Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Bruno Bezerra Lima

Advanced Cardiac Imaging · VUMC

View

Dr. Sara Ersozlu

Advanced Cardiac Imaging · MGH

View

Dr. Christopher Kramer

Advanced Cardiac Imaging · University of Virginia

View

Dr. Filipe Moura

Advanced Cardiac Imaging · Yale College

View

Dr. Robert Hendel

Advanced Cardiac Imaging · Tulane University

View

Dr. Venkatesh Murthy

Advanced Cardiac Imaging · UMHS

View

Dr. Dorairaj Prabhakaran

General Cardiology

View

Dr. Pugazhendhi Vijayaraman

EP · Geisinger

View

Dr. Robert Schaller

EP · HUP

View

Dr. Trey Woods

General Cardiology · Emory Healthcare

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Race- and Sex-Specific Age at Which Coronary Artery Calcium Becomes Detectable Among Young Adults.

    Mar 3, 2025DOI: 10.1161/circimaging.124.016599Journal ArticleAHA

    Written by: Yvette A Yeboah-KordiehEllen BoakyeAlbert D OseiOmar DzayeZeina A DardariJoao A C LimaAlan RozanskiDaniel S BermanMatthew J BudoffMichael D MiedemaKhurram NasirJohn A RumbergerLeslee J ShawDavid R JacobsMichael J Blaha

    Find these authors on Roon

    YY

    Dr. Yvette Yeboah-Kordieh

    General Pulmonology · ISMMS

    AO

    Dr. Albert Osei

    General Cardiology · UPMC Mercy

    JR

    Dr. John Rumberger

    Advanced Cardiac Imaging

    Introduction:

    An estimated 127.9 million US adults aged ≥20 years have cardiovascular disease which includes coronary heart disease(CHD), according to data from the National Health and Nutrition Examination Survey(NHANES 2017–2020).1 Among young adults aged 18 – 44 years in the United States, the prevalence of CHD was estimated to be 0.5% in 2023.2Atherosclerosis is the primary pathophysiologic mechanism underlying CHD.3 The rate of progression of atherosclerosis is driven by the presence and number of risk factors such as hypertension, dyslipidemia, diabetes, cigarette smoking, and a family history of atherosclerotic cardiovascular disease(ASCVD).4,5 Therefore, early identification of subclinical CHD is essential to allow timely intervention to slow the progression to overt CHD such as myocardial infarction.Coronary artery calcium(CAC) is a marker of subclinical coronary atherosclerosis and vascular age.6,7 CAC effectively stratifies ASCVD risk across different populations.8 The absence of CAC is one of the strongest negative risk markers for future cardiovascular events which has led to the concept of “Power of Zero”.6,9 Conversely, individuals with non-zero CAC, particularly those with higher CAC burden(≥100 Agatston Units[AU]) are at an increased risk of CHD, cardiovascular death, and all-cause mortality.10,11Our understanding of the predictive utility of CAC among young adults is evolving. Significant differences in CAC distribution have been observed among diverse race and sex groups, even when they share similar risk factor profiles. Higher CAC scores are associated with increasing age, male gender, and White race, emphasizing the importance of independently studying predicted ages across different age and sex categories.12–14 Additionally, individuals with any of the ASCVD risk factors have consistently demonstrated higher CAC scores compared to counterparts without such risk factors in historical analyses.15,16We hypothesized that the age at which CAC becomes detectable in a broadly representative population of young adults would vary by sex, race, and the presence and number of ASCVD risk factors. Therefore, using a combined cohort(CARDIA and CAC Consortium), we aimed to 1) assess the prevalence of CAC by the number of ASCVD risk factors; and 2) model the race-and sex-specific ages for CAC conversion among young adults based on the presence and number of ASCVD risk factors.

