• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Oct 9, 2007; 104(41): 16263–16268.
Published online Oct 8, 2007. doi:  10.1073/pnas.0700933104
PMCID: PMC2042195
Sustainability Science, Medical Sciences

Educational status and cardiovascular risk profile in Indians


The inverse graded relationship of education and risk factors of coronary heart disease (CHD) has been reported from Western populations. To examine whether risk factors of CHD are predicted by level of education and influenced by the level of urbanization in Indian industrial populations, a cross-sectional survey (n = 19,973; response rate, 87.6%) was carried out among employees and their family members in 10 medium-to-large industries in highly urban, urban, and periurban regions of India. Information on behavioral, clinical, and biochemical risk factors of CHD was obtained through standardized instruments, and educational status was assessed in terms of the highest educational level attained. Data from 19,969 individuals were used for analysis. Tobacco use and hypertension were significantly more prevalent in the low- (56.6% and 33.8%, respectively) compared with the high-education group (12.5% and 22.7%, respectively; P < 0.001). However, dyslipidemia prevalence was significantly higher in the high-education group (27.1% as compared with 16.9% in the lowest-education group; P < 0.01). When stratified by the level of urbanization, industrial populations located in highly urbanized centers were observed to have an inverse graded relationship (i.e., higher-education groups had lower prevalence) for tobacco use, hypertension, diabetes, and overweight, whereas in less-urbanized locations, we found such a relationship only for tobacco use and hypertension. This study indicates the growing vulnerability of lower socioeconomic groups to CHD. Preventive strategies to reduce major CHD risk factors should focus on effectively addressing these social disparities.

Keywords: coronary heart disease, socioeconomic status

Because cardiovascular disease has become the leading cause of mortality worldwide, coronary heart disease (CHD) is now contributing to large and rising burdens of death and disability in many developing countries (1). The relationship of socioeconomic status (SES) and CHD has varied across different populations, when concurrently studied, and within each population, when studied over time (2). In populations where the CHD epidemic has matured over several decades, it has been observed that the epidemic of CHD appears to emerge first in higher socioeconomic groups and declines first in the same groups (3, 4). Studies conducted in developed countries over the past three decades provide convincing evidence of an inverse relationship between SES and CHD (59). Additionally, the lowest socioeconomic group is reported to have increased prevalence of subclinical CHD compared with those in the highest socioeconomic group (10, 11). However, when multiple countries are compared, the relationship is quite variable, depending on the level of health transition in each country. It has been suggested that studies of CHD risk factors in heterogeneous populations of developing countries may help us understand the multifactorial nature of CHD causation (2).

In India, a large developing country, the relationship of SES to CHD has not been clear, although an evolution from a direct to an indirect relationship has been predicted to occur as the epidemic advances. Even two decades ago, McKeigue and Sevak (12, 13) predicted that an inverse association between SES and CHD will finally emerge among South Asians, based on several studies carried out among migrant South Asians in the United Kingdom during that period. Later, Bhopal et al. (14) observed this relationship in South Asian migrants and most clearly in migrant Indians.

Studies in India over the past half century have revealed a similar trend toward a progressive reversal of the social gradient for CHD. Although studies conducted from the 1960s to the early 1990s suggested a direct relationship between income and CHD risk, studies conducted in the last decade have reported an inverse relationship between education and/or income with prevalent or incident CHD (1520). A large case-control study conducted by us (20) revealed that the risk of developing myocardial infarction was two times higher in those with the lowest when compared with the highest level of education. Studies of CHD risk factors in Indians have revealed variable associations with SES, reporting an inverse graded relationship of education with tobacco consumption and hypertension, with no clear relationship identified for other risk factors (21, 22).

