Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study

Raj Bhopal, Nigel Unwin, Martin White, Julie Yallop, Louise Walker, KGMM Alberti, Jane Harland, Sheila Patel, Naseer Ahmad, Catherine Turner, Bill Watson, Dalvir Kaur, Anna Kulkarni, Mike Laker, Anna Tavridou

Raj Bhopal, professor

Martin White, senior lecturer in public health medicine

Nigel Unwin, senior lecturer in epidemiology

Julie Yallop, research associate

Louise Walker, research associate

Jane Harland, research associate

Sheila Patel, research associate

Naseer Ahmad, research associate

Catherine Turner, research nurse

Bill Watson, research associate

Dalvir Kaur, study administrator

Anna Kulkarni, research nurse

KGMM Alberti, professor

Mike Laker, reader

Anna Tavridou, PhD student

Correspondence to: R Bhopal Public Helath Sciences, Medical School, Edinburgh EH8 9AG Raj.Bhopal{at}ed.ac.uk

Abstract

Objective To compare coronary risk factors and disease prevalence among Indians, Pakistanis, and Bangladeshis, and in all South Asians (these three groups together) with Europeans.

Design Cross sectional survey.

Setting Newcastle upon Tyne.

Participants 259 Indian, 305 Pakistani, 120 Bangladeshi, and 825 European men and women aged 25-74 years.

Main outcome measures Social and economic circumstances, lifestyle, self reported symptoms and diseases, blood pressure, electrocardiogram, and anthropometric, haematological, and biochemical measurements.

Results There were differences in social and economic circumstances, lifestyles, anthropometric measures and disease both between Indians, Pakistanis, and Bangladeshis and between all South Asians and Europeans. Bangladeshis and Pakistanis were the poorest groups. For most risk factors the Bangladeshis (particularly men) fared the worst: smoking was most common (57%) in that group, and Bangladeshis had the highest concentrations of triglycerides (2.04 mmol/l) and fasting blood glucose (6.6 mmol/l), and the lowest concentration of high density lipoprotein cholesterol (0.97 mmol/l). Blood pressure, however, was lowest in Bangladeshis. Bangladeshis were the shortest (men 164 cm tall v 170 cm for Indians and 174 cm for Europeans). A higher proportion of Pakistani and Bangladeshi men had diabetes (22.4% and 26.6% respectively) than Indians (15.2%). Comparisons of all South Asians with Europeans hid some important differences, but South Asians were still disadvantaged in a wide range of risk factors. Findings in women were similar.

Conclusion Risk of coronary heart disease is not uniform among South Asians, and there are important differences between Indians, Pakistanis, and Bangladeshis for many coronary risk factors. The belief that, except for insulin resistance, South Asians have lower levels of coronary risk factors than Europeans is incorrect, and may have arisen from combining ethnic subgroups and examining a narrow range of factors. A coronary heart disease strategy based on a wide range of risk factors which accounts for the heterogeneity among South Asians is needed.

Introduction

Coronary heart disease is apparently commoner in South Asians in Britain than in the general population ⁽¹⁾ despite lower levels of several classic coronary risk factors. ⁽²⁾⁽³⁾ Insulin resistance is proposed to be the underlying factor in high rates of coronary heart disease among South Asians worldwide and has been related to exercise and obesity. ^{(2) (3) (4) (5) (6)} Bhopal, and Shaukat and de Bono, however, emphasised a wide range of risk factors including smoking and poverty. ⁽⁷⁾⁽⁸⁾ Williams et al concluded that South Asians had a higher prevalence of a broad range of non-biochemical risk factors than the general population. ⁽⁹⁾ Nazroo showed that the prevalence of self reported coronary heart disease was higher in Bangladeshis and Pakistanis combined, and lower in Indians, than in the white population after standard of living was adjusted for. ⁽¹⁰⁾ The Newcastle heart project compared coronary heart disease risk factors in Indians, Pakistanis, and Bangladeshis and also compared South Asians as a whole with Europeans. ^(11)(12)

Participants and methods

The methods and some data on the European study have already been published. ^(11)(12) South Asians are defined as Newcastle residents with ancestral origins in India, Pakistan, or Bangladesh and who had three or more grandparents born there. Indians, Pakistanis, and Bangladeshis self identified as such at interview, using 1991 census categories of ethnic group. Europeans are defined as Newcastle residents with ancestral origins in European countries and were identified by excluding people from ethnic minority populations. In referring to published work we generally use the authors’ terms (white, general population, etc).

