Prevalence and determinants of hypertension in the Indian social class and heart survey

Journal of Human Hypertension (1997) 11, 51 56 1997 Stockton Press. All rights reserved 0950-9240/97 $12.00 Prevalence and determinants of hypertension in the Indian social class and heart survey RB Singh 1, JP Sharma 1, V Rastogi 1, MA Niaz 1 and NK Singh 2 1 Heart Research Laboratory, Medical Hospital and Research Centre, Moradabad; and 2 Institute of Medical Sciences, BHU, Varanasi, India To determine the association of socio-economic status 1.05 1.14; women 1.08, 95% CI 1.05 1.13), body mass (SES) and prevalence of hypertension and its risk fac- index (odds ratio: men 1.12, 1.08 1.18; women 1.11, tors in a rural population, a cross sectional survey was 1.06 1.16) and sedentary lifestyle (odds ratio: men 1.45, conducted in two randomly selected villages in the 1.32 1.58; women 1.38, 1.26 1.49). Only weak but sig- Moradabad district in North India. There were 1935 resi- nificant associations were observed with smoking, alcodents aged over 25 (984 men and 951 women) who were hol and salt intake. The association of hypertension with randomly selected and categorised into social classes social class was reduced after adjustment of body mass 1 4 depending upon SES based on occupation, housing index, sedentary lifestyle, smoking and salt intake (odds conditions, land holding, total per capita income, owner- ratio: men 0.96, 0.81 1.14; women 0.73, 0.54 1.04). ship of consumer durables and education. The preva- There was an increase in the prevalence of hypertension lence of hypertension diagnosed by JNC V criteria and age-specific blood pressure (BP) with increasing ( 140/90 mm Hg) was significantly higher among social age in both sexes. The overall prevalence of hypertenclass 1 and 2 and showed positive relation with SES in sion by WHO criteria ( 160/95) was 4.6% and by JNC V both sexes. Among social class 1 and 2 subjects, there criteria 20.8%, and the rates were comparable in both was a higher prevalence of overweight and obesity and sexes. Social class 1 and 2 subjects in rural North India sedentary lifestyle. Logistic regression analysis with have a higher prevalence of hypertension and its risk adjustment of age showed that SES had a positive factors of overweight and sedentary lifestyle. relation with hypertension (odds ratio: men 1.09, 95% CI Keywords: sedentary lifestyle; physical activity; body mass index; socio-economic status; education Introduction lence of coronary artery disease (CHD) and its risk factors, for the first time to our knowledge in India. 10 In developed countries there is strong inverse In this paper we examine the association between association between cardiovascular disease and hypertension and its risk factors with social classes social status, whether measured by occupation, in a rural population from North India. income or education. 1 The prevalence of coronary risk factors and blood pressure (BP) levels are also higher in the lower social classes. 1 3 However, studies from developing countries have shown no such The subjects and methods of this study have been Patients and methods correlation. Studies in Indian migrants to Britain described previously. 10 In brief, we selected two vilreported a higher prevalence of hypertension, higher lages at random from the Moradabad district which stroke mortality and higher BP levels and body mass has a population of 0.1 million (census, 1991). We index (BMI) than in their siblings in India. This sug- randomly selected two to three blocks from each gests that socioeconomic factors and lifestyle may street of each village. Each block contained 100 300 be important in the development of hypertension. 4,5 adults and we selected 2000 subjects for this study. A few studies from India which have examined this Details of population in the two villages were question suggest that BP increases with social class obtained from the voters lists (1991). The total and that hypertension is more common in higher study sample was 2235 subjects 25 years of age, of income groups. 