Evidence for distinct genetic effects on obesity and lipid-related CVD risk factors in diabetic compared to nondiabetic American Indians: the Strong Heart Family Study

BACKGROUND

The relationship between diabetes and increased cardiovascular disease (CVD) risk is well established. Previous data from the Strong Heart Family Study (SHFS) have demonstrated significant heritabilities for CVD risk factors and implicated diabetes as an important predictor of several of the phenotypes in three large groups of American Indians in Arizona (AZ), Oklahoma (OK), and the Dakotas (DK). However, the genetic architecture of this relationship is not well understood. The purpose of the present paper is to determine whether the genetic effects on CVD risk factors in the SHFS are different in diabetic and nondiabetic individuals.

METHODS

Approximately 950 men and women, 18 years or older, in 32 extended families, were examined between 1997 and 1999. Interaction between genotype and diabetes status was estimated for nine CVD risk factors [body fat mass (FAT), body mass index (BMI), high-density lipoprotein cholesterol (HDL-C), log-transformed fibrinogen, log-transformed triglycerides, log-transformed urinary albumin : creatinine ratio (LACR), plasminogen activator inhibitor 1 (PAI-1), systolic blood pressure (SBP), and waist-to-hip ratio (WHR)], using a maximum likelihood variance decomposition technique.

RESULTS

We found evidence that genetic factors influencing BMI, FAT, log-transformed triglycerides, HDL-C, and WHR are differentially expressed in individuals with diabetes compared to those without diabetes.

CONCLUSION

Thus, in the environmental and genetic milieu of diabetes, there may be distinct genetic determinants of various CVD risk factors.

Genetic and environmental contributions to cardiovascular disease risk in American Indians: the strong heart family study

The aims of the Strong Heart Family Study are to clarify the genetic determinants of cardiovascular disease (CVD) risk in American Indians and to map and identify genes for CVD susceptibility. The authors describe the design of the Strong Heart Family Study (conducted between 1998 and 1999) and evaluate the heritabilities of CVD risk factors in American Indians from this study. In the first phase of the study, approximately 950 individuals, aged 18 years or more, in 32 extended families, were examined. The examination consisted of a personal interview, physical examination, laboratory tests, and an ultrasound examination of the carotid arteries. The phenotypes measured during the physical examination included anthropometry, lipoproteins, blood pressure, glycemic status, and clotting factors. Heritabilities for CVD risk factor phenotypes were estimated using a variance component approach and the program SOLAR. After accounting for the effects of covariates, the authors detected significant heritabilities for many CVD risk factor phenotypes (e.g., high density lipoprotein cholesterol (heritability = 0.50) and diastolic blood pressure (heritability = 0.34)). These results suggest that heredity explains a substantial proportion of the variability of CVD risk factors and that these heritabilities are large enough to warrant a search for major risk factor genes.

Alcohol intake and glycemia in American Indians: the strong heart study

The goal of this study was to explore the relationship between alcohol intake and glycemia and type 2 diabetes in American Indians aged 45 to 74 years. Data were obtained from participants in the Strong Heart Study, a longitudinal study of 13 American Indian communities in 3 geographic areas in the United States. Alcohol consumption was determined by self-reported alcohol intake history. Participants previously diagnosed with diabetes were excluded from the analysis. Analysis of covariance (ANCOVA) was used to estimate the adjusted means of blood glucose for alcohol intake categories. Logistic regression was used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs) to estimate the association between alcohol intake and type 2 diabetes in the cross-sectional analysis and between alcohol intake and glucose intolerance using longitudinal data. Fasting and 2-hour plasma glucose concentrations showed an inverse J-shaped curve across categories of alcohol intake. Using never drinkers as the referent group in cross-sectional analysis, light drinkers had a significantly lower risk of having diabetes (OR, 0.66; 95% CI, 0.44 to 0.99); among drinkers, heavy drinkers had a higher, although not significant, prevalence of diabetes. Longitudinal analysis showed no significant worsening of glucose tolerance across levels of alcohol intake. Subanalyses stratified by body mass index (BMI) did not show differences between obese and non-obese participants in the relationship between alcohol intake and glucose tolerance. Although plasma glucose concentration showed a shallow, inverse J-shaped association across levels of increasing alcohol intake in American Indians aged 45 to 74 years, alcohol intake did not appear to significantly increase the risk for worsening glucose tolerance. Thus, alcohol intake does not appear to be a determinant of diabetes risk in this population.

Prevalence of diabetes is higher among female than male Zuni indians

OBJECTIVE

Test the hypothesis that diabetes and related risk factors are more common among female than male Zuni Indians.

RESEARCH DESIGN AND METHODS

We conducted a population-based, cross-sectional survey of the Zuni Indians aged > or =5 years. We used households within neighborhood clusters as the sampling frame. We administered a questionnaire, collected blood and urine, and measured height and weight. Self-reported diabetes was used to assess previously diagnosed diabetes. Participants without a prior history of diabetes were classified as having newly diagnosed diabetes if they had HbA(1c) >7.0% or random glucose > or =11.1 mmol/l during the survey.

RESULTS

The prevalence of previously diagnosed diabetes among Zuni Indians aged > or =5 years (n = 1,503) was higher among female Zuni Indians (16.7% [95% CI 14.1-19.3]) than male Zuni Indians (9.7% [7.4-12.1]) (P < 0.001). The prevalence of newly diagnosed diabetes was similar among female Zuni Indians (2.4% [1.4-3.4]) and male Zuni Indians (2.4% [1.2-3.6]). The prevalence of previously and newly diagnosed diabetes was higher among female Zuni Indians (19.1% [16.4-21.9]) than male Zuni Indians (12.2% [9.5-14.8]) (P < 0.001). The prevalence of obesity was higher among female Zuni Indians (34.3% [30.9-37.7]) than male Zuni Indians (21.5% [18.4-24.7]) (P < 0.001). Obesity was associated with diabetes among female and male Zuni Indians. Physical inactivity was more common among female Zuni Indians (44.2% [40.7-47.8]) than male Zuni Indians (35.1% [31.5-38.7]) (P < 0.001). However, physical inactivity was not associated with diabetes among either female or male Zuni Indians. Gestational diabetes was a risk factor among female Zuni Indians.

CONCLUSIONS

Among the Zuni Indians, the prevalence of diabetes was 57% higher among female than male members of the population. Culture, tradition, and lifestyle differences may contribute to the higher prevalence of diabetes and obesity among female Zuni Indians.

Risk factors for sudden infant death syndrome among northern plains Indians

CONTEXT

Sudden infant death syndrome (SIDS) is a leading cause of postneonatal mortality among American Indians, a group whose infant death rate is consistently above the US national average.

OBJECTIVE

To determine prenatal and postnatal risk factors for SIDS among American Indians.

DESIGN, SETTING, AND PARTICIPANTS

Population-based case-control study of 33 SIDS infants and 66 matched living controls among American Indians in South Dakota, North Dakota, Nebraska, and Iowa enrolled from December 1992 to November 1996 and investigated using standardized parental interview, medical record abstraction, autopsy protocol, and infant death review.

MAIN OUTCOME MEASURES

Association of SIDS with maternal socioeconomic and behavioral factors, health care utilization, and infant care practices.

RESULTS

The proportions of case and control infants who were usually placed prone to sleep (15.2% and 13.6%, respectively), who shared a bed with parents (59.4% and 55.4%), or whose mothers smoked during pregnancy (69.7% and 54.6%) were similar. However, mothers of 72.7% of case infants and 45.5% of control infants engaged in binge drinking during pregnancy. Conditional logistic regression revealed significant associations between SIDS and 2 or more layers of clothing on the infant (adjusted odds ratio [aOR], 6.2; 95% confidence interval [CI], 1.4-26.5), any visits by a public health nurse (aOR, 0.2; 95% CI, 0.1-0.8), periconceptional maternal alcohol use (aOR, 6.2; 95% CI, 1.6-23.3), and maternal first-trimester binge drinking (aOR, 8.2; 95% CI, 1.9-35.3).

CONCLUSIONS

Public health nurse visits, maternal alcohol use during the periconceptional period and first trimester, and layers of clothing are important risk factors for SIDS among Northern Plains Indians. Strengthening public health nurse visiting programs and programs to reduce alcohol consumption among women of childbearing age could potentially reduce the high rate of SIDS.

Heritability of carotid artery structure and function: the Strong Heart Family Study

OBJECTIVE

Alterations in carotid artery structure and function may represent phenotypic measures of vascular disease that contain information beyond that which can be inferred from conventional cardiovascular disease risk assessment. However, apart from their associations with cardiovascular disease risk factors and outcome, the genetic basis of variations in carotid artery structure and function is largely unknown. The purpose of this study was to examine the genetic and environmental contributions to carotid artery structure and function in 3 large groups of American Indians.

METHODS AND RESULTS

Approximately 950 men and women, aged > or =18 years, in 32 extended families were examined between 1997 and 1999. By use of a variance component approach and the program Sequential Oligogenic Linkage Analysis Routines, heritabilities for carotid artery structure and function phenotypes were estimated. After accounting for the effects of covariates (sex, age, diabetes, impaired glucose tolerance, smoking, cholesterol, body surface area, and hypertension), we detected significant heritabilities (given as h2 values) for common carotid artery diastolic diameter (h2<0.44), intimal-medial wall thickness (h2<0.21), vascular mass (h2<0.27), arterial stiffness (h2<0.23), and the augmentation index (h2<0.18).

