- Email: [email protected]
Diet, central obesity and prevalence of hypertension in the urban population of South India
ELSEVIER
International
Journal
of Cardiology
51 (1995)
183-191
Diet, central obesity and prevalence of hypertension in the urban population of South India Raheena Beegom*a, Razia Beegoma, Mohammad
A. Niaza,b3c, Ram B. Singha.bTc
“Department of Home Science, Trivandrum. bCentre of Nutrition Research, Moradabad, ‘Heart Research Laboratory. Moradabnd, Received
7 December
1994; revision
accepted
India India India 30 May
1995
Abstract Central obesity is a strong predictor of higher prevalence of diabetes, hypertension and coronary artery disease among Indian immigrants to Britain. To test this hypothesis in Indians, 1569 adults, between 25 and 64 years of age, from 750 randomly selected households (representative of 0.52 million population of Trivandrum city, Kerala) were selected for this study. The response rate was roughly 95% and the sample consisted of 1497 individuals (737 males and 760 females). The survey methods included dietary diaries for 7-day food intake record, blood pressure measurements using a mercury sphygmo-manometer and anthropometric measurements. The prevalence rates of hypertension between 25 and 64 years was 189/1000 (95% confidence limits 85-360) and between 45 and 64 years was 33YlOOO (95% confidence limits 210-460) which is higher than in Western populations. The prevalence was higher in males than females in the younger age groups and comparable in both sexes in the upper age groups. The prevalence of central obesity was significantly higher among male (77.2 vs. 48.9%) and female (84.0 vs. 51.4%) hypertensives compared to non-hypertensive subjects; however, mean body weight, body mass index and waist-hip ratio (WHR) were lower among Indian men compared to a British comparison group. Thus, comparison of Indian men with Britons showed that obesity, salt and alcohol intake, sedentariness, smoking and dietary fat intake do not explain the cause of higher prevalence of hypertension among South Indian men from Kerala. However, further analysis of data in relation to WHR showed that male subjects with higher quintiles of WHR (>0.85 and females >0.81) had significantly higher prevalence of hypertension and diabetes compared to the lowest quintile of WHR. Energy expenditure during daily physical activity for each upper quintile of WHR was lower, and dietary fat energy intake was the higher compared to the lowest quintile of WHR. These findings suggested that Indians could benefit by decreasing total fat intake to 21% kcal/day and by increasing physical activity with the aim of decreasing central obesity, to prevent hypertension in the community. Keywords: Diet; Central obesity; Hypertension;
* Corresponding author, Trivandrum-14. India.
1 TC Green
0167-5273/95/$09.50 0 1995 Elsevier SSDI 0167-5273(95)02402-I
Park
Science
Diabetes
Villa,
Jagathy,
Ireland
Ltd.
All rights
reserved
184
R. Beegom
et al. /International
Journal
1. Introduction
Epidemiologic studies [l-3] suggest that some component of animal products, possibly protein or fat may influence blood pressure in well nourished populations. As with obesity, there is clinical evidence of inter-individual variations in susceptibility to hypertension but the long-term interactions with dietary factors are less easy to discern [1,2]. A recent large scale multinational study [4] showed that a high body mass index and high atsohol intake had strong independent effects on olood pressure. In one study, more than 50% of patients with raised blood pressure had obesity [5]. In Indian immigrants to Britain, central obesity in association with insulin resistance are strong predictors of hypertension, diabetes and coronary artery disease [6]. Among Indian immigrants, Sikhs had higher median systolic blood pressure than Gujrati and Muslims [6]. In one study [7] from Delhi, the prevalence of hypertension was 127.5 per 1000 adults and in another study from Kerala 181, the prevalence was 161 per 1000. In a recent study [9] on the urban population from North India, the prevalence was 164 per 1000 adults. Males have a higher prevalence than females in the age group 25-64, after which the prevalence may be similar 17-91. In the Delhi survey, hypertension was present in 24.8% of obese adults compared to 8.2% in non-obese subjects [lo]. To our knowledge, there is no study from India on a random sample showing the relation of diet and central obesity with prevalence of hypertension. 2. Material
and methods
The geographical area of this study was Trivandrum city, South India. The sampling frame consisted of the final population total of the city, which is roughly 0.52 million given by the census of India (Census, 1991) [9]. After excluding 10 suburban streets, roughly 50 out of 226 streets were randomly selected by persons unconnected with this study, by blindly selecting 50 random numbers (with the help of computer) out of 226 numbers, each representing a street. Assuming from other studies that the prevalence
of Cardiology
51 (1995)
183-191
of hypertension is at least 5% in the community [9], we would need a sample of at least 1500 to detect with 90% confidence, at 5% significance, a relative risk of 1.54 for prevalent hypertension in India [ll]. This calculation is based on the assumption that the subjects have been selected by a simple random sampling. Since each household was expected to contain at least two adults above 25 years of age, our sample included 750 households. Thus with a sample of at least 1500 persons or 750 households, we would be able to estimate the prevalence to within 2% error on either side, which was deemed satisfactory as a first estimate. 2. I. Study design
The household was decided as the primary sampling unit and the block containing 50-100 households as a cluster. Each street contained 6- 12 blocks or clusters. Of 50 streets randomly selected, six were further selected for this study and from each of these, 12 blocks or clusters were randomly selected to obtain 750 households. The survey team consisted of nutritionists, health workers, research fellows and a physician. All were trained and briefed regarding the details of questionnaires before starting the survey. Before any household was declared non-contact or non-responder, at least three calls were made, one in the morning, one in the evening (around 18:00 h) and the last was made over the weekend. Of 750 households selected and personally contacted, 26 were noncontact or had been replaced by a family from another town. We successfully contacted 724 householders, comprising of 1569 persons between 25 and 64 years of age, of which 72 (4.6%) subjects refused to give a detailed history and examination. In the remaining 1497 subjects, detailed interviews were performed with the help of a pre-tested and validated [9] questionnaire by the nutritionists and a physician to obtain detailed information on dietary intake, age, socioeconomic status and past and family history of hypertension, diabetes, hypercholesterolemia, alcohol intake and smoking. There were 737 males and 760 females and the mean ages were comparable between the two groups (44.1 + 14.5 vs. 44.5 + 15.1 years). Mean body weight was significantly higher in males than
R. Beegom
et al. /International
females (64.1 f 5.6 vs. 56.4 f 4.6 kg). The socioeconomic status of the family was arbitrarily classified as high, middle and low based on housing condition, occupational income and number of dependents. Per capita income was calculated by dividing the total income of the family by total number of family members. Dietary intake was obtained for all individuals in the household by using food measures, food models and food portions; details were recorded in weekly diet diaries. A cross-check questionnaire was completed based on the weekly diet diaries by asking probing questions to assess exact intake of foods recorded in the diaries. All subjects were asked to record details of every food item taken at meals and between meals, around the time of eating. Detailed instructions were given to all subjects to maintain diet diaries and how to estimate portion sizes. Fruit, vegetables and legumes were weighed by all subjects before recording. Salt intake was assessed with the help of salt measures based on salt mixed in the food and added during eating. The criteria of diagnosis of risk factors and of hypertension were based on those of the World Health Organization [ 111. Essential hypertension was diagnosed in the presence of either systolic, diastolic or both blood pressures > 160/95 mmHg; overweight was considered for body mass index > 25 kg/m’, obesity for an index > 30 kg/m2 and central obesity in the presence of waist-hip ratio > 0.88 units. Alcohol intake was considered for an intake of > 60 g/week of alcohol and physical inactivity by < 1 km/day of walking during routine activity. Excess intake of salt was considered when salt intake exceeded 8.0 g/day. Diabetes mellitus was diagnosed by a positive glucose tolerance test showing fasting blood glucose > 140 mg/dl and post-prandial blood glucose, 2 h after 75 g oral glucose, > 200 mg/dl. A smoker (tobacco user) was defined as one smoking > 15 cigarettes or biri per day, or smoking hukka (pipe) > 15 times per day, or chewing tobacco; other smokers were considered mild smokers. Hypertriglyceridemia was considered in the presence of serum triglycerides > 185 mg/dl and lower high density lipoprotein cholesterol in the presence of <35 mg/dl. Hypercholesterolemia was considered in the presence of serum cholesterol >250 mg/dl.
