https://orcid.org/0009-0003-5156-3532
Abstract
Investigate sex-specific associations between total cholesterol, non-HDL cholesterol (non-HDL-C), and the burden of atherosclerosis assessed by coronary artery calcium (CAC) score.
A total of 10 049 participants (women: 958, men: 9091) aged 49–75 years, without known cardiovascular disease (CVD) or current use of lipid-lowering medication, were included from the Danish Risk Score study and the Danish Cardiovascular Screening Trial cohorts. Logistic regression models and zero-inflated negative binomial regression models were used to estimate odds ratio (OR), the incidence rate ratio (IRR), and 95% confidence intervals (CIs) for the association between total cholesterol, non-HDL-C, and CAC presence (CAC > 0) and extent. All analyses were adjusted for age, body mass index, diabetes, smoking, hypertension, and family history of CVD. The OR for presence of CAC and total cholesterol was 1.09 (95% CI: 0.94–1.27) in women and 1.26 (95% CI: 1.19–1.33) in men. The OR for presence of CAC and non-HDL-C was 1.12 (95% CI: 0.96–1.29) in women and 1.25 (95% CI: 1.18–1.33) in men. No significant association between increased total cholesterol and extent of CAC was found, regardless of sex (women: IRR: 0.99; 95% CI: 0.83–1.19; men: IRR: 1.04; 95% CI: 0.997–1.07). Non-HDL-C was significantly associated with extent of CAC in men (IRR: 1.04; 95% CI: 1.001–1.08) but not in women (IRR: 0.93; 95% CI: 0.78–1.12).
Total cholesterol was associated with presence of CAC, and non-HDL-Cs were associated with presence and extent of the CAC score in men. No association by total cholesterol or non-HDL-C was found among women.
Lay Summary
In both men and women, high cholesterol increases the risk of having a heart attack. It is unclear whether the significance of high cholesterol is as serious in women as it is in men regarding the development of atherosclerosis. In this study, we investigated the relationship between cholesterol and atherosclerosis in 10 049 Danish men and women, respectively. Atherosclerosis was measured by cardiac computed tomography (CT) scans. We found that:
High total cholesterol and non-HDL-C were both associated with calcifications in men.
High total cholesterol and non-HDL-C were not associated with calcifications in women.
Introduction
Cardiovascular disease (CVD) is the most common cause of death in the Western World, where it alone accounts for 45% of all deaths in both men and women, although women develop CVD later than men.1 The established modifiable risk factors are high cholesterol, smoking, hypertension, diabetes mellitus, increased alcohol consumption, and obesity.1–5 Although sex and age are very strong non-modifiable risk factors, evidence suggests sex-based differences in risk persist until the age of about 60, after which women catch up to the risk of men indicating a potential interaction between age and sex in the risk of CVD.6
Coronary artery calcification (CAC) is a specific marker of coronary artery atherosclerosis7 and as such a strong predictor of atherosclerotic CVD, using the Agatston method to quantify the extent of calcified plaque and expressed as the CAC score.7 Having a CAC score of zero means no presence of calcified plaques in the coronary arteries and carries a very low risk of future coronary events.8–10 The CAC score is currently the best-established imaging marker to improve the stratification of CVD risk.4
Until 2021, total cholesterol was used in the European Systemic Coronary Risk Estimation (SCORE) chart.11 However, the updated guidelines introduced in 2021 now advocate the adoption of the Systemic Coronary Risk Estimation 2 (SCORE2) and SCORE2-Older Person (SCORE2-OP) for estimating the 10-year risk of cardiovascular events (both fatal and non-fatal) within the population. These updated guidelines integrate non-HDL cholesterol (non-HDL-C) instead of total cholesterol, calculated as the difference between total cholesterol and HDL-C, into the algorithm.4
Women experience a pronounced increase in total cholesterol during the menopause resulting in higher average concentrations than men,6,12 but men have a higher CAC score than women at any age, and men develop cardiovascular diseases before women.13–16 A sex- and age-specific investigation of the association between total cholesterol, non-HDL-C, and CAC score may therefore bring knowledge that could facilitate CVD prevention strategies.
Thus, we aimed to investigate sex-specific associations between total blood cholesterol, non-HDL-C concentrations, and the burden of atherosclerosis assessed by the CAC score.
Methods
Study population
The current study is based on data from participants in the Danish Risk Score study (DanRisk) and the Danish Cardiovascular Screening Trial (DANCAVAS) cohorts.13,14,17 All the participants were randomly selected from the general population of Denmark by the national civil registry where all the residents of Denmark are assigned a unique personal identification number. The inclusion criteria for the DanRisk study were men and women born in 1949 (49–50 years) and 1959 (59–60 years) living in the Southern region of Denmark. In total, 1257 (participation rate 69%) accepted the invitation to participate in the DanRisk study in 2009/2010.13 The inclusion criteria for the DANCAVAS pilot study were men and women aged 65–74 years living in the Northern part of Funen and the city of Odense. In total, 1318 (participation rate 64%) accepted the invitation to participate in the DANCAVAS pilot study in 2014/2015.14 The inclusion criteria for the DANCAVAS trial were men aged 60–74 years living on the island of Funen, or in the surrounding communities of Vejle, Silkeborg, or Nykøbing Falster. Approximately 15 000 men (participation rate 62%) accepted the invitation to participate in the DANCAVAS study performed from 2014 to 2019.14,17 There were no exclusion criteria in the DanRisk and DANCAVAS studies. A total of 16 241 men and women (women, n = 1398, 8.6%) aged 49–74 years participated in these studies. The inclusion of participants older than 74 years in this study is attributed to the fact that individuals from the DANCAVAS studies were initially invited to participate at the age of 74; however, their examinations were conducted after reaching the age of 75. The inclusion of participants, stratified by sex in the DanRisk and DANCAVAS studies, is depicted in Figure 1.
Flowchart of the inclusions and exclusions of participants from the DanRisk study (n = 1257) and the DANCAVAS studies (n = 14 984) for the present study. n, number of participants.
All the participants filled in questionnaires about risk factors, relevant history of CVD, family history of CVD, smoking history, and medication.13,14,17 The participants were examined at 1 of 6 different centres (Odense, Esbjerg, Svendborg, Vejle, Nykøbing, or Silkeborg). The participant’s weight, height, and blood pressure were measured on the day of the examination. Non-fasting blood samples were collected to determine LDL cholesterol (LDL-C), HDL-C, triglycerides, and total cholesterol in DanRisk and DANCAVAS. Furthermore, blood glucose was used in DanRisk to measure for diabetes, whereas DANCAVAS utilized HgbA1c. The blood samples were collected before a non-contrast cardiac computed tomography (NCCT) scan was performed for all participants.13,14,17
For the present study, we excluded participants using lipid-lowering medication (n = 5132) or having known CVD (myocardial infarction, percutaneous coronary intervention, coronary artery bypass graft, heart valve surgery, peripheral arterial surgery, or stroke; n = 637). The exclusion of participants with known CVD was done to minimize the potential uncertainties during the CAC score assessments performed in the DanRisk and DANVACAS studies, as objects such as stents in the coronaries cause artefacts in the NCCT scans, which would make precise measurements of the CAC score difficult. Furthermore, 423 participants were excluded due to missing data in the covariables used in the analysis. This led to a total study population of 10 049 participants (n = 958 women, 9.5%). A flowchart of the inclusion and exclusions is shown in Figure 1. To examine the effect of excluding participants using lipid-lowering medication or with known CVD on the results, sensitivity analyses were performed where these participants were incorporated.
