) 62 (22.5) 70 (25.4) 116 (42.2) 26 (15.2) 55 (32.3) 84 (49.4) 24 (19.8) 35 (28.9) 57 (47.1) 200 (23.5) 230 (27.0) 357 (42.0)

270

VOL. 18, NO. 3

Body Fatness and Coronary Disease in India

Table 2. Body Fat Percent and Prevalence of Coronary Artery Disease in Men
Body fat percent Categories n (%) High-fat ( 25%) Over-fat (2025%) Normal-fat (1020%) Under-fat ( 10%) Total Mentel-Haenzel X2 p-value No. of Subjects 357 230 200 63 850 Clinical (known Rose Questionnaire) 45 (12.6) 13 (5.6) 6 (3.0) 64 (7.5) 15.63 0.001 Electrocardiographic (Silent) 22 (6.1) 5 (2.2) 4 (2.0) 2 (3.2) 33 (3.9) 12.15 0.01 Total

67 (18.7) 18 (7.8) 10 (5.0) 2 (3.2) 97 (11.4) 18.54 0.001

Table 3. Prevalence of Coronary Risk Factors in Relation to Body Fat Percent in Men
Body fat percent High-fat ( 25%) Over-fat (2025%) Normal-fat (1020%) Under-fat ( 10%) Total Mentel-Haenzel X2 p-value No. of subjects 357 230 200 63 850 10.52 Cholesterol ( 5.18 mmol/L) 177 (49.5) 95 (41.3) 25 (12.5) 4 (6.3) 301 (35.4) 11.68 0.01 Blood pressure ( 140/90 mmHg) 131 (36.6) 74 (32.1) 16 (8.1) 3 (4.7) 224 (26.3) 16.12 0.01 Diabetes mellitus 36 (10.0) 17 (7.3) 8 (4.0) 2 (3.1) 63 (7.4) 5.92 0.001 Smoking 88 (24.6) 57 (24.7) 55 (27.5) 26 (41.2) 226 (26.5) 9.12 0.05 Sedentary lifestyle 255 (71.4) 125 (54.3) 80 (40.0) 4 (6.3) 464 (54.5) 0.01

Table 4. Risk Factor Levels in Relation to Body Fat Percent

Total cholesterol (mmol/L) 5.42 0.8 5.08 0.6 4.56 0.6 3.62 0.5 0.058 2.76* HDL cholesterol (mmol/L) 1.16 0.3 1.17 0.2 1.18 0.2 1.13 0.1 0.088 2.46* Blood Pressures (mmHg) Systolic High-fat ( 25%) Over-fat (2025%) Normal-fat (1020%) Under-fat ( 10%) Kendalls T t-value 24.8 4.3 23.7 3.8 22.0 3.1 21.1 2.7 0.081 3.37** 1.92 0.6 1.66 0.5 1.34 0.5 1.26 0.4 0.098 2.61* 130 16 127 15 121 13 118 10 0.089 2.71* Diastolic 87 11 84 9 81 9 77 7 0.096 1.05

Body fat percent

Body mass index (kg/m )

Triglycerides (mmol/L)

Values in parentheses are mean 1 standard deviation. HDL High density lipoprotein, * p 0.01, ** p 0.001.

excess body fat [1,2]. There is increasing evidence that dietary fat is particularly more conducive to weight gain than complex carbohydrates. We have reported that Indian rural subjects consuming low-fat diets and having physically demanding occupations have significantly lower body-mass indices and lower rates of obesity than urban subjects consuming relatively higher fat diets [79]. It is possible that hepatic handling of insulin is modulated by habitual fat intake, and, on a high-fat diet, insulin secretion may rise to suppress hepatic glucose production resulting in insulin resistance characterized by greater body fat percent, particularly in the abdomen [79]. Higher fat diet appears to be a risk factor of higher body fat percent [11]. We found a positive correlation of age with body-fat percent (Table 1). The criteria for normal body fat may change with age. A high body-fat content for a 20-year-old man may be considered normal for someone who is 50 years old.

