Distinct Component Profiles and High Risk Among African Americans With Metabolic Syndrome: The Jackson Heart Study

The JHS is a prospective population-based investigation of the predictors and outcomes of CVD and related disorders among African Americans. The original design and sampling plan for the JHS, which have been described in detail previously (14), included only those aged 35–84 years. The current JHS cohort comprises 5,302 participants in total, consisting of 5,036 adults aged 35–84 years and an additional 251 aged 21–34 years and 15 aged >85 years. The latter group entered the study as part of the nested JHS Family Study (15,16). Comprehensive baseline information was obtained. Demographics, socioeconomic status data (e.g., educational attainment and annual household income), multiple lifestyle data (including extensive dietary and physical activity information), and other sociocultural and psychological parameters as well as other interview data were obtained by certified interviewers through home interviews and during the clinic visit. Additionally, certified technicians and nurses conducted clinic interviews and testing that included medical history, anthropometrics, vital signs, venipuncture (for a comprehensive panel of analyses), electrocardiogram (ECG), echocardiography and carotid ultrasonography, and other evaluations.

Samples of glucose and lipids were collected according to standardized procedures and processed at the Central Laboratory (University of Minnesota) described in previous publications (14,15). Sitting blood pressure was measured twice at 5-min intervals; the average of two measurements was used for analysis. Two measures of the waist (at the level of the umbilicus and in the upright position) were averaged to determine baseline waist circumference for each participant. For the present analysis, we excluded individuals who did not complete data collection on any components of metabolic syndrome.

The ATP III has identified five components of metabolic syndrome that related to CVD (abdominal obesity, high triglyceride concentration, low HDL cholesterol concentration, elevated blood pressure, and elevated fasting glucose concentration) (10). In consonance with the ATP III definition, the subjects were considered to have metabolic syndrome if they had three or more of following: 1) abdominal obesity, defined by waist circumference ≥102 cm (40 inches) in men and ≥88 cm (35 inches) in women; 2) elevated blood pressure, defined as a systolic blood pressure of 130 mmHg or greater and/or a diastolic blood pressure of 85 mmHg or greater, physician diagnosis, and/or use of antihypertensive drugs; 3) elevated plasma triglyceride concentration (≥150 mg/dl) or treatment with fibrate medication; 4) Low HDL cholesterol concentration (men <40 mg/dl and women <50 mg/dl); and 5) elevated fasting glucose concentration (≥100 mg/dl) or diagnosed type 2 diabetes. Only individuals who contributed fasting blood specimens, with exception of those under treatment for type 2 diabetes, were included in this report.

Prevalent coronary heart disease (CHD) was defined as evidence of a previous myocardial infarction by ECG based on Minnesota Code criteria (Codes 1.1 and 1.2 plus 4.1–4.2, or 5.1–5.2), a history of physician-diagnosed myocardial infarction, or either percutaneous coronary intervention or coronary bypass surgery. Cerebrovascular disease (CBD) was defined as a history of CBD (by personal history, CBD signs and symptoms ascertained by standardized questionnaires), transient ischemic attack, or carotid endarterectomy and/or angioplasty. CVD was defined as having either CHD or CBD. All medical history obtained at baseline was self-reported.

To determine the cross-sectional association of metabolic syndrome with morbid vascular conditions, we calculated the odds of prevalent cardiovascular disease manifestations (CHD, CBD, and CVD) in those with metabolic syndrome compared with the reference group, defined as participants who did not have metabolic syndrome.

Educational attainment was divided into four categories: less than high school education), high school or GED, more than high school but less than college, and college or greater. Annual incomes were divided into five categories according to the participant's household income for the previous year: <$16,000, $16,000–24,999, $25000–49,999, $50,000–99,999, and >$100,000. Alcohol consumption and cigarette smoking were analyzed as current practice (self-reported as user in past 12 months versus nonuser).

Data were analyzed using the SAS system for Windows, version 9.1 (SAS Institute, Cary NC). Contingency tables were compiled to calculate the sample prevalence of metabolic syndrome and its possible components (elevated blood pressure, abdominal obesity, low HDL cholesterol, and elevated glucose and triglyceride concentrations). Prevalence was calculated separately for age, sex, current alcohol consumption and smoking status, marital status, education, and annual household income. The χ² test was performed to assess the significance of subgroup differences.

To assess the prevalence of each metabolic component, age- and sex-specific prevalence for each metabolic component was calculated separately from the entire JHS cohort. Logistic regression models were used to calculate odds ratios (ORs) and their 95% CI to assess the odds of CVD, CHD, and CBD being present in a participant with metabolic syndrome versus in a participant without metabolic syndrome. Statistical tests of hypothesis with two-directional alternatives that resulted in P ≤ 0.05 were considered statistically significant.