Sample

We selected European subjects from the sample of 6448 people aged 25-74 identified from the family health services authority register for the Newcastle health and lifestyle survey. ⁽¹³⁾ People with South Asian sounding names were selected from the full register. ⁽¹⁴⁾ Detailed lists of South Asian and English sounding names were drawn up (details from authors) and used in a computer program, which identified 7459 South Asian sounding names of people aged 25-74 years. Where there was doubt about a name it was placed in the South Asian list. The European and South Asian sampling frames were each divided into 10 year age and sex strata and equal numbers from each stratum randomly selected.

Recruitment

European sample—Europeans were screened between April 1993 and October 1994. Subjects were sent a letter about the study followed by an invitation to attend for screening. Up to three invitations were sent to non-respondents. Those who still did not respond were sent a letter by recorded delivery, and this informed us if they were still resident at the address.

South Asian sample—South Asians were screened between May 1995 and March 1997. South Asian community leaders were contacted, and the study was promoted through the local media. Material relating to the study was produced in five languages (English, Punjabi, Urdu, Bengali, and Hindi), and the research team had the necessary language skills and cultural knowledge. Subjects were sent a recorded delivery letter about the project, a reply slip, and a reply paid envelope. If recorded delivery letters were returned by the Post Office, the general practitioner and family health services authority were contacted to check accuracy of addresses. Recorded delivery reminder letters were sent to updated addresses. Non-respondents were contacted by telephone or, when this was not possible, by home visiting on up to three occasions at different times of different days.

Sample size

Our target sample size for both the South Asian and European samples was 600 men and 600 women aged 25-74 years. This target was partially determined by the resources available. However, it was also based on power calculations that showed this sample size would achieve reasonably precise estimates of the prevalence and mean values of key risk factors and would detect differences of expected magnitude between Europeans and South Asians in the prevalence of risk factors. For example, based on a study power of 80% and a significance level of 5% (and SD 1.25 mmol/l) 612 people in each group are needed to detect a mean difference in blood cholesterol of 0.2 mmol/l. To detect a mean difference of 0.5 mmol/l requires 98 people per group. For this variable, our achieved sample size was sufficient to reliably detect small differences between South Asians and Europeans but not between the Indians, Pakistanis, and Bangladeshis. Observations of no difference between Indians, Pakistanis, and Bangladeshis tend therefore to be statistically unreliable.

Data collection

Participants attended the Royal Victoria Infirmary, Newcastle, between 0800 and 1000. They had received written information about procedures, and this was repeated verbally on attendance. Informed consent was obtained.

Biochemical measurements

Participants who did not require insulin fasted from 2200 the night before. People with diabetes taking hypoglycaemic drugs or insulin were classified as diabetic irrespective of their blood glucose concentration. Participants without diabetes took an oral glucose tolerance test. They drank 75 g of glucose in 388 ml of water, and venous blood was taken after 1 and 2 hours. Plasma was separated by centrifugation either immediately or within an hour (after refrigeration of the sample). Plasma glucose concentration was measured by the glucose oxidase method with automated colorimetry on a Hitachi 717 analyser. ⁽¹⁵⁾ Diabetes was defined as a glucose concentration ³ 11.1 mmol/l after 2 hours and impaired glucose tolerance as a concentration ³ 7.8 and <11.1 mmol/l. Insulin was measured by an enzyme linked immunosorbent assay (ELISA; DAKO Diagnostics, Ely).

Lipid concentrations were measured in blood samples taken before the glucose tolerance test. Cholesterol concentration was measured by a cholesterol oxidase-peroxidase method with calibrants traceable to the Centres for Disease Control definitive method. Triglyceride concentrations were estimated by a lipase-glycerol kinase method and high density lipoprotein cholesterol by measuring the supernatant cholesterol concentration after precipitation of apolipoprotein B-containing lipoproteins with heparin and manganese. Low density lipoprotein cholesterol concentrations were calculated by the Friedewald formula. ⁽¹⁶⁾ Until May 1994 the lipid analyses were performed on a Cobas Bio centrifugal analyser (Roche Products, Welwyn Garden City), and after this date a DAX analyser was used (Bayer, Basingstoke). The data for triglycerides confirmed a positive bias with the DAX analyser, and therefore results obtained with the DAX were adjusted as previously described. ⁽¹¹⁾

