Donate Help Contact The AHA Sign In Home
American Heart Association
Circulation
Search: search_blue_button Advanced Search
« Previous Article | Table of Contents | Next Article »
Circulation. 2007;115:2722-2730
Published online before print May 14, 2007, doi: 10.1161/CIRCULATIONAHA.106.674143
CLINICAL PERSPECTIVE
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
115/21/2722    most recent
CIRCULATIONAHA.106.674143v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kronmal, R. A.
Right arrow Articles by Burke, G. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kronmal, R. A.
Right arrow Articles by Burke, G. L.
Right arrowPubmed/NCBI databases
Medline Plus Health Information
*Coronary Artery Disease
Related Collections
Right arrow Epidemiology
Right arrow Risk Factors
Right arrow Imaging
Right arrow Computerized tomography and Magnetic Resonance Imaging
Right arrowRelated Article

(Circulation. 2007;115:2722-2730.)
© 2007 American Heart Association, Inc.

Epidemiology

Risk Factors for the Progression of Coronary Artery Calcification in Asymptomatic Subjects

Results From the Multi-Ethnic Study of Atherosclerosis (MESA)

Richard A. Kronmal, PhD; Robyn L. McClelland, PhD; Robert Detrano, MD; Steven Shea, MD, MS; João A. Lima, MD; Mary Cushman, MD, MS; Diane E. Bild, MD, MPH; Gregory L. Burke, MD

From the Department of Biostatistics, University of Washington, Seattle (R.A.K., R.L.M.); Division of Cardiology, Harbor–UCLA Medical Center, Los Angeles, Calif (R.D.); Departments of Medicine and Epidemiology, Columbia University, New York, NY (S.S.); Department of Cardiology, Johns Hopkins University, Baltimore, Md (J.A.L.); Department of Medicine, University of Vermont, Burlington (M.C.); National Institutes of Health, Bethesda, Md (D.E.B.); and Department of Public Health Sciences, Wake Forest University, Winston-Salem, NC (G.L.B.).

Correspondence to Robyn McClelland, Collaborative Health Studies Coordinating Center, Department of Biostatistics, University of Washington, Bldg 29, Suite 310, 6200 NE 74th St, Seattle, WA 98115. E-mail rmcclell{at}u.washington.edu

Received November 6, 2006; accepted March 26, 2007.


*    Abstract
up arrowTop
 *Abstract
down arrowIntroduction
 down arrowMethods
 down arrowResults
 down arrowDiscussion
 down arrowConclusions
 down arrowReferences
 
Background— The Multi-Ethnic Study of Atherosclerosis (MESA) provides an opportunity to study the association of traditional cardiovascular risk factors with the incidence and progression of coronary artery calcium (CAC) in a large community-based cohort with no evidence of clinical cardiovascular disease.

Methods and Results— Follow-up CAC measurements were available for 5756 participants with an average of 2.4 years between scans. The incidence of newly detectable CAC averaged 6.6% per year. Incidence increased steadily across age, ranging from <5% annually in those <50 years of age to >12% in those >80 years of age. Median annual change in CAC for those with existing calcification at baseline was 14 Agatston units for women and 21 Agatston units for men. Most traditional cardiovascular risk factors were associated with both the risk of developing new incident coronary calcium and increases in existing calcification. These included age, male gender, white race/ethnicity, hypertension, body mass index, diabetes mellitus, glucose, and family history of heart attack. Factors also existed that were related only to incident CAC risk, such as low- and high-density lipoprotein cholesterol and creatinine. Diabetes mellitus had the strongest association with CAC progression for blacks and the weakest for Hispanics, with intermediate associations for whites and Chinese.

Conclusions— This is the first large multiethnic study reporting on the incidence and progression of CAC. Standard coronary risk factors were generally related to both CAC incidence and progression. Whites had more incident CAC and CAC progression than the other 3 racial/ethnic groups. Except for diabetes mellitus, risk factor relationships were similar across racial/ethnic groups.


Key Words: arteries • calcium • coronary disease • epidemiology • imaging • risk factors


*    Introduction
up arrowTop
 up arrowAbstract
 *Introduction
down arrowMethods
 down arrowResults
 down arrowDiscussion
 down arrowConclusions
 down arrowReferences
 
Coronary artery calcium (CAC) has been shown to be associated with the risk of coronary events.1–9 A handful of studies to date have examined risk factors for the progression of CAC.10–19 In a retrospective study of asymptomatic individuals, Yoon et al10 reported that hypertension and diabetes mellitus were significantly associated with a greater magnitude of progression. Budoff et al11 reported on asymptomatic but referred subjects and found that statin users had significantly less progression compared with untreated hypercholesterolemics. Similarly, Callister et al12 reported an apparent progression-reducing effect of statins and a positive association between low-density lipoprotein (LDL) cholesterol level and progression. In contrast, Wong et al13 found that CAC progression was inversely related to high-density lipoprotein (HDL) cholesterol but not to LDL cholesterol or lipid-lowering therapy. Cassidy et al14 found that among those at lower coronary heart disease risk, various obesity measures were associated with progression of CAC. Bursztyn et al15 found that among high-risk hypertensive patients, mild renal dysfunction was associated with faster CAC progression.