CONCLUSIONS

These results suggest that the additive effects of genes explain a moderate proportion of the variability of carotid artery structure and function.

Using HbA(1c) to improve efficacy of the american diabetes association fasting plasma glucose criterion in screening for new type 2 diabetes in American Indians: the strong heart study

OBJECTIVE

To find an optimal critical line in the fasting plasma glucose (FPG)-HbA(1c) plane for identifying diabetes in participants with impaired fasting glucose (IFG) and thereby improve the efficacy of using FPG alone in diabetes screening among American Indians.

RESEARCH DESIGN AND METHODS

We used FPG, 2-h postload glucose (2hPG), and HbA(1c) measured in the 2,389 American Indians (aged 45-74 years, without diabetes treatment or prior history of diabetes) in the Strong Heart Study (SHS) baseline (second) examination. Participants were classified as having diabetes if they had either FPG > or =126 mg/dl or 2hPG > or =200 mg/dl, as having IFG if they had 110 < or = FPG < 126 mg/dl, and as having normal fasting glucose (NFG) if they had FPG <110, according to the American Diabetes Association (ADA) definition. Logistic regression models were used for identifying diabetes (2hPG > or =200 mg/dl) in IFG participants. The areas under the receiver operating characteristic (ROC) curves generated by different logistic regression models were evaluated and compared to select the best model. A utility function based on the best model and the cost-to-benefit ratio was used to find the optimal critical line. The data from the second examination were used to study the effect of the time interval between the successive diabetes screenings on both the FPG criterion and the optimal critical line.

RESULTS

A total of 37% of all subjects with new diabetes at baseline and 55.2% of those in the second exam had 2hPG > or =200 but FPG <126. There was a very large portion of IFG participants with diabetes (19.3 and 22.9% in the baseline and second exam, respectively). Among the areas under the ROC curves, the area generated by the logistic regression model on FPG plus HbA(1c) is the largest and is significantly larger than that based on FPG (P = 0.0008). For a cost-to-benefit ratio of 0.23888, the optimal critical line that has the highest utility is: 0.89 x HbA(1c) + 0.11 x FPG = 17.92. Those IFG participants whose FPG and HbA(1c) were above or on the line were referred to take an oral glucose tolerance test (OGTT) to diagnose diabetes. The optimal critical line is lower if a successive diabetes screening will be conducted 4 years after the previous screening.

CONCLUSIONS

FPG > or =126 and 2hPG > or =200, as suggested by the ADA, are used independently to define diabetes. The FPG level is easy to obtain, and using FPG alone is suggested for diabetes screening. It is difficult to get physicians and patients to perform an OGTT to get a 2hPG level because of the many drawbacks of the OGTT, especially in those patients who already have FPG <126. It is also impractical to conduct an OGTT for everyone in a diabetes screening. Our data show that 37% of all subjects with new diabetes in the SHS baseline exam and 55.2% of those in the second exam have 2hPG > or =200 but FPG <126. These cases of diabetes cannot be detected if FPG is used alone in a diabetes screening. Therefore, although the small portion of diabetes in the NFG group (4.7% in the baseline and 6.9% in the second exam) may be ignored, those cases of diabetes among IFG participants ( approximately 20% in our data) need further consideration in a diabetes screening. It may be worthwhile for those IFG participants identified by the optimal critical line to take an OGTT. The optimal critical line and time interval between successive diabetes screenings need further study.

Prevalence of gallbladder disease in American Indian populations: findings from the Strong Heart Study

American Indians are believed to be at high risk of gallbladder disease (GBD), but there has been no systematic evaluation of its prevalence among diverse groups of American Indians. Therefore, we determined the prevalence of GBD and associated risk factors among specified American Indian populations using ultrasonography of the gallbladder and standardized diagnostic criteria. Enrolled members, aged 47 years and older, of 13 American Indian tribes or communities in Arizona, Oklahoma, and South and North Dakota who participated in the Strong Heart Study were analyzed. GBD was the sum of gallstones (determined by ultrasound examination) and cholecystectomy (determined by ultrasound and self-report). The proportion of American Indian heritage was based on the heritage of the grandparents of participants. GBD prevalence was determined among 3,296 participants at the 3 sites. Among women, 17.8% had gallstones, and 46.3% had evidence of a cholecystectomy, for a total of 64.1% with GBD. Among men, 17.4% had gallstones, and 12.1% had evidence of a cholecystectomy, for a total of 29.5% with GBD. When figures were adjusted for age and Indian heritage, there was no significant difference in GBD prevalence across the 3 geographical areas. In multivariate logistic regression analysis, age, American Indian heritage, and waist circumference were associated with GBD among men, and age, American Indian heritage, diabetes, and parity were associated with GBD among women. Body mass index was not independently associated with GBD in either sex. In conclusion, GBD was found in epidemic proportions in diverse American Indian populations.

Genetic influences on aortic root size in American Indians: the Strong Heart Study

Aortic root dilatation is a major pathophysiological mechanism for aortic regurgitation and predisposes the aortic root to dissection or rupture. However, only a small proportion of the variance of aortic root size can be explained by its known clinical and demographic correlates. The present study was undertaken to determine the heritability of echocardiographically derived aortic root diameter in the American Indian participants in the second Strong Heart Study examination. Echocardiograms were analyzed in 1373 SHS participants who had > or =1 family member in the cohort. Heritability calculations were performed by using variance component analysis as implemented in SOLAR, a computer analysis program. In a polygenic model, the variables entered and identified as covariates of larger aortic root diameter were older age, male sex, and center (P<0.001), which accounted for 35% of the overall variability of aortic root diameter. After simultaneous adjustment was made for these significant covariates, the proportion of phenotypic variance due to additive genetic contribution or residual heritability (h2) was 0.51 (SE=0.08, P<0.001). Additionally, simultaneous adjustment for height, weight, and systolic and diastolic BPs yielded slightly lower residual h2 of aortic root diameter (h2=0.44, SE=0.08, P<0.001), which accounted for 26% of the overall variance of aortic root size. Because center effects were identified as significant covariates in the analyses, h2 analyses were performed separately in Arizona, Oklahoma, and North/South Dakota centers, which confirmed that a significant proportion of the phenotypic variance of aortic root diameter is due to additive genetic contribution. Heredity explains a substantial proportion of the variability of aortic root size that is not accounted for by age, sex, body size, and blood pressure. Echocardiographic screening of family members with aortic root dilatation may identify other individuals predisposed to aortic dissection or rupture.

Associations of insulin levels with left ventricular structure and function in American Indians: the strong heart study

We evaluated the association of insulin and echocardiographic left ventricular (LV) measurements in 1,388 (45% men) nondiabetic American Indian participants in the Strong Heart Study (SHS). Significant (all P < 0.05) relations were found in men and women between log(10) fasting insulin and LV mass (r = 0.24 and 0.26), left atrial diameter (r = 0.25 and 0.28), posterior wall thickness (r = 0.20 and 0.26), septal thickness (r = 0.19 and 0.24), LV diameter (r = 0.17 and 0.16), and cardiac output (r = 0.20 and 0.24) and in women relative wall thickness (r = 0.11) and peripheral resistance (r = -0.17). In regression analyses, adjusting for BMI, age, height, and systolic pressure, fasting insulin was independently correlated with cardiac output in men and relative wall thickness and septal thickness in women (all P < 0.05). The 97th percentiles of fasting insulin (25 microU/ml for men, and 23 microU/ml for women) in 163 apparently normal (BMI <26; blood pressure <140/90; and absence of diabetes, valvular disease, LV wall motion abnormality, or antihypertensive treatment) SHS participants were used to separate normal from elevated fasting insulin levels. Adjusting for age, BMI, and height, men with elevated insulin levels had larger LV diameters (5.41 vs. 5.16 cm; P = 0.05), higher cardiac output (5.5 vs. 4.9 l/min; P < 0.001), and lower peripheral resistance (1,487 vs. 1,666; P = 0.01), paralleling results of regression analyses. Positive relations between insulin and heart size in nondiabetic adults are largely due to associations with body size; after adjustments for covariates, fasting insulin levels are related to greater LV size and cardiac output in men and more concentric LV geometry in women.

Mitral ratio of peak early to late diastolic filling velocity as a predictor of mortality in middle-aged and elderly adults: the Strong Heart Study

BACKGROUND

With aging, left ventricular filling tends to decrease in early diastole, reducing the mitral ratio of peak early to late diastolic filling velocity (E/A). However, the prognostic significance of low or high E/A in older adults remains to be elucidated in population-based samples.

METHODS AND RESULTS

Doppler echocardiograms were analyzed in 3008 American Indian participants in the second Strong Heart Study examination who had no more than mild mitral or aortic regurgitation. Participants were followed for a mean of 3 years after Doppler echocardiography to assess risks of all-cause and cardiac death associated with E/A <0.6 or >1.5; 2429 (81%) participants had normal E/A ratio, 490 (16%) had E/A <0.6, and 89 (3%) had E/A >1.5. All-cause mortality was higher with E/A <0.6 or E/A >1.5 (12% and 13% versus 6%), as was cardiac mortality (4.5% and 6.5% versus 1.6%; both P<0.001). Adjusting for age, sex, body mass index, systolic blood pressure, HDL and LDL cholesterol, smoking, hypertension, diabetes, coronary heart disease, left ventricular hypertrophy, and low ejection fraction (<40%), the relative risk of all-cause death with E/A >1.5 was 1.73 (95% CI, 0.99 to 3.03; P=0.05); the relative risk of cardiac death was 2.8 (95% CI, 1.19 to 6.75; P<0.05). E/A <0.6 was not independently associated with increased all-cause or cardiac mortality (P=0.19 and 0.31, respectively) after adjusting for covariates.

