Cardiovascular disease risk assessment for primary prevention: Risk calculators

INTRODUCTION — Cardiovascular disease (CVD), predominantly atherosclerotic cardiovascular disease (ASCVD), is the leading cause of global mortality and a major cause of morbidity [1,2].

This topic discusses risk calculators (risk models) that estimate risk of CVD (mainly ASCVD) for patients without known prior CVD (ASCVD).

The following issues are presented separately:

●Identification of candidates for ASCVD risk assessment and an approach to assessing risk. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach" and "Overview of primary prevention of cardiovascular disease" and "Aspirin for the secondary prevention of atherosclerotic cardiovascular disease" and "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease", section on 'CVD risk assessment'.)

●Primary prevention measures based upon estimated ASCVD risk. (See "Aspirin in the primary prevention of cardiovascular disease and cancer" and "Goal blood pressure in adults with hypertension" and "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease", section on 'CVD risk assessment'.)

●Risk factors for heart failure (HF). (See "Epidemiology of heart failure", section on 'Risk factors for heart failure'.)

KEY FEATURES OF RISK MODELS — Key features of multivariable risk calculators for estimating risk of incident ASCVD events in individuals without ASCVD include derivation dataset characteristics, risk factor inputs, and clinical endpoints. All risk models have advantages and disadvantages, and no single risk model is appropriate for all patients [3,4].

Clinically helpful features and limitations of risk calculators are discussed below. (See 'Features promoting clinical use' below and 'Limitations of risk calculators' below.)

Dataset characteristics — Characteristics of datasets include the number and diversity of included patients, the type of dataset (research or clinical), the geographic regions in which the patients reside, and the era (years) during which the patients' baseline and endpoint data were collected.

For estimation of risk of incident ASCVD, datasets typically include data on risk factors and outcomes from adults without overt baseline ASCVD who are in outpatient settings.

Role of geographic region — The geographic area in which a dataset's participants reside is a commonly used proxy for social drivers of health (including economic, political, social, and healthcare conditions). (See "Use of race and ethnicity in medicine", section on 'Social and cultural drivers of health'.)

The geographic region may be used in two ways:

●For calculator selection – Large geographic areas (such as countries) in which the derivation dataset participants resided are commonly used to identify the appropriate risk model to use for individual patients. The selected risk model or calculator is ideally derived from (or indexed to) a population in the country or geographic region where the patient resides. Thus, risk calculators are generally used in the countries in which they are derived. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'Choosing a risk calculator'.)

Some risk calculators include separate risk models for different risk regions. As an example, the SCORE2 2021 and SCORE2-OP 2021 risk models are calibrated for four European and United Kingdom risk regions (low, moderate, high, and very high risk). (See 'Europe' below.)

●For risk stratification – Another approach is to include geographic regions (as identified by country, zip code, or post code) as input variables within a risk model to stratify risk associated with social drivers of health (PREVENT 2023, WHO 2019, China-PAR 2016 [region and urbanization], and QRISK3 2018). (See 'PREVENT (2023)' below and 'QRISK3 (2018)' below.)

Risk factors — Each risk model is based upon a set of variables, or risk factors, that are associated with risk of incident ASCVD events in individuals without baseline ASCVD. Many risk factors included in multivariate ASCVD risk models cause graded increases in risk (eg, age, hypertension, serum low-density lipoprotein-cholesterol [LDL-C]) [5].

Generally accepted risk factors — Variables in ASCVD risk models typically include:

●Common patient descriptors – Most risk models include variables for age, sex, cholesterol (total, high-density lipoprotein [HDL], with or without direct measurement of LDL), systolic blood pressure, body mass index calculated from height and weight, diabetes mellitus, and current smoking.

The use of calculated versus measured LDL levels is discussed separately. (See "Measurement of blood lipids and lipoproteins", section on 'LDL cholesterol'.)