Clinical Perspective p 2730

These previous studies on CAC progression have been limited by issues such as small sample sizes,10–12,15–19 retrospective designs,10–12,17 and/or restricted populations such as high-risk hypertensives15,16 or familial hypercholesterolemics.18 Certain studies consisted predominantly of men,11,13,18 and many focused only on particular risk factors of interest such as cholesterol levels or statin use,12,13,18 obesity,14 or renal dysfunction.15 The Multi-Ethnic Study of Atherosclerosis (MESA) provides a unique opportunity to study the risk factors for progression of coronary artery calcification in a large, gender-balanced, multiethnic, nonreferred, asymptomatic cohort. We consider here traditional cardiovascular risk factors as they relate to the risk of developing incident detectable CAC and to the magnitude of CAC progression given that some calcium is present.


*    Methods
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 *Methods
down arrowResults
 down arrowDiscussion
 down arrowConclusions
 down arrowReferences
 
Study Sample
MESA is designed to study the prevalence, risk factors, and progression of subclinical cardiovascular disease in a multiethnic cohort. A detailed description of the study design and methods has been published previously.20 Briefly, 6814 participants 45 to 84 years of age who identified themselves as white, black, Hispanic, or Chinese were recruited from 6 US communities between 2000 and 2002. All participants were free of clinically apparent cardiovascular disease (having physician-diagnosed heart attack, angina, stroke, transient ischemic attack, or heart failure; having current atrial fibrillation; taking nitroglycerin; or having undergone angioplasty, coronary artery bypass graft, valve replacement, pacemaker or defibrillator implantation, or any surgery on the heart or arteries). The communities were Forsyth County, NC; Northern Manhattan and the Bronx, NY; Baltimore and Baltimore County, Md; St Paul, Minn; Chicago, Ill; and Los Angeles County, Calif. The source population for each field center varied in size and ethnic composition. Each field center developed its recruitment procedures according to the characteristics of its community and available resources, including lists of residents, dwellings, and telephone exchanges. Supplemental resources, including lists of Medicare beneficiaries and referrals by participants, were used in the last stages of enrollment to meet goals for the elderly and minorities. Each site recruited an approximately equal number of men and women, according to prespecified age and race/ethnicity proportions. All participants gave informed consent, and the study protocol was approved by the institutional review board at each site.

Measurement of CAC
CAC was measured with electron-beam computed tomography (EBCT) at 3 field centers and multidetector row helical computed tomography (MDCT) at 3 field centers. Each participant was scanned twice consecutively, and these scans were read independently at a centralized reading center. We collimated and reconstructed at least 40 scan slices using a 3.0-mm slice thickness for EBCT scanners and 2.5-mm slice thickness for MDCT scanners. The methodology for acquisition and interpretation of the scans has been reported previously.21 The results of the 2 scans were averaged to provide a more accurate point estimate of the amount of calcium present. The amount of calcium was quantified with the Agatston scoring method.22 Calcium scores were adjusted with a standard calcium phantom that was scanned along with the participant.23 The phantom contained 4 bars of known calcium density and provided a way to calibrate the x-ray attenuation level between measurements conducted on different machines. This was important because scanners were changed between baseline and follow-up at 3 of the 6 sites. Any detectable calcium was defined as a CAC score >0; a minimum focus of calcification was based on at least 4 contiguous voxels, which resulted in identification of calcium of 1.15 mm3 for the MDCT scanners and 1.38 mm3 for the EBCT scanners.21 The nominal section thickness was 3.0 mm for EBCT scanners and 2.5 mm for MDCT scanners. Interobserver agreement and intraobserver agreement were found to be very high ({kappa}=0.93 and 0.90, respectively). Follow-up CAC measurements were performed on half the cohort (randomly selected) at a second examination (September 2002 through January 2004) and the other half of the cohort at a third examination (March 2004 through July 2005) at an average of 1.6 and 3.2 years after the participant’s first examination, respectively. As for the baseline examination, the results from 2 consecutive scans were averaged. A full characterization of the distribution of CAC at baseline by age, gender, and race has been published previously.24

Measurement of Covariates
Information on demographics, smoking, medical conditions, and family history was collected by questionnaire at the initial examination. Height and weight were also measured at the baseline examination, and blood was drawn for measurements, including lipids, inflammation, fasting glucose, fibrinogen, and creatinine. Resting blood pressure was measured 3 times in the seated position, and the average of the last 2 measurements was used in the analysis. Medication use was determined by questionnaire. Additionally, the participant was asked to bring to the clinic containers for all medications used during the 2 weeks before the visit. The interviewer then recorded the name of each medication, the prescribed dose, and frequency of administration from the containers. Further details and additional references have been published previously.20

Statistical Analysis
All participants with both a baseline and a follow-up CAC measurement were included in the analysis. The presence of CAC was defined as an Agatston score >0. Sensitivity analyses revealed that using various alternative cut points (such as 10 Agatston units) did not materially alter the risk factor relationships (data not shown). Progression of CAC was defined in 2 ways: incident CAC defined as detectable CAC at the follow-up examination (either examination 2 or 3) in a participant free of detectable CAC at examination 1 and change in CAC score in participants who had detectable CAC at examination 1. These 2 end points were modeled separately. Risk factors considered include age, gender, race/ethnicity, education, income, systolic and diastolic blood pressures, use of antihypertensive medications, diabetes status (normal, impaired fasting glucose 100 to 125 mg/dL, untreated diabetes mellitus [>125 mg/dL fasting glucose], and treated diabetes mellitus), smoking (never, former, current), pack-years of smoking, body mass index, LDL and HDL cholesterol, triglycerides, use of lipid-lowering medication, fibrinogen, creatinine, and C-reactive protein (CRP). Triglycerides and CRP were highly skewed and were log-transformed. For creatinine, there was some evidence of nonlinearity; hence, creatinine was categorized with cut points selected empirically on the basis of observed risk of incident CAC. Three groups were defined as creatinine ≤0.9, 1.0, and ≥1.1 mg/dL.

