Metabolic syndrome (MetS) and diabetes (DM) predict coronary heart disease (CHD) events and mortality (1- 3). Subclinical atherosclerosis as evidenced by coronary artery calcium (CAC) (4- 9) and carotid intima-media thickness (10- 11) is increased in MetS and DM, but no study has compared the incidence and progression of CAC across these conditions. Progression of CAC may be clinically important because persons experiencing CHD events have greater progression of CAC (12) and, recently, progression of CAC has been shown to predict all-cause mortality (13). The population-based MESA (Multiethnic Study of Atherosclerosis) has demonstrated that most standard CHD risk factors are associated with the incidence and progression of CAC (14).

In this report, we compared, in MESA, the incidence and progression of CAC among persons with MetS (but no DM) and DM (with and without MetS) relative to those with neither condition. Our hypothesis was that MetS subjects would be associated with future development and progression of CAC greater than those subjects without MetS would, but less than those with DM would.

The design of MESA, a prospective epidemiologic study of the prevalence, risk factors, and progression of subclinical cardiovascular disease has been previously published (15). Briefly, 6,814 participants 45 to 84 years of age, who were free of clinical cardiovascular disease and identified as White, African American, Hispanic, or Chinese were recruited from 6 U.S. communities (Forsyth County, North Carolina; Northern Manhattan and the Bronx, New York; Baltimore City and Baltimore County, Maryland; St. Paul, Minnesota; Chicago, Illinois; and Los Angeles County, California) from 2000 to 2002. Recruitment included lists of residents, dwellings, telephone exchanges, lists of Medicare beneficiaries, and referrals by participants. Similar numbers of men and women were recruited according to pre-specified age and race/ethnicity quotas. All participants gave informed consent, and the study protocol was approved by the Institutional Review Board at each site.

This report includes 5,662 subjects with both baseline (Exam 1) and follow-up (at Exams 2 or 3) computed tomography (CT) scans, available data to define DM or MetS, and with no incident CHD event occurring between baseline and follow-up CT. This resulted in excluding 1,056 subjects who did not have follow-up scans (or were out of protocol), 26 with incomplete data to define DM or MetS, and 70 who had an intervening CHD event. Diabetes was defined as having a fasting glucose ≥7.0 mmol/l (126 mg/dl) or being on insulin or oral hypoglycemic medications. Among nondiabetics, MetS was defined to be present if ≥3 of the following were present: 1) abdominal obesity based on waist circumference >88 cm (35 inches) for women and >102 cm (40 inches) for men; 2) high-density lipoprotein cholesterol <1.0 mmol/l (40 mg/dl) for men or <1.3 mmol/l (50 mg/dl) for women; 3) fasting triglycerides ≥1.7 mmol/l (150 mg/dl); 4) blood pressure of ≥130 mm Hg systolic or ≥85 mm Hg diastolic, or on treatment; or 5) impaired fasting glucose defined as a fasting glucose of 5.55 to 6.99 mmol/l (100 to 125 mg/dl), based on the American Heart Association/National Heart, Lung, and Blood Institute definition (16).

CAC was measured by electron-beam (3 sites) or multidetector (3 sites) CT. Participants were scanned twice consecutively and scans were read by a trained physician-reader at a centralized reading center (Los Angeles Biomedical Research Institute, Torrance, California). The methodology for acquisition and interpretation of the scans has been published (17). Calcium volume scores (17) and Agatston scores (18) were based on averaging results from each scan and adjusted using a standard calcium phantom (scanned with the participant) to calibrate X-ray attenuation between measurements conducted on different machines (19). Detectable calcium was defined as a CAC score >0. A second scan was performed on one-half of the cohort (randomly selected) at a second exam (September 2002 to January 2004) and on the other one-half at a third exam (March 2004 to July 2005), averaging 1.6 and 3.2 years after the first scan, respectively (average 2.4 years between). The distribution of CAC in MESA at baseline by age, sex, and race has been published previously (20).

Information on demographics, smoking, medical conditions, and family history was obtained by questionnaire. Height, weight, total and high-density lipoprotein cholesterol, triglycerides, and fasting glucose levels were determined. Resting blood pressure was measured 3 times, with the average of the last 2 measurements used in analysis. Use of cholesterol, blood pressure, and diabetes medications was determined by questionnaire and from medication containers (15).