Measurement of coronary artery calcification
The CAC score was assessed based on the NCCT scans followed by calculation of the Agatston score. All scans performed in the DanRisk study were done using 64-slice computed tomography (CT) scanners.13 The NCCT scans performed in the DANCAVAS studies were all done using either 128, 256, or 320-slice CT scanners. All scans were performed during an inspiratory breath-hold, and imaging was prospectively electrocardiographic-triggered at an interval between 50 and 75% of the R–R interval, depending on the CT scanner used at the different centres.13,14,17 The types of CT scanners and the exact scanning parameters have been thoroughly described for both the DanRisk and DANCAVAS studies previously.18 Experienced cardiologists and trained radiographers estimated the CAC score from the NCCT scans. Each region of interest found in the coronary arteries was measured and summed up to give the CAC score for each of the participants.13,14,17
Assessment of serum total cholesterol
In accordance with established procedural guidelines, trained healthcare personnel collected blood samples after the completion of the NCCT scans in all the three studies. These specimens underwent analysis using an automated enzymatic method to determine the serum total cholesterol concentrations.13,14,17 LDL-C, HDL-C, and triglycerides were analysed through routine laboratory measurements in conjunction with the determination of the serum total cholesterol concentration.13,14,17 Non-HDL-C was derived from the total cholesterol variable by subtracting the HDL value from each participant’s total blood cholesterol.4
Assessment of covariates
All the participants responded to a comprehensive questionnaire sent to them before the examination day and received a follow-up interview on the day of the examination about their current health status including symptoms like angina, dyspnoea, palpitations, claudication, or prior diseases including CVD. Self-reported use of medication classified as thrombocyte aggregation inhibitor, anticoagulant, lipid-lowering, antiarrhythmic, thiazide, β-blocker, angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker, calcium blocker, potassium-saving diuretic, loop diuretic, antidiabetics, and others.13,14,17
Cardiovascular disease was defined as a self-reported medical history of previous stroke, atrial fibrillation, myocardial infarction, percutaneous coronary intervention, coronary artery bypass graft, heart valve surgery, aneurysm, or peripheral artery disease. Participants who had one or more of these conditions were categorized as having CVD. Smoking status was self-reported and categorized as never, former, or current smoker. A family history of CVD was reported as having a parent, sibling, or child suffering from stroke or myocardial infarction before age 55 (first-degree man) or 65 (first-degree woman). Height in metres without shoes and weight in kilograms in light clothes were measured on the day of the NCCT scan and the body mass index (BMI) was calculated. Systolic and diastolic blood pressures were measured three times after 5 min of supine rest, where the last two values were averaged for the analysis by technicians or trained medical students.13,14,17 The participants with one or more of the following criteria were defined as having hypertension; a systolic blood pressure ≥140 mmHg, a diastolic blood pressure ≥90 mmHg, or using antihypertensive medication. The participants with one or more of the following criteria were defined as having diabetes mellitus; self-reported diabetes mellitus, or a fasting plasma blood glucose level ≥7.0 mmol/L on two separate days,13 a haemoglobin A1c level >48 mmol/mol,17 or use of anti-diabetic medication. The participants with one or more of the following criteria were defined as having hypercholesterolaemia, self-reported hypercholesterolaemia, or a total plasma cholesterol concentration ≥5 mmol/L. The CAC score was stratified into five categories: CAC score of 0, > 0−< 10, ≥10 − < 100, ≥100−< 400, and ≥ 400.
Statistical analysis
Participant characteristics were stratified by sex and presented as means and standard deviations of continuously distributed covariates (age, height, BMI, systolic blood pressure, diastolic blood pressure, LDL-C, HDL-C, triglycerides, total cholesterol, and non-HDL-C). The distribution of categorical covariates (smoking, family history of CVD, hypertension, use of antihypertensive medication, hypercholesterolaemia, diabetes, use of diabetes medication, and the CAC score) was presented in terms of counts and percentages. All results for total cholesterol, non-HDL-C, LDL-C, HDL-C, and triglycerides are presented in millimoles per litre.
Two different statistical models were used to examine the association between total cholesterol, non-HDL-C, and CAC presence and extent, respectively. Logistic regression models using the CAC score as a dichotomized variable (i.e. having a CAC score above zero; no, yes) were used to estimate the odds ratio (OR) with corresponding 95% confidence intervals (CI) for the association between total cholesterol, non-HDL-C, and a CAC score >0. The assumptions for the logistic regression models were met.
The distribution of the continuous CAC score variable was right-skewed with a large proportion of zeroes; hence, it was not possible to meet the assumptions for linear regression models by performing a 1 + log transformation. Therefore, a zero-inflated negative binomial regression model was chosen for the examination of the association of total cholesterol, non-HDL-C, and extent of the CAC score as a continuous variable. Using the zero-inflated negative binomial regression model on count data with excess zeroes, it was possible to examine whether the participants with a CAC score of zero and those with non-zero scores were generated by related mechanisms.19 The zero-inflated negative binomial regression model distinguishes the zero CAC scores from each other by including different factors that can affect how the zero-response occurred. In this analysis, the age variable was inflated in the regression model to distinguish the zero-responses from each other due to age having the strongest correlation with a non-zero prevalence beside sex.20 Subsequently, analyses with an inflated age variable and adjusted for the potential confounders were performed. The results from the zero-inflated negative binomial regression models were presented as an incidence rate ratio (IRR) to make the results easier for interpretation.
All analyses were performed stratified by sex due to significant interaction between sex and total cholesterol and non-HDL-C observed in the various analyses (see Supplementary material online, Table S1).
All models were adjusted for the same set of covariates chosen a priori based on previously published literature and consisting of the established factors known to be associated with the CAC score; age, BMI, smoking status, diabetes mellitus, hypertension, and family history of CVD.21
Additional analyses conducted on LDL-C and its association with the CAC score can be found in the Supplementary materials.
Finally, we performed four sensitivity analyses. First, we included all the potential confounders used in the adjusted models in the prediction algorithm of the zero-inflated negative binomial regression model. The reasoning behind this strategy was to consider the differentiated effects of confounders on both zero and count values, as well as to investigate latent associations that might remain veiled when solely implementing adjustment procedures. Furthermore, we included the participants who reported use of lipid-lowering medication or with known CVD at baseline and adjusted further for these two covariates. In the final sensitivity analysis, we repeated the analyses stratified by cohort and sex and stratified by age (over/under 55 years) and sex.
All analyses were carried out using Stata/BE 17.0 (StataCorp, College Station, TX, USA).