Relative risk regression25 was used to model the probability of incident-detectable CAC among those free of CAC at examination 1. That is, the probability of incident CAC was modeled as a function of covariates using a generalized linear model with log link and binomial error distribution. In cases in which the model failed to converge with the binomial error ({approx}5% of the models), we substituted gaussian error and used robust standard error estimates. We used relative risk regression rather than logistic regression because the incidence of new calcification is not rare (occurring in >16% of participants during the follow-up), so the odds ratio is an overestimate of the relative risk. Age- and gender-adjusted models for each risk factor were estimated, followed by a multivariable model constructed via a backward elimination variable selection process. The backward selection process started with all the candidate variables (regardless of statistical significance) and sequentially removed variables with P>0.05, beginning with the least significant variable. The time between scans was included as a covariate in all models, and interactions of each risk factor with gender and race/ethnicity were tested.

Among those with some detectable CAC at examination 1, we defined progression as the absolute difference between follow-up and examination 1 CAC, and this was treated as a continuous end point. The modeling strategy for this end point was analogous to the models for incident CAC; however, outliers in the calcium measurements posed an additional analytic challenge. To overcome this, we fit robust linear regression models. The robust regression algorithms in Stata26 were used, which iteratively downweight observations with large residuals.27 In these models, {approx}300 participants (10.7%) were given extremely low or zero weight in the analysis. Generally, these were participants with very fast progression relative to their predicted progression based on risk factors. Scanner changes at some of the sites also influenced progression magnitude, and a term for scanner pair was included in all the models for progression. The multivariable model was again constructed via a backward elimination variable selection process as described above for the incident CAC analysis. Our primary analysis for absolute progression did not control for baseline CAC score. In a secondary analysis, we explored which risk factor associations persisted after adjustment for baseline score.

In addition to the raw change in calcium, it was also of interest to study the relative change in calcium (relative to baseline). An initial analysis considered the percentage change in CAC; however, this end point is highly influenced by very small baseline calcium scores, for which a very small absolute increase in CAC may result in a huge percentage increase. A common approach to this problem is to exclude those with small initial calcium, say CAC <10; however, this would have resulted in exclusion of >400 selectively low-risk participants. We expressed relative change in CAC as the change from baseline to follow-up in the log of CAC plus a constant: ie, ln(CACFU+25)–ln(CACBL+25). The constant was chosen so that the resulting change was fairly symmetric and normally distributed. The addition of the constant prevents the very small baseline CAC scores from leading to unreasonably large relative increases, and use of the log-scale downweights the exceptionally large follow-up CAC scores.

The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written. All statistical analyses were performed with either Stata version 9.126 or R version 2.3.1.28


*    Results
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 *Results
down arrowDiscussion
 down arrowConclusions
 down arrowReferences
 
After exclusion of 1058 participants without a follow-up calcium score, 5756 participants (84%) remained for analysis; 48% were male, and the average age was 62 years. The sample was 40% white, 12% Chinese, 27% black, and 21% Hispanic, with almost 84% having completed high school, 37% having completed college or graduate school, and roughly 25% having a total gross family income below $25 000. The average body mass index was 28.3 kg/m2, 35% were taking antihypertensive medications, and 16% were taking lipid-lowering medication. Half the cohort had never smoked, 12% were current smokers, {approx}9% had treated diabetes mellitus, and 40% had a family history of heart attack. Participants who were excluded because of a lack of follow-up calcium were slightly older (64 versus 62 years), were more likely to have CAC at baseline (56% versus 49%), had higher systolic blood pressure (130 versus 126 mm Hg), and were more likely to be diabetic (19% versus 13%) or current smokers (16.6% versus 12.4%).

Among the 2948 participants (51.2%) without detectable CAC at baseline, 475 (16.1%) went on to develop incident CAC over an average of 2.4 years of follow-up, for an estimated incidence rate of 6.6% per year. Figure 1 illustrates the linearly increasing yearly rate of incident CAC as a function of baseline age. The incidence was <5% per year at younger ages and >12% per year at the highest ages. Age differences were similar across gender and race/ethnicity subgroups. Table 1 provides yearly rates of incident CAC by gender and race/ethnicity. In women, black participants had the highest yearly rate of incident CAC at 6.3%, whereas Chinese women had the lowest at 4.7%. White and Hispanic women were intermediate at 5.6% and 6.0%, respectively. After age adjustment, these racial/ethnic differences among women were nonsignificant. With the exception of Chinese, men had higher incidence rates than women (P<0.001). White men had the highest rate at 10.3%, followed by blacks at 8.2% (P=0.032 compared with white men after age adjustment) and Hispanics at 7.2% (age-adjusted P=0.039 compared with white men). Chinese men were again the lowest, with a rate of only 4.4% (age adjusted P<0.001 compared with white men).