CONCLUSIONS

In a population-based sample of middle-aged and elderly adults, mitral E/A >1.5 at baseline Doppler echocardiography is associated with 2-fold increased all-cause and 3-fold increased cardiac mortality independent of covariates; mitral E/A <0.6 was also associated with 2-fold increased all-cause and cardiac mortality but not independent of covariates.

Respondent bias in the collection of alcohol and tobacco data in American Indians: the Strong Heart Study

This study addresses the impact of assessment method (interviewer-administered questionnaire vs. self-administered questionnaire) and interviewers demographic characteristics (gender, ethnicity, and residency) on responses to alcohol and tobacco questions. The study population included 1,522 men and women aged 45 to 74 from the Dakota Center of the Strong Heart Study (SHS), a multi-center study of cardiovascular disease in American Indians. Assessment method effects were greater for alcohol than tobacco but did not differ by interviewer characteristics.

Recruitment of American Indians in epidemiologic research: the Strong Heart Study

This paper describes the methods used to recruit American Indian (AI) populations for the Strong Heart Study (SHS), a community-based study of cardiovascular disease (CVD) and its risk factors in AI men and women. Recruitment strategies included personal contact by recruiters and drivers/recruiters in remote areas, SHS staff participation in community activities, and mass media. A total of 4,549 participants aged 45-74 years were recruited from 13 American Indian tribes and communities. Overall participation rates were 72&percnt, 55&percnt, and 62&percnt, respectively, for the three study centers (Arizona, the Dakotas, and Oklahoma). Participant feedback and educational material related to risk factor reduction and promoting a healthy lifestyle were emphasized. Participants were likely to be female, young, and nonsmokers. Barriers to recruitment included lack of telephones in a large proportion of households, conflicting beliefs about health/health care/research, fears, taboos, and occasional rumors about study examination procedures. Participants were referred for follow-up of health problems detected by the study. The strong commitment of the participating communities helped to insure the success of the SHS, which can be considered a model for recruitment in future American Indian population-based studies. Success was facilitated by the use of a variety of recruitment techniques.

The Aberdeen Indian Health Service infant mortality study: design, methodology, and implementation

Of all Indian Health Service areas, the Aberdeen Area has consistently had the highest infant mortality rate. Among some tribes in this area the rate has exceeded 30/ 1000 live birth and half the infant deaths have been attributed to Sudden Infant Death Syndrome,a rate four to five times higher than the national average. The Indian Health Service, Centers for Disease Control and Prevention, National Institute of Child Health and Human Development, and the Aberdeen Area Tribal Chairmen's Health Board collaborated to investigate these high rates with the goals of refining the ascertainment of the causes of death, improving cause-specific infant mortality rates and identifying factors contributing to the high rates. Ten of the 19 tribes or tribal communities, representing 66%of the area population, participated in a 4-year prospective case-control study of infants who died after discharge from the hospital. Infant care practices and socio-demographic, economic, medical, health care, and environmental factors were examined. The study included parental interviews, death scene investigations, autopsies, neuropathology studies, medical chart abstractions, blood cotinine assays, and a surveillance system for infant deaths. Controls were the previous and subsequent infants born on the case mother's reservation. From December 1,1992 until November 30,1996,72 infant deaths were investigated. This report describes the study methods and the model employed for involving the community and multiple agencies to study the problem of infant mortality among Northern Plains Indians. Data gathered during the investigations are being analyzed and will be published at a later date.

Lipoprotein(a) in American Indians is low and not independently associated with cardiovascular disease. The Strong Heart Study

PURPOSE

To evaluate the distribution of lipoprotein(a) (Lp(a)) and assess its association to cardiovascular disease (CVD) in American Indians.

METHODS

Lp(a) was measured in 3991 American Indians (aged 45-74 years with no prior history of CVD at baseline) from 13 communities in Arizona, Oklahoma, and South/North Dakota. They were followed prospectively from 1989 to 1997 for CVD. The distribution of Lp(a) was examined by center, sex, and diabetic status. Spearman correlation coefficients and Cox regression models were used to evaluate the association of Lp(a) to CVD.

RESULTS

A total of 388 participants subsequently developed CVD. Median Lp(a) concentration in American Indians was 3.0 mg/dl. This was almost half of that in whites and one sixth in blacks from the CARDIA study measured by the same method. Nondiabetic participants had significantly higher Lp(a) levels than diabetic participants for both genders. Lp(a) levels were higher in women than in men for nondiabetic participants, but there was no gender difference for diabetic participants. Correlation analysis showed Lp(a) was significantly negatively correlated with the degree of Indian heritage, insulin, triglycerides (TG), fasting plasma glucose (FPG), and 2-hour plasma glucose (2hPG), and positively with low-density lipoproteins (LDL), apoprotein B (apoB), and fibrinogen (FIB). In Cox regression models, adjusting for other risk factors, Lp(a) was no longer a significant predictor of CVD in either diabetic or nondiabetic participants.

CONCLUSIONS

The lower concentration of Lp(a) in American Indians and the high correlation with Indian heritage confirm the concept that Lp(a) concentration is in large part genetically determined. Lp(a) concentration is not an independent predictor of CVD among American Indians; it is higher in those who develop CVD because of its positive correlation with LDL, apoB, and FIB.

Changes in cardiovascular disease risk factors among American Indians. The Strong Heart Study

PURPOSE

This study describes changes in cardiovascular disease (CVD) risk factors in older American Indians over a 4-year period.

METHODS

The Strong Heart Study, a longitudinal population-based study of CVD and CVD risk factors among American Indians aged 45-74 years, measured CVD risk factors among 3638 members of 13 tribes in three geographic areas during examinations in 1989 to 1991 and 1993 to 1995.

RESULTS

Changes in mean low-density lipoprotein (LDL) cholesterol and the prevalence of elevated LDL cholesterol were inconsistent. Mean high- density lipoprotein (HDL) cholesterol decreased, and the prevalence of low HDL cholesterol increased throughout. Mean systolic blood pressure and hypertension rates increased in nearly all center-sex groups, and hypertension awareness and treatment improved. Smoking rates decreased but remained higher than national rates except among Arizona women. Mean weight and percentage body fat decreased in nearly all center-sex groups but the prevalence of obesity did not change significantly in any group. Diabetes and albuminuria prevalence rates increased throughout the study population. The prevalence of alcohol use decreased, but binge drinking remained common in those who continued to drink.

CONCLUSIONS

Improvements in management and prevention of hypertension, diabetes, renal disease, and obesity, and programs to further reduce smoking and alcohol abuse, are urgently needed.

A population-based, cross-sectional survey of the Zuni Pueblo: a collaborative approach to an epidemic of kidney disease

Approximately 80% of the Zuni Indians live in Zuni Pueblo in rural New Mexico. The Zuni are relatively endogamous and differ culturally and ethnically from neighboring tribes. The Zuni Pueblo is experiencing an epidemic of renal disease. The prevalence of end-stage renal disease (ESRD) among the Zuni Indians, adjusted for age and gender, is 6-fold and 21-fold higher than the prevalence rates for Native Americans and European Americans. Almost all Zuni tribal members have a relative with ESRD. This has led to strong public support for renal research. In response, the Zuni community has partnered with the Indian Health Service (IHS), University of New Mexico Health Sciences Center (UNMHSC), Southwest Foundation for Biomedical Research (SFBR), Dialysis Clinic Inc (DCI), and the National Institutes of Health (NIH) to establish a research consortium, the Zuni Kidney Project (ZKP). The goal of the ZKP is to reduce the burden of renal disease in the Zuni community. The ZKP combines traditional, population-based, epidemiologic methods with modern techniques of population genetics. The foundation of the ZKP is a population-based, cross-sectional survey (PBCSS). The PBCSS has 3 specific aims: (1) estimate the prevalence of renal disease in Zuni; (2) assess risk factors for renal disease susceptibility; and (3) identify participants for planned case-control, longitudinal cohort, and family studies designed to identify environmental, familial, and genetic risk factors for the susceptibility and/or progression of renal disease. Preliminary results of the PBCSS confirm that the epidemic of renal disease in Zuni Indians is due to high rates of both diabetic and nondiabetic renal disease.

Accuracy of lipoprotein lipids and apoproteins in predicting coronary heart disease in diabetic American Indians. The Strong Heart Study

PURPOSE

To evaluate the accuracy of lipoprotein lipids and apoproteins in predicting coronary heart disease (CHD) in diabetic American Indians.

METHODS

This study included 2099 diabetic participants of the Strong Heart Study, which is a longitudinal study of cardiovascular disease (CVD) and its risk factors in American Indians in Arizona, Oklahoma, and South and North Dakota. Diabetic participants with incident CHD (N = 126) were selected as the case group, and those without CHD or any cardiovascular events were the control group (N = 1732). Previous vascular events such as stroke were the sole exclusion criterion (N = 241). Baseline measurements of lipoprotein lipids and apoproteins were used to predict CHD diagnosed at the 4-year follow-up examination by using Receiver-Operating Characteristic (ROC) curve analysis.