Ethics
The DanRisk project was approved by the Regional Scientific Ethical Committee for Southern Denmark (S-20080140)13, and the DANVACAS projects were approved by the Southern Denmark Region Committee on Biomedical Research Ethics (S-20140028) and the Data Protection Agency.14,17 All participants signed informed consent forms at the interview before any examinations were performed.13,14,17
Results
Characteristics of the population stratified by sex are shown in Table 1. The distribution between women and men was balanced in DanRisk (women 52%, men 48%), whereas 95.5% of DANVACAS participants were men, resulting in a study population consisting of 958 women and 9091 men in the present study. The women were on average 5 years younger than the men and were more likely to be non-smokers. The women had more cases of family history of CVD but were less likely to have a history of hypertension and diabetes. Table 2 shows the characteristics of the population stratified by sex and age groups (age in years: ≤50; 51−≤ 60; 61−≤ 70; 71−≤ 75). The women were more likely to have a CAC score of zero compared with men. A gradual increase over the age span in total cholesterol, non-HDL-C, LDL-C, and triglycerides was observed among the women while the men had more stable concentrations of these lipoproteins and lipids. However, although HDL-C was higher through the ages for women compared with the men, for both sexes the concentrations were stable over the age span. The women were more often reported to have hypercholesterolaemia in the last three age groups (51−≤ 60, 61−≤ 70, and 71−≤ 75 years) compared with men.
Baseline characteristics of the study population (n = 10 049) from the DanRisk and DANCAVAS studies stratified by sex
| Women | Men | |||
|---|---|---|---|---|
| n = 958 | n = 9091 | |||
| DanRisk | 554 | 57.8% | 512 | 5.6% |
| DANCAVAS | 404 | 42.2% | 8579 | 94.4% |
| Age (year) | 60.6 | (7.8) | 66.2 | (4.7) |
| Smoking | ||||
| Non-smokers | 484 | 50.5% | 3272 | 36.0% |
| Former smokers | 303 | 31.6% | 4329 | 47.6% |
| Active smokers | 171 | 17.8% | 1490 | 16.4% |
| Family history of CVD (yes) | 192 | 20.0% | 1099 | 12.1% |
| Height (cm) | 165.6 | (6.1) | 177.7 | (6.6) |
| BMI (kg/m2) | 26.2 | (5.0) | 27.5 | (4.1) |
| Systolic blood pressure (mmHg) | 142.2 | (22.3) | 148.7 | (18.9) |
| Diastolic blood pressure (mmHg) | 81.5 | (10.2) | 83.2 | (10.0) |
| Hypertensiona | 565 | 59.0% | 6809 | 74.9% |
| Use of antihypertensive medicineb | 222 | 23.2% | 2709 | 29.8% |
| LDL-C (mmol/L) | 3.4 | (0.9) | 3.3 | (0.8) |
| HDL-C (mmol/L) | 1.7 | (0.5) | 1.4 | (0.4) |
| Triglycerides (mmol/L) | 1.5 | (0.9) | 1.8 | (1.1) |
| Total cholesterol (mmol/L) | 5.8 | (1.0) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)c | 4.1 | (1.0) | 4.0 | (1.0) |
| Hypercholesterolaemiad | 798 | 83.3% | 6842 | 75.3% |
| Diabetese | 26 | 2.7% | 435 | 4.8% |
| Use of diabetes medicine | 14 | 1.5% | 249 | 2.7% |
| CAC score | ||||
| 0 | 571 | 59.6% | 1990 | 21.9% |
| > 0–< 10 | 104 | 10.9% | 950 | 10.4% |
| ≥ 10–< 100 | 171 | 17.8% | 2354 | 25.9% |
| ≥ 100–< 400 | 80 | 8.4% | 2129 | 23.4% |
| ≥ 400 | 32 | 3.3% | 1668 | 18.3% |
| Women | Men | |||
|---|---|---|---|---|
| n = 958 | n = 9091 | |||
| DanRisk | 554 | 57.8% | 512 | 5.6% |
| DANCAVAS | 404 | 42.2% | 8579 | 94.4% |
| Age (year) | 60.6 | (7.8) | 66.2 | (4.7) |
| Smoking | ||||
| Non-smokers | 484 | 50.5% | 3272 | 36.0% |
| Former smokers | 303 | 31.6% | 4329 | 47.6% |
| Active smokers | 171 | 17.8% | 1490 | 16.4% |
| Family history of CVD (yes) | 192 | 20.0% | 1099 | 12.1% |
| Height (cm) | 165.6 | (6.1) | 177.7 | (6.6) |
| BMI (kg/m2) | 26.2 | (5.0) | 27.5 | (4.1) |
| Systolic blood pressure (mmHg) | 142.2 | (22.3) | 148.7 | (18.9) |
| Diastolic blood pressure (mmHg) | 81.5 | (10.2) | 83.2 | (10.0) |
| Hypertensiona | 565 | 59.0% | 6809 | 74.9% |
| Use of antihypertensive medicineb | 222 | 23.2% | 2709 | 29.8% |
| LDL-C (mmol/L) | 3.4 | (0.9) | 3.3 | (0.8) |
| HDL-C (mmol/L) | 1.7 | (0.5) | 1.4 | (0.4) |
| Triglycerides (mmol/L) | 1.5 | (0.9) | 1.8 | (1.1) |
| Total cholesterol (mmol/L) | 5.8 | (1.0) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)c | 4.1 | (1.0) | 4.0 | (1.0) |
| Hypercholesterolaemiad | 798 | 83.3% | 6842 | 75.3% |
| Diabetese | 26 | 2.7% | 435 | 4.8% |
| Use of diabetes medicine | 14 | 1.5% | 249 | 2.7% |
| CAC score | ||||
| 0 | 571 | 59.6% | 1990 | 21.9% |
| > 0–< 10 | 104 | 10.9% | 950 | 10.4% |
| ≥ 10–< 100 | 171 | 17.8% | 2354 | 25.9% |
| ≥ 100–< 400 | 80 | 8.4% | 2129 | 23.4% |
| ≥ 400 | 32 | 3.3% | 1668 | 18.3% |
Numbers are n (%), mean (SD).
n, number of participants; BMI, body mass index; CAC, coronary artery calcification; CVD, cardiovascular disease; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
aBased on participants with one or more of the following criteria: systolic blood pressure ≥140 mmHg; diastolic blood pressure ≥90 mmHg; using antihypertensive medication.
bThiazid, beta-blockers, angiotensin converting enzyme inhibitors, or angiotensin ll receptor blockers, calcium antagonists.
cTotal blood cholesterol with HDL-C subtracted.
dBased on self-reported hypercholesterolaemia or total cholesterol >5 mmol/L.
eBased on self-reported diabetes, or a fasting plasma blood glucose level ≥7.0 mmol/L on two separate days.