Journal
of Cardiology
51 (1995)
183-191
185
2.2. Data collection
Clinical data and drug intake were recorded for all subjects. Blood pressures (systolic and diastolic phase V of Korotkoff) were measured in the right arm, after 5 min rest lying comfortably in the supine position, by a single mercury sphygmomano-meter and by the same physician for patients suspected of hypertension during the routine survey. However, during the routine survey, blood pressures were recorded in the sitting position. Three measurements 10 min apart were taken to represent systolic and diastolic blood pressure and the means of the last two readings were taken to represent actual systolic and diastolic blood pressure. By taking three recordings and using only the last two, we eliminated to a large extent the initial recording, which was likely to be higher as a result of apprehension or anxiety. Body weights (in light underclothes) were measured to the nearest 0.1 kg, by the health workers independently. Waist and hip girths were measured in the 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. Thigh girth was measured as the maximum girth around the upper thigh with the right foot resting on a chair so as to flex the hip and knee to right angles. Smoking habits were recorded in terms of weekly consumption of tobacco. Occupational and spare time physical activities were assessedin terms of weekly activities with the help of Indian classification of activities with the use of a 4-point scale; energy cost of the activities were calculated from the tables [3]. Nutrient intakes were calculated per 24 h by computation of the Indian food composition table [12], based on 7-day food intake record and the weight of fruit, vegetable and legume intake. 2.3. Laboratory
data
A venous blood sample was obtained after an overnight fast, for analysis of hemoglobin, blood counts, urea, glucose, albumin, triglycerides [ 131, total cholesterol [14] and high density lipoprotein cholesterol [14] in all subjects with the diagnosis of essential hypertension. Each patient was asked to drink 75 g anhydrous glucose in 200 ml of water
186 Table 1 Prevalence
R. Beegom
of hypertension
Age (years)
et al. /International
by age and sex in the urban
Males
25-34 35-44 45-54 55-64 Total Total between 45-64 years P-value obtained *P < 0.05.
Journal
population
Subjects
Hypertension W)
Subjects
Hypertension (“W
274 185 150 128 737 278
23 28 50 44 145 94
285 190 154 131 760 285
18 25 54 41 138 95
by comparison
(8.39)* (15.1) (33.3) (34.3) (19.6) (33.8)
of males with
(6.31) (13.1) (35.0) (31.2) (18.1) (33.3)
The prevalence of essential hypertension among
in patients
with
With
Total subjects Central obesity (WHR >0.88) Alcohol intake (> 60 g/week) Salt intake (a8 g/day) Sedentariness (< 1 km/day) Family history (by record) Higher and middle socioeconomic status Smoking ( > 10 cigarettes/day) P-value obtained subjects. *P < 0.05; **p
by comparison < 0.01.
51 (1995)
and without
India
Total subjects
Total
559 375 304 259 1497 563
41 53 104 85 283 189
hypertension,
expressed
Females
Total
145 112 62 61 110 22 125
138 116 (84.0)**
283 228 62 126 222 40 235
of prevalence
(%)
(7.3) (14.1) (34.2) (32.8) (18.4) (33.5)
as value
(%)
Without
Males
86 (59.3)*
hypertension
1497 adults (25-64 years) from Kerala, South India was 189/1000 (95% confidence limits 85-360). Half of the patients (n = 140) were known hypertensives. The prevalence of hypertension among 278 males and 285 females between 45 and 64 years of age were comparable (33.8 vs. 33.3%). The prevalence rate among all 563 subjects was 33.5% (Tables 1,2). The prevalence of central obesity and sedentariness were significantly higher among hypertensives than in non-hypertensive subjects
hypertension
(77.2)** (42.8)* (42.0) (75.8)** (15.1)* (86.2)**
183-191
females.
3. Results
of risk factors
of South
Females
and a second blood sample was collected after 2 h for analysis of glucose. Statistical analysis was performed using the paired t-test by comparing the variables in patients with hypertension and subjects without any risk factors. A P-value of c 0.05 was considered significant.
Table 2 Prevalence
of Cardiology
65 112 18 110
(47.1) (81.1)** (13.0)* (79.7)**
rate of risk factors
(80.5)** (21.9)* (44.5) (78.4)* (14.1)* (83.0)**
86 (30.3) in male and female
hypertension
Males
Females
Total
592 290 162 285 308 30 305
622 320 (51.4)
1214 610 162 61 I 643 55 637
(48.9) (27.3) (48.1) (52.0) (5.0) (51.5)
326 335 25 332
(52.4) (53.8) (4.0) (53.3)
290 (48.9) hypertensives
(50.2) (13.3) (50.3) (52.9) (4.5) (52.4)
290 (23.8) compared
to non-hypertensive
R. Beegom
et al. /International
Journal
and the majority of hypertensives were from higher and middle socioeconomic groups. Since we know that WHR is very different in men and women and varies by age, and the prevalence of risk factors also varies by age and sex, we have calculated sex-specific data in subjects between 45 and 64 years of age having maximum prevalence of hypertension. A comparison is made with British subjects of similar age and sex to determine the association of risk factors with hypertension in Indian and British population groups (Table 3). Indians are more often vegetarians than Britons. Energy intake in Indians and Britons were comparable but in Britons, the energy value of alcohol, Table 3 Clinical and anthropometric
data, risk factors
and nutrient
intake
of Cardiology
Number of subjects Age (years) Body weight (kg) Body mass index (kg/m2) Waist-hip girth ratio Thigh girth (cm) Suprailiac skinfold (mm) Median blood pressures (mmHg) Systolic Diastolic Prevalence of risk factors (n(%)) Hypertension (> 160/95 mmHg) Central obesity Smoking ( > 1O/day) Alcohol intake (> 60 g/week) Diabetes melhtus (self report) Non-vegetarians Dietary intake Total energy (kcal/day) Total carbohydrates (% kcabday) Total proteins (% kcabday) Total fat (% kcal/day) Saturated fat (% kcahday) Polyunsaturated/saturated fat ratio Total cholesterol (mg/day) Total Iibre (g/day) Total salt (g/day) Alcohol (g/week) Caffeine (mg/day) Total fruit, vegetable and legume (g/day)
by sex in urban
aNutrient
intakes
are average
urban
between
45 and 64 years of age and Britons
Britisha
Indian
1515 40-60 78.7 25.9 0.94 56.7 21.0
285 45-64 56.3 (4.2) 22.2 (2.6) 0.86 (0.12) 58.7 (5.4) 35.0 (8.8)
128 (124-130) 82 (80-84) (33.8) (50.3) (48.5) (34.8) (10.8) (74.1)
2180 (185) 56.5 (5.6) 14.5 (2.7) 29.0 (3.5) 14.2 (3.1) 0.41 (0.05) 225.1 (20.5) 24.0 (3.6) 9.9 (1.2) 46.5 (5.4) 250.0 (22.5) 150.5 (14.6)
of male and female.