It is likely that the relationship between cardiovascular disease risk factors and SES in Indian population groups is influenced by the stage of health transition. At the midpoint of health transition, an urbanized population would reveal a reversal of the social gradient (the presence of a graded inverse relationship of SES with CHD risk factors), whereas in a relatively rural population group, the relationship of SES and CHD risk factors would still show a direct relationship. Because different regions of India are at different stages of epidemiological transition, we hypothesize that (i) the relationship of CHD risk factors with SES will vary depending on the level of urbanization; and (ii) the pattern of reversal in social gradient for CHD risk factors, in Indian population groups, will be different from that presently observed in Western societies. To test these hypotheses, we analyzed data of individuals who participated in the multicenter sentinel surveillance for CHD risk factors in Indian industrial workers and their families (23). Of the several measures of SES, educational attainment has been reported to be a valid and easily measurable indicator of SES and considered suitable for social ranking across many populations at different stages of development (2). The present study reports the associations of educational status with different CHD risk factors in several Indian industrial population groups at different levels of urbanization.


Demographic Data.

A total of 19,973 individuals consented to participate in our study in the age group of 20–69 years. The response rate was 87.6%. Data from 19,969 individuals were used for analysis, because the database did not capture the educational status of four individuals. The general characteristics of the study population are published elsewhere (23). The mean age in our study population was 39.8 ± 11.9 years. A third of the study population (32.8%) were individuals occupying high-end executive positions. Fourteen percent of the study group were occupying low-end jobs (mainly manual workers), and less than one-tenth (8%) of the individuals were unemployed. The remaining had jobs that were intermediate between executive positions and manual workers. More than one-fifth (22.6%) of the study group were uneducated or had education only up to primary school. The majority of the individuals (43.3%) had education above primary school and up to secondary school. A small proportion (12.9%) of the study group were postgraduates. The remaining 21.2% were either college graduates or had studied beyond the secondary-school level.

The age group and occupational status of the study group across different educational group are presented in Table 1. The low-educational-status group was significantly older compared with the high-educational-status group. As expected, the job position occupied by the individuals was commensurate with their level of education.

Table 1.
General characteristics of study population

Prevalence of CHD Risk Factors Stratified by Education.

The mean levels of major CHD risk factors across various educational groups are given in Table 2. The prevalence and prevalence ratio with 95% confidence interval (across educational group) of categorical variables are given separately for men and women in Table 3. Tobacco use (P < 0.001) and hypertension prevalence (P = 0.05) showed a significant inverse relationship with education status in men, even after adjustment for age and occupation. Similarly, tobacco use (P < 0.001), hypertension (P < 0.001), diabetes (P < 0.01), and metabolic syndrome (P = 0.04) were inversely related to education level among women. The inverse relationship of educational status and leisure-time physical activity was also observed both in men and women (P < 0.001). However, in men, although diabetes did not show any trend of increasing prevalence with educational status, the metabolic syndrome was high among those with a lower level of education, although not reaching statistical significance levels. In contrast, we observed a direct graded relationship of dyslipidemia prevalence and level of education in both men (P = 0.009) and women (P = 0.04).

Table 2.
Mean levels of CHD risk factors stratified by education and gender (mean ± SD)
Table 3.
Prevalence and prevalence ratio with 95% CI of coronary risk factors stratified by gender and education

Knowledge and Awareness of CHD Risk Factors Stratified by Education.

Treatment-seeking behavior and data on optimum management of risk factors are given in Table 4. Despite the high prevalence of hypertension, significantly fewer men in the low- compared with the high-educational group sought treatment (P = 0.01). However, this relationship was exactly the opposite among women, because significantly more individuals in the low-educational group sought treatment for hypertension compared with the high-educational group (P = 0.05). Optimal control of blood pressure (≤140/90 mmHg; 1 Hg = 133 Pa) was significantly lower in the low-educational groups. Although there was no graded relationship for awareness of diabetes both among men and women, optimal control of diabetes was low across all four educational groups.

Table 4.
Management of elevated CHD risk factors and educational status (%, OR; 95% CI)

Level of Urbanization, Educational Class, and Prevalence of Risk Factors.