We also found a higher prevalence of hypertension, hypercholesterolemia, diabetes mellitus and sedentary lifestyle in association with greater body-fat percent. It is possible that higher body-fat percent predisposes these risk factors, mainly by causing insulin resistance and central obesity. However abdominal obesity based on waist-hip ratio among Thai women did not influence blood lipids which were related to body-fat percent [18]. It seems that excess body-fat percent in conjunction with a sedentary lifestyle, directly predisposes individuals to these risk factors. It is possible that excess of body-fat percent acts conjointly with other variables in increasing the prevalence of CAD. Presence of coronary risk factors among high-fat and over-fat subjects may increase the prevalence of CAD and the excess burden of ill health noted in other studies [23]. Magnetic resonance imaging, computarised axial tomography and body density derived from underwater weighing are

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

271

Body Fatness and Coronary Disease in India

Table 5. Mean Levels of Clinical and Biochemical Risk Factors and their Correlation with Body Fat Percent (Spearmans Rank Correlation)
4. Men Risk Factors Mean (standard deviation) 25.8 42.8 17.8 46.6 74.2 64.4 165.6 23.6 0.90 126 84 4.92 4.6 10 3.2 3.3 6.7 7 16 4.1 0.18 17 12 0.8 r 0.45* 0.08* 0.02 0.04 0.07 0.03 0.15* 0.14* 0.09* 0.08 0.04 0.06 0.08* 0.12* 10. 5. Susan BS, Hankinson SE, Hennekens CH, Speizer FE: Body weight and mortality among women. N Engl J Med 333:677685, 1995. Mckeigue PM, Ferrie JE, Pierpoint T, Marmot MG: Association of early onset coronary heart disease in south Asian men with glucose intolerance and hyperinsulinemia Circulation 87:152161, 1993. Mckeigue PM, Shah B, Marmot MG: Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in south Asians. Lancet 337:382386, 1991. Mckeigue PM, Pierpoint T, Ferrie JE, Marmot MG: Relationship of glucose intolerance and hyperinsulinemia to body fat patterns in south Asians and Europeans. Diabetologia 35:785791, 1992. Snehlatha C, Ramachandran A, Vijay V, Vishwanathan M: Differences in plasma insulin responses in urban and rural Indians: A study in south Indians. Diab Med 11:445448, 1994. Singh RB, Ghosh S, Niaz MA, Gupta S, Bishnoi I, Sharma JP: Epidemiologic study of diet and coronary risk factors in relation to central obesity and insulin levels in rural and urban populations of north India. Int J Cardiol 47:245255, 1995. Singh RB, Niaz MA, Rastogi V, Ghosh S, Beegom R, Rastogi SS, Postiglione A: Prevalence of coronary artery disease and coronary risk factors in the elderly rural and urban populations of north India. The Indian lifestyle and heart study in elderly. Cardiol Elder 4:111117, 1996. Kushner RF, Schoeller DA: Estimation of total body water by bioelectrical impedance analysis. Am J Clin Nutr 44:417424, 1996. Tuckner LA, Kano MJ: Dietary fat and body fat: A multivariate study of 205 adult females. Am J Clin Nutr 56:616622, 1992. Singh RB, Beegom R, Ghosh S, Niaz MA, Rastogi V, Rastogi SS, Singh NK: Epidemiological study of hypertension and its determinants in an urban population of north India. J Human Hyper 11:679685, 1997. Rose G, Blackburn H, Gillum RF, Prineas RJ: Cardiovascular Survey Methods. Geneva: World Health Organization, 1982. Indian Consensus Group. Indian consensus for prevention of hypertension and coronary artery disease: A joint scientific statement of Indian Society of Hypertension and International College of Nutrition. J Nutr Environ Med 6:309318, 1996. The Fifth Report of the WHO/ISH Mild Hypertension Liaison Committee: 1993, Guideline for the management of mild hypertension: Memorandum from a World Health Organization/ International Society of Hypertension meeting. Hypertension, 22: 392403, 1993. Singh RB, Ghosh S, Niaz MA, Rastogi V: Validation of physical activity and socioeconomic status questionnaire in relation to food intakes for the Five City Study and proposed classifications for Indians. J Asso Phys India 45:603607, 1997. Gupta R, Prakash H, Majumdar S, Sharma S, Gupta VP: Prevalence of coronary heart disease and coronary risk factors in an urban population of Rajasthan. Ind Heart J 47:331338, 1995. Leelahagul P, Soipet S, Achariyont P, Pakpeankitvalana R, Tanphaichitra V: Influence of body composition on risk factors for coronary heart disease in Thai women. Asia Pac J Clin Nutr 4:7980, 1995. Kissebah AH, Vydelingum N, Murray R: Relation of body fat distribution to metabolic complications of obesity. J Clin Endocrinol Metab 54:254260, 1962.

