Author + information
- Received March 27, 2015
- Revision received June 2, 2015
- Accepted June 11, 2015
- Published online December 1, 2015.
- Seamus P. Whelton, MD, MPH∗∗ (, )
- Michael G. Silverman, MD†,
- John W. McEvoy, MB, BCh, BAO, MHS∗,
- Matthew J. Budoff, MD‡,
- Ron Blankstein, MD§,
- John Eng, MD‖,
- Roger S. Blumenthal, MD∗,
- Moyses Szklo, MD, DRPH¶,
- Khurram Nasir, MD, MPH∗,#∗∗,†† and
- Michael J. Blaha, MD, MPH∗,¶
- ∗Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Baltimore, Maryland
- †Division of Cardiology, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
- ‡Los Angeles Biomedical Research Institute at Harbor, University of California—Los Angeles, Los Angeles, California
- §Cardiovascular Division and Department of Radiology, Brigham and Women's Hospital, Boston, Massachusetts
- ‖Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, Maryland
- ¶Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
- #Center for Healthcare Advancement & Outcomes, Baptist Health South Florida, Miami, Florida
- ∗∗Department of Medicine Herbert Wertheim College of Medicine & Department of Epidemiology, Robert Stempel College of Public Health, Florida International University, Miami, Florida
- ††Miami Cardiac & Vascular Institute (MCVI), Baptist Health South Florida, Miami, Florida
- ↵∗Reprint requests and correspondence:
Dr. Seamus Whelton, Division of Cardiology, Johns Hopkins School of Medicine, 1800 Orleans Avenue, 7th Floor, Baltimore, Maryland 21287.
Objectives This study sought to determine the predictors of healthy arterial aging.
Background Long-term nondevelopment of coronary artery calcification (persistent CAC = 0) is a marker of healthy arterial aging. The predictors of this phenotype are not known.
Methods We analyzed 1,850 participants from MESA (Multi-Ethnic Study of Atherosclerosis) with baseline CAC = 0 who underwent a follow-up CAC scan at visit 5 (median 9.6 years after baseline). We examined the proportion with persistent CAC = 0 and calculated multivariable relative risks and area under the receiver operating characteristic curve for prediction of this healthy arterial aging phenotype.
Results We found that 55% of participants (n = 1,000) had persistent CAC = 0, and these individuals were significantly more likely to be younger, female, and have fewer traditional risk factors (RF). Participants with an ASCVD (Atherosclerotic Cardiovascular Disease Risk Score) risk score <2.5% were 53% more likely to have healthy arterial aging than were participants with an ASCVD score ≥7.5%. There was no significant association between the Healthy Lifestyle variables (body mass index, physical activity, Mediterranean diet, and never smoking) and persistent CAC = 0. The area under the receiver operating characteristic curve incorporating age, sex, and ethnicity was 0.65, indicating fair to poor discrimination. No single traditional RF or combination of other risk factors increased the area under the receiver operating characteristic curve by more than 0.05.
Conclusions Whereas participants free of traditional cardiovascular disease RF were significantly more likely to have persistent CAC = 0, there was no single RF or specific low-risk RF phenotype that markedly improved the discrimination of persistent CAC = 0 over demographic variables. Therefore, we conclude that healthy arterial aging may be predominantly influenced by the long-term maintenance of a low cardiovascular disease risk profile or yet to be determined genetic factors rather than the absence of any specific RF cluster identified in late adulthood.
Traditionally, aging has been thought of as a linear, chronological process. However, genetics and cumulative risk factor exposure can lead to significant variations in biological age for individuals of the same chronological age (1–3). This has prompted some to suggest substituting an individual’s biological or vascular age in place of chronological age for risk prediction models (4–6). Although the predictors of unhealthy arterial aging are well established, the predictors of healthy arterial aging are less certain, and it is unclear whether there are specific protective factors or a low-risk phenotype that exist for the maintenance of healthy arterial aging.
