Beruflich Dokumente
Kultur Dokumente
PAPER
High prevalence of diabetes, obesity and
dyslipidaemia in urban slum population in northern
India
A Misra1*, RM Pandey2, J Rama Devi1, R Sharma3, NK Vikram1 and Nidhi Khanna3
1
Department of Medicine, All India Institute of Medical Sciences, New Delhi, India; 2Department of Biostatistics, All India Institute
of Medical Sciences, New Delhi, India; and 3Department of Dietetics, All India Institute of Medical Sciences, New Delhi, India
BACKGROUND AND AIMS: In this study, a prevalence survey of various atherosclerosis risk factors was carried out on hitherto
poorly studied rural – urban migrants settled in urban slums in a large metropolitan city in northern India, with the aim of
studying anthropometric and metabolic characteristics of this population in socio-economic transition.
DESIGN: A cross-sectional epidemiological descriptive study.
SUBJECTS: A total of 532 subjects (170 males and 362 females) were included in the study (response rate approximately 40%).
METHODS AND RESULTS: In this study, diabetes mellitus was recorded in 11.2% (95% CI 6.8 – 16.9) of males and 9.9% (95%
CI 7.0 – 13.5) of females, the overall prevalence being 10.3% (95% CI 7.8 – 13.2). Based on body mass index (BMI), obesity was
more prevalent in females (15.6%; 95% CI 10.7 – 22.3) than in males (13.3%; 95% CI 8.5 – 19.5). On the other hand, classifying
obesity based on percentage body fat (%BF), 10.6% (95% CI 6.4 – 16.2) of males and 40.2% (95% CI 34.9 – 45.3) of females
were obese. High waist – hip ratio (WHR) was observed in 9.4% (95% CI 5.4 – 14.8) of males and 51.1% (95% CI 45.8 – 56.3) of
the females. All individual skinfolds and sum of skinfolds were significantly higher in females (P < 0.001). In both males and
females above 30 y of age, there was a steep increase in the prevalence of high WHR, and in females, %BF was very high
(particularly in %BF quartile > 30%). Furthermore, total cholesterol and low-density lipoprotein cholesterol were high in both
males and females. Stepwise multiple linear regression analysis showed that for both males and females BMI, WHR and %BF
were positive predictors of biochemical parameters, except for HDL-c, for which these parameters were negatively associated.
CONCLUSIONS: Appreciable prevalence of obesity, dyslipidaemia, diabetes mellitus, substantial increase in body fat, general-
ised and regional obesity in middle age, particularly in females, need immediate attention in terms of prevention and health
education in such economically deprived populations.
International Journal of Obesity (2001) 25, 1722 – 1729
