Coronary artery calcium scoring (CAC): Overview and clinical utilization

INTRODUCTION — The association between vascular calcification and cardiovascular disease (CVD) is well established. In asymptomatic patients without established atherosclerotic cardiovascular disease (ASCVD), the presence of coronary artery calcification (CAC) on computed tomography (CT) scans is a well-validated measure of subclinical atherosclerosis. Indeed, the presence of CAC should prompt consideration of aggressive risk factor modification for primary prevention of ASCVD events (including myocardial infarction [MI], stroke, and death from coronary heart disease [CHD]).

The indications for and approach to obtaining a CAC score, along with the diagnostic and prognostic implications of CAC scoring, will be reviewed here.

The use of multidetector CT and other modalities for noninvasive coronary angiography are discussed separately. (See "Cardiac imaging with computed tomography and magnetic resonance in the adult".)

WHAT IS CAC?

Pathophysiology — CAC occurs concomitant with the progression of atherosclerotic plaque and is thought to result from smooth muscle cell apoptosis [1]. CAC begins as microscopic calcifications (0.5 to 15 micrometers), and these can grow into larger calcium sheet-like deposits (ie, >3 mm). Most CAC occurs in the intimal layer of the coronary artery, although in some cases it occurs in the medial layer of the blood vessel. Medial vessel calcification is more commonly seen in renal disease, hyperparathyroidism, and other noncoronary conditions. (See "Vascular calcification in chronic kidney disease".)

The earliest form of CAC appears as intimal thickening under microscopic imaging, whereas the calcification at later stages is usually clustered deep within the lipid core of a plaque near the internal elastic lamina of the coronary artery. The center of an atherosclerotic plaque does not always contain calcific deposits. The pattern and extent of calcification does not have a linear correlation with the size of an atherosclerotic plaque or its necrotic core. As such, the presence and extent of CAC can predict the presence of coronary artery stenoses, but in general it is a better marker of the extent of coronary atherosclerosis than it is of disease severity.

Younger patients tend to have more noncalcified plaque than older patients [2].

CAC quantification

●Agatston method – On breath-hold electrocardiogram (ECG) gated CT scans done for the purpose of CAC quantification, CAC is most commonly quantified using the Agatston method. Typically, a 2.5 to 3.0 mm slice thickness CT dataset is used to calculate the Agatston score. The area of calcified atherosclerosis (defined as an area with at least 1 mm² with a CT density >130 Hounsfield units [HU] [3]) is multiplied by a density weighting factor. Then, CAC is summed for the entire coronary artery tree. Other methods of quantifying CAC that do not account for lesion density, such as volume or mass scores, may have higher reproducibility, but are less validated for atherosclerotic cardiovascular disease (ASCVD) outcomes and rarely used clinically.

Most studies evaluating the predictive value of CAC have used absolute values of the Agatston score, which sums the area and density of the CAC. Anatomic features such as number of vessels involved and presence of proximal coronary artery disease add to the Agatston score in predicting major coronary heart disease (CHD) events [4,5]. These measures may enable refinement of models using CAC to stratify patients for cardiovascular disease (CVD) risk reduction.

A number of studies have used the following definitions to relate the CAC score to coronary plaque burden:

•0 Agatston units – No identifiable disease

•1 to 99 Agatston units – Mild disease

•100 to 399 Agatston units – Moderate disease

•≥400 Agatston units – Severe disease

●Incidental CAC – Incidental CAC can be found on non-gated chest CT scans that are obtained for other reasons (ie, lung cancer screening). We describe incidental CAC in a semi-quantitative manner in order to fully inform providers and patients of ASCVD risk and to guide preventive strategies [6]. Our approach is consistent with the Society of Cardiovascular Computed Tomography (SCCT) and Society of Thoracic Radiology (STR); they recommend that the presence, severity, and extent of CAC be reported in all non-contrast chest studies [6].

•Correlation with Agatston score – There is high correlation between ECG-gated and non-gated Agatston scoring. Non-gated scans are not typically reconstructed at 2.5 to 3.0 mm slices and may be acquired at variable kVp, making formal Agatston scoring impractical. However, several methods of assessing CAC severity using ordinal scores have been shown to correlate to Agatston scoring and have prognostic importance.

•Scoring incidental CAC – There is not adequate data upon which to base a recommendation on what type of scoring method to use for assessment of CAC on non-gated chest CT.

Possible methods are summarized below:

-An Agatston score may be calculated on non-gated CT scans using software. Moderate to high correlations between Agatston scores in gated and non-gated scans in the same people have been shown. We prefer to use this method for incidental CAC scoring.

-A visual estimation requires the least effort and equipment but has not been well studied (ie, subjective assignment of none, mild, moderate, or severe CAC would correlate with Agatston scores of 0, 1 to 99, 100 to 399, and >400 as described above).