Indians
Women
278 45-64 63.7 (5.0) 22.8 (3.0) 0.91 (0.15) 55.8 (5.8) 22.4 (6.5)
94 140 135 97 30 206
187
183-191
confectionary and soft drinks eaten outside were not included. Comparison of dietary intake in Indians and Britons [15] are shown in Table 3, which showed that there were more adverse factors in the Britons’ diet compared to that of Indians. It is clear from Table 4 that the prevalence of hypertension in the males was significantly higher in the subjects with greater WHR (> 0.88) than in the lower WHR groups and the prevalence rate remained higher after adjustment for diabetes and family history of hypertension. Table 5 shows that similar observations were made in females but the subjects with WHR > 0.88 were 1.5 times less common than among males. Underlying higher preva-
Men Indian
51 (1995)
121 78
urban
British 246 40-60 65.8 25.2 0.76 57.9 23.0
125 (123-128) 80 (78-82)
-
95 (33.3) 150 (52.6) 38.0
120 16 33.0
4.8 100
25 (8.7) 190 (66.6)
2210 44.9 12.0 42.2 18.5 0.28 405.0 17.9 9.6
2012 (180) 58.5 (6.2) 14.0 (2.6) 27.5 (3.4) 13.6 (3.2) 0.41 (0.05) 216.6 (18.5) 23.5 (3.5) 8.8 (1.1)
200 100
245 (24.0) 141.5 (12.2)
2.3 100
188
R. Beegom
Table 4 Prevalence Risk
of risk factors
in relation
et al. /International
to waist-hip
factor
ratio
Waist-hip
P-value obtained % (O/O). bMean (SD.). *P < 005. . . **p
by z-score
52 28 8 51 18 46
ratio
C (0.85-0.88)
65 32 IO 60 24 58
58 18 5 54 22 51
7 (13.4)* 1180 (llO)* 32.8 (4.2)*
(49.2)** (15.3)# (92.3)** (36.9) (89.2)*
8 (12.3)* 1292 (125)* 29.5 (3.8)*
by comparison
D (0.8 l-0.84)
8 (13.7)* 1380 (152) 24.6 (2.9)
of prevalence
53 12 5 25 16 40
(31.0) (8.6) (93.0) (37.9) (87.9)*
of risk factors
;
(22.6) (9.4) (47.1) (30.1) (75.4)
5 (9.4) 1510 (166) 21.4 (2.6) in group
50 4 2 14 17 25
(8.0) (4.0) (28) (34.0) (50.0)
2 (4.0) 1520 (170) 18.2 (2.1)
A and B with group
C or D.
< 0.01.
of risk factors
in relation
to waist-hip
ratio
Waist-hip
Risk factor
in female
ratio
A (>0.92) Sample size Hypertension (> 160/95 mmHg)” Diabetes mellitus (self report)” Sedentariness (< 1 km/week)” Alcohol intake (> 60 g/week)a Higher and middle socioeconomic status (questionnaire)” Family history (by record)a Energy expenditure (kcal/day)b Dietary fat intake (% kcal/day)b P-value obtained 52 (O/O). bMean (SD.). *P < 0.05; **p
183-191
45 and 64 years of age
1.89-0.92)
lence of hypertension among subjects with higher WHR in both sexes, was a higher prevalence of sedentariness and a lower expenditure of mean energy during daily physical activity. Hypertension was more common in the middle and higher
Table 5 Prevalence
between
51 (1995)
(n = 278)
(53.8)** (15.3)* (98.0)** (34.6) (68.4)*
test for proportions
of Cardiology
in male subjects
A (>0.92) Sample size Hypertension ( > 160/95 mmHg)B Diabetes meilitus (questionnaire)” Sedentariness (< 1 km/week)a Alcohol intake (> 60 g/week)a Higher and middle socioeconomic status (questionnaire)” Family history (by record)” Energy expenditure (kcal/day)b Dietary fat intake (% kcal/day)b
Journal
by z-score
< 0.01.
test proportion
socioeconomic groups who were also consuming higher amounts of energy from total fat calories compared to lower socioeconomic groups, who ate less calories from fat and had a lower prevalence of hypertension.
subjects
between
45 and 64 years of age
(n = 285) (Bo.89-0.92)
27 14 (51.9)** 5 (18.6)** 25 (92.6)*
C (0.85-0.88)
;.81-0.84)
54 27 (50.0) 7 (12.9)* 51 (94.4)*
56 17 (30.3) 4 (7. I) 50 (89.2)*
22 (81.6)*
45 (83.3)*
47 (83.9)*
7 (25.9)* 1095 (105); 30.8 (3.8)*
6 (I 1.1) 1190 (96)* 27.6 (3.5)*
-
90 31 (34.4) 6 (6.6) 76 (84.4)
58 6 (10.3) 3 (5.1) 25 (43.1)
73 (81.1)
32 (55. I)
-
-
by comparison
E <0.81
of group
6 (10.7) 1225 (126) 25.6 (3.0) A and B with
group
4 (4.4) 1305 (180) 23.4 (3.0) C or D.
2 (3.4) 1415 (128) 19.1 (2.8)
R. Beegom
et al. /International
Pearson correlation coefficients obtained for correlation betweeen dietary constituents and blood pressure showed that both systolic and diastolic blood pressures were positively related to total fat intake (I = 0.73, 0.86, P < 0.01). Central obesity was positively related to total fat intake (I = 0.41, P < 0.05) and negatively to fruit, vegetable and legume consumption (r = -0.35, P < 0.05). Salt consumption was not related to blood pressure. 4. Discussion
We assessed the prevalence of hypertension according to World Health Organization Criteria [l 11. All patients who were taking antihypertensive drugs were also included as hypertensive subjects. The prevalence rate between 25 and 64 years was 189/1000 (confidence limits 85-360) which was higher than the prevalence rate (161/1000) found in a village near Trivandrum city [S]. Males had a slightly higher prevalence than females in the age range 25-44 years, thereafter the prevalence was comparable in both sexes (Table 1). In the rural study [8], the gender pattern in the younger age groups was partly similar to our study but was reversed after the age of 54 years. In the Delhi survey [7], the prevalence was 127.5/1000 and in Ludhiana professionals, the prevalence rate between 30 and 60 years was 159/1000 [ 161. In a recent study from North India [20], the prevalence was 164/1000. These studies indicate that the prevalence rate of hypertension in the urban population of Kerala was greater than in North Indian cities. The prevalence rate between 45 and 64 years was higher and comparable in both sexes (335/1000, confidence limits 210-460). It is clear that the prevalence of hypertension increased with advancing age in both sexes in both South and North India. In the United States, the prevalence of hypertension in persons aged 45-64 years was 214/1000, which is lower than South India [16]. Prevalence of risk factors of hypertension such as central obesity was significantly higher, but the salt intake was similar among hypertensives compared to non-hypertensive subjects (Table 2). However, comparison of a group of urban Indians between 45 and 64 years of age with a British com-
Journal
of Cardiology
51 (1995)
183-191
189
parison group of similar age showed that mean body weight and body mass index were lower in Indian males and females. However in males, the WHR was not that much lower in Indian males compared to European males, (Tables 2, 3). The Indian female showed a pronounced tendency for central obesity evident in higher WHR, high girth and thicker suprailiac skin folds. Almost half of the Indian males (50.3%) and females (52.6%) had central obesity. The prevalence of central obesity and hypertension were not mentioned in the Britons, although median systolic and diastolic blood pressures were higher in Indians in both sexes than in Britons (Table 3). The prevalence of hypertension in a relatively younger age group (35-64 years) in Glasgow [ 171 was 32.0% in males and 25.4% in females. Smoking and alcohol intake were absent in Indian females and most Indians were vegetarians. When the nutrient intake of the Indians were compared with that of the British subjects, the most striking differences were the lower consumption of total and saturated fat and cholesterol and higher intake of polyunsaturated fat, tibre, caffeine and fruit, vegetable and legumes in the Indians. Similar comparisons of dietary intake were made between diets of Indian immigrants to Britain and British subjects in one study on the relation of diet and risk of coronary artery disease [ 151. The authors concluded that the diet survey did not explain why Indian immigrants should have a higher rate of coronary artery disease than the general British population. In our study, since dietary salt intake is comparable and fat intake is lower, in association with lower waisthip girth ratios in Indian urbans compared to Britons, it is logical to conclude according to British studies [6,15], that dietary salt and fat intake, alcohol intake and central obesity do not explain the cause of higher rate of hypertension in Indians (at least in males) than in the British population. However, in view of the marked differences in dietary intakes, mean body weight and body mass index and differences in prevalence of hypertension in the rural and urban populations of India [ 18-201, conclusions drawn from such comparisons may not be valid [19] and may require further analysis of data in the same population. There was a higher prevalence of non-insulin de-