The cohort was followed for incident CHD events for a mean of 4.9 ± 1.3 years following the second scan. At intervals of 9 to 12 months, a telephone interviewer inquired about interim hospital admissions, cardiovascular diagnoses, and deaths. An adjudication committee received copies of all death certificates and medical records for hospitalizations and outpatient cardiovascular diagnoses and conducted next-of-kin interviews. Two physicians independently classified and assigned incidence dates. For disagreements, a full mortality and morbidity review committee made the final classification. We followed participants for occurrence of all CHD endpoints, which included myocardial infarction, angina, resuscitated cardiac arrest, or CHD death. CHD death was based on review of hospital records and interviews with families. The reviewers were blinded to CT scan and cardiac magnetic resonance results and used pre-specified criteria.

Subjects were classified as having: 1) neither MetS nor DM; 2) MetS without DM; 3) DM without MetS; and 4) DM with MetS. The first 2 groups were also classified by number of MetS risk factors (0, 1, 2, 3, and 4 to 5). To assess bivariate associations between these groups, risk factors, and CAC score/volume measures, the chi-square test (for categorical covariates) or an F test from analysis of variance (for continuous covariates) was used. Incident CAC was defined among those without baseline CAC (n = 2,927) as those who developed detectable CAC at the follow-up scan. Absolute progression of CAC was defined among those with CAC at baseline (n = 2,735) as the difference between the CAC volume score on follow-up (CAC_FU) and that at baseline (CAC_BL) (14). We used relative risk (RR) regression (21) to obtain asymptotically unbiased estimates of the RR of incident CAC among those free of CAC at baseline. This involved modeling the probability of incident CAC score as an exponential function of risk factors (including each MetS/DM classification relative to the reference group) and performing nonlinear least-squares estimation. To account for misspecification of the variance, we computed model-robust (Huber-White) standard errors. To estimate the absolute progression of CAC among those with detectable CAC at baseline, we used robust linear regression, down-weighting the influence of participants with very large progression to increase model robustness. We also present our findings as relative progression, defined as the median annualized percentage of change in CAC from baseline to follow-up scan. Our analyses also adjusted for time between scans, age, sex, ethnicity, baseline total cholesterol, lipid-lowering medication use, smoking status, and family history of myocardial infarction. Adjusting for the time between scans in a progression model of the absolute change in CAC implicitly standardizes CAC change with respect to time and is equivalent to directly modeling the annualized absolute CAC change. Absolute progression analyses (but not relative progression) were additionally adjusted for the scanner pair used at baseline and follow-up to account for scanner changes over time. As a sensitivity analysis, we also performed progression analyses additionally adjusted for baseline calcium volume score. We also investigated the independent contribution of each MetS component to predicting the incidence or progression of CAC, and we evaluated incidence and progression models that included each of the 5 separate MetS components in the model together. To examine if the composite of MetS/DM still predicted incidence and progression of CAC after accounting for the individual MetS components, we added this in a final model. Finally, Cox proportional hazards regression was used to examine the relation of progression of CAC to the incidence of total CHD events within each disease group separately. All statistical analyses were performed using SAS (version 9.1, SAS Institute, Cary, North Carolina) (22).

Overall, 5,662 subjects were included (51% women, mean age 61.0 ± 10.3 years): 3,528 (62.3%) had neither MetS nor DM; 1,426 (25.2%) had MetS without DM; 198 (3.5%) had DM without MetS; and 510 (9.0%) had both MetS and DM. Subjects excluded (n = 70) because of intervening CHD events between baseline and follow-up scans were more likely to have both MetS and DM (23%) and less likely to have neither condition (40%); also, 31% had MetS (without DM) and 6% had DM without MetS. Scanners used for the initial and follow-up scan were electron-beam CT (n = 2,852, 50.4%), multidetector (n = 2,630, 46.4%), and the remainder were electron-beam multidetector (n = 180, 3.2%). (Table 1) shows the distribution of demographic and clinical risk factors and calcium scores by the 4-category MetS/DM classification. Systolic and diastolic blood pressure, triglycerides, and waist circumference were highest and high-density lipoprotein cholesterol was lowest in those with MetS, and fasting glucose levels were highest in those with DM (all p < 0.01). The prevalence of CAC ranged from 44% to 62% by MetS/DM classification (p < 0.01). Among those with CAC, baseline volume score was highest in those with DM.