Smoking history is self-reported, and a major focus in CVD risk prediction has been on current smoking habits over the past year. Some risk algorithms include past history of smoking and qualitative estimates of how much the person has smoked. For instance, QRISK3 additionally stratifies smoking according to light, moderate, heavy, non-, and ex-smokers. (See "Cardiovascular risk of smoking and benefits of smoking cessation".)

●Treatment variables – Risk models may also include treatment variables, such as whether the patient is receiving antihypertensive medication (PREVENT 2023, ACC/AHA 2013, QRISK3 2018, China-PAR 2016), a statin (PREVENT 2023, ASCVD Risk Estimator Plus 2018), or aspirin (ASCVD Risk Estimator Plus 2018). The ASCVD Risk Estimator Plus 2018 provides an estimate of the effect of primary prevention therapies. (See 'ASCVD Risk Estimator Plus (2018)' below.)

Additional risk factors for ASCVD that may or may not be included in risk calculators are discussed below. (See 'Exclusion of relevant risk factors' below.)

Consideration of race and ethnicity — The use of race or ethnicity variables in ASCVD models continues to be examined. Because population studies have identified differences in ASCVD between race-defined groups, some ASCVD models have included race or ethnicity variables (eg, ACC/AHA pooled cohort equations and QRISK3).

However, evidence from the United States suggests that social drivers of health are important determinants of cardiovascular disease (CVD) mortality. Social drivers of health include an array of economic, political, social, and healthcare conditions. (See "Use of race and ethnicity in medicine", section on 'Use of race in medical algorithms and decision aids'.)

As an example, in a study of 50,808 United States adults, race and ethnicity were associated with increased CVD mortality [6]. Age- and sex-adjusted rates of CVD mortality were highest among Black adults, with a hazard ratio of 1.54 (95% CI 1.34-1.77) compared with White adults. However, adjustment for metabolic, behavioral, and social risk factors significantly attenuated this Black-White difference in CVD mortality (HR 0.91, 95% CI 0.79-1.05). Metabolic risk factors (obesity, hypertension, and diabetes), behavioral risk factors (smoking and lack of leisure-time physical activity), and social risk factors (unemployment, lower family income, food insecurity, not owning a home, and not married or living with a partner) were all independently associated with CVD death.

Endpoints — This topic focuses primarily on models estimating ASCVD risk, as this risk guides ASCVD primary prevention measures. Each ASCVD (or CVD) risk calculator is based upon endpoint data for only a selected subset of ASCVD (or CVD) outcomes.

●CVD – Clinical CVD encompasses a wide variety of congenital and acquired conditions, including heart disease, cerebrovascular disease, disease of the aorta, peripheral artery disease, and venous disease [1].

CVD risk calculators typically estimate risk over time for the development of a specific CVD diagnosis or diagnoses in persons without heart or vascular disease at the beginning of the observation period. As an example, the PREVENT calculator defines incident CVD based on selected incident ASCVD and/or HF events. (See 'PREVENT (2023)' below.)

●ASCVD – ASCVD is a subset of CVD caused by atherosclerosis; it includes atherosclerotic coronary heart disease (CHD) as well as atherosclerotic cerebrovascular, aortic, and peripheral artery disease. ASCVD is a much narrower category of disease than CVD. CHD is the most common manifestation of ASCVD and, globally, causes approximately half of cardiovascular deaths [2]. CHD clinically manifests as fatal or nonfatal myocardial infarction (MI) and angina pectoris, although most calculators do not include angina as an endpoint. In some contemporary publications, adults who present with an acute coronary syndrome and undergo nonelective percutaneous coronary interventions or coronary artery bypass grafting are classified as having CHD.

Contemporary risk calculators have focused primarily on "hard" ASCVD events (commonly defined as first occurrence of nonfatal MI, CHD death, or fatal or nonfatal stroke) [7,8]. Hard ASCVD endpoints are generally considered more reliable than endpoints that may be subjective (eg, angina) or influenced by provider preferences and practice patterns (eg, elective revascularization, hospitalization). As an example, the PREVENT calculator defines incident ASCVD as the initial presentation of MI, fatal CHD, or stroke. (See 'PREVENT (2023)' below.)

