Strong Heart Study - Strong Heart Study


Population-based survey of cardiovascular risk factors and prevalent and incident cardiovascular disease in 13 American Indian tribal communities in Arizona, Oklahoma, and South and North Dakota.


The goals of this longitudinal study were to investigate and quantify rates of CHD morbidity and mortality in geographically diverse native American Indian tribes using a standardized protocol. A second goal of this study was to identify and compare CHD risk factors among tribes from different geographic regions in the U.S. to explain possible differences in the occurrence of CHD. The third objective of this study was to estimate the prevalence of diabetes in these diverse tribes and assess the importance of this metabolic disorder as a risk factor for CHD.

Study Design

Study Design:

Patients Enrolled: 4549
Mean Follow Up: 7 years
Female: 60

Patient Populations:

Members 45 to 74 years of age in 3 tribes in central Arizona, 7 in southwestern Oklahoma, and 3 in South and North Dakota were recruited from tribal members living on reservations or (in Oklahoma) in a defined geographic area (overall participation rate=62%) for a first examination in 1989 to 1992. Standardized measurements of seated brachial BP; aspects of body habitus, including body mass index, waist-to-hip ratio, and percent body fat by bioelectric impedance; and fasting and 2-hour postload glucose and glycosylated hemoglobin levels were obtained. DM was diagnosed by World Health Organization criteria if fasting glucose was >140 mg/dL, 2-hour postchallenge glucose was >200 mg/dL, or participants were receiving hypoglycemic medication; normal glucose tolerance was identified by fasting sugar levels <140 mg/dL and 2-hour postchallenge glucose <200 mg/dL. Individuals with impaired glucose tolerance were not considered in the present study to maximize the contrast between groups. The second SHS examination was conducted in 1993 to 1995 to assess change over time in body habitus, BP, and other baseline measures and to add echocardiography. A total of 3630 surviving SHS enrollees participated in the second examination, an 89% return rate. Echocardiograms were performed in 3501 participants (97%), with the remainder missed because of delay in initiating echocardiography in 2 field centers.

Primary Endpoints:

Mortality Cardiovascular mortality Cerebrovascular mortality Non-fatal MI Non-fatal stroke

Drug/Procedures Used:

Interviews of tribal community health representatives were conducted to ascertain the location and vital status of potential study participants. Baseline (phase I) and follow-up (phase II) examinations consisted of personal interviews and physician examinations. Family history, degree of acculturation, socioeconomic status, diet, alcohol and tobacco consumption, and physical activity were assessed in personal interviews. Measurements of body fat, body circumferences, blood pressure, examination of the heart and lungs, assessment of peripheral vascular disease, and a 12 lead ECG were carried out in a standardized physical examination. Fasting blood samples were obtained for the measurement of serum lipids, lipoproteins, insulin, plasma creatinine, plasma fibrinogen, and glycohemoglobin. An oral glucose tolerance test was performed and urinary albumin excretion was estimated. DNA was extracted from lymphocytes and stored for future genetic studies. A baseline (phase I) examination between 1988 and 1991 was followed by a second examination (phase II) between 1993 and 1995. Examination participation rates were approximately 70% in Oklahoma and Arizona and 53% in the Dakotas. Tribal and Indian Health Service medical records and death certificates were reviewed to ascertain fatal and nonfatal CVD events. This information was reviewed by the Strong Heart Study morbidity and mortality review committee to establish the specific CVD diagnosis. Percentage of Indian heritage was calculated from the reported degree of Indian heritage for each parent and grandparent.

Principal Findings:

Rates of nonfatal and fatal CHD were higher for men than for women and the rates of fatal stroke were lower than fatal CHD. In men, incidence rates of fatal and nonfatal CHD were lowest in Arizona, highest in South/North Dakota and intermediate in Oklahoma. Some differences in these patterns were seen in women with the lowest rates observed in Oklahoma. In terms of risk factors for fatal and nonfatal CHD, hypertension, HDL cholesterol (inverse), diabetes, albuminuria, and fibrinogen were associated with CHD in both men and women. Rates of CHD were also compared to those in two other U.S. population studies, namely CHS and ARIC, which used similar methods of ascertainment. Stroke rates were lower in American Indian men than in men from these other studies but the rates for CHD in American Indian men and women were almost 2 fold higher. The most likely explanation for the high rates of CVD in American Indians was the high prevalence of diabetes – 56% of the CVD events in men and 78% of the events in women occurred in those with diabetes. In Arizona, approximately 70% of study subjects had diabetes, while more than 40% of individuals from the other study centers had diabetes. Diabetic men had a more than 2 fold increased risk of CVD while diabetic women had a 3.5 fold increased risk compared with non-diabetics. Albuminuria was a strong independent risk factor for CVD in men and women. Obesity and body fat distribution were also associated with increased risk of CVD in men and women. No association with CVD risk was found when full blooded Indians were compared with those with non-Indian admixture.


The study was well designed and carefully conducted. Participation rates of all eligible tribal members ranged from 55% to 72% in the 3 study centers. Non-participants were reported to be similar to participants in terms of age and self-reported history of diabetes. This provides some reassurance as to the comparability of participants and nonparticipants, for purposes of generalizability, though no further comparison data were reported. The published findings pertain to individuals aged 45-74 years. The American Indian population is relatively young: only 17% are between the ages of 45-74 compared with more than 1 quarter of U.S. whites. The study went to considerable length to validate the occurrence of study endpoints through the review of medical records and death certificates by independent physician reviewers. The study used a standardized protocol in each of the 3 geographic regions which allowed for more reliable comparisons to be drawn. An important strength of the study lies in the comparisons between participating Strong Heart Study centers and 2 longitudinal studies of the general U.S. population. The study results suggest an apparent reversal of the low prevalence and mortality rates of CHD seen in earlier studies of Pima-Indians in Arizona, which had given rise to hypotheses regarding an “innate protection” from atherosclerosis for American Indians. The results of tis prospective study suggest that the rising prevalence of diabetes is very important to address in the control of CHD in this population.


Circulation 2000;101:2271-6. Circulation 1999;99:2389-2395. Am J Epidemiol 1995;142:254-268.

Clinical Topics: Diabetes and Cardiometabolic Disease, Dyslipidemia, Noninvasive Imaging, Prevention, Vascular Medicine, Atherosclerotic Disease (CAD/PAD), Lipid Metabolism, Nonstatins, Echocardiography/Ultrasound, Diet, Hypertension

Keywords: DNA, Follow-Up Studies, Electric Impedance, Death Certificates, Risk Factors, Creatinine, Electrocardiography, Advisory Committees, Hemoglobin A, Glycosylated, Glucose Intolerance, Waist-Hip Ratio, Glucose Tolerance Test, Motor Activity, Obesity, Cholesterol, HDL, Tobacco Use, Fibrinogen, Population Groups, Hypertension, Echocardiography, Insulin, Stroke, Atherosclerosis, United States Indian Health Service, Medical Records, Social Class, Peripheral Vascular Diseases, Body Mass Index, Lymphocytes, Hypoglycemic Agents, Fasting, Diabetes Mellitus, Body Fat Distribution

< Back to Listings