The total CAC score (Agatston score) using noncontrast computed tomography is a recognized estimation of atherosclerosis in asymptomatic adults with at least moderate risk of cardiovascular disease [
The Multiethnic Study of Atherosclerosis (MESA) is a prospective observational evaluation of 6,814 men and women, aged 45 to 84 years, from four racial/ethnic groups (White, Asian, Hispanic, and Black) in the United States. At the time of enrollment, participants had no known cardiovascular disease.
Participants were enrolled between July 2000 and September 2002 at field centers located in Forsyth County, North Carolina; St. Paul, Minnesota; Chicago; New York City; Baltimore, Maryland; and Los Angeles. Institutional review boards approved the study protocol at each study center. Further details regarding the MESA study have been detailed elsewhere [
Study participants provided information about cardiovascular risk factors. A central laboratory (University of Vermont, Burlington, Vermont) measured levels of total and high-density lipoprotein cholesterol, triglycerides, plasma glucose, and high-sensitivity C-reactive protein (CRP) after a 12-hour fast. The coronary calcium scan was done twice in each participant to increase accuracy. For each scan, the participant was asked to remain still and momentarily hold his/her breath twice, each time for 20 to 30 seconds, in order to get good quality pictures. Participants’ CAC scores were reported as average CAC Agatston scores. Vessel-specific CAC scores were calculated in 6,540 MESA participants (96%). Of those, 6,479 MESA participants with coronary calcium CT scans at baseline and vessel-specific CAC distribution (left main, left anterior descending, left circumflex, and right) were screened. Included in the final study cohort were 4,917 participants, after exclusion of 1562 participants who were taking lipid-lowering medication as well as diabetics and those who lacked measurements of dyslipidemia and/or CAC. Individuals with triglycerides > 500 were not specifically excluded from the analysis. However, in those cases it is usually not possible to calculate LDL-c if using the Friedewald formula, and patients without LDL-c values would have been excluded from our analysis.
Cardiovascular risk factors were measured or collected, and included height, weight, and waist circumference, medical history including presence of diabetes (using the 2003 American Diabetes Association criteria), hypertension (defined as systolic blood pressure > 139 mm Hg at baseline visit, or diastolic blood pressure > 89 mm Hg, or by a history of physician diagnosed hypertension and taking a medication for hypertension), and assessment of personal habits such as alcohol and tobacco use [
The following cardiovascular risk factors were collected at MESA field centers: height, weight, waist circumference, alcohol and tobacco use, family history of heart attack, CAC score, diabetes, and hypertension (systolic blood pressure ≥ 140 mmHg at baseline visit, diastolic blood pressure ≥ 90 mmHg, or a history of taking an antihypertensive medication). Age and race/ethnicity were self-reported. Lipids, including total and high‐density lipoprotein cholesterol, triglycerides, inflammatory markers, and glucose levels, were measured from fasting plasma samples in a central laboratory (University of Vermont, Burlington, VT). Venous blood samples were collected after a 12 h fast by certified technicians using standardized venipuncture procedures. Samples were then centrifuged at 2000
Table
Lipid categories defined by HDL, LDL, and triglyceride levels.