Keywords: Asian Indian people; diabetes mellitus; body fat; hypercholesterolaemia; waist – hip ratio
Males Females
Body mass index 165 20.0 4.1 359 20.7 4.5 (NS)
Skinfolds (mm) and ratios
Waist – hip ratio 170 0.86 0.11 362 1.02 3.95 (NS)
Biceps 170 5.5 3.8 362 8.3 5.4*
Triceps 170 9.7 6.6 362 14.8 7.8*
Subscapular 170 13.1 7.3 362 17.1 9.5*
Suprailiac 169 14.3 7.5 361 19.9 11.3*
S4SF (sum of four skinfolds) 169 42.3 21.2 361 60.0 30.9*
Central skinfolds (sum of 169 27.2 13.0 361 36.9 19.8*
subscapular and suprailiac skinfolds)
Peripheral skinfolds (sum of triceps and biceps skinfolds) 170 15.2 9.6 362 23.1 12.6*
Central – peripheral skinfold ratio 169 2.01 0.69 361 1.67 0.59*
SS=TR (Subscapular – triceps skinfold ratio) 170 1.54 0.64 362 1.23 0.7*
Body fat (%) 169 17.5 6.2 361 28.4 7.6*
Table 3 Percentage prevalence of diabetes mellitus and various measures of obesity stratified in quartiles according to age groups
2
% Body fat quartiles, % (n) BMI (kg=m ) quartiles, % (n)
High
Age group (y) Sex DM Up to 18 > 18 – 25 > 25 – 30 > 30 Total Up to 18 > 18 – 23 > 23 – 24.9 25 Total WHR
18 – 30 M 6.6 (4) 63.9 (389) 24.6 (15) 6.6 (4) 4.9 (3) 100 (61) 39.0 (23) 47.5 (28) 8.5 (5) 5.0 (3) 100 (59) 0.0 (0)
F 7.1 (13) 17.4 (32) 32.6 (60) 20.1 (37) 29.9 (55) 100 (184) 33.0 (60) 52.0 (91) 5.5 (10) 11.5 (21) 100 (182) 41.8 (77)
Total 6.9 (17) 29.0 (71) 30.6 (75) 16.7 (41) 23.7 (58) 100 (245) 34.4 (83) 49.4 (119) 6.2 (15) 10.0 (24) 100 (241) 31.4 (77)
31 – 50 M 10.7 (9) 61.9 (52) 27.4 (23) 8.3 (7) 2.4 (2) 100 (84) 28.0 (23) 41.5 (34) 14.6 (12) 15.9 (13) 100 (82) 15.5 (13)
F 10.0 (15) 7.3 (11) 22.7 (34) 19.3 (29) 50.7 (76) 100 (150) 25.5 (38) 40.9 (61) 14.1 (21) 19.5 (29) 100 (149) 60.0 (90)
Total 10.3 (24) 26.9 (63) 24.4 (57) 15.4 (36) 33.3 (78) 100 (234) 26.4 (61) 41.1 (95) 14.3 (33) 18.2 (42) 100 (231) 44.0 (103)
51 and above M 24.0 (6) 71.0 (17) 21.0 (5) 8.0 (2) 0.0 (0) 100 (24) 41.7 (10) 37.5 (9) 16.7 (4) 4.1.0 (1) 100 (24) 12.0 (3)
F 29.6 (8) 18.5 (5) 22.2 (6) 7.4 (2) 51.9 (14) 100 (27) 35.7 (10) 32.1 (9) 10.7 (3) 21.5 (6) 100 (28) 66.7 (18)
Total 26.9 (14) 44.2 (23) 21.2 (11) 7.7 (4) 26.9 (14) 100 (52) 39.3 (20) 33.3 (17) 13.7 (7) 13.7 (7) 100 (51) 40.4 (21)
Total M 11.2 (19) 64.0 (108) 25.4 (43) 7.6 (13) 3.0 (5) 100 (169) 34.0 (56) 43.0 (71) 12.7 (21) 13.3 (17) 100 (165) 9.4 (16)
F 9.9 (36) 13.3 (48) 27.7 (100) 18.8 (68) 40.2 (145) 100 (361) 30.2 (108) 44.7 (160) 9.5 (34) 15.6 (56) 100 (359) 51.1 (185)
Total 10.3 (55) 29.4 (156) 27.0 (143) 15.3 (81) 28.3 (150) 100 (530) 31.3 (164) 44.2 (231) 10.5 (55) 13.9 (73) 100 (523) 37.8 (201)
DM: diabetes mellitus; high WHR (waist – hip ratio), males > 0.95, females > 0.80.
