Community Prevalence of Ideal Cardiovascular Health, by the American Heart Association Definition, and Relationship With Cardiovascular Disease Incidence

Study design and subjects

The ARIC study is a prospective cohort study of atherosclerotic diseases in 4 U.S. communities: Forsyth County, North Carolina; Jackson, Mississippi; Washington County, Maryland; and the northwest suburbs of Minneapolis, Minnesota (10). The cohort comprised, at baseline in 1987 to 1989, 15,792 men and women age 45 to 64 years who were selected by list or area probability sampling. Only African Americans were recruited in the Jackson study center. The baseline home interview and clinic examination measured various risk factors, health behaviors, and cardiovascular conditions. The ARIC study protocol was approved by the institutional review board of each participating university.

Baseline examination

The ARIC study protocols were described previously (7,8,10). Diet was assessed by a slightly modified, 66-item Harvard food frequency questionnaire (11). We first excluded persons with extreme energy intake of <600 or >4,200 kcal/day for men or <500 or >3,600 kcal/day for women (approximate lower and upper 1 percentiles). We then categorized achievement of the 5 AHA ideal cardiovascular health items: ≥4.5 cups/day of fruits and vegetables (approximated as ≥4.5 servings/day in the ARIC study); ≥3.5 oz servings/week of fish (approximated as ≥2 3- to 5-oz servings/week); ≥3 1-oz servings/day of whole grains (approximated as ≥3 servings/day); sodium (<1,500 mg/day); and ≤36 oz/week of sugar-sweetened beverages (approximated as ≤4 glasses/week). The Baecke questionnaire (12) asked participants to report the frequency of participation in as many as 4 sports and in walking in the previous year; this was converted to minutes per week of moderate or vigorous exercise (13). Smoking status (current, former, or never smokers) was derived from interviews. Use of antihypertensive, cholesterol-lowering, and glucose-lowering medications within the past 2 weeks of baseline interview were self-reported or taken from prescription bottles. Fasting plasma total cholesterol was measured by enzymatic methods. Serum glucose was measured by a hexokinase/glucose-6-phosphate dehydrogenase method. Sitting blood pressure was measured 3 times using a random-zero sphygmomanometer after a 5-min rest. The mean of the last 2 measurements was used for analysis. Body mass index (kg/m²) was computed from weight in a scrub suit and standing height. Using the AHA definitions of cardiovascular health (9), we classified each cardiovascular health metric at baseline into ideal, intermediate, or poor categories, as shown in Table 1.

Pre-existing heart failure at baseline was defined as the following: 1) an affirmative response to “Were any of the medications you took during the last 2 weeks for heart failure?” and 2) stage 3 or “manifest heart failure” by Gothenburg criteria (14,15). Pre-existing coronary heart disease (CHD) at baseline was defined by self-reported previous physician diagnosis of myocardial infarction (MI) or coronary revascularization, or by prevalent MI by 12-lead electrocardiography. Pre-existing stroke was defined by any self-reported previous physician diagnosis of stroke.

Incident CVD events

Incident CVD events comprised heart failure, definite or probable MI, definite fatal CHD, and definite or probable stroke. We followed all participants from the baseline examination in 1987 to 1989 to the date of the CVD event, death, loss to follow-up, or otherwise through December 31, 2007. CVD events in the ARIC study were ascertained by contacting participants annually, identifying hospitalizations and deaths during the previous year, and surveying discharge lists from local hospitals and death certificates from state vital statistics offices for potential CVD events (15–17). Incident heart failure in the ARIC study was defined as the first occurrence of either a hospitalization that included an International Classification of Diseases-9th Revision (ICD-9) discharge code of 428 (428.0 to 428.9) among the primary or secondary diagnoses or else a death certificate with an ICD-9 code of 428 or an ICD-10 code of I50 among the listed or underlying causes of death (15). The ARIC study did not further validate ICD-9 code 428 for heart failure; validation studies using physician review of the ARIC study hospital records beginning in 2005 indicated the positive predictive value of ICD-9 428 to be 93% for acute decompensated heart failure and 97% for chronic heart failure.