    Methods:

    A subset of the data may be available upon reasonable request.In this cross-sectional study, we used combined data from the CAC Consortium and the CARDIA study. The CAC Consortium is a multicenter retrospectively assembled cohort of persons aged 18 years or older without baseline CVD who were referred for CAC scan for cardiovascular risk stratification. The individuals in the CAC Consortium were recruited from four(4) centers in the US between 1991 and 2010: Cedars-Sinai Medical Center(Los Angeles, California), PrevaHealth Wellness Diagnostic Center(Columbus, Ohio), Harbor UCLA Medical Center(Torrance, California), and Minneapolis Heart Institute(Minneapolis, Minnesota).17 We included 13,473 individuals aged 30–50 years from the CAC Consortium who had reported CAC scores in this study. For the CAC Consortium, participants were excluded if they had missing scan identifiers, impossible age/scan date, non-dedicated CAC score, other non-CAC CT scans, or insufficient data for death ascertainment.The CARDIA study is an NHLBI-funded prospective population based study that recruited individuals aged 18 to 30 years from 1985 to 1986 with a 35 year follow-up to study CHD risk factor evolution among young adults.18,19 Participants in the CARDIA study were approximately equally distributed in the different race(Black and White) and sex categories across 4 study sites in Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California.18 Coronary artery calcium was measured at study years 15, 20, and 25.18We included 3,812 participants aged 30–50 years from CARDIA and we used only risk factor data obtained at the time of their first CAC scans. Participants from both cohorts were asymptomatic at the time of CAC testing.In CARDIA and CAC Consortium, a standard protocol was used to quantify CAC using non-contrast, ECG-gated cardiac CT at all participating centers.13,20 CAC was measured in the CAC Consortium using mostly electron-beam CT scans(93%) and multi-detector CT scans(7%). In CARDIA, electron-beam CT scans were used at two(2) sites and multi-detector CT scans were used at the other two(2) sites. Previous studies show no significant difference in CAC results obtained from electron-beam CT scans versus multi-detector CT scans.21 CAC was defined as lesions with attenuation ≥130 Hounsfield units[HU] and area ≥1mm2 at each slice level in any of the coronary arteries and quantified using the Agatston score(in Agatston Units, AU).20Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, prior diagnosis of hypertension, or treatment with anti-hypertensive medication.17,22 Blood pressure taken at the time of CAC scanning was not used to override the diagnosis of hypertension in the CAC Consortium.17 Diabetes was defined as a fasting glucose level ≥126 mg/dL, a 2-hour glucose level of ≥200 mg/dL, a prior diagnosis of diabetes or treatment with oral hypoglycemic drugs or insulin.17,22 Dyslipidemia was defined by a prior diagnosis of dyslipidemia, or prior diagnosis of primary hyperlipidemia or treatment with any lipid-lowering medication, or LDL-C >160mg/dL, fasting triglycerides ≥150 mg/dL, high-density lipoprotein(HDL-C) <35 mg/dL in men and <45 mg/dL in women in CARDIA, or HDL-C <40 mg/dL in men and <50 mg/dL in women in the CAC Consortium.17,22 Smoking was categorized as current smoking or not. Family history of CHD was determined by the presence of a first-degree relative with a history of CHD. However CAC Consortium, PrevaHealth Wellness Diagnostic Center(Columbus, Ohio) and Cedars-Sinai Medical Center(Los Angeles, California) used a definition of premature family history of CHD(<55 years in a male relative and <65 years in a female relative).17Baseline characteristics of the 17,285 study participants were summarized by race-sex groups. Additionally, we graphically presented the number of ASCVD risk factors, and then also the CAC prevalence by the number of ASCVD risk factors, first overall and then for each race-sex group.For the remaining analyses, the 2 cohorts(CAC Consortium and CARDIA) were weighted using raking/sample harmonization so that each cohort contributed approximately half of the final analytic results. Doing this provided a more representative sample of community-based CARDIA and referred CAC Consortium population. Due to an insufficient number of Asians, Hispanics, and individuals from other races, these groups were excluded from the analysis. The weighted values were used throughout to achieve balanced contributions from the cohort, although this was not fully possible with White men due to the larger imbalance in population size between the cohorts.In each race group, two main multivariable logistic regression models were generated to estimate the probability of CAC >0, one based on age and each ASCVD risk factor, and the other based on age and the number of ASCVD risk factors(0, 1, 2, and ≥3 risk factors). Then, within each of the 8 distinct regression analyses, we modeled the predicted probability of CAC >0 as a function of age(continuous) and the number or the profile of ASCVD risk factors present.To obtain the estimated age to assess for CAC, the derived equations were used to back-calculate the age at which a prespecified proportion of young adults, in this case, 25%(number needed to scan [NNS] of 4), would have CAC >0, based on our prevalence data and prior studies.20,23,24 This was done by substituting the estimated probability of 0.25 into the logistic regression equation and substituting 1(if the risk factor is present) or 0(if the risk factor is absent) for the multiplier of the risk factor coefficients or number of risk factors to calculate the age. The coefficients for the equations derived from the logistic regression models for each race-sex group and example calculations for suggested age for CAC conversion based on the equations are provided in Supplemental Tables 1 and 2. We calculated 95% confidence intervals(CI) by using the margins command in stata after fitting each individual regression model and evaluating at a 25% prevalence of CAC>0. Finally, we defined an offset period to predict the age at which an individual with a set of risk factors would have detectable CAC compared to an individual with no risk factor.23 This was obtained mathematically by the difference in predicted ages between an individual with a specified risk factor profile and an individual of the same race and sex with no risk factors.23All statistical analyses were conducted using STATA 17(Stata Corp). IRB approval was obtained.