RESULTS

The ratio of high-density lipoprotein (HDL) to total cholesterol had the highest area under the ROC curve (0.69 +/- 0.02). The areas for the ratios of HDL to low-density lipoprotein (LDL) cholesterol (0.68 +/- 0.02), apo AI/B (0.66 +/- 0.02), and the single component of total cholesterol (0.64 +/- 0.03) and LDL cholesterol (0.63 +/- 0.05) were not significantly different from the area for HDL/total cholesterol. However, the areas for apo B (0.64 +/- 0.02), HDL cholesterol (0.62 +/- 0.03), triglycerides (0.58 +/- 0.03), and apo AI (0.57 +/- 0.05) were significantly lower than the area for HDL/total cholesterol. Logistic regression analysis indicates that only HDL and LDL cholesterol were significant independent lipoprotein lipid and apoprotein predictors for CHD. The other significant predictors in the model were study center, age, gender, and albuminuria. The ROC area for this model is 0.75.

CONCLUSIONS

HDL and LDL cholesterol were the most important independent predictors for incident CHD in diabetic American Indians. The ratios of HDL/total cholesterol, HDL/LDL cholesterol, and apo AI/B had higher accuracy for predicting CHD. Although the values for all lipoprotein lipids and apoproteins and their ratios were not large enough to definitely predict CHD, they can be used as screening tools for CHD in diabetic American Indians.

Incidence of diabetes in American Indians of three geographic areas: the Strong Heart Study

OBJECTIVE

To estimate incidence rates of diabetes and associated risk factors among participants of the Strong Heart Study.

RESEARCH DESIGN AND METHODS

Of the 4,549 Strong Heart Study participants examined at baseline, 3,638 returned for a similar examination after an average of 4 years. The 1985 World Health Organization criteria for diabetes were used to identify new diabetes cases. Rates of diabetes among participants who did not have diabetes at baseline examination were determined. The relationships between the incidence rates of diabetes and a number of risk factors measured at baseline examination were studied.

RESULTS

Significant variables associated with the development of diabetes included triglycerides, obesity, fasting plasma glucose, insulin, and degree of American Indian blood among participants with NGT at baseline. For those with IGT at baseline, significant predictors included fasting plasma glucose, 2-h glucose, BMI, degree of American Indian blood, and albuminuria.

CONCLUSIONS

The high incidence rates found in this study were alarming. To slow down the rapid increase of this disease in the American Indian population, preventive programs must be designed and implemented. Patients with IGT should be treated with diabetes medication or put on a rigid weight-reduction program to reduce the risk of progression to diabetes.

Prevalence and correlates of mitral valve prolapse in a population-based sample of American Indians: the Strong Heart Study

PURPOSE

Mitral valve prolapse is heritable and occurs frequently in the general population despite associations with mitral regurgitation and infective endocarditis, suggesting that selective advantages might be associated with mitral valve prolapse.

SUBJECTS AND METHODS

Clinical examination and 2-dimensional and color Doppler echocardiography were performed in 3340 American Indian participants in the Strong Heart Study.

RESULTS

Mitral valve prolapse (clear-cut billowing of one or both mitral leaflets across the mitral anular plane in 2-dimensional parasternal long-axis recordings or >2-mm late systolic posterior displacement of mitral leaflets by M mode) occurred in 37 (1.8%) of 2077 women and 20 (1.6%) of 1263 men (P = 0.88); 32 (3.5%) of 907 patients with normal glucose tolerance, 11 (2.3%) of 486 patients with impaired glucose tolerance, and 13 (0.7%) of 1735 patients with diabetes (P <0.0001). Participants with mitral valve prolapse had lower mean (+/- SD) body mass index (28 +/- 5 kg/m(2) vs. 31 +/- 6 kg/m(2), P = 0.001) and blood pressure (124/71 +/- 19/10 mm Hg vs. 130/75 +/- 21/10 mm Hg, P <0.05), as well as lower levels of fasting glucose, triglycerides, serum creatinine, and log urine albumin/creatinine ratio (all P <0.001), than did those without mitral valve prolapse, although all subjects were similar in age (60 +/- 8 years). Participants with mitral valve prolapse had lower ventricular septal (0.87 +/- 0.08 cm vs. 0.93 +/- 0.13 cm) and posterior wall thicknesses (0.82 +/- 0.08 cm vs. 0.87 +/- 0.10 cm), mass (38 +/- 7 g/m(2.7) vs. 42 +/- 11 g/m(2.7)), and relative wall thickness (0.33 +/- 0.04 vs. 0.35 +/- 0.05), and increased stress-corrected midwall shortening (all P <0.01). Mitral valve prolapse was associated with a higher prevalence of mild (16 of 57 [28%] vs. 614 of 3283 [19%]) and more severe mitral regurgitation (5 of 57 [9%] vs. 48 of 3283 [1%], P <0.0001). Regression analyses showed prolapse was associated with low ventricular relative wall thickness, high midwall function, and low urine albumin/creatinine ratio, independent of age, sex, body mass index, and diabetes.

CONCLUSIONS

Mitral valve prolapse is fairly common and is strongly associated with mitral regurgitation in the general population. However, it is also associated with lower body weight, blood pressure, and prevalence of diabetes; a more favorable metabolic profile and ventricular geometry; and better myocardial and renal function.

Hypertension treatment patterns in American Indians: the strong heart study

Pharmacologic treatment patterns for hypertensive American Indians from 13 communities in Arizona, Oklahoma, South Dakota, and North Dakota were assessed. Participants (2254 women and 1384 men, aged 48 to 79 years) completed a clinical examination between July 1993 and December 1995. The mean of two blood pressure (BP) measurements and detailed medication histories were obtained. The observed prevalence of hypertension was 46.7% (n=1698). In participants taking antihypertensive medications (n=1114), four principal drug classes were evaluated: diuretics, calcium channel blocking agents, beta-blocking agents, and angiotensin-converting enzyme (ACE) inhibitors. Among treated hypertensive participants, 71.4%, 24.6%, and 4.0% received one, two, and three medications, respectively. Among single drug regimens, ACE inhibitors (n=340) were used most often (49.4%), with calcium channel blocking agents and diuretics accounting for 24.2% and 19.9%, respectively. Although multiple drug class therapies varied, the combination of a diuretic and ACE inhibitor (n=120) accounted for 47.4% of dual therapy use. Hypertension control (SBP < 140 mm Hg, DBP < 90 mm Hg) rates were highest for those on dual therapies (65.4%), followed by participants on single (53.8%) and triple (43.6%) therapies. Among monotherapies, diuretics exhibited the best overall hypertension control rate in both diabetics (63.0%) and nondiabetics (68.0%), versus 47% to 61% for other remaining agents. The frequent use of ACE inhibitors, used singly or in combination, reflects the high prevalence of diabetes among American Indians. ACE inhibitors, combined with diuretics, were particularly useful in achieving BP control in this population.

Spirometry reference values for American Indian adults: results from the Strong Heart Study

STUDY OBJECTIVE

To derive spirometry normative values from a large population of American Indian adults and compare them to reference values for white adults.

DESIGN

Pulmonary function was assessed using spirometry in participants of the Strong Heart Study, a multicenter, community-based, prospective study of cardiovascular risk factors and disease in American Indians, utilizing American Thoracic Society guidelines and a vigorous quality assurance program.

SETTING

Central Arizona, southwestern Oklahoma, central South Dakota, and northeastern North Dakota.

PARTICIPANTS

Acceptable spirometry results were obtained from 1,619 women and 1,005 men aged 45 to 74 years.

RESULTS

Internal reference values and normal ranges for FEV(1), FVC, and the FEV(1)/FVC ratio were derived from a healthy subgroup of 253 women and 190 men, identified by excluding participants with factors associated with a lower FEV(1). Ten percent of the entire cohort (269 of 2,624 subjects) had airways obstruction, as defined by an FEV(1)/FVC below the lower limit of the normal (LLN) using the internal reference equations. After allowing for measurement "noise," 31 participants were below the LLN using reference equations for white adults from the large National Health and Nutrition Examination Study (NHANES) III study but were normal using the internal reference equations (1.3% false-positive), while 27 participants were classified as normal using NHANES III equations but had airways obstruction using the internal reference equations (1.2% false-negative). Similarly low misclassification rates were seen for a low FVC (prevalence, 17.6%).

CONCLUSION

For clinical purposes, NHANES III spirometry reference equations for white adults may be used when testing American Indian women and men aged 45 to 74 years.

Separate and joint effects of systemic hypertension and diabetes mellitus on left ventricular structure and function in American Indians (the Strong Heart Study)

Although the association of systemic hypertension (SH) with diabetes mellitus (DM) is well established, the cardiac features and hemodynamic profile of patients with SH and DM diagnosed by American Diabetes Association criteria have not been elucidated. To address this issue, echocardiograms were analyzed in 1,025 American Indian participants of the Strong Heart Study with neither DM nor SH, 642 with DM alone, 614 with SH alone, and 874 with SH and DM. In analyses that adjusted for age, gender, body mass index, and heart rate, DM and SH were associated with increased left ventricular (LV) wall thicknesses, with the greatest impact of DM on LV relative wall thickness and of the combination of DM and SH on LV mass (both p <0.001). LV fractional shortening was reduced with SH and SH + DM, midwall shortening was reduced with DM, SH, and their combination, and was reduced in both diabetic groups compared with their nondiabetic counterparts (p <0.001). DM alone was associated with lower measures of LV pump performance (stroke volume, cardiac output, and their indexes) than SH alone. Pulse pressure/stroke index, an indirect measure of arterial stiffness, was elevated in participants with DM or SH alone and most in those with both conditions. There were progressive increases from the reference group to DM alone, SH alone, and DM + SH with regard to prevalences of LV hypertrophy (12% to 19%, 29% and 38%) and subnormal LV myocardial function (7% to 10%, 11% and 18%, both p <0.001). In conclusion, DM and SH each have adverse effects on LV geometry and function, and the combination of SH and DM results in the greatest degree of LV hypertrophy, myocardial dysfunction, and arterial stiffness.