-Ordinal scoring is semi-quantitative and uses a simple integer score. For example, one such scoring system assigns absent, mild, moderate, or severe CAC scores of 0, 1, 2, or 3 to each coronary artery studied (corresponding to CAC of 0, 1 to 99,100 to 333, and >400, respectively). In this system, mild CAC may be assigned if the CAC is present in less than one-third of the coronary arteries assessed.

CAC percentiles and calculators — Following the results of CAC scoring, providers may compare a patient's CAC score to patients of similar age, sex, and ethnicity using the Multiethnic Study of Atherosclerosis (MESA) calculator. This calculator can be used in people 45 to 84 years of age who do not have diabetes or known CVD. Using the percentiles and other information from the MESA calculator can help guide treatment decisions; this information may also help providers convey risk to their patients regarding the amount of subclinical coronary atherosclerosis compared with their peers. This information may also motivate patients to make important therapeutic lifestyle changes (diet, exercise) and willingness to take preventive therapies (statins, antihypertensive agents, etc) [7].

●The MESA calculator was derived from a study of 6110 participants from the MESA who did not have diabetes or CVD and who underwent CAC scans and Agatston scoring [8]. These participants were between 45 and 84 years of age, and identified themselves as White, African American, Hispanic American, or Chinese American. The current tool is thus applicable only for these four race/ethnicity categories and within this age range.

●An online calculator is available for CAC comparisons in younger patients, ages 30 to 45 [9]. This calculator was derived from pooled data on 19,725 patients from three cohorts (Coronary Artery Risk Development in Young Adults [CARDIA], the CAC consortium, and the Walter Reed Cohort). The majority of patients were either Black or White, and thus the calculator allows for calculation of age- and race-specific percentiles of CAC for these two groups of patients.

The absolute CAC score is the primary driver of short-term risk. For example, a younger patient with a CAC of 40 is at low short-term risk but high lifetime CVD risk if this score places them in the category of >75^th percentile for their age and sex; a patient with CAC >400 is at high CVD risk regardless of the time horizon.

The prevalence of CAC in specific race/ethnic groups and in the context of other social determinants of health are described in more detail below. (See 'Race-ethnicity' below.)

Accuracy for detecting coronary artery disease — In both males and females, CAC detected by CT is highly sensitive for the presence of ≥50 percent angiographic stenosis, but only moderately specific, especially in individuals over 60 years of age [10-15]. In a review of 16 studies, the sensitivity and specificity of CAC for angiographic stenosis (≥50 percent diameter stenosis) detection were 91 and 49 percent, respectively [16]. Both sensitivity and specificity for the presence of ≥50 percent angiographic stenosis vary with the amount of CAC; as CAC increases, sensitivity falls but specificity increases.

IMAGING CAC — Advantages of CAC imaging compared with invasive coronary angiography or coronary CT angiography (CCTA) include minimal requirements for patient preparation, no requirement for iodinated intravenous contrast, and relatively low effective radiation doses.

●Patient preparation – CAC scoring may be performed without any patient preparation. Unlike for invasive coronary angiography or CCTA, patients do not need an intravenous catheter, intra-arterial catheter, beta blockers, nitroglycerin, or other medications, nor does CAC scoring require iodinated contrast or fasting. Patients should ideally be able follow instructions asking them to hold their breath for typically three to five seconds and to lie still during scan acquisition.

●Imaging acquisition – The CAC scan is a rapidly acquired, prospective, ECG-triggered, non-contrast CT image of the heart. Initially validated using electron beam CT, contemporary CAC scoring is now performed using multidetector CT scanning systems. Overall, the image acquisition for a CAC scan differs from that of a chest CT done for other indications (eg, lung cancer screening). (See 'Incidental CAC' below.)

●Radiation dose – CAC scans should be performed in accordance with contemporary professional society guidelines. The acquisition of the CAC scan should have an effective radiation dose of no more than 1 mSv; this dose is comparable to that of screening mammography (approximately 0.8 mSv) [17]. CAC scans can be accurately performed using contemporary dual-source CT scanners using prospective high-pitch helical acquisition at an even lower radiation dose (<0.4 mSv) [18], and CAC scoring using iterative reconstruction techniques may allow for even further dose reductions. Performing CAC using 100 kVp rather than 120 kVp also results in significant dose reduction, but requires modification of the Hounsfield units (HU) threshold for CAC detection in order to perform accurate Agatston scoring and is therefore not routinely performed [16].

OVERVIEW OF CLINICAL UTILITY — Two primary uses of screening for CAC include prediction of atherosclerotic cardiovascular disease (ASCVD) and selection of patients for treatment with aggressive ASCVD risk factor modification (ie, statin therapy).

These suggested uses of CAC scoring, including any therapy based on the results of the CAC scan, are generally in agreement with those of professional societies [19-22].