Concentrations of apolipoproteins A-I and B were estimated by immunoturbidometry with a DPA analyser (Bayer). Throughout the study period the laboratory participated in an external quality assurance scheme for lipids, and there was no change in inaccuracy of total cholesterol measurement over this period relative to the external standard. Lp(a) lipoprotein was measured by a latex agglutination method, and fibrinogen by an Instrumentation Laboratory Automated Coagulation Analyser 3000.

Anthropometric measurements, blood pressure, pulse, and electrocardiography

The same measurement protocols, training, and quality control measures were used throughout the study. These included monthly checks on the procedures followed by all observers and a mandatory refresher course on all measurements for all observers every 6 months. Height was measured without shoes to the nearest 5 mm, with the head positioned so that the eye and the external auditory meatus were level, and weight was measured to the nearest 100 g with the subject lightly clothed. With the subject standing, waist circumference was measured to the nearest centimetre at the midpoint between the lower costal margin and the superior iliac crest, and the hips over the greater trochanters. Waist circumference was measured with the waist unclothed and hip circumference over the subject’s underwear.

Blood pressure and pulse were measured following the method of the British Hypertension Society ⁽¹⁷⁾ with a mercury sphygmomanometer and a 12.5? 35 cm cuff. Blood pressure and pulse were measured twice after the subject had been seated for at least 20 minutes after venepuncture. The average systolic (Korotkoff phase I) and diastolic (Korotkoff phase V) blood pressures, based on the mean of the two measurements, were used for analysis.

We used established criteria to define risk factors for coronary heart disease. ^(18)(19) Participants with a carbon monoxide concentration >8 ppm on the Bedfont smokerlyzer carbon monoxide monitor ⁽²⁰⁾ and who did not admit to smoking were counted as carbon monoxide adjusted smokers.

A 12 lead resting electrocardiogram was recorded. The traces were Minnesota coded by two trained independent coders, and discrepant reports were passed to a third independent coder. Probable coronary heart disease was defined as Minnesota coding 1.1-1.2 (large Q and QS) and possible coronary heart disease as Minnesota coding 1.3 (small Q and QS), 4.1-4.4 (ST-T depression), 5.1-5.3 (T wave inversion and flat), or 7-1-1 (complete left bundle branch block). ⁽²¹⁾

Questionnaire

Participants completed a questionnaire that included questions on state of health, health behaviour (exercise, diet, smoking, and alcohol consumption), socioeconomic circumstances, experience of racism, and family history of cardiovascular diseases. The questionnaire included the Rose questionnaire for angina and chest pain, and this was interpreted as recommended. ⁽²²⁾ Europeans self completed the questionnaire at screening, with help from staff if needed.

The questionnaire was translated from English into four South Asian languages then translated back into English by another person. The original English version and the back translated version were compared question by question. If there was doubt about equivalence of meaning translators and researchers conferred until agreement was reached. The questionnaire was completed at interview in the participants’ homes and in their preferred language by South Asian interviewers of the same sex as the participant. Copies of the questionnaire are available from the authors on request.

Analysis of data

We used SPSS software for analysis. ⁽²³⁾ Direct age standardisation was to the 1991 England and Wales population. Differences between Europeans and South Asians for continuous variables were assessed by using independent samples t tests, and differences between Indians, Pakistanis, and Bangladeshis by analyis of variance. We adjusted data on income, for men only, for household composition using the formula income/(1+0.7? adults+0.5? children). ⁽²⁴⁾ As triglycerides, Lp(a) lipoprotein, and insulin had skewed distributions log transformations were used in analysis, and geometric means are presented. Age adjusted variances were calculated for categorical variables. ⁽²⁵⁾

Ethics

Newcastle upon Tyne joint ethics committee approved the study. Informed consent was obtained from participants. If participants did not consent to three venepunctures, our priority was baseline and then 2 hour venepuncture. For example, 20 Indian, 26 Pakistani, and 22 Bangladeshi women did not consent to a 2 hour sample.