A finding of no coronary artery calcium (CAC) in a middle age or older adult is a marker of healthy arterial aging that is a powerful predictor of long-term all-cause survival and very low risk of cardiovascular disease (CVD) (7–9). In contrast, compared with an individual without CAC, individuals with a CAC score as low as 1 to 10 Agatston units are at more than 3× the risk for a hard coronary heart disease (CHD) event (10). Moreover, elderly patients without CAC have a better survival rate than do younger individuals with an elevated CAC score (11). As such, long-term, persistent maintenance of CAC = 0 can be considered an extreme manifestation of healthy arterial aging.
Just as it is important to clearly define high-risk phenotypes, it is also important to understand the predictors of healthy arterial aging in order to promote continued healthy aging. In this analysis, we sought to identify a phenotype for healthy arterial aging using routinely collected clinical variables.
We used data from the MESA (Multi-Ethnic Study of Atherosclerosis): a community-based, multiethnic, cohort free of CVD at baseline, which has been described in greater detail elsewhere (12).
CAC was measured in all participants at MESA visit 1 (2000 to 2002) and one-half of the participants were selected randomly for a CAC scan at MESA visit 5 (2010 to 2012). We excluded participants with any CAC at MESA visit 1 (n = 3,399) and those without a repeat CAC scan at MESA visit 5 (n = 1,566) for a total of 1,850 participants included in this analysis.
CAC scans were performed using an electron beam computed tomography scanner at the Chicago, Los Angeles, and New York field centers and using multidetector computed tomography at the Baltimore, Forsyth County, and St. Paul field centers. Participants were scanned twice during MESA visit 1 and once during MESA visit 5. Calcium scores were calculated using the Agatston method and calcium phantoms scanned alongside the participants were used to standardize results between field centers (13,14). Scans were uploaded and read at the Harbor-University of California, Los Angeles Research and Education Institute. There was an interobserver kappa value of 0.90 and interscan kappa value of 0.92 for the presence of coronary artery calcification (15).
Participant covariates were obtained from the baseline MESA clinical examination. We defined hypertension as a systolic blood pressure ≥140 mm Hg, a diastolic blood pressure ≥90 mm Hg, or use of a blood pressure–lowering medication; diabetes was defined as a fasting blood glucose ≥126 mg/dl, self-report of diabetes mellitus, or use of a blood glucose–lowering medication; smoking was classified as current, former, or never. Optimal body mass index (BMI) was defined as ≥18.5 kg/m2 to <25 kg/m2. Regular physical activity was defined as more than 150 min per week of moderate intensity physical activity or more than 75 min per week of vigorous intensity physical activity. We used a Mediterranean diet score as previously defined by Trichopoulou et al. (16). A family history of CHD was classified as a parent, sibling, or child with a history of CHD at any age. We defined socioeconomic status based on the highest level of achieved education and household income. The Lifestyle Score is the sum of optimal BMI, regular physical activity, adherence to a Mediterranean diet, and never smoking (17).
Persistent CAC = 0, our definition of healthy arterial aging, was the primary outcome for this analysis. In order to determine whether we could identify a healthy arterial aging phenotype, we calculated the absolute number and percentage of participants with persistent CAC = 0 stratified by high- and low-risk groups. We also used hazard ratios (HRs) and associated 95% confidence intervals (CIs) to assess whether these relationships were independent of other associated cardiovascular risk factors. The proportional hazards assumption was assessed using Schoenfeld residual testing and was not violated (18). Model 1 was adjusted for age, sex, and race. Model 2 was adjusted for age, sex, race, BMI, waist circumference, systolic and diastolic blood pressure, antihypertensive medication use, smoking status, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, lipid-lowering medication use, family history of CHD, household income, educational attainment, and reported physical activity. In addition, we calculated the area under the receiver operating characteristic curve (AUC) to further investigate whether we could identify and predict a persistent CAC = 0 phenotype.
We also performed sensitivity analyses incorporating the following novel risk factors to determine whether they helped identify a CAC = 0 phenotype: high-sensitivity C-reactive protein, N-terminal pro–B-type natriuretic peptide, creatinine, microalbuminuria, homocysteine, and fibrinogen.