An interesting relationship was observed when a comparison was made based on the location of industries (Table 5). In highly urbanized locations, we observed a reversal of the social gradient (the inverse relationship of the prevalence of risk factors to the level of education) for hypertension, diabetes, tobacco use, and overweight. Although the difference in prevalence of tobacco use, hypertension, and diabetes were marked (P value for trend <0.01 for all three risk factors), the differences for overweight were modest (P value for trend = 0.02). On the contrary in periurban locations, we observed a clear reversal of social gradient for tobacco use (P value for trend <0.001) and hypertension (P value for trend = 0.03) with a direct relationship (higher prevalence among those with higher levels of education) for diabetes (P value for trend <0.001) abdominal obesity (P value for trend <0.001) and dyslipidemia (P value for trend = 0.02).

Table 5.
CHD risk factor prevalence across educational group based on level of urbanization


The relationship of SES to CHD has been observed to change as the epidemic evolves (3, 4). Initially, the urban, affluent, and educated sections of a population (early adopters) use their higher disposable incomes to experiment with risk-prone behaviors and, therefore, are at a greater risk of CHD. Later, as the mediators of risk (tobacco, unhealthy foods, and automated transport) become widely available for mass consumption, all social classes are affected. In the advanced phase of the CHD epidemic, the urban, educated, and affluent sections acquire health information, adopt healthy behaviors, and access health care more efficiently. As CHD rates decline in that group, risk factors and disease burdens of CHD become higher in the less-educated and low-income groups and finally even in rural populations. Different CHD risk factors are likely to experience reversal of the social gradient at different times as health transition advances. The associations observed in our cross-sectional study need to be interpreted in the context of that evolutionary profile of the CHD epidemic.

We observed reversal of the social gradient for tobacco use and hypertension across the whole population. When stratified by level of urbanization, in industrial populations located at highly urbanized centers, we observed reversal of social gradients for tobacco use, hypertension, diabetes, and overweight, whereas in less-urbanized locations, we found such a reversal only for tobacco use and hypertension. Prevalence of diabetes and abdominal obesity was directly associated with educational status in the less-urbanized locations. Awareness and treatment status both for diabetes and hypertension were low across all categories. Thus, we have demonstrated a high and largely untreated burden of CHD risk factors in a relatively young population. These burdens vary, depending on the levels of urbanization and education.

Educational level and CHD risk factors have an inverse relationship in several populations (2427). This is demonstrated by the higher burden of CHD among the less educated in several populations (28, 29). However, in countries such as India, which are experiencing rapid epidemiological transition, the results have been varied. In case-control studies from large tertiary-level hospitals, a higher risk of myocardial infarction has been reported by us and others among the poor and less educated (19, 20). However, to the best of our knowledge, there are no reports evaluating this relationship in secondary and primary care centers, which are located predominantly in less-urbanized rural settings. With regard to CHD risk factors (which indirectly predict the burden of CHD), the results have been heterogeneous. Although smoking and hypertension have been consistently reported to be more prevalent in those who are poor and less educated, data are sparse and inconclusive with regard to other CHD risk factors (21, 22). Our study, which collected data from several parts of India using standardized methods, fills these lacunae and demonstrates the importance of targeting those with lower levels of education when planning CHD risk-factor prevention programs.