190
R. Beegom
et al. /International
Journal
pendent diabetes mellitus in Indians of both sexes than in Britons, although the diagnosis of diabetes in our study was based on self reporting. In view of the existing evidence on the presence of a syndrome of glucose intolerance, hyperinsulinaemia, hypertension, low plasma high density lipoprotein cholesterol and high triglyceride, in association with central obesity in South Asian immigrants to Britain [6] and in Indians [21], further analysis of data was done to determine the relation of risk factors with WHR and to find out whether hypertension could be a manifestation of insulin resistance which is usually associated with sedentariness and central obesity (Tables 4, 5). It is clear that the prevalence of hypertension and diabetes mellitus were significantly higher in the group of subjects with higher WHR in both sexes than in subjects with lower WHR. It seems that the tendency for central obesity or higher WHR in Indian urban men is relatively modest, compared to Indian immigrants to Britain (0.91 vs. 0.98 units) because Indians are also less overweight than immigrants. It is possible that higher prevalence of hypertension and diabetes in association with a tendency for central obesity in Indians may be a manifestation of insulin resistance and reduced physical activity. Assay of plasma insulin levels may confirm our suggestion. Higher prevalence of diabetes [22], hypertension [8,19], and higher plasma insulin responses in Indian urbans have also been described in other studies [19,23]. However, this finding contrasts with other populations in whom a high prevalence of diabetes and hypertension occur in association with general obesity [l-6]. It seems that the increase in blood pressure is more likely to be due to the combination of increased insulin resistance and reduced physical activity rather than central obesity. The exact mechanism how central obesity develops in Indians and causes hypertension is not known. It seems that hyperinsulinemia may be a sort of compensatory mechanism [20] due to rapid increase in fat intake and reduced physical activity resulting in central obesity in populations adapted to survival under scarcity [6]. Even mild increase in plasma insulin may favour the development of hypertension and its haemodynamic consequences through insulin mediated increased catecholamine
of Cardiology
51 (1995)
183-191
and sympathetic activity. It is clear from Tables 4 and 5 that in both sexes subjects with higher WHR showed lower energy expenditure in the routine physical activity and higher consumption of energy from dietary fat (25-32 kcal!day) compared to subjects with lower WHR. These findings suggest that in Indians [l, 12,201 even modest increase in fat intake (which is common with better socioeconomic status) and reduced physical activity and, as a result, modest tendency of central obesity may not allow the physiological mechanisms to adapt effectively, resulting into rapid development of insulin resistance, hypertension, diabetes, coronary artery disease and atherosclerosis. These populations may benefit by increasing physical activity and reducing the consumption of fat to roughly 22% kcal/day [24]. It seems that the amount of fat consumption which is considered safe in Western populations (< 30% kcal/day) may have adverse effect in Indians and therefore fat intake should be 21-24% kcal/day of total energy intake [20,24]. References
111 World
Health Organization Study Group. Diet, nutrition and prevention of chronic diseases. Geneva: World Health Organization, 1990. 121 World Health Organization Scientific Group. Cardiovascular disease risk factors: new areas for research. Geneva: World Health Organization, 1993. Expert Committee. Rehabili[31 World Health Organization tation after cardiovascular diseases with special emphasis on developing countries. Geneva: World Health Organization, 1993. Research Group. Interrsalt: an in[41 Intersalt Cooperative ternational study of electrolyte excretion and blood pressure. Results of 24 h urinary sodium and potassium excretion. Br Med 3 1988; 297: 319-328. PI MC Mohan S, Cutler J, Britain E, Higgins M. Obesity and hypertension: epidemiological an clinical issues. Eur Heart J 1987; 7 (Suppl B): 57-70. 161 Mckeigue PM, Shah B, Marmot MC. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancer 1991; 337: 382-386. S, Ramachandran K, Kaul [71 Chadda SL, Radhakrishnan U, Gopinath V. Prevalence, awareness and treatment status of hypertension in urban population of Delhi. Indian J Med Res 1990; 92: 233-240. KG, Jayasree AK, 181 Raman Kutty V, Balakrishnan Thomas J. Prevalence of coronary heart disease in the
R. Beegom
et al. /International
rural population of Thiruvanananthapuram district, Kerala, India. Int J Cardiol 1993; 39: 59-70. [9] Beegum R, Singh RB. Prevalence of coronary heart disease and its risk factors in the urban population of South and North India. Acta Cardiol 1994. In press. [IO] Gopinath N, Chadda SL, Jain P, Shekhawat S, Tandon R. An epidemiological study of obesity in adults in the urban population of Delhi. J Assoc Phys India 1994; 42: 212-21s. 1111 Rose GA, Blackbum H, Gillum RF, Prineas RJ. Cardiovascular survey methods. Geneva: World Health Organization, 1982. [12] Narsingrao BS, Deosthale YG, Pant KC. Nutrient composition of Indian foods. Hyderabad: National Institute of Nutrition, 1989. [13] Vanhandle E, Zilversmitt BD. Micromethod in the direct estimation of serum triglycerides. J Lab Clin Med 1957; 60: 152-159. [14] Wilson DE, Spiger MJ. A dual precipitation method for quantitative plasma lipoprotein measurement without centrifugation. J Lab Clin Med 1973; 82: 413-482. [15] Mckeigue PM, Marmot MG, Adolstein AM. Diet and risk factors for coronary heart disease in Asians in northwest London. Lancet 1985; ii: 1086-1090. [16] Allan C, Brotman H. Chart book on aging in America. Washington, DC: The 1981 White House Conference on Aging, 1981. (171 The WHO MONICA Project. Geographical variations in the major risk factors of coronary heart disease in men and women aged 35-64 years. World Health Statist 1988; 41: 115-140.
Journal
of Cardiology [IS]
51 (1995)
183-191
191
Nissinen A, Bothig S, Granroth H, Lopez AD. Hypertension in developing countries. World Health Statist 1988; 41: 141-154. [19] Sharma BK, Arora OP, Bansal BL, Sagar S, Khurana SP. Hypertension among industrial workers and professional classes in Ludhiana, Punjab. Indian Heart J 1985; 37: 380-385. [20] Singh RB, Ghosh S, Naiz MA, Gupta S, Bishnoi I, Sharma JP, Agarwal P, Rastogi SS, Beegam R, Chibo H, Shoumin Z. Epidemidogic study of diet and coronary risk factors in relation to central obesity and insulin levels in rural and urban population of North India. Int J Cardiol 1995;47: 245-255. [21] Bhatnagar MK, Ganeh R, Goel A, Verma NPS, Chauhan UPS. A study of hyperinsulinemia in Indian hypertensive subjects. J Assoc Phys India 1994; 42: 522-524. 1221 Verma NPS, Mehta SL, Madhu S, Mather HM, Keen H. Prevalence of known diabetes in urban Indian environment: the Darya Ganj diabetes survey. Br Med J 1986; 293: 423-424. [23] Snehlatha C, Ramchandran A, Vijaya V, Viswanathan M. Differences in plasma insulin responses in urban and rural Indians: a study in Southern Indians. Diabetic Med 1994; 11: 445-448. [24] Singh RB, Niaz MA, Ghosh S. Effect on central obesity and associated disturbances of low energy, fruit and vegetable-enriched prudent diet in North Indians. Postgrad Med J 1994; 70: 895-900.