Prediction of ASCVD risk — CAC scores are usually interpreted in conjunction with other ASCVD risk scoring to guide therapeutic decisions. Various risk calculators and methods of risk assessment are discussed separately. (See "Cardiovascular disease risk assessment for primary prevention: Risk calculators" and "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach".)

●ASCVD risk scoring – Categories of ASCVD risk based on risk calculators are a common way of assessing patient risk. One such set of categories is presented below:

•Low-risk – Less than 5 percent 10-year ASCVD risk

•Borderline elevated risk – 5 to 7.4 percent 10-year ASCVD risk

•Intermediate risk – 7.5 to 10 percent 10-year ASCVD risk

•High-risk – 10 percent or greater 10-year ASCVD risk

•Very high-risk – 20 percent or greater 10-year ASCVD risk

●Incremental benefit of CAC – Studies have shown that CAC is accurate in predicting future ASCVD. A CAC score can also add incremental value when used in addition to validated risk calculators (eg, American College of Cardiology/American Heart Association [ACC/AHA] pooled cohort equation, Framingham Risk Score, Joint British Societies, etc) [23,24].

The prognostic implications of CAC have now been demonstrated in the prospective Multiethnic Study of Atherosclerosis (MESA). In MESA, CAC was measured in 6814 asymptomatic males and females without baseline cardiovascular disease (CVD) [25]. Participants were followed for 3.8 years for incident coronary heart disease (CHD) death and nonfatal coronary events. Compared with patients who have a CAC score of 0, hazard ratios (HR) for a coronary event were 7.7 for those with a CAC between 101 and 300 and 9.7 for those with CAC >300.

Among patients in the CAC consortium cohort, a retrospective multicenter study of 66,636 asymptomatic patients without known CVD, 4.3 percent of patients had a CAC score of >1000. These patients had an increase in CVD-related mortality (ie, from myocardial infarction or revascularization) compared with lower CAC scores [CAC scores of 0 (HR 5.0, 95% CI 3.92-6.48); 1 to 399 (HR 1.77, 95% CI 1.43-2.18); and 400 to 999 (HR 3.09, 95% CI 2.41-3.97)] [26]. Patients with CAC >1000 also had increases in all-cause mortality compared with lower CAC scores.

●Comparison with other ASCVD risk enhancers – The serum concentration of biomarkers, including high-sensitivity cardiac troponin T (hs-cTnT) and high-sensitivity C-reactive protein (hsCRP), have predictive value for CVD in a variety of clinical settings. (See "Elevated cardiac troponin concentration in the absence of an acute coronary syndrome", section on 'Elevation in the general population' and "C-reactive protein in cardiovascular disease".)

However, in asymptomatic patients, the absence of CAC (CAC = 0) has been shown to be the strongest "negative risk factor" as compared with normal or negative values of multiple other novel risk markers, to include carotid intima-medial thickness (CIMT), absence of carotid plaque, family history, ankle brachial index, B-type natriuretic peptide (BNP), albuminuria, family history, and hsCRP. This "power of zero" provides the strongest degree of individual "de-risking" available (ie, when compared with other traditional and other novel biomarkers) (figure 1) [27].

CAC is a stronger predictor of cardiac risk than most other serum biomarkers, including hsCRP. In a study of 1286 patients, the log of CAC increased the C-statistic for cardiac events, whereas multiple biomarkers including hsCRP, interleukin-6, myeloperoxidase, BNP, and plasminogen activator-1 did not [28]. The C-statistic is a statistical measure of how well a predictor(s) can discriminate between an event (eg, cardiac event) or non-event. In the MESA study, CAC was compared with hsCRP, CIMT, ankle brachial index, brachial flow-mediated dilation, and family history of CHD in 1330 individuals without diabetes [24]. Each of the individual markers improved the C-statistic over Framingham Risk Score, but CAC demonstrated the highest increment. The net reclassification improvement with CAC for incident CHD was far superior to any of the other markers.

Treatment based on risk reclassification — Obtaining a CAC score in specific patient groups can often help us decide in whom we would recommend more aggressive lifestyle or pharmacologic therapy for CVD risk factor modification including statin therapy.

Many patients who would not be considered high risk may have a high CAC score that would warrant consideration for statin therapy. In the MESA study, among patients at intermediate CVD risk (based on the Framingham Risk Score), 6.8 percent had a CAC score >400; these patients were not otherwise considered candidates for lipid-lowering therapy. (See "Cardiovascular disease risk assessment for primary prevention: Risk calculators".)

CAC may be better than age at reclassifying ASCVD risk. (See 'Older adults' below.)

PATIENT SELECTION FOR CAC SCREENING

Patients we screen — We screen for CAC in selected asymptomatic adults ≥40 years of age at intermediate to high risk (7.5 to <20 percent 10-year atherosclerotic cardiovascular disease [ASCVD] risk) by the American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort risk calculators (calculator 1) [20]. In these patients, CAC scoring may serve as an arbiter of risk and also help to guide clinical-patient risk discussion regarding statin and possibly aspirin therapy. If the decision to treat with a statin has already been decided, we would not obtain a CAC scan in such patients.