    Results

    Of the 17,285 individuals with a mean age of 43.3(SD ± 4.7) years included in this study, 66.9% were men, 87.5% were White, and 12.4% were Black. Dyslipidemia(51.7%) and family history of CHD(47.5%) were the most prevalent ASCVD risk factors, followed by hypertension(16.1%), current smoking(13.4%), and diabetes(3.4 %)(Table 1). Men had higher mean CAC score and a greater proportion of individuals with hypertension, and dyslipidemia compared to women. Also, the number of ASCVD risk factors varied by race and sex with White men having the greatest proportion of individuals with at least 1 risk factor(Figure 1a).Approximately 30.1 %(5,196 out of 17,285) had CAC > 0 AU, and of them, the median CAC was 21 AU(IQI: 4.6–81 AU)(Table 1). Most of the participants with CAC had CAC scores between 1 and 100 AU, with a small proportion having CAC >300 AU(Table 1). The probability of CAC increased significantly with age across all race-sex groups and was higher in men than in women(Figure 2). Similarly, the prevalence and predicted probabilities of CAC increased with increasing number of ASCVD risk factors for each race-sex group(Figure 1b and Figure 1c). Compared to Black women, White women had a higher predicted probability of CAC with each ASCVD risk factor except diabetes(Figure 3 and Figure S1).The estimated age at which CAC became detectable differed by sex, race, and ASCVD risk factor burden. Increasing number of ASCVD risk factors was associated with an earlier age for CAC conversion(Table 2, and Figure S2). The estimated age at which non-zero CAC became common for White men, Black men, White women, and Black women without any of the specified ASCVD risk factors was 43.0 years(95% CI: 42.1–43.9 years), 50.2 years(95% CI: 47.8–52.2 years), 64.1 years(95% CI: 59.7–67.8 years), and 65.9 years(95% CI: 57.1–72.0 years), respectively(Table 2). For individuals with at least three(3) risk factors, the estimated ages were earlier: 32.4 years(95% CI: 30.9 – 33.7 years), 36.7 years(95% CI: 34.6–38.6 years), 41.5 years(95% CI: 39.5 – 43.3 years), and 42.6 years(95% CI: 40.2 – 44.8 years) respectively for White men, Black men, White women, and Black women(Table 2). White men generally had an earlier estimated age for CAC testing than Black men, with the presence of any of the other ASCVD risk factors.The single risk factor that predicted the earliest age at CAC conversion differed among the race-sex groups. Current smoking, as a single risk factor, predicted the earliest age at which CAC became detectable among Black men(41.8 years, 95% CI: 40.5–43.1 years) while hypertension, as a single risk factor, predicted the earliest age for CAC conversion among White women(51.9 years, 95% CI: 49.2–54.2 years)(Table 2). White men had the earliest estimated CAC conversion age among those without any of the specified ASCVD risk factors. Risk factor combinations produced a higher offset in general in women compared to men. For instance, for hypertension and diabetes, White women, and Black women had an offset age of −15.5, and −21.2 respectively, compared to −6.1 and −9.2 for White men and Black men respectively(Table 2).