International
Journal
of Cardiology
51 (1995)
183-191
Diet, central obesity and prevalence of hypertension in the urban population of South India Raheena Beegom*a, Razia Beegoma, Mohammad
A. Niaza,b3c, Ram B. Singha.bTc
“Department of Home Science, Trivandrum. bCentre of Nutrition Research, Moradabad, ‘Heart Research Laboratory. Moradabnd, Received
7 December
1994; revision
accepted
India India India 30 May
1995
Abstract Central obesity is a strong predictor of higher prevalence of diabetes, hypertension and coronary artery disease among Indian immigrants to Britain. To test this hypothesis in Indians, 1569 adults, between 25 and 64 years of age, from 750 randomly selected households (representative of 0.52 million population of Trivandrum city, Kerala) were selected for this study. The response rate was roughly 95% and the sample consisted of 1497 individuals (737 males and 760 females). The survey methods included dietary diaries for 7-day food intake record, blood pressure measurements using a mercury sphygmo-manometer and anthropometric measurements. The prevalence rates of hypertension between 25 and 64 years was 189/1000 (95% confidence limits 85-360) and between 45 and 64 years was 33YlOOO (95% confidence limits 210-460) which is higher than in Western populations. The prevalence was higher in males than females in the younger age groups and comparable in both sexes in the upper age groups. The prevalence of central obesity was significantly higher among male (77.2 vs. 48.9%) and female (84.0 vs. 51.4%) hypertensives compared to non-hypertensive subjects; however, mean body weight, body mass index and waist-hip ratio (WHR) were lower among Indian men compared to a British comparison group. Thus, comparison of Indian men with Britons showed that obesity, salt and alcohol intake, sedentariness, smoking and dietary fat intake do not explain the cause of higher prevalence of hypertension among South Indian men from Kerala. However, further analysis of data in relation to WHR showed that male subjects with higher quintiles of WHR (>0.85 and females >0.81) had significantly higher prevalence of hypertension and diabetes compared to the lowest quintile of WHR. Energy expenditure during daily physical activity for each upper quintile of WHR was lower, and dietary fat energy intake was the higher compared to the lowest quintile of WHR. These findings suggested that Indians could benefit by decreasing total fat intake to 21% kcal/day and by increasing physical activity with the aim of decreasing central obesity, to prevent hypertension in the community. Keywords: Diet; Central obesity; Hypertension;
* Corresponding author, Trivandrum-14. India.
1 TC Green
0167-5273/95/$09.50 0 1995 Elsevier SSDI 0167-5273(95)02402-I
Park
Science
Diabetes
Villa,
Jagathy,
Ireland
Ltd.
All rights
reserved
184
R. Beegom
et al. /International
Journal
1. Introduction
Epidemiologic studies [l-3] suggest that some component of animal products, possibly protein or fat may influence blood pressure in well nourished populations. As with obesity, there is clinical evidence of inter-individual variations in susceptibility to hypertension but the long-term interactions with dietary factors are less easy to discern [1,2]. A recent large scale multinational study [4] showed that a high body mass index and high atsohol intake had strong independent effects on olood pressure. In one study, more than 50% of patients with raised blood pressure had obesity [5]. In Indian immigrants to Britain, central obesity in association with insulin resistance are strong predictors of hypertension, diabetes and coronary artery disease [6]. Among Indian immigrants, Sikhs had higher median systolic blood pressure than Gujrati and Muslims [6]. In one study [7] from Delhi, the prevalence of hypertension was 127.5 per 1000 adults and in another study from Kerala 181, the prevalence was 161 per 1000. In a recent study [9] on the urban population from North India, the prevalence was 164 per 1000 adults. Males have a higher prevalence than females in the age group 25-64, after which the prevalence may be similar 17-91. In the Delhi survey, hypertension was present in 24.8% of obese adults compared to 8.2% in non-obese subjects [lo]. To our knowledge, there is no study from India on a random sample showing the relation of diet and central obesity with prevalence of hypertension. 2. Material
and methods
The geographical area of this study was Trivandrum city, South India. The sampling frame consisted of the final population total of the city, which is roughly 0.52 million given by the census of India (Census, 1991) [9]. After excluding 10 suburban streets, roughly 50 out of 226 streets were randomly selected by persons unconnected with this study, by blindly selecting 50 random numbers (with the help of computer) out of 226 numbers, each representing a street. Assuming from other studies that the prevalence
of Cardiology
51 (1995)
183-191
of hypertension is at least 5% in the community [9], we would need a sample of at least 1500 to detect with 90% confidence, at 5% significance, a relative risk of 1.54 for prevalent hypertension in India [ll]. This calculation is based on the assumption that the subjects have been selected by a simple random sampling. Since each household was expected to contain at least two adults above 25 years of age, our sample included 750 households. Thus with a sample of at least 1500 persons or 750 households, we would be able to estimate the prevalence to within 2% error on either side, which was deemed satisfactory as a first estimate. 2. I. Study design
The household was decided as the primary sampling unit and the block containing 50-100 households as a cluster. Each street contained 6- 12 blocks or clusters. Of 50 streets randomly selected, six were further selected for this study and from each of these, 12 blocks or clusters were randomly selected to obtain 750 households. The survey team consisted of nutritionists, health workers, research fellows and a physician. All were trained and briefed regarding the details of questionnaires before starting the survey. Before any household was declared non-contact or non-responder, at least three calls were made, one in the morning, one in the evening (around 18:00 h) and the last was made over the weekend. Of 750 households selected and personally contacted, 26 were noncontact or had been replaced by a family from another town. We successfully contacted 724 householders, comprising of 1569 persons between 25 and 64 years of age, of which 72 (4.6%) subjects refused to give a detailed history and examination. In the remaining 1497 subjects, detailed interviews were performed with the help of a pre-tested and validated [9] questionnaire by the nutritionists and a physician to obtain detailed information on dietary intake, age, socioeconomic status and past and family history of hypertension, diabetes, hypercholesterolemia, alcohol intake and smoking. There were 737 males and 760 females and the mean ages were comparable between the two groups (44.1 + 14.5 vs. 44.5 + 15.1 years). Mean body weight was significantly higher in males than
R. Beegom
et al. /International
females (64.1 f 5.6 vs. 56.4 f 4.6 kg). The socioeconomic status of the family was arbitrarily classified as high, middle and low based on housing condition, occupational income and number of dependents. Per capita income was calculated by dividing the total income of the family by total number of family members. Dietary intake was obtained for all individuals in the household by using food measures, food models and food portions; details were recorded in weekly diet diaries. A cross-check questionnaire was completed based on the weekly diet diaries by asking probing questions to assess exact intake of foods recorded in the diaries. All subjects were asked to record details of every food item taken at meals and between meals, around the time of eating. Detailed instructions were given to all subjects to maintain diet diaries and how to estimate portion sizes. Fruit, vegetables and legumes were weighed by all subjects before recording. Salt intake was assessed with the help of salt measures based on salt mixed in the food and added during eating. The criteria of diagnosis of risk factors and of hypertension were based on those of the World Health Organization [ 111. Essential hypertension was diagnosed in the presence of either systolic, diastolic or both blood pressures > 160/95 mmHg; overweight was considered for body mass index > 25 kg/m’, obesity for an index > 30 kg/m2 and central obesity in the presence of waist-hip ratio > 0.88 units. Alcohol intake was considered for an intake of > 60 g/week of alcohol and physical inactivity by < 1 km/day of walking during routine activity. Excess intake of salt was considered when salt intake exceeded 8.0 g/day. Diabetes mellitus was diagnosed by a positive glucose tolerance test showing fasting blood glucose > 140 mg/dl and post-prandial blood glucose, 2 h after 75 g oral glucose, > 200 mg/dl. A smoker (tobacco user) was defined as one smoking > 15 cigarettes or biri per day, or smoking hukka (pipe) > 15 times per day, or chewing tobacco; other smokers were considered mild smokers. Hypertriglyceridemia was considered in the presence of serum triglycerides > 185 mg/dl and lower high density lipoprotein cholesterol in the presence of <35 mg/dl. Hypercholesterolemia was considered in the presence of serum cholesterol >250 mg/dl.
