ORIGINAL ARTICLES
Cardiovascular risk factor prevalence among men in a large industry of northern India
D. Prabhakaran, Pankaj Shah, Vivek Chaturvedi, Lakshmy Ramakrishnan, Ajay Manhapra, K. Srinath Reddy

ABSTRACT
    Background.Industrial settings, with their intramural resources and healthcare infrastructure, are ideal for initiating preventive activities to increase the awareness and control of cardiovascular diseases (CVD). However, there are no reliable estimates of CVD and risk factor burden, nor of its awareness and treatment status in urban Indian industrial settings. We aimed to evaluate the prevalence of CVD and its risk factors, and to assess the status of awareness and control of CVD risk factors among a large industrial population of northern India.
    Methods.We conducted a cross-sectional survey among all employees aged 20–59 years of a large industry near Delhi (n=2935), to evaluate their cardiovascular risk profile—by employing a structured questionnaire and clinical and biochemical estimations. The presence of coronary heart disease was ascertained by evidence of its treatment, Rose angina questionnaire and Minnesota coded electrocardiograms.
    Results.The results for 2122 men, in whom complete information was available, are reported here. The mean age was 42 years and 90% of the men were below 50 years of age. The prevalence of major CVD risk factors (95% CI) was: hypertension 30% (28%–32%), diabetes 15% (14%–17%), high serum total cholesterol/HDL ratio (>4.5) 62% (60%–64%) and current smoking 36% (34%–38%). Forty-seven per cent of the respondents had at least two of these risk factors. Another 44% (95% CI: 42%–46%) had pre-hypertension (JNC VII criteria) and 37% (95% CI: 35%–39%) had evidence of either impaired fasting glucose or impaired glucose tolerance. Thirty-five per cent (95% CI: 33%–37%) of the individuals were overweight (BMI >25 kg/m2) while 43% (95% CI: 40%–45%) had central obesity (waist circumference >90 cm). The metabolic syndrome was present in 28%–35% of the individuals depending on the diagnostic criteria used. The prevalence of several risk factors and the metabolic syndrome was high with increasing age, BMI and waist circumference. A third of those who had hypertension (31.5%) and diabetes (31%) were aware of their status. Among those aware, adequate control of blood pressure and blood glucose was present in only 38% of those with hypertension and 31% of those with diabetes, respectively. Coronary heart disease was present in 7.3% of the individuals while 0.3% had a history of stroke.
    Conclusion.This study demonstrates the high prevalence of CVD and its risk factors against a background of poor awareness and control among a comparatively young male population in a north Indian industrial setting.
Natl Med J India 2005;18:59–65

INTRODUCTION
The cardiovascular disease (CVD) burden of India is expected to double in the next two decades, making it the single largest cause of death and the second largest cause of disability by the year 2020.1 This will be characterized by an enormous burden of CVD among urban communities. Further, the prevalence of CVD in rural and semi-urban areas is expected to increase substantially.2,3 While the exact aetiology of this predisposition to CVD in Indians is still debated, from a public health point of view it is clear that the rapid transition in diet and lifestyles with urbanization has contributed to increasing levels of potentially reversible CVD risk factors.4 Data from several cross-sectional studies confirm the high prevalence of risk factors such as smoking, type 2 diabetes, high blood pressure, dyslipidaemia and obesity5–11 in urban Indians. Despite this high burden, there is poor awareness among Indians, in addition to low detection and control rates.12–18 The reasons for this include low literacy, lack of access to healthcare and competing priorities such as infectious and nutritional diseases. Given this background and with the expected overall improvement in literacy and socioeconomic status, it would be useful to study the awareness of CVD risk factors and lifestyle patterns in industrial settings, where the access to healthcare is usually better than that among the general community.
    Recent research has highlighted the importance of a continuum of risk with variables such as blood pressure and glucose.19–23 There is robust evidence of increased vascular risk even in the presence of pre-hypertension (or high-normal blood pressure) and pre-diabetes (impaired fasting glucose and impaired glucose tolerance).24,25 These risk categories have a high rate of progression to hypertension and type 2 diabetes. They are also components of the metabolic syndrome, which confers an almost 2-fold risk for future diabetes, CVD mortality and morbidity.26,27 However, detailed data regarding the burden of these risk factors in India are limited.
    We report the findings of a survey done among the employees of a large organized sector industry located close to Delhi. The main objectives of the study were (i) to study the risk factor and treatment status in individuals with respect to CVD, (ii) to estimate the prevalence of risk factors for coronary heart disease (CHD), both traditional and intermediary, and (iii) to estimate the prevalence of CHD using a history of previously diagnosed disease, Rose angina questionnaire and electrocardiogram (ECG).