    Discussion

    In this cross-sectional analysis using a large cohort of young adults, we observed that approximately one-third had CAC > 0 AU, and among them, the majority had CAC scores ranging from 1 to 100 AU. The predicted probabilities of non-zero CAC increased with increasing age and the number of ASCVD risk factors. Also, the estimated age at which CAC transitioned to non-zero(i.e., 25% CAC prevalence) differed by race, sex, and specific ASCVD risk profiles. Generally, White men had an earlier estimated age of CAC conversion compared to Black men with the same risk factor profile except for current smoking. Similarly, White women had earlier age for CAC conversion compared to Black women with the same risk factor profiles except for diabetes.We assumed a 25% yield was “ideal”, which may be viewed by some as conservative. For example, some clinicians and some patients might prefer to detect CAC earlier(i.e., at a NNS >4), while others may prefer to only order a CAC test when the expected yield would be higher(i.e., NNS<4). This decision has implications for groups that may be at higher or lower than expected risk beyond CAC results, for example Black individuals who generally have higher event rates despite lower average CAC scores(see Table S3 for additional calculations using NNS=5 for Black individuals). On an individual basis in the clinical setting, this should be guided by an evidence-based patient-physician discussion, which can begin with the data we have presented here. The absence of CAC should not be used as a rationale to avoid treating early onset risk factors in groups known to be at especially high risk, for example hypertension in younger Black people.In a recent study by Dzaye et al. using data from young adults aged 30–50 years in the CAC Consortium, a predominantly White clinical cohort, the predicted probability of CAC conversion at any given age was higher among men than women.23 Among those with no ASCVD risk factor, CAC conversion was predicted to occur at 42.3 years and 57.6 years in men and women, respectively.23 Similar to the aforementioned study by Dzaye et al., we found that the predicted age for the CAC conversion in women was, on average, 12 years later than in men with similar risk factor profiles. This correlates with findings that the development of CAC tends to lag by an average of 10 years in women compared to men.16 Although the study by Dzaye et al. contributed to filling a significant knowledge gap in our understanding of the suggested age at which CAC transitions from zero among young adults, the suggested age by race and number of risk factors were not assessed. The suggested ages differed significantly for Black individuals per our study. Existing data confirms differences in the burden and progression of CAC across race/ethnicity.25,26 This may contribute to the different ages for CAC conversion by race and hence it is essential that the age at which CAC is first detected be stratified by not only sex but also by race.The predicted probability of developing CAC varied across risk factor groups and may highlight the differential effects of the various ASCVD risk factors on the development of premature atherosclerosis.27,28 A study among patients undergoing coronary angiography found that of all the traditional ASCVD risk factors, systolic blood pressure displayed the strongest association with the degree of CAC.29 Another study in a predominantly White population showed that diabetes had the greatest association with conversion to non-zero CAC.23 However, we found that no one single risk factor was most strongly associated with earlier CAC conversion in all four race-sex groups, reflecting the multiple risk factor nature of atherosclerosis in general.The age for CAC conversion among young adults by race, sex, and ASCVD risk factor profile has not been extensively studied. Combining data from the CAC Consortium and CARDIA allowed the generation of sex-and race-specific estimates of the suggested ages at which CAC becomes first detectable in a racially diverse cohort of young adults aged 30–50 years with no known history of cardiovascular disease. The inclusion of the more ethnically diverse CARDIA study participants allowed us to extend the modeling of ages for CAC conversion from zero in Black individuals. This cohort also combines participants from a general community-based population(CARDIA) and a clinical cohort referred for CAC scanning(CAC Consortium), making our results more generalizable to the general population. By combining the two data sets, we leverage on their inherent heterogeneity to produce