Journal
of Cardiology
51 (1995)
183-191
185
2.2. Data collection
Clinical data and drug intake were recorded for all subjects. Blood pressures (systolic and diastolic phase V of Korotkoff) were measured in the right arm, after 5 min rest lying comfortably in the supine position, by a single mercury sphygmomano-meter and by the same physician for patients suspected of hypertension during the routine survey. However, during the routine survey, blood pressures were recorded in the sitting position. Three measurements 10 min apart were taken to represent systolic and diastolic blood pressure and the means of the last two readings were taken to represent actual systolic and diastolic blood pressure. By taking three recordings and using only the last two, we eliminated to a large extent the initial recording, which was likely to be higher as a result of apprehension or anxiety. Body weights (in light underclothes) were measured to the nearest 0.1 kg, by the health workers independently. Waist and hip girths were measured in the 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. Thigh girth was measured as the maximum girth around the upper thigh with the right foot resting on a chair so as to flex the hip and knee to right angles. Smoking habits were recorded in terms of weekly consumption of tobacco. Occupational and spare time physical activities were assessedin terms of weekly activities with the help of Indian classification of activities with the use of a 4-point scale; energy cost of the activities were calculated from the tables [3]. Nutrient intakes were calculated per 24 h by computation of the Indian food composition table [12], based on 7-day food intake record and the weight of fruit, vegetable and legume intake. 2.3. Laboratory
data
A venous blood sample was obtained after an overnight fast, for analysis of hemoglobin, blood counts, urea, glucose, albumin, triglycerides [ 131, total cholesterol [14] and high density lipoprotein cholesterol [14] in all subjects with the diagnosis of essential hypertension. Each patient was asked to drink 75 g anhydrous glucose in 200 ml of water
186 Table 1 Prevalence
R. Beegom
of hypertension
Age (years)
et al. /International
by age and sex in the urban
Males
25-34 35-44 45-54 55-64 Total Total between 45-64 years P-value obtained *P < 0.05.
Journal
population
Subjects
Hypertension W)
Subjects
Hypertension (“W
274 185 150 128 737 278
23 28 50 44 145 94
285 190 154 131 760 285
18 25 54 41 138 95
by comparison
(8.39)* (15.1) (33.3) (34.3) (19.6) (33.8)
of males with
(6.31) (13.1) (35.0) (31.2) (18.1) (33.3)
The prevalence of essential hypertension among
in patients
with
With
Total subjects Central obesity (WHR >0.88) Alcohol intake (> 60 g/week) Salt intake (a8 g/day) Sedentariness (< 1 km/day) Family history (by record) Higher and middle socioeconomic status Smoking ( > 10 cigarettes/day) P-value obtained subjects. *P < 0.05; **p
by comparison < 0.01.
51 (1995)
and without
India
Total subjects
Total
559 375 304 259 1497 563
41 53 104 85 283 189
hypertension,
expressed
Females
Total
145 112 62 61 110 22 125
138 116 (84.0)**
283 228 62 126 222 40 235
of prevalence
(%)
(7.3) (14.1) (34.2) (32.8) (18.4) (33.5)
as value
(%)
Without
Males
86 (59.3)*
hypertension
1497 adults (25-64 years) from Kerala, South India was 189/1000 (95% confidence limits 85-360). Half of the patients (n = 140) were known hypertensives. The prevalence of hypertension among 278 males and 285 females between 45 and 64 years of age were comparable (33.8 vs. 33.3%). The prevalence rate among all 563 subjects was 33.5% (Tables 1,2). The prevalence of central obesity and sedentariness were significantly higher among hypertensives than in non-hypertensive subjects
hypertension
(77.2)** (42.8)* (42.0) (75.8)** (15.1)* (86.2)**
183-191
females.
3. Results
of risk factors
of South
Females
and a second blood sample was collected after 2 h for analysis of glucose. Statistical analysis was performed using the paired t-test by comparing the variables in patients with hypertension and subjects without any risk factors. A P-value of c 0.05 was considered significant.
Table 2 Prevalence
of Cardiology
65 112 18 110
(47.1) (81.1)** (13.0)* (79.7)**
rate of risk factors
(80.5)** (21.9)* (44.5) (78.4)* (14.1)* (83.0)**
86 (30.3) in male and female
hypertension
Males
Females
Total
592 290 162 285 308 30 305
622 320 (51.4)
1214 610 162 61 I 643 55 637
(48.9) (27.3) (48.1) (52.0) (5.0) (51.5)
326 335 25 332
(52.4) (53.8) (4.0) (53.3)
290 (48.9) hypertensives
(50.2) (13.3) (50.3) (52.9) (4.5) (52.4)
290 (23.8) compared
to non-hypertensive
R. Beegom
et al. /International
Journal
and the majority of hypertensives were from higher and middle socioeconomic groups. Since we know that WHR is very different in men and women and varies by age, and the prevalence of risk factors also varies by age and sex, we have calculated sex-specific data in subjects between 45 and 64 years of age having maximum prevalence of hypertension. A comparison is made with British subjects of similar age and sex to determine the association of risk factors with hypertension in Indian and British population groups (Table 3). Indians are more often vegetarians than Britons. Energy intake in Indians and Britons were comparable but in Britons, the energy value of alcohol, Table 3 Clinical and anthropometric
data, risk factors
and nutrient
intake
of Cardiology
Number of subjects Age (years) Body weight (kg) Body mass index (kg/m2) Waist-hip girth ratio Thigh girth (cm) Suprailiac skinfold (mm) Median blood pressures (mmHg) Systolic Diastolic Prevalence of risk factors (n(%)) Hypertension (> 160/95 mmHg) Central obesity Smoking ( > 1O/day) Alcohol intake (> 60 g/week) Diabetes melhtus (self report) Non-vegetarians Dietary intake Total energy (kcal/day) Total carbohydrates (% kcabday) Total proteins (% kcabday) Total fat (% kcal/day) Saturated fat (% kcahday) Polyunsaturated/saturated fat ratio Total cholesterol (mg/day) Total Iibre (g/day) Total salt (g/day) Alcohol (g/week) Caffeine (mg/day) Total fruit, vegetable and legume (g/day)
by sex in urban
aNutrient
intakes
are average
urban
between
45 and 64 years of age and Britons
Britisha
Indian
1515 40-60 78.7 25.9 0.94 56.7 21.0
285 45-64 56.3 (4.2) 22.2 (2.6) 0.86 (0.12) 58.7 (5.4) 35.0 (8.8)
128 (124-130) 82 (80-84) (33.8) (50.3) (48.5) (34.8) (10.8) (74.1)
2180 (185) 56.5 (5.6) 14.5 (2.7) 29.0 (3.5) 14.2 (3.1) 0.41 (0.05) 225.1 (20.5) 24.0 (3.6) 9.9 (1.2) 46.5 (5.4) 250.0 (22.5) 150.5 (14.6)
of male and female.