6 8 Social classes and the prevalence which 300 (13.4%) refused to give a detailed history of risk factors of hypertension have not been studied and examination. All subjects included in the study adequately in other developing countries. 9 In the had lived in the area since birth. All 1935 subjects, Indian Social Class and Heart Survey, we reported (984 males and 951 females) were invited for partician association between social classes and the preva- pation in the study. A physician and social worker administered questionnaire was prepared based on the guidelines of Correspondence: Dr RB Singh, Hon Professor, Preventive Cardi- World Health Organization (WHO), 11 the Indian ology, Civil Lines, Moradabad-10 (UP) 244001, India Received 3 August 1996; revised 4 November 1996; accepted 11 Council of Medical Research and other Indian stud- November 1996 ies. 8 The questionnaire was validated 8 and included

52 information on age, sex, education, socio-economic Statistical analysis status (SES), physical activity, past and family his- Assuming from other studies that the prevalence of tory of hypertension, smoking and alcohol conhypertension is about 5% in the community, we sumption. The SES measurement scale was conwould need a sample of at least 1500 subjects to estistructed based on earlier studies from India and was mate with 90% confidence to detect at 5% signifigraded into social classes 1 4 based on classicance a relative risk of 1.54 for prevalence of hyperfication of Raman Kutty and coworkers 12 (see tension in Indians. We proposed a larger sample size Appendix). In this method scores were assigned to to increase the level of confidence and to estimate each subject based on education, occupation, housthe prevalence to within 2% error on either side ing conditions, ownership of consumer durables, which was deemed satisfactory as a first estimate. land holding and per capita total income of the fam- Multivariate analysis to determine the overall ily members. Per capita income was calculated from relation of social class with hypertension and risk the total income of the family from all sources divfactors was performed by logistic regression. The ided by number of family members. Educational dependent variables were presence of hypertension, status was assessed by the number of years of edusalt intake, alcohol, smoking, physical activity and cation calculated from the highest class achieved. obesity. Independent variables were social class, age Housing conditions and consumer durables were and physical measurements. Age, which is the major assessed as described by Raman Kutty. 12 Social class confounding factor, was then added to the equation 1 was considered highest and 4 lowest. Salt intake and odds ratios determined and tabulated. Odds was assessed with the help of a salt measure by mearatios were calculated for hypertension last, after suring the salt used in cooking of food and added adding BMI, salt and alcohol intakes, smoking and during eating by each subject. sedentary lifestyle to the equation. All P values were Physical examination included measurements of two-tailed and significance was taken as P 0.05. height, weight and BP. Body weights were measured by the dietitian independently in light underclothes to the nearest 0.5 kg. Height was measured in standing Results position. BP (systolic and diastolic phase V The overall prevalence of hypertension 25 years of Korotkoff) was measured while seated in the right age using the JNC V criteria ( 140/90 mm Hg) was arm after 5 min rest with a standard mercury man- 20.8% each in men and women. There was an age ometer by the same physician in all subjects. When dependent increase in the prevalence of hyperten- a high BP ( 140/90 mm Hg) was noted, a final read- sion in both sexes. When classified according to ing was recorded in the lying position after a 5 min WHO criteria ( 160/95 mm Hg), the prevalence of rest as per WHO guidelines. 