●Heart failure – HF is a complication of many heart disorders, including, but not limited to, CHD. Thus, HF is a subset of CVD that partially overlaps ASCVD. (See "Determining the etiology and severity of heart failure or cardiomyopathy", section on 'Etiology'.)

HF has not been included as an endpoint in most ASCVD risk calculators due to concerns about inconsistent reporting of HF among various studies [7]. Additionally, since HF has many causes other the CHD, its incidence is only partially a reflection of ASCVD. Consequently, the calculation of HF risk is separate from the calculation of ASCVD risk (as in the PREVENT calculator) [8]. (See "Determining the etiology and severity of heart failure or cardiomyopathy", section on 'Etiology'.)

FEATURES PROMOTING CLINICAL USE — Calculator features that promote widespread clinical use include:

●Ease of use – Calculator use is facilitated by reliable availability and performance (eg, by mobile application, website, and integration into electronic medical records). Clinical data to input into the calculator should be easily measurable and commonly available. One way to include some less commonly available clinical variables in a risk calculator is to include these as optional inputs.

●Accuracy – Calculator equations should accurately and precisely estimate risk of well-defined incident ASCVD (or CVD) events. Risk calculators should undergo external validation to determine accuracy and utility for different population groups. (See "Clinical prediction rules", section on 'Validation'.)

●Applicability – Calculators should estimate major clinical endpoints that are important at both the individual and population level. They should be derived and externally validated using diverse contemporary datasets that resemble the populations in which they will be used [8].

●Major society endorsement – Major professional society and health organization guidelines and statements have included recommendations for use of specific risk calculators to estimate risk of incident ASCVD or CVD [7-10].

The selection of risk calculators for individual patients is discussed separately. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'Choosing a risk calculator'.)

UNITED STATES — In the United States (US), the main risk calculators are the 2023 American Heart Association (AHA) Predicting Risk of cardiovascular disease EVENTs (PREVENT) calculator and the 2018 ASCVD Risk Estimator Plus (based upon the American College of Cardiology/American Heart Association [ACC/AHA] 2013 pooled cohort equations [PCE] and the 2016 Million Hearts Longitudinal ASCVD Risk Assessment Tool).

PREVENT (2023) — The 2023 AHA PREVENT calculator was developed, in part, because the 2013 PCE overestimate the risk of incident ASCVD in some contemporary populations [8,11]. The PREVENT calculator was derived and validated with more contemporary data from over 6.6 million patients [12]:

●Derivation dataset (data from 1992 to 2017) – 3,281,919 individuals with 106,661 incident total CVD events in 25 datasets including ARIC, CHS, CARDIA, FHS, JHS, MESA, and clinical data (from Geisinger Health and a random 50 percent selection of health systems in the Optum Labs Data Warehouse).

●External validation dataset (data from 1992 to 2017) – 3,330,085 individuals with 104,854 incident total CVD events in 21 datasets including REGARDS, Chronic Renal Insufficiency Cohort, Rancho Bernardo Study, and contemporary clinical data (the remaining 50 percent of health systems in the Optum Labs Data Warehouse).

●Risk factors – The PREVENT base model includes age, total cholesterol, HDL cholesterol, systolic blood pressure, body mass index, estimated glomerular filtration rate (eGFR), diabetes, current smoking, antihypertensive therapy, statin therapy.

The PREVENT add-on models include up to three optional risk factors: urine albumin-to-creatinine ratio (UACR), glycated hemoglobin (HbA1c), and US zip code of residence (for estimating social deprivation index).

●Clinical endpoints – The PREVENT calculator estimates risks of incident ASCVD (MI, fatal CHD, and stroke), HF, and CVD (defined as ASCVD and/or HF) [8,12]. Ten-year risks for these endpoints are estimated for adults aged 30 to 79; 30-year risks are estimated for adults aged 30 to 59.

