Asian inmigrants to England and Wales have high mortality from coronary heart disease (McKeigue & Marmot, 1985) which is higher than the England and Wales average (Marmot, 1984a, b). Evidence shows that Asian inmigrants suffer more severe coronary heart disease than Asians who remain in their own countries (Tunstall-Pedoe, 1975; Balarajan, 1991; Miall, 1972a, b, c) The high incidence of CHD is not explained by known risk factors (blood pressure, serum lipids, smoking) to determine CHD susceptibility (Lowry, 1983). It has been suggested that the process of migration and associated psychosocial difficulties in cultural adaptation may be implicated in the mental and physical health of migrants, and that the form of CHD may indeed vary in different ethnic populations (Fox, 1988). In addition, the explanation of high CHD in overseas Asian populations who differ in religious, cultural, geographic, and genetic backgrounds, must include (yet to be determined) some factor common to them (McKeigue, 1988), with CHD in Asian migrants tending to occur in more severe forms (Lowry, 1983).
1. Who are the Asian populations?
The term ‘Asian’ is one of convenience that requires accurate definition (Shaunak, 1986). It is a mistake to ascribe ‘immigrant’ to all people of Asian origin or descent living in the UK. It is more correct to say about “…one million people of Indian subcontinent origin live in Britain.” (Kholi, 1988). There are four main ethnic groups of Indian subcontinental origin living in Britain – only some of whom are in-migrants – Gujuratis, Punjabis, Southerners, and Muslims (Balarajan, 1984).
Immigration from the Indian subcontinent occurred in two main periods – that prior to 1954 and that during the 1960’s. The earlier period comprised mainly ex-colonial British subjects. In 1951 it was estimated that of 121,884 residents in England and Wales born in India and Pakistan, 86,000 (71%) were ‘white’. Subsequent migration was largely of Asians. In 1966, Indian and Pakistan born population in England and Wales had increased to an estimated 305,340 – of this total the estimated `white’ population had declined to 73,000 or 24% (Marmot, 1984a).
`Asian’ populations in the UK show a regional distribution both in numbers and ethnic groupings. Bangladeshi populations (who prefer to call themselves Bengali) come mostly from the Sylhet district and are settled mainly in East London. Pakistanis originate (mainly) from Mirpur in Free Kashmir north of Punjab, most are Muslim and speak Mirpuri (a dialect of Bengali) and Urdu. Most live in Yorkshire, Greater Manchester, Lancashire, the West Midlands, Glasgow and Cardiff. All areas of the UK with high CHD mortality rates. Most Indians arrived during the late 1950’s and early 1960’s from two main areas – Punjab and Gujurat. The Punjabis are mainly Sikhs whereas Gujuratis are Hindu. Large Punjabi and Gujurati communities live in Leeds, West London, West Midlands and Glasgow. People of Gujurati origin also migrated via East Africa and live mainly in North and South London, Leicester, Coventry and Greater Manchester. About one fifth of ‘Asian’ origin populations travelled from Kenya, Tanzania, Uganda, Malawi, and Zambia. In addition, a small number of people from Southern India are present (Tamiland, Kerala) but who have no common language with people of northern Indian origin.
2. Coronary heart disease in mainland India
Studies in India (Malhotra, 1967a, b, c) on 1.5 million railway employees found an occupational group/social class gradient in CHD with a spatial distribution of incidence. Highest mortality was found in southern India (135/100,000) compared to northern India (20/100,000). It is noteworthy that southern Indians living in the UK also show a high incidence of CHD despite their small numbers – myocardial infarction is reported commonest in southern India (Balarajan, 1984). CHD was reported highest in higher socioeconomic groups (Malhotra, 1967a) and nearly seven times more common in the lower group (Class IV). Moreover, it was also apparent that CHD mortality was lower in less active sedentary clerks than more active physical occupations of fitters and sweepers (Indian Railway employees). This contrasts with the opposite situation in so-called ‘western’ affluent or industrialised countries of today but not with early 20th century class distributions for CHD. (Baldry, 1971; Susser, 1975, 1985; Doll, 1987; Marmot, 1989b). Dietary differences exist between Indian socioeconomic groups which may explain the social gradient. Higher income groups have a higher intake of refined carbohydrates, and also a higher intake of animal fats. Furthermore, lower CHD populations of the north had 19 times greater intake of fat than higher CHD groups of the south. An example where CHD incidence is at variance with risk factor distribution for CHD amongst Indian origin populations. The view being that “…environmental factors are important in the causation of this disease.” (Malhotra, 1967a).