The effects of urbanization of CHD have been well studied in Western populations (30). Urbanization leads to lifestyle changes, resulting in an increased consumption of energy-rich foods, a decrease in energy expenditure (through less physical activity), and loss of social support that is available for rural societies, all of which lead to higher rates of obesity, raised population mean levels of blood pressure, serum cholesterol, and blood glucose, and a decrease in insulin sensitivity (31). Although urbanization has been a relatively uniform process in prosperous Western economies, it occurs in an unplanned and sometimes chaotic manner, with the establishment of large slums, in India and other developing counties. Even then, an ascending gradient for CHD risk factors has been observed in a comparison of rural, urban slum and nonslum dwelling populations in India (32). As in Western populations, tobacco consumption has been the first CHD risk factor to demonstrate a reversal of the social gradient. Although the reversal of the social gradient for hypertension has occurred along with or has followed the reversal of the social gradient for obesity in Western populations (33), it has occurred even in relatively nonobese population groups in India. This is exemplified in our results by the low body mass index and waist circumference observed among the less educated, who nevertheless have higher levels of blood pressure and hypertension. Although the causes of this early reversal for hypertension have yet to be ascertained, diets high in salt and low in fruit and vegetables may be responsible. Given that Indians have a high propensity to develop metabolic abnormalities with lower levels of body weight and abdominal obesity (also evident in our results), the higher burden of CHD risk factors among the urban poor is a cause for major concern, and urgent and concerted prevention policies are necessary to reduce this burden.

Employees and their family members in all participating industries were eligible for on- or off-site health care (worksite dispensaries, reimbursement of medical expenses, or medical insurance). Despite this, levels of awareness and control of CHD risk factors, such as hypertension and diabetes, were low. Even with such low overall levels of awareness and risk factor control, the higher-ducation groups had better treatment-seeking behaviors and risk-factor management. The reasons for this could be many. For example, lack of education could adversely influence health-seeking behaviors or access to health care. In addition, these findings could be a result of the emphasis on curative clinical care rather than preventive programs in participating industries. It is in the interest of sustainable health of all population groups in India that policies and program for CHD prevention are designed to protect persons in all SES groups and are delivered with special attention to the needs of the less-educated and more vulnerable groups.


We have discussed some of the limitations of our work elsewhere (23). Briefly, the study population was mainly composed of industrial employees and may not be representative of the general population. Although the prevalence of risk factors could be different in the general population, we expect a similar social gradient in risk factors across different educational groups, as suggested by small and localized community studies (22). However, the results of this large study, which has a good mix of different types of industries in India, can be generalized to the working population in the organized sector, which employs >30 million people.


The detailed methods of the study are described elsewhere (23). We provide a brief description of the methods below.



The objective of the paper was to examine whether CHD risk factors are predicted by level of education.

Study Setting.

This was a cross-sectional study conducted in 10 industries across India. Ten medium-to-large industries (defined as industries employing 1,500–5,000 people) in the organized sector were selected from different sites spread across India, from both public and private sectors, based on their willingness to participate in the study and proximity to an academic medical institution.


All of the employees and their family members between the ages of 20 and 69 were eligible to be included in the survey. At each participating center, detailed data were obtained from randomly selected employees and their eligible family members (n = 2,000 at each center). Further, from this group, we chose 1,000 individuals per center by stratified random subsampling for biochemical analysis.

Study Variables.

The study involved collection of data related to the demographic profile, individual characteristics related to major risk factors of CHD, past medical history, clinical and anthropometric profile, and biochemical parameters.

Quality Control Measures.

To ensure the accuracy, completeness, and comparability of blood pressure and anthropometric measurements and of interviewee responses across the 10 study sites, several quality-control measures were included in the study protocol. More details are available in the methodology paper published earlier (23). Briefly, the study followed a common study protocol and a manual of operation. Common measurement techniques by trained study staff using structured pretested proformae were used in all participating centers. Ten percent of the biochemical samples were reanalyzed at the central coordinating center laboratory. The analysis of the results of this 10% sample from all participating sites yielded <5% coefficient of variation between the central coordinating laboratory results and the individual laboratory results.


Educational status and occupation.

Educational status was assessed in terms of highest educational level and was stratified into four categories. The current primary occupation was taken as the employment status of the individual. The industries were classified as highly urban, urban, and periurban/rural based on the location of the industry.

Current tobacco use.

Current tobacco use was defined as use of any form of tobacco products in the previous 30 days.

Physical activity level.