Relationship of impaired glucose tolerance to left ventricular structure and function: The Strong Heart Study

BACKGROUND

We have identified increased left ventricular (LV) mass, wall thickness, relative wall thickness, and reduced systolic function in diabetic individuals after adjusting for blood pressure and body mass index. However, the cardiovascular correlates of impaired glucose tolerance (IGT), a precursor of diabetes, are unknown.

METHODS

We compared LV measurements between 457 American Indian participants in the Strong Heart Study with IGT (34% men) by World Health Organization criteria and 888 participants (49% men) with normal glucose tolerance.

RESULTS

Participants with IGT were older (60 vs 59 years, P < .01), more overweight (body mass index, 32 +/- 6 vs 29 +/- 5 g/m(2)), and had higher systolic blood pressure (129 +/- 20 vs 124 +/- 18 mm Hg, P < .001) and heart rate (67 +/- 10 vs 66 +/- 11 beats/min, P = .011). In univariate analyses, women but not men with IGT had higher LV mass (mean, 150 vs 138 g, P < .001) and cardiac index (2.6 vs 2.5 L/min/m(2), P < .05). LV wall thicknesses and relative wall thickness were greater in women and men with IGT. Regression analysis, adjusting for multiple covariates in the entire study population, identified independent associations of IGT with higher LV relative wall thicknesses, LV mass/height(2.7), and cardiac output/height(1.83).

CONCLUSIONS

IGT is associated with increased LV wall thickness, mass, and cardiac output independent of effects of relevant covariates.

Reference values for echocardiographic measurements in urban and rural populations of differing ethnicity: the Strong Heart Study

Discrepancies in reported reference values for left ventricular (LV) dimensions and mass may be due to imaging errors with early echocardiographic methods or effects of subject characteristics and inclusion criteria. To determine whether contemporary echocardiographic methods provide stable normal limits for left ventricular measurements in different populations, M-mode/2-dimensional echocardiography was applied in 176 American Indian participants in the Strong Heart Study and 237 New York City residents who were clinically normal. No consistent difference in any measure of LV size or function existed between populations. Upper normal limits (98th percentile) for LV mass were 96 g/m(2) in women and 116 g/m(2) in men and 3.27 cm/m for LV chamber diameter normalized for height. Thus contemporary M-mode/2D echocardiography provides reference ranges for LV measurements that approximate necropsy measurements and have acceptable stability in apparently normal white, African-American/Caribbean, and American Indian populations.

The impact of diabetes on left ventricular filling pattern in normotensive and hypertensive adults: the Strong Heart Study

OBJECTIVES

We sought to determine the effect of diabetes mellitus (DM) on left ventricular (LV) filling pattern in normotensive (NT) and hypertensive (HTN) individuals.

BACKGROUND

Diastolic abnormalities have been extensively described in HTN but are less well characterized in DM, which frequently coexists with HTN.

METHODS

We analyzed the transmitral inflow velocity profile at the mitral annulus in four groups from the Strong Heart Study: NT-non-DM (n = 730), HTN-non-DM (n = 394), NT-DM (n = 616) and HTN-DM (n = 671). The DM subjects were further divided into those with normal filling pattern (n = 107) and those with abnormal relaxation (AbnREL) (n = 447).

RESULTS

The peak E velocity was lowest in HTN-DM, intermediate in NT-DM and HT-non-DM and highest in the NT-non-DM group (p < 0.001), with a reverse trend seen for peak A velocity (p < 0.001). In multivariate analysis, E/A ratio was lowest in HTN-DM and highest in NT-non-DM, with no difference between NT-DM and HTN-non DM (p < 0.001). Likewise, mean atrial filling fraction and deceleration time were highest in HTN-DM, followed by HTN-non-DM or NT-DM and lowest in NT-non-DM (both p < 0.05). Among DM subjects, those with AbnREL had higher fasting glucose (p = 0.03) and hemoglobin A1C (p = 0.04).

CONCLUSIONS

Diabetes mellitus, especially with worse glycemic control, is independently associated with abnormal LV relaxation. The severity of abnormal LV relaxation is similar to the well-known impaired relaxation associated with HTN. The combination of DM and HTN has more severe abnormal LV relaxation than groups with either condition alone. In addition, AbnREL in DM is associated with worse glycemic control.

A population-based assessment of left ventricular systolic dysfunction in middle-aged and older adults: the Strong Heart Study

BACKGROUND

Although clinical congestive heart failure (CHF) is increasingly common, few data document the prevalence and correlates of underlying left ventricular (LV) systolic dysfunction (D) in population-based samples.

METHODS

Echocardiography was used in the second Strong Heart Study (SHS) examination to identify mild and severe LVD (LV ejection fraction [EF] 40%-54% and <40%, respectively) in 3184 American Indians.

RESULTS

Mild and severe LVD were more common in men than women (17.4% vs 7.2% and 4.7% vs 1.8%) and in diabetic than nondiabetic participants (12.7% vs 9.1% and 3.5% vs 1.6%). Stepwise increases were observed from participants with normal EF to those with mild and severe LVD in age (mean 60 vs 61 and 63 years, P <.001), prevalence of overt CHF (2% vs 6% and 28%) and definite coronary heart disease (3% vs 11% and 32%), systolic pressure (129 vs 135 and 136 mm Hg), serum creatinine level (0.98 vs 1.34 and 2.16 mg/dL), and log urinary albumin/creatinine level (3.2 vs 3.7 and 4.7); a negative relation was seen with body mass index (31.1 vs 31.0 and 28.4 kg/m(2)) (all P <.001). In multivariate analyses lower LVEFs were independently associated with clinical CHF and coronary heart disease, lower myocardial contractility, male sex, hypertension, overweight, arterial stiffening (higher pulse pressure/stroke volume) and renal dysfunction (higher serum creatinine level), higher LV mass, and lower relative wall thickness.

CONCLUSIONS

LVD, present in approximately 14% of middle-aged to elderly adults, is independently associated with overt heart failure and coronary heart disease, male sex, hypertension, overweight, arterial stiffening, and renal target organ damage and, less consistently, with older age and diabetes.

Relations of stroke volume and cardiac output to body composition: the strong heart study

BACKGROUND

Although cardiac output (CO) plays the vital role of delivering nutrients to body tissues, few data are available concerning the relations of stroke volume (SV) and CO to body composition in large population samples.

METHODS AND RESULTS

Doppler and 2D echocardiography and bioelectric impedance in 2744 Strong Heart Study participants were used to calculate SV and CO and to relate them to fat-free body mass (FFM), adipose mass, and demographic variables. Both SV and CO were higher in men than women and in overweight than normal-weight individuals, but these differences were diminished or even reversed by normalization for FFM or body surface area. In both sexes, SV and CO were more strongly related to FFM than adipose mass, other body habitus measures, arterial pressure, diabetes, or age. In multivariate analyses using the average of Doppler and left ventricular SV to minimize measurement variability, FFM was the strongest correlate of SV and CO; other independent correlates were adipose mass, systolic pressure, diabetes, age, and use of digoxin and calcium channel and beta-blockers.

CONCLUSIONS

In a population-based sample, SV and CO are more strongly related to FFM than other variables; increased FFM may be the primary determinant of increased SV and CO in obesity.

Prevalence and correlates of mitral regurgitation in a population-based sample (the Strong Heart Study)

Evidence suggesting that mitral regurgitation (MR) may be induced by appetite suppressant medications heightens the importance of understanding the prevalence and correlates of MR, especially its relation to obesity, in population-based samples. MR was assessed by color Doppler echocardiography in 3,486 American Indian participants in the Strong Heart Study. Mild (1+) MR was present in 19.2%, moderate (2+) MR in 1.6%, moderately severe (3+) in 0.3%, and severe (4+) in 0.2% of participants. In univariate analyses, MR was unrelated to gender, diabetes, or lipid levels, but was more frequent in North/South Dakota (28.3%) than in Oklahoma (21.6%) or Arizona (14.3%) (p <0.001). MR was related to lower body mass index (BMI) (p <0.001), older age (p <0.001), higher systolic blood pressure (p = 0.003), higher serum creatinine (p <0.001), and higher urine albumin/creatinine ratio (p <0.001). In multivariate analyses, the presence and severity of MR were independently associated with higher serum creatinine, lower BMI, mitral stenosis, prior myocardial infarction, female gender, mitral valve prolapse and, variably, older age. In conclusion, MR, mostly mild, is detected by color Doppler echocardiography in >20% of middle-aged and older adults. MR is independently associated with female gender, lower BMI, older age, and renal dysfunction, as well as with prior myocardial infarction, mitral stenosis, and mitral valve prolapse. It is not related to dyslipidemia or diabetes.