11 A 12-lead electrocar- hypertension was less than one quarter than the diogram was recorded in all subjects. rates according to JNC V criteria in both men (20.8 vs 5.1%) as well as in women (20.8 vs 4.2%). Only 22 patients (12 males and 10 females) were known hypertensives, and of these only one third were taking Diagnostic criteria regular treatment (Table 1). The mean systolic and diastolic BPs were significantly associated with The diagnosis of hypertension was made when systolic age in both sexes (Table 2). BP (SBP) was 140 mm Hg or more or diastolic A lowering of SES from social class 1 to social BP (DBP) was 90 mm Hg or more as per guidelines class 4 was associated with a significant declining of the US National Health and Nutrition assessment trend in the prevalence of hypertension and its risk survey. 13 Figures for the older WHO criteria factors (Tables 3 5). No such trend was observed for ( 160/95 mm Hg) were also recorded. BMI was cal- salt, alcohol and tobacco intake. There was a significant culated and obesity defined as a BMI of 27 kg/m 2, positive rank correlation of the social class and overweight as a BMI 25 kg/m 2. Figures for cri- with BMI (r = 0.12, P 0.01 for men and r = 0.09, teria laid down by the Indian Consensus Group 14 for P 0.05 for women) and sedentary lifestyle overweight ( 23 kg/m 2 ) were also calculated. (r = 0.13, P 0.01 for men, and r = 0.11, P 0.05 In India, cigarette, beedies, Indian pipes, raw tobacco for women). and chewing tobacco are commonly consumed Multivariate logistic regression analysis was done and people use tobacco in more than one form. We to determine the association of social class with therefore categorised users of any form of tobacco as hypertension ( 140/90) and the prevalence of its smokers as was done in other studies. 12 Beedies are risk factors (Table 6). The results showed a signifi- a type of cigarette in which leaf is used for paper. cant positive association of level of social class with Subjects who admitted to drinking alcohol more the age adjusted prevalence of hypertension, and than once a week (at least 300 ml beer or 60 ml indi- risk factors of hypertension obesity and overweight ginous or factory made whisky) were categorized as and sedentary lifestyle in both men and women. alcohol consuming. Physical activity was assessed Only weak but significant association was observed from occupational and spare time activities. According with salt, smoking and alcohol intake. Adding BMI, to Paffenberger et al, 15 a person is considered sedentary lifestyle, smoking and salt intake to the as leading a sedentary lifestyle if he or she walks equation negated the association of social classes 14.5 km a week, climbs fewer than 20 flights of with the prevalence of hypertension in both males stairs a week and performs no moderately vigorous (odds ratio 0.96, 95% CI 0.81, 1.14, and females physical activity on 5 days a week. (odds ratio 0.73, 95% CI 0.54 1.04).

Table 1 Age distribution of subjects with hypertension 53 Age groups Men Women Subjects WHO n (%) JNC V n (%) Subjects WHO n (%) JNC V n (%) 25 34 300 8 (2.6) 25 (8.3) 350 6 (1.7) 29 (8.2) 35 44 296 12 (4.0) 45 (15.2) 264 7 (2.6) 44 (16.6) 45 54 178 12 (6.7) 46 (25.8) 165 10 (6.6) 42 (25.4) 55 64 120 10 (8.3) 40 (33.3) 96 8 (8.3) 36 (37.5) 64 90 8 (8.8) 49 (54.4) 76 8 (10.5) 47 (61.8) Total between 25 64 years 894 42 (4.7) 156 (17.4) 875 32 (3.6) 151 (17.2) Total 984 50 (5.1) 205 (20.8) 951 40 (4.2) 198 (20.8) 2 for trend 180.6 188.2 172.4 182.5 P 0.001 0.001 0.001 0.001 JNC V = Joint National Committee V guidelines. Table 2 Blood pressure levels in various age groups Age groups (years) Men Women No Systolic Diastolic No Systolic Diastolic 25 34 300 114 (10) 72 (7) 350 112 (9) 70 (7) 35 44 296 118 (11) 75 (8) 264 115 (10) 73 (8) 45 54 178 123 (12) 78 (9) 165 120 (11) 76 (9) 55 64 120 128 (12) 84 (10) 96 124 (12) 81 (10) 64 90 134 (16) 85 (12) 76 130 (14) 83 (11) Total between 25 64 894 121 (15) 76 (13) 875 116 (16) 73 (12) Total 984 123 (18) 77 (15) 951 117 (17) 74 (11) Pearson s r 0.33 0.14 0.28 0.12 Regression coefficient 4.82 2.80 4.56 2.65 Kendall s 0.302 0.232 0.288 0.215 Values are expressed as mean (s.d. mm Hg. P 0.001 for all differences. Table 3 Social class and prevalence of hypertension Mean (n = 984) Women (n = 951) No WHO JNC