Author + information
- Received September 7, 2005
- Revision received May 16, 2006
- Accepted May 23, 2006
- Published online September 19, 2006.
- Umed A. Ajani, MBBS, MPH⁎ ( and )
- Earl S. Ford, MD, MPH
- ↵⁎Address correspondence and reprints requests to:
Dr. Umed A. Ajani, Centers for Disease Control and Prevention, 4770 Buford Highway NE, Mailstop K-50, Atlanta, Georgia 30341.
Objectives The objective of this study was to compare the 10-year risk of developing coronary heart disease (CHD) among U.S adults during the years 1988 to 1994 with that among U.S. adults during the years 1999 to 2002.
Background A decline in deaths as the result of CHD has been reported. Data about changes in actual risk of developing CHD among U.S. adults are sparse.
Methods Data for noninstitutionalized U.S. residents ages 20 to 79 years who participated in the National Health and Nutrition Examination Survey (NHANES)-III (1988 to 1994) or NHANES 1999 to 2002 were examined to compute 10-year risk of developing CHD using modified Framingham risk score, as adopted by the National Cholesterol Education Program, Adult Treatment Panel III.
Results Most participants in both surveys had a low (<10%) 10-year risk of developing CHD. The proportion of participants at intermediate (10% to 20%) and high (>20%) 10-year risk of developing CHD also was similar.
Conclusions Data from national surveys conducted approximately a decade apart showed no appreciable difference in the distribution of 10-year risk of developing CHD. Greater efforts are needed to reduce the risk of developing CHD among U.S. adults.
The mortality from coronary heart disease (CHD) has declined during last few decades in the U.S. and elsewhere (1–4). Despite the decline, CHD remains the single largest killer of American men and women (5). A substantial portion of this decline has been attributed to changes in risk factors for CHD. Efforts are in place to further reduce the prevalence of modifiable CHD risk factors, such as high blood pressure, high blood cholesterol levels, and smoking. The prevalence of smoking has been declining (6,7), but the prevalence of high blood pressure has increased in the United States (8). However, corresponding data on changes in overall risk for CHD are sparse.
The Framingham Heart Study published multivariate models and a risk equation to estimate 10-year risk of developing CHD (9). The risk score equation, with some modification, was adopted by the National Cholesterol Education Program Expert Panel in their third report on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III; NCEP/ATPIII) (10). The algorithm has been used successfully for CHD risk assessment and formed the basis of clinical guidelines issued by a European task force on coronary prevention and British recommendations on prevention of CHD (11,12). The distribution of 10-year risk for CHD among U.S. adults during the years 1988 to 1994 was reported previously (13). Since then, however, the prevalence of risk factors has changed, and various public health efforts have been made to reduce the burden of CHD. Therefore, it is important to compute updated distribution of CHD risk among U.S. adults and determine whether this distribution has changed. We used data from the National Health and Nutrition Examination Survey (NHANES)-III and NHANES 1999 to 2002 to calculate modified Framingham risk score and to compare the prevalence of 10-year risk of CHD among U.S. adults.
The NHANES collects data from a representative sample of the noninstitutionalized U.S. civilian population using a stratified multistage sampling design. The NHANES-III was conducted from 1988 to 1994 (14). A more recent NHANES study started in 1999 and has completed two 2-year cycles (1999 to 2000 and 2001 to 2002). For this study, we used combined data for 4 years (15,16). The survey received human subject approval from the Centers for Disease Control and Prevention, all participants provided informed consent, and data were approved by the Centers for Disease Control and Prevention’s Institutional Review Board for confidentiality. Participants in NHANES were interviewed by trained interviewers at the participants’ home and were asked to attend an examination center, where they completed additional questionnaires, underwent various examinations, and had blood samples collected.
To estimate participants’ 10-year risk of developing CHD, we used the modified Framingham algorithm as adopted by NCEP/ATPIII, thus including participants aged 20 to 79 years. The modified algorithm included participants’ age, gender, systolic blood pressure, current treatment for high blood pressure, current smoking status, serum total cholesterol concentrations, and serum high-density lipoprotein cholesterol concentrations. Separate risk scores were computed for men and women.
Information about participants’ age, gender, smoking status, and current use of medication for high blood pressure were collected by interview. Systolic blood pressure measurements were recorded during participants’ visit to the examination center. Participants were considered current smokers if they reported smoking at least 100 cigarettes during their life and were currently smoking. Total cholesterol and high-density lipoprotein cholesterol concentrations were measured by enzymatic assays on a Hitachi 704 Analyzer (Roche Diagnostics, Indianapolis, Indiana).
