The goals of the American College of Cardiology (ACC) and the American Heart Association (AHA) are to prevent cardiovascular (CV) diseases, improve the management of people who have these diseases through professional education and research, and develop guidelines, standards and policies that promote optimal patient care and CV health. Toward these objectives, the ACC and AHA have collaborated with the National Heart, Lung, and Blood Institute (NHLBI) and stakeholder and professional organizations, including those in the National Program to Reduce Cardiovascular Risk (NPRCR), to develop clinical practice guidelines for assessment of CV risk, lifestyle modifications to reduce CV risk, and management of blood cholesterol, overweight and obesity in adults.

In 2008, the NHLBI initiated these guidelines by sponsoring rigorous systematic evidence reviews for each topic by expert panels convened to develop critical questions (CQs), interpret the evidence and craft recommendations. In response to the 2011 report of the Institute of Medicine on the development of trustworthy clinical guidelines (1), the NHLBI Advisory Council (NHLBAC) recommended that the NHLBI focus specifically on reviewing the highest quality evidence and partner with other organizations to develop recommendations (2,3). Accordingly, in June 2013 the NHLBI initiated collaboration with the ACC and AHA to work with other organizations to complete and publish the 4 guidelines noted above and make them available to the widest possible constituency. Recognizing that the expert panels did not consider evidence beyond 2011 (except as specified in the methodology), the ACC, AHA, and collaborating societies plan to begin updating these guidelines starting in 2014.

The joint ACC/AHA Task Force on Practice Guidelines (Task Force) appointed a subcommittee to shepherd this transition, communicate the rationale and expectations to the writing panels and partnering organizations and expeditiously publish the documents. The ACC/AHA and partner organizations recruited a limited number of expert reviewers for fiduciary examination of content, recognizing that each document had undergone extensive peer review by representatives of the NHLBAC, key Federal agencies and scientific experts. Each writing panel responded to comments from these reviewers. Clarifications were incorporated where appropriate, but there were no substantive changes as the bulk of the content was undisputed.

Although the Task Force led the final development of these prevention guidelines, they differ from other ACC/AHA guidelines. First, as opposed to an extensive compendium of clinical information, these documents are significantly more limited in scope and focus on selected CQs in each topic, based on the highest quality evidence available. Recommendations were derived from randomized trials, meta-analyses, and observational studies evaluated for quality, and were not formulated when sufficient evidence was not available. Second, the text accompanying each recommendation is succinct, summarizing the evidence for each question. The Full Panel Reports include more detailed information about the evidence statements that serves as the basis for recommendations. Third, the format of the recommendations differs from other ACC/AHA guidelines. Each recommendation has been mapped from the NHLBI grading format to the ACC/AHA Class of Recommendation/Level of Evidence (COR/LOE) construct (Table 1) and is expressed in both formats. Because of the inherent differences in grading systems and the clinical questions driving the recommendations, alignment between the NHLBI and ACC/AHA formats is in some cases imperfect. Explanations of these variations are noted in the recommendation tables, where applicable.

In consultation with NHLBI, the policies adopted by the writing panels to manage relationships of authors with industry and other entities (RWI) are outlined in the methods section of each panel report. These policies were in effect when this effort began in 2008 and throughout the writing process and voting on recommendations, until the process was transferred to ACC/AHA in 2013. In the interest of transparency, the ACC/AHA requested that panel authors resubmit RWI disclosures as of July 2013. Relationships relevant to this guideline are disclosed in Appendix 5. None of the ACC/AHA expert reviewers had relevant RWI (Appendix 6).

Systematic evidence reports and accompanying summary tables were developed by the expert panels and NHLBI. The guideline was reviewed by the ACC/AHA Task Force and approved by the ACC Board of Trustees, the AHA Science Advisory and Coordinating Committee, and the governing bodies of partnering organizations. In addition, ACC/AHA sought endorsement by other stakeholders, including professional organizations and members of the NPRCR. It is the hope of the writing panels, stakeholders, professional organizations, NHLBI, and the Task Force that the guidelines will garner the widest possible readership for the benefit of patients, providers and the public health.

Guidelines attempt to define practices that meet the needs of patients in most circumstances and are not a replacement for clinical judgment. The ultimate decision about care of a particular patient must be made by the healthcare provider and patient in light of the circumstances presented by that patient. As a result, situations might arise in which deviations from these guidelines may be appropriate. These considerations notwithstanding, in caring for most patients, clinicians can employ the recommendations confidently to reduce the risks of atherosclerotic cardiovascular disease (ASCVD) events.