Figure 1183778
View larger version (16K):
[in this window]
[in a new window]

  		Figure 1. Incidence rate of newly detectable CAC by age. Points indicate average observed percent with new CAC for each year of age, calculated by dividing the number of new cases by the total person-years of follow-up for participants of that age at baseline. One point is included for each year of age represented in MESA. The solid line is a local regression smoother through the points. This smoother weights each point equally and does not account for differential variability across the points.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Yearly Incidence Rate of CAC (Among Those Free of CAC at Baseline)

Table 2 displays the associations of traditional cardiovascular risk factors with the risk of incident detectable CAC. Each additional 10 years of age was associated with a 39% higher risk (95% CI, 28 to 52), similar to comparisons of men relative to women (43% higher rate for men adjusted for age and length of follow-up; 95% CI, 22 to 67). After adjustment for age, gender, and length of follow-up, many traditional cardiovascular risk factors were associated with an increased risk of incident CAC, including white race, higher body mass index, high blood pressure, abnormal lipids (higher LDL or triglycerides, lower HDL, or use of lipid-lowering medication), glucose disorders (impaired fasting glucose, untreated or treated diabetes mellitus), family history of heart attack, higher fibrinogen, higher CRP, and low creatinine. Education, family income, and pack-years of smoking were not related to incident CAC risk. Backward selection through all the variables resulted in a model including age, gender, race/ethnicity, body mass index, use of antihypertensive medications, LDL cholesterol, triglycerides, use of lipid-lowering medications, diabetes status, family history of heart attack, and creatinine group. Current and former smokers had higher incident CAC rates than never smokers, but this difference was not statistically significant once other risk factors were considered. Fibrinogen and CRP were no longer significantly associated with CAC incidence after adjustment for body mass index. If triglycerides were not included as a covariate, HDL cholesterol entered instead with a protective association. The risk of incident CAC did not differ by type of scanner (MDCT versus EBCT) or by study site (data not shown).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Relative Risk Regression Models for the Risk of Incident CAC (Among Those Free of CAC at Baseline)

Table 3 provides summary characteristics of annualized CAC progression. Three-hundred-fifty-eight (12.8%) participants revealed a negative calcium change on follow-up, and 691 (24.6%) had a very small amount of progression (<10 Agatston units per year). The majority of participants (1367 or 48.7%) had modest progression (between 10 and 99 Agatston units per year). Two-hundred-forty-nine (8.9%) had substantial progression (between 100 and 199 Agatston units annually) and 143 (5.1%) had extremely large progression (between 200 to >700 Agatston units per year). Median progression was 18 Agatston units overall (14 for women, 21 for men), with a mean of 46 Agatston units (36 for women, 54 for men). The distribution was heavily skewed because of the presence of several large progressors. Figure 2 is a histogram showing the distribution of annual change in CAC. Figure 3 illustrates a smoothed plot of annual change in calcium by age among the 2808 participants with detectable CAC at baseline. The increasing trend is linear across age and was similar for men and women, although men had {approx}17 to 20 Agatston units more progression per year on average than women at any given age.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Summary of Average Annual CAC Progression in Those With Detectable CAC at Baseline


Figure 2183778
View larger version (23K):
[in this window]
[in a new window]

  		Figure 2. Distribution of annual CAC progression among those with detectable coronary calcium at baseline. Histogram of the annual change in CAC for those with positive CAC at baseline. For clarity, the x axis has been truncated at –100 and 250 Agatston units. Eight participants had an annual change <–100 Agatston units and 104 participants had an annual change >250 Agatston units.


Figure 3183778
View larger version (14K):
[in this window]
[in a new window]

  		Figure 3. Average annual change in CAC by age among those with detectable coronary calcium at baseline. Points indicate the average annual change in phantom-adjusted CAC for participants of a particular age at baseline. Annual change in CAC for each participant is calculated as CAC at follow-up minus CAC at baseline divided by the time between scans in years. This is then averaged for all participants of a given age at baseline. The solid line is a local regression smoother through the points. This smoother weights each point equally and does not account for differential variability across the points. Participants with no detectable CAC at baseline have been excluded.

Table 4 displays the age-, gender-, scanner type–, and follow-up time–adjusted associations of risk factors with CAC progression, followed by the multivariable results. Each additional year of follow-up was associated with an extra 17.5 Agatston units of progression on average (95% confidence interval, 14.6 to 20.4), and each additional 10 years of baseline age was associated with 8.8 more Agatston units of progression (95% confidence interval, 6.4 to 11.2). Men had an average of 10.9 more Agatston units of progression than women (95% confidence interval, 6.3 to 15.5). Other risk factors associated with more progression included white race/ethnicity, greater body mass index, higher blood pressure (particularly systolic), higher triglycerides or use of lipid-lowering medication, pack-years of smoking, diabetes mellitus, higher CRP, and a family history of heart attack. In the multivariable models, factors associated with increased progression included time between scans, age at baseline, male gender, white race/ethnicity, higher body mass index, high blood pressure, family history of heart attack, and diabetes mellitus. The amount of progression did not differ by type of scanner (MDCT versus EBCT) or by study site (data not shown). The R2 for the multivariable model is {approx}12%; thus, much of the variation remains unexplained. There was also a significant interaction between race and diabetes mellitus (P=0.0002), indicating that the association of diabetes mellitus with progression was strongest for blacks than for the other races and weakest among Hispanics, with whites and Chinese intermediate. Treated diabetes mellitus was associated with an estimated 48 more Agatston units of progression in blacks, 28 more Agatston units in whites, 19 more Agatston units in Chinese, and only 7 more Agatston units in Hispanics.