The NCEP/ATPIII considered patients with existing CHD or CHD equivalents in the group at the highest risk for CHD (>20%), the other two categories being 10-year CHD risk of 10% to 20% and <10%. Stroke and diabetes mellitus were considered CHD equivalents in these analyses. Self-reports of heart attack were used to identify CHD cases. A history of stroke was identified by self-reports, and diabetes was identified, as in other studies (4,17), by a fasting blood glucose level ≥126 mg/dl, nonfasting blood glucose ≥200 mg/dl, or participants’ self-reported history of diabetes mellitus.
Our analyses were restricted to participants ages 20 to 79 years who attended the examination center (15,331 participants in NHAHES-III and 8,747 participants in NHANES 1999 to 2002). After excluding participants whose CHD/CHD-equivalent status could not be determined, we were left with 15,143 participants in NHANES-III and 8,726 participants in NHANES 1999 to 2002. Complete information on all risk factors needed to calculate a modified Framingham risk score was available for 13,888 participants for NHANES-III and 7,909 participants in NHANES 1999 to 2002. Of the remaining participants with some missing information (1,255 in NHANES-III and 817 in NHANES 1999 to 2002), we included 818 NHANES-III participants and 542 participants in NHANES 1999 to 2000 because their risk group remained the same using the highest (or worst) possible score for missing components. Therefore, our final study sample consisted of 14,706 participants in NHANES-III and 8,451 participants in NHANES 1999 to 2002.
We categorized the study population from both surveys into three levels of 10-year risk for developing CHD: low (<10%), intermediate (10% to 20%), and high (>20%) and computed the proportions of participants at each risk level separately for men and women. We also present the results for three race/ethnicity groups: white, African American, and Mexican American. The racial/ethnic composition of the fourth group (other) was not clear and because a relatively small number of participants were classified as “other,” estimates for that group were unstable. Therefore, we did not present separate results for this group. Means and proportions were computed separately for each component risk factor. We also computed the proportion of participants at each risk level after excluding those with a CHD or CHD equivalent to estimate the proportion of the population at high risk because of risk factors only. We used sampling weights and Software for the Statistical Analysis of Correlated Data (SUDAAN) (18) to compute population estimates of CHD risk distribution that accounted for the complex sampling design.
Most participants from NHANES-III (76.5%) and NHANES 1999 to 2002 (75.6%) had a 10-year CHD risk <10%. The proportion of participants in the two surveys who were at intermediate (11.2% and 11.4%) and high (12.3% and 13.0%) 10-year CHD risk also was similar. Among men, approximately 15% to 16% of participants in each survey had high 10-year risk of CHD (>20%) in both survey with overlapping confidence intervals. Among women, approximately 10% to 11% had high 10-year risk in the two surveys. Among the 3 racial/ethnic groups, African Americans had the highest proportion of participants in the high-risk group. Although the percentage of white participants in high 10-year CHD risk group was about the same in both surveys (approximately 12%), a slightly increased prevalence of high 10-year risk was observed among African-American participants in more recent survey (14.4% in NHANES-III and 17.1% in NHANES 1999 to 2002) (Table 1).
We also examined the distribution of 10-year CHD risk among participants without existing CHD or CHD equivalent conditions (Table 2).Overall, approximately 2% of participants without CHD, stroke, or diabetes had a high 10-year risk of developing CHD. The estimates were similar for the 2 surveys with overlapping confidence interval. Greater proportion of participants among men had high 10-year risk of CHD (4% and 3.2% in the 2 surveys) than women (0.7% and 0.9%), but the estimates across surveys for men and women were similar. The prevalence of high 10-year risk in the two surveys also was similar among white (2.5% and 2.2%), African-American (2.0% and 1.5%), and Mexican-American (1.1% and 1.2%) participants.
The distribution of individual risk factors included in the modified Framingham score was examined separately for men and women (Table 3).Participants in NHANES 1999 to 2002 were slightly older and reported less smoking than those in NHANES-III. Total cholesterol and high-density cholesterol concentrations among the participants in the two surveys were very similar for men, whereas total cholesterol concentrations were slightly lower and high-density cholesterol concentrations were slightly higher among women in NHANES 1999 to 2002. Systolic blood pressure and the prevalence of treatment for high blood pressure were both greater among participants in NHANES 1999 to 2002 and the increase in systolic blood pressure was much greater for women than for men. The prevalence of myocardial infarction decreased slightly among both men and women, the prevalence of stroke also increased among both men and women, with a slightly greater increase among women, and the prevalence of diabetes showed much greater increase among men (6.4% and 8.5%) but not among women (7.4% and 7.2%).