See Tables 2 and ​and33 for an explanation of the NHLBI recommendation grading methodology.

1. Introduction

2. Risk Assessment: Recommendations

3. Approach to Risk Assessment

In addressing its charge, the Work Group recognized the need for a risk assessment approach that was based on the types of data that primary care providers could easily collect and that could be implemented in routine clinical practice. After deliberation, the Work Group endorsed the existing and widely employed paradigm of matching the intensity of preventive efforts with the individual’s absolute risk (24,25). The Work Group acknowledges that none of the risk assessment tools or novel risk markers examined in the present document have been formally evaluated in randomized controlled trials of screening strategies with clinical events as outcomes. Nevertheless, this approach balances an understanding of an individual’s absolute risk for CVD and potential treatment benefits against the potential absolute risks for harm from therapy. Using this framework, treatment can be targeted to those most likely to benefit without undue risk for harm, in the context of a “risk discussion.” A risk discussion could include the assessment of the patient’s risk for ASCVD, and potential benefits, negative aspects, risks, and patient preferences regarding initiation of relevant preventive therapies.

By its nature, such an approach requires a platform for reliable quantitative estimation of absolute risk based upon data from representative population samples. It is important to note that risk estimation is based on group averages that are then applied to individual patients in practice. This process is admittedly imperfect; no one has 10% or 20% of a heart attack during a 10-year period. Individuals with the same estimated risk will either have or not have the event of interest, and only those patients who are destined to have an event can have their event prevented by therapy. The criticism of the risk estimation approach to treatment-decision making also applies to the alternative, and much less efficient approach, of checking the patient’s characteristics against numerous and complex inclusion and exclusion criteria for a potentially large number of pertinent trials. Only a small fraction of trial participants have events, and only a fraction of these events are prevented by therapy. Using either approach, the clinician must apply the average results obtained from groups of patients to the individual patient in practice.

Given the modification and adoption of the Framingham 10-year risk score for CHD risk assessment by the “Third Report of the National Cholesterol Education Program Expert Panel on Diagnosis, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III)” (25), and the uptake of this algorithm by practice sites across the United States, the Work Group began by discussing the value of retaining this algorithm. In collaboration with other NHLBI panels, the Work Group decided not to use this algorithm in its 2013 recommendations, because of its derivation in an exclusively White sample population and the limited scope of the outcome (in determining CHD alone). Rather, the Work Group derived risk equations from community-based cohorts that are broadly representative of the U.S. population of Whites and African Americans, and focused on estimation of first hard ASCVD events (defined as first occurrence of nonfatal myocardial infarction or CHD death, or fatal or nonfatal stroke) as the outcome of interest because it was deemed to be of greater relevance to both patients and providers. The focus on hard ASCVD, rather than CHD alone, is also consistent with evidence reviewed in a statement from the AHA/American Stroke Association calling for the inclusion of ischemic stroke in the outcome of interest for CVD risk assessment (26).

Numerous multivariable risk scores/equations have been derived and published (Appendix 3, and for more details, the Full Work Group Report Supplement (http://jaccjacc.cardiosource.com/acc_documents/2013_FPR_S5_Risk_Assessment.pdf)). As part of its deliberations, the Work Group considered previously published risk scores with validation in NHLBI cohort data as 1 possible approach. However, a number of persistent concerns with existing risk equations were identified including nonrepresentative or historically dated populations, limited ethnic diversity, narrowly defined endpoints, endpoints influenced by provider preferences (e.g., revascularizations), and endpoints with poor reliability (e.g., angina and heart failure [HF]). Given the inherent limitations of existing scores, the Work Group judged that a new risk score was needed to address some of the deficiencies of existing scores, such as utilizing a population sample that approaches, to the degree possible, the ideal sample for algorithm development and closely represents the U.S. population.

Data are sparse regarding usage and impact of absolute risk scores in clinical practice in primary prevention settings (27). Two systematic reviews, based on few studies, support the conclusion that risk assessment, combined with counseling, is associated with favorable but modest changes in patient knowledge and intention to change, and with provider prescribing behavior and risk factor control (28,29). No data are available on hard event outcomes. The Work Group specifically calls for research in this area (Section 8).