Dyslipidemia category | HDL (mg/dL) | LDL-c (mg/dL) | Triglycerides (mg/dL) |
---|---|---|---|
Normolipidemia | >40 men | <160 | <150 |
>50 women | |||
Combined hyperlipidemia | No cutoff | ≥160 | ≥150 |
Hypercholesterolemia | No cutoff | ≥160 | <150 |
Dyslipidemia compatible with metabolic syndrome (MetS) | ≤40 men | <160 | ≥150 |
≤50 women | |||
Low HDL-c | ≤40 men | <160 | <150 |
≤50 women | |||
Hypertriglyceridemia | >40 men | <160 | ≥150 |
>50 women |
We created these dyslipidemia groups using criteria based on current National Cholesterol Education Program (NCEP)/Adult Treatment Panel- (ATP-) III guidelines that define LDL-C, HDL-C, and triglyceride thresholds as abnormal [
Participants were classified based on the most severe dyslipidemia. For example, someone would be classified in the only MetS if the person had low HDL-c and elevated triglycerides; the person would not be in the low HDL-c group. To define the subtype of dyslipidemia appropriately, we had to exclude diabetes and lipid-lowering therapy from this analysis. Diabetes is considered a coronary heart disease risk equivalent and there is a strong independent relationship of diabetes with low HDL-c, elevated triglycerides, and CAC extent. Participants receiving lipid-lowering therapy were excluded because lipid lowering has a substantial impact on all lipid parameters as well as on CAC. Fasting triglycerides were measured in plasma using a glycerol blanked enzymatic method developed by Trig/GB (Roche Diagnostics, Indianapolis, Indiana). Cholesterol and HDL-c were measured in plasma on the Hitachi 911 using a cholesterol esterase, cholesterol oxidase reaction (Chol R1, Roche Diagnostics). For triglyceride levels < 400 mg/dL, the LDL-c was calculated using the Friedewald formula; otherwise nuclear magnetic resonance spectroscopy was used for triglycerides > 400 mg/dl [
Electron-beam computed tomography (EBT) or multidetector row helical computed tomography (MDCT) was used to measure CAC, defined by a minimum of 130 Hounsfield units [
Extent of CAC was analyzed according to the number of main coronary arteries (left main, left anterior descending, left circumflex, and right) with calcification ranging from 0 to 4. Multivessel CAC was defined as involvement of at least 2 coronary arteries. This included three-vessel CAC, which was defined as involvement of the left main or left anterior descending coronary artery in addition to CAC in the left circumflex and the right coronary arteries. Single vessel CAC was classified as a distinct entity apart from no CAC and from multivessel disease. Our statistical analysis focused on the relationship between dyslipidemias and specifically multivessel disease, as relationships between dyslipidemias and CAC score in general have previously been described [
A cross-sectional sample of participants from the MESA cohort was classified into 6 mutually exclusive dyslipidemia categories (including “normal” as reference group) based on their levels of LDL-c, HDL-c, and triglycerides (Table
The baseline characteristics of the study cohort are shown in Table
Baseline characteristics and cardiac risk factors of MESA cohort excluding those on lipid-lowering medications and diabetics (collected 2000–2002).
Variable | All | Normal | Combined | HC | MetS | Low HDL | HTG | |
---|---|---|---|---|---|---|---|---|
| 4917 | 2329 (47.4) | 172 (3.5) | 345 (7.0) | 793 (16.1) | 907 (18.5) | 371 (7.6) | |
| 61.6 (10.3) | 62.3 (10.5) | 61.9 (9.9) | 61.4 (9.5) | 60.8 (10.3) | 60.8 (10.6) | 61.5 (9.4) | 0.015 |
| 2311 (47.0) | 1054 (45.3) | 74 (43.0) | 155 (44.9) | 438 (55.2) | 406 (44.8) | 184 (49.6) | <0.001 |
| <0.001 | |||||||