These results showed that the overall distribution of 10-year risk for developing CHD among U.S. adults changed little between 1988 to 1994 and 1999 to 2002. During both periods, about 1 in every 8 U.S. adults ages 20 to 79 years had a >20% 10-year risk of developing CHD. No apparent change in 10-year risk of CHD was observed among men and women. Although most of the 10-year risk was explained by presence of prevalent CHD, stroke, and diabetes, the overall CHD risk distribution among persons without these conditions was also similar in the 2 surveys approximately 10 years apart: in each, approximately 2% of such persons were at high risk for CHD and approximately 13% were at intermediate risk.
Our results underscore the importance of using an overall risk assessment, in addition to targeting specific risk factors. During the last decade or so, important public health efforts have been made to reduce the prevalence of known risk factors for CHD. It is reasonable to expect that the risk of CHD would decrease over time in conjunction with a reduction in the prevalence of known CHD risk factors. Our results support the findings of changes in prevalence of some risk factors, such as decreased smoking over last decade or so. Somewhat surprisingly, however, such changes were not reflected in the overall 10-year risk for CHD, which remained essentially unchanged during the 10-year period.
One explanation of this finding is that increases in prevalence of some risk factors may have offset the benefits of decrease in prevalence of other risk factor. The increasing age of U.S. population may shift a greater proportion of U.S. adults into the higher risk group. However, our study was limited to those ages 20 to 79 years, and the change in average age of participants over the 2 surveys was <2 years. Other changes include increase in prevalence of hypertension as reported previously (8) and reflected by higher systolic blood pressure and increased proportion of participants reporting treatment for hypertension in more recent survey. Another risk factor trend resulting in unchanged 10-year CHD risk is the prevalence of diabetes. Consistent with an earlier report (19), an increased prevalence of diabetes was observed in more recent survey. At least part of this increase may be the result of the increasing prevalence of obesity (defined as body mass index ≥30 kg/m2). We observed increased prevalence of obesity among men (19.7% in NHANES-III and 27.3% in NHANES 1999 to 2002) and among women (25.0% in NHANES-III and 33.8% in NHANES 1999 to 2002). Similarly, increased prevalence of abdominal obesity (as used by NCEP/ATP III, waist circumference; for men >102 cm, for women >88 cm) was observed in more recent survey. The prevalence of abdominal obesity among men increased from 27.5% to 37.4%, whereas that among women, the increase was from 45.8% to 57.1% in NHANES-III and NHANES 1999 to 2002 data. These findings are consistent with increases in prevalence from other surveys (20). Although the prevalence of some harmful risk factors has increased over the course of time, a decrease in the prevalence of others, such as smoking, has been reported from national survey (7). This trend also was observed in these data. Although the mortality from CHD has declined in recent years, our finding that the risk of developing CHD has not declined over last decade may be a reason for concern. This may result in a constant burden or, in a worst-case scenario, increase the community burden of CHD in the coming years.
Our finding of no significant change in 10-year risk for developing CHD among U.S. adults does not explain declining CHD mortality rate. However, previous studies have suggested that reduction in case-fatalities, probably because of improved therapeutics, and improved survival of CHD patients may have driven the recent decline in CHD mortality rates (1,2,21). It is not clear, however, whether the decline in CHD mortality is entirely due to case-fatality reduction because some studies also have reported declining incidence (21,22). If reduction in case-fatalities were, in fact, resulting in lower CHD mortality, it can not be ruled out in our study by its design and would not be reflected in 10-year risk of developing CHD. Another possible explanation of our finding could be limitations of modified Framingham scoring, as adopted by National Cholesterol Education Program and used in our study. The coronary prediction algorithm was developed using a middle-aged white population sample. The different distribution of demographic characteristics of U.S. population in recent years may have resulted in misclassification of risk scores. Although the surveys were conducted a decade apart, inclusion of people ages 20 to 79 years resulted in a slight increase in average age of participant. Earlier studies have reported some limitations in using the Framingham risk score in different socioeconomic groups (23), and some socioeconomic disparities in clustering of cardiovascular disease risk factors also has been reported (24). In our study, CHD risks by socioeconomic status as reflected by level of education (high school or less, more than high school) were similar to overall findings. The proportion of people at higher risk was greater among those with education level of high school or lower in both surveys. However, the results across surveys were similar (data not shown). Another possible explanation can be the level of risk predicted by individual risk factor using the Framingham risk score. Although same risk factors are considered at 2 different points in time, difference in prevalence of risk factor and underlying rates of CHD events may result in need of some recalibration, as suggested by an earlier report (25).