The Work Group notes that the “2009 ACCF/AHA Performance Measures for the Primary Prevention of CVD” specifically recommended use of global CVD risk estimation in clinical practice (30). Likewise, the U.S. Preventive Services Task Force recommendations for aspirin (31), NHLBI Adult Treatment Panel III recommendations (25), and European (32) and Canadian (33,34) guidelines for primary prevention of CVD, among others, have all recommended the use of absolute risk assessment for decision making about the intensity of lifestyle and pharmacological preventive interventions. Risk scores have been implemented in practice through paper scoring sheets, and increasingly through websites and downloadable applications. The electronic medical record can be adapted to estimate absolute risks automatically using patient data and published equations, and it is anticipated that risk estimation using this technology will become a mainstream application of the current and future risk algorithms.

4. Development of New Pooled Cohort ASCVD Risk Equations

Having made the decision to develop new equations to estimate the 10-year risk for developing a first ASCVD event, the Work Group used the best available data from community-based cohorts of adults, with adjudicated endpoints for CHD death, nonfatal myocardial infarction, and fatal or nonfatal stroke. Cohorts that included African-American or White participants with at least 12 years of follow-up were included. Data from other race/ethnic groups were insufficient, precluding their inclusion in the final analyses. The final pooled cohorts included participants from several large, racially and geographically diverse, modern NHLBI-sponsored cohort studies, including the ARIC (Atherosclerosis Risk in Communities) study (8), Cardiovascular Health Study (5), and the CARDIA (Coronary Artery Risk Development in Young Adults) study (7), combined with applicable data from the Framingham Original and Offspring Study cohorts (4,6).

The Work Group used state-of-the-art statistical methods to derive and internally validate the Pooled Cohort Equations, which provide sex-and race-specific estimates of the 10-year risk for ASCVD for African-American and White men and women 40 to 79 years of age. The variables that statistically merit inclusion in the risk assessment equations are age, total and HDL-cholesterol, systolic BP (including treated or untreated status), diabetes, and current smoking status.

An expanded description of the derivation and validation of the Pooled Cohort Equations, as well as the means for implementing them in clinical practice, are provided in Appendix 4. Additional details are provided in the Full Report of the Work Group (http://jaccjacc.cardiosource.com/acc_documents/2013_FPR_S5_Risk_Assessment.pdf). A specific clinical vignette is also provided as an example in Appendix 4. In the clinical vignette, the 10-year risk is calculated for a patient 55 years of age who is a nonsmoker without diabetes, and with total cholesterol 213 mg/dL, HDL–cholesterol 50 mg/dL, and untreated systolic BP 120 mm Hg. Using these values in the Pooled Cohort Equations, the predicted 10-year ASCVD risks are 2.1% for White women, 3.0% for African-American women, 5.3% for White men, and 6.1% for African-American men.

Numerous other potential risk markers were considered for inclusion in the Pooled Cohort Equations, but for many there was no additional utility demonstrated upon their inclusion; for others, data were insufficient at the present time to determine their additional value. The equations were also assessed in external validation studies using data from other available cohorts. Other than the Framingham CHD risk score (and its derivative ATP-III risk assessment profile) and the European SCORE (System for Cardiac Operative Risk Evaluation) algorithm for CVD death, these equations have been subjected to more rigorous validation than other currently available equations, and they are the only risk assessment equations that include significant numbers of African Americans and focus on estimation of 10-year risk for the clinically relevant endpoint of ASCVD. The Work Group specifically calls for further research to develop similar equations applicable to other ethnic groups, to validate the utility of the Pooled Cohort Equations in diverse primary prevention settings, and to assess the potential benefit of novel risk markers when added to these equations, so that the equations may be modified or expanded over time as new data become available.

5. Implications for Risk Assessment

A range of estimated 10-year risk for a first hard ASCVD event is illustrated in the Full Work Group Report Supplement (Tables 8–11), across a broad range of risk factor burdens for selected combinations of the risk factors in sex-race groups (African-American and White women and men) (http://jaccjacc.cardiosource.com/acc_documents/2013_FPR_S5_Risk_Assessment.pdf). The estimated risks are specific to defined combinations of the risk factors, and demonstrate how they vary over a broad spectrum of potential profiles. Risk factor levels that are more adverse than those shown in these tables should always be associated with a higher estimated risk. For example, if a given risk factor combination indicates an estimated 10-year risk for hard ASCVD of 8%, but a patient has a higher level of systolic BP or total cholesterol, or a lower level of high-density lipoprotein cholesterol, than shown for that cell, then the estimated risk would be ≥8%. Because the estimated probabilities can become unstable when approaching the limits of the sample data, the risk probabilities are truncated at 1% and 30%. The proportion of the U.S. adult population, 40 to 79 years of age, in selected strata of estimated 10-year risk for hard ASCVD events, are shown overall and by sex and race in Table 5. When compared with nonHispanic Whites, estimated 10-year risk for ASCVD is generally lower in Hispanic-American and Asian-American populations and higher in American-Indian populations (35,36); hence, the lack of ethnic-specific risk algorithms are an important gap in our efforts to understand and prevent ASCVD in these populations. While the development of algorithms specific to these race/ethnic groups is encouraged, in the interim, providers may consider using the equations for nonHispanic Whites for these patients. When doing so, it is important to remember that the estimated risks may be over-estimates, especially for Hispanic- and Asian-Americans.