Physical activity levels were assessed by using a structured questionnaire. We obtained data on occupation-related physical activity in a semiquantitative way. Occupational-related physical activity levels were classified into four categories: very light (walking, having a job involving desk work, and watching television), light (standing all day working and housework such as cooking and cleaning in the house), moderate (gardening, agricultural work, walking long distances up and down hills, and climbing >20 steps in a day), and heavy (lifting heavy weights, a job involving labor, and running). Leisure-time physical activity was assessed by the number of minutes of activity as well as the type of activity. In addition, we obtained data on physical activity expended toward travel to work. We also obtained data on sedentariness (amount of time spent watching television, working with the computer, and reading).


Hypertension (stages I and II combined) was defined as either a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg and/or being on drug treatment for hypertension (34).


Diabetes was defined as a fasting blood glucose value ≥126 mg/dl and/or being on treatment for diabetes (35). Impaired fasting glucose was defined as a fasting blood glucose value 110–126 mg/dl and not being on any drug therapy

Metabolic syndrome (MS).

The definition of MS was based on National Cholesterol Education Program Adult Treatment Panel III criteria (36).


Dyslipidemia was defined as fasting total cholesterol to high-density cholesterol (high-density lipoprotein) ratio >5.

Statistical Analysis.

The analysis focused on assessing both the direction and magnitude of the relationship of educational level and CHD risk factors. Differences in means of CHD risk factors were compared across the four educational groups, using analysis of variance, after adjusting for differences in age and occupation. For categorical variables, logistic regression was used to calculate the risk ratio and 95% confidence interval of risk ratio, after adjustment for age and occupation. The level of statistical significance was set at a P value of <0.05 without adjustment for multiple comparisons. The data were analyzed by using the Statistical Package for Social Sciences Version 13 (SPSS, Inc., Chicago, IL).


The current study stresses the scale and seriousness of the emerging challenge of CHD risk factors in India, with particular emphasis on socially deprived groups. The differences in risk factor levels observed in this study may contribute to social disparities in morbidity and mortality because of CHD, especially because awareness and treatment are at very low levels. Strategies to reduce major CHD risk factors should focus on socially disadvantaged groups.


We acknowledge the financial support provided by the Ministry of Health, the Government of India, and the World Health Organization. We also acknowledge the infrastructural support provided by the participating industries.


coronary heart disease
socioeconomic status
educational status.


This paper is part of a special series on Sustainable Health. See the related editorial on page 15969 and accompanying articles on pages 16038, 16044, and 16194.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.