Our results reflect some of the possible issues with current mechanisms of health promotion and disease prevention. During the last several years, an increasing amount of public health resources and efforts have been directed to reducing the burden of CHD. Several improvements can be made in health-promotion campaigns. The results of this study suggest that a more holistic public health approach to reduce CHD risk may be needed. Although targeting individual risk factors to reduce its prevalence is important, the objective should be to reduce overall risk for disease, which will require a more global approach of multifaceted programs targeting several risk factors. Health campaigns targeted toward a single CHD risk factor may not result in overall reduction in risk, especially if the prevalence of other CHD risk factor or behavior is on the rise. The proportion of U.S. adults engaged in healthy lifestyle affecting multiple behaviors is small (26). Persuading people to adopt a multidimensional healthy lifestyle to reduce the 10-year risk of developing CHD will require serious and concerted efforts from health care providers and public health professionals. Appropriate screening for CHD risk, active counseling, and timely treatment by health care providers and multifaceted community interventions and motivational campaigns by public health professionals will play a major role in overall risk reduction. In addition, a truly multidimensional effort focused on primary and secondary prevention and wide variety of activities from reduction in prevalence of multiple risk factors to reducing health disparities by demographic characteristics is likely to require partnership with many outside the traditional health community, such as policy makers and media. A call to action for dealing with this issue of controlling heart disease in the form of an action plan has been developed (27). A successful execution of such plan and effective implementation of overall risk reducing programs can play major role in reducing public health burden from CHD.
Our findings are based on large study samples from 2 surveys, representative of the U.S. population ages 20 to 79 years, and thus are applicable to the general U.S. population within this age group. However, because the study was cross sectional in design and used self-reported data about health behaviors; medication use; and diagnoses of CHD, stroke, and diabetes, some misclassification cannot be ruled out. Some participants were excluded from analyses because of incomplete data for all components necessary to compute CHD risk. However, on the basis of available characteristics, participants excluded from analyses appeared to be at greater risk of CHD. Although the reason for exclusion was incomplete data, such exclusion may have resulted in more conservative risk estimates.
In summary, our results provide important information about recent CHD risk distribution among U.S. adults. We found no appreciable change in 10-year risk distribution of developing CHD, although the prevalence of individual risk factors did change over time. The lack of change in the 10-year risk distribution for CHD among U.S. adults suggests a need for renewed commitments and more holistic public health efforts at the individual and population level to reduce the overall risk of CHD.
This research was conducted at Centers for Disease Control. The findings and conclusions in this paper are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
- Abbreviations and Acronyms
- coronary heart disease
- National Cholesterol Education Program/Adult Treatment Panel
- National Health And Nutrition Examination Survey
- Received September 7, 2005.
- Revision received May 16, 2006.
- Accepted May 23, 2006.
- American College of Cardiology Foundation
- Levi F.,
- Lucchini F.,
- Negri E.,
- La V.C.
- Fox C.S.,
- Sullivan L.,
- D’Agostino R.B. Sr.,
- Wilson P.W.
- American Heart Association
- Wilson P.W.,
- D’Agostino R.B.,
- Levy D.,
- Belanger A.M.,
- Silbershatz H.,
- Kannel W.B.
- ↵(2002) Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 106:3143–3421.
- Joint British Recommendations on Prevention of Coronary Heart Disease in Clinical Practice
- Ford E.S.,
- Giles W.H.,
- Mokdad A.H.
- ↵Centers for Disease Control and Prevention. NHANES-III Data Files. Available at: http://www.cdc.gov/nchs/about/major/nhanes/nh3data.htm. 2005. Accessed May, 2006.
- ↵Centers for Disease Control and Prevention. NHANES 1999–2000 Data Files. Available at: http://www.cdc.gov/nchs/about/major/nhanes/nhanes99_00.htm. 2005. Accessed May, 2006.
- Centers for Disease Control and Prevention. NHANES 2001–2002 Data Files. Available at: http://www.cdc.gov/nchs/about/major/nhanes/nhanes99_00.htm. 2005. Accessed May, 2006.
- Research Triangle Institute
- Brindle P.M.,
- McConnachie A.,
- Upton M.N.,
- Hart C.L.,
- Davey S.G.,
- Watt G.C.
- U.S. Department of Health and Human Services