For patients hospitalized with a potential MI, trained abstractors recorded the presenting symptoms and related clinical information, including cardiac biomarkers, and photocopied up to 3 12-lead electrocardiograms for Minnesota coding (18). Out-of-hospital deaths were investigated by means of death certificates and, in most cases, by an interview with ≥1 next of kin and a questionnaire completed by the patient's physician. Coroner reports or autopsy reports, when available, were abstracted for use in validation. A CHD event was defined as a validated definite or probable MI requiring hospitalization or a definite CHD death. The criteria for definite or probable MI were based on combinations of chest pain symptoms, electrocardiographic changes, and cardiac enzyme levels (16). The criteria for definite fatal CHD were based on chest pain symptoms, history of CHD, underlying cause of death from the death certificate, and any other associated hospital information or medical history, including that from an ARIC study clinic visit (16).

The diagnostic classification of stroke was described previously (17). In brief, for patients hospitalized for potential strokes, the abstractors recorded signs and symptoms and photocopied neuroimaging (computed tomography or magnetic resonance imaging) and other diagnostic reports. Using criteria adopted from the National Survey of Stroke (19), definite or probable strokes were classified by computer algorithm and separate review by a physician, with disagreements resolved by a second physician.

Statistical analyses

Of 15,792 ARIC study participants at baseline, we excluded, due to small numbers, participants who were neither white nor African-American subjects (n = 48). We further excluded 1,969 who at baseline had a history of heart failure, CHD, or stroke or could not be classified based on history; 417 who were not fasting ≥8 h; 271 had extreme energy intake; and 299 who did not have complete information on risk factors or health behaviors, leaving 12,744 participants.

Prevalences of each cardiovascular health metric at baseline, or combinations of them, were computed. The cumulative incidence of CVD was calculated via a life-table approach. Age-, sex-, and/or race-adjusted incidence rates of CVD were calculated according to cardiovascular health groups using Poisson regression. The hazard ratios of incident CVD in relation to cardiovascular health groups were estimated from Cox proportional hazard models adjusted for age, sex, and sometimes race. The proportional hazards assumption was confirmed by examining whether the ln(-ln) survival curves for the cardiovascular health groups were parallel.

Prevalence of ideal cardiovascular health

The ARIC study cohort free of CVD in 1987 to 1989 had a mean age of 54 years. As shown in Table 1, the proportions of participants in the total ARIC study sample who had ideal levels of individual cardiovascular health metrics were as follows: smoking, 72.2%; body mass index, 34.5%; physical activity, 37.7%; healthy diet, 5.3%; total cholesterol, 37.5%; blood pressure, 43.8%; and fasting plasma glucose, 53.2%. Except for total cholesterol, more whites than African Americans met ideal levels of each metric. The healthy diet components were poorly met, but were met more often for sugar-sweetened beverages (69%) and sodium (58%) than for fruits and vegetables (30%), fish (28%), and whole grains (9%).

As shown in Table 2, only 0.1% of ARIC study participants (n = 17) had all 7 cardiovascular health metrics in the ideal range, and, thus, almost no one had “ideal cardiovascular health.” Approximately 12.2% of participants had 5 to 7 ideal health metrics, and this percentage varied by age, race, and sex: 15.2% for ages 45 to 54 years and 8.8% for ages 55 to 64 years, 3.6% for African-American men, 4.2% for African-American women, 10.5% for white men, and 18.7% for white women.

The prevalence of having “intermediate” cardiovascular health (at least 1 intermediate metric and no poor metrics) was also low: 17.4% (Table 3). In contrast, 82.5% of participants had “poor” cardiovascular health (at least 1 poor health metric). In fact, 27.9% had ≥3 poor health metrics.