Congestive heart failure despite normal left ventricular systolic function in a population-based sample: the Strong Heart Study

In selected clinical series, > or = 50% of adults with congestive heart failure (CHF) do not have left ventricular (LV) systolic dysfunction. Little is known of the prevalence of this phenomenon in population samples. Therefore, clinical examination and echocardiography were used in the second examination of the Strong Heart Study (3,184 men and women, 47 to 81 years old) to identify 95 participants with CHF, 50 of whom had normal LV ejection fraction (EF) (> 54%), 19 of whom had mildly reduced EF (40% to 54%), and 26 of whom had EF < or = 40%. Compared with those with no CHF, participants with CHF and no, mild, or severe decrease in EF had higher creatinine levels (2.34 to 2.85 vs 1.01 mg/dl, p < 0.001) and higher prevalences of diabetes (60% to 70% vs 50%) and hypertension (75% to 96% vs 46%, p < 0.05). Compared with those with no CHF, participants with CHF and normal EF had prolonged deceleration time (233 vs 204 ms, p < 0.05) and a reduced E/A, whereas those with CHF and EF < or = 40% had short deceleration time (158 ms, p < 0.05) and high E/A (1.70, p < 0.001); patients with CHF and normal EF had higher LV mass (98 vs 84 g/m2, p < 0.001) and relative wall thickness (0.37 vs 0.35, p < 0.05) than those without CHF. Patients with CHF with normal EF were, compared with those without CHF or with CHF and EF < or = 40%, disproportionately women (mean 84% vs 63% and 42%, p < 0.001), older (mean 64 vs 60 years and 63 years, respectively, p < 0.01), had higher body mass index (mean 33.1 vs 31.0 and 27.7 kg/m2, p < 0.05), and higher systolic blood pressure (mean 137 vs 130 and 128 mm Hg, both p < 0.05). Thus, in a population-based sample, patients with CHF and normal LV EF were older and overweight, more often women, had renal dysfunction, impaired early diastolic LV relaxation, and concentric LV geometry, whereas patients with CHF and severe LV dysfunction were more often men, had lower body mass index, a restrictive pattern of LV filling, and eccentric LV hypertrophy.

Effects of obesity and body fat distribution on lipids and lipoproteins in nondiabetic American Indians: The Strong Heart Study

OBJECTIVES

To examine the relationship between obesity and lipoprotein profiles and compare the effects of total obesity and central adiposity on lipids/lipoproteins in American Indians.

RESEARCH METHODS AND PROCEDURES

Participants were 773 nondiabetic American Indian women and 739 men aged 45 to 74 years participating in the Strong Heart Study. Total obesity was estimated using body mass index (BMI). Central obesity was measured as waist circumference. Lipoprotein measures included triglycerides, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, apolipoprotein AI (apoAI), and apolipoprotein B (apoB). Partial and canonical correlation analyses were used to examine the associations between obesity and lipids/ lipoproteins.

RESULTS

Women were more obese than men in Arizona (median BMI 32.1 vs. 29.2 kg/m2) and South Dakota and North Dakota (28.3 vs. 28.0 kg/m2), but there was no sex difference in waist circumference. Men had higher apoB and lower apoAI levels than did women. In women, when adjusted for center, gender, and age, BMI was significantly related to HDL cholesterol (r = -0.24, p < 0.001). There was a significant but weak relation with apoAI (r = -0.14, p < 0.001). Waist circumference was positively related to triglycerides (r = 0.14, p < 0.001) and negatively related to HDL cholesterol (r = -0.23, p < 0.001) and apoAI (r = -0.13, p < 0.001). In men, BMI was positively correlated with triglycerides (r = 0.30, p < 0.001) and negatively correlated with HDL cholesterol (r = -0.35, p < 0.001) and apoAI (r = -0.23, p < 0.001). Triglycerides increased with waist circumference (r = 0.30, p < 0.001) and HDL cholesterol decreased with waist circumference (r = -0.36, p < 0.001). In both women and men there was an inverted U-shaped relationship between obesity and waist with LDL cholesterol and apoB. In canonical correlation analysis, waist circumference received a greater weight (0.86) than did BMI (0.17) in women. However, the canonical weights were similar for waist (0.46) and BMI (0.56) in men. Only HDL cholesterol (-1.02) carried greater weight in women, whereas in men, triglycerides (0.50), and HDL cholesterol (-0.64) carried a large amount of weight. All the correlation coefficients between BMI, waist circumference, and the first canonical variable of lipids/lipoproteins or between the individual lipid/lipoprotein variables and the first canonical variable of obesity were smaller in women than in men. Triglycerides and HDL cholesterol showed clinically meaningful changes with BMI and waist circumference in men. All lipid/lipoprotein changes in women in relation to BMI and waist circumference were minimal.

DISCUSSION

The main lipoprotein abnormality related to obesity in American Indians was decreased HDL cholesterol, especially in men. Central adiposity was more associated with abnormal lipid/lipoprotein profiles than general obesity in women; both were equally important in men.

Prevalence and correlates of aortic regurgitation in American Indians: the Strong Heart Study

OBJECTIVES

We sought to determine the prevalence and correlates of aortic regurgitation (AR) in a population-based sample group.

BACKGROUND

Concern over induction of AR by weight loss medication highlights the importance of assessing the prevalence and correlates of AR in unselected patient groups.

METHODS

Aortic regurgitation was assessed by color flow Doppler echocardiography in 3,501 American Indian participants age 47 to 81 years during the second Strong Heart Study.

RESULTS

Mild (1+) AR was present in 7.3%, 2+ AR in 2.4% and 3+ to 4+ AR in 0.3% of participants, more frequently in those > or =60 years old than in those <60 years old (14.4% vs. 5.8%, p<0.001); AR was unrelated to gender. Compared with participants without AR, those with mild AR had a lower body mass index (p<0.004) and higher systolic pressure (p<0.003). Participants with AR had larger aortic root diameters (3.6+/-0.4 vs. 3.4+/-0.4 cm, p<0.001), higher creatinine levels (1.3+/-1.3 vs. 1.0+/-1.0 mg/dl, p<0.001) and higher urine albumin/creatinine levels (3.6+/-2.3 vs. 3.3+/-2.0 log, p<0.001), as well as higher prevalences of aortic stenosis (AS) or mitral stenosis (MS) (p<0.001). Regression analysis showed that AR was independently related to older age and larger aortic roots (p<0.0001), AS and absence of diabetes (p = 0.002), MS (p = 0.003) and higher log urine albumin/creatinine (p = 0.005).

CONCLUSIONS

Aortic regurgitation occurred in 10% of a sample group of middle-aged to older adults and was related to older age, larger aortic root diameter, aortic and mitral stenosis and albuminuria. There was no association of AR with being overweight and a negative association of AR with diabetes.

Relation of generalized and central obesity to cardiovascular risk factors and prevalent coronary heart disease in a sample of American Indians: the Strong Heart Study

OBJECTIVE

To examine the hypothesis linking measures of obesity including body mass index (BMI), waist circumference (waist) and percentage body fat to coronary heart disease (CHD) prevalence and its risk factors in American Indians.

DESIGN

The Strong Heart Study assesses the prevalence of CHD and its risk factors in American Indians in Arizona, Oklahoma and South/North Dakota. Participants underwent a physical examination and an electrocardiogram; anthropometric and blood pressure measurements were taken, as were measurements of glucose, lipoproteins, fibrinogen, insulin, hemoglobin A1c and urinary albumin.

PARTICIPANTS

Data were available for 4549 men and women between 45 and 74 y of age.

MEASUREMENTS

Obesity, measured using body mass index, waist circumference and percentage body fat, was correlated with prevalent CHD and its risk factors.

RESULTS

More than 75% of participants were overweight (BMI>25 kg/m2). Measures of obesity were greater in women than in men, in younger than in older participants, and in participants with diabetes than in nondiabetic participants. CHD risk factors were associated with measures of obesity but, except for insulin concentration, changes in metabolic variables with increasing obesity were small. Associations were not stronger with waist than with BMI. The prevalence of CHD in those whose BMI and/or waist measurements lay in the lowest and highest quintiles, by gender and diabetic status, was similar.

CONCLUSIONS

Although CHD risk factors are associated with obesity in American Indians, distribution of obesity (ie waist) is no more closely related to risk factors than is generalized obesity (ie BMI), and changes in CHD risk factors with obesity were small. Thus, the relations among obesity, body fat distribution and CHD risk may differ in this population.

Low national breast and cervical cancer-screening rates in American Indian and Alaska Native women with diabetes

BACKGROUND

The cervical cancer mortality rate for American Indian and Alaska Native women is twice that of all races in the United States. To date the only published national breast and cervical cancer-screening rates for American Indian and Alaska Native women are based on self-reported data. When the Indian Health Service (IHS) conducts an annual audit on patients with diabetes, it includes cancer screening. This observational study presents national breast and cervical cancer-screening rates for American Indian and Alaska Native women with diabetes.

METHODS

Cancer-screening rates were extracted from the 1995 diabetic audit for the 12 IHS areas. These rates were compared with rates for women without diabetes of the same age, 50 to 69 years, by chart review, at four IHS hospitals in the Aberdeen IHS area.

RESULTS

Screening rates for women with diabetes in the 12 areas varied: mammogram (ever) 35% to 78%; clinical breast examination (last year) 28% to 70%, and Papanicolaou smear (last year) 26% to 69%. The Aberdeen IHS area women with diabetes had 51% more clinic visits per year than women without diabetes, but the groups had similar screening rates.

CONCLUSION

Cancer-screening rates for American Indian and Alaska Native women vary by region. In the Aberdeen IHS area, women with diabetes had more visits (missed opportunities) but similar screening rates as women without diabetes. The diabetic audit could be used to monitor national IHS cancer-screening trends for women with diabetes and in the Aberdeen IHS area for all women aged 50 to 69 years.