Among asymptomatic individuals, the CAC score has prognostic value in predicting ASCVD and mortality; this is particularly true for those at borderline and intermediate risk (ie, 5 to 20 percent 10-year ASCVD risk) [25,26,29-34]. While these studies have demonstrated the prognostic impact of CAC scoring across a wide range of clinical risk, age, sex, and ethnicity, the value of cardiovascular disease (CVD) risk reclassification appears to be most significant among individuals at borderline to intermediate risk (5 to 20 percent 10-year ASCVD risk) based on traditional risk factors.

Consistent with these findings, the Heinz Nixdorf Recall Study of over 4400 participants without known coronary heart disease (CHD) had CAC testing and were followed for over five years for the development of CHD [29]. Among the intermediate CVD risk category (as assessed by the Framingham Risk Score), 2.1 percent of patients experienced a CHD event. The presence of high CAC helped refine ASCVD risk among patients in this risk category; CHD event rates were highest in the CAC >400 group (6.8 percent) compared with the CAC 100 to 399 (2.4 percent) and CAC <100 (1.4 percent) groups. Nearly 22 percent of patients in the intermediate CVD risk category were correctly reclassified into higher- or lower-risk groups if CAC <100 was used to place them in the low-risk category and CAC ≥400 to place them in the high-risk category.

Patients we may screen — We screen some patients with borderline elevated ASCVD risk (5 to 7.4 percent 10-year ASCVD risk). For instance, in those with a family history of premature ASCVD in the borderline ASCVD risk category, CAC may serve as an arbiter of risk; the presence of CAC may change decision-making or patient acceptance with regards to statin, possibly aspirin treatment and intensity of ASCVD risk factor modification. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'Identify risk-enhancing factors'.)

Patients we do not screen

●Patients at low (<5 percent 10-year risk) or very high (≥20 percent 10-year risk) ASCVD risk – When low or very high risk is indicated by the ACC/AHA pooled cohort risk calculators (calculator 1), the results are generally unlikely to change our approach to treatment in these populations [20]. In patients at very high risk, we will most always pursue primary prevention lifestyle and pharmacologic therapy. (See "Overview of primary prevention of cardiovascular disease".)

There may be incremental prognostic benefit to obtaining a CAC in patients with 10-year ASCVD risk of <5 percent [31,32]; however, the absolute risk is too low to warrant therapy unless there are other concerning clinical risk factors, such as family history or chronic inflammatory conditions, which inform shared decision-making. This was illustrated in the Coronary Artery Risk Development in Young Adults (CARDIA) study of subjects 32 to 46 years old [32]. In this study, the number needed to screen (NNS) to detect a CAC score of ≥100 was calculated for low- and higher-risk patients (using the Framingham Risk Score). CAC score ≥100 was observed in 1.3, 2.4, 3.5, and 17.2 percent of those with Framingham Risk Scores of 0 to 2.5 percent, 2.6 to 5 percent, 5.1 to 10 percent, and >10 percent, respectively (NNS = 79, 41, 29, 6).

●Patients with symptoms of myocardial ischemia – We recommend against CAC scoring alone as a primary diagnostic tool in patients with symptoms concerning for myocardial ischemia (such as angina, dyspnea or effort intolerance). Although CAC has prognostic value in symptomatic patients [35,36], a score of 0 does not carry the same high negative predictive value as it does in asymptomatic patients. In patients with symptoms, CAC has decreased specificity for predicting significant coronary artery disease and the relatively high prevalence of CAC would necessitate much additional testing.

However, in symptomatic patients who are undergoing coronary CT angiography(CCTA) for diagnostic purposes, CAC scoring may be concurrently performed as it may provide additional prognostic information. (See "Noninvasive testing and imaging for diagnosis in patients at low to intermediate risk for acute coronary syndrome" and "Clinical use of coronary computed tomographic angiography".)

Studies suggest that 13 to 16 percent of people with symptoms and CAC = 0 have hemodynamically significant coronary artery disease; 3.5 percent of such patients have non-hemodynamically significant coronary artery disease (<50 percent stenosis) [2,37,38]. Among symptomatic patients in the multicenter PROMISE study, CCTA was superior to CAC scoring for CHD event prediction; 16 percent of patients with CAC = 0 were shown to have non-calcified plaque on CCTA. During approximately two years of follow-up, 16 percent of all events occurred in those with CAC = 0 [38].

MANAGEMENT BASED ON CAC

General measures in all patients — We counsel patients to make therapeutic lifestyle changes (eg, healthy diet, smoking cessation, regular exercise, etc), and we institute atherosclerotic cardiovascular disease (ASCVD) risk factor modification (eg, optimize blood pressure, smoking cessation, etc) in all patients. The presence of CAC may provide additional motivation for patients to make these therapeutic lifestyle changes [7]. (See "Overview of primary prevention of cardiovascular disease".)