A survey on dietary lipids and atherosclerosis in Dehli (Padmavati, 1959), showed no cases of CHD in low income groups. The higher social income group had a higher intake of fat coupled with a higher rate of CHD. This 1959 study varies with the 1967 survey and may indicate developments within Indian populations for CHD susceptibility. Serum cholesterol levels were significantly higher in higher income groups. Indian results were compared internationally with other surveys. Indian low socioeconomic group serum cholesterol levels resembled those found for similar strata in Spain, Italy, South Africa. Similar TC values were equated for higher income groups in the same countries. But – the dietary fat intake of Indian higher groups was the same as for Cape Europeans, Slough (England), industrial workers, but lower for similar groups in Boston, Minnesota, Swedish firemen, Madrid professionals.
CHD appeared at a lower TC level in India compared to the optimum for UK and USA populations (Padmavati, 1959). CHD appeared at 4% in the higher income, higher serum cholesterol males, whereas it occurred at 3% in industrial worker Indians with lower TC values (UK levels of similar range are non-symptomatic or sub-clinical). It would appear that the possible threshold effect of serum lipids for atherosclerosis may be lower in Indian populations. Considering ‘Asian’ dietary practices in the UK (low fat/high fibre) it seems that there is a variance for lipids as a CHD risk variable – but not if the pathological process of atheroma is initiated at a lower threshold level in Indians.
CHD has shown great antiquity (Brothwell, 1967) and non-specified heart disease was known in ancient India in the Vedic scripts of the Hindus (Sigerist, 1961; Zimmer, 1948; Jayne, 1925). Other Indian studies (Bhargava, 1966) have shown CHD comprised 13.57% of circulatory disease admissions to hospital. Again, the CHd incidence began early – highest in men of 41-60 years group. The clinical pattern showed 12.48% of cases were angina, but in common with ‘Asian’ CHD patterns overseas AMI was more prevalent at 53.86% (505 patients). The obvious severity of CHD in Indians is demonstrated by the excess of infarction over angina. In India (Bhargava, 1966) nearly 16% of CHD occurred in age group 21-40, with a peak incidence for both sexes in age decile 40-50. Similar peak incidences have been found in other Indian studies, e.g., 41-50 years (Vakil, 1962, 1949), 51-60 years (Mathur, 1960; Malhotra, 1958), and 41-60 (Bhargava, 1966). The commonest pattern was for acute myocardial infarction (Mathur, 1963; Vytilingham, 1964).
Indian CHD was related to the ABO blood group system and compared to a similar sample in Edinburgh (Bhargava, 1966) with 54% of Indian men (with angina) being group 0 (A being 39%), myocardial infarction men gave 50% group 0 (37% for A). The Edinburgh sample showed no significant deviation from the Indian data. An Indian study (9257 healthy Hindus and Muslims and 198 IHD cases) found blood group A more susceptible to CHD (Srivastava, 1966) which confirmed earlier reports of excess A over B and 0 in 81 cases of AMI (Gertler, 1954) in the USA, and A and B over 0 in South Africa (Bronte-Stewart, 1962). In India (Srivastava, 1966) “…only the group A excess and 0-deficiency was to be significant (value 0.5).” whereas “No statistically significant difference in distribution could be found between Hindus and Muslims or between two sexes.” Analysis of 325 healthy Indians in Singapore (Banerjee, 1969) detected no significant association between serum cholesterol and ABO blood groups but found highest mean TC in CHD patients with group A. In Australia (Denborough, 1969) an excess of A in AMI patients was found who, interestingly, were mostly British, some Jewish, others East European in origin. Framingham (Havlik, 1969) found in men aged 39-72 a lower incidence of non-fatal CHD in 0 compared to A (most effect for age 50-59) with the trend reversed in women – most CHD in group O. In Scotland (Oliver, 1962; 1969) group A men had higher mean TC compared to groups 0 and B with a “…deficit of group 0 relative to A in patients who survived myocardial infarction but not in those with angina.” (Oliver, 1969). In the British 24 towns study (Whincup, 1980) towns with higher population incidence of 0 had higher incidence of AMI but in individuals IHD had a higher incidence amongst group A’s.