Participants had a mean age of 57.1 ± 8.6 years, 37% were men, and there was a median of 9.6 years between CAC scans. We found that 55% of participants (n = 1,000) had persistent CAC = 0. Participants with persistent CAC = 0 were more likely to be younger, female, have a lower burden of CVD risk factors, and a lower ASCVD (Atherosclerotic Cardiovascular Disease) score (Table 1). They also were more likely to have an optimal BMI, never smoked cigarettes, and a higher Healthy Lifestyle score, but were not significantly more likely to report regular physical activity or consume a Mediterranean diet. Among the 45.9% of participants with CAC = 0 who developed incident CAC, the median CAC score was 25 (interquartile range: 9 to 65). There were 2 (0.2%) hard CHD events among participants with persistent CAC = 0 and 11 (1.3%) hard CHD events among participants with incident CAC.
Table 2 shows the percentage of participants with persistent CAC = 0 stratified by risk factors groups. In the fully adjusted model, participants <55 years of age at baseline were 39% more likely to have persistent CAC = 0 than were participants ≥65 years of age (p < 0.001) and women were 18% more likely to have persistent CAC = 0 than were men (p < 0.004) (Table 2). Participants with no traditional CVD risk factors were more likely to have persistent CAC = 0 and had a HR of 0.69 (95% CI: 0.54 to 0.90) than were participants with ≥3 risk factors. Individuals without diabetes mellitus were more likely to have persistent CAC = 0 and had a HR of 0.80 (95% CI: 0.68 to 0.94) than were individuals with diabetes mellitus. In the fully adjusted model, there was no significant difference in the likelihood of persistent CAC = 0 for any of the individual lifestyle score components or for the Lifestyle Score: HR 1.06 (95% CI: 0.88 to 1.27).
There was significant heterogeneity in persistent CAC = 0 of individuals who would be traditionally classified as having a high or low CVD risk. Among participants who would traditionally be classified as high risk for the development of CVD: 38% of participants age ≥65 years old, 36% of participants with an ASCVD ≥7.5%, and 36% of participants with ≥3 CVD risk factors had persistent CAC = 0 (Table 3). Conversely, among participants who would traditionally be classified as low risk for the development of CVD, 35% of participants were <55 years old; 40% had a low ASCVD of <2.5%; and 30% of participants with no traditional CVD risk factors did develop CAC (Figure 1).
Age, sex, and race were moderately predictive for persistent CAC = 0 with a C-statistic of 0.65 (Table 4). The addition of a participant’s ASCVD score or their CVD risk factors had a small improvement in the C-statistic of 0.011 (p < 0.001) and 0.051 (p < 0.001), respectively. There was no significant change after adding the Lifestyle score (p = 0.07). When added individually, the absence of any single CVD risk factor did not significantly increase the C-statistic (p > 0.05). Adding high-sensitivity C-reactive protein and other novel CVD risk factors to the model had a small, but statistically significant improvement in the C-statistic of 0.027 (p < 0.001).
There was no association between persistent CAC = 0 and a carotid intima-medial thickness <75% percentile; HR 0.95 (95% CI: 0.85 to 1.06) (Table 5). However, participants without carotid plaque by ultrasound were 11% more likely to have persistent CAC = 0 than were those with a carotid plaque (HR: 0.89; 95% CI: 0.80 to 0.99) and participants without thoracic artery calcification were 18% more likely to have persistent CAC = 0 than were individuals with extracoronary calcification of HR 0.82 (95% CI: 0.73 to 0.92).
In our sensitivity analyses we found was no significant difference in high-sensitivity C-reactive protein levels between participants with and without persistent CAC = 0 (Online Table 1). In addition, using the visit 5 variables in place of the baseline variables, the AUC estimates were essentially the same (Online Table 2).