Results

Of 2160 people with South Asian sounding names, 1050 people were eligible and contacted; 288 refused and 53 completed only the interview, leaving 709 (67.5% of 1050). Of these, 684 classified themselves as Indian, Pakistani, or Bangladeshi. Of 1744 people sampled from the Newcastle health and lifestyle survey, 1308 were contacted and 840 were screened (64.2%). Fourteen were South Asian and one of African origin, leaving 825 Europeans.

Population characteristics

Table w1 shows the demographic and socioeconomic characteristics of the men. Bangladeshis were the youngest and Europeans the oldest. Most South Asian men were born abroad and migrated as adults. Bangladeshis were the most recent immigrants. South Asians included Muslims, Hindus, and Sikhs. Indians were most, and Bangladeshis least, educated. There were more Indians in social classes I, II, and IIIN (70%) than any other ethnic group and fewest Bangladeshis (26%). Levels of reported racism relating to work or housing were low (<10%) in all three South Asian groups, with Pakistanis reporting most racism at work. Indians and Europeans had the highest median income and Bangladeshis the lowest.

Table w1 Background, social and economic characteristics of men in study. Values are numbers (percentages) and denominators with 10 of n unless stated otherwise
 

Bangladeshi and Pakistani women were younger than the other groups (table w2). Most Bangladeshis (76%) migrated after 1975. More Indian (13%) than European women (8%) had attended higher education and no Bangladeshi women had higher education. There were fewer Pakistani and Bangladeshi women in social classes I, II, and IIIN than European or Indian women.

Table w2Background, social and economic characteristics of women in study. Values are numbers (percentages) and denominators with 10 of n unless stated otherwise

Lifestyle

Table w3 shows the highest prevalence of smoking was in Bangladeshi men. Pakistanis and Indians were most likely to eat fruit or vegetables daily. Few Pakistanis and Bangladeshi men drank alcohol; most Indians did. Indians were the most physically active South Asians, and Bangladeshis the least. Large differences were seen between Europeans and South Asians except for smoking.

Table w3 Smoking, diet, alcohol consumption, and exercise among men and women. Values are numbers (percentages) weighted to 1991 England and Wales population. Denominators are within 10 of n unless stated otherwise

Few Indian, Pakistani, or Bangladeshi women smoked or drank alcohol (table w3). Daily consumption of fruit or vegetables was commoner in Indians than in Bangladeshis and Europeans. Bangladeshis were the least, and the Indians the most, physically active South Asians. The differences between Europeans and South Asians were large.

Biological characteristics

Among South Asian men, Indians were the tallest and heaviest and Bangladeshis shortest and lightest (table w4). Indians had the highest body mass index and largest waists and hips, and the Bangladeshis had the smallest (waist: hip ratios were similar). South Asians had more central obesity than Europeans.

Table w4 Biological characteristics of men weighted to 1991 England and Wales population. Figures are means (SD) and denominators are within 10 of n unless stated otherwise

Glucose and insulin measurements were similar among Indians, Pakistanis, and Bangladeshis. South Asians had higher concentrations than Europeans. Among South Asians, Indians had the highest total cholesterol and high density lipoprotein cholesterol concentrations and Bangladeshis the lowest; triglyceride concentrations were highest in Bangladeshis. South Asians and Europeans had similar concentrations of cholesterol. Apolipoprotein A concentrations were higher in Indians and Pakistanis and lower in Bangladeshis. Bangladeshis had higher Lp(a) lipoprotein concentrations than Indians or Pakistanis.

White cell count was highest in Bangladeshis. Bangladeshis had lower blood pressure than Indians and Pakistanis. Europeans had higher blood pressure than South Asians.

Bangladeshis women were shortest and lightest, and Indian women heaviest (table w5). Waist:hip ratios (lowest in Indians) differed among South Asian groups. Europeans had a lower waist:hip ratio than South Asians. Fasting glucose and 2 hour insulin concentrations were higher in Bangladeshi women than Indians and Pakistanis (insulin not significant), and higher in South Asians than Europeans.