White (%) | 1930 (39.3) | 960 (41.2) | 70 (40.7) | 135 (39.1) | 311 (39.2) | 282 (31.1) | 172 (46.4) | |
Asian (%) | 616 (12.5) | 279 (12.0) | 20 (11.6) | 31 (9.0) | 126 (15.9) | 117 (12.9) | 43 (11.6) | |
Black (%) | 1285 (26.1) | 700 (30.1) | 24 (14.0) | 113 (32.8) | 96 (12.1) | 303 (33.4) | 49 (13.2) | |
Hispanic (%) | 1086 (22.1) | 390 (16.8) | 58 (33.7) | 66 (19.1) | 260 (32.8) | 205 (22.6) | 107 (28.8) | |
| 4079 (83.0) | 2029 (87.1) | 136 (79.1) | 285 (82.6) | 609 (76.8) | 730 (80.5) | 290 (78.2) | <0.001 |
| 690 (14.0) | 296 (12.7) | 21 (12.2) | 46 (13.3) | 143 (18.0) | 129 (14.2) | 55 (14.8) | 0.011 |
| 1903 (38.7) | 874 (37.5) | 62 (36.1) | 127 (36.8) | 339 (42.8) | 336 (37.1) | 165 (44.5) | 0.014 |
| 96.8 (14.0) | 93.6 (14.4) | 100.2 (12.4) | 97.0 (12.7) | 101.4 (12.6) | 99.3 (13.5) | 99.0 (13.1) | <0.001 |
| 91.1 (19.2) | 88.5 (13.6) | 95.8 (26.0) | 92.0 (21.1) | 96.4 (27.9) | 91.8 (18.9) | 92.0 (19.3) | <0.001 |
| 7.9 (11.4–5.8) | 6.7 (9.1–5.0) | 9.1 (13.6–6.9) | 7.7 (11.0–5.8) | 10.9 (14.9–8.0) | 9.1 (12.6–6.5) | 8.6 (12.5–6.5) | <0.001 |
| 1.9 (4.2–0.8) | 1.5 (3.7–0.7) | 2.4 (5.7–1.2) | 1.9 (4.2–1.0) | 2.3 (4.4–1.1) | 2.2 (4.8–0.9) | 2.4 (5.8–1.1) | 0.0001 |
| 0.95 (0.22) | 0.94 (0.20) | 0.94 (0.19) | 0.96 (0.19) | 0.96 (0.24) | 0.94 (0.26) | 0.94 (0.23) | 0.352 |
| 197.1 (35.3) | 190.4 (27.1) | 262.9 (23.2) | 250.0 (22.2) | 197.1 (33.4) | 175.7 (26.5) | 212.0 (30.6) | <0.0001 |
| 120.4 (31.0) | 112.7 (24.4) | 177.2 (18.4) | 177.2 (18.8) | 113.8 (25.7) | 114.6 (24.0) | 116.3 (25.1) | <0.0001 |
| 51.5 (15.0) | 60.7 (14.2) | 44.7 (8.9) | 51.9 (11.6) | 37.4 (6.0) | 40.3 (6.1) | 54.2 (10.4) | <0.0001 |
| 108 (156–76) | 82 (105–63) | 189 (216–171) | 106 (126–85) | 203 (255–172) | 105 (127–83) | 185 (221–162) | <0.0001 |
| 0 (0–6252) | 0 (0–6252) | 4.08 (0–2791) | 3.74 (0–2348) | 0 (0–2946) | 0 (0–3358) | 0 (0–2867) | <0.001 |
| 2778 (56.50) | 1397 (59.9) | 82 (47.6) | 161 (46.6) | 412 (51.9) | 516 (56.8) | 210 (56.6) | <0.001 |
| <0.0001 | |||||||
| 2778 (56.4) | 1397 (59.9) | 82 (47.6) | 161 (46.6) | 412 (51.9) | 516 (56.8) | 210 (56.6) | |
| 889 (18.0) | 401 (17.2) | 36 (20.9) | 79 (22.9) | 139 (17.5) | 158 (17.4) | 76 (20.4) | |
| 622 (12.6) | 264 (11.3) | 18 (10.4) | 54 (15.6) | 118 (14.8) | 131 (14.4) | 37 (9.9) | |
| 480 (9.7) | 204 (8.7) | 27 (15.7) | 42 (12.1) | 89 (32.8) | 82 (9.0) | 36 (9.7) | |
| 148 (3.0) | 63 (2.7) | 9 (5.2) | 9 (2.6) | 35 (4.4) | 20 (2.2) | 12 (3.2) | |
| 1250 (25.4) | 531 (22.8) | 54 (31.4) | 105 (30.4) | 242 (30.5) | 233 (25.6) | 85 (22.9) | <0.0001 |
The majority of the cohort consisted of white (39%) and female (53%) participants, with an average age of 61.6 years. Most participants completed at least a high-school level of education (83%) and currently used alcohol (58%). Only 14% of participants were current smokers. Approximately 39% of the cohort had a history of hypertension, and 53% of the cohort had one of the five types of dyslipidemia.
The low HDL-c dyslipidemia group was the most common type of dyslipidemia, followed by the MetS group. The latter had the largest waist circumference (Table
A comparison was performed using Chi-Square test between the lipid groups and the number of affected vessels with CAC, showing significant between-group differences (
Number of coronary vessels with calcification as a function of lipid profiles. The between-group differences in the number of affected vessels with CAC, showing significant between-group differences
The normolipidemia group had the highest percentage of individuals with zero affected vessels (60%), whereas the combined hyperlipidemia group had the highest percentage of those with three- and four-vessel calcification (16% and 5%, resp.).