The prevalence of metabolic syndrome in baseline JHS participants is summarized in Table 1. The overall (aged 21–94 years) prevalence of metabolic syndrome in the JHS cohort was 39.4%. Large percentages of participants with metabolic syndrome were found to have elevated blood pressure (70.4%), abdominal obesity (64.6%), and low HDL cholesterol concentration (37.2%). Among those aged 35–84 years, the age-adjusted prevalence was 44.8% in women and 33.4% in men (P < 0.0001) (Table 2). Women with metabolic syndrome had a comparatively higher prevalence of abdominal obesity and low HDL cholesterol concentration than men (P < 0.0001 for both comparisons).

For the cohort as a whole, prevalence of elevated blood pressure, abdominal obesity, and elevated triglyceride and glucose levels increased with age whereas low HDL cholesterol concentration was about equally prevalent across all age-groups. The prevalence of elevated triglyceride and low HDL cholesterol concentrations were significantly higher in smokers compared with nonsmokers. In addition, participants with higher educational attainment and household income and absence of alcohol consumption have lower prevalence of metabolic syndrome, elevated blood pressure, abdominal obesity, and elevated glucose levels (Table 2).

The prevalences of CVD, CHD, and CBD among those with metabolic syndrome were 12.8, 8.7, and 5.8%, respectively. In logistic regression analyses adjusted for age and sex, metabolic syndrome–associated ORs for prevalent CVD, CHD, and CBD were about 1.7 (Table 3).

The prevalence of metabolic syndrome reported by different studies in the U.S. has varied widely. Differences in the characteristics of study populations, the definition of metabolic syndrome, study geographic locations, and study time periods have all contributed to this variation. Data from population-based cohorts such as the Framingham Offspring Study and San Antonio Heart Study during the early to mid-1990s reported high prevalence estimates (24 and 30%, respectively) among participants who did not have diabetes (12). Baseline data (obtained 1987–1989) from the Atherosclerosis Risk in Communities Study indicated that the prevalence of metabolic syndrome was 28% in African-American men, 18% in women, and 30% overall in a middle-aged cohort (17). During 1988–1994, NHANES III (a national probability sample) using NCEP ATP III criteria showed that the unadjusted prevalence of metabolic syndrome was 21.8% among the U.S. general population (11). More recently, NHANES III, using consistent sampling methodology and case definition, found a prevalence of 26.7% in the year 2000 (18). Increases in the prevalence of obesity and hypertension and to a lesser degree in hypertriglyceridemia and low HDL cholesterol concentration most likely accounted for much of this increase (18).

Our analyses of recently obtained (2000–2004) baseline data from JHS, the largest population-based epidemiological study of CVD in African Americans ever undertaken, demonstrate extraordinarily high metabolic syndrome prevalence and typical component profiles compared with the general metabolic syndrome literature. Among those aged 35–84 the prevalence was 44.8% in women and 33.4% in men, far above corresponding national rates. The prevalence of metabolic syndrome is significantly and inversely associated with socioeconomic status and is higher in women than in men at similar degrees of education and socioeconomic status. Both in men and women, the prevalence significantly decreases with increases in household income and educational attainment.

Although limited data are available regarding the prevalence of various metabolic syndrome profiles in the African-American population, our study suggests that African Americans may have distinct cardiometabolic risk patterns within the rubric of metabolic syndrome. In studies examining differences in prevalence of metabolic syndrome profiles, abdominal obesity, elevated blood pressure, and elevated triglyceride concentration predominated as metabolic syndrome components in European and Mexican Americans (11,19–22). Among the JHS cohort, however, elevated blood pressure, abdominal obesity, and low HDL cholesterol concentration were more frequent than the triad of conditions most commonly seen in metabolic syndrome among other groups. The higher frequency of low HDL cholesterol concentration is notable for at least two reasons. First, owing to a multitude of studies that have reported high HDL cholesterol concentration for African Americans, there is a prevailing notion that low HDL cholesterol concentration is relatively uncommon among African Americans (11,23,24). However, NHANES highlighted the fact that the race-specific rates of low HDL cholesterol have risen among African Americans from 1989–1994 to 1999–2000 (11,25). Our data seem to reflect this trend. Second, the usual reciprocal elevation in triglyceride concentration that accompanies low HDL cholesterol is not evident and the prevalence of a low HDL cholesterol level stays high across age spectra in our cohort. This finding seems to corroborate indirectly the findings of the Triglyceride and Cardiovascular Risk in African American (TARA) Study of the National Institutes of Health, which reported that insulin-resistant African-American subjects do not demonstrate impairment of postheparin–lipoprotein lipase activity commonly seen with insulin resistance in other ethnic groups. This impairment is inversely associated with serum triglyceride concentration (26). A low frequency of impairment of postheparin–lipoprotein lipase in African Americans with metabolic syndrome (and presumed insulin resistance) would help explain why triglyceride levels remain low, even in a population with high rates of depressed HDL cholesterol levels. This finding could explain our observation that triglyceride concentration abnormalities contribute little to the prevalence of metabolic syndrome in this African-American cohort. In addition, it is worthy to note that the prevalence of metabolic syndrome tends to be lower in current smokers, whose triglyceride level is elevated and HDL cholesterol concentration is depressed. Alcohol drinking also was associated with a lower prevalence of metabolic syndrome with decreased prevalence of low HDL cholesterol concentration. These findings are generally consistent with the expected and widely replicated inverse association of HDL and smoking. The positive association of triglyceride concentration and smoking has also been reported. In contrast, the lower prevalence of metabolic syndrome in those who are not current drinkers suggests that drinking may have modest differential influences on the individual metabolic syndrome components.