ASCVD Risk Estimator Plus (2018) — The ASCVD Risk Estimator Plus incorporates the 2013 ACC/AHA PCE (calculator 1), the 2016 Million Hearts Longitudinal ASCVD Risk Assessment Tool, and ACC/AHA guideline recommendations [7,13,14]. The ASCVD Risk Estimator Plus provides estimates of the potential benefit of specific risk-lowering interventions (statins, antihypertensive medication, aspirin, smoking cessation), and enables updating and tracking of changes in risk over time based on a patient's actual response to interventions.

Features of the ACC/AHA PCE included in the ASCVD Risk Estimator Plus include [7,15]:

●Derivation dataset – 20,338 non-Hispanic White and Black individuals (without baseline ASCVD) with 2689 events in several US cohorts:

•Framingham Heart Study (FHS) original cohort (1948 to 2014)

•FHS offspring cohort (1971 to 2014)

•Coronary Artery Risk Development in Young Adults Study (CARDIA; 1983 to 2006)

•Cardiovascular Health Study (CHS; 1989 to 1999)

•Atherosclerosis Risk In Communities Study (ARIC; 1987 to 2011)

●External validation dataset – 37,759 non-Hispanic White and Black individuals (without baseline ASCVD) with 1715 events in several US cohorts:

•ARIC – Cycles closest to 2013 with 10-year follow-up

•FHS – Cycles closest to 2013 with 10-year follow-up

•Multi-Ethnic Study of Atherosclerosis (MESA, started in 2000 [16]) with six-year follow-up

•REasons for Geographic And Racial Differences in Stroke study (REGARDS; started with 2003 [17]) with four-year follow-up

●Risk factors – Race, sex, age, total cholesterol, HDL cholesterol, systolic blood pressure, antihypertensive treatment, diabetes mellitus, smoking.

●Clinical endpoints – The model estimates 10-year risk of incident ASCVD event for individuals aged 40 to 79 and 30-year (or lifetime) risk for individuals aged 20 to 59 years (calculator 1). An ASCVD event is defined as CHD death, nonfatal MI, fatal stroke, or nonfatal stroke.

EUROPE — The 2021 European Society of Cardiology guidelines on CVD prevention recommend use of the SCORE2 model for individuals in Europe aged 40 to 69 years old, and SCORE2-OP for individuals 70 years and older [18].

SCORE2 (2021) — The Systematic COronary Risk Evaluation 2 (SCORE2) risk model was published in 2021 [19].

●Derivation dataset – 677,584 individuals with 30,121 CVD events in 45 cohorts in 13 countries (European countries including the United Kingdom (UK), as well as the United States and Canada).

●External validation dataset – 1,133,181 individuals with 43,492 CVD events in 25 cohorts in 15 European countries including the UK.

●Risk regions – The model is calibrated for four European risk regions (low, moderate, high, and very high) using estimated age- and sex-specific incidences calculated by rescaling country-specific World Health Organization (WHO) CVD mortality data from 49 countries using multipliers calculated from UK and European data for 10.78 million individuals.

●Risk factors – Age, sex, systolic blood pressure, total cholesterol, HDL-C, LDL-C (optional), current smoker (y/n). LDL-C is not used for the risk calculation but for guideline-based recommendations.

●Clinical endpoints – This model estimates 10-year risk of first onset of CVD for individuals aged 40 to 69 years old without prior ASCVD or diabetes mellitus. CVD is defined as a composite of nonfatal MI, nonfatal stroke, and CVD mortality. CVD mortality is defined to include death from CHD, HF, stroke, and sudden death. Since this definition of CVD mortality includes HF and sudden death, this includes some deaths not caused by ASCVD.

SCORE2-OP (2021) — The SCORE2-OP algorithm was developed to estimate the risk of cardiovascular events in adults ≥70 years old [20].

●Derivation dataset – The model was derived from individuals over 65 years old without baseline ASCVD from the Cohort of Norway (28,503 individuals, 10,089 CVD events) [21].

