Abstract
Background:
Current guidelines recommend coronary artery calcium (CAC) scoring for stratification of atherosclerotic cardiovascular disease (ASCVD) risk only in patients with borderline to intermediate risk score by the pooled cohort equation with low density lipoprotein cholesterol (LDL-C) of 70–190 mg/dL. It remains unknown if CAC or thoracic aorta calcification (TAC), detected on routine chest computerized tomography (CT), can provide further risk stratification in patients with LDL-C≥190 mg/dL.
Methods:
From a multi-site medical center, we retrospectively identified all patients from March 2005 to June 2021 age≥ 40 years, without established ASCVD and LDL-C≥190 mg/dL who had non-gated non-contrast chest CT within 3 years of LDL-C measurement. Ordinal CAC and TAC scores were measured by visual inspection. Kaplan Meier curves and multivariable Cox-regression models were built to ascertain the association of CAC and TAC scores with all-cause mortality.
Results:
We included 811 patients with median age 59 [53–68] years, 262 (32.3%) were male, and LDL-C median level was 203 [194–217] mg/dl. Patients were followed for 6.2 [3.29–9.81] years, and 109 (13.4%) died. Overall, 376 (46.4%) of patients had CAC=0 and 226 (27.9%) had TAC=0. All-cause mortality increased with any CAC and moderate to severe TAC. In a multivariate model, patients with CAC had a significantly higher mortality compared to those without CAC: mild HR= 1.71 [1.03–2.83], moderate HR= 2.12 [1.14–3.94], and severe HR= 3.49 [1.94–6.27]. Patients with moderate TAC (HR= 2.34 [1.19–4.59]) and those with severe TAC (HR= 3.02 [1.36–6.74]) had higher mortality than those without TAC.
Conclusions:
In patients without history of ASCVD and LDL-C≥190 mg/dL, the presence and severity of CAC and TAC are independently associated with all-cause mortality.
Keywords: LDL-C≥190, coronary artery calcification, thoracic artery calcification
Subject terms: Computerized Tomography (CT), prognosis, lipids and cholesterol
Introduction
Patients with primary severe hypercholesterolemia, defined as LDL-C≥ 190 mg/dL, represent a higher risk group amongst patients with hyperlipidemia. This condition is not rare with an estimated prevalence of 5–7% of the Western population1, is associated with high lifetime risk of atherosclerotic cardiovascular disease (ASCVD) and has been linked to familial hypercholesterolemia (FH) phenotype2. As per American College of Cardiology (ACC)/ American Heart Association (AHA) cholesterol treatment guidelines, the pooled cohort equation (PCE), that represents an essential physicians’ tool for ASCVD risk stratification, should not be used in this population and these patients should be considered high-risk without further risk stratification3. It has been suggested that diverse genetic background and lifetime exposure to additional risk factors may introduce heterogeneity in this group. It is known that amongst patients with LDL-C≥ 190 mg/dL, carriers of monogenic mutations for FH have a higher risk of coronary disease compared to noncarriers1.
Coronary artery calcium (CAC) is associated with the total burden of coronary atherosclerosis and it predicts both cardiovascular and all-cause mortality independent of traditional risk factors across diverse populations4. Moreover, visual calcium scoring from non-gated non-contrast chest CT can predict events in a similar fashion in multiple clinical scenarios 5–7. Additionally, incidental detection of thoracic aortic calcification (TAC) is independently associated with death and may provide additional prognostic and reclassification value beyond CAC mostly for non-coronary events 8,9.
Current ACC/AHA cholesterol treatment guidelines recommend (CAC) evaluation for further ASCVD risk stratification for patients with borderline to intermediate ASCVD risk by the PCE with LDL-C (70 – 190 mg/dL) when the risk decision remains uncertain3. Whether CAC is predictive of cardiovascular events in asymptomatic patients with LDL-C≥ 190 mg/dL or FH has been evaluated in small studies and a meta-analysis with promising results10–12. A recent study evaluated CAC in statin treated patients with heterozygous FH form French and Spanish Registries showing improved cardiovascular risk stratification13. Moreover, a Danish study found in symptomatic patients referred for CT coronary angiogram that absence of plaque was associated with low event rates across all LDL-C levels including those with LDL-C≥ 190 mg/dL14. However, there is no data on the prognostic value of TAC and no studies to date have evaluated the role of CAC and/or TAC derived from routine chest CT in all-cause mortality risk stratification in patients with LDL-C≥ 190 mg/dL. Accordingly, we aimed to establish the clinical significance of CAC and TAC in patients with LDL-C≥ 190 mg/dL.