METHODS
Study design and setting
We did a cross-sectional survey of all employees in the 20–59 years’ age group in an organized sector industry, during 1995–98. A total of 2935 employees (344 women and 2591 men) were eligible to participate. All potential participants received a letter inviting them to participate in the study. Approval for conducting the study was obtained both from the management and the employee representatives. Participation in the study was voluntary and all the measurements were carried out in the medical unit located within the industry. Informed consent was obtained from each participant.

Study questionnaire
We used a simple structured questionnaire which was administered by research assistants, who had received training in the methodology of field-based cardiovascular surveys. The questionnaire was used to elicit information from each study participant for the following variables: (i) demographic characteristics such as age, marital status, profession, education and socioeconomic status; (ii) lifestyle-related factors such as dietary habits, physical activity, tobacco and alcohol consumption; (iii) the health status of the individual and any history related to heart disease, stroke, hypertension and diabetes. The presence of angina was assessed using the Rose angina questionnaire. If a person reported the presence of hypertension or diabetes or CVD (defined as the presence of CHD or stroke or angina), a detailed questionnaire was administered to elicit information on control, awareness, treatment and limitations due to the illness. Each question was structured to gather as much information as possible without increasing the duration of the interview, which usually lasted about 30 minutes.

Clinical measurements and anthropometry
A 30 second pulse rate was measured in each individual followed by 2 blood pressure measurements using a random zero sphygmomanometer by standard techniques.28 The measurements were obtained half an hour apart, either before or 1 hour after the first blood sampling. The lower of the two measurements was used for this analysis. Height was measured in centimetres using a stadiometer, which was calibrated to the nearest 0.5 cm, and weight was measured using a bathroom scale, which was validated on a daily basis with known weights.
    Waist and hip circumferences were measured in the following manner. The subject was asked to stand with the arms by the sides and to breathe normally. The following points were marked on the right side: (i) the subcostal margin in the mid-axillary line, (ii) the highest point of the iliac crest in the mid-axillary line. The centre of these two points was marked. At this point the waist circumference was measured with a fibreglass tape after applying a tension of 600 g (done with the help of a spring balance). The measurement was made up to the nearest millimetre. Adequate care was taken to ensure that the subject was breathing normally at the time the measurement was taken. Hip circumference was measured at the level of the greater trochanters with the subject standing and breathing normally in a manner similar to that during measurement of the waist circumference. All the measurements were taken without intervening clothes, with the subject wearing a loose-fitting surgical gown.
    An ECG was obtained after the clinical measurements. All ECGs were read by a cardiologist and coded using the Minnesota coding system. Ten per cent of randomly selected ECGs were checked by two different cardiologists for agreement. There was complete agreement between the original and fresh coding. (Kappa was not estimated in view of the complete agreement.)

Laboratory measurements
Blood samples were drawn by trained personnel, centrifuged and stored for later analysis. Laboratory measurements included estimation of fasting blood glucose, 2-hour post-load blood glucose (all post-load measurements were made 2 hours after a 75 g glucose solution was administered orally under supervision), plasma insulin (both fasting and post-load), total blood cholesterol, high density lipoprotein (HDL) cholesterol (HDL-C), and triglycerides (both fasting and post-load). Glucose was analysed by the glucose oxidase method (GOD-PAP), cholesterol by the CHOD-PAP method, and triglycerides were estimated by the GPOD-PAP method. The intra- and interassay coefficient of variation (CV) for glucose was 1.5% and 2%, respectively; for cholesterol and triglycerides these were <2% and <2.5%, respectively. HDL-C was estimated by the precipitation method (phosphotungstate/Mg) and the inter- and intra-assay CVs were <2.5%. Insulin concentrations were measured by radioimmunoassay (Coat-a-Count insulin kit, Diagnostic Products). The intra- and interassay CVs were <5% and <7.5%, respectively.

Definition of risk factors
The definitions used for the categorization of risk levels among the participants are given in Box 1.

Statistical analysis
Statistical analysis was performed using the SPSS Version 9.0 package. Continuous data are summarized as means and standard deviations and categorical data as proportions. We used the Student t test for comparing means and a chi-square test for categorical variables.