View this table:
[in this window]
[in a new window]

  		TABLE 4. Robust Regression Models for the Change in CAC Over Time (Among Those With CAC >0 at Baseline)

Forcing in the same risk factors as the fully adjusted robust regression model (model 2 in Table 4), adjustment for baseline CAC removed the association of age, gender, race/ethnicity, smoking, and systolic blood pressure; body mass index, family history of heart attack, and diabetes mellitus remained significant. Within the subset of participants who had only a small amount of CAC (1 to 100 Agatston units) at baseline, factors that remained strong modifiable predictors of progression included blood pressure and glucose (data not shown). Factors that were associated with a higher relative increase in CAC (expressed relative to their baseline CAC) were longer time between scans and higher systolic blood pressure, diabetes mellitus, and CRP (even after adjustment for body mass index, which was not significant).

After exclusion of those taking medication for diabetes mellitus (335 who had detectable CAC at baseline and a follow-up CAC measurement), glucose level (log transformed) was significantly associated with progression. Among the 327 participants with treated diabetes mellitus, self-reported information on diabetes mellitus duration was available for 271. In this subset, diabetes duration was positively associated with calcium progression, with each additional 10 years of diabetes duration associated with 24 more Agatston units of progression (95% confidence interval, 7 to 416; P=0.005) after adjustment for length of follow-up, scanner, age, and gender. Aside from age and follow-up time, none of the other risk factors were significant in this subset.


*    Discussion
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 *Discussion
down arrowConclusions
 down arrowReferences
 
In the MESA cohort, many traditional cardiovascular risk factors were associated with both the risk of developing incident coronary calcium and increases in existing calcification. These included older age, male gender, white race/ethnicity, hypertension, higher body mass index, diabetes mellitus, and family history of heart attack. Certain factors also appeared to be related only to the risk of incident CAC but not to the progression of existing calcification. For instance, LDL cholesterol and HDL cholesterol were not associated with the progression of existing CAC, even in unadjusted models, but both were associated (HDL negatively) with the risk of incident CAC. Triglycerides were associated with both end points in models adjusted for age, gender, and follow-up time, but in fully adjusted models, this association became nonsignificant for CAC progression while remaining significant for incident CAC. Surprisingly, low creatinine was associated with higher risk of incident CAC, although creatinine was not associated with progression of existing CAC.

Diabetes mellitus was the strongest risk factor for CAC progression. In fact, glucose was associated with higher risk of incident CAC even at levels well below the standard thresholds for diabetes mellitus or impaired fasting glucose and among nondiabetic participants. Among treated diabetics, duration of diabetes was the only significant risk factor for CAC progression after adjustment for age, gender, scanner, and duration of follow-up. However, because of the small number of such patients (n=281 with both CAC change and diabetes duration), our power for detecting other associations was limited.

The associations that persisted after adjustment for baseline CAC were body mass index, family history of heart attack, diabetes mellitus, and glucose. This indicates that these factors have an association with CAC that is not yet completely reflected by their lifetime CAC accumulation. This is in contrast to the more chronic associations of age, gender, race/ethnicity, hypertension, and smoking, the influence of which is already manifest in the baseline CAC level. Although both fibrinogen and CRP were related to an increased risk of incident CAC, this became nonsignificant after adjustment for body mass index.

In the present study, we observed a modest U-shaped relationship of creatinine with the risk of incident CAC. Specifically, those with creatinine levels of 1 mg/dL had the lowest risk, those with creatinine <1 mg/dL had significantly higher risk, and those with creatinine >1 mg/dL had higher risk, but not significantly so. Renal dysfunction has been associated with a greater amount and progression of coronary calcium.15,29,30 Very few participants in MESA have renal dysfunction, which may explain why the association of higher creatinine levels with risk was nonsignificant. Low levels of creatinine can indicate the presence of severe liver disease or decreased muscle mass, but it seems unlikely that this explains the increased risk of incident CAC in this relatively healthy cohort. This finding was data driven and may simply be a false-positive. Further study is needed to confirm or refute this association.

Comparison of our results with previous studies on progression is challenging because of differing analytic strategies. Almost all previous studies have controlled for baseline calcium in the models, which may obscure important associations. These studies tend to conclude that the strongest (or even the only) predictor of calcium change is baseline calcium level and that factors such as age and gender are not associated.10,31 The rate of progression before baseline will be a strong determinant of the baseline CAC score. Furthermore, it is reasonable to assume that the rate of progression before baseline is predictive of future progression rates. In other words, baseline CAC is not a confounder but part of the causal pathway for the relationship between long-term risk factors and CAC progression. For example, despite the fact that men clearly have more progression than women, once the baseline CAC is controlled for, no gender difference exists. The reason is that they already have been progressing more and already have more CAC. They also will continue to progress more rapidly. It is difficult to interpret the results from such a model because the effects that the variables have had up until baseline are conditioned out. If baseline CAC is included in the model, only associations that are acute (eg, a new medication, recent infection, newly diagnosed condition) or possibly those that have a strong association that changes over time will be easily detectable.

The choice of scale for the analysis of continuous change in CAC score is a problematic analytic issue, one that makes comparison across various studies in the literature difficult. Different choices (eg, raw change versus percentage change versus change in log calcium plus a constant) yield different conclusions. We found certain standard analyses, eg, simply modeling raw change or percent change, to be heavily influenced by outliers. To overcome this issue, we used robust regression techniques when modeling absolute change and a logarithmic transformation of calcium plus a constant when modeling relative change. The addition of the constant reduces the impact of small baseline calcium, resulting in very large relative changes, and the log scale decreases the association of the extremely large follow-up calcium measurements.