Body fat percent Mean age (years) Body fat weight (kg) Lean weight (kg) Lean mass percent Body weight (kg) Height (cm) Body mass index (kg/m2) Waist/hip ratio Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Total cholesterol (mmol/L) Low density lipoprotein cholesterol (mmol/L) High density lipoprotein cholesterol (mmol/L) Triglycerides (mmol/L)
* p 0.01, ** p 0.001.

6.

7.

8.

2.81 0.7 1.15 0.3 1.67 0.5

9.

other more accurate methods for measuring the body fat percent. However these methods are difficult and expensive for a population study when compared to bioelectrical impedance analysis [2,18]. We also observed that smoking was more common among under-fat subjects. Smoking is known to cause loss of appetite and decrease energy intake [24]. It is not clear why under-fat subjects, despite greater smoking, had a lower prevalence of CAD. It is possible that greater physical activity due to physically demanding occupations and low-fat diet may have protected these subjects against risk of CAD. We can not exclude possible bias; selection, nonresponse and measurement error, confounding by variables not measured, chance variation in a small sample and lack of generalizability to other populations or to women are the major limitations of this study. In conclusion, the findings of our study indicate that bodyfat percent was positively associated with CAD and the coronary risk factors hypertension, body mass index, diabetes mellitus and sedentary lifestyle. It is possible that increased physical activity and a decrease in body weight may be of benefit in the prevention of excess body-fat percent and CAD.

11. 12.

13. 14.

15.

16.

17.

REFERENCES
18. 1. World Health, Organization Study Group: Diet, Nutrition and Prevention of Chronic Diseases. Geneva: World Health Organization, 1990. 2. Bjorntorp P: Visceral obesity: A civilization syndrome. Obes Res 1:206222, 1993. 3. Manson JE, Willet WC, Stamfer MJ, Colditz GA, Hunter DJ,

19.

272

VOL. 18, NO. 3

Body Fatness and Coronary Disease in India

20. Despres JP, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C: Regional distribution of body fat, plasma lipoproteins and cardiovascular disease. Arteriosclerosis 10:497502, 1990. 21. Lee IM, Manson JE, Hennekens CH, Paffenbarger RS Jr. Body weight and mortality: A 27 year followup in the middle aged men. JAMA 270:28232828, 1993. 22. Krentz AJ: Insulin resistance. Br Med J 313:13851389, 1996. 23. Lean MEJ, Han TS, Seidell JC: Impairment of health and quality of life in people with large waist circumference. Lancet 351:853 856, 1998. 24. Singh RB, Ghosh S, Niaz MA, Rastogi V and Wander GS: Validation of tobacco and alcohol intake for the Five City Study and a proposed classification for Indians. J Asso Phys India 46:587592, 1998.

Received April 1998; revision accepted December 1998.

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

273