We do not recommend repeat/serial CAC scanning to assess the efficacy of lifestyle changes.

When making management decisions, we prefer to use CAC percentiles based on age, sex, and race. However, we recognize that these will not generalize to all patients (ie, if an available calculator does not account for a specific age or race ethnic group). In this case, a CAC score alone can be used to guide management.

CAC score greater than or equal to 100 or >75 percentile for age, sex, and race — For patients in this moderate or more category of CAC, we recommend treating with a statin if the low density lipoprotein cholesterol (LDL-C) is between 100 and 190 mg/dL. This is consistent with professional society guidelines that recommend statin therapy in those with CAC >100 or >75^th percentile for age, sex, and race and consider statins in those with any measurable CAC [20]. (See "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)

●CAC ≥100 is associated with ASCVD risk – Among 6814 Multiethnic Study of Atherosclerosis (MESA) participants of all demographic groups, those with CAC ≥100 were uniformly above an ASCVD risk of 7.5 percent, which is an accepted threshold for initiation of ASCVD preventive therapies [25]. (See "Cardiovascular disease risk assessment for primary prevention: Risk calculators" and "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach".)

●CAC ≥100 or >75^th percentile predicts ASCVD – When these CAC score and percentile criteria were applied retrospectively to patients in the intermediate ASCVD risk category (7.5 to <20 percent 10-year ASCVD risk) in Framingham Heart Study participants, they predicted ASCVD [39]. Among 389 such participants, those with CAC score ≥100 or ≥75^th percentile for age and sex were observed to have a substantially greater risk for ASCVD (ASCVD events 21.2 versus 4.9 percent, hazard ratio [HR] 5.0, 95% CI 2.1-12.7) compared with patients with CAC = 0. Participants with CAC >75^th percentile and a CAC score of 1 to 99 had higher ASCVD event rates compared with participants with those with CAC = 0 (12.2 versus 9.0 percent) and substantially lower ASCVD event rates compared with those with CAC ≥100 (ASCVD events 24.4 percent).

Aspirin appears to have a net clinical benefit (which outweighs the risk) for ASCVD primary prevention in patients under 70 years old who also have CAC ≥100, regardless of their ASCVD risk category. Shared decision-making discussions between patient and their provider(s) are warranted prior to initiating therapy [40]. (See "Aspirin in the primary prevention of cardiovascular disease and cancer".)

We do not recommend repeat/serial CAC scanning to assess the efficacy of these therapies.

In multiple observational studies, the presence and severity of CAC has been shown to positively affect the initiation, treatment intensification, and patient adherence to aspirin, statins, antihypertensive agents, and exercise [7].

Direct evidence of improved patient outcomes from changes in statin, aspirin, and other preventive therapies prescribed according to CAC score has been limited to mainly small randomized studies and observational data.

●Evidence for combined aspirin and statin use – Treatment studies focused specifically on individuals with CAC >75^th percentile for age, sex, and race are limited. However, one trial in the St. Francis Heart Study included people and CAC scores ≥80^th percentile for age and sex (from an internal database). In this study, 1000 such people aged 50 to 70 years were treated with low-dose aspirin and randomly assigned to taking atorvastatin (20 mg/day), vitamin C (1 g/day), and vitamin E (1000 U/day) or matching placebos [41]. The primary endpoint was a composite of all ASCVD events.

Over a mean follow-up of 4.3 years, the group that received statin and vitamins had a nonsignificant trend towards a fewer ASCVD events (6.9 versus 9.9 percent). In a non-prespecified analysis, ASCVD event rates were directly related to the baseline CAC score and may have been more markedly reduced in the subgroup of patients with a baseline CAC score >400 (8.7 versus 15.0 percent). The group assigned to statin and vitamin therapy lowered their LDL-C by 43 percent from a mean baseline value of 146 mg/dL (3.8 mmol/L) to <80 mg/dL (2.1 mmol/L).

●Evidence for statin use – In general, the cumulative available data strongly suggest that patients with a CAC score >100 benefit from long-term statin therapy. A study from MESA longitudinal observational cohort illustrates that treating people with CAC >0 with a statin has benefits on reducing cardiovascular events that are more pronounced in persons with higher CAC scores (figure 2) [42].

•In 5534 statin-naïve MESA subjects followed for 7.6 years, the projected number needed to treat (NNT) with a statin to prevent an event was 30 in those with no lipid abnormalities and CAC score >100; the NNT was 154 for people with three risk factors and a CAC score of 0 (table 1) [42].

•In a large-scale analysis from a single institution retrospective cohort study, 13,664 patients without known ASCVD underwent CAC scoring and were followed for a median of 9.4 years for ASCVD outcomes. Using propensity score risk adjustment, patients with any CAC treated with statins had significantly reduced ASCVD outcomes compared with non-statin treated patients (HR 0.76, 95% CI 0.60-0.95) [43]. Furthermore, patients without CAC derived no benefit from long-term statin treatment unless they had high baseline ASCVD risk.