RESULTS
Of the 2935 eligible individuals (11.7% women), 2548 (86.8% of total, 266 women and 2282 men) agreed to participate in the study. We report here the results of the men, as the number of women who were eligible and participated in the study was small (10.4%). Of the 2282 men, 14 declined to undergo clinical examination or give blood samples. Complete data for all the risk variables were available for 2122 subjects.

Demographic characteristics
The mean age of the participants was 42 years. Nearly 60% of the participants were in the 41–50 years’ age group (Table I). A large

majority were educated with 66.4% having graduate, postgraduate or professional qualifications. Twenty-one per cent had secondary level education and the remaining 12.6% were illiterate (Table I). The majority of employees were either skilled workers (47%) or professionals (26%). Semi-skilled and unskilled workers constituted 14% and 2.4% of the workforce, respectively. The remaining were clerical staff (11.5%).

Risk factors
The mean values of various clinical, laboratory and anthropometric measurements are given in Table I. The age-stratified and total prevalence of CVD risk factors are given in Table II. Normal blood pressure (systolic pressure <120 mmHg and diastolic pressure <80 mmHg, respectively) was observed in a quarter (25.6%) of the individuals surveyed. The prevalence of pre-hypertension and hypertension as per the Seventh Joint National Committee (JNC VII) criteria was 44% (95% CI: 42%–47%) and 30% (95% CI: 28%–32%), respectively. Among those with hypertension only one-third (31.5%) were aware of their blood pressure status. Of these, only 38% (12% of the total who were hypertensive) had their blood pressure under control (<140/90 mmHg). The proportion of individuals with hypertension who were aware of their disease increased with age (from 19% of all those with hypertension in the 20–29 years’ age group to 40.5% in the 50–59 years’ age group). Diabetes was present in 15% (95% CI: 14%–17%) and another 30.7% (95% CI: 29%–33%) had evidence of abnormal glucose homeostasis, denoted by impaired fasting glucose. Impaired glucose tolerance, an indicator of insulin resistance, was observed in 13.2% (95% CI: 12%–15%) of participating individuals (Table II). While 31% of those with diabetes were aware of their status, in only 17% of these (5% of the people with diabetes) was the blood glucose controlled to normoglycaemic levels (fasting plasma glucose <100 mg/dl and postprandial glucose <140 mg/dl)). The proportion of individuals with diabetes who were aware of their disease increased with age (from 20% of all those with diabetes in the 20–29 years’ age group to 42.1% in the 50–59 years’ age group).

Hypercholesterolaemia (total cholesterol ³200 mg/dl) was observed in 30.1% (95% CI: 28%–32%) of individuals, and hypertriglyceridaemia (serum fasting triglycerides ³150 mg/dl) in 38% (95% CI: 38%–42%) of those surveyed (Table II). A large proportion (67.2% [95% CI: 65%–69%]) of individuals had low HDL-C levels (serum HDL-C <40 mg/dl), which resulted in nearly two-thirds (62%) of individuals having a total cholesterol/HDL-C ratio of ³4.5. Overall, this population had a high rate of tobacco use (48.3%); 36% of the individuals were current smokers (cigarettes 21.5% and beedis 19.7%), and one-quarter of the population used non-smoking forms (Table II). The commonest non-smoking form of tobacco use was chewing tobacco or pan masala mixed with tobacco.
One-third of the population (35% [95% CI: 33%–37%]) was overweight using the traditional cut-off of BMI ³25 kg/m2. However, as per the recent WHO recommendations for defining health risks associated with BMI in Asians, 12.2% were in the ‘high to very high’ risk (BMI ³27.5 kg/m2) category and 58.5% in the ‘moderate to high risk’ (BMI ³23 kg/m2) category29 (Table II). Similarly, very few individuals were centrally obese by the National Cholesterol Education Program (NCEP) criteria of waist circumference >102 cm (7.2%). However, 43% (95% CI: 40%–45%) of the individuals had a waist circumference ³90 cm, the proposed criteria for central obesity in men in the Asia-Pacific region,30 while 26.5% (95% CI: 25%–28%) had a waist circumference >94 cm, the proposed first action level in the Caucasian population.31