Table w5 Biological characteristics of women weighted to 1991 England and Wales population. Figures are means (SD) and denominators are within 10 of n unless stated otherwise
 

Bangladeshis had lower high density lipoprotein cholesterol concentrations but higher triglyceride concentrations than Indians and Pakistanis, but low density lipoprotein cholesterol concentration was similar. Europeans had higher concentrations of high density lipoprotein cholesterol than South Asians. Among South Asians, apolipoprotein A concentration was lowest and apolipoprotein B highest in Bangladeshis. Lp(a) lipoprotein concentrations were higher in Pakistanis than Indians and Bangladeshis, and higher in South Asians than Europeans. Fibrinogen concentration was higher in Indians than in Bangladeshis or Pakistanis. Systolic blood pressure was lower in Bangladeshis than in Pakistanis or Indians.

Prevalence of clinical problems

Table w6 shows non-significant variation in diabetes between the three male South Asian groups, which collectively had a five times higher prevalence of diabetes than Europeans. There were important differences between the three South Asian groups in total cholesterol:high density lipoprotein cholesterol ratio and triglyceride concentration. South Asians had lower high density lipoprotein cholesterol concentration, higher total cholesterol:high density lipoprotein cholesterol ratio and higher triglyceride concentrations than Europeans.

Table w6 Prevalence of glucose intolerance, dyslipidaemia, obesity, hypertension, and coronary heart disease, in men. Values are numbers (percentages) age adjusted to 1991 England and Wales population

Obesity was commoner in Pakistanis and Indians than in Bangladeshis. Differences in central obesity between South Asian groups were small. Waist:hip ratios of ³ 0.95 were commoner in South Asians than Europeans. Differences in hypertension between Indians, Pakistanis, and Bangladeshis were not significant, but hypertension was least common in Bangladeshis, and less common in South Asians than Europeans.

Prevalence of Rose angina was similar among the South Asian groups and between them and Europeans. Possible myocardial infarction was higher (not significant) in Bangladeshis than Indians or Pakistanis. Evidence of coronary heart disease on electrocardiography was similar among the three South Asian groups, which combined had more probable heart disease than Europeans on electrocardiography.

A higher proportion of Indian women had diabetes than Pakistanis and Bangladeshis (not significant). Diabetes was four to five times commoner in South Asians than in Europeans (table w7). Among South Asians, Bangladeshis were most likely to have low high density lipoprotein cholesterol, high cholesterol:high density lipoprotein cholesterol ratio and high triglyceride concentrations. South Asians had a less favourable lipid profile than Europeans with the exception of total cholesterol.

Table w7 Prevalence of glucose intolerance, dyslipidaemia, obesity, hypertenion, and coronary heart disease in women. Values are numbers (percentages) age adjusted to 1991 England and Wales population

Obesity was commoner in Indian and Pakistani women than in Bangladeshis. High waist: hip ratios were commoner in Pakistanis and Bangladeshis than Indians, and about four times commoner in South Asians than Europeans.

Coronary heart disease was too rare for reliable comparisons among South Asian groups, but on both questionnaire and electrocardiography there was slightly less coronary heart disease in Bangladeshis than Indians and Pakistanis. South Asians had more possible myocardial infarction on questionnaire and probable coronary heart disease on electrocardiography than Europeans.

Table 8 shows that Bangladeshi men and women had the highest risk profiles for 9 of 15 coronary heart disease risk factors. Indians were advantaged by comparison. South Asians had higher levels than Europeans in 10 of 15 factors.

Table 8 Ethnic group with highest risk profile for various coronary heart disease risk factors: comparison of Indians, Pakistanis, Bangladeshis, and Europeans separately and of South Asians combined with Europeans

Discussion

Newcastle South Asians, mostly from the north of the Indian subcontinent and Sylhet, have a mix of religions, languages, and lifestyles similar to those described nationally. ^{(26) (27) (28) (29) (30) (31) (32)} Our findings that patterns of coronary risk factors are different in Indians, Pakistanis, and Bangladeshis and that South Asians combined are disadvantaged in comparison with Europeans probably apply elsewhere. The heterogeneity of South Asian populations ^{(33) (34) (35)} has rarely been taken into account in the context of coronary heart disease. ^{(2) (3) (4) (5) (6)(36) (37) (38) (39)} New and larger studies are needed to assess whether the incidence and prevalence of coronary heart disease and diabetes differs between Indians, Pakistanis, and Bangladeshis, as suggested here and elsewhere.