Methods
Patient selection
We conducted a retrospective cohort analysis of all patients ≥ 40 years of age, with at least one measurement of LDL-C level ≥190 mg/dL assessed between March 2005 and June 2021 without adjustment for statin use, who had a chest CT without contrast for any clinical indication within 3 years from the LDL-C blood test at Montefiore. We chose the 3-year window based on prior work by Dzaye et al. establishing the rate of conversion from zero to non-zero CAC and warranty period for the highest risk group at 3.6 years15. From the initially identified group consisting of 1,963 subjects, we excluded 560 patients with possible secondary causes of hyperlipidemia (biliary obstruction 24.8%, nephrotic syndrome 9.7% and hypothyroidism 65.5%), 443 patients with previously established ASCVD and 149 patients with history of cancer. Patients were included regardless of their statin use. A total of 811 subjects were therefore finally included in the study (Figure 1). Baseline demographic, clinical, laboratory data and outcomes were retrieved from our electronic medical records system. The primary outcome was all-cause mortality. The investigators had direct access to primary data. Indications for the CT scan were: lung cancer screening/nodule follow up (44%), respiratory symptoms (36%), infection (10%), vascular/cardiac (5%), trauma (1%) and unknown (4%). The study was approved by our Institutional Review Board (Office of Human Research Affairs at Albert Einstein College of Medicine) and was HIPAA compliant. Informed consent was waived due to retrospective study design. As per AHA Journals implementation of Transparency and Openness Promotion (TOP), we will make the data, methods used in the analysis, and materials used to conduct the research available upon request.
Figure 1. Flow chart.
Patient selection.
CT = Computerized tomography; CVA = Cerebrovascular accident; LDL-c = Low-density lipoprotein cholesterol; MI = Myocardial infarction; PAD = Peripheral artery disease.
CT acquisition
Patients were imaged on any of the following CT scanners without contrast administration or electrocardiographic gating: GE Optima 660, GE Lightspeed VCT, Siemens Somatom Sensation 16, and GE Optima CT 540. Images were reconstructed at 2.5 mm slice thickness.
Coronary and Thoracic calcium scoring
Four readers (J.A., E.L., P.N. and P.I.) blinded from clinical data calculated ordinal scores from standard non-gated non-contrast chest CT studies using the methods described by Shemesh et al.5. Briefly, the right, left main, left anterior descending, and left circumflex coronary arteries were each given a score of 0–3 for presence and extent of calcium with 0 (none), 1 (< 1/3 of the artery length calcified), 2 (≥ 1/3 to < 2/3 calcified), and 3 (≥ 2/3 calcified). These scores were summed across the 4 arteries, providing a range of possible scores from 0 to 12. Patients were further characterized in groups based on the total CAC as none (0), mild (1–3), moderate (4–6) or severe (7–12). TAC was scored in a similar fashion dividing the aorta into 4 segments (ascending, arch, descending up to heart level, descending up to the abdominal aorta) and each segment was scored as 0 (none), 1 (< 1/3 of the artery length calcified), 2 (≥ 1/3 to < 2/3 calcified), and 3 (≥ 2/3 calcified). Patients were further characterized in groups based on the total TAC as none (0), mild (1–3), moderate (4–6) or severe (7–12). Calcification in the aortic root and annulus was included in the TAC score. Calcification of the aortic valve and mitral annulus was not included.