The metabolic syndrome was present in 28.1% (95% CI: 26%–30%) of the study population (Table III) by the NCEP Adult Treatment Panel (ATP) III criteria (see Box 1).32 Using the criteria for genetically susceptible populations (waist circumference >94 cm instead of 102 cm), the prevalence of the metabolic syndrome was higher (35%). Using the modified WHO criteria33 for defining the metabolic syndrome (see Box 1 and Table III), the prevalence was 29% (95% CI: 27%–31%). The prevalence of the metabolic syndrome using the modified NCEP definition was 29% after individuals with diabetes were excluded, and 21.7% when only individuals with normoglycaemia were considered. Among 1798 individuals without diabetes, fasting serum insulin levels were available for 1552 persons. The mean and median values of insulin were 8.8 IU/L and 5.0 IU/L, respectively. After log transformation of insulin values to make the distribution normal, the mean values of insulin were consistently higher in those with the metabolic syndrome as compared to those without (1.9 and 1.67, respectively, among those with and without the metabolic syndrome, p for difference <0.001) irrespective of the way the metabolic syndrome was defined.
Only 15% of the study population were free from any of the risk factors (hypertension, diabetes, current smoking and total cholesterol/HDL-C ratio ³4.5). While 85% of the study population had at least 1 of these risk factors, about 47% of the employees had at least 2 co-existing risk factors.
Effect of age on CVD risk factors
While diabetes, hypertension, hypertriglyceridaemia, overweight, central obesity and the metabolic syndrome, as well as the presence of multiple risk factors showed rising trends with age, the prevalence of smoking, low HDL and adverse total cholesterol/HDL-C ratio was roughly similar in all the age groups (Table II). Even the youngest age group of 20–29 years had a high prevalence of risk factors such as dyslipidaemia, hypertension and diabetes.
Effect of adiposity on CVD risk factors
There was a continuous graded increase in mean values as well as risk factor prevalence with increasing BMI (Fig. 1) and waist circumference (Fig. 2). The risk of developing obesity-associated risk factors started to rise much below the defined thresholds for western populations (BMI >25 kg/m2 and waist circumference >94 cm).
Prevalence of CVD

Fig 1. Prevalence of risk factors in the four quartiles of body mass index (all values in kg/m2) Quartile I: 13.9–21.2; Quartile II: 21.3–23.7; Quartile III: 23.8–25.8; Quartile IV: 25.9–37.2 *TC:HDL ratio total cholesterol:high density lipoprotein cholesterol ratio † Metabolic syndrome as defined by the NCEP-ATP III but excluding waist circumference from the definition

Fig 2. Prevalence of risk factors in the four quartiles of waist circumference (all values in cm) Quartile I: 51.4–81.0; Quartile II: 81.1–88.0; Quartile III: 88.1–94.5; Quartile IV: 94.6–131 *TC:HDL ratio total cholesterol:high density lipoprotein cholesterol ratio † Metabolic syndrome as defined by the NCEP-ATP III but excluding waist circumference from the definition