Indians
Women
278 45-64 63.7 (5.0) 22.8 (3.0) 0.91 (0.15) 55.8 (5.8) 22.4 (6.5)
94 140 135 97 30 206
187
183-191
confectionary and soft drinks eaten outside were not included. Comparison of dietary intake in Indians and Britons [15] are shown in Table 3, which showed that there were more adverse factors in the Britons’ diet compared to that of Indians. It is clear from Table 4 that the prevalence of hypertension in the males was significantly higher in the subjects with greater WHR (> 0.88) than in the lower WHR groups and the prevalence rate remained higher after adjustment for diabetes and family history of hypertension. Table 5 shows that similar observations were made in females but the subjects with WHR > 0.88 were 1.5 times less common than among males. Underlying higher preva-
Men Indian
51 (1995)
121 78
urban
British 246 40-60 65.8 25.2 0.76 57.9 23.0
125 (123-128) 80 (78-82)
-
95 (33.3) 150 (52.6) 38.0
120 16 33.0
4.8 100
25 (8.7) 190 (66.6)
2210 44.9 12.0 42.2 18.5 0.28 405.0 17.9 9.6
2012 (180) 58.5 (6.2) 14.0 (2.6) 27.5 (3.4) 13.6 (3.2) 0.41 (0.05) 216.6 (18.5) 23.5 (3.5) 8.8 (1.1)
200 100
245 (24.0) 141.5 (12.2)
2.3 100
188
R. Beegom
Table 4 Prevalence Risk
of risk factors
in relation
et al. /International
to waist-hip
factor
ratio
Waist-hip
P-value obtained % (O/O). bMean (SD.). *P < 005. . . **p
by z-score
52 28 8 51 18 46
ratio
C (0.85-0.88)
65 32 IO 60 24 58
58 18 5 54 22 51
7 (13.4)* 1180 (llO)* 32.8 (4.2)*
(49.2)** (15.3)# (92.3)** (36.9) (89.2)*
8 (12.3)* 1292 (125)* 29.5 (3.8)*
by comparison
D (0.8 l-0.84)
8 (13.7)* 1380 (152) 24.6 (2.9)
of prevalence
53 12 5 25 16 40
(31.0) (8.6) (93.0) (37.9) (87.9)*
of risk factors
;
(22.6) (9.4) (47.1) (30.1) (75.4)
5 (9.4) 1510 (166) 21.4 (2.6) in group
50 4 2 14 17 25
(8.0) (4.0) (28) (34.0) (50.0)
2 (4.0) 1520 (170) 18.2 (2.1)
A and B with group
C or D.
< 0.01.
of risk factors
in relation
to waist-hip
ratio
Waist-hip
Risk factor
in female
ratio
A (>0.92) Sample size Hypertension (> 160/95 mmHg)” Diabetes mellitus (self report)” Sedentariness (< 1 km/week)” Alcohol intake (> 60 g/week)a Higher and middle socioeconomic status (questionnaire)” Family history (by record)a Energy expenditure (kcal/day)b Dietary fat intake (% kcal/day)b P-value obtained 52 (O/O). bMean (SD.). *P < 0.05; **p
183-191
45 and 64 years of age
1.89-0.92)
lence of hypertension among subjects with higher WHR in both sexes, was a higher prevalence of sedentariness and a lower expenditure of mean energy during daily physical activity. Hypertension was more common in the middle and higher
Table 5 Prevalence
between
51 (1995)
(n = 278)
(53.8)** (15.3)* (98.0)** (34.6) (68.4)*
test for proportions
of Cardiology
in male subjects
A (>0.92) Sample size Hypertension ( > 160/95 mmHg)B Diabetes meilitus (questionnaire)” Sedentariness (< 1 km/week)a Alcohol intake (> 60 g/week)a Higher and middle socioeconomic status (questionnaire)” Family history (by record)” Energy expenditure (kcal/day)b Dietary fat intake (% kcal/day)b
Journal
by z-score
< 0.01.
test proportion
socioeconomic groups who were also consuming higher amounts of energy from total fat calories compared to lower socioeconomic groups, who ate less calories from fat and had a lower prevalence of hypertension.
subjects
between
45 and 64 years of age
(n = 285) (Bo.89-0.92)
27 14 (51.9)** 5 (18.6)** 25 (92.6)*
C (0.85-0.88)
;.81-0.84)
54 27 (50.0) 7 (12.9)* 51 (94.4)*
56 17 (30.3) 4 (7. I) 50 (89.2)*
22 (81.6)*
45 (83.3)*
47 (83.9)*
7 (25.9)* 1095 (105); 30.8 (3.8)*
6 (I 1.1) 1190 (96)* 27.6 (3.5)*
-
90 31 (34.4) 6 (6.6) 76 (84.4)
58 6 (10.3) 3 (5.1) 25 (43.1)
73 (81.1)
32 (55. I)
-
-
by comparison
E <0.81
of group
6 (10.7) 1225 (126) 25.6 (3.0) A and B with
group
4 (4.4) 1305 (180) 23.4 (3.0) C or D.
2 (3.4) 1415 (128) 19.1 (2.8)
R. Beegom
et al. /International
Pearson correlation coefficients obtained for correlation betweeen dietary constituents and blood pressure showed that both systolic and diastolic blood pressures were positively related to total fat intake (I = 0.73, 0.86, P < 0.01). Central obesity was positively related to total fat intake (I = 0.41, P < 0.05) and negatively to fruit, vegetable and legume consumption (r = -0.35, P < 0.05). Salt consumption was not related to blood pressure. 4. Discussion
We assessed the prevalence of hypertension according to World Health Organization Criteria [l 11. All patients who were taking antihypertensive drugs were also included as hypertensive subjects. The prevalence rate between 25 and 64 years was 189/1000 (confidence limits 85-360) which was higher than the prevalence rate (161/1000) found in a village near Trivandrum city [S]. Males had a slightly higher prevalence than females in the age range 25-44 years, thereafter the prevalence was comparable in both sexes (Table 1). In the rural study [8], the gender pattern in the younger age groups was partly similar to our study but was reversed after the age of 54 years. In the Delhi survey [7], the prevalence was 127.5/1000 and in Ludhiana professionals, the prevalence rate between 30 and 60 years was 159/1000 [ 161. In a recent study from North India [20], the prevalence was 164/1000. These studies indicate that the prevalence rate of hypertension in the urban population of Kerala was greater than in North Indian cities. The prevalence rate between 45 and 64 years was higher and comparable in both sexes (335/1000, confidence limits 210-460). It is clear that the prevalence of hypertension increased with advancing age in both sexes in both South and North India. In the United States, the prevalence of hypertension in persons aged 45-64 years was 214/1000, which is lower than South India [16]. Prevalence of risk factors of hypertension such as central obesity was significantly higher, but the salt intake was similar among hypertensives compared to non-hypertensive subjects (Table 2). However, comparison of a group of urban Indians between 45 and 64 years of age with a British com-
Journal
of Cardiology
51 (1995)
183-191
189
parison group of similar age showed that mean body weight and body mass index were lower in Indian males and females. However in males, the WHR was not that much lower in Indian males compared to European males, (Tables 2, 3). The Indian female showed a pronounced tendency for central obesity evident in higher WHR, high girth and thicker suprailiac skin folds. Almost half of the Indian males (50.3%) and females (52.6%) had central obesity. The prevalence of central obesity and hypertension were not mentioned in the Britons, although median systolic and diastolic blood pressures were higher in Indians in both sexes than in Britons (Table 3). The prevalence of hypertension in a relatively younger age group (35-64 years) in Glasgow [ 171 was 32.0% in males and 25.4% in females. Smoking and alcohol intake were absent in Indian females and most Indians were vegetarians. When the nutrient intake of the Indians were compared with that of the British subjects, the most striking differences were the lower consumption of total and saturated fat and cholesterol and higher intake of polyunsaturated fat, tibre, caffeine and fruit, vegetable and legumes in the Indians. Similar comparisons of dietary intake were made between diets of Indian immigrants to Britain and British subjects in one study on the relation of diet and risk of coronary artery disease [ 151. The authors concluded that the diet survey did not explain why Indian immigrants should have a higher rate of coronary artery disease than the general British population. In our study, since dietary salt intake is comparable and fat intake is lower, in association with lower waisthip girth ratios in Indian urbans compared to Britons, it is logical to conclude according to British studies [6,15], that dietary salt and fat intake, alcohol intake and central obesity do not explain the cause of higher rate of hypertension in Indians (at least in males) than in the British population. However, in view of the marked differences in dietary intakes, mean body weight and body mass index and differences in prevalence of hypertension in the rural and urban populations of India [ 18-201, conclusions drawn from such comparisons may not be valid [19] and may require further analysis of data in the same population. There was a higher prevalence of non-insulin de-