●External validation dataset – The model was validated with data on 338,615 individuals with 33,219 CVD events from six study populations.

●Risk regions – Four European regions of CVD risk (low, moderate, high, and very high risk) were defined, as for the SCORE2 model.

●Clinical endpoints – The model estimates 5- and 10-year risk of CVD in individuals over 70 years old. CVD is defined as a composite of CVD mortality, nonfatal MI, and nonfatal stroke. CVD mortality is defined to include death from CHD, HF, stroke, and sudden death. Since this definition of CVD mortality includes HF and sudden death, this includes some deaths not caused by ASCVD.

UK — The most commonly used risk calculator for the United Kingdom (UK) is the QRISK3 calculator [22,23].

QRISK3 (2018) — The 2023 National Institute for Health and Care Excellence (NICE) guidelines recommend use of the QRISK3 calculators (QRISK3-2018 and QRISK3-lifetime calculator) [22-25] to estimate the 10-year and lifetime ASCVD risk for adults aged between 25 and 84 (including those with type 2 diabetes) without CVD [26].

●Derivation dataset – 7,889,803 patients (without baseline ASCVD) with 365,565 ASCVD events in 982 QResearch general practices in England.

●External validation dataset – 2,671,298 patients (without baseline CVD) in 328 QResearch general practices in England.

●Risk factors – Age, sex, ethnicity, UK postcode, smoking (non-, ex-, light, moderate, heavy), diabetes status (none, type 1, or type 2), angina or heart attack in a first-degree relative <60 years old, chronic kidney disease (stage 3, 4, or 5), atrial fibrillation, antihypertensive treatment, migraines, rheumatoid arthritis, systemic lupus erythematosus, severe mental illness, on atypical antipsychotic medications, on steroid tablets, diagnosis or treatment of erectile dysfunction, cholesterol/HDL ratio, systolic blood pressure, standard deviation of at least two most recent systolic blood pressure readings, height, and weight.

●Clinical endpoints – The ASCVD endpoint was defined as a composite of CHD (including angina, MI, and chronic ischemic heart disease), ischemic stroke, or transient ischemic attack. The QRISK3-2018 calculator estimates a 10-year QRISK3 score for individuals aged 25 to 84, which is the 10-year risk of incident ASCVD. The QRISK3-lifetime calculator for individuals in the same age range estimates an individual's lifetime ASCVD risk (up to age 99 years) and compares this risk with the individual's projected risk with good control of the following risk factors: smoking, body mass index, cholesterol/HDL ratio, and systolic blood pressure.

ASCVD is defined as a composite of CHD (including angina and MI), ischemic stroke, and transient ischemic attack. The model also provides a QRISK3 Healthy Heart Age (which is the age at which a healthy person of the specified sex and ethnicity has the same 10-year QRISK score).

CHINA-PAR RISK PREDICTOR (2016) — The China-PAR (Prediction for ACSVD Risk in China) calculator was published in 2016 [27]. The China-PAR risk estimator was found to be more accurate for individuals living in China than the ACC/AHA pooled cohort equations.

●Derivation dataset – 21,320 participants (with no baseline ASCVD) with 1048 first ASCVD events in two studies.

●External validation dataset – External validation was performed with two cohorts of 70,838 participants with no baseline ASCVD in two studies.

●Risk factors – Included risk factors are age, systolic blood pressure, total cholesterol, HDL-C, current smoking, and diabetes mellitus.

•Additional variables for men: waist circumference, geographic region, urbanization, and family history of ASCVD

•Additional variables for women: waist circumference and geographic region

●Clinical endpoints – The risk model predicts the 10-year risk of incident ASCVD, defined as nonfatal acute MI, CHD death, or nonfatal or fatal stroke.

WHO CVD RISK 2019 — The World Health Organization (WHO) CVD updated risk charts were published in 2019. These are available as color-coded charts for 21 worldwide regions [9,28].