Both antihypertensive medication and lipid-lowering medication use were associated with a significantly increased risk of incident CAC and significantly more CAC progression. This does not imply that these medications have a negative impact; more likely, their use indicates more severe underlying hypertension or dyslipidemia, which may have an impact beyond that of the concurrently measured blood pressure or cholesterol values. These may be normal as a result of treatment despite underlying disease. Several previous studies have found evidence that use of statins slows, or in some cases even reverses, the progression of calcium11,12,32 among those with high cholesterol. Those studies were based on relatively small sample sizes, and although the participants were free of clinical coronary artery disease, they were referred for assessment of their CAC. These participants tended to have more risk factors and more calcium at baseline than the participants in MESA. Additionally, the findings of statin benefit refer to those with hypercholesterolemia, and the untreated participants in MESA do not necessarily have high cholesterol.

A unique feature of our data is the ability to study racial/ethnic differences. Whites had far more incident CAC and CAC progression than the other 3 racial/ethnic groups. Risk factor relationships, however, were largely the same across racial/ethnic groups. For incident CAC, there was no evidence that the effect of risk factors differed by race. For CAC progression, although most risk factor relationships were similar across races, the effect of diabetes mellitus was strongest in blacks and weakest in Hispanics. This is consistent with at least 1 previous study33 reporting decreased congestive heart disease and cardiovascular disease mortality in Hispanic diabetics compared with non-Hispanic white diabetics.

Strengths of the MESA study include the large sample size, inclusion of 4 racial/ethnic groups, and the community-based (as opposed to referral-based) nature of the sample. Additionally, the prospective nature of the study allowed calcium measurement and risk factor assessment to proceed in a standardized manner. The sample did not include participants with clinical cardiovascular disease; hence, results on progression of calcium cannot be generalized to that population. Participants without a follow-up CAC tended to have a somewhat worse risk factor profile on average than included participants. This may have weakened certain risk factor associations. Some differences in scanning methodology exist between our study and previously published studies. The field of view for the MESA scans was 350 mm, which is larger than previous studies, resulting in larger pixels. We also required calcium to be present in 4 adjacent pixels to register as a calcified area, whereas other studies required between 2 and 4 pixels. We used backward selection to construct our multivariable models, and a risk exists that ≥1 confounding variables (ie, variables that might influence the association of other variables despite not being statistically significant themselves) were removed. Many statistical comparisons were performed in this study, and the some of the associations may be false-positives. A considerable amount of nonbiological variability exists in the measurement of CAC because of computed tomography technical factors. As a result, we may have been unable to detect certain more modest associations as a result of the presence of noise. In part, because of this variability, the R2 for our multivariable model of CAC change was relatively low at 12%. However, we were able to demonstrate several factors that are strongly associated with CAC progression.


*    Conclusions
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 up arrowDiscussion
 *Conclusions
down arrowReferences
 
This is the first large multiethnic study to report on the incidence and progression of CAC. We found that the incidence of CAC increased steadily across age, regardless of gender and race. Annual amount of CAC progression also increased linearly across age, with a similar trend for men and women. Additionally, we found that age, male gender, white race, and standard coronary risk factors generally were related to both the risk of incident coronary calcification and the progression of calcification. Whites had more incident CAC and CAC progression than the other 3 racial/ethnic groups. With the exception of diabetes mellitus, risk factor relationships were largely the same across racial/ethnic groups.


*    Acknowledgments
 
We 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.

Sources of Funding

This research was supported by contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169 from the National Heart, Lung, and Blood Institute.

Disclosures

None.


*    References
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMethods
 up arrowResults
 up arrowDiscussion
 up arrowConclusions
 *References
 