Our analysis demonstrates that participants with persistent CAC = 0 had an overall lower cardiovascular risk factor profile. The healthy arterial aging group had a significantly lower level for all of the individual cardiovascular risk factors and a higher level for all of the healthy lifestyle factors, except Mediterranean diet. The absence of traditional CVD risk factors was also associated with persistent CAC = 0 when compared with participants with ≥3 CVD risk factors. However, there was no single modifiable traditional cardiovascular risk factor whose absence was strongly associated with healthy arterial aging. This is similar to a large cross-sectional cohort study of almost 17,000 participants by Boutouyrie et al. (19) in which sex, dyslipidemia, and smoking were not significantly associated with arterial aging, which was classified as an increased central arterial stiffness. In addition, Lehmann et al. (20) found a significant inverse trend between the number of cardiovascular risk factors and healthy arterial aging.
Although our results did not identify an individual predictor or specific phenotype of healthy arterial aging, it is important to understand that this does not exclude the existence of such predictors or phenotypes. Our analysis focused on using traditional CVD risk factors measured in late adulthood to examine healthy arterial aging. These risk factors are well established to be individually and collectively predictive of CVD events. Our analysis showed that the absence of these risk factors collectively was associated with healthy arterial aging, but that the absence of the risk factors individually was not associated with healthy arterial aging.
This finding suggests that although modifying CVD risk factors individually may lead to a lower burden of CAC and, hence, risk of CVD; we did not identify a specific CVD risk factor measured in late adulthood in which an improvement would be expected to predict continued healthy arterial aging. This does not preclude the existence of a predictor of healthy arterial aging, but it suggests that there may be other factors that we did not examine, such as genetics.
We did not observe the same consistent difference in the Lifestyle score variables between the healthy and nonhealthy arterial aging groups. Neither the individual Lifestyle score variables nor the aggregate Healthy Lifestyle score were significantly associated with healthy arterial aging after multivariable adjustment. It is possible that these variables measured at MESA visit 1 may not be representative of the participants’ true lifetime risk status or the self-reported physical activity, and Mediterranean diet variables may incompletely define those variables. However, CVD risk factors earlier in life are predictive of CAC in middle age and a lower Healthy Lifestyle was associated with a reduced risk of CHD in a separate analysis (17,21).
Our analysis showed that whereas a lower overall CVD risk profile was associated with healthy arterial aging, there was still a large amount of heterogeneity within individuals who would traditionally be classified as low or high risk for CVD. For example, among participants with no CVD risk factors, approximately one-third did not have continued healthy arterial aging over the course of the almost 10-year follow-up, and among participants with ≥3 cardiovascular risk factors, more than one-third had healthy arterial aging during the same follow-up. Furthermore, among individuals ≥65 years old more than one-third had persistent CAC = 0 during follow-up, at which time they had a mean age of approximately 79 years compared with approximately one-third of participants <55 years old who had unhealthy arterial aging by the mean age of 59 years at MESA visit 5.
Regardless of the definition of high and low risk, approximately one-third of the high-risk participants had healthy arterial aging versus one-third of low-risk participants who had unhealthy arterial aging. Within the majority of each high- and low-risk group there were also minimal differences in the number of CVD risk factors, mean ASCVD scores, and Lifestyle score variables. Further emphasizing the challenge of predicting healthy arterial aging, we found that age, sex, and race provided the vast majority of the prediction in AUC models, although with an AUC of 0.65 these demographic variables were only poorly predictive of healthy arterial aging. The addition of further traditional and novel variables to the models provided statistically significant improvements, but the absolute improvements were small enough that the models remained poorly predictive.
Even when we examined imaging of other vasculature as a predictor of healthy arterial aging there was still significant heterogeneity as 42% of participants with any carotid plaque had persistent CAC = 0 and only 59% of those without carotid plaque had healthy arterial aging. Although we found that the association between imaging of other vasculature beds was more strongly associated with healthy arterial aging than traditional risk factors, the imaging variables had a smaller improvement in the AUC score than did traditional risk factors as a continuous variable.