V No WHO JNC V Table 4 Prevalence of risk factors of hypertension in men and women Men Women Total (n = 984) (n = 951) (n = 1935) Social class Body mass index (kg/m 2 ) 1 160 20 (12.5) 80 (50.0) 130 17 (13.0) 63 (48) 27 50 (5.0) 50 (5.2) 100 (5.2) 2 164 16 (9.7) 75 (45.7) 137 14 (10.2) 65 (47) 25 106 (10.8) 110 (11.6) 216 (11.2) 3 317 10 (3.1) 32 (10) 323 7 (2.2) 45 (14) 23 230 (23.4) 232 (24.4) 462 (23.8) 4 343 4 (1.2) 18 (5) 361 2 (0.6) 25 (7) Sedentary lifestyle 140 (14.3) 232 (24.4) 372 (19.2) Mantel- 10.5 16.1 8.6 14.7 Alcohol intake 47 (4.8) 47 (2.4) Haenzel 2 ( once/week) P value 0.05 0.001 0.05 0.001 Salt intake ( 8 g/day) 652 (66.2) 610 (64.1) 1262 (65.2) Tobacco users 320 (32.5) 60 (6.3) 380 (19.6) Family history 6 (0.6) 5 (0.5) 11 (0.5) Regular moderate physical activity Discussion Spare time 106 (10.8) 50 (5.2) 156 (8.0) Occupational 452 (45.9) 550 (57.8) 1002 (51.8) This study showed that the prevalence of hypertension ( 140/90) and its risk factors were significantly associated with level of SES in a cohort of rural population in North India. Social class 1 and 2 were India because in most studies, only economic status associated with a higher prevalence of JNC V hypertension. has been considered as the sole attribute of social This relation persisted after adjustment of class. 6 8,16 In one study from South India all other age but declined after the addition of other lifestyle attributes of social class were duly considered to characteristics in a multivariate analysis. Over- classify the SES. 12 However, no attempt was made weight and obesity and sedentary lifestyle were also to classify hypertension and its risk factors accord- more prevalent among social class 1 and 2 subjects. ing to social class. One study 6 from central India in However, physical activity was greater among social 448 subjects showed that hypertension was more class 3 and 4 subjects. prevalent among high income group and education The results of the Indian Social Class and Heart was not associated with hypertension. Another Survey cannot be compared with other studies from study 16 among 3340 industrial workers and 1008

54 Table 5 Prevalence of risk factors of hypertension in relation to socioeconomic status No BMI Sedentary Salt intake Alcohol Tobacco users ( 23 kg/m 2 ) lifestyle ( 8 g/day) ( once/week) Males n (%) Social class 1 160 100 (62) 60 (37) 106 (66) 6 (4.0) 50 (31) 2 164 95 (58) 45 (27) 110 (67) 7 (4.2) 55 (33) 3 317 22 (7) 20 (6) 218 (69) 16 (5.0) 105 (33) 4 343 13 (4) 15 (4) 218 (64) 18 (5.2) 110 (32) Mantel-Haenzel 2 10.34 8.68 0.68 1.26 2.26 P value 0.001 0.001 0.22 0.09 0.16 Females n (%) Social class 1 130 105 (81) 106 (82) 96 (74) 10 (8) 2 137 97 (71) 94 (69) 100 (73) 8 (6) 3 323 20 (6) 20 (6) 203 (63) 20 (3) 4 361 10 (3) 12 (3) 213 (59) 022 (5) Mantel-Haenzel 2 8.76 9.55 1.35 1.51 P value 0.001 0.001 0.27 0.09 Table 6 Multivariate logistic regression analysis after adjustment buy expensive animal foods, sugar, fat and other of age for association of various risk factors with prevalence of luxury foods and have physically demanding occuhypertension explained by socioeconomic status pations. However, social class 1 and 2 subjects have Men Women easy access to above foods, are mostly sedentary and Odds ratio Odds ratio therefore coronary risk factors are more common (95% CI) (95% CI) among higher social classes. Multivariate analysis of our data after inclusion of physical characteristics Hypertension 1.09 (1.05, 1.14)* 1.08 (1.05, 1.13)* and smoking nullified the association of social class Body mass index 1.12 (1.08, 1.18)** 1.11 (1.06, 1.16)** with the prevalence of hypertension. The findings Sedentary lifestyle 1.45 (1.32, 1.58)* 1.38 (1.26, 1.49)* Alcohol 1.27 (1.07, 1.56) indicate that any apparent SES differences were Salt intake 1.31 (1.08, 1.62) 1.28 (1.06, 1.72) mediated by other variables and not through SES per Smoking 1.35 (1.02, 1.88) 1.82 (0.85, 1.57) se in increasing the risk of hypertension. We observed that sedentary lifestyle and higher BMI *P 0.01; **P 0.001; P = 0.05. were significantly associated with social class and were significant risk factors of hypertension. The prevalence of hypertension ( 160/95) in this professionals showed that hypertension ( 160/95) study is comparable with other studies from rural was significantly higher among professionals than areas. 