Incidence rates of CVD

The median duration of follow-up was 18.7 years (maximum, 21.1 years), during which 3,063 incident CVD events occurred. The first CVD event was CHD in 49%, heart failure in 30%, stroke in 16%, and multiple outcomes simultaneously in 5%. The 17 participants with all 7 factors rated as ideal had no CVD events. In contrast, the CVD incidence rate was 7.5 per 1,000 person-years (95% confidence interval: 6.4 to 8.4) for those with intermediate cardiovascular health, and it was 14.6 (95% confidence interval: 14.0 to 15.2) for those with poor cardiovascular health.

There was a strong gradient of cumulative CVD incidence during follow-up according to the number of ideal health metrics met (Fig. 1). The 2.8% of the ARIC study cohort with 6 ideal factors had a relatively low cumulative CVD incidence of approximately 6%. In contrast, the 2.5% of the ARIC study cohort with no ideal health factors had approximately a 50% cumulative incidence.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Cumulative Incidence of Cardiovascular Disease According to the Number of Ideal Cardiovascular Health Metrics

Cumulative incidence of cardiovascular disease according to the number of ideal cardiovascular health metrics, ARIC Study, 1987 to 2007.

After adjustment for age, race, and sex (Table 4), the overall CVD incidence rate in the ARIC study was 13.3 per 1,000 person-years. The rate was relatively low for those with 5 to 7 ideal health metrics (<7 per 1,000 person-years) and less than one fifth that of those with no ideal health metrics (32.1 per 1,000 person-years). Table 4 also shows an overall higher incidence rate of CVD in African Americans (16.5 per 1,000 person-years) than whites (12.2 per 1,000 person-years). However, within each stratum of ideal cardiovascular health counts, CVD rates were relatively similar between African Americans and white participants.

As shown in Figure 2, both ideal health behaviors (nonsmoking, body mass index, physical activity, healthy diet score) and ideal health factors (total cholesterol, blood pressure, glucose) contributed to lower CVD risk.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Incidence Rate of Cardiovascular Disease According to the Number of Ideal Health Behaviors and Health Factors

Age-, sex-, and race-adjusted incidence rate of cardiovascular disease according to the number of ideal cardiovascular health behaviors (nonsmoking, body mass index, physical activity, healthy diet score) and health factors (total cholesterol, blood pressure, glucose), ARIC (Atherosclerosis Risk in Communities), 1987 to 2007.

Study limitations

Drawbacks of our study warrant consideration. First, the ARIC study sample is community based but not nationally representative. Most of the African Americans were from 1 center, so their lower prevalence of ideal cardiovascular health than that of ARIC study whites, although consistent with national patterns (23), might be due to geographic or socioeconomic differences and should not be attributed to race per se. Furthermore, the ARIC study has no cardiovascular health information on other minority groups. Such information needs to be documented in minority cohort studies and national surveys. Second, although measurement of major risk factors is well standardized and therefore quite generalizable from study to study, measurement of diet and physical activity is not. The instruments that we used were validated (11,12) but brief. For example, “fish” included deep-fried and other types of fish that probably have widely varying health effects. In addition, sodium intake was likely quite underestimated by this brief food frequency questionnaire. Of course, if the ARIC study had asked about additional food items or activities, the prevalences of adults meeting ideal diet metrics might have been different. Nevertheless, our estimates for healthy diet and physical activity were close to national estimates (9). Third, we used a single measure of cardiovascular health. Changes in risk factor levels undoubtedly occurred over 2 decades of follow-up and would have typically led to underestimation of the true biological associations between cardiovascular health metrics and CVD incidence. This underestimation tends to be greater for behavioral factors than for biological risk factors. Fourth, we studied risk factors in the period from 1987 to 1989. Yet, the 1987 to 1989 prevalence estimates for ARIC are not greatly different from those for the U.S. population currently (9). Using a more recent value, for example, ARIC study visit 4 in 1996 to 1998 may have given different estimates of the prevalence of ideal health. Because the cohort would have been 9 years older then, likely the prevalence of ideal cardiovascular health would have been even lower. Finally, using the later ARIC study examination data also would have shortened the follow-up time for CVD events, resulting in poorer precision of incidence rates.