Baseline characteristics of the study population (n = 10 049) from the DanRisk and DANCAVAS studies stratified by sex
| Women | Men | |||
|---|---|---|---|---|
| n = 958 | n = 9091 | |||
| DanRisk | 554 | 57.8% | 512 | 5.6% |
| DANCAVAS | 404 | 42.2% | 8579 | 94.4% |
| Age (year) | 60.6 | (7.8) | 66.2 | (4.7) |
| Smoking | ||||
| Non-smokers | 484 | 50.5% | 3272 | 36.0% |
| Former smokers | 303 | 31.6% | 4329 | 47.6% |
| Active smokers | 171 | 17.8% | 1490 | 16.4% |
| Family history of CVD (yes) | 192 | 20.0% | 1099 | 12.1% |
| Height (cm) | 165.6 | (6.1) | 177.7 | (6.6) |
| BMI (kg/m2) | 26.2 | (5.0) | 27.5 | (4.1) |
| Systolic blood pressure (mmHg) | 142.2 | (22.3) | 148.7 | (18.9) |
| Diastolic blood pressure (mmHg) | 81.5 | (10.2) | 83.2 | (10.0) |
| Hypertensiona | 565 | 59.0% | 6809 | 74.9% |
| Use of antihypertensive medicineb | 222 | 23.2% | 2709 | 29.8% |
| LDL-C (mmol/L) | 3.4 | (0.9) | 3.3 | (0.8) |
| HDL-C (mmol/L) | 1.7 | (0.5) | 1.4 | (0.4) |
| Triglycerides (mmol/L) | 1.5 | (0.9) | 1.8 | (1.1) |
| Total cholesterol (mmol/L) | 5.8 | (1.0) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)c | 4.1 | (1.0) | 4.0 | (1.0) |
| Hypercholesterolaemiad | 798 | 83.3% | 6842 | 75.3% |
| Diabetese | 26 | 2.7% | 435 | 4.8% |
| Use of diabetes medicine | 14 | 1.5% | 249 | 2.7% |
| CAC score | ||||
| 0 | 571 | 59.6% | 1990 | 21.9% |
| > 0–< 10 | 104 | 10.9% | 950 | 10.4% |
| ≥ 10–< 100 | 171 | 17.8% | 2354 | 25.9% |
| ≥ 100–< 400 | 80 | 8.4% | 2129 | 23.4% |
| ≥ 400 | 32 | 3.3% | 1668 | 18.3% |
| Women | Men | |||
|---|---|---|---|---|
| n = 958 | n = 9091 | |||
| DanRisk | 554 | 57.8% | 512 | 5.6% |
| DANCAVAS | 404 | 42.2% | 8579 | 94.4% |
| Age (year) | 60.6 | (7.8) | 66.2 | (4.7) |
| Smoking | ||||
| Non-smokers | 484 | 50.5% | 3272 | 36.0% |
| Former smokers | 303 | 31.6% | 4329 | 47.6% |
| Active smokers | 171 | 17.8% | 1490 | 16.4% |
| Family history of CVD (yes) | 192 | 20.0% | 1099 | 12.1% |
| Height (cm) | 165.6 | (6.1) | 177.7 | (6.6) |
| BMI (kg/m2) | 26.2 | (5.0) | 27.5 | (4.1) |
| Systolic blood pressure (mmHg) | 142.2 | (22.3) | 148.7 | (18.9) |
| Diastolic blood pressure (mmHg) | 81.5 | (10.2) | 83.2 | (10.0) |
| Hypertensiona | 565 | 59.0% | 6809 | 74.9% |
| Use of antihypertensive medicineb | 222 | 23.2% | 2709 | 29.8% |
| LDL-C (mmol/L) | 3.4 | (0.9) | 3.3 | (0.8) |
| HDL-C (mmol/L) | 1.7 | (0.5) | 1.4 | (0.4) |
| Triglycerides (mmol/L) | 1.5 | (0.9) | 1.8 | (1.1) |
| Total cholesterol (mmol/L) | 5.8 | (1.0) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)c | 4.1 | (1.0) | 4.0 | (1.0) |
| Hypercholesterolaemiad | 798 | 83.3% | 6842 | 75.3% |
| Diabetese | 26 | 2.7% | 435 | 4.8% |
| Use of diabetes medicine | 14 | 1.5% | 249 | 2.7% |
| CAC score | ||||
| 0 | 571 | 59.6% | 1990 | 21.9% |
| > 0–< 10 | 104 | 10.9% | 950 | 10.4% |
| ≥ 10–< 100 | 171 | 17.8% | 2354 | 25.9% |
| ≥ 100–< 400 | 80 | 8.4% | 2129 | 23.4% |
| ≥ 400 | 32 | 3.3% | 1668 | 18.3% |
Numbers are n (%), mean (SD).
n, number of participants; BMI, body mass index; CAC, coronary artery calcification; CVD, cardiovascular disease; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
aBased on participants with one or more of the following criteria: systolic blood pressure ≥140 mmHg; diastolic blood pressure ≥90 mmHg; using antihypertensive medication.
bThiazid, beta-blockers, angiotensin converting enzyme inhibitors, or angiotensin ll receptor blockers, calcium antagonists.
cTotal blood cholesterol with HDL-C subtracted.
dBased on self-reported hypercholesterolaemia or total cholesterol >5 mmol/L.
eBased on self-reported diabetes, or a fasting plasma blood glucose level ≥7.0 mmol/L on two separate days.
Baseline characteristics of the study population (n = 10 049) from the DanRisk and DANCAVAS studies stratified by age groups and sex
| Women | Men | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Age (years) | ≤ 50 n = 81 | 51 to ≤ 60 n = 292 | 61 to ≤ 70 n = 487 | 71 to ≤ 75 n = 98 | ≤50 n = 66 | 51 to ≤ 60 n = 596 | 61 to ≤ 70 n = 6682 | 71 to ≤ 75 n = 1747 | ||||||||
| CAC score | ||||||||||||||||
| 0 | 63 | 77.8% | 227 | 77.7% | 251 | 51.5% | 30 | 30.6% | 34 | 51.5% | 265 | 44.5% | 1476 | 22.1% | 215 | 12.3% |
| > 0–< 10 | 9 | 11.1% | 32 | 11.0% | 48 | 9.9% | 15 | 15.3% | 16 | 24.2% | 75 | 12.6% | 725 | 10.9% | 134 | 7.7% |
| ≥10–< 100 | 8 | 9.9% | 20 | 6.8% | 117 | 24.0% | 26 | 26.5% | 9 | 13.6% | 133 | 22.3% | 1832 | 27.4% | 380 | 21.8% |
| ≥ 100–< 400 | 1 | 1.2% | 9 | 3.1% | 51 | 10.5% | 19 | 19.4% | 4 | 6.1% | 89 | 14.9% | 1519 | 22.7% | 517 | 29.6% |
| ≥400 | 0 | 0.0% | 4 | 1.4% | 20 | 4.1% | 8 | 8.2% | 3 | 4.6% | 34 | 5.7% | 1130 | 16.9% | 501 | 28.7% |
| Smoking | ||||||||||||||||
| Non-smokers | 46 | 56.8% | 131 | 44.9% | 247 | 50.7% | 60 | 61.2% | 28 | 42.4% | 239 | 40.1% | 2377 | 35.6% | 628 | 35.9% |