6,7 According to the JNC V criteria, hypertenindustrial workers (14.1 vs 2.6%). Mean systolic and sion was prevalent in 20.8% each in men and diastolic BPs were significantly higher among higher women. Cross-sectional surveys in countries with income subgroups compared to lower income sub- different cultures and at various stages of economic groups in both industrial and professional subjects. development have shown a consistent relationship The professionals had higher income and education between age and BP. This study showed a rising than the industrial workers. In a recent study among trend in the prevalence of hypertension and age-spe- 595 elderly urban subjects, the prevalence of hyper- cific BP with increase in age in both sexes. However, tension ( 160/95) was significantly higher among contrary to developed populations 20 the average SBP middle and higher income groups than lower in the seventh decade were 134 mm Hg in men and income groups. 8 130 mm Hg in women. In our study, the DBP did Contrary to these studies, a recent study 17 from not decrease after the fifth decade, instead it showed West India reported that illiteracy and lack of edu- some increase, more so in females (Table 2). In unaccation were independently associated with hyper- cultured societies, BP may not change with age. tension and CHD. These findings are consistent with They acquire a predisposition to age related increase the observations made in developed countries. 1 3 A after adoption of lifestyle resembling that of the recent American Heart Association scientific state- affluent societies. It seems that in rural population ment on socioeconomic factors and cardiovascular of India the transition from poverty to affluence has disease reiterated that coronary risk factors have just started. become more prevalent among uneducated people In conclusion, the present study has shown that and people of low social class. 18 Studies from other the prevalence of hypertension was significantly developing countries have not analysed data accord- associated with higher social classes. BMI and seding to social class. 9 In Sri Lanka, 19 lower social class entary lifestyle were significant risk factors of hyperwas not associated with CHD and its relation with tension. The prevalence of hypertension has become hypertension was not examined. In developing a public health problem in the rural population of countries, people in lower social class are unable to North India.

References aged males in a defined population in central Sri Lanka. Int J Cardiol 1994; 46: 135 142. 1 Rogot E, Sorlie PD, Johnson N. A mortality study of 1.3 20 Whelton PK. Epidemiology of hypertension. Lancet million persons by demographic, social and economic 1994; 344: 101 106. factors: 1979 to 1985 follow-up. Bethesda, MD, National Heart, Lung and Blood Institute, NIH Publication No. 92-3297, 1992. Appendix 2 Pappas G, Gueen S, Hadden W, Fisher G. The increas- Classification of socioeconomic status in a rural ing disparity in mortality between socio-economic population of India groups in the United States, 1960 and 1986. N Engl J Med 1993; 329: 103 109. The following characteristics were used for this 3 Rose G, Marmot MJ. Social class and coronary heart classification. disease. Br Heart J 1981; 45: 13 19. 4 Keil JE et al. Hypertension in Punjab females. Com- (a) Education parisons between migrants to London and natives in (b) Occupation India. Human Biol 1980; 52: 423 433. (c) Per capita income 5 Bhatanagar D et al. Coronary risk factors in people (d) Land holding from Indian subcontinent living in West London and (e) Ownership of consumer durables their siblings in India. Lancet 1995; 345: 405 409. (f) Housing condition 6 Joshi PP, Kate SK, Shegokar V. Blood pressure trends and lifestyle risk factors in rural India. J Asso Phys (a) Education India 1993; 41: 579 581. 