Table 4 Anthropometric measurements (mean s.d.), glucose intolerance and dyslipidaemia (% (n)), according to BMI quartiles
Variables Sex Total I: Up to 18 II: > 18 – 23 III: > 23.1 – 24.9 IV: > 25 P value
c e
Waist – hip ratio M 170 0.83 0.05 (55) 0.86 0.07 (70) 0.87 0.07 (21) 0.97 0.26 , (17) P < 0.001
a b c e
F 362 0.79 0.07 (108) 0.81 0.06 (161) 0.84 0.06 (34) 0.84 0.07 , (56) P < 0.001
Percentage body fat M 169 13.9 4.5 (56) 18.4 6.1a (71) 19.4 5.7b (21) 23.6 6.2c,e (17) P < 0.001
F 361 23.7 6.5 (108) 28.3 6.5a (161) 30.1 6.6b (34) 36.7 6.1c,e,f (56) P < 0.001
Biceps skinfold M 170 4.6 3.8 (56) 4.9 2.7 (71) 7.9 4.6b,d (21) 9.2 3.8c,e (17) P < 0.001
a b c e f
F 362 5.7 3.3 (108) 7.7 4.1 (161) 9.1 4.8 (34) 14.3 7.4 , , (56) P < 0.001
b c e
Triceps skinfold M 170 6.9 4.9 (56) 10.0 7.1 (71) 11.9 4.8 (21) 15.9 7.1 , (17) P < 0.001
a b c e f
F 362 10.6 5.3 (108) 14.4 6.3 (161) 16.0 5.4 (34) 23.9 9.2 , , (56) P < 0.001
a b c e f
Subscapular skinfold M 170 9.3 4.4 (56) 13.5 6.0 (71) 15.4 8.6 (21) 22.1 9.1 , , (17) P < 0.001
F 362 11.5 6.4 (108) 16.4 7.7a (161) 19.8 7.3b (34) 28.2 10.3c,e,f (56) P < 0.001
Suprailiac skinfold M 169 10.3 5.0 (56) 14.9 6.7a (71) 16.6 6.2b (21) 24.4 8.7c,e,f (16) P < 0.001
F 361 13.9 8.0 (108) 19.3 9.6a (161) 22.2 8.6b (34) 32.7 12.7c,e,f (55) P < 0.001
Impaired fasting glucose M 23 30.4 (7) 43.5 (10) 8.7 (2) 17.4 (4) NS
F 56 25.0 (14) 39.3 (22) 10.7 (6) 25.0 (14) P ¼ 0.004
Diabetes mellitus M 19 21.1 (4) 47.4 (9) 21.1 (4) 10.4 (2) NS
F 35 28.6 (10) 25.7 (9) 11.4 (4) 34.3 (12) P ¼ 0.004
Hypercholesterolaemia M 42 21.4 (9) 45.3 (19) 23.8 (10) 9.5 (4) P ¼ 0.05
F 92 25.0 (23) 39.1 (36) 15.2 (14) 20.7 (19) P ¼ 0.03
Hypertriglyceridaemia M 26 19.2 (5) 61.5 (16) 7.7 (2) 11.6 (3) NS
F 42 11.9 (5) 42.9 (18) 16.7 (7) 28.5 (12) P ¼ 0.003
Low high-density lipoprotein-cholesterol M 24 25.0 (6) 54.2 (13) 4.1 (1) 16.7 (4) NS
F 55 29.1 (16) 49.1 (27) 10.9 (6) 10.9 (6) P ¼ 0.70
High low-density lipoprotein-cholesterol M 38 23.7 (9) 42.1 (16) 23.7 (9) 10.5 (4) NS
F 77 31.2 (24) 37.7 (29) 10.4 (8) 20.7 (16) P ¼ 0.26
a
I vs II; bI vs III; cI vs IV; dII vs III; eII vs IV; fIII vs IV at P < 0.05; NS statistically not significant.
a b a
Total cholesterol 158.7 þ 0.65(Age ) þ 1.1(BMI) þ 37.4(WHR) 85.1 þ 0.69(Age ) þ 1.63(BMI ) þ 47.5(WHR)
a a b a
Serum triglycerides 92.6 þ 0.01(Age) þ 2.18(%BF ) 7 64.2 þ 0.68(Age ) þ 3.1(BMI ) þ 81.5(WHR) þ 1.2(%BF )
High-density lipoprotein-cholesterol 43.6 þ 0.01(Age) 7 3.7(WHR) 7 0.06(%BF) 50.9 þ 0.04(Age) 7 0.05(BMI) 7 13.5(WHR)
Low-density lipoprotein-cholesterol 121.8 þ 0.65(Agea) þ 0.52(%BF) 61.8 þ 0.46(Age) þ 56.2(WHR)
a b c
P < 0.05; P < 0.01; P < 0.001.