| Former smokers | 20 | 24.7% | 97 | 33.2% | 161 | 33.1% | 25 | 25.5% | 15 | 22.7% | 219 | 36.7% | 3181 | 47.6% | 914 | 52.3% |
| Active smokers | 15 | 18.5% | 64 | 21.9% | 79 | 16.2% | 13 | 13.3% | 23 | 34.8% | 138 | 23.2% | 1124 | 16.8% | 205 | 11.7% |
| BMI (kg/m2) | 26.2 | (5.5) | 26.2 | (4.6) | 26.2 | (5.0) | 25.9 | (5.5) | 25.9 | (3.3) | 27.6 | (4.5) | 27.6 | (4.1) | 27.3 | (3.9) |
| LDL-C (mmol/L) | 2.8 | (0.7) | 3.2 | (0.8) | 3.6 | (0.8) | 3.6 | (0.8) | 3.4 | (0.9) | 3.3 | (0.9) | 3.3 | (0.8) | 3.2 | (0.8) |
| HDL-C (mmol/L) | 1.6 | (0.4) | 1.7 | (0.5) | 1.7 | (0.5) | 1.7 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.5 | (0.4) |
| Triglycerides (mmol/L) | 1.2 | (0.6) | 1.3 | (0.9) | 1.6 | (0.9) | 1.8 | (0.9) | 1.7 | (1.0) | 1.8 | (1.2) | 1.8 | (1.1) | 1.7 | (0.9) |
| Total cholesterol (mmol/L) | 5.0 | (0.8) | 5.6 | (0.9) | 6.0 | (0.9) | 6.1 | (1.0) | 5.6 | (1.0) | 5.4 | (1.0) | 5.5 | (0.9) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)a | 3.4 | (0.8) | 3.9 | (1.0) | 4.3 | (1.0) | 4.4 | (1.0) | 4.2 | (1.1) | 4.0 | (1.0) | 4.0 | (1.0) | 3.9 | (0.9) |
| Hypercholesterolaemiab | 41 | 50.6% | 228 | 78.1% | 441 | 90.6% | 88 | 89.8% | 50 | 75.8% | 433 | 72.7% | 5092 | 76.2% | 1267 | 72.5% |
| Hypertensionc | 22 | 27.2% | 127 | 43.5% | 333 | 68.4% | 83 | 84.7% | 30 | 45.5% | 334 | 56.0% | 5000 | 74.8% | 1445 | 82.7% |
| Family history of CVD (yes) | 14 | 17.3% | 83 | 28.4% | 84 | 17.2% | 11 | 11.2% | 15 | 22.7% | 103 | 17.3% | 816 | 12.2% | 165 | 9.4% |
| Diabetesd | 2 | 2.5% | 5 | 1.7% | 11 | 2.3% | 8 | 8.2% | 1 | 1.5% | 18 | 3.0% | 308 | 4.6% | 108 | 6.2% |
| Women | Men | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Age (years) | ≤ 50 n = 81 | 51 to ≤ 60 n = 292 | 61 to ≤ 70 n = 487 | 71 to ≤ 75 n = 98 | ≤50 n = 66 | 51 to ≤ 60 n = 596 | 61 to ≤ 70 n = 6682 | 71 to ≤ 75 n = 1747 | ||||||||
| CAC score | ||||||||||||||||
| 0 | 63 | 77.8% | 227 | 77.7% | 251 | 51.5% | 30 | 30.6% | 34 | 51.5% | 265 | 44.5% | 1476 | 22.1% | 215 | 12.3% |
| > 0–< 10 | 9 | 11.1% | 32 | 11.0% | 48 | 9.9% | 15 | 15.3% | 16 | 24.2% | 75 | 12.6% | 725 | 10.9% | 134 | 7.7% |
| ≥10–< 100 | 8 | 9.9% | 20 | 6.8% | 117 | 24.0% | 26 | 26.5% | 9 | 13.6% | 133 | 22.3% | 1832 | 27.4% | 380 | 21.8% |
| ≥ 100–< 400 | 1 | 1.2% | 9 | 3.1% | 51 | 10.5% | 19 | 19.4% | 4 | 6.1% | 89 | 14.9% | 1519 | 22.7% | 517 | 29.6% |
| ≥400 | 0 | 0.0% | 4 | 1.4% | 20 | 4.1% | 8 | 8.2% | 3 | 4.6% | 34 | 5.7% | 1130 | 16.9% | 501 | 28.7% |
| Smoking | ||||||||||||||||
| Non-smokers | 46 | 56.8% | 131 | 44.9% | 247 | 50.7% | 60 | 61.2% | 28 | 42.4% | 239 | 40.1% | 2377 | 35.6% | 628 | 35.9% |
| Former smokers | 20 | 24.7% | 97 | 33.2% | 161 | 33.1% | 25 | 25.5% | 15 | 22.7% | 219 | 36.7% | 3181 | 47.6% | 914 | 52.3% |
| Active smokers | 15 | 18.5% | 64 | 21.9% | 79 | 16.2% | 13 | 13.3% | 23 | 34.8% | 138 | 23.2% | 1124 | 16.8% | 205 | 11.7% |
| BMI (kg/m2) | 26.2 | (5.5) | 26.2 | (4.6) | 26.2 | (5.0) | 25.9 | (5.5) | 25.9 | (3.3) | 27.6 | (4.5) | 27.6 | (4.1) | 27.3 | (3.9) |
| LDL-C (mmol/L) | 2.8 | (0.7) | 3.2 | (0.8) | 3.6 | (0.8) | 3.6 | (0.8) | 3.4 | (0.9) | 3.3 | (0.9) | 3.3 | (0.8) | 3.2 | (0.8) |
| HDL-C (mmol/L) | 1.6 | (0.4) | 1.7 | (0.5) | 1.7 | (0.5) | 1.7 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.5 | (0.4) |
| Triglycerides (mmol/L) | 1.2 | (0.6) | 1.3 | (0.9) | 1.6 | (0.9) | 1.8 | (0.9) | 1.7 | (1.0) | 1.8 | (1.2) | 1.8 | (1.1) | 1.7 | (0.9) |
| Total cholesterol (mmol/L) | 5.0 | (0.8) | 5.6 | (0.9) | 6.0 | (0.9) | 6.1 | (1.0) | 5.6 | (1.0) | 5.4 | (1.0) | 5.5 | (0.9) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)a | 3.4 | (0.8) | 3.9 | (1.0) | 4.3 | (1.0) | 4.4 | (1.0) | 4.2 | (1.1) | 4.0 | (1.0) | 4.0 | (1.0) | 3.9 | (0.9) |
| Hypercholesterolaemiab | 41 | 50.6% | 228 | 78.1% | 441 | 90.6% | 88 | 89.8% | 50 | 75.8% | 433 | 72.7% | 5092 | 76.2% | 1267 | 72.5% |
| Hypertensionc | 22 | 27.2% | 127 | 43.5% | 333 | 68.4% | 83 | 84.7% | 30 | 45.5% | 334 | 56.0% | 5000 | 74.8% | 1445 | 82.7% |
| Family history of CVD (yes) | 14 | 17.3% | 83 | 28.4% | 84 | 17.2% | 11 | 11.2% | 15 | 22.7% | 103 | 17.3% | 816 | 12.2% | 165 | 9.4% |
| Diabetesd | 2 | 2.5% | 5 | 1.7% | 11 | 2.3% | 8 | 8.2% | 1 | 1.5% | 18 | 3.0% | 308 | 4.6% | 108 | 6.2% |
Numbers are n (%), mean (SD).
n, number of participants; BMI, body mass index; CAC, coronary artery calcification; CVD, cardiovascular disease; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
aTotal blood cholesterol with HDL-C subtracted.
bBased on self-reported hypercholesterolaemia, or total cholesterol >5 mmol/L.
cBased on participants with one or more of the following criteria: systolic blood pressure ≥140 mmHg; diastolic blood pressure ≥90 mmHg; using antihypertensive medication.
dBased on self-reported diabetes, or a fasting plasma blood glucose level ≥7.0 mmol/L on two separate days.