When stratified according to CAC severity, statin treatment was associated with the most significant reduction in events in subjects with CAC >100, with a NNT of 12 subjects with CAC >100 to prevent one incident ASCVD outcome over 10 years.

It may not be feasible to conduct a randomized trial of statins based upon baseline CAC as it may be unethical to withhold statins from those with CAC >100 [40,44]. (See "Aspirin in the primary prevention of cardiovascular disease and cancer" and "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)

●Evidence for aspirin use – Aspirin use is estimated to result in net clinical benefit (eg, reduced ASCVD events and an acceptable bleeding risk) in adults under 70 years old with CAC score ≥100 regardless of 10-year ASCVD risk. The MESA study suggested that individuals with CAC ≥100 had a net benefit with aspirin regardless of other risk factors, whereas individuals with a score of 0 had potential for more harm than benefit (table 1) [42]. The net clinical benefit of aspirin in adults >70 years old with elevated CAC has not been well studied.

CAC score 1 to 99 or <75th percentile for age, sex, and race — We consider statins for patients with LDL-C between 100 and 190 mg/dL if their CAC scores are mild, between 1 to 99 or <75^th percentile based on patient and provider shared decision-making. The benefit of lipid lowering in patients with this level of CAC is small. Some clinicians and patients might decide that this degree of benefit is still worthwhile. (See "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)

We generally do not use aspirin for primary prevention of ASCVD in patients with a CAC of <100 regardless of 10-year ASCVD risk.

CAC of zero — We do not routinely use statin therapy in patients in whom CAC is 0. These patients usually have an established low 10-year ASCVD risk and thus we recommend against statin therapy for up to five years (at which point repeat risk stratification should be performed) unless there are other compelling risk factors (eg, diabetes, active smoking, hypertension, family history of premature coronary artery disease) [19,43]. (See "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)

We generally do not use aspirin for primary prevention of ASCVD in patients with a CAC of zero regardless of 10-year ASCVD risk.

Specific patient examples of when having a CAC score of 0 may impact therapeutic decisions regarding therapeutic lifestyle changes and modification of all cardiovascular risk factors are presented in a table (table 1). (See "Screening for coronary heart disease".)

Incidental CAC — For a patient who has CAC found incidentally (eg, on a diagnostic chest CT), we use the same treatment approach compared with patients who have CAC scoring done intentionally. These patients will usually have had non-gated chest CT, making Agatston scoring impossible. We would treat incidentally noted CAC the same as we would intentionally measured CAC.

FOLLOW-UP BASED ON CAC SCORE

CAC of 0 at baseline — Among patients with a baseline CAC score of 0, we use their baseline atherosclerotic cardiovascular disease (ASCVD) risk to determine when repeat CAC testing should be performed. A repeat scan is done if it may change treatment recommendations. The following ASCVD risk categories determine when the repeat scan should be done:

●In patients at low 10-year ASCVD risk (<5 percent), we repeat CAC scanning in five to seven years.

●In patients at intermediate risk for ASCVD (5 to 10 percent), we repeat CAC scanning in three to five years.

●In patients with diabetes, we repeat CAC scanning in three years.

This approach was recommended by the authors of a Multiethnic Study of Atherosclerosis (MESA) study of 3116 individuals with baseline CAC score of 0 who underwent repeat CAC scanning [45]. Among study participants, the time for conversion to CAC >0 varied according to baseline ASCVD risk and particularly the presence of diabetes.

CAC progression and risk — We do not recommend serial CAC scoring in patients with established CAC (>0), particularly those with CAC >100, as repeated measurement will rarely alter management.

●Although rapid CAC progression is associated with increased risk of cardiovascular events [46-48], similar to the association between rapid angiographic progression of coronary atherosclerosis and cardiovascular events [49], progression does not add to risk prediction. In the Heinz Nixdorf Recall study of 3281 participants who underwent baseline and follow-up CAC scoring at a mean of 5.1 years, CAC progression added little to baseline risk prediction [50].

We also recommend against serial CAC scoring to assess the efficacy of cholesterol lowering therapy.

●It is important to note that according to a meta-analysis of four prospective studies, statins do not appear to significantly reduce the rate of CAC progression [51]. Conversely, studies suggest that statins may increase the calcified content of individual plaques through plaque stabilization [52], while reducing non-calcified plaque volume.

PROGNOSTIC VALUE BY PATIENT GROUP — There are differences in the prevalence of CAC in different patient groups; however, the management based on CAC score for these patients is generally the same as for other patients.

