Author + information
- Ashim Hajra, MD†∗ (, )
- Yan Li, MD, PhD‡,
- Stanton Siu, MD§,
- Natalia Udaltsova, PhD⋮,
- Mary Anne Armstrong, MA⋮,
- Gary D. Friedman, MD⋮ and
- Arthur L. Klatsky, MD⋮
- †Department of Medicine Kaiser Permanente, Oakland, California
- ‡Department of Hematology and Oncology Kaiser Permanente, Oakland, California
- §Department of Pulmonary and Critical Care Medicine, Kaiser Permanente, Oakland, California
- ⋮Division of Research, Kaiser Permanente, Oakland, California
- ↵∗Department of Medicine, Kaiser Permanente, 280 West MacArthur Boulevard, Oakland, California 94611
To the Editor:
Defined as persons of Indian, Pakistani, Bangladeshi, or Sri Lankan ancestry, the South Asian (SA) population is believed to have a higher risk of coronary artery disease (CAD) than that of other ethnic groups (1–6). Substantial evidence comes mostly from prevalence data or case-control comparisons. The few published cohort analyses with clinical CAD endpoints (1–4) include our small-scale report (1). We present here an updated long-term follow-up with comparisons to multiple ethnic groups. We believe this is the most detailed cohort study to date of CAD in the SA population.
We studied 126,905 people who voluntarily underwent a health examination offered by a northern California prepaid health plan between 1978 and 1985. The procedure included health measurements and queries about sociodemographic status, medical history, and habits. The questionnaire item “What is your race?” identified the following groups: white (n = 69,911; 55.1%), African American (black; n = 34,956; 27.5%), Asian American (Asian; n = 13,585; 10.7%), Hispanic (n = 5,659; 4.5%), and other (n = 2,794; 2.2%). Among the people in the Asian group, 6,050 (45%) were self-identified as Chinese, 4,236 (31%) as Filipino, 1,707 (13%) as Japanese, and 878 (6%) as other Asian (largely Korean or Vietnamese). Before analyses of endpoints, SA people (n = 714; 5%) were identified on the basis of birthplace, surname, sex, marital status, and religion among those self-identified as other Asian or other. Independently, 2 investigators performed this review with adjudication of disagreements.
Subsequent hospital admissions to health plan centers were identified from computerized sources. The primary discharge diagnosis ascertained hospitalizations for CAD, defined by codes 410 to 414 in the International Classification of Diseases, Ninth Revision, Clinical Modification. Each subject was followed up from the date of examination to the end of plan membership, to hospitalization for CAD, or through 2009, whichever came first. Mean follow-up was 17.4 years, with a total of 2,158,856 person-years. Cox proportional hazards models with 8 covariates (Table 1) yielded hazard ratios (HRs), 95% confidence intervals (CIs), and p values.
Ethnicity was studied in: 1) all people with white as referent plus black, Asian, Hispanic, and other subgroups; 2) all people with white as referent plus black, Hispanic, other, and Asian subgroups; 3) all people with black or Hispanic as referent; and 4) Asian people only with, in different models, Chinese, SA, Japanese, or Filipino as referent plus other Asian ethnicities. Adjusted models showing various comparisons are provided in Table 1.
Age- and sex-adjusted models with white as referent yielded HR estimates ranging from 0.7 in Chinese people to 2.3 in SA people (p < 0.001 for each). With further adjustment for the 8 covariates in our basic multivariate model, these HRs were 0.8 in Chinese people and 2.4 in SA people (Table 1). Addition of systolic (or diastolic) blood pressure, total cholesterol level, blood glucose level, and leukocyte count separately or together had little effect on these HRs (Table 1).
The risk of CAD in SA people was substantially higher than that in any other ethnic group (all p < 0.001) (Table 1). This increased risk was consistent in most strata, including men, “yes” or “no” with respect to a cardiorespiratory composite inquiry, smoking categories, <50 years of age at baseline, and diagnosis of acute myocardial infarction or other CAD (all p < 0.01). For women, with only 9 cases of CAD, the p value exceeded 0.05.
All covariate traits in our multivariate models had significant relationships to risk of CAD. There was an increased risk associated with smoking, higher BMI, the cardiorespiratory composite, blood pressure, cholesterol level, glucose level, and leukocyte count; whereas decreased risk was associated with alcohol drinking, college graduation, and never married status.
Our main finding were the substantially higher risk of CAD in SA people when compared with any of 7 other ethnic groups in northern California. The persistence of a similarly increased risk after correction for multiple risk factors and the general consistency in multiple strata strengthen the likelihood that the increased risk of CAD in SA people is a true independent relationship. Relative to the increased risk of SA, other interethnic disparities are modest. These data are compatible with prior prospective analyses (1–4).
An explanation for the high risk of CAD is not evident in our data or those of others (2–4). In our cohort, SA people did not have a higher prevalence of conventional CAD risk factors at baseline. Furthermore, after adjustment for age, covariate control had little effect on our CAD risk estimates. Proposed plausible mechanisms include the increased prevalence of the metabolic syndrome, increased genetic susceptibility to conventional risk factors, nonconventional dyslipidemia traits, and lifestyle traits.
Our study is limited by the use of a 1978 to 1985 northern California cohort devoid of more recent SA immigrants that might have different traits related to risk of CAD. Other limitations include absence of follow-up covariate measurements and data on lipid subfractions, waist/hip circumference, dietary habits, and exercise. Strengths include the large study population, long follow-up, large number of ethnic groups for comparison, and control for many CAD risk traits.
These data highlight the importance of defining Asian subgroups during health data collection (5) and improving risk assessment tools in the SA population. Despite the absence of proven mechanisms for high risk of CAD in SA people, it is appropriate to pursue aggressive management of traditional risk factors and to alert healthcare providers and SA patients.
Please note: This work was supported by a Community Budget Grant from The Kaiser Foundation Research Institute to Dr. Li. Data collection from 1978 to 1985 was supported by a grant from the Alcoholic Beverage Medical Research Foundation to Dr. Klatsky. All other authors have reported that they have no relationships relevant to the content of this paper to disclose.
- American College of Cardiology Foundation
- Lee J.,
- Heng D.,
- Chia K.S.,
- Chew S.K.,
- Tan B.Y.,
- Hughes K.
- Miller G.J.,
- Beckles G.L.A.,
- Maude G.H.,
- et al.
- Tillin T.,
- Hughes A.D.,
- Mayet J.J.,
- et al.
- Gupta M.,
- Singh N.,
- Verma S.