Baseline characteristics of the study population (n = 10 049) from the DanRisk and DANCAVAS studies stratified by age groups and sex
| Women | Men | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Age (years) | ≤ 50 n = 81 | 51 to ≤ 60 n = 292 | 61 to ≤ 70 n = 487 | 71 to ≤ 75 n = 98 | ≤50 n = 66 | 51 to ≤ 60 n = 596 | 61 to ≤ 70 n = 6682 | 71 to ≤ 75 n = 1747 | ||||||||
| CAC score | ||||||||||||||||
| 0 | 63 | 77.8% | 227 | 77.7% | 251 | 51.5% | 30 | 30.6% | 34 | 51.5% | 265 | 44.5% | 1476 | 22.1% | 215 | 12.3% |
| > 0–< 10 | 9 | 11.1% | 32 | 11.0% | 48 | 9.9% | 15 | 15.3% | 16 | 24.2% | 75 | 12.6% | 725 | 10.9% | 134 | 7.7% |
| ≥10–< 100 | 8 | 9.9% | 20 | 6.8% | 117 | 24.0% | 26 | 26.5% | 9 | 13.6% | 133 | 22.3% | 1832 | 27.4% | 380 | 21.8% |
| ≥ 100–< 400 | 1 | 1.2% | 9 | 3.1% | 51 | 10.5% | 19 | 19.4% | 4 | 6.1% | 89 | 14.9% | 1519 | 22.7% | 517 | 29.6% |
| ≥400 | 0 | 0.0% | 4 | 1.4% | 20 | 4.1% | 8 | 8.2% | 3 | 4.6% | 34 | 5.7% | 1130 | 16.9% | 501 | 28.7% |
| Smoking | ||||||||||||||||
| Non-smokers | 46 | 56.8% | 131 | 44.9% | 247 | 50.7% | 60 | 61.2% | 28 | 42.4% | 239 | 40.1% | 2377 | 35.6% | 628 | 35.9% |
| Former smokers | 20 | 24.7% | 97 | 33.2% | 161 | 33.1% | 25 | 25.5% | 15 | 22.7% | 219 | 36.7% | 3181 | 47.6% | 914 | 52.3% |
| Active smokers | 15 | 18.5% | 64 | 21.9% | 79 | 16.2% | 13 | 13.3% | 23 | 34.8% | 138 | 23.2% | 1124 | 16.8% | 205 | 11.7% |
| BMI (kg/m2) | 26.2 | (5.5) | 26.2 | (4.6) | 26.2 | (5.0) | 25.9 | (5.5) | 25.9 | (3.3) | 27.6 | (4.5) | 27.6 | (4.1) | 27.3 | (3.9) |
| LDL-C (mmol/L) | 2.8 | (0.7) | 3.2 | (0.8) | 3.6 | (0.8) | 3.6 | (0.8) | 3.4 | (0.9) | 3.3 | (0.9) | 3.3 | (0.8) | 3.2 | (0.8) |
| HDL-C (mmol/L) | 1.6 | (0.4) | 1.7 | (0.5) | 1.7 | (0.5) | 1.7 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.5 | (0.4) |
| Triglycerides (mmol/L) | 1.2 | (0.6) | 1.3 | (0.9) | 1.6 | (0.9) | 1.8 | (0.9) | 1.7 | (1.0) | 1.8 | (1.2) | 1.8 | (1.1) | 1.7 | (0.9) |
| Total cholesterol (mmol/L) | 5.0 | (0.8) | 5.6 | (0.9) | 6.0 | (0.9) | 6.1 | (1.0) | 5.6 | (1.0) | 5.4 | (1.0) | 5.5 | (0.9) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)a | 3.4 | (0.8) | 3.9 | (1.0) | 4.3 | (1.0) | 4.4 | (1.0) | 4.2 | (1.1) | 4.0 | (1.0) | 4.0 | (1.0) | 3.9 | (0.9) |
| Hypercholesterolaemiab | 41 | 50.6% | 228 | 78.1% | 441 | 90.6% | 88 | 89.8% | 50 | 75.8% | 433 | 72.7% | 5092 | 76.2% | 1267 | 72.5% |
| Hypertensionc | 22 | 27.2% | 127 | 43.5% | 333 | 68.4% | 83 | 84.7% | 30 | 45.5% | 334 | 56.0% | 5000 | 74.8% | 1445 | 82.7% |
| Family history of CVD (yes) | 14 | 17.3% | 83 | 28.4% | 84 | 17.2% | 11 | 11.2% | 15 | 22.7% | 103 | 17.3% | 816 | 12.2% | 165 | 9.4% |
| Diabetesd | 2 | 2.5% | 5 | 1.7% | 11 | 2.3% | 8 | 8.2% | 1 | 1.5% | 18 | 3.0% | 308 | 4.6% | 108 | 6.2% |
| Women | Men | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Age (years) | ≤ 50 n = 81 | 51 to ≤ 60 n = 292 | 61 to ≤ 70 n = 487 | 71 to ≤ 75 n = 98 | ≤50 n = 66 | 51 to ≤ 60 n = 596 | 61 to ≤ 70 n = 6682 | 71 to ≤ 75 n = 1747 | ||||||||
| CAC score | ||||||||||||||||
| 0 | 63 | 77.8% | 227 | 77.7% | 251 | 51.5% | 30 | 30.6% | 34 | 51.5% | 265 | 44.5% | 1476 | 22.1% | 215 | 12.3% |
| > 0–< 10 | 9 | 11.1% | 32 | 11.0% | 48 | 9.9% | 15 | 15.3% | 16 | 24.2% | 75 | 12.6% | 725 | 10.9% | 134 | 7.7% |
| ≥10–< 100 | 8 | 9.9% | 20 | 6.8% | 117 | 24.0% | 26 | 26.5% | 9 | 13.6% | 133 | 22.3% | 1832 | 27.4% | 380 | 21.8% |
| ≥ 100–< 400 | 1 | 1.2% | 9 | 3.1% | 51 | 10.5% | 19 | 19.4% | 4 | 6.1% | 89 | 14.9% | 1519 | 22.7% | 517 | 29.6% |
| ≥400 | 0 | 0.0% | 4 | 1.4% | 20 | 4.1% | 8 | 8.2% | 3 | 4.6% | 34 | 5.7% | 1130 | 16.9% | 501 | 28.7% |
| Smoking | ||||||||||||||||
| Non-smokers | 46 | 56.8% | 131 | 44.9% | 247 | 50.7% | 60 | 61.2% | 28 | 42.4% | 239 | 40.1% | 2377 | 35.6% | 628 | 35.9% |
| Former smokers | 20 | 24.7% | 97 | 33.2% | 161 | 33.1% | 25 | 25.5% | 15 | 22.7% | 219 | 36.7% | 3181 | 47.6% | 914 | 52.3% |
| Active smokers | 15 | 18.5% | 64 | 21.9% | 79 | 16.2% | 13 | 13.3% | 23 | 34.8% | 138 | 23.2% | 1124 | 16.8% | 205 | 11.7% |
| BMI (kg/m2) | 26.2 | (5.5) | 26.2 | (4.6) | 26.2 | (5.0) | 25.9 | (5.5) | 25.9 | (3.3) | 27.6 | (4.5) | 27.6 | (4.1) | 27.3 | (3.9) |
| LDL-C (mmol/L) | 2.8 | (0.7) | 3.2 | (0.8) | 3.6 | (0.8) | 3.6 | (0.8) | 3.4 | (0.9) | 3.3 | (0.9) | 3.3 | (0.8) | 3.2 | (0.8) |
| HDL-C (mmol/L) | 1.6 | (0.4) | 1.7 | (0.5) | 1.7 | (0.5) | 1.7 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.4 | (0.4) | 1.5 | (0.4) |
| Triglycerides (mmol/L) | 1.2 | (0.6) | 1.3 | (0.9) | 1.6 | (0.9) | 1.8 | (0.9) | 1.7 | (1.0) | 1.8 | (1.2) | 1.8 | (1.1) | 1.7 | (0.9) |
| Total cholesterol (mmol/L) | 5.0 | (0.8) | 5.6 | (0.9) | 6.0 | (0.9) | 6.1 | (1.0) | 5.6 | (1.0) | 5.4 | (1.0) | 5.5 | (0.9) | 5.4 | (0.9) |
| Non-HDL-C (mmol/L)a | 3.4 | (0.8) | 3.9 | (1.0) | 4.3 | (1.0) | 4.4 | (1.0) | 4.2 | (1.1) | 4.0 | (1.0) | 4.0 | (1.0) | 3.9 | (0.9) |
| Hypercholesterolaemiab | 41 | 50.6% | 228 | 78.1% | 441 | 90.6% | 88 | 89.8% | 50 | 75.8% | 433 | 72.7% | 5092 | 76.2% | 1267 | 72.5% |
| Hypertensionc | 22 | 27.2% | 127 | 43.5% | 333 | 68.4% | 83 | 84.7% | 30 | 45.5% | 334 | 56.0% | 5000 | 74.8% | 1445 | 82.7% |
| Family history of CVD (yes) | 14 | 17.3% | 83 | 28.4% | 84 | 17.2% | 11 | 11.2% | 15 | 22.7% | 103 | 17.3% | 816 | 12.2% | 165 | 9.4% |
| Diabetesd | 2 | 2.5% | 5 | 1.7% | 11 | 2.3% | 8 | 8.2% | 1 | 1.5% | 18 | 3.0% | 308 | 4.6% | 108 | 6.2% |
Numbers are n (%), mean (SD).