Recently, the concept of the metabolic syndrome has been challenged on several fronts. Variation in its definition by different professional and scientific bodies (27), ethnic specificity of some component cut points (28), and the additive rather than synergistic impact of risk factors on CVD outcomes (29) have led to controversy in assigning the label “syndrome” to this grouping of metabolic abnormalities. NCEP ATP III does not mandate any specific metabolic component as a necessary condition for the diagnosis of metabolic syndrome. This issue is controversial, however, and data suggesting that a common antecedent or underlying perturbation in metabolic function produces the syndrome have been reported (30). In the JHS cohort, although elevated blood pressure, abdominal obesity, and low HDL cholesterol concentration are particularly prevalent in metabolic syndrome, the frequency of elevated glucose is still substantial both in men and women, indicating that hyperglycemia is also an important contributor to metabolic syndrome prevalence. This result is consistent with the fact that insulin resistance is the core pathological feature of the metabolic syndrome (31), a notion further supported by other studies (29,32).

Growing evidence for metabolic syndrome as a major cause for both diabetes and CVD raises the question of whether differences in metabolic syndrome prevalence and/or profiles between African and European Americans might help explain racial disparities in the prevalence of diabetes and cardiovascular complications. Although our study does not permit a direct cross-racial comparison, the facts that the JHS cohort demonstrated a high prevalence of CHD, CBD, and overall CVD and the presence of metabolic syndrome increased the odds of these diseases are substantially comparable to the >1.6-fold increase in risk found by others (21,23). Our demonstration of high OR and higher prevalence suggest a high attributable risk in the African-American population for metabolic syndrome. This finding further suggests that control/elimination of metabolic syndrome would result in a substantial reduction in manifest CVD among African Americans.

We also found that the risk for CVD attributable to these cardiometabolic abnormalities is greater in women than in men. Recent mortality data from the San Antonio Heart Study demonstrated that the association between metabolic syndrome and CVD death was more than twice as large in women as in men (33). The authors suggested that a possible explanation for this difference is the stronger diabetes-CVD association in women. In the JHS, increased prevalence of metabolic syndrome in women was associated with their greater proportion of abdominal obesity and low HDL cholesterol concentration. This result could reflect a sex-specific impact of obesity on glucose homeostasis, lipid profiles, and atherosclerosis (34).

The principal limitation of this analysis is the use of cross-sectional data, necessitated by the use of the initial data from this new cohort. Thus, causal pathways underlying the observed relationships cannot be inferred. Also, owing to the design of the initial JHS examination used for this analysis, we relied significantly, although not exclusively, on self-reported medical history data. Although CBD history was obtained exclusively by self-report, the CHD definition (and consequently the CVD definition) combined self-reported data with ECG findings for corroboration and additional case ascertainment (e.g., “silent myocardial infarction”). Another limitation of this analysis is that we were not able to fully account for all potential covariates, such as changes in physical activity, dietary information, and weight gain. Therefore, future research should investigate the impact of these covariates, especially dietary and fitness differences among the participants with or without metabolic syndrome.

Although not a limitation specific to this analysis, it should be noted that the NCEP ATP III cut points for abnormal HDL and triglyceride concentration may lead to an underestimation of metabolic risk among African Americans. As noted above, at least one prior investigation indicated that African Americans, despite being more insulin resistant than their European-American counterparts, paradoxically exhibited lower triglyceride and higher HDL cholesterol concentrations (35). This finding suggests that applying a uniform cut point for both groups for triglyceride and HDL cholesterol concentrations may result in significant misclassification (36). Adjustment of cut point values (using CVD incidence data by triglyceride and HDL concentrations) or measures of lipoprotein subclass profiles (e.g., VLDL, small LDL, and HDL particle concentrations) may provide better estimates of lipid-related metabolic risk in African Americans than sole use of the lipid criteria from the NCEP ATP III (26). However, our data affirm that abdominal obesity, elevated blood pressure, and low HDL cholesterol concentration are key components of metabolic syndrome in the African-American population.

Funding for H.T. was provided under Contracts N01-HC-95170, N01-HC-95171, and N01-C-95172 from the National Heart, Lung, and Blood Institute and the National Center for Minority Health and Health disparities.

The authors thank the staff, interns, and participants in the Jackson Heart Study for their long-term commitment and important contributions to the study. We also thank Robert Garrison, PhD, for his insightful comments on an early draft of manuscript.