Older adults — In older adults, CAC is an independent predictor of long-term risk for coronary heart disease (CHD) and may be used to reclassify risk among asymptomatic individuals [53-56]. The Rotterdam Study followed 2028 asymptomatic adults (mean age 69) over a median of 9.2 years [55]. In the intermediate risk group (10-year Framingham Risk Score 10 to 20 percent), CAC reclassified 52 percent as either high-risk (if CAC score was >615) or low-risk (if CAC score was <50). This was similar to data from the Multiethnic Study of Atherosclerosis (MESA) study suggesting CAC was better than age at accurately reclassifying atherosclerotic cardiovascular disease (ASCVD) risk [54]. In another study of 3570 patients over the age of 70, having a CAC <400 helped accurately reclassify high-risk patients (ie, with ≥3 risk factors) into a lower-risk category [56].

Females — Whereas females have less CAC as compared with males of similar age and risk factor burden, the prognostic value of CAC is equal in both sexes. However, females generally have lower 10-year ASCVD risk according to risk scores. In a meta-analysis of five large-scale population-based studies involving 6739 middle-aged females at low to borderline 10-year ASCVD risk (<7.5 percent), the prevalence of any CAC was 36 percent. CAC >0 was associated with a twofold increased risk of ASCVD outcomes over 7 to 12 years of follow-up. The presence of CAC reclassified (net reclassification index) 20 percent of females to more appropriate risk categories and improved overall accuracy for event prediction (the C-statistic increased from 0.73 to 0.77) [57].

Diabetes — Data from a CAC consortium study of over 22,000 patients suggest that the optimal age for a potential first scan is earlier for males and for people with diabetes [29]. Specifically:

●Males with diabetes: 35 to 38 years

●Females with diabetes: 49 to 52 years

●Males without diabetes: 41 to 44 years

●Females without diabetes: 56 to 60 years

Compared with individuals without risk factors, those with diabetes developed CAC 6.4 years earlier on average, whereas smoking, hypertension, dyslipidemia, and a family history of CHD were individually associated with developing CAC 3.3 to 4.3 years earlier.

Race-ethnicity — The prevalence and amount of CAC is higher among White individuals as compared with Chinese American, Hispanic American, and African American individuals even though the predictive value of CAC is similar across race-ethnic groups [8,58-60].

●Prevalence of CAC

•In a separate study in MESA of 6814 persons (mean age 63 years), CAC was evaluated in those free of cardiovascular disease (CVD). Among males, the prevalence of CAC was significantly higher in White compared with Chinese, Hispanic, or African Americans (70 versus 59, 56, and 52 percent, respectively) [58]. A similar trend was noted in females, although CAC was less common (45 versus 42, 35, and 37 percent, respectively).

•Among 8700 participants in a nested study of the PROMISE trial of coronary CT angiography (CTTA) versus functional testing for stable chest pain, CAC prevalence was evaluated among Black and White participants [60]. Despite having a lower burden of hypertension and diabetes, White participants had higher prevalence of CAC (63 versus 45 percent), as compared with Black participants. Notably, there was a lower proportion of females in the White versus Black subgroups (52 versus 60 percent), which could have accounted for some of the higher CAC among White participants.

●Extent of CAC

•Among males in the MESA study, CAC scores were highest in White and Hispanic Americans; the scores were lowest in African Americans at younger ages and Chinese persons at older ages [8]. For females, CAC scores were highest in White Americans and generally lowest in Hispanic American, although Chinese American females had the lowest scores in the oldest age group.

●Role of acculturation and socioeconomic status – Acculturation is defined as a person’s assimilation into a dominant culture, often in the context of immigration or movement into a new country or geography. In an additional study from MESA [59], the following findings were noted, which could not be explained by differences in smoking, body mass index, lipid levels, diabetes, or hypertension:

•After adjustment for age, sex, and income, not being born in the United States was associated with a lower prevalence of CAC in both Hispanic and African Americans (adjusted relative prevalence [RP] 0.75 and 0.89, respectively).

•For Chinese participants not born in the United States, the prevalence of CAC increased with the duration of residence in the United States (adjusted RP 1.06 for each 10 years of residence).

•Low education (no high school compared with completed college) was associated with a higher prevalence of CAC in White individuals (adjusted RP 1.17) but a trend toward a lower prevalence of CAC in Hispanic Americans.

Kidney disease — Vascular calcification is highly prevalent in patients with chronic kidney disease. Among patients with chronic kidney disease, although vascular calcification is associated with increased cardiovascular risk and mortality, its additional predictive value to other cardiovascular risk factors is not clear. This is discussed in detail separately. (See "Vascular calcification in chronic kidney disease", section on 'Coronary artery calcification'.)

Smoking — Among patients who smoke, CAC has only limited added utility beyond knowledge of established ASCVD risk factors. Furthermore, it is uncertain whether knowledge of CAC on lung cancer screening CT scans has clinical utility in those with a heavier smoking history who qualify for this type of screening. Among 6814 participants in MESA, the following findings were noted [61]:

●CAC was predictive of ASCVD but only modestly improved the C-statistic (discrimination of future events) above the pooled cohort ASCVD risk equations.