4, if all members of family uneducated; 7 Hussain SA, Nayak KC, Gupta A. A study of preva- 3, if one or more educated up to class V; lence of hypertension with reference to economic, educational, environmental and hereditary factors in gen- 2, if educated up to class X; eral population of North-West Rajasthan. Indian Heart 1, if all members educated beyond class X. J 1988; 40: 148 151. 8 Singh RB et al. Epidemiological study of coronary (b) Occupation artery disease and its risk factors in an elderly urban 4, if all members of family are labourers; population of North India. J Am Coll Nutr 1995; 14: 3, if any one member is a skilled worker; 628 634. 2, if any one member supervisors and all other mem- 9 INCLEN Multicentre Collaborative Group. Risk factors bers work in farms; for cardiovascular disease in the developing world. A 1, if the subject is professional or supervises farmmulticentre collaborative study in the International Clinical Epidemiology Network (INCLEN). J Clin Epidemiol ing. 1992; 45: 841 847. 10 Singh RB et al. Social class and coronary artery disease (c) Per capita monthly income in a rural population of North India: The Indian Social 4, if irregular income; too poor to buy food; Class and Heart Survey. Euro Heart J 1997; (in press). 3, if Rs. 300; 11 Rose G, Blackburn H, Gillum R, Prineas RJ. Cardio- 2, if Rs. 300 600; vascular Survey Methods, Geneva, World Health 1, if Rs. 600. Organization, 1982. 12 Raman Kutty V, Balakrishnan KG, Jayasree AK, (d) Land holding Thomas J. Prevalence of coronary heart disease in the rural population of Thiruananthapuram district Ker- 4, if family owned 0.11 acre of land; ala, India. Int J Cardiol 1993; 39: 59 70. 3, if family owned 0.11 to 5 acres of land; 13 The Fifth Report of the Joint National Committee on 2, if family owned 5 12 acres of land; detection, evaluation and treatment of high blood 1, if family owned 12 acres of land. pressure (JNC V). Arch Intern Med 1993; 153: 154 183. 14 Indian Consensus Group Indian Consensus for preven- (e) Ownership of consumer durables tion of hypertension and coronary artery disease: A 4, if family had no consumer durables; scientific statement of the Indian Society of Hyperten- 3, if family had cycle/sewing machine/radio, 2 3 sion and International College of Nutrition. J Nutr animals for farm work or milk; Environ Med 1996; 6: 309 318. 15 Paffenberger RS et al. The association of changes in 2, if family had motor cycle/television/tractor and physical activity level and other lifestyle character- 2 5 animals for farm work and milk; istics with mortality among men. N Engl J Med 1993; 1, if family had tractor/jeep and several animals 328: 538 545. for milk. 16 Sharma BK et al. Hypertension among the industrial workers and professional classes in Ludhiana, Punjab. (f) Housing condition Indian Heart J 1985; 37: 380 385. 4, if 1 2 room accommodation made of mud and 17 Gupta R, Gupta VP, Ahluwalia NS. Educational status, thatched; coronary heart disease and coronary risk factor preva- 3, if 1 2 room house made of bricks and roof made lence in rural population of India. Br Med J 1993; 309: of wood and mud and animal living in the same 1332 1336. place; 18 Kaplan GA, Keil JE. Socioeconomic factors and cardiovascular disease: a review of the literature. AHA 2, if 2 5 rooms house made of bricks and separate medical/scientific statement. Circulation 1993; 88: compound for animals; 1973 1998. 1, if house had more than 2 rooms made of bricks 19 Mendis S, Ekanayake EMTKB. Prevalence of coronary with concrete roof and separate compound for heart disease and cardiovascular risk factors in middle animals. 55

56 Hypertension and social class in India If family or subject was eligible for a high score, we ignored the lower scorings. Each attribute was given Socioeconomic status Total score equal weightage. The score of each subject or house- (SES) groups hold as a fraction of maximum possible and for each characteristic calculated to construct socioeconomic SES-1 13 16 groups from total scores. Maximum score was 16 SES-2 9 12 SES-3 5 8 and minimum 4 indicating social class 1 4. SES-4 4,