Coronary heart disease incidence among non-Western immigrants compared to Danish-born people: effect of country of birth, migrant status, and income

Abstract

Background

Increasing global migration has made immigrants’ health an important topic worldwide. We examined the effect of country of birth, migrant status (refugee/family-reunified) and income on coronary heart disease (CHD) incidence.

Design

This was a historical prospective register-based cohort study.

Methods

The study cohort consisted of immigrants above 18 years from non-Western countries who had obtained a residence permit in Denmark as a refugee (n = 29,045) or as a family-reunified immigrant (n = 28,435) from 1 January 1993–31 December 1999 and a Danish-born reference population (n = 229,918). First-time CHD incidence was identified from 1 January 1993–31 December 2007. Incidence ratios for 11 immigrant groups were estimated using Cox regression analysis.

Results

Immigrants from Afghanistan, Iraq, Turkey, Eastern Europe and Central Asia, South Asia, the Former Yugoslavia, and the Middle East and North Africa had significantly higher incidences of CHD (hazard ratio (HR) = 1.36; 95% confidence interval (CI): 1.05–1.75 to HR = 2.86; 95% CI: 2.01–4.08) compared with Danish-born people. Immigrants from Somalia, South and Middle America, Sub-Saharan Africa and women from East Asia and the Pacific did not differ significantly from Danish-born people, whereas immigrant men from East Asia and the Pacific had a significantly lower incidence (HR = 0.32; 95% CI: 0.17–0.62). When also including migrant status, the higher incidences were reduced. Refugee men (HR = 1.35; 95% CI: 1.11–1.65) and women (HR = 1.33; 95% CI: 1.08–1.65) had a significantly higher incidence of CHD than family-reunified immigrants. When migrant status and income were included simultaneously, the incidences decreased to an insignificant level for most immigrant groups.

Conclusions

Most non-Western immigrant groups had a higher incidence of CHD than Danish-born people. The study revealed that migrant status and income are important underlying mechanisms of the effect of country of birth on CHD.

Introduction

Increasing global migration has made immigrants’ health an important topic worldwide. In Denmark, the number of immigrants has increased, especially during the past 30 years, and by 1 January 2013 the number of first generation non-Western immigrants was 266,666, corresponding to 5.3% of the total Danish population.¹

Coronary heart disease (CHD) is one of the leading causes of hospitalisation and death in Denmark^2,3 and worldwide.⁴ Several studies document that immigrants have a higher risk of CHD than local-born people in Western countries.^5–11 This raises concerns about inequity in health and about health care costs. Moreover, with an increasing average age among immigrants, the scope of these problems can be expected to grow. The immigrant population is heterogeneous regarding health, health behaviour, social structures, and cultures. In Denmark, evidence about CHD incidence is lacking for most immigrant groups.

Existing epidemiologic research on immigrants’ health has focused on ethnicity, often measured by country of birth, and has mainly explained variability in health outcomes with differences in culture, language, tradition, religion, biology and genes. However, the health of immigrants can also be influenced by their migration process, including exposures before, during, and after the migration.¹² Recently, it has been suggested that the legal ground for immigrants’ residence permit (migrant status e.g. refugee/family-reunified) can be used as a proxy measure of differences in the migration process, because certain exposures tend to accumulate within one or the other group.^13–16 Refugees may experience a more burdensome migration because their migration process is often initiated by factors driving the individual out of the country of origin whereas family-reunified migrants are likely to be attracted towards the recipient country. Including migrant status together with country of birth in epidemiologic studies has the potential to show the health consequences of the migration process independently of ethnic background.

Low socioeconomic status (SES) is a well-established risk factor for CHD and, in Denmark, immigrants generally have a lower SES than Danish-born people.¹ Therefore, SES is an important factor to consider.¹⁷ In line with leading researchers in this field, this study considers SES as a mediator.^6,13,18

The aim of this study was to investigate the incidence of CHD among 11 groups of non-Western immigrants compared with Danish-born people and to examine the effect of migrant status and income on CHD.