Inter-reader agreement
In order to evaluate the agreement between the four readers’ scoring of CAC and TAC group classification, we selected a random subset (n=140, 17% of full cohort) of the CT chest scans and asked each reader to determine the CAC and TAC groups. Readers were not aware of the other investigators’ results. We utilized Kendall’s W test to assess if there was agreement. With a Kendall’s W of 1 indicating complete agreement, the values observed reflect strong agreement between our readers16. The Kendall’s W for the 4 readers in their assessments of ordinal calcium score groups was 0.819, p-value < 0.001 and a W = 0.849, p-value < 0.001, for CAC score group and TAC score group classification, respectively, reflecting strong agreement between the readers.
Mortality ascertainment
Patient mortality status was extracted from our electronic medical records. In addition to deaths occurred while in the hospital, deaths occurred elsewhere and notified to our institution were also included.
Statistical Analysis
Normality of data was assessed with visual inspection of frequency histograms and probability-probability plots. Continuous data are presented as median (25% – 75% interquartile range). Categorical data are presented as percentages. Characteristics among CAC and TAC groups were compared with ANOVA or Kruskal–Wallis tests for continuous variables and Chi-Square tests for categorical variables, as appropriate. We utilized Kaplan–Meier curves and log-rank tests to estimate the cumulative incidence of all-cause mortality. Patients who were not reported dead on July 22, 2021 were censored. A p-value <0.05 was considered statistically significant. Statistical analysis was performed with Stata (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC).
Multivariable regression model
A multivariable Cox proportional hazards regression model was used to estimate the adjusted HRs of all-cause mortality for the CAC and TAC groups. An initial Cox regression model was built based on demographic and clinical variables that were independently associated with CAC groups and all-cause mortality with a p-value <0.2 on univariate analysis. Additionally, age, gender, and statin use were forced into the model. A backward stepwise approach was utilized to build the reduced regression model. Variables with p<0.05 were retained in the final model. Additionally, we also ran the model with CAC score as a continuous variable, rather than having it categorized into groups. In a similar fashion, we generated a second model utilizing TAC group instead of CAC grouping. The final model included TAC group, age, gender, statin use, LDL-C, and history of hypertension. No significant interactions were found. Proportional hazards assumptions for the variables of interest were graphically assessed.
Results
Baseline characteristics
A total of 811 patients were included in the analysis (Figure 1). Median age was 59 [53 – 68] years, 262 (32.3%) were male and median LDL-C level was 203 [194 – 217] mg/dl. Patients were followed for 6.2 [3.29 – 9.81] years and 109 (13.4%) died during the follow-up period. A total of 703 patients (86.7%) had at least another LDL-C measurement during the follow-up period. The median of the lowest LDL-C achieved during follow-up was 99 [76 – 133] mg/dl with 352 patients (50.1%) reaching a LDL-C level <100 mg/dl and 133 patients (14.7%) reaching a LDL-C level <70 mg/dl. Demographics, lipid panel values, and past medical history stratified by CAC and TAC scores are presented in Table 1 and Table 2, respectively. Our study included a large percentage of usually underrepresented minorities with 36.6% Blacks and 34.8% Hispanics. Most patients (72%) had hypertension, 20.5% were diabetics, and 14.8% had chronic kidney disease (CKD). Most patients (81.6%) were on statin therapy within one year of their LDL-C value. There were no differences between percentages of patients on statin therapy among CAC and TAC severity groups (p=0.93 and 0.47, respectively) or on the values of LDL-C (p=0.14 and 0.09), HDL-C (p=0.98 and 0.24) or triglycerides (TGs, p=0.56 and 0.42). Overall, 376 (46.4%) of patients had CAC score equal to zero and 226 (27.9%) had TAC score equal to zero. One hundred and eighty-three patients (22.6%) had both TAC and CAC scores equal to zero. When different CAC and TAC groups were compared, age and proportion of male patients were higher in the moderate and severe groups (p<0.001 and 0.004, respectively). Prevalence of hypertension, heart failure, dementia, diabetes, and CKD increased as severity of calcium score increased.
Table 1.