n, number of participants; BMI, body mass index; CAC, coronary artery calcification; CVD, cardiovascular disease; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.
aTotal blood cholesterol with HDL-C subtracted.
bBased on self-reported hypercholesterolaemia, or total cholesterol >5 mmol/L.
cBased on participants with one or more of the following criteria: systolic blood pressure ≥140 mmHg; diastolic blood pressure ≥90 mmHg; using antihypertensive medication.
dBased on self-reported diabetes, or a fasting plasma blood glucose level ≥7.0 mmol/L on two separate days.
The interaction analyses between sex and total cholesterol and non-HDL-C for the CAC score showed significant interaction for all crude analyses for both the logistic and the zero-inflated negative binomial regression models. Similarly, there was a significant interaction for all the adjusted interaction models, except between sex and non-HDL-C in the logistic analyse (see Supplementary material online, Table S1).
When assessing the association between total cholesterol and having a CAC score >0 stratified by sex, we found that among men odds of a CAC score >0 increased with increasing total cholesterol (OR: 1.26; 95% CI: 1.19–1.33) but found no such association for women (OR: 1.09; 95% CI: 0.94–1.27; Figure 2). However, when examining the association between total cholesterol and the extent of the CAC score, we found no evidence of such association for both men and women (IRR: 1.04; 95% CI: 0.997–1.07, IRR: 0.99; 95% CI: 0.83–1.19, respectively; Figure 3).
Odds ratio and 95% confidence intervals for having a coronary artery calcium score >0 by total cholesterol and non-HDL cholesterol (non-HDL-C) in men and women from the DanRisk and DANCAVAS studies (n = 10 049). n, number of participants. aAdjusted for age, body mass index, diabetes, smoking, hypertension, and family history of cardiovascular disease.
Incidence rate ratio and 95% confidence intervals of the extent of the coronary artery calcium score by total cholesterol and by non-HDL cholesterol (non-HDL-C) in men and women from the DanRisk and DANCAVAS studies (n = 10 049). n, number of participants. aAdjusted for age, body mass index, diabetes, smoking, hypertension, and family history of cardiovascular disease.
When assessing the association between non-HDL-C and having a CAC score >0 stratified by sex, we found that among men odds of a CAC score >0 increased with increasing non-HDL-C (OR: 1.25; 95% CI: 1.18–1.33) but again found no such association for women (OR: 1.12; 95% CI: 0.96–1.29; Figure 2). When examining the association between non-HDL-C and the extent of the CAC score, we found that among men IRR increased with increasing non-HDL-C (IRR: 1.04; 95% CI: 1.001–1.08) but found no such association for women (IRR: 0.93; 95% CI: 0.78–1.12; Figure 3).
In the sensitivity analyses in which we further inflated BMI, smoking status, diabetes mellitus, hypertension, and family history of CVD in the zero-inflated negative binomial regression models, the associations were not qualitatively different from the results in the main analyses (results not shown).
Association between increasing LDL-C and risk of having a CAC score >0 was observed for both men and women (see Supplementary material online, Table S2). However, when examining the association between LDL-C and the extent of the CAC score, we found no evidence of such association for both sexes (see Supplementary material online, Table S3). The inclusion of participants using lipid-lowering medication or with known CVD did not result in qualitatively different results for the association between total cholesterol and the CAC score (see Supplementary material online, Table S4); however, the association between non-HDL-C and extend of the CAC score for men was no longer present (see Supplementary material online, Table S5). Separating the study populations into the DanRisk and DANCAVAS studies did not lead to significantly different results (see Supplementary material online, Tables S6 and S7). The same applied to stratification by age (over/under 55 years). However, it was not possible to perform the zero-inflated negative binomial regression model on women under 55 years, as there were insufficient women with CAC in this age group. Consequently, only the results for the logistic regression are presented (see Supplementary material online, Table S8).
Discussion
Our study included 10 049 women and men randomly selected from the general population of Denmark. They had no prior CVD and were without lipid-lowering medication. We found no associations between total cholesterol and non-HDL-C with presence and extent of the CAC score, respectively, among women. Opposed to that, we found that increased total cholesterol concentration was associated with presence but not the extent of CAC among men. Furthermore, increased non-HDL-C was associated with both presence and extent of CAC among men.
Our results challenge to some degree the findings in previous studies.13,20,22,23 However, some of these studies were performed in unstratified populations13,22; hence, the studies would not be able to show the sex difference in the association between total cholesterol and presence of CAC. Furthermore, the previous studies were performed in populations of younger age, i.e. mean age below 60 years.20,22 Unstratified by sex and younger age may potentially give too much weight to the effect of the women, which consequently demonstrates no association between total cholesterol and the CAC score in the general population. It could of course be questioned whether results from our cohort are valid;, however, as we demonstrate in our supplementary analysis for LDL-C, we confirmed results found in previous studies on the positive association between LDL-C and presence of CAC.19–22,24–26 We, therefore, believe that similar results indicate that the study population and the results are valid for investigating the sex difference in the effect of total cholesterol and non-HDL-C on the CAC score.
The present study demonstrates a strong association between total cholesterol, non-HDL-C, and the presence of CAC in men, while this association is not observed in women. This could, of course, be due to the low number of women in the study and might have been significant with a larger sample size. This study additionally identified an association between non-HDL-C and extent of CAC for men. However, with an IRR association of 1.04, and the sensitivity analysis including more participants showing no association, this association is considered to lack clinical significance.