Unadjusted univariate Poisson regression analysis showed that combined hyperlipidemia, simple hypercholesterolemia, and dyslipidemia of metabolic syndrome had a statistically significant likelihood of having a multivessel CAC as compared to the normolipidemia reference group (Table
Unadjusted prevalence ratio of multivessel CAC as a function of lipid groups.
Lipid group | PR (95% CI) |
---|---|
Normolipidemia | Ref group |
Combined hyperlipidemia | 1.37 (1.04–1.82) |
Hypercholesterolemia | 1.33 (1.08–1.64) |
Metabolic syndrome dyslipidemia | 1.33 (1.14–1.55) |
Low HDL-c | 1.12 (0.96–1.31) |
Hypertriglyceridemia | 1.00 (0.79–1.26) |
By contrast, there was no statistically significant difference between the low HDL-c, hypertriglyceridemia, and normolipidemia groups in the prevalence of multivessel CAC in the unadjusted model. Subsequent multivariate Poisson regression analysis adjusting for the demographic and cardiac risk factors, including Agatston calcium score, showed that the same lipid groups (combined hyperlipidemia, simple hypercholesterolemia, and dyslipidemia of metabolic syndrome) and the multivessel CAC maintained statistical significance in the model, compared to the normolipidemia reference group (Table
Adjusted prevalence ratio of multivessel CAC as a function of lipid groups.
Adjusted PR (95% CI) | |
---|---|
| |
Normolipidemia | Ref group |
Combined hyperlipidemia | 1.41 (1.06–1.87) |
Hypercholesterolemia | 1.55 (1.26–1.92) |
Metabolic syndrome dyslipidemia | 1.28 (1.09–1.51) |
Low HD-c | 1.20 (1.02–1.40) |
Hypertriglyceridemia | 1.05 (0.83–1.33) |
| 1.05 (1.05–1.06) |
| 1.71 (1.49–1.96) |
| |
White | Ref group |
Asian | 1.23 (1.04–1.45) |
Black | 1.11 (0.88–1.39) |
Hispanic | 0.89 (0.74–1.07) |
| 1.05 (0.89–1.23) |
| 1.32 (1.12–1.55) |
| 1.19 (1.05–1.34) |
| 1.01 (1.0004–1.010757) |
| 0.99 (0.99–1.00) |
| 1.00 (0.99–1.01) |
| 1.00 (0.99–1.01) |
| 0.96 (0.74–1.24) |
| 1.00 (1.0004–1.0005) |
Final model adjusting for age, gender, race, high school education, smoking, hypertension, waist circumference, serum glucose level, serum insulin, serum CRP level, and Agatston’s calcium score. All variables are adjusted simultaneously.
Previous literature has shown that different types of dyslipidemia have varying association with CAC scores [
Similar to Paramsothy et al., we found that isolated hypertriglyceridemia was not associated with CAC extent [
In contrast to the Paramsothy study where no significant association was found between low HDL-c and prevalent CAC, we found an association between low HDL-c and increased extent of CAC. The prevalence of low HDL-c group and adjusted risk factors were identical in both studies. In their study of 6093 participants, Allison and Wright showed that the individuals with an HDL-c level < 40 mg/dl had significantly higher calcium scores while increases in HDL-c were associated with a significant reduction in risk for the presence of any calcified plaque. Multivariate logistic regression revealed that HDL-c is predictive of calcified plaque development independent of LDL-c. However, sensitivity and positive predictive values for HDL-c were low [
Combined hyperlipidemia, simple hypercholesterolemia, and participants with MetS had significantly increased risk of multivessel involvement of the coronary arteries in patients with subclinical CAD [
Atherogenic dyslipidemia is largely underdiagnosed and undertreated in clinical practice per the findings of Fruchart et al. in the Residual Risk Reduction Initiative (R3i). These are prevalent in patients with type 2 diabetes, metabolic syndrome, and/or established CVD. Especially in MetS, these patients commonly had elevated ApoB levels, smaller LDL particle size, and elevated ApoCIII levels and as such metabolic syndrome is associated with residual CVD risk [
Gender may also contribute to CAC extent, as we found CAC extent was significantly increased among men. Male gender is a recognized independent risk factor for coronary heart disease by the Framingham Risk Score. In addition, in their study of 6814 participants, McClelland et al. found that men had greater calcium levels than women [
Coronary artery calcium, a known marker of coronary atherosclerotic plaque, has been consistently associated with cardiovascular morbidity and mortality [
Although the absence of CAC is not reassuring in symptomatic patients, the CAC score may be associated with myocardial perfusion defects in asymptomatic patients. Studies have shown that CAC burden and extent predict future coronary revascularization procedures [