●Derivation dataset – 85 prospective cohorts in Europe, North America, Japan, and Australia with 376,177 individuals aged 40 to 80 years and 19,333 ASCVD events. Baseline survey dates were 1960 to 2013. Data in the derivation model were mostly from resource-abundant countries.

●Recalibration by region – The algorithms were adjusted to different global regions using age- and sex-specific incidence rates and risk-factor values from the Global Burden of Disease (GBD) Study 2017 [29] and the NCD Risk Factor Collaboration (NCD-RisC) [30].

●External validation dataset – External validation was performed using individual-participant data from a further 19 prospective cohort studies with 1,096,061 individuals.

●Risk factors

•Laboratory-based model – Age, sex, smoking, systolic blood pressure, total cholesterol, and diabetes mellitus.

•Nonlaboratory-based model – Age, sex, systolic blood pressure, smoking, and body mass index.

●Clinical endpoints – 10-year risk of nonfatal or nonfatal MI, nonfatal or fatal stroke, or fatal CHD.

LIMITATIONS OF RISK CALCULATORS — Each risk calculator has limited accuracy and precision in estimating ASCVD risk for individual patients. The accuracy and precision of risk models are limited by the characteristics of the derivation database (era, diversity and size of study populations), included risk factors, and specified outcomes.

Incomplete risk stratification — Risk models that stratify patient risk according to the presence and number of defined risk factors (eg, diabetes mellitus) can identify high-risk individuals, but they may falsely reassure individuals with subthreshold abnormalities and may not adequately stratify risk among patients with higher risk [31,32]. (See "Overview of established risk factors for cardiovascular disease", section on 'Prevalence and impact of cardiovascular risk factors'.)

One way to address this problem is to include additional risk factors that identify subthreshold abnormalities and/or stratify risk among higher risk patients. As an example, the PREVENT 2023 risk base model includes a variable for presence/absence of diabetes and also includes estimated glomerular filtration rate (eGFR); optional variables include glycated hemoglobin (HbA1c) level and urine albumin-to-creatinine ratio (UACR) [8,12]. HbA1c can identify individuals with prediabetes. HbA1c, UACR, and eGFR are variables associated with increased risk of ASCVD in individuals with diabetes [33]. (See "Prevalence of and risk factors for coronary heart disease in patients with diabetes mellitus".)

Exclusion of relevant risk factors — Risk models commonly exclude some ASCVD risk factors because of insufficient data on some risk factors in the datasets used to derive and validate these models. Also, inclusion of infrequently assessed clinical risk factors as mandatory input variables in a risk calculator may limit the clinical use of that calculator. (See 'Features promoting clinical use' above.)

Recognized risk-enhancing factors — The following list highlights some recognized ASCVD risk-enhancing factors (table 1) [13]. Only some of these variables are included in commonly used risk calculators. If a patient has an ASCVD risk-enhancing factor (table 1) not included as a variable in the ASCVD risk calculator chosen for that patient, one approach is to consider this risk factor as an ASCVD risk-enhancing factor, as described separately. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach".)

●Family history of premature ASCVD – Some risk models (eg, SCORE2 2013) include family history of early ASCVD (males, age <55 years; females, age <65 years) given its association with increased risk of ASCVD.

Other risk models (eg, PREVENT 2023 and ASCVD Risk Estimator Plus 2018) have not included this variable due to inconsistent or uncertain ascertainment of family history in clinical settings and failure of family history to improve risk estimates in some cohort studies [8].

●Kidney disease – Variables that reflect kidney disease include eGFR and UACR (PREVENT 2023) and history of chronic kidney disease (QRISK3 2018).

●Inflammatory disorders – Chronic inflammatory disorders associated with ASCVD risk include psoriasis, rheumatoid arthritis, systemic lupus erythematosus, and chronic human immunodeficiency virus (HIV) infection (QRISK3 2018).