High levels of haemoglobin (Hb) as a ‘risk factor’ for AMI has been suggested (Bottiger, 1972) but criticised on the basis that the paradigm Hb + CHD and CHD + lipids equates with CHD + Hb is not statistically valid (Miller, 1972). Primary hyperlipoproteinaemia Type II associated with 56% group A subjects (Polychronopolou, 1974) suggesting ABO groups correlated with lipoprotein profiles rather than TC level. Higher TC has been claimed for group A (Kark, 1984) with TC slightly higher in A’s in the British Regional Heart Study’s 24 towns (Whincup, 1980). It was considered that the association between the ABO locus and serum lipid profile was at the beta-lipoprotein fraction (Fox, 1986). In addition, group A is associated with slightly higher mean SBP and DBP (Kark, 1984), diastolic BP consistently associated with higher levels in 0 compared to lower levels among B group (Fox, 1986), with increased hypertension among group A (Kark, 1984). Moreover, physical fitness and HDL levels
show an inverse relationship to white blood cells (WBC) with WBC raised in smokers and elevated WBC being higher in CHD (Friedman, 1990). Other research (Hansen, 1990) showed elevated leucocytes were independently associated with other CHD risk factors (including smoking) with strong inverse associations for HDL, positive for TG and positive for LDL.
3. Coronary Heart Disease in the UK in ‘Asian’ Populations.
As with Indian subcontinent populations ‘Asian’ in-migrant and their descendants in the UK show considerable ‘premature’ and ‘severe’ forms of CHD. Similar patterns are seen for Indian descent populations who migrated to other areas – Caribbean, Malaya, Africa, Singapore. Explanations vary from environmental causation theories to whether or not Indian populations are genetically susceptible to premature CHD and early atherosclerosis. One key may be the higher incidence of diabetes mellitus (with its concomitant macrovascular and microvascular affects) amongst Indians.
In-migrants from the Indian subcontinent have a high mortality from CHD, hypertensiv@rand cerebrovascular disease but ‘white’ in-migrants had the same magnitude of CHD incidence as Indian ethnic groups (Marmot, 1984a). Mortality due to CHD was high in all ethnic groups from the Indian subcontinent but highest in Moslems (Balarajan, 1984). More male Punjabis died from cerebrovascular disease (CVD) with Gujuratis more affected by diabetes.
There are about 1.3 million people’of Indian origin in Britain of whom 40% are aged 40 or more (OPCS, 1983; 1986b). Also, Indians show a great cultural and geographical diversity (Lancet, 1987) and in Britain the Asian communities differ as similarly as do European populations (McKeigue, 1988). Europeans show different rates of CHD incidence. High rates of CHD have been shown by Gujuratis, Punjabis, Southern Indians, and especially Moslems. This suggests a common factor is shared by these diverse communities (McKeigue, 1988), that the problem is not related to the country of immigration but to shared genetic and environmental factors (Fox, 1988). This does exclude, however, the possibility that migration and the discontinities it entails is not an additive factor in disease susceptible groups.
Bangladeshis in Tower Hamlets (Tunstall-Pedoe, 1975) were found to experience higher than average myocardial infarction rates. This ethnic group had high cigarette consumption and high dietary fat intake (57% of energy) which was twice the national average. In essence this community was exhibiting a high risk profile for CHD in two major areas which may explain the 30% higher rate of infarction compared with other residents of the borough (Silman, 1985a, b). An investigation of a Gujurati community in Brent and Harrow, North West London (McKeigue, 1985) found significant excess hospital admissions for myocardial infarction. Further studies analysed CHD data from Asian communities in London (McKeigue, 1988) and found CHD mortality among Asians in England and Wales had increased by about 25% since 1970 to 1972. Some 77% of the Asian population in Brent and Harrow (McKeigue, 1985) were Gujuratis, were comparatively affluent and imainly vegetarian had very low smoking rates among women. The opposite applied to Moslem communities of Waltham Forest and Tower Hamlets. It is striking therefore that Asian men and women in each borough should share a common CHD mortality 50% higher than the national average (McKeigue, 1988).
A Leicester survey (Donaldson, 1983) found 2.2 times as many hospital admissions for AMI among Asians – results of country of birth analysis of Asians in Leicester gave UK at 45%, 31% India, Pakistan or Bangladesh, 19% East Africa, 5% other countries. In Birmingham (Lowry, 1983) a survey detected more severe CHD among Asians referred for investigation, though the pattern of distribution for ‘Asians’ and `whites’ was similar. CHD among Asians in Birmingham was severe even though their cholesterol levels, body weight and cigarette consumption were lower than comparable ‘whites’. In Birmingham the Asian population comes from Pakistan, Punjab, with a few Bangladeshis, with a mean residence in the UK of 13 years (Lowry, 1983). AMI amongst Asians and `whites’ in Birmingham (Lawrence, 1985) found that the incidence of ventricular arrythmias was greater in Asians indicating increased risk of fatal AMI and sudden cardiac death – nonetheless, incidence and complications of AMI were similar in matched ‘whites’.