●A CAC score of 0 was less "protective" than in a general population; nearly 7 percent of smokers had an ASCVD event during follow-up.

●Among the higher-risk smokers who qualified for a lung cancer screening examination, CAC was not able to improve discrimination (ie, improve the C-statistic) of an ASCVD event over established risk factors. This finding is relevant because CAC can be obtained from the lung cancer screening examination even though the scan is not intended for CAC screening.

Further studies are needed to determine whether knowledge of CAC on lung cancer screening CT scans has clinical utility among this high-risk group. (See 'Incidental CAC' above and "Screening for lung cancer".)

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: Multimodality cardiovascular imaging appropriate use criteria".)

SUMMARY AND RECOMMENDATIONS

●Background – The presence of coronary artery calcification (CAC) in an asymptomatic patient is a validated measure of subclinical atherosclerosis. (See 'Introduction' above.)

•Advantages of CAC imaging compared with invasive or coronary CT angiography (CCTA) include minimal patient preparation, no iodinated intravenous contrast, and low effective radiation doses. (See 'Imaging CAC' above.)

•The Agatston score is used to quantify a CAC score. (See 'CAC quantification' above.)

•Providers can obtain the patient’s CAC percentile based on others of similar age, sex, and ethnicity using the Multiethnic Study of Atherosclerosis (MESA) calculator. (See 'CAC percentiles and calculators' above.)

●Clinical utility – CAC has clinical utility in atherosclerotic cardiovascular disease (ASCVD) risk prediction and can help guide ASCVD prevention treatment (eg, aspirin, statins, aggressive lifestyle modification). (See 'Overview of clinical utility' above.)

●Patients we screen for CAC – We screen for CAC in select asymptomatic adults ≥40 years of age at intermediate to high risk (7.5 to <20 percent 10-year ASCVD risk) by the American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort risk calculators (calculator 1). (See 'Patient selection for CAC screening' above.)

In select patients at borderline risk (5 to 7.5 percent 10-year ASCVD risk), CAC can be an arbiter of risk, especially for those in whom the presence of CAC may change decision-making or patient acceptance with regard to statin treatment and intensity of ASCVD risk factor modification. (See "Atherosclerotic cardiovascular disease risk assessment for primary prevention in adults: Our approach", section on 'Identify risk-enhancing factors'.)

●Patients we do not screen for CAC

•We do not screen patients at low (<5 percent) or very high risk (≥20 percent).

•We do not use CAC scoring as a primary diagnostic tool in patients with symptoms concerning for myocardial ischemia. (See 'Patient selection for CAC screening' above and "Noninvasive testing and imaging for diagnosis in patients at low to intermediate risk for acute coronary syndrome".)

●Management – We counsel all our patients to modify lifestyle and we manage ASCVD risk factors.

•In patients with CAC (Agatston) scores ≥100 or >75 percentile for age, sex, and race, we recommend treating with statins if the low density lipoprotein cholesterol (LDL-C) is between 100 and 190 mg/dL (Grade 1B). In these patients, we also consider low-dose aspirin therapy. (See 'CAC score greater than or equal to 100 or >75 percentile for age, sex, and race' above and "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease" and "Overview of primary prevention of cardiovascular disease".)

•We consider statins for patients with CAC scores between 1 to 99 or <75^th percentile and with LDL-C between 100 and 190 mg/dL based on patient and provider shared decision-making. (See 'CAC score 1 to 99 or <75th percentile for age, sex, and race' above.)

•Patients with a calcium score of 0 and no other compelling risk factors (eg, diabetes, active smoking, hypertension, family history of premature coronary disease) have a low 10-year ASCVD risk and do not require statin and aspirin therapy. We repeat risk stratification at five years. (See 'Follow-up based on CAC score' above and "Low-density lipoprotein cholesterol-lowering therapy in the primary prevention of cardiovascular disease".)

•For a patient who has incidental CAC noted on a non-gated chest CT obtained for other reasons, we use the same treatment approach compared with patients who have CAC scoring done intentionally. (See 'Incidental CAC' above.)

●Follow-up based on CAC score (see 'Follow-up based on CAC score' above)

•We repeat CAC scanning only if it will change treatment recommendations.

•Among patients with a baseline CAC score of 0, repeat testing may be performed in three to five years based on ASCVD risk. (See 'CAC of 0 at baseline' above.)

•We do not recommend serial CAC scoring in patients with established CAC (>0), particularly those with CAC >100, as repeated measurement will rarely alter management. We also recommend against serial CAC scoring to assess the efficacy of cholesterol lowering therapy.

●Prognostic value based on patient group – There are differences in the prevalence but not prognostic value of CAC in different patient groups. (See 'Prognostic value by patient group' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Thomas Gerber, MD, PhD, FACC, FAHA, who contributed to an earlier version of this topic review.