We acknowledge potential bias because Europeans self completed the questionnaire whereas South Asians had home interviews and Europeans and South Asians were studied sequentially. These decisions were pragmatic and resource driven; self completion of questionnaires by South Asians was inappropriate, and interviewing Europeans beyond our resources. The South Asian study needed staff with appropriate languages and cultural knowledge so screening them separately was more practical. Changes in disease and risk factors would be small between 1993-4 (European study, midpoint January 1994) and 1995-7 (South Asian study, midpoint August 1996). This paper, moreover, focuses on variations among South Asian subgroups, for whom data were collected simultaneously.

Hypotheses for the high rates of coronary heart disease in South Asians include use of ghee and other cooking oils ^(40)(41); non-vegetarian diets ⁽⁴²⁾; subclinical hypothyroidism ⁽⁴³⁾; stress, racism, and poverty ^(9)(44); deprivation in infancy and childhood ^(9)(45); and insulin resistance. ^(2)(36)(46) The insulin resistance hypothesis has overshadowed other explanations. ^{(3)(36)(46)(47)} This study draws attention to a wide range of risk factors and shows that combining data for South Asians is misleading.

Our observations emphasise the importance of poverty (among Pakistanis and Bangladeshis), smoking (among Bangladeshis, Pakistanis, and European men), high blood pressure (among Europeans and Indians), obesity (in all groups), and lack of exercise (in all groups). Our study supports infant deprivation (South Asians were shorter, an indicator of poorer early life nutrition), ⁽⁴⁵⁾ central obesity and insulin resistance (all South Asians), abstinence from alcohol (especially Pakistanis and Bangladeshis), and chronic inflammation (Pakistani and Bangladeshi men) as potentially important causes of coronary heart disease. ⁽⁴⁸⁾

Strategies to control coronary heart disease in South Asians should emphasise all important factors including social and environmental ones such as employment and poverty, ⁽⁴⁹⁾ propose linguistic and cultural adaptations, and consider the heterogeneity of Indians, Pakistanis, and Bangladeshis.

Key messages

· South Asians have more coronary heart disease than Europeans despite apparently lower levels of risk factors

· This study shows that Indians, Pakistanis and Bangladeshis differ in a wide range of coronary risk factors and combining their data is misleading

· Among South Asians, Indians were least and Bangladeshis most disadvantaged in a range of coronary risk factors. South Asians were disadvantaged in comparison with Europeans

· Future research and prevention strategies for coronary heart disease in South Asians should acknowledge a broad range of risk factors, the heterogeneity of these populations, linguistic and cultural needs, and environmental factors such as poverty
 

We thank Margaret Miller, Mavis Brown, Amanda McEwan, Heather Armstrong, Afroz Qureshi, Ayesha Motala, Kaushik Ramaiya, and Dilip Singh for help at the screening sessions; Peter Stevenson for both screening and analysis of electrocardiographs; Nan Keen for coding electrocardiographs; Denise Howel for statistical advice; David Whiting for writing the name search programme; Linda Ashworth for measuring insulin; Sheinaz Mughal for advice; and Carole Frazer for preparing the manuscript. The interviews were done by a team of 20 interviewers, whom we thank. They were Jusna Ahmed, Mushtaq Ahmed, Shuhel Ahmed, Rafiqul Alam, Masooma Ali, Showkat Ali, Santokh Bamrah, Parul Begum, Manju Chandra, Malik Chaudhry, Shubh Ghai, Amarjit Ghura, Sultana Kimti, Raminder Pal Singh, Rajindar Ghura, Salma Hasan, Nasir Iqbal, Surbhi Khanna, Pardeep Lally, Rakesh Prasad, Arati Roy, Nasim Shafiq, Baldev Singh, Salah Uddin, Anita Sarkar, and Prehlad Kanwar.