Studies of hypertension in Asians, whites and blacks in Birmingham (Cruickshank, 1983) screened factory workers for blood pressure levels. In contrast to USA findings, black West Indian origin males, Asians, and local whites did not differ for mean SBP and DBP. Asian men did have a slightly DBP but Birmingham blacks did not show the higher mean BP’s shown by blacks longer resident in the USA. In Southall, Punjabi women showed an increase in BP with increasing age of clinical significance (Keil, 1980). A study of BP among Bengalis in East London showed their increased incidence of AMI was unlikely related to BP (Silman, 1987). The Bengalis had lower mean SEP and DBP – but both local whites and the Bengalis showed incremental BP elevations with advancing age and that migration did not explain the increase. Bengali BP increases were more likely due to body weight increase, and suggests that both local whites and Bengalis share a common risk factor for increasing BP and CHD.
A survey relating deaths from stroke to past maternal mortality (Barker, 1987a) plus a study correlating infant mortality, childhood nutrition, and ischaemic heart disease (Barker, 1986) in England and Wales raised some pertinent issues, and numerous correspondences have discussed the matter (Fox and Shapiro, 1988; Head, 1988). A link was suggested (Head, 1988) between early environment and risk of atheromatous disease in later life – UK prevalence of CHD more common among patients whose early life was spent in areas where poverty was more common. Thus experiencing adverse social effects in childhood increases susceptibility to IHD – in-migrants from the Indian subcontinent coming from an area of known poverty. Dietary implications suggested (Goldberg, 1986) that Asians had a low intake of omega-polyunsaturated fatty acids and thus increased risk of CHD.
4. Coronary Heart Disease in Asian Migrants outside the UK
Indians in South Africa show a mortality rate for CHD similar to South Africans of European origin (Walker, 1963) and that atheroma is very prevalent (Tejeda, 1968). In Durban both large and small vessel disease was found to be very common amongst Indian diabetics – commonly in heart, brain, and legs. Similarly, Durban Indians had atherosclerosis comparable to New Orleans whites (West, 1978). In East Africa the CHD rate for Indians (and diabetes) is higher than for Africans. In relation to diabetes and CHD it is considered that impaired glucose tolerance and IHD were connected, appearing to be a “…characteristic feature of ischaemic heart disease among Indians.” (Miall, 1972). In Kampala, Indians have high levels of serum cholesterol, high fat diets and high rates of atherosclerosis (West, 1978), but this contrasts with lifestyle factors for Indians in the UK.
Surveys in South Africa amongst diabetics show high CHD rates in both Indians and whites, but peripheral vascular disease (a common complication of diabetes) is less common in Indians (Jackson, 1972). In Durban Indian diabetics there are high serum cholesterol levels and also for triglyceride values (Asmal, 1975) – both regarded independent risk variables for CHD.
In Guyana CHD is common in Indian diabetics but who show a lower rate compared to European and USA whites, but higher than the local black population (Weinstein, 1962). In Fiji, hypercholesterolaemia was a common finding in Indians with AMI – the majority values above 7.75 mmo1/1 or 300 mg/dl (KBakami, 1975) compared to mean values of only 5.17 mmo1/1 (200 mg/dl) found in Indian higher socioeconomic groups in Dehli (Padmavati, 1959). Incidence of CHD in Fiji Indians is some 30x higher than the indigenous population. In Fiji, 28% of Indian cases of AMI are known diabetics. Similar high Indian rates for diabetes and concomitant CHD have been found in Fiji (Cassidy, 1967), Singapore (Cheah, 1975), Trinidad and Malaya, whereas in India Bengali Hindus have more diabetes than Moslems. One might expect therefore a similar pattern in the UK – in Southall and East London diabetes in Asians was 3.8 times higher than in Europeans (Mather, 1985) with similar findings appearing in Leicester (Samanta, 1986, 1987).