BMI, body mass index, WHR, waist – hip ratio, %BF, percentage body fat.
of 20 – 21, the prevalence of obesity, as defined by BMI, is BMI in this group. There is, however, a marked discordance
high. Moreover, high WHR is noteworthy, both in males between the two measures of obesity in females; 40.2% show
and, particularly, females. %BF more than 30%, while only 15.6% show BMI higher
The prevalence of obesity recorded in the current study is than 25. The most striking observation is the increase in
lower compared to that recorded (27.8%) in an urban sample body fat with advancing age (about 50% had high body fat
of adults in New Delhi.25 Unfortunately, few comparable above the age of 30 y). The conspicuous feature in women,
studies performed in the urban slums are available in the therefore, is under-representation of obesity when defined by
literature. Studies from developed countries indicate higher BMI alone. These observations are of considerable practical
prevalence of atherosclerosis risk factors among men and relevance, questioning BMI as a valid epidemiological tool in
women with lower socio-economic status.8 In developing Asian Indian population, particularly in females. Further, all
countries, the problem of poverty and scarcity of food is the metabolic complications of excess fat are likely to be
particularly acute in the slum areas. In the Klong Toey slums present in these women with high body fat, even when their
of Bangkok, 25.5% of the subjects were overweight and 10% BMI is in a normal range. Such women, when postmeno-
were obese.10,11 A study involving women in a low-income pausal and without the benefit of hormonal protection, are
area in Karachi showed 42 and 8% of women to be over- more likely to develop occlusive consequences of athero-
weight and have abdominal obesity respectively,12 an obser- sclerosis.
vation similar to the current study. Sawaya et al reported Variable prevalence (3 – 11.2%) of diabetes has been
high prevalence of obesity and malnutrition coexisting in reported from urban areas of India depending upon on the
the same population.13 region, caste and type of survey, diagnostic tool and diag-
Estimations of individual skinfolds in the current study nostic terminology.6,7 In the urban population of Delhi, the
show comparatively lower values when compared to those prevalence of diabetes mellitus ranged from 1.6 to 9%, being
recorded in South Asian men and women in the UK, having more common in obese subjects.25 In the rural Indian
mean BMI in the range of 25 – 26,1 indicating more subcu- population, it is reported to be in the range of 1 – 5%.7,26,27
taneous fat in the immigrant Asian Indian people in the UK. In one of the well-designed studies in rural areas of North
Although desirable and much needed, data on skinfolds of Arcot District in south India, the prevalence of impaired
rural and urban native Indian population are not available. glucose tolerance (IGT) and diabetes mellitus was 6.6 and
In the current study, the ratio of central vs peripheral skin- 4.9%, respectively, as investigated by 2 h post-75 g oral glu-
folds is greater in males due to excess truncal fat, and this is cose load values.27 Similarly, in a rural community in the
also reflected by higher SS=TR ratios. However, WHR was state of Punjab, only 4.6% were diabetic.28 It is also note-
higher in females, indicating excess abdominal fat even in worthy that the prevalence of IFG is greater in females as
females. Mean WHR is lower in males, and higher in females compared to males in the current study, although it was
as compared to Asian Indian people in the UK.1 Of further statistically insignificant. In a recent study from Kashmir
interest is the remarkable increase in the prevalence of high valley, authors record higher prevalence of IGT and diabetes
WHR in middle-aged females, with no appreciable increase mellitus in females as well.29 The observed prevalence of IFG
in their BMI. In males, WHR increased with increase in age and diabetes mellitus in the present study is higher than that
from 30 to 50 y (Table 3). More than a third of the population reported from the rural population, and equal to or more
over 30 y of age having abdominal obesity is of considerable than that observed in the urban population. However, it is
concern because of associated metabolic and cardiovascular difficult to compare most studies done in India, since diag-
consequences. nostic criteria of diabetes, methodology of tests, and sam-
The detailed observations on anthropometric measure- pling modes were different. Nonetheless, the prevalence of
ments and %BF have rarely been explored in the studies IFG and type 2 diabetes mellitus in this sample of urban slum
done on Asian Indian people living in India. The most population is impressive.
striking observation in the current study is high %BF in Similarly, in the present study, the high prevalence of
females. In males, only 10.6% had excess %BF, a figure dyslipidaemia is striking. Studies from various parts of India
generally approximating the percentage prevalence of high again reveal differences in the prevalence. For example,