Materials and methods

Study population

This study is a historical prospective register-based cohort study. The open cohort included all immigrants above 18 years of age who obtained a residence permit in Denmark as a refugee or family-reunified immigrant from 1 January 1993–31 December 1999 (n = 62.461), and a reference population of Danish-born residents with Danish-born parents (n = 249,839) matched individually 1:4 on age and sex. Matching was not necessary for this study, but performed as the cohort was established for another study. Immigrants and their comparisons were followed from the commencement of residence permission of the immigrant until one of the following four events: (a) date of first CHD incident, (b) date of death by causes other than CHD, (c) date of emigration, or (d) study closure (31 December 2007). Immigrants were defined as people born abroad and not having Danish-born parents. For each person, data on country of birth, date of residence permit, migrant status (refugee/family-reunified), and personal identification number were available. For further information about the cohort see Norredam et al.¹⁶ We excluded immigrants from Western countries (n = 4826) and their comparisons (n = 19,301). Due to incomplete registration of cause of death, 155 immigrants and 620 Danish-born people were excluded. The final study population comprised 29,045 refugees, 28,435 family-reunified and 229,918 Danish-born comparisons.

Data collection and variables

Outcome variable

The outcome was CHD incidence identified between 1 January 1993 and 31 December 2007. Incidents were defined as first-time hospitalisation with a CHD diagnosis or out-of-hospital death by CHD given no previous hospitalisation by CHD. The personal identification number of individuals in the study cohort was linked to the Danish National Patient Registry (DNPR), the National Causes of Death Registry (NCDR) and the Registry for Population Statistics (RPS). We obtained information on hospital contacts from DNPR, out-of-hospital death from NCDR, and missing dates of death in NCDR from RPS. Hospitalisation due to CHD was based on the patient’s diagnosis upon discharge including both primary and secondary diagnosis. Underlying cause of death of CHD was used to identify out-of-hospital deaths. This approach has been validated¹⁹ and used in other register-based studies.^5,6,8,9 Hospitalisation and death were registered according to the 8^th (until 31 December 1993) and 10^th revision (1 January 1994 and onwards) of the World Health Organisation (WHO) International Classification of Diseases (ICD). The included ICD-8 codes were: 410–414, and ICD-10 codes were: DI20–DI25.

Explanatory variables

Country of birth was used as an indicator of ethnicity. The three largest groups of immigrants (Iraq, Somalia and Turkey) were analysed separately, as were immigrants from Afghanistan because of an increased migration from Afghanistan since 2000. The remaining immigrants were grouped by geographical region according to the World Bank categorisation of countries.²⁰ Individuals from Afghanistan, Iraq, Somalia, and Turkey were not included in their respective regions. Migrant status was classified as either refugee or family-reunified based on the background of residence permit. Personal average income was used as an indicator for SES, and was obtained from the Integrated Database for Labour Market Research. It was calculated as an average income beginning one year after entry in the study until the year before onset of CHD, emigration, or death by other causes than CHD.

Statistical analysis

Analysis of variance (ANOVA) was used to investigate whether mean age at entry into the cohort and mean personal annual income differed between Danish-born people and each immigrant group and t-test to compare all immigrants to Danish-born people for men and women respectively. Cox proportional hazards regression models were used to analyse the time from entry to a CHD incident. The models estimated incidence ratios of CHD for immigrants relative to Danish-born people. All analyses were stratified by gender, and adjusted for age, by using age (continuous) as the time-dependent variable. Delayed entry was used because individuals entered the study at different ages.

We estimated age-adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for the 11 groups of non-Western immigrants comparing them to the Danish-born people, and then explored the plausible confounding effect of migration status and mediating effect of SES. The analysis that included income was restricted to individuals between 30 and 64 years, because income is an invalid indicator of SES in younger and older ages. This reduced the population by 148,341 individuals. In addition 2327 persons were excluded due to missing information on income. Income was included as a piecewise linear function to properly account for the influence of income over the whole income span. The plausible interaction term between migration status and income was tested. Data was analysed using SAS statistical software version 9.2 (SAS Institute Inc., Cary, NC, USA). The study was approved by the Danish Data Protection Agency. Further ethical approval regarding registry-based research is not required in Denmark.²¹

Results

For both men and women, most refugees were from the Former Yugoslavia, while most family-reunified men were from Turkey and most family-reunified women were from East Asia and the Pacific (Table 1). Mean age at entry for the immigrant groups was between 25 and 38 years. All immigrant groups had a significantly lower annual income than Danish-born people. Mean follow-up times for immigrant groups were between 8.0 and 10.4 years, and incidence rates varied between 63 and 622 CHD incidents per 10⁵ person years. The incidence ratios for all immigrants compared to the age-matched Danish-born population were 1.90 for men and 1.45 for women.