Differences in echocardiographic findings and systemic hemodynamics among non-diabetic American Indians in different regions: The Strong Heart Study

PURPOSE

This study was undertaken to determine whether differences in left ventricular (LV) and systemic hemodynamic findings exist between American Indians in different regions that might contribute to known differences in cardiovascular morbidity rates among American Indians.

METHODS

We compared echocardiography results in 290 non-diabetic Strong Heart Study (SHS) participants in Arizona, 595 in Oklahoma and 572 in North/South Dakota (ND/SD).

RESULTS

Participants in the 3 regions were similar in age and gender but those in Arizona had the highest body mass indices and lowest heart rates while those in ND/SD had the lowest diastolic blood pressures (BP). In analyses that adjusted for significant covariates, ND/SD participants had larger aortic (Ao) anular, Ao root, and LV chamber size as well as higher cardiac output and lower peripheral resistance, whereas Arizona participants had increased LV wall thickness and mass and reduced LV myocardial contractility. These findings may contribute to the known high rates of cardiovascular events in ND/SD Indians and to the proportionately higher rate of cardiovascular death than of non-fatal cardiovascular events that has been recently documented in Arizona Indians.

CONCLUSIONS

Application of echocardiography to non-diabetic SHS participants reveals that LV chamber and arterial size are larger in ND/SD Indians and that LV wall thicknesses and mass are higher and LV myocardial contractility lower in Arizona Indians, possibly contributing to the higher than expected rates of cardiovascular morbidity and mortality among Indians in Arizona.

Impact of diabetes on cardiac structure and function: the strong heart study

BACKGROUND

Whether diabetes mellitus (DM) adversely affects left ventricular (LV) structure and function independently of increases in body mass index (BMI) and blood pressure is controversial.

METHODS AND RESULTS

Echocardiography was used in the Strong Heart Study, a study of cardiovascular disease in American Indians, to compare LV measurements between 1810 participants with DM and 944 with normal glucose tolerance. Participants with DM were older (mean age, 60 versus 59 years), had higher BMI (32.4 versus 28.9 kg/m(2)) and systolic blood pressure (133 versus 124 mm Hg), and were more likely to be female, to be on antihypertensive treatment, and to live in Arizona (all P<0.001). In analyses adjusted for covariates, women and men with DM had higher LV mass and wall thicknesses and lower LV fractional shortening, midwall shortening, and stress-corrected midwall shortening (all P<0.002). Pulse pressure/stroke volume, a measure of arterial stiffness, was higher in participants with DM (P<0.001 independent of confounders).

CONCLUSIONS

Non-insulin-dependent DM has independent adverse cardiac effects, including increased LV mass and wall thicknesses, reduced LV systolic chamber and myocardial function, and increased arterial stiffness. These findings identify adverse cardiovascular effects of DM, independent of associated increases in BMI and arterial pressure, that may contribute to cardiovascular events in diabetic individuals.

LDL cholesterol as a strong predictor of coronary heart disease in diabetic individuals with insulin resistance and low LDL: The Strong Heart Study

Diabetes has been shown to increase the risk of coronary heart disease in all populations studied. However, there is a lack of information on the relative importance of diabetes-associated risk factors for cardiovascular disease (CVD), especially the role of lipid levels, because low density lipoprotein (LDL) cholesterol often is not elevated in diabetic individuals. The objective of this analysis was to evaluate CVD risk factors in a large cohort of diabetic individuals and to compare the importance of dyslipidemia (ie, elevated triglycerides and low levels of high density lipoprotein [HDL] cholesterol) and LDL cholesterol in determining CVD risk in diabetic individuals. The Strong Heart Study assesses coronary heart disease and its risk factors in American Indians in Arizona, Oklahoma, and South/North Dakota. The baseline clinical examinations (July 1989 to January 1992) consisted of a personal interview, physical examination, and drawing of blood samples for 4549 study participants (2034 with diabetes), 45 to 74 years of age. Follow-up averaged 4.8 years. Fatal and nonfatal CVD events were confirmed by standardized record review. Participants with diabetes, compared with those with normal glucose tolerance, had lower LDL cholesterol levels but significantly elevated triglyceride levels, lower HDL cholesterol levels, and smaller LDL particle size. Significant independent predictors of CVD in those with diabetes included age, albuminuria, LDL cholesterol, HDL cholesterol (inverse), fibrinogen, and percent body fat (inverse). A 10-mg/dL increase in LDL cholesterol was associated with a 12% increase in CVD risk. Thus, even at concentrations well below the National Cholesterol Education Program target of 130 mg/dL, LDL cholesterol is a strong independent predictor of coronary heart disease in individuals with diabetes, even when components of diabetic dyslipidemia are present. These results support recent recommendations for aggressive control of LDL cholesterol in diabetic individuals, with a target level of <100 mg/dL.

Prevalence of undiagnosed diabetes in three American Indian populations. A comparison of the 1997 American Diabetes Association diagnostic criteria and the 1985 World Health Organization diagnostic criteria: the Strong Heart Study

OBJECTIVE

In 1997, the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus of the American Diabetes Association (ADA) recommended three new sets of criteria for the diagnosis of diabetes that were different from those established by the World Health Organization (WHO) in 1985. One of these three methods was based on a fasting plasma glucose value only. This article compares ADA criteria with WHO criteria by applying them to three subgroups of American Indians in the Strong Heart Study who had no known diabetes.

RESEARCH DESIGN AND METHODS

The Strong Heart Study is a prospective epidemiological study of vascular disease in three American Indian populations aged 45-74 years. During the baseline examination from 1988 to 1991, participants without diagnosed diabetes underwent a fasting glucose test and a 2-h oral glucose tolerance test. These values were used to compare the ADA and WHO diagnostic criteria.

RESULTS

By using fasting and 2-h glucose values, prevalence rates of undiagnosed diabetes were 15.9% according to WHO criteria and 14.4% according to ADA criteria. The overall agreement rate was 65%, and the weighted kappa statistic was 0.474, which indicates moderate agreement. The age-specific analysis showed that, among participants between 45 and 54 years of age, the prevalence rates of undiagnosed diabetes were 13.4% according to WHO criteria and 12.7% according to ADA criteria. Among those aged 55-74 years, the rates were 18.7% according to WHO criteria and 16.3% according to ADA criteria. Thus, the difference in the prevalence rates when using WHO and ADA criteria, although generally small in this population, was three times higher in the older group (2.4%) than the difference in the younger group (0.7%).

CONCLUSIONS

The Strong Heart Study found that prevalence rates of undiagnosed diabetes determined by ADA criteria and WHO criteria were similar in its American Indian population. The data suggest that the difference between the two criteria may increase as age increases. Longitudinal data will be needed to evaluate further the utility of the two criteria.

Assessment of QT interval and QT dispersion for prediction of all-cause and cardiovascular mortality in American Indians: The Strong Heart Study

BACKGROUND

Both a prolonged QT interval and increased QT interval dispersion (QTD) have been proposed as surface ECG markers of vulnerability to ventricular arrhythmias and potential predictors of mortality.

METHODS AND RESULTS

The predictive values of QT prolongation and QTD were assessed in 1839 participants in the Strong Heart Study, a prospective study of cardiovascular disease in American Indians. ECGs were acquired at 250 Hz; QT intervals were measured by computer in all 12 leads and corrected for heart rate (QTc) by use of Bazett's formula. QTD was calculated as the difference between the maximum and minimum QTc. After a mean follow-up of 3.7+/-0.9 years, there were 188 deaths from all causes, including 55 cardiovascular deaths. In univariate Cox analyses, prolonged QTc and increased QTD were significant predictors of all-cause mortality (chi(2)=53.0, P<0.0001; chi(2)=11.3, P=0.0008) and cardiovascular mortality (chi(2)=14.7, P=0.0001; chi(2)=26.5, P<0.0001). In multivariate Cox regression analyses controlling for risk factors, QTc remained a strong predictor of all-cause mortality (chi(2)=16.5, P<0.0001) and a weaker predictor of cardiovascular mortality (chi(2)=5.8, P=0.016); QTD remained a significant predictor of cardiovascular mortality only (chi(2)=12.5, P=0.0004).

CONCLUSIONS

These findings support the value of computerized measurements of QTc and QTD in noninvasive risk stratification and suggest that these surface ECG variables may reflect different underlying abnormalities of ventricular repolarization.

Glycemic control in diabetic American Indians. Longitudinal data from the Strong Heart Study

OBJECTIVE

To describe glycemic control and identify correlates of elevated HbA1c levels in diabetic American Indians participating in the Strong Heart Study, which is a longitudinal study of cardiovascular disease in American Indians in Arizona, Oklahoma, South Dakota, and North Dakota.

RESEARCH DESIGN AND METHODS

This analysis is based on data from the baseline (1989-1992) and first follow-up (1994-1995) examinations of the Strong Heart Study. The 1,581 diabetic participants included in this analysis were aged 45-74 years at baseline, were diagnosed with diabetes before and at baseline, and had their HbA1c levels measured at follow-up. HbA1c was used as the index of glycemic control. Characteristics that may affect glycemic control were evaluated for cross-sectional and longitudinal relationships by analysis of covariance and multiple regression.

RESULTS

There was no significant difference between median HbA1c at baseline (8.4%) and at follow-up (8.5%). Sex, age (inversely), and insulin and oral hypoglycemic agent therapy were significantly related to HbA1c levels in both the cross-sectional and longitudinal analyses. Current smoking, prior use of alcohol, and duration of diabetes were significant only for the cross-sectional data. Baseline HbA1c significantly and positively predicted HbA1c levels at follow-up. Comparison of HbA1c by therapy type shows that insulin therapy produced a significant decrease in HbA1c between the baseline and follow-up examinations.