  1. Greenland P, LaBree L, Azen SP, Doherty TM, Detrano RC. Coronary artery calcium score combined with Framingham score for risk prediction in asymptomatic individuals. JAMA. 2004; 291: 210–215.[Abstract/Free Full Text]
  2. Arad Y, Goodman K, Roth M, Newstein D, Guerci AD. Coronary calcification, coronary disease risk factors, C-reactive protein, and atherosclerotic cardiovascular disease events: the St Francis Heart Study. J Am Coll Cardiol. 2005; 46: 158–165.[Abstract/Free Full Text]
  3. O’Malley PG, Taylor AJ, Jackson JL, Doherty TM, Detrano RC. Prognostic value of coronary electron-beam computed tomography for coronary heart disease events in asymptomatic populations. Am J Cardiol. 2000; 85: 945–948.[CrossRef][Medline] [Order article via Infotrieve]
  4. Kondos GT, Hoff JA, Sevrukov A, Daviglus ML, Garside DB, Devries SS, Chomka EV, Liu K. Electron-beam tomography coronary artery calcium and cardiac events: a 37-month follow-up of 5635 initially asymptomatic low- to intermediate-risk adults. Circulation. 2003; 107: 2571–2576.[Abstract/Free Full Text]
  5. Raggi P, Cooil B, Callister TQ. Use of electron beam tomography data to develop models for prediction of hard coronary events. Am Heart J. 2001; 141: 375–382.[CrossRef][Medline] [Order article via Infotrieve]
  6. Wong ND, Hsu JC, Detrano RC, Diamond G, Eisenberg H, Gardin JM. Coronary artery calcium evaluation by electron beam computed tomography and its relation to new cardiovascular events. Am J Cardiol. 2000; 86: 495–498.[CrossRef][Medline] [Order article via Infotrieve]
  7. Lamonte MJ, Fitzgerland SJ, Church TS, Barlow CE, Radford NB, Levine BD, Pippin JJ, Gibbons LW, Blair SN, Nichaman MZ. Coronary artery calcium score and coronary heart disease events in a large cohort of asymptomatic men and women. Am J Epidemiol. 2005; 162: 421–429.[Abstract/Free Full Text]
  8. Taylor AJ, Bindeman J, Fuerstein I, Cao F, Brazaitis M, O’Malley PG. Coronary calcium independently predicts incident premature coronary heart disease over measured cardiovascular risk factors: mean three-year outcomes in the Prospective Army Coronary Calcium (PACC) project. J Am Coll Cardiol. 2005; 46: 807–814.[Abstract/Free Full Text]
  9. Vliegenthart R, Oudkerk M, Hofman A, Oei HH, van Dijck W, van Rooij FJ, Witteman JC. Coronary calcification improves cardiovascular risk prediction in the elderly. Circulation. 2005; 112: 572–577.[Abstract/Free Full Text]
  10. Yoon HC, Emerick AM, Hill JA, Gjertson DW, Goldin JG. Calcium begets calcium: progression of coronary artery calcification in asymptomatic subjects. Radiology. 2002; 224: 236–241.[Abstract/Free Full Text]
  11. Budoff MJ, Lane KL, Bakhsheshi H, Mao S, Grassmann BO, Friedman BC, Brundage BH. Rates of progression of coronary calcium by electron beam tomography. Am J Cardiol. 2000; 86: 8–11.[Medline] [Order article via Infotrieve]
  12. Callister TQ, Raggi P, Cooil B, Lippolis NJ, Russo DJ. Effect of HMG-CoA reductase inhibitors on coronary artery disease as assessed by electron-beam computed tomography. N Engl J Med. 1998; 339: 1972–1978.[Abstract/Free Full Text]
  13. Wong ND, Kawakubo M, LaBree L, Azen SP, Xiang M, Detrano R. Relation of coronary calcium progression and control of lipids according to National Cholesterol Education Program guidelines. Am J Cardiol. 2004; 94: 431–436.[CrossRef][Medline] [Order article via Infotrieve]
  14. Cassidy AE, Bielak LF, Zhou Y, Sheedy PF, Turner ST, Breen JF, Araoz PA, Kullo IJ, Lin X, Peyser PA. Progression of subclinical coronary atherosclerosis: does obesity make a difference? Circulation. 2005; 111: 1877–1882.[Abstract/Free Full Text]
  15. Bursztyn M, Motro M, Grossman E, Shemesh J. Accelerated coronary artery calcification in mildly reduced renal function of high-risk hypertensives: a 3-year prospective observation. J Hypertens. 2003; 21: 1953–1959.[CrossRef][Medline] [Order article via Infotrieve]
  16. Shemesh J, Apter S, Stolero D, Itzchak Y, Motro M. Annual progression of coronary artery calcium by spiral computed tomography in hypertensive patients without myocardial ischemia but with prominent atherosclerotic risk factors, in patients with previous angina pectoris or healed acute myocardial infarction, and in patients with coronary events during follow-up. Am J Cardiol. 2001; 87: 1395–1397.[CrossRef][Medline] [Order article via Infotrieve]
  17. Chironi G, Simon A, Denarie N, Vedie B, Sene V, Megnien J, Levenson J. Determinants of progression of coronary artery calcifications in asymptomatic men at high cardiovascular risk. Angiology. 2002; 53: 677–683.[Medline] [Order article via Infotrieve]
  18. Hoffman U, Derfler K, Haas M, Stadler A, Brady TJ, Kostner K. Effects of combined low-density lipoprotein apheresis and aggressive statin therapy on coronary calcified plaque as measured by computed tomography. Am J Cardiol. 2003; 91: 461–463.[CrossRef][Medline] [Order article via Infotrieve]
  19. Bielak LF, Sheedy PF, Peyser PA. Coronary artery calcification measured at electron-beam CT: agreement in dual scan runs and change over time. Radiology. 2001; 218: 224–229.[Abstract/Free Full Text]
  20. Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR, Greenland P, Jacobs DR, Kronmal R, Liu K, Clark Nelson J, O’Leary D, Saad MF, Shea S, Szklo M, Tracy RP. Multi-Ethnic Study of Atherosclerosis: objectives and design. Am J Epidemiol. 2002; 156: 871–881.[Abstract/Free Full Text]
  21. Carr JJ, Nelson JC, Wong ND, McNitt-Gray M, Arad Y, Jacobs DR Jr, Sidney S, Bild DE, Williams OD, Detrano R. Calcified coronary artery plaque measurement with cardiac CT in population-based studies: standardized protocol of Multi-Ethnic Study of Atherosclerosis (MESA) and Coronary Artery Risk Development in Young Adults (CARDIA) Study. Radiology. 2005; 234: 35–43.[Abstract/Free Full Text]
  22. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990; 15: 827–832.[Abstract]
  23. Nelson JC, Detrano R, Kronmal RA, Carr JJ, McNitt-Gray MF, Wong N, Loria C, Goldin JG, Williams DO. Measuring coronary calcium on CT images adjusted for attenuation differences. Radiology. 2005; 235: 403–414.[Abstract/Free Full Text]
  24. McClelland RL, Chung H, Detrano R, Post W, Kronmal RA. The distribution of coronary artery calcium by race, gender and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2006; 113: 30–37.[Abstract/Free Full Text]
  25. Lumley T, Kronmal RA, Ma S. Relative risk regression in medical research: models, contrasts, estimators and algorithms. Seattle, Wash: University of Washington; 2006. UW Biostatistics Working Paper Series, paper 293. Available at: http://www.bepress.com/uwbiostat/paper293. Accessed July 19, 2006.
  26. Stata Corp. Stata Statistical Software: Release 9. College Station, Tex: Stata Corp LP; 2005.
  27. Hamilton LC. How robust is robust regression? Stata Tech Bull. 1991; 2: 21–26.
  28. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2004.
  29. Goodman WG, Goldin J, Kuizon BD, Yoon C, Gales B, Sider D, Wang Y, Chung J, Emerick A, Greaser L, Elashoff RM, Salusky IB. Coronary-artery calcification in young adults with end-stage renal disease who are undergoing dialysis. N Engl J Med. 2000; 342: 1478–1483.[Abstract/Free Full Text]
  30. Braun J, Oldendorf M, Moshage W, Heidler R, Zeitler E, Luft FC. Electron beam computed tomography in the evaluation of cardiac calcification in chronic dialysis patients. Am J Kidney Dis. 1996; 27: 394–401.[Medline] [Order article via Infotrieve]
  31. Hsia J, Klouj A, Prasad A, Burt J, Adams-Campbell LL, Howard BV. Progression of coronary calcification in healthy postmenopausal women. BMC Cardiovasc Disord. 2004; 4: 21.[CrossRef][Medline] [Order article via Infotrieve]
  32. Achenbach S, Ropers D, Pohle K, Leber A, Thilo C, Knez A, Menendez T, Maeffert R, Kusus M, Regenfus M, Bickel A, Haberl R, Steinbeck G, Moshage W, Daniel WG. Influence of lipid-lowering therapy on the progression of coronary artery calcification: a prospective evaluation. Circulation. 2002; 106: 1077–1082.[Abstract/Free Full Text]
  33. Swenson CJ, Trepka MJ, Rewers MJ, Scarbro S, Hiatt WR, Hamman RF. Cardiovascular disease mortality in Hispanics and non-Hispanic whites. Am J Epidemiol. 2002; 156: 919–928.[Abstract/Free Full Text]
 