CHD incidence

In the model only adjusted for age (see Model 1 in Table 2) immigrant men and women from Afghanistan, Iraq, Turkey, Eastern Europe and Central Asia, South Asia, the Former Yugoslavia, as well as the Middle East and North Africa, had a significantly higher incidence of CHD than Danish-born people, with HRs ranging from 1.36 (95% CI: 1.05–1.75) to 2.86 (95% CI: 2.01–4.08). Immigrant men and women from Somalia, South and Middle America and Sub-Saharan Africa did not differ significantly from Danish-born people. Immigrant men from East Asia and the Pacific had a significantly lower incidence than Danish-born men with a HR of 0.32 (95% CI: 0.17–0.62). A decreased incidence was also observed in immigrant women from East Asia and the Pacific, though not significant. In the model also adjusted for migrant status (see Model 2 in Table 2), refugee men (HR = 1.35; 95% CI: 1.11–1.65) and women (HR = 1.33; 95% CI: 1.08–1.65) had a significantly higher incidence of CHD than family-reunified immigrants. For immigrant men from Afghanistan and the Former Yugoslavia the difference decreased to an insignificant level after inclusion of migrant status.

When including personal annual income, we found a significant interaction with migrant status (Table 3). Among men, the differences in incidence for country of birth became insignificant in six out of seven groups when including the interaction term between migrant status and SES and in five out of seven groups among women.

Discussion

We found that most non-Western immigrant groups in Denmark have a higher incidence of CHD than Danish-born people (seven out of 11 groups). Interestingly, both refugee men and women have a significantly higher incidence of CHD than family-reunified immigrants, when analysing country of birth and migration status simultaneously. Finally, we found that the effect of SES mediates some of the country of birth effect on CHD, and this effect is modified by migrant status.

Our study revealed a large heterogeneity in CHD incidence among immigrants. The finding of a higher incidence among refugees compared to family-reunified immigrants within the same country of birth indicates that these two groups have different risk patterns even when originating from the same country. This suggest that differences in CHD incidence between immigrant groups are not all due to ethnic characteristics such as culture and genes, but that migration processes are also influential.

The higher incidence among immigrant groups became insignificant for most groups when income and migration status were included simultaneously. This does not mean that there is no effect of country of birth on CHD. Rather, the interpretation should be that the effect of country of birth is mediated by SES, as factors such as language and discrimination can lead to low SES, which again leads to a high incidence of CHD.²²

Other Scandinavian studies of either CHD or acute myocardial infarction (AMI) incidence have found higher incidence among immigrants from South Asia, the Middle East, Northern Africa, Turkey, Poland and, to some extent, also immigrants from other Eastern European countries^5,6,8,11 than the local born people, which is consistent with our results. Studies show inconsistent results regarding immigrants from the Former Yugoslavia, the Baltics and Russia.^5,6,8,11 Immigrants from Latin America seem to have a lower incidence, which might also be the case for immigrants from South East Asia and sub-Saharan Africa.^8,11

Few previous epidemiological studies have examined the effect of migrant status, though one found that refugees have a higher all-cause and cardiovascular disease mortality than family-reunified immigrants.¹⁵ Regarding SES, other studies have also found a mediating effect on CHD and AMI.^6,11

The explanation for the differences in the health of immigrants is a result of a lifetime of accumulated risks factors.^12,23 Risk factors include, for example, environmental exposures, socioeconomic deprivation, discrimination, health beliefs, health behaviour, genetics, access to health care, life stress and lack of social support.^22,23 Traditionally, it is hypothesised that risk of lifestyle-related diseases mainly develops among non-Western immigrants after migration.^12,24 However, a recent study showed that, after five years of arrival to Denmark, CHD risk was higher among immigrants than among Danish-born people,²⁴ indicating that immigrants had already accumulated a considerable risk before arrival in Denmark.