Thus, the heterogeneity of arterial aging was one of the most important findings of this analysis. It may be attributable to the fact that even though the presence of a CVD risk factor is associated with the development of CVD, its absence is not necessarily associated with healthy arterial aging. This is reflected in most cardiovascular risk scores by the virtue that CVD risk factors add points, but typically only an elevated high-density lipoprotein cholesterol is used to subtract points. It may also be that despite using nontraditional cardiovascular risk factors including noninvasive imaging methods, we may not be measuring the true determinants of arterial aging.
Strengths of this analysis include that participants in this analysis were middle to older age at baseline and there was a nearly 10-year median follow-up time allowing for the investigation of long-term arterial aging. In addition, we examined both traditional and novel cardiovascular risk factors and noninvasive imaging modalities.
One limitation of this study is that the follow-up CAC scans were performed approximately 10 years apart, and we do not know precisely when during the follow-up time individuals developed CAC. We also did not examine the initiation of new medications or directly investigate incident CVD. However, <2% of participants had a CVD event over the nearly 10-year follow-up and the goal of our analysis was to identify predictors of long-term healthy arterial aging measured by CAC.
Whereas the risk factors for the development of CVD are firmly established, the predictors of healthy arterial aging are less well defined. In this analysis we demonstrate that a low CVD risk factor burden is associated with an increased likelihood of healthy arterial aging as measured by CAC. However, the absence of any one individual CVD risk factor was not strongly predictive of healthy arterial again and there was significant heterogeneity among the participants with healthy arterial aging. Consequently, healthy arterial aging may be predominantly influenced by the long-term maintenance of a low CVD risk profile rather than the absence of any specific RF or low-risk phenotype identified in adulthood. The prediction of individuals who will continue to have healthy arterial aging remains elusive given the current available risk factors and further work using other markers such as genetics is needed to determine whether a healthy aging phenotype can be identified.
COMPETENCY IN MEDICAL KNOWLEDGE: The long-term nondevelopment of CAC is a marker of healthy arterial aging. Participants without a single traditional cardiovascular risk factor were more likely to have healthy arterial aging, although there was no single risk factor whose absence significantly improved the predication of healthy arterial aging.
TRANSLATIONAL OUTLOOK: There was no single traditional cardiovascular risk factor or group of risk factors whose absence significantly improved the prediction of healthy arterial aging. Future studies are needed to determine whether there are other factors such as genetics that may improve the prediction of healthy arterial aging.
The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.
For supplemental tables, please see the online version of this article.
This research was supported by contracts N01-HC-95159 through N01-HC-95169 from the National Heart, Lung, and Blood Institute and by grants UL1-RR-024156 and UL1-RR-025005 from National Center for Research Resources. This publication was developed under a STAR research assistance agreement, number RD831697 (MESA Air), awarded by the U.S. Environmental Protection Agency (EPA). It has not been formally reviewed by the EPA. The views expressed in this document are solely those of the authors and the EPA does not endorse any products or commercial services mentioned in this publication. Dr. Budoff is on the Speakers Bureau of GE Healthcare. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- area under the receiver operating characteristic curve
- body mass index
- coronary artery calcium
- coronary heart disease
- confidence interval
- cardiovascular disease
- hazard ratio
- Received March 27, 2015.
- Revision received June 2, 2015.
- Accepted June 11, 2015.
- 2015 American College of Cardiology Foundation
- Blaha M.,
- Budoff M.J.,
- Shaw L.J.,
- et al.
- Hermann D.M.,
- Gronewold J.,
- Lehmann N.,
- et al.,
- for the Heinz Nixdorf Recall Study Investigative Group
- McClelland R.L.,
- Chung H.,
- Detrano R.,
- Post W.,
- Kronmal R.A.
- Agatston A.S.,
- Janowitz W.R.,
- Hildner F.J.,
- Zusmer N.R.,
- Viamonte M. Jr..,
- Detrano R.
- Carr J.J.,
- Nelson J.C.,
- Wong N.D.,
- et al.
- Hartiala O.,
- Magnussen C.G.,
- Kajander S.,
- et al.