There are numerous factors at play in the development of CAC and CVD in both women and men. Factors such as differences in sex hormones where women’s oestrogen production is known to delay the development of CVD,6 but also differences in gene expression patterns within the atherosclerotic tissues. Genes that are more active in women are linked to mesenchymal and endothelial cells, while genes with increased activity in men are specifically linked to immune system functions.6,27,28 General physiological differences such as women’s smaller diameter of the major blood vessels or their greater intrinsic heart rate compared with men’s may also be of importance.6 Additionally, a study has demonstrated that the size of LDL-C particles, which varies among individuals, has a significant impact on the development of CAC, even when accounting for total LDL-C levels, further adding to the complexity.29 The combination of these diverse factors emphasizes that the results of this study alone cannot independently explain the development of CAC and CVD. However, it provides insight into sex differences regarding the influence of cholesterols on CAC development and the necessity to stratify by sex when performing analysis related to CAC, total cholesterol, and non-HDL-C.
The findings from this study suggest that total cholesterol and non-HDL-C are associated with the presence of CAC but not extent of the CAC score, at least among men. This might be attributed to the biological onset of CAC. Cholesterol, specifically LDL-C, is implicated in initiating plaque development, yet once the plaque initiation occurs, the immune system becomes the predominant player in its further advancement.28 Additionally, there are limitations in the CT modalities’ detection of the early stages of CAC. This implies that the CAC score measured through CT scans, for instance, may not detect microcalcifications that are strongly correlated with plaque instability.30 A study using coronary CT angiography has demonstrated that women exhibit significantly higher proportions of mixed and non-calcified plaques compared with men, suggesting differences in plaque composition between the two sexes.31 The lack of association between total cholesterol and non-HDL-C in women may also be attributed to the protective effect of HDL-C against the presence of CAC, as demonstrated in previous studies.19,21,22 Table 2 indicates that women in the study generally have higher HDL-C levels than men across all age groups, which could influence the relationship with total cholesterol. When comparing the results for total cholesterol and non-HDL-C, it becomes apparent that the association strengthens and moves towards a more significant result when HDL-C is excluded from the analysis.
The findings of this study could also be influenced by its observational design, where there was no follow-up data included. Had follow-up data been included in the study, it would have provided a better opportunity to observe the influence of total cholesterol and non-HDL-Cs effect on the progression of CAC, by allowing more time for the potential microcalcifications to develop. This, in turn, would enhance the capability of the CT scans to detect more CAC. Furthermore, the women would have been older, thereby moving them further away from the menopause and its preventive effect on the development of CAC. In previous studies with longitudinal data, LDL-C and total cholesterol have been identified as independent indicators of CAC growth in individuals without CAC at baseline.21,24 A longitudinal study conducted on healthy adults aged 32–46 years with minimal risk factors showed that total cholesterol, LDL-C, and non-HDL-C are associated with the presence, incidence, and progression of CAC even when lipid parameters are optimal, suggesting that these lipids play a significant role in the development of CAC throughout life.32 However, further studies investigating the association between total cholesterol and non-HDL-C with CAC progression are necessary before reaching a definitive conclusion.
The strength of this study is the large, randomly selected population-based cohort from different areas of residence in Denmark, representing the general population, with no specific selection criteria regarding health except known CVD and use of lipid-lowering agents in the present analysis. The limited selection criteria and the representation of the general population makes this cohort distinctive from other large studies like MESA and the Heinz Nixdorf Recall Study.15,16,33 Furthermore, the CT scans in the present study were performed using contemporary cardiac CT scans, which increase the generalizability, whereas many previous studies used outdated 4-sliced CT scanners or electron-beam scanners,15,16,20–23 which according to a phantom study, do not result in the same CAC scores when compared with a modern 64-slice CT scanner.34 Additionally, we used the zero-inflated negative binomial regression model when analysing the association between total cholesterol, non-HDL-C, and their association with extent of the CAC score. This allows the CAC score to remain as a continuous variable thus retaining information that could otherwise have been lost. A previous study has shown that using a zero-inflated model on CAC score data was a better statistical model compared with the CAC score transformation used in other studies when a log transformation was not possible.19
This study has several limitations, often attributable to the limited number of women in the study sample. This disparity predominantly arises from the study design of DANCAVAS, which predominantly recruited men. The participation rates were 69% for the DanRisk study13 and 62% for the DANCAVAS study.17 Although higher than other studies25,35 the relatively low participation rates could cause selection bias.
To evaluate the impact of increased female representation in the study, we performed sensitivity analyses including participants using lipid-lowering medication or with known CVD. The results of these sensitivity analyses showed minimal deviation from the main analyses. It has previously been demonstrated that the non-participants in the DanRisk study had a lower socioeconomic status than the participants.36 Even though living in a low socioeconomic status area is associated with an increased incidence for presence of CAC, it is mostly due to an increased burden of CVD risk factors.37 However, given the adjustment for the potential confounders in the models and the comparable cardiovascular events and mortality rates between non-participants and participants in the DanRisk study,36 along with the minimal changes observed in the inclusion of participants using lipid-lowering medication or with known CVD, we consider our findings representative of the general population.38 This study is a cross-sectional study in which participants had their lipid biomarkers measured only once. This could impact the study’s results, as lipid biomarkers can fluctuate, potentially leading to overlooked associations. To minimize uncertainty, all participants using lipid-lowering medication and those with established CVD were excluded from the study population. Furthermore, because the findings of this study are consistent with those of other studies, we do not consider the potential fluctuation to have a significant effect on the results. Through questionnaires and interviews, the women reported having more cases of family history of CVD compared with the participating men. However, due to the nature of questionnaires and interviews, information bias could occur resulting in misclassifications of the participants and thus potentially affecting the results.39 However, the potential misclassification arising from the questionnaires and interviews is confined to the covariables in this study. Participants completed the questionnaires prior to the CT scans, without knowledge of their CAC score, which renders these misclassifications non-differential. Consequently, this might lead to an underestimation of the association between total cholesterol, non-HDL-C and the CAC score. Finally, the study participants were primarily white and of European ancestry. Therefore, generalisability to other ethnicities cannot be assumed.
Conclusions
In conclusion, this study showed associations between total cholesterol, non-HDL-C, and the presence of CAC for men but not women, in two Danish cohorts of middle-aged and older men and women. No association between total cholesterol and extent of the CAC score was found for neither men nor women and the observed association between non-HDL-C and extent of the CAC score in men was considered to lack clinical significance. The result of this study contributes to the existing knowledge of total cholesterols’ and non-HDL-Cs’ influence on the development of arteriosclerosis. However, further studies are required to explore the association between total cholesterol, non-HDL-C and the development of atherosclerosis especially in older women. These studies would help ascertain the impact of total cholesterol and non-HDL-C on women at a heightened risk of developing CVD or experiencing cardiovascular events.
Supplementary material
Supplementary material is available at European Journal of Preventive Cardiology.
Acknowledgements
The whole or parts of the work have not previously been presented.
Author contribution
K.W.L. and A.C.D. contributed to the design, acquisition, analysis, interpretation of data for the work, and drafted the manuscript. C.D. contributed to the design, analysis, and interpretation of data for the work, and drafted the manuscript. O.G. contributed to the statistical analysis of data and critically revised the manuscript. J.S.L., J.L., L.F., M.K., K.E., and M.B. contributed to acquisition and critically revised the manuscript. All gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy.
Funding
The authors did not receive any funding in relation to this work.
Data availability
The data used in the current study are available from the corresponding author upon reasonable request.
References
Author notes
Conflict of interest: none declared.
Comments