a final analytic dataset that may be more balanced in terms of the relative potential selection biases, with the possibility of bias towards higher CAC in referred patients from CAC Consortium, and a reverse bias of healthier population in CARDIA.28 Though weighting does not remove the impact of all heterogeneity, it is likely that there is no perfect data set for our research question.As hypothesized, we observed significant differences in the age for the conversion of CAC by race and sex as well as by the number and type of ASCVD risk factors. Separate prediction models based on the number and different combinations of ASCVD risk factors were derived to allow for the versatile use of these models. The derived equations in this study, based on sex, race, type, and the number of risk factors, provided the predicted probability of non-zero CAC and the suggested age to at which CAC transitions to >0 among young adults. Given that age significantly influences the ASCVD risk estimated by the pooled cohort equation, this may underestimate the cardiovascular risk among young adults. Our finding of an earlier age at which CAC becomes detectable based on sex, race, and risk factor burden can be incorporated into future guidelines to identify patients at greatest risk of early or premature atherosclerosis based on their CAC scores.Our study had some limitations. First, we did not have enough data to estimate the ages for CAC conversion in Hispanics, East Asians, South Asians, Native Americans, or other racial-ethnic groups, although differences in the prevalence of CAC among these groups have been previously described.26 We recommend additional studies in diverse populations with similar age groups to compare our estimated ages for CAC screening as part of cross validation. Also, we did not have information on lipoprotein(a), polygenic risk, and other risk enhancing factors that may modulate the decision of when to assess CAC in individual patients. Most of the CAC scanning sample was done using electron beam CT, although most modern scanning is done on multi-detector row CT(MDCT). Additionally, the pre-defined variables in the dataset used in this study limited our ability to choose more stringent cutoffs in defining the CVD risk factors. For example, LDL-C threshold of >160 was used reflecting studies that show that maintaining LDL-C levels >160 is a risk enhancer for ASCVD.Furthermore, in a select small group of young patients, there may be coronary artery stenosis from non-calcified plaque for which CAC may not suitably predict their risk for major adverse cardiovascular events. Existing studies have showed that among those with zero CAC, less than 0.8% had significant stenosis, as defined by more than 50% stenosis on CT coronary angiogram.30–32 This may however vary among symptomatic patients and in other study populations. While further testing in asymptomatic younger patients with zero CAC score should be tailored to the specific clinical scenario, additional testing in the general clinic population is not recommended. CAC, especially, a CAC of zero vs. non-zero, remains a great risk stratification tool to identify populations that would most benefit from early CVD prevention strategies such as statin initiation.Finally, the estimated ages for CAC conversion suggested in this paper should be interpreted cautiously, as these ages may not necessarily be fully correlated with the onset of major adverse cardiac events in different races(i.e., in select situations, clinicians may choose to test for CAC sooner in Blacks, who may experience higher event rates at a given CAC score). Conversely, like many screening tests, there is a risk of potential overdiagnosis and psychological impacts of false positives. It is crucial therefore to engage in evidence-based patient-physician discussion and individualize risk before testing.

    Conclusion

    Our results suggest that race, sex, and ASCVD risk factor burden can inform the estimated age for CAC conversion in young asymptomatic adults and will be useful in clinical practice and future guidelines to identify patients at increased risk of early or premature atherosclerosis to guide initiation of statin therapy.However, since the absence of CAC does not preclude the presence of the earliest manifestations of atherosclerosis, future studies should examine whether other subclinical measures of atherosclerosis provide any additional meaningful value for identifying young patients with early non-calcified atherosclerosis.
    JRDr. John Rumberger

    Dr. John Rumberger

    Cardiology · Advanced Cardiac Imaging
    johnr