The unadjusted incidence of CAC increased progressively according to MetS/DM status (Figure 1). Among men with DM and women with DM plus MetS, incidence of CAC was highest. Incidence was unexpectedly low, however, in women with DM but without MetS (although uncertain because of a low sample size).

(Table 2) shows the adjusted RRs for incident CAC for the 4- and 7-category MetS/DM classifications. Compared with those with neither MetS nor DM, those with MetS (without DM) and those with DM (regardless of the presence of MetS) had a greater incidence of CAC. In gender-stratified analyses, results were generally similar to the overall group, except for those with DM (without MetS) where findings were not significant for women. In analyses stratified by ethnicity (results not shown), the RR for incident CAC was significantly greater for those with both DM and MetS among Chinese (RR = 3.7, p < 0.05), Hispanics (RR = 2.2, p < 0.01), and African Americans (RR = 1.8, p < 0.05) and was also greater for those with MetS without DM in all ethnic groups (RR = 1.7 to 2.1, p < 0.05 to p < 0.01). Only in African Americans was incident CAC significantly greater (RR = 2.1, p < 0.05) for those with DM (without MetS) versus those without MetS or DM. When examining CAC incidence by number of MetS risk factors, compared with those with no MetS risk factors, those with as few as 2 MetS risk factors had an increased incidence of CAC (RR = 1.5, p < 0.05), with increases to RRs of 2.0 or greater in those with 3 or 4 to 5 MetS risk factors, DM without MetS, or both DM and MetS (p < 0.01 overall), with stronger associations seen in women.

Among those with baseline CAC, progression of CAC (median annualized percent change in CAC) also increased directly according to MetS/DM status (Figure 2). Robust linear regression (Table 3) showed those with both MetS and DM as well as those with DM but no MetS to have the greatest progression of CAC, and those with MetS but not DM to have an intermediate level of progression. Women with DM and no MetS and men with both DM and MetS had the greatest degree of progression of CAC. Among each individual ethnic group, progression of CAC was greatest in those with both MetS and DM (mean adjusted volume score differences of 15.3 to 27.1, p < 0.01, compared with those with neither MetS nor DM), being highest in Caucasians (23.4) and African Americans (27.1). Of other MetS/DM groups, only DM without MetS in African Americans (31.7, p < 0.01) and Chinese (17.5, p < 0.05) and MetS without DM in Caucasians (11.0, p < 0.01) had progression significantly greater than those with neither MetS nor DM did. In analyses by number of MetS risk factors (Table 4), compared with those with 0 MetS risk factors, progression was greater only for those with ≥3 MetS risk factors, DM, or both DM and MetS. The greatest increases were seen for those with DM with MetS. Although in sex-stratified analyses, this was the case for men, but not for women where DM without MetS had the greatest progression. Results presented according to Agatston score showed similar findings with the expected greater magnitude of differences due to absolute Agatston scores being higher than volume scores.

When additionally adjusting for baseline volume score, our findings regarding progression of CAC were not substantially affected and remained statistically significant. Mean differences (compared with those subjects with neither MetS nor DM) in volume score change were 6.2 (95% confidence interval [CI]: 3.0 to 9.4) for those with MetS without DM, 13.4 (95% CI: 6.5 to 20.4) for those with DM without MetS, and 14.6 (95% CI: 10.1 to 19.1) for those with DM and MetS (all p < 0.01).

We also evaluated the relation of individual MetS components to the incidence and progression of CAC. With all components in the model simultaneously, only increased waist circumference (adjusted RR = 1.4, p < 0.001) and increased glucose (adjusted RR = 1.25, p < 0.01) predicted CAC incidence. Progression of CAC was driven most strongly by elevated glucose (volume score change of 10.7, p < 0.001) and blood pressure (volume score change of 5.8, p < 0.05).