190
R. Beegom
et al. /International
Journal
pendent diabetes mellitus in Indians of both sexes than in Britons, although the diagnosis of diabetes in our study was based on self reporting. In view of the existing evidence on the presence of a syndrome of glucose intolerance, hyperinsulinaemia, hypertension, low plasma high density lipoprotein cholesterol and high triglyceride, in association with central obesity in South Asian immigrants to Britain [6] and in Indians [21], further analysis of data was done to determine the relation of risk factors with WHR and to find out whether hypertension could be a manifestation of insulin resistance which is usually associated with sedentariness and central obesity (Tables 4, 5). It is clear that the prevalence of hypertension and diabetes mellitus were significantly higher in the group of subjects with higher WHR in both sexes than in subjects with lower WHR. It seems that the tendency for central obesity or higher WHR in Indian urban men is relatively modest, compared to Indian immigrants to Britain (0.91 vs. 0.98 units) because Indians are also less overweight than immigrants. It is possible that higher prevalence of hypertension and diabetes in association with a tendency for central obesity in Indians may be a manifestation of insulin resistance and reduced physical activity. Assay of plasma insulin levels may confirm our suggestion. Higher prevalence of diabetes [22], hypertension [8,19], and higher plasma insulin responses in Indian urbans have also been described in other studies [19,23]. However, this finding contrasts with other populations in whom a high prevalence of diabetes and hypertension occur in association with general obesity [l-6]. It seems that the increase in blood pressure is more likely to be due to the combination of increased insulin resistance and reduced physical activity rather than central obesity. The exact mechanism how central obesity develops in Indians and causes hypertension is not known. It seems that hyperinsulinemia may be a sort of compensatory mechanism [20] due to rapid increase in fat intake and reduced physical activity resulting in central obesity in populations adapted to survival under scarcity [6]. Even mild increase in plasma insulin may favour the development of hypertension and its haemodynamic consequences through insulin mediated increased catecholamine
of Cardiology
51 (1995)
183-191
and sympathetic activity. It is clear from Tables 4 and 5 that in both sexes subjects with higher WHR showed lower energy expenditure in the routine physical activity and higher consumption of energy from dietary fat (25-32 kcal!day) compared to subjects with lower WHR. These findings suggest that in Indians [l, 12,201 even modest increase in fat intake (which is common with better socioeconomic status) and reduced physical activity and, as a result, modest tendency of central obesity may not allow the physiological mechanisms to adapt effectively, resulting into rapid development of insulin resistance, hypertension, diabetes, coronary artery disease and atherosclerosis. These populations may benefit by increasing physical activity and reducing the consumption of fat to roughly 22% kcal/day [24]. It seems that the amount of fat consumption which is considered safe in Western populations (< 30% kcal/day) may have adverse effect in Indians and therefore fat intake should be 21-24% kcal/day of total energy intake [20,24]. References
111 World
Health Organization Study Group. Diet, nutrition and prevention of chronic diseases. Geneva: World Health Organization, 1990. 121 World Health Organization Scientific Group. Cardiovascular disease risk factors: new areas for research. Geneva: World Health Organization, 1993. Expert Committee. Rehabili[31 World Health Organization tation after cardiovascular diseases with special emphasis on developing countries. Geneva: World Health Organization, 1993. Research Group. Interrsalt: an in[41 Intersalt Cooperative ternational study of electrolyte excretion and blood pressure. Results of 24 h urinary sodium and potassium excretion. Br Med 3 1988; 297: 319-328. PI MC Mohan S, Cutler J, Britain E, Higgins M. Obesity and hypertension: epidemiological an clinical issues. Eur Heart J 1987; 7 (Suppl B): 57-70. 161 Mckeigue PM, Shah B, Marmot MC. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancer 1991; 337: 382-386. S, Ramachandran K, Kaul [71 Chadda SL, Radhakrishnan U, Gopinath V. Prevalence, awareness and treatment status of hypertension in urban population of Delhi. Indian J Med Res 1990; 92: 233-240. KG, Jayasree AK, 181 Raman Kutty V, Balakrishnan Thomas J. Prevalence of coronary heart disease in the
R. Beegom
et al. /International
rural population of Thiruvanananthapuram district, Kerala, India. Int J Cardiol 1993; 39: 59-70. [9] Beegum R, Singh RB. Prevalence of coronary heart disease and its risk factors in the urban population of South and North India. Acta Cardiol 1994. In press. [IO] Gopinath N, Chadda SL, Jain P, Shekhawat S, Tandon R. An epidemiological study of obesity in adults in the urban population of Delhi. J Assoc Phys India 1994; 42: 212-21s. 1111 Rose GA, Blackbum H, Gillum RF, Prineas RJ. Cardiovascular survey methods. Geneva: World Health Organization, 1982. [12] Narsingrao BS, Deosthale YG, Pant KC. Nutrient composition of Indian foods. Hyderabad: National Institute of Nutrition, 1989. [13] Vanhandle E, Zilversmitt BD. Micromethod in the direct estimation of serum triglycerides. J Lab Clin Med 1957; 60: 152-159. [14] Wilson DE, Spiger MJ. A dual precipitation method for quantitative plasma lipoprotein measurement without centrifugation. J Lab Clin Med 1973; 82: 413-482. [15] Mckeigue PM, Marmot MG, Adolstein AM. Diet and risk factors for coronary heart disease in Asians in northwest London. Lancet 1985; ii: 1086-1090. [16] Allan C, Brotman H. Chart book on aging in America. Washington, DC: The 1981 White House Conference on Aging, 1981. (171 The WHO MONICA Project. Geographical variations in the major risk factors of coronary heart disease in men and women aged 35-64 years. World Health Statist 1988; 41: 115-140.
Journal
of Cardiology [IS]
51 (1995)
183-191
191
Nissinen A, Bothig S, Granroth H, Lopez AD. Hypertension in developing countries. World Health Statist 1988; 41: 141-154. [19] Sharma BK, Arora OP, Bansal BL, Sagar S, Khurana SP. Hypertension among industrial workers and professional classes in Ludhiana, Punjab. Indian Heart J 1985; 37: 380-385. [20] Singh RB, Ghosh S, Naiz MA, Gupta S, Bishnoi I, Sharma JP, Agarwal P, Rastogi SS, Beegam R, Chibo H, Shoumin Z. Epidemidogic study of diet and coronary risk factors in relation to central obesity and insulin levels in rural and urban population of North India. Int J Cardiol 1995;47: 245-255. [21] Bhatnagar MK, Ganeh R, Goel A, Verma NPS, Chauhan UPS. A study of hyperinsulinemia in Indian hypertensive subjects. J Assoc Phys India 1994; 42: 522-524. 1221 Verma NPS, Mehta SL, Madhu S, Mather HM, Keen H. Prevalence of known diabetes in urban Indian environment: the Darya Ganj diabetes survey. Br Med J 1986; 293: 423-424. [23] Snehlatha C, Ramchandran A, Vijaya V, Viswanathan M. Differences in plasma insulin responses in urban and rural Indians: a study in Southern Indians. Diabetic Med 1994; 11: 445-448. [24] Singh RB, Niaz MA, Ghosh S. Effect on central obesity and associated disturbances of low energy, fruit and vegetable-enriched prudent diet in North Indians. Postgrad Med J 1994; 70: 895-900.