Large numbers of migrants of Indian origin have settled in the Caribbean islands. Indian men in Trinidad appeared to be at increased risk for CHD than male counterparts in India (Miller, 1984), because LDL values in Trinidad are high and not accompanied by protective levels of HDL. LDL levels were found to be higher in Trinidad, intermediate in North India, and lowest in Madras (Southern India). Triglycerides were high in Trinidad, intermediate in South India and lowest in North India. HDL levels are associated with levels of apolipoprotein AI and low ApoAI levels in Trinidad Indians may suggest a population difference between them and Indian subcontinent populations. Apolipoprotein B is the major protein component of LDL and evidence is suggestive of a variable distribution for the ApoB gene between Trinidad and subcontinent Indians. ApoAI showed a mean of 1.28 mg/dl for Indians compared to a mean of 1.41 mg/dl in African origin populations – suggesting a possibly genetically determined higher level of protective HDL in Africans (who have lower rates of CHD).
Another survey in Trinidad (Miller, 1982) found that Indians had a relative risk of CHD of 3/1 compared to other ethnic groups. Indians had much lower HDL cholesterol levels and significantly higher LDL levels as well as elevated serum VLDL concentrations (also a diabetic feature). HDL was 14% lower in Indians than Africans and 9% lower than that for other groups. Serum LDL was 9% higher than in Africans, 5% higher than others. This indicates that non-Indian, non-African populations occupy an intermediate position for CHD risk. Yet, Trinidadian Indians come mainly from North India where CHD rates are lower. It is suggested that CHD in associated with a broad disorder of lipoprotein metabolism rather than a specific lipoprotein abnormality (Miller, 1982). Diabetes is more common in Trinidad Indians than others (Poon-King, 1968). The ethnic differences in Trinidad for CHD could not be explained by serum lipids, BP, smoking habits, or fasting glucose (Heckles, 1980; 1986) therefore diabetes (19% in Indians) explained the prevalence of CHD, but again, HDL concentrations were also low in Indians. In Trinidad populations of mixed descent showed low all-cause and cardiovascular mortality, and it was thus suggested that Trinidad Indian CHD susceptibility (Miller, 1988) arose out of an underlying disposition coupled to diabetes prevalence.
5. Migration and CHD in Indian Populations
It has been stressed that “…the process of migration and the psychosocial difficulties immigrants have in adapting to cultural change have been implicated in the physical and mental health of migrants.” (Fox, 1988). Mechanisms of physiological stress are therefore involved. However, Indians exhibit diffuse distal coronary artery disease (more indicative of diabetes) which may not be related to a specific ethnic pattern. Indians also appear more vulnerable to severe atherosclerosis and myocardial infarction – more common in proximal coronary artery disease and seen in AMI and SCD. The disease susceptibility alleles model (McKeigue & Marmot, 1988) provides an explanation of Asian CHD. Asian CHD also fits with a migrant mortality increase model of CHD susceptibility operating within a framework of spatial mobility, with `transitional inconsistencies’ and ‘extranormal incongruities’.
Migrants may take their disease susceptibilities with them but such diseases and underlying mechanisms (like migrants) are not static. CHD in migrants demonstrates changes in form and severity. CHD is the result of (and is itself) a dynamic process – just as population movement. Indians have been shown to have high rates of CHD wherever they migrate to, but equally evident are the variations in prevalence and severity of CHD within disparate Indian migrant groups, e.g., London, Birmingham, Trinidad, Africa, Singapore, South East Asia, Malaya, Fiji, as well as North and South Indian populations.
Migrational stress coupled with problems of adaptation may make a susceptible population differentially vulnerable to CHD in whatever circumstances they find themselves. In this respect psychosocial transitions consequent upon geographical and spatial movement resemble those found operating within the framework of social mobility. It may appear that genetic mechanisms determine Indian CHD rates. This is too simple a rationale because environmental factors play a crucial role. The fundamental processes of CHD pathology are common to all ethnic groups – only prevalence, incidence and severity vary. This indicates that a common pathology underlies human variability for susceptibility to CHD. The biological substrate operates within a matrix of cultural, social and environmental factorsfand it is this dynamic that creates the within and between population variability for prevalence and risk for clinically overt CHD. In this respect spatial mobility can be seen as an added factor in the overall vulnerability quotient for CHD.
Originally Appendix 178 of my MPhil entitled Population Variation for Risk Factors in Ischaemic Heart Disease. CNAA at Oxford Polytechnic and Oxford Brookes University, 1987-1992.
References to follow.