6. CQs and Systematic Evidence Review

7. Implementation Considerations for Risk Assessment

A suggested approach for incorporating these recommendations into clinical practice is shown in Figure 1. For patients 20 to 79 years of age who are free from clinical ASCVD, the first step is to assess ASCVD risk factors. Whereas it is reasonable to assess ASCVD risk factors in younger and older individuals, limitations in available data prevented the development of robust risk assessment algorithms in these populations. Hence, for patients outside this age range, providers should refer to applicable clinical practice guidelines (i.e., pediatric (53) and adult primary prevention guidelines (54,55)). Risk assessment should be repeated every 4 to 6 years in persons who are found to be at low 10-year risk (<7.5%). Beginning at age 40, formal estimation of the absolute 10-year risk for ASCVD is recommended. Long-term or lifetime risk estimation is recommended for all persons who are between 20 to 39 years of age and for those between 40 to 59 years of age who are determined to be at low 10-year risk (<7.5%). As shown in Figure 1, all patients should receive applicable risk information and appropriate lifestyle counseling. The 10-year risk estimates provided by the new Pooled Cohort Equations differ from those generated by the Adult Treatment Panel III algorithm in several respects (25) as discussed in detail in the Full Work Group Report (http://jaccjacc.cardiosource.com/acc_documents/2013_FPR_S5_Risk_Assessment.pdf). To summarize, based on the risk estimation algorithm recommended by Adult Treatment Panel III, approximately 31.9% of the ASCVD-free, nonpregnant U.S. population between 40 and 79 years of age have a 10-year risk of a first hard CHD event of at least 10% or have diabetes. Based on the new Pooled Cohort Equations described here, approximately 32.9% have a 10-year risk of a first hard ASCVD of at least 7.5%. The outcomes and thresholds of these 2 approaches are different, but the overlap of these 2 means for defining high-risk groups is substantial, at roughly 75%. Nonetheless, these important differences make simple linear conversions imprecise. We recommend that healthcare organizations convert to these new Pooled Cohort Equations as soon as practical (Appendix 4). A downloadable spreadsheet enabling estimation of 10-year and lifetime risk for ASCVD and a web-based calculator are available at http://my.americanheart.org/cvriskcalculator and http://www.cardiosource.org/science-and-quality/practice-guidelines-and-quality-standards/2013-prevention-guideline-tools.aspx.

Open in a separate window

Figure 1

Implementation of Risk Assessment Work Group Recommendations

ACC indicates American College of Cardiology; AHA, American Heart Association; and ASCVD, atherosclerotic cardiovascular disease.

8. Evidence Gaps and Future Research Needs

The Work Group strongly recommends continued research to fill gaps in knowledge regarding short- and long-term ASCVD risk assessment and outcomes in all race/ethnic groups, across the age spectrum, and in women and men. Future research should include analyses of short- and long-term risk in diverse groups; optimal communication of ASCVD risk information; utility of short-and long-term risk assessment for motivating behavioral change and adherence to therapy; utility of short-and long-term risk assessment for influencing risk factor levels and clinical outcomes; utility of differential information conveyed by short- and long-term risk assessment; and utility of novel risk markers in short- and long-term risk assessment.

9. Conclusions

The Work Group’s approach to risk assessment represents a step forward in ASCVD prevention that is large enough to justify the challenges inherent in implementing a new approach, rather than staying with the CHD risk assessment approach recommended previously. The final recommendations are summarized in Table 4 and Figure 1. Two major advantages of this approach are the ability to estimate risk for a broader based ASCVD outcome that is more relevant to additional segments of the population, including women and African Americans, and the ability to provide risk estimates specific to African Americans. Promoting lifetime risk estimation may represent an additional step forward in supporting lifestyle behavior change counseling efforts. Periodic updating of the guidelines should address numerous issues related to risk assessment.

Acknowledgment

Appendix 1

Appendix 2

Appendix 4

Appendix 6

Appendix 7

Footnotes

References