A past history of diagnosed CHD and stroke was identified in 16 (0.8%) and 7 (0.3%) individuals, respectively. The prevalence of CHD, defined as history of CHD (identified by the modified Rose questionnaire) or presence of ECG abnormalities suggestive of CHD (by specified Minnesota codes), was noted in 7.3% of men. Abnormal ECGs suggestive of CHD in the form of Q waves, T wave and ST–T abnormalities were observed in 2.5% of the population studied.
DISCUSSION
Our study on the cardiovascular risk profile in a industrial workplace setting in India, among a relatively young urban population, found the prevalence of traditional risk factors to be high. Even in the 20–29 years’ age group, only one-quarter of the individuals had normal blood pressure, only half had normal glucose tolerance, more than half had dyslipidaemia, one-fifth were overweight, and more than one-fifth had at least 2 major risk factors. Awareness and control of hypertension and diabetes was poor, indicating low detection and poor management of major CVD risk factors. The high prevalence of ‘risk markers’ such as impaired fasting glucose and glucose tolerance, and pre-hypertension suggests that there is a large vulnerable population which can develop an overt adverse risk profile and CVD in the future.
Community-based cross-sectional surveys and other studies carried out in the past few decades have revealed a high and increasing prevalence of CVD and its risk factors in individuals of South Asian origin residing, in India as well as in other countries.5–11,34–37 However, not many of these have evaluated the CVD risk profile in younger populations. The disconcertingly low level of awareness of CVD and its treatment further increases the risk for future morbidity and mortality due to CVD. The low awareness and control noted in our population is similar to data published from other developing countries such as China. Tao et al.38 from their nationwide survey of blood pressure of more than 950 000 men and women, found that of all the hypertensives, only 25% were aware of their hypertensive status and, among those aware, only half were on any blood pressure lowering medications. Overall, in only 3% was the blood pressure controlled to normal levels, although this had improved by 2000–01.39 Further, a recent report highlights the urban–rural differences in the awareness, treatment and control of hypertension. The Hypertension Study Group40 while noting similar low rates of awareness and control, identified visit to a physician in the previous 1 year, higher educational attainment and female gender to be important correlates of awareness among a cross-section of subjects with hypertension from India and Bangladesh. Therefore, while opportunistic screening may identify individuals with undetected hypertension and diabetes, we need specific strategies to improve the control and therapy of those with CVD risk factors.
The importance of insulin resistance in determining the high mortality due to CVD among Indians has been highlighted by McKeigue41 and Hughes.42 Based largely on data obtained from migrant Indians and comparing them to indigenous populations, Indians were believed to have an enhanced propensity to insulin resistance and CVD.41 Though others have questioned the role of insulin resistance and emphasized the importance of conventional risk factors such as lipids, the high prevalence of the metabolic syndrome and its association with hyperinsulinaemia indirectly underscores the importance of insulin resistance among Indians. We explored the causes for the high prevalence of the metabolic syndrome in our study population and found overweight and central obesity to be major correlates of these risk markers (Figs. 1 and 2). As there is a pronounced urban–rural gradient in obesity among Asian Indians,43 we believe that the high prevalence of risk markers such as impaired glucose tolerance, the metabolic syndrome and high-normal blood pressure, all of which are influenced by body fat, is a reflection of the current urbanization and adaptation of a western lifestyle among the Indian population. The prevention and management of the metabolic syndrome largely involves lifestyle modifications. Recently published clinical trials such as the Diabetes Prevention Programme44 and the Finnish diabetes prevention studies45 have established the importance of lifestyle measures in the prevention of diabetes. Similarly, a low level of physical activity and a diet rich in saturated fats and high glycaemic foods have been implicated in the pathogenesis of the metabolic syndrome.46
While such a high prevalence of risk factors is a cause for concern, organized sector industries provide a unique opportunity for carrying out prevention programmes. It is estimated that there are nearly 6 million people working in such large organized sector industries in India. While most of these have their own primary healthcare facilities, they also provide for the healthcare of employees and their dependents in higher medical institutions when required. With the rising burden of CVD, the expenditure on such healthcare programmes by every industry is likely to increase enormously. Thus, there is a lot of scope for and benefit in initiating comprehensive, low-cost CVD prevention programmes at the workplace for employees and their dependants. Such onsite programmes have been found to be modestly successful in the West through increased awareness, health education and risk reduction interventions, and modified models could easily be adapted to Indian settings. This is likely to lead to a healthier workforce as well as a decrease in expenditure by the industry on treatment costs and increased absenteeism. We hope that our study will provide the stimulus for initiating such surveillance and preventive activities.
The prevalence rates that we have reported are based on western cut-offs and we require a longitudinal evaluation of our study population on CVD morbidity and mortality to fully understand the implications of our findings. To address the above issues, and to document changes in risk factors and prospectively identify determinants of CVD, we are continuing our study in this industrial population.
Limitations
As the study was carried out in an industrial setting, it may not be representative of the general population. Though we have provided age-specific prevalence rates, the industrial population is different from the general population in terms of its socioeconomic profile. However, this study is broadly generalizable to similar large, organized sector workforce involving nearly 6 million employees. Further, our study population comprised predominantly young men. Also, women comprised only 10% of the workforce. The results of this study cannot be generalized to individuals older than 60 years, in whom the risk factor as well as disease burden will be even higher than that reported in this population.
Conclusion
We have demonstrated a high prevalence of CVD risk factors, its markers such as the metabolic syndrome and risk determinants such as obesity (both generalized and central) among a group of young men. Though the study population was not representative of the national population (older, only men, higher educational and socioeconomic strata), we believe that it does represent the rising trend of CVD in urban India. Despite the availability of onsite healthcare and a high level of education as compared to the general population, the awareness and treatment of hypertension and diabetes is low. Therefore, there is a pressing need to initiate onsite preventive programmes to identify and manage individuals at high risk for future CVD.
ACKNOWLEDGEMENTS
The International Clinical Epidemiology Network (INCLEN) provided partial financial support for this study.
We gratefully acknowledge the cooperation of the employees and the management of Bharat Electronics Limited for participating in this study, Dr S. K. Puri and Dr Medha Joshi for providing clinical support at the industrial site, A. Karthika for data management, and all the research and support staff, without whom this study would not have been possible.