●Metabolic syndrome – This syndrome is defined as the co-occurrence of metabolic risk factors for type 2 diabetes mellitus and ASCVD (abdominal obesity, hyperglycemia, dyslipidemia, and hypertension). The National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) is the most widely used definition of metabolic syndrome (table 2). The China-PAR 2016 risk model includes waist circumference. (See "Metabolic syndrome (insulin resistance syndrome or syndrome X)".)

●Lipids/biomarkers – Some lipid measurements (eg, lipoprotein(a)) and biomarkers (eg, high-sensitivity C-reactive protein) associated with increased ASCVD risk (table 1) are not included in widely used ASCVD risk calculators.

Coronary artery calcification — Coronary artery calcification (CAC) by computed tomography is a validated marker of atherosclerosis but is not included in the table of risk-enhancing factors. The role of CAC screening in selected asymptomatic patients and use of the Multiethnic Study of Atherosclerosis (MESA) calculator to identify the CAC percentile is discussed separately. (See "Coronary artery calcium scoring (CAC): Overview and clinical utilization", section on 'Patient selection for CAC screening'.)

Timeframe considerations — Most risk models provide estimates of 10-year risk of ASCVD. This approach does not consider lifetime risk, which might be substantially higher and amenable to aggressive risk factor reduction [31,34-36]. To address this issue, some risk calculators offer 30-year or lifetime risk estimates. The PREVENT 2023 calculator estimates 30-year risk for adults aged 30 to 59. The ASCVD Risk Estimator Plus estimates lifetime risk for individuals aged 20 to 59 years. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach".)

Risk factor levels and therapeutic interventions may change over time. Periodic risk reassessment is part of overall healthcare. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'How often should ASCVD risk be reassessed?'.)

Requirement for updating — ASCVD risk calculators reflect the risks prevalent during the eras of their derivation cohorts and may require revision or replacement over time. Risk calculators derived using older clinical databases (such as the 2013 ACC/AHA pooled cohort equations) have been found to overestimate 10-year risk of ASCVD risk in some contemporary cohorts [11]. The reasons for this overestimation of risk are not known but may include secular changes in risk factors, greater use of cardiovascular prevention medicines and strategies, and a general decline in ASCVD morbidity and mortality over time.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Assessment of cardiovascular risk".)

SUMMARY AND RECOMMENDATIONS

●Categories of disease – Cardiovascular disease (CVD), predominantly atherosclerotic cardiovascular disease (ASCVD), is the leading cause of global mortality and a major cause of morbidity. (See 'Endpoints' above.)

•CVD encompasses a wide variety of heart and vascular diseases.

•ASCVD is a subset of CVD caused by atherosclerosis. Each ASCVD risk calculator is based upon endpoint data for only a subset of ASCVD outcomes (eg, coronary heart disease [CHD] death, myocardial infarction [MI], and stroke).

•Heart failure (HF) – HF is a complication of many heart disorders, including, but not limited to, CHD.

●Role of risk calculators – In individuals without ASCVD, risk calculators are used to guide primary prevention measures. The selection of risk calculators for individual patients is discussed separately. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'Choosing a risk calculator'.)

●Key features of risk calculators – Features include derivation and validation dataset characteristics, included risk factors, and the type of clinical risk estimates provided. The geographic region for calculator datasets impacts risk model selection as well as risk stratification within some risk models. Calculators that have been derived and validated using diverse contemporary datasets reflecting the populations in which they will be used are more likely to provide accurate estimates. (See 'Key features of risk models' above.)

●Characteristics promoting clinical use – Calculator features that promote widespread use include ease of use (reliable availability and inclusion of commonly available clinical input variables), accuracy (tested by external validation), applicability (estimation of major clinical endpoints), and major professional society endorsement. (See 'Features promoting clinical use' above.)

●Limitations of risk calculators – Limitations include incomplete risk stratification, exclusion of relevant risk factors, limited timeframe for risk predictions, and need for updating of risk calculator models. (See 'Limitations of risk calculators' above.)