The few studies on CHD risk factors among immigrant groups document differences in the distribution of cardiac risk factors.^25–29 The results generally support our findings. For example, the studies find high rates of hypertension, dyslipidaemia, diabetes and low rates of physical activity in south Asian immigrants,^25,28,29 high smoking rates and rates of hypertension and obesity among Turkish immigrants,^26–28 high smoking rates, high blood pressure, and overweight among eastern European immigrants,^27,29 and finally, high levels of high-density lipoprotein cholesterol and low hypertension rates in east Asian immigrants.²⁸ Studies on risk accumulated before migration are needed to enhance the understanding of the differences in the CHD incidence in migrant populations.

To our knowledge, no studies have investigated the relative distribution of CHD risk factors between refugees and family-reunified immigrants, thus it is necessary to look into general differences in migration processes between the two groups to identify explanations to the observed differences. Refugees are more likely than family-reunified immigrants to have experienced war, torture, loss of family and friends, stays in refugee camps, long asylum processes and long periods with poor access to health care. One study has documented high rates of traumatisation and mental disease among refugees,³⁰ which might affect cardiac health directly or affect personal resources for a healthy lifestyle. Family-reunified immigrants can obtain their residence permits already in their country of origin and are received by a settled and supportive network, which can be protective to health. On the other hand, refugees receive a health assessment upon arrival in Danish asylum centres, which family-reunified immigrants do not. The assessment has, however, focused on acute and infectious diseases rather than on prevention of chronic diseases and is therefore unlikely to affect cardiac risk. Furthermore, living conditions in asylum centres can negatively affect mental health.³¹

Methodological strengths and limitations

This study included all individuals obtaining residence permit in Denmark as refugee or family-reunified immigrant from 1 January 1993–1 December 1999, thereby omitting problems of selection due to sampling. A large study population allowed us to acknowledge heterogeneity, and the personal identification number made it possible to base all measurements on individual data. The registers used in this study have nearly complete coverage,³² and for acute diagnoses the internal validity of CHD has been found to be high in the DNPR.¹⁹ A limitation of the register-based design is the lack of information on risk factors for CHD which is not routinely collected in Danish national registers. Also, the study design did not allow for comparison to the population in countries of origin, which could have provided explanations for the observed differences.²²

Bias in relative risk estimates could occur if the number of cases not leading to either hospitalisation or death from CHD differs between immigrant groups and Danish-born people. For acute cases of CHD we expect bias to be minimal, but for non-acute cases minor bias is possible.

It has been argued that self-reported ethnicity offers a more precise and nuanced measurement,³³ as country of birth does not capture the extent to which individuals conform to socio-cultural patterns, customs, and changes in these factors within a lifetime. However, country of birth produces well-defined groups which are suitable for tailored health-promoting interventions.

Personal income is likely to be an inadequate indicator of SES because it only captures some aspects of SES, thus it would have been valuable to include other indicators such as education, employment status, neighbourhood deprivation etc.

Conclusion

Our study reveals a serious public health problem, identifying vulnerable immigrant groups with a need for tailored preventive initiatives. The higher incidence among refugees and among immigrants with lower SES calls for preventive strategies which focus on factors such as trauma, life stress, social support, mental diseases, unemployment, discrimination and living conditions. Also, health professionals should take migration history and social conditions into consideration in risk assessment and searching for diagnoses in immigrant patients. Future research should address immigrants as a heterogeneous group and include measures of the migration process and SES.

Acknowledgements

The authors would like to acknowledge the Danish Immigration Service and Statistics Denmark for their help in establishing the cohort, and Morten Hulvej Rod and Morten Grønbæk for useful comments on the text. Regarding author contributions, AB and LZ have contributed equally. They conceived the study, conducted the statistical analysis, compiled and interpreted the data and drafted the manuscript. MN participated in the design and planning process. JHP supervised the statistical analysis. All authors participated in the interpretation of data, critically reviewed and contributed in writing the manuscript and approved the final version.

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Conflict of interest

The authors declare that there is no conflict of interest.

References

Author notes