The significant association between dyslipidemia types, except hypertriglyceridemia, and the extent of CAC may prompt more aggressive treatment and prevention of elevated LDL-c, low HDL-c, combined hyperlipidemia, and MetS. Simple hypercholesterolemia may have the greatest impact in determining the severity of atherosclerotic disease, especially in those with diabetes and taking lipid-lowering medications. As others have shown, LDL-c is the dominant lipid determinant of atherosclerotic disease [
As the MESA study gathered data from a large, multiethnic population, the results may be widely generalizable. The imaging and laboratory procedures were standardized at a common institution. Our study also has several limitations. Although we attempted to adjust for all possible confounding factors in our model, the residual confounding by unevaluated factors cannot be completely ruled out. As we examined cross-sectional associations, the possibility of temporal and selection biases may exist. Participants on statin therapy and with diabetes were excluded because these factors could potentially misclassify the lipid categories and confound the relationship between dyslipidemia and CAC. We used the term dyslipidemia compatible with metabolic syndrome instead of dyslipidemia of metabolic syndrome. To define this category, we did not factor in obesity or blood pressure to isolate the impact of dyslipidemia on the extent of subclinical atherosclerosis.
The results of this study further elucidate the role of CAC in cardiovascular disease. Clinicians already use the CAC score to predict future cardiovascular morbidity and mortality in asymptomatic patients with moderate risk factors. Callister et al. [
The association between multivessel coronary artery disease and the cardiovascular risk factors, including dyslipidemia, had been widely studied in patients with documented clinical CVD, using invasive coronary angiography [
Further research is needed in examining coronary artery calcium extent, not simply the score, as a potential tool in prognosticating, treating, and perhaps preventing subclinical atherosclerosis in otherwise low-risk populations. Other lipid parameters such as non-HDL-c and lipoprotein(a) (Lp(a)) have been proposed as independent risk factors for coronary heart disease and their impact on the extent of multivessel CAC needs to be explored further. A previous study using the MESA population showed that non-HDL-c is independently associated with increased CAC in patient populations without CAC at baseline and especially non-HDL-c > 190 is associated with a significant progression of CAC [
Genetic studies and multiple epidemiologic studies have identified Lp(a) as a risk factor for atherosclerotic diseases such as coronary heart disease and stroke [
In a population-based cohort, the extent of multivessel CAC was associated with different dyslipidemia types except for hypertriglyceridemia. The results were still significant even after controlling for CAC score. Future research should focus on the mechanistic understanding of this relationship for disease prevention and investigate the association between other promising lipid parameters and CAC extent in asymptomatic adults with subclinical atherosclerosis.
Areas under receiver operating curve
Coronary artery calcium
Coronary artery disease
Coronary calcium score
Carotid intima media thickness
C-reactive protein
Cardiovascular disease
Electron-beam computed tomography
Framingham Risk Score
High-density lipoprotein cholesterol
Lipoprotein(a)
Low density lipoprotein cholesterol
Multidetector row helical computed tomography
Dyslipidemia compatible with metabolic syndrome
Multiethnic Study of Atherosclerosis
Triglycerides.
The authors declare no conflicts of interest.
Moshrik Abd alamir is the principal investigator. Michael Goyfman contributed to study design and is an editor. Adib Chaus contributed to literature research and manuscript format. Firas Dabbous contributed to statistics. Leslie Tamura contributed to drafting initial manuscript. Veit Sandfort contributed to study design and is an editor. Alan Brown is the lipid expert of the study. Mathew Budoff is the senior author.
This research was supported by the National Heart, Lung, and Blood Institute (R01-HL-071739 and Contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169). 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