When examining the relation of CAC progression to total CHD events in each disease group, total CHD events per 1,000 person years increased progressively according to extent of change in volume score in those with neither MetS nor DM, those with MetS and no DM, and those with both MetS and DM (Figure 3). Corresponding hazard ratios, adjusted for age, sex, ethnicity, and risk factors, compared with those subjects with no or negative change, were increased in those subjects in the second and third tertiles of positive CAC change: 4.5 (95% CI: 2.2 to 9.4), p < 0.01 and 7.6 (95% CI: 3.7 to 15.5), p < 0.01, respectively in those subjects with neither MetS nor DM, 2.3 (95% CI: 1.0 to 4.9), p < 0.05 and 4.1 (95% CI: 2.0 to 8.5), p < 0.01, respectively, in those subjects with MetS and no DM and 4.0 (95% CI: 1.1 to 14.9), p < 0.05 and 4.9 (95% CI: 1.3 to 18.4), p < 0.05, respectively, in those subjects with both MetS and DM. After additional adjustment for baseline CAC, these estimates were similar: 4.5 (95% CI: 2.2 to 9.4), p < 0.01 and 7.0 (3.4 to 14.7), p < 0.01, respectively, in those subjects with neither MetS nor DM, 2.3 (95% CI: 1.1 to 5.0), p < 0.05 and 3.5 (95% CI: 1.6 to 7.3), p < 0.01, respectively, in those subjects with MetS and no DM and 3.9 (95% CI: 1.0 to 14.8), p < 0.05 and 4.0 (95% CI: 0.95 to 16.0), p = NS, respectively, in those with both MetS and DM.

In the MESA, persons with MetS or DM have a greater incidence and progression of CAC than those subjects without MetS, and those with MetS (without DM) have an intermediate incidence and progression. Also, insulin resistance (23) and DM (24- 25) have been shown in smaller or selected cohorts to relate to progression of CAC, and in those with DM, a glycated hemoglobin ≥7% predicted progression of CAC (26). Progression of CAC has also been shown to predict total mortality over baseline risk factors and CAC (13). In our study, we also found increased progression of CAC in persons with MetS and DM to predict future CHD events.

The baseline calcium score, a strong predictor of CAC progression, is important to understanding the relationship of MetS and DM to progression of CAC (23,27). Because MetS is associated with an intermediate level and DM the highest level of CAC, we might expect a similar pattern for progression. Whereas baseline CAC could be considered a confounder, it can also be considered part of the causal pathway between risk factors such as MetS and DM and the progression of CAC. Such persons likely had more rapid progression of CAC to begin with and continued to show greater future progression; hence, including baseline CAC in the model could condition out the effects that variables of interest (in this case, MetS and DM) may have up to baseline (14). In our study, however, secondary analyses additionally adjusted for baseline calcium volume showed only a slight attenuation of our findings, which remained largely significant.

In addition, the choice of the scale for continuous progression (e.g., absolute vs. relative change) and the failure to account for outlying progressors may markedly influence the results obtained (14). Similar to Kronmal et al. (14), we use absolute progression as our primary outcome and account for outlying progressors using robust regression modeling techniques, which limits the influence of outlying observations (e.g., fast progressors).

Strengths of MESA include its large sample size, ethnic diversity, and community-based recruitment. The prospective design allows for assessment of baseline factors including MetS and DM in relation to development and progression of CAC, as well as the evaluation of progression of CAC in relation to subsequent CHD events. In addition, MESA had standardized protocols for scanning and interpretation of scans. Importantly, our estimate of progression of CAC was based only on the change between 2 scans done an average of about 2 years apart, from which progression was then annualized. However, this assumes a linear relation of progression with time, which may or may not be the case, hence results could be different had more measures been available and/or if the time between scans was greater. Also, exclusion of a small number of individuals (n = 70) who had intervening CHD events may have influenced the results. Such persons were more likely to have DM and MetS, progression of CAC, as well as CHD events, so our findings relating progression to events could have been underestimated (thus conservative). However, our intention was to look at the natural progression of CAC not interrupted by CHD events, thus we kept our group homogenous by excluding such individuals. As there is controversy of which definition may be most appropriate, or even whether MetS should be considered a syndrome (28- 30), our findings may have differed had other definitions for MetS been used.

Persons with both MetS and DM have the greatest incidence and degree of progression of CAC. Those with MetS without DM have an incidence and degree of progression of CAC intermediate between those with DM and without these conditions. Moreover, in those with MetS or DM, progression predicts future CHD event risk.

The authors thank the other investigators, staff, and 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. The authors also thank Ms. Yanting Luo for her assistance with data analysis and manuscript preparation.