Contributors: RB contributed to the study hypotheses and design, supervision of project, planning and interpretation of data analyses, and was the lead writer. NU and MW contributed to the study hypotheses and design, supervision, screening, planning, and interpretation of data. JY participated in management of screening in South Asian study, analysis of data, and drafting methods section. LW participated in preparation and analysis of data, and drafting methods, results, and tables of manuscript. KGMMAcontributed to the study hypotheses and design and supervision of project. JH participated in development of study design, methods, and questionnaire for European study, management and screening of European population, and data coding. SP participated in screening South Asians, community liaison and recruitment to study, and data coding. NA contributed to development of South Asian questionnaire, translation, sampling, screening, recruitment, and data preparation. NU, MW, JY, LW, KGMMA, JH, SP, and NA commented on the manuscript. CT participated in screening, development of questionnaire, recruitment, and community liaison. BW managed and participated in screening for European and South Asian samples and helped with entry and preparation of data for analysis. DK participated in coordination of recruitment and data collection for the South Asian study, data analysis on response rates, and drafting text. AK participated in screening, development of questionnaire, community liaison, and recruitment. ML advised on biochemical methods, supervised laboratories doing biochemical tests, and had responsibility for lipoprotein(a) assays. AT participated in the lipoprotein(a) study and provided data and advice on lipoprotein(a). RB, GA, NU, and MW are the study guarantors.

Funding: Barclay Trust, British Diabetic Association, Newcastle Health Authority, research and development directorate of the Northern Regional Health Authority, Department of Health, and British Heart Foundation.

Competing interests: None declared.

