Finding the Right Age for CAC Testing: How Low Should You Go?

Tasneem Z. Naqvi, MD, MMM; Tamar S. Polonsky, MD, MSCI


J Am Coll Cardiol. 2021;78(16):1584-1586. 

"The aim of medicine is to prevent disease and prolong life; the ideal of medicine is to eliminate the need of a physician."
—William J. Mayo (1)

Atherosclerotic cardiovascular disease (ASCVD) risk assessment guidelines have focused on adults 40 to 75 years of age.[2] Age is one of the strongest risk factors for myocardial infarction (MI), and clinical risk stratification methods are heavily biased for age. However, the proportion of hospitalizations for acute MI attributable to younger adults (ages 35–54) has increased over time.[3] These trends are particularly pronounced among young women, who have experienced an increase in the prevalence of obesity and diabetes, and in whom the incidence of MI is increasing.[4] Screening strategies before middle age are therefore needed to help identify which patients are most vulnerable and would potentially benefit from early preventive treatment.

Coronary artery calcium testing (CAC) is recommended (Class IIa) in American guidelines for intermediate-risk adults (estimated 10-year ASCVD risk 7.5%–20%) for whom decisions about statin therapy for primary prevention are unclear. Given the typically low prevalence of CAC among adults with an estimated ASCVD risk <5%, it is usually not recommended as a routine screening tool in low-risk groups. However, data from population-based cohort studies suggest that CAC testing may be useful among selected populations of younger adults. For example, among 3,043 participants in the CARDIA (Coronary Artery Risk Development in Young Adults) study with a mean age 40.3, a CAC prediction score demonstrated that CAC prevalence was strongly associated with the burden of traditional risk factors.[5] During 12.5 years of follow-up, even low CAC scores were associated with the hazard of coronary heart disease (CHD) events. Those with CAC scores 1 to 19 experienced a 2.6-fold higher risk, and those with CAC scores 20 to 99 a 5.8-fold higher risk. The results highlight that although a CAC score >100 is considered higher risk for older adults, lower values have strong prognostic significance in younger adults.

In this issue of the Journal, Dzaye et al[6] sought to determine the ideal age at which CAC testing could be recommended among younger adults. The authors evaluated 22,346 patients aged 30 to 50 who underwent CAC testing between 2001 and 2010 for ASCVD risk stratification at 4 high-volume centers. Importantly, the study cohort was not a population-based sample but rather a group of younger adults enriched with traditional risk factors (49.3% with a family history of CHD, 49.6% with dyslipidemia, 11% who smoked, 20% with hypertension, 4% with diabetes), likely resembling those seen in a clinical setting. Because of their age, 93% had an estimated 10-year ASCVD risk <5%, but 34.4% had CAC scores >0 (by contrast, only 10.2% of CARDIA participants had CAC). The majority had scores <50.

The authors performed multivariable adjusted logistic regression stratified by sex, with age as the independent variable modeled as a restricted cubic spline. These associations were then graphed and stratified by the absence or presence of individual risk factors to demonstrate the age required to observe a 25% probability of CAC scores >0. Sex-stratified multivariable logistic regression models with traditional ASCVD risk factors were constructed, from which calculators were derived to compute the predicted probability of CAC score >0 according to age and risk factor burden. The authors then calculated the "offset period," which was how many years in advance an individual with a specific risk factor would experience a CAC score >0 compared with the same person without ASCVD risk factors.

The authors found that for a positive testing yield of 25% (4 patients needed to find one CAC score >0), the optimal age for a first scan for men without risk factors was 42 years; for women it was 58 years. The offset period for those with diabetes was 6.4 years, meaning that women with diabetes have a 1 in 4 chance of a CAC score >0 at age 50, whereas men with diabetes had the same chance as at age 36 to 37. Smoking, hypertension, dyslipidemia, and a family history of CHD were individually associated with offset periods of 3.3 to 4.3 years. The authors also provided estimates for ideal scanning ages based on combinations of risk factors: age 49 for women with a family history of CHD plus hypertension, age 45 for women with diabetes plus hyperlipidemia, age 38 for men with smoking plus hypertension, and age 34 for men with diabetes plus hyperlipidemia. The study has filled an important clinical gap, providing highly actionable data that could help guide clinical decision making for ASCVD prevention.

There were important limitations to the study, many of which the authors acknowledged. The study was cross sectional, and each participant had a single CAC scan; hence, data on first "conversion" of a CAC score 0 to a CAC score >0 were derived from different study participants. Only 25% of the cohort were women, 12.3% were non-White, and the study cohort was likely of higher socioeconomic status. It is therefore unclear how generalizable the data are to the broader population of the United States or non-Western countries. The type of diabetes among participants was unknown, and the prevalence of diabetes was low.

In addition, CAC screening at the authors' recommended ages could still miss a substantial number of young women with incident MI. The Atherosclerosis Risk In Communities Surveillance study demonstrated that among adults ages 35 to 54 hospitalized with MI, the mean age among women was 48.[3] Clinical variables such as obstetric history (pre-eclampsia, gestational diabetes) and conditions such as lupus or other rheumatologic diseases could prompt providers to screen even earlier. CAC develops in women with pre-eclampsia about 5 years earlier than in women without pre-eclampsia.[7] The authors acknowledged that a 25% testing yield is somewhat arbitrary, and one could use lower cutoffs. Future study should incorporate these nontraditional but important risk factors.

Exposure to ionizing radiation with CAC should be considered, particularly in young women. Recent technical advancements, however, have significantly reduced the exposure with CAC testing to a one-time radiation dose as low as 0.2 mSv, which is lower than that used in screening mammography for breast cancer (0.7 mSv) or CT scan for lung cancer (1 to 2 mSv). Alternatively, ultrasonography avoids radiation altogether and can detect plaque earlier than the development of CAC. Thus, in a low-risk population referred for CAC, 41% had CAC scores >0, whereas 59% had ultrasonography-detected carotid plaque, and 52% of participants with CAC scores of 0 among all risk groups had carotid plaque.[8] Accordingly, assessment of carotid artery and femoral plaque with ultrasound is given a Class IIa recommendation in the 2019 European Society of Cardiology guidelines on cardiovascular disease and a Class IIb recommendation for CAC[9] and diabetes.[10] However, plaque imaging by ultrasonography is not yet incorporated in the American guidelines. Techniques to quantitate 3-dimensional plaque burden by ultrasonography reliably, accurately, and rapidly are emerging and may provide an additional and safer risk stratification tool in younger adults. Ultrasonography has the additional advantage of repeated imaging to detect "conversion" to plaque among those with multiple risk factors and even "regression" of plaque in those using preventive treatment.

The use of preventive therapies among older adults who are at high risk for ASCVD has contributed to a substantial decline in the incidence and mortality from MI. Previous studies have shed light on potential long-term gains from prevention. Using data from the National Health Nutrition and Examination Survey, Pencina et al[11] estimated that the initiation of statin therapy in adults 40 to 49 years old with non–HDL-C ≥160 mg/dL could reduce their average predicted 30-year risk of 17.1% to 6.5% (absolute risk reduction 10.6%). However, the absolute risk reduction decreased to 7.2% if treatment was delayed until ages 50 to 59. The work by Dzaye et al[6] provides additional concrete information about how to target appropriate prevention strategies among younger adults who may derive the greatest benefit.