CONCLUSIONS

Glycemic control was poor among diabetic American Indians participating in the Strong Heart Study. Women, patients taking insulin or oral hypoglycemic agents, and younger individuals had the worst control of all the participants. Baseline HbA1c, and weight loss predicted worsening of control, whereas insulin therapy predicted improvement in control. Additional therapies and/or approaches are needed to improve glycemic control in this population.

Rising tide of cardiovascular disease in American Indians. The Strong Heart Study

BACKGROUND

Although cardiovascular disease (CVD) used to be rare among American Indians, Indian Health Service data suggest that CVD mortality rates vary greatly among American Indian communities and appear to be increasing. The Strong Heart Study was initiated to investigate CVD and its risk factors in American Indians in 13 communities in Arizona, Oklahoma, and South/North Dakota.

METHODS AND RESULTS

A total of 4549 participants (1846 men and 2703 women 45 to 74 years old) who were seen at the baseline (1989 to 1991) examination were subjected to surveillance (average 4.2 years, 1991 to 1995), and 88% of those remaining alive underwent a second examination (1993 to 1995). The medical records of all participants were exhaustively reviewed to ascertain nonfatal cardiovascular events that occurred since the baseline examination or to definitively determine cause of death. CVD morbidity and mortality rates were higher in men than in women and were similar in the 3 geographic areas. Coronary heart disease (CHD) incidence rates among American Indian men and women were almost 2-fold higher than those in the Atherosclerosis Risk in Communities Study. Significant independent predictors of CVD in women were diabetes, age, obesity (inverse), LDL cholesterol, albuminuria, triglycerides, and hypertension. In men, diabetes, age, LDL cholesterol, albuminuria, and hypertension were independent predictors of CVD.

CONCLUSIONS

At present, CHD rates in American Indians exceed rates in other US populations and may more often be fatal. Unlike other ethnic groups, American Indians appear to have an increasing incidence of CHD, possibly related to the high prevalence of diabetes. In the general US population, the rising prevalence of obesity and diabetes may reverse the decline in CVD death rates. Therefore, aggressive programs to control diabetes and its risk factors are needed.

Validation of a self-administered questionnaire to screen for prenatal alcohol use in Northern Plains Indian women

OBJECTIVE

This study among American Indian prenatal patients was conducted to validate a self-administered questionnaire (SAQ) designed to (1) identify women who had consumed alcohol during pregnancy, (2) identify women who may be at risk of drinking during pregnancy, and (3) determine the quantity and frequency of alcohol and other substance use just before and during pregnancy.

METHODS

The validation involved three components: (1) review of the SAQ responses by a public health nurse; (2) structured patient interview with the research nurse; and (3) medical record abstraction postpartum.

RESULTS

Compared to extensive interview and medical record data, the SAQ is sensitive (76.6%) and specific (92.8%) in detecting pregnant women who had consumed alcohol during pregnancy.

CONCLUSION

The SAQ is a useful screening tool for alcohol use in this population.

Prevalence of peripheral arterial disease and associated risk factors in American Indians: the Strong Heart Study

Studies of peripheral arterial disease (PAD) in minority populations provide researchers with an opportunity to evaluate PAD risk factors and disease severity under different types of conditions. Examination 1 of the Strong Heart Study (1989-1992) provided data on the prevalence of PAD and its risk factors in a sample of American Indians. Participants (N = 4,549) represented 13 tribes located in three geographically diverse centers in the Dakotas, Oklahoma, and Arizona. Participants in this epidemiologic study were aged 45-74 years; 60% were women. Using the single criterion of an ankle brachial index less than 0.9 to define PAD, the prevalence of PAD was approximately 5.3% across centers, with women having slightly higher rates than men. Factors significantly associated with PAD in univariate analyses for both men and women included age, systolic blood pressure, hemoglobin A1c level, albuminuria, fibrinogen level, fasting glucose level, prevalence of diabetes mellitus, and duration of diabetes. Multiple logistic regression analyses were used to predict PAD for women and men combined. Age, systolic blood pressure, current cigarette smoking, pack-years of smoking, albuminuria (micro- and macro-), low density lipoprotein cholesterol level, and fibrinogen level were significantly positively associated with PAD. Current alcohol consumption was significantly negatively associated with PAD. In American Indians, the association of albuminuria with PAD may equal or exceed the association of cigarette smoking with PAD.

Relations of left ventricular mass to fat-free and adipose body mass: the strong heart study. The Strong Heart Study Investigators

BACKGROUND

It is unclear whether increased left ventricular (LV) mass in overweight individuals is related to their adiposity or to greater fat-free mass (FFM).

METHODS AND RESULTS

We compared echocardiographic LV mass to FFM and adipose body mass by bioelectric impedance and to anthropometric measurements in 3107 American Indian participants in the Strong Heart Study. In men and women, the relations of LV mass and FFM (r=0.37 and 0.38, P<0.001) were closer (P<0.05 to <0.001) than they were with adipose mass, waist/hip ratio, body mass index, systolic blood pressure, height, or height2.7. Regression analyses showed that in men LV mass had the strongest independent relation with FFM, followed by systolic blood pressure and age (all P<0.001); in women, LV mass was related to FFM more strongly than it was to systolic blood pressure, age (all P<0. 001), and diabetes (P=0.012). Adipose mass had no independent relation to LV mass. When waist/hip ratio or body mass index were substituted for adipose mass, LV mass was independently related to FFM (P<0.001) and body mass index (P=0.02) but not to waist/hip ratio in men and was independently related to FFM and waist/hip ratio (both P<0.001) but not to body mass index in women. Using 97.5 percentile gender-specific partitions for LV mass/FFM in reference individuals, we found that LV hypertrophy occurred in 20.8% of Strong Heart Study participants with hypertension, overweight, or diabetes compared with 10.5% and 16.7% by LV mass indexed for body surface area or height2.7.

CONCLUSIONS

LV mass is more strongly related to FFM than to adipose mass, waist/hip ratio, body mass index, or height-based surrogates for lean body weight; LV mass/FFM criteria may increase sensitivity to detect LV hypertrophy.

Plasma lipoproteins and the incidence of abnormal excretion of albumin in diabetic American Indians: the Strong Heart Study

Animal studies suggest that lipids are risk factors for kidney diseases. Some prospective studies and clinical trials have reported predictive effects of lipoproteins on different stages of diabetic nephropathy in humans. We examined lipoprotein abnormalities to determine if they predict abnormal urinary excretion of albumin (> or = 30 mg albumin/g creatinine), using logistic regression. We followed 671 American Indians (211 men, 460 women) with Type II diabetes for a mean of 3.9 years (range 1.7-6.2). Participants were aged 45-74 years. They had normal excretion of albumin and normal serum creatinine at baseline. 67 men and 144 women developed abnormal excretion of albumin. In models controlled for age, treatment with oral hypoglycaemic agents or insulin, HbA1c, study site, degree of Indian heritage, mean arterial blood pressure, albumin excretion at baseline and duration of diabetes, a high HDL cholesterol was a protector for abnormal excretion of albumin in women [odds ratio (OR) comparing the 90th with the 10th percentile = 0.56, 95% confidence interval (CI) = 0.32-0.98], but not in men (OR = 1.5, 95% CI = 0.66-3.4). Further adjustment for obesity, insulin concentration, alcohol consumption or physical activity did not change the results. There was a tendency for high values of VLDL and total triglyceride and small LDL size to predict abnormal excretion of albumin in women only. We conclude that low HDL cholesterol was a risk factor for abnormal excretion of albumin in women, but not in men. Sex hormones may be responsible for sex differences in the association between HDL cholesterol and abnormal excretion of albumin.

Adverse effects of diabetes on multiple cardiovascular disease risk factors in women. The Strong Heart Study

OBJECTIVE

Many studies have shown that diabetes increases the risk of cardiovascular disease (CVD) in women to a greater extent than in men. One explanation could be that diabetes has more adverse effects on CVD risk factors in women than in men. We compared diabetes-associated differences in CVD risk factors in men and women in the Strong Heart Study, a population-based study of CVD and its risk factors in American Indians.

RESEARCH DESIGN AND METHODS

A total of 1,846 men and 2,703 women between the ages of 45 and 74 years from 13 American Indian communities in three geographic areas underwent an examination that included a medical history; an electrocardiogram; anthropometric and blood pressure measurements; an oral glucose tolerance test; and measurements of fasting plasma lipoproteins, fibrinogen, insulin, HbA1c, and urinary albumin.

RESULTS

Statistically significantly greater adverse differences in those with diabetes versus those without diabetes were observed in women than in men for waist-to-hip ratio, HDL cholesterol, apolipoprotein (apo)B, apoA1, fibrinogen, and LDL size. In multiple linear regression models adjusting for age, center, sex, and diabetes, the diabetes by sex interaction terms were statistically significant for waist-to-hip ratio, LDL cholesterol, HDL cholesterol, apoB, apoA1, fibrinogen, and LDL size.

CONCLUSIONS

Compared with diabetes-associated differences in men, diabetes in women was related to greater adverse differences in levels of several CVD risk factors. Although the magnitude of the individual diabetes-related differences between men and women was not large, the combined effects of these risk factor differences in diabetic women may be substantial. The apparent greater negative impact of diabetes on CVD risk factors in women may explain, in part, the greater risk for CVD in diabetic women.