1. Balarajan R. Ethnicity and variations in mortality from coronary heart disease. Health Trends 1996;28:45-51.
2. McKeigue P, Sevak L. Coronary heart disease in South Asian communities. London: Health Education Authority, 1994.
3. McKeigue PM. Coronary heart disease in South Asians overseas: a review. J Clin Epidemiol 1989;42:597-609.
4. Knight T, Smith Z, Lockton JA, Sahota P, Bedford A, Toop M, et al. Ethnic differences in risk markers for heart disease in Bradford and implications for preventive strategies. J Epidemiol Community Health 1993;47:89-95.
5. Hughes LO, Cruickshank JK, Wright J, Raftery EB. Disturbances of insulin in British Asian and white men surviving myocardial infarction. BMJ 1989;299:537-40.
6. Enas EA, Yusuf S, Mehta JL. Prevalence of coronary artery disease in Asian Indians. Am J Cardiol 1992;70:945-9.
7. Bhopal RS. Several key facts need to be considered. BMJ 1996;312:375.
8. Shaukat N, de Bono DP. Are Indo-origin people especially susceptible to coronary artery disease. Postgrad Med J 1994;70:315-8.
9. Williams R, Bhopal R, Hunt K. Coronary risk in a British Punjabi population: comparative profile of non-biochemical factors. Int J Epidemiol 1994;23:28-37.
10. Nazroo J. The health of Britain's ethnic minorities. London: Policy Studies Institute, 1997.
11. Harland JO, Unwin N, Bhopal RS, White M, Watson B, Laker M, et al. Low levels of cardiovascular risk factors and coronary heart disease in a UK Chinese population. J Epidemiol Community Health 1997;51:636-42.
12. Unwin N, Harland J, White M, Bhopal RS, Winocour P, Stephenson P, et al. Body mass index, waist circumference, waist-hip ratio, and glucose intolerance in Chinese and Europid adults in Newcastle, UK. J Epidemiol Community Health 1997;51:160-6.
13. Harrington B, White M, Foy C, Raybould S, Harland J. The Newcastle health and lifestyle survey 1991: health and lifestyles in Newcastle. Newcastle upon Tyne: Newcastle Health Authority, Department of Epidemiology and Public Health, University of Newcastle, 1993.
14. Coldman A, Braun T, Gallagher R. The classification of ethnic status using name information. J Epidemiol Community Health 1988;42:390-5.
15. Burrin J, Price C. Measurement of blood glucose. Ann Clin Biochem 1985;22:327-42.
16. Betteridge D, Dodson P, Durrington P, Hughes EA, Laker MF, Nicholls DP, et al. Management of hyperlipidaemia: guidelines of the British Hyperlipidaemia Association. Postgrad Med J 1993;69:359-69.
17. Petrie J, O'Brien E, Littler W, de Sweit H, Padfield PL, Dillon MJ. Recommendations on blood pressure measurement. London: BMJ Publications, 1990.
18. WHO MONICA Project. Geographical variation in the major risk factors of coronary heart disease in men and women aged 35-64 years. World Health Org Q 1988;41:115-38.
19. Flegal KM, Carroll MD, Kuuczmarski RJ, Johnson CL. Overweight and obesity in the United States: prevalence and trends, 1960-1994. Int J Obesity Relat Metab Disord 1998;22:39-47
20. Bedfont Scientific. Operator's manual for mini and micro smokerlyzers. Upchurch: Bedfont Scientific, 1993.
21. Prineas R, Crow R, Blackburn H. The Minnesota code manual of electrocardiographic findings. London: John Wright, 1982.
22. Rose G, Blackburn A, Gillum R, Prineas R. Cardiovascular survey methods. 2nd ed. Geneva: World Health Organisation, 1982.
23. SPSS. Statistical package for social sciences (SPSS/PC+) version 6.0. Chicago: SPSS, 1996.
24. Coombs M, Raybould S, Long CA. Index of deprivation. London: Department of Environment, 1992.
25. Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Oxford: Blackwell Science, 1994:436.
26. Coleman D, Salt J, eds.. Demographic characteristics of the ethnic minority populations. Ethnicity in the 1991 census. London: HMSO, 1996.
27. Ratcliffe P, ed. Social geography and ethnicity in Britain: geographical spread, spatial concentration and internal migration. Ethnicity in the 1991 census. London: HMSO, 1996.
28. Peach C, ed. The ethnic minority populations of Great Britain. Ethnicity in the 1991 census. London: HMSO, 1996.
29. Karn V, ed. Employment, education and housing among the ethnic minority populations of Britain. Ethnicity in the 1991 census. London: HMSO, 1996.
30. Church J, Summerfield C. Social focus on ethnic minorities. London: Office of National Statistics, 1996.
31. Rudat K. Black and minority ethnic groups in England. London: Health Education Authority,1994.
32. Modood T, Berthoud R, eds. Ethnic minorities in Britain. London, Policy Studies Institute, 1997.
33. Shaunak S, Lakhani SR, Abraham R, Maxwell JD. Differences among Asian patients. BMJ 1986;293:1169.
34. Bhopal RS, Phillimore P, Kohli HS. Inappropriate use of the term "Asian": an obstacle to ethnicity and health research. J Pub Health Med 1991;13:244-6.
35. Ramaiya KL, Swai, ABM, McLarty DG, Bhopal RS, Alberti K. Prevalences of diabetes and cardiovascular disease risk factors in Hindu Indian subcommunities in Tanzania. BMJ 1991;303:271-303.
36. McKeigue PM, Shah B, Marmot MG. Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 1991;337:382-6.
37. McKeigue PM, Marmot MG, Adelstein AM. Diet and risk factors for coronary heart disease in Asians in Northwest London. Lancet 1985;ii:1086-90.
38. McKeigue PM. Mortality from coronary heart disease in Asian communities in London. BMJ 1988;297:903-7.
39. Marmot MG, Adelstein AM, Bulusu L. Immigrant mortality in England and Wales 1970-1978. Causes of death by country. Studies on medical and population subjects. London: HMSO, 1984.
40. Jacobson MS. Cholesterol oxides in Indian ghee: possible cause of unexplained high risk of atherosclerosis in Indian immigrant populations. Lancet 1987;ii:656-8.
41. Nath BS, Murthy R. Cholesterol in Indian Ghee. Lancet 1988;ii:39.
42. Pais P, Pogue J, Gerstein H, Zachariah E, Savitha D, Jayprakash S, et al. Risk factors for acute myocardial infarction in Indians: a case-control study. Lancet 1996;348:358-63.
43. Fowler PBS. Diet and risk factors for coronary heart disease in Asians in north west London. Lancet 1985;ii:1363.
44. Williams R, Bhopal R, Hunt K. Health of a Punjabi ethnic minority in Glasgow: a comparison with the general population. J Epidemiol Community Health 1993;47:96-102.
45. Barker DJP. Mothers, babies and disease in later life. London: BMJ Publishing, 1994.
46. Knight TM, Smith Z, Wittles A, Sahota P, Lockton JA, Hogg G, et al. Insulin resistance, diabetes, and risk markers for ischaemic heart disease in Asian men and non-Asian men in Bradford. Br H J 1992;67:343-50.
47. Gupta S, de Belder A, Hughes LO. Avoiding premature coronary deaths in Asians in Britain. BMJ 1995;311:1035-6.
48. Mendall M. Inflammatory responses and coronary heart disease. BMJ 1998;316:953-4.
49. Acheson D. Independent inquiry into inequalities in health. London: HMSO, 1998.