CLINICAL PERSPECTIVE

Coronary artery calcium (CAC) has been shown to be associated with the risk of clinical cardiovascular events. In this prospective multiethnic cohort study we found that many of the traditional risk factors for clinical cardiovascular disease are also associated with the incidence and progression of subclinical cardiovascular disease as measured by CAC. In serial scanning, one can expect to see newly detectable CAC at an average rate of {approx}6.6% per year, ranging from <5% per year below age 50 to >12% per year at ages 80 and over. Median annual change in CAC for those with existing calcification at baseline was 14 Agatston units for women and 21 for men. In addition, although whites have more progression of CAC than do blacks, Hispanics, or Chinese, risk factor associations with progression are largely similar across different races/ethnicities.


Related Article:

Issue Highlights
Circulation 2007 115: 2683. [Full Text]



This article has been cited by other articles:


Home page
 Anesth. Analg.Home page
M. M. Vigoda, L. I. Rodriguez, E. Wu, K. Perry, R. Duncan, D. J. Birnbach, and D. A. Lubarsky
The Use of an Anesthesia Information System to Identify and Trend Gender Disparities in Outpatient Medical Management of Patients with Coronary Artery Disease
Anesth. Analg., July 1, 2008; 107(1): 185 - 192.
[Abstract] [Full Text] [PDF]

Home page
 CirculationHome page
I. Gottlieb and J. A.C. Lima
Screening High-Risk Patients With Computed Tomography Angiography
Circulation, March 11, 2008; 117(10): 1318 - 1332.
[Full Text] [PDF]

Home page
 Eur Heart JHome page
S. Schroeder, S. Achenbach, F. Bengel, C. Burgstahler, F. Cademartiri, P. de Feyter, R. George, P. Kaufmann, A. F. Kopp, J. Knuuti, et al.
Cardiac computed tomography: indications, applications, limitations, and training requirements: Report of a Writing Group deployed by the Working Group Nuclear Cardiology and Cardiac CT of the European Society of Cardiology and the European Council of Nuclear Cardiology
Eur. Heart J., February 2, 2008; 29(4): 531 - 556.
[Abstract] [Full Text] [PDF]

Home page
 J Am Coll CardiolHome page
D. V. Anand, E. Lim, D. Darko, P. Bassett, D. Hopkins, D. Lipkin, R. Corder, and A. Lahiri
Determinants of Progression of Coronary Artery Calcification in Type 2 Diabetes: Role of Glycemic Control and Inflammatory/Vascular Calcification Markers
J. Am. Coll. Cardiol., December 4, 2007; 50(23): 2218 - 2225.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
115/21/2722    most recent
CIRCULATIONAHA.106.674143v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kronmal, R. A.
Right arrow Articles by Burke, G. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kronmal, R. A.
Right arrow Articles by Burke, G. L.
Right arrowPubmed/NCBI databases
Medline Plus Health Information
*Coronary Artery Disease
Related Collections
Right arrow Epidemiology
Right arrow Risk Factors
Right arrow Imaging