1. Reddy KS. N Engl J Med. 2004;350:2438–2440. [PubMed]
2. Chang CL, Marmot MG, Farley TM, Poulter NR. J Clin Epidemiol. 2002;55:741–747. [PubMed]
3. Pearson TA, Jamison DT, Trejo-Gutierrez H. In: Disease Control Priorities in Developing Countries. Jamison DT, editor. New York: Oxford Univ Press; 1993. pp. 577–599.
4. Reddy KS, Yusuf S. Circulation. 1998;97:596–601. [PubMed]
5. Marmot MG, Adelstein AM, Robinson N, Rose GA. Bone Miner J. 1978;ii:1109–1112. [PMC free article] [PubMed]
6. Butcher HC, Regland DR. Am J Public Health. 1995;85:1231–1236. [PMC free article] [PubMed]
7. Elo IT, Preston SH. Soc Sci Med. 1996;42:47–57. [PubMed]
8. Smith GD, Neaton JD, Wentworth D, Stamler R, Stamler J. Am J Public Health. 1996;86:486–496. [PMC free article] [PubMed]
9. Emberson JR, Whincup PH, Morris RW, Walker M. Int J Epidemiol. 2004;33:289–296. [PubMed]
10. Rosvall M, Ostergren PO, Hedblad B, Isacsson SO, Janzon L, Berglund G. Soc Sci Med. 2006;62:1785–1798. [PubMed]
11. Nordstrom CK, Diez Roux AV, Jackson SA, Gardin JM. Soc Sci Med. 2004;59:2139–2147. [PubMed]
12. McKeigue P, Sevak L. Coronary Heart Disease in South Asian Communities. London: Health Education Authority; 1994. pp. 21–22.
13. McKeigue PM. J Clin Epidemiol. 1989;42:597–609. [PubMed]
14. Bhopal R, Hayes L, White M, Unwin N, Harland J, Ayis S, Alberti G. J Public Health Med. 2002;24:95–105. [PubMed]
15. Chadha SL, Radhakrishnan S, Ramachandran K, Kaul U, Gopinath N. Indian J Med Res. 1990;92:423–430.
16. Padmavati S. Circulation. 1962;25:711–717. [PubMed]
17. Sarvotham SG, Berry JN. Circulation. 1968;37:939–953. [PubMed]
18. Chadha SL, Gopinath N, Ramachandran K. Indian J Med Res. 1992;96:115–121. [PubMed]
19. Pais P, Pogue S, Gerstein H, Zacharia E, Savitha D, Jayaprakash S, Yusuf S. Lancet. 1996;348:358–363. [PubMed]
20. Rastogi T, Reddy KS, Vaz M, Spiegelman D, Prabhakaran D, Willett WC, Stampfer MJ, Ascherio A. Am J Clin Nutr. 2004;79:582–592. [PubMed]
21. Gupta R, Gupta VP, Ahluwalia NS. Bone Miner J. 1994;309:1332–1336. [PMC free article] [PubMed]
22. Gupta R, Gupta VP, Sarna M, Prakash H, Rastogi S, Gupta KD. J Assoc Physicians India. 2003;51:470–477. [PubMed]
23. Reddy KS, Prabhakaran D, Chaturvedi V, Jeemon P, Thankappan KR, Ramakrishnan L, Mohan BVM, Pandav CS, Ahmed FU, Joshi PP, et al. Bull World Health Org. 2006;84:461–469. [PMC free article] [PubMed]
24. Hoeymans N, Smit HA, Verkleij H, Kromhout D. Eur Heart J. 1996;17:518–525. [PubMed]
25. Choiniere R, Lafontaine P, Edwards AC. Can Med Assoc J. 2000;162:S13–24. [PMC free article] [PubMed]
26. Osler M, Gerdes LU, Davidsen M, Bronnum-Hansen H, Madsen M, Jorgensen T, Schroll M. J Epidemiol Commun Health. 2000;54:108–113. [PMC free article] [PubMed]
27. Yu Z, Nissinen A, Vartiainen E, Song G, Guo Z, Zheng G, Tuomilehto J, Tian H. Bull World Health Org. 2000;78:1296–1305. [PMC free article] [PubMed]
28. Mackenbach JP, Cavelaars AE, Kunst AE, Groenhof F. Eur Heart J. 2000;21:1141–1151. [PubMed]
29. Mackenbach JP, Bos V, Andersen O, Cardano M, Costa G, Harding S, Reid A, Hemstrom O, Valkonen T, Kunst AE. Int J Epidemiol. 2003;32:830–837. [PubMed]
30. Yusuf S, Reddy S, Ounpuu S, Anand S. Circulation. 2001;104:2746–2753. [PubMed]
31. Reddy KS. Public Health Nutr. 2002;5:231–237. [PubMed]
32. Reddy KS, Prabhakaran D, Shah P, Shah B. Obes Rev. 2002;3:197–202. [PubMed]
33. Colhoun HM, Hemingway H, Poulter NR. J Hum Hypertens. 1998;12:91–110. [PubMed]
34. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ. J Am Med Assoc. 2003;289:2560–2572. [PubMed]
35. Alberti KG, Zimmet PZ. Diab Med. 1998;15:539–553. [PubMed]
36. Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. J Am Med Assoc. 2001;285:2486–2497. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Cited in Books
    Cited in Books
    PubMed Central articles cited in books
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...