Demographics and CAC.
| CAC | ||||||
|---|---|---|---|---|---|---|
| Total | None | Mild | Moderate | Severe | p-value for CAC group comparison | |
| n | 811 | 376 (46) | 265 (33) | 94 (12) | 76 (9) | |
| Age, years | 59 (53 – 68) | 56 (50 – 62) | 62 (56 – 69) | 68 (60 – 75) | 66.5 (57.3 – 74.8) | <0.001 |
| Income, US dollars | 50832 (30977 – 57160) | 48540 (33932 – 57160) | 50832 (31715.8 – 57160) | 50832 (30977 – 57507.8) | 49686 (28921 – 58551) | 0.96 |
| Race/ethnicity | 0.52 | |||||
| White, n (%) | 118 (15) | 44 (37) | 37 (31) | 19 (16) | 18 (15) | |
| Black, n (%) | 297 (37) | 140 (47) | 98 (33) | 32 (11) | 27 (9) | |
| Hispanic, n (%) | 282 (35) | 136 (48) | 93 (33) | 30 (11) | 23 (8) | |
| Asian, n (%) | 10 (1) | 4 (40) | 4 (40) | 1 (10) | 1 (10) | |
| Other, n (%) | 104 (13) | 52 (50) | 33 (32) | 12 (12) | 7 (7) | |
| Male sex, n (%) | 262 (32) | 122 (32) | 68 (26) | 41 (44) | 31 (41) | 0.004 |
| LDL-C, mg/dl | 203 (194 – 217) | 202 (194 – 215.8) | 205 (195 – 218) | 201.5 (195.8 – 215.8) | 206.5 (197 – 220.8) | 0.14 |
| HLD-C, mg/dl | 55±17.1 | 54.6±17.6 | 55.4±16.5 | 55.2±18.2 | 55.5±16.3 | 0.98 |
| TGs, mg/dl | 150 (114 – 210) | 151 (115.3 – 210.8) | 149 (113.5 – 205) | 141 (108.5 – 214.3) | 159.5 (119.5 – 219.5) | 0.56 |
| Total Cholesterol, mg/dl | 294 (278 – 312) | 292 (277 – 312) | 294 (278.5 – 312.5) | 292 (282.5 – 310.3) | 299.5 (284 – 315.5) | 0.32 |
| History of Hypertension, n (%) | 584 (72) | 241 (64) | 195 (74) | 79 (84) | 69 (91) | <0.001 |
| History of Heart Failure, n (%) | 91 (11) | 28 (7) | 33 (13) | 15 (16) | 15 (20) | 0.003 |
| History of Dementia, n (%) | 10 (1) | 2 (1) | 3 (1) | 3 (3) | 2 (3) | 0.073 |
| History of Asthma/COPD, n (%) | 205 (25) | 97 (26) | 56 (21) | 31 (33) | 21 (28) | 0.13 |
| History of Liver disease, n (%) | 53 (7) | 22 (6) | 21 (8) | 7 (7) | 3 (4) | 0.55 |
| History of Diabetes, n (%) | 166 (21) | 53 (14) | 55 (21) | 28 (30) | 30 (40) | <0.001 |
| History of CKD, n (%) | 120 (15) | 43 (11) | 38 (14) | 16 (17) | 23 (30) | <0.001 |
| On Statin, n (%) | 662 (82) | 305 (81) | 215 (81) | 79 (84) | 63 (83) | 0.93 |
| Mortality, n (%) | 109 (13) | 29 (8) | 36 (13.6) | 20 (21) | 24 (32) | <0.001 |
Table shows demographics, comorbidities, pertinent lab values and mortality for the overall population and each CAC severity group. CAC = Coronary artery calcification; CKD = Chronic kidney disease; COPD = Chronic obstructive pulmonary disease; HDL-c = High-density lipoprotein cholesterol; LDL-c = Low-density lipoprotein cholesterol; TGs = Triglycerides
Table 2.
Demographics and TAC
| TAC | ||||||
|---|---|---|---|---|---|---|
| Total | None | Mild | Moderate | Severe | p-value for TAC group comparison | |
| n | 811 | 226 (28) | 389 (48) | 150 (19) | 46 (6) | |
| Age, years | 59 (53 – 68) | 53 (47 – 58) | 60 (55 – 67) | 68.5 (60 – 75) | 71 (65.8 – 79.3) | <0.001 |
| Income, US dollars | 50832 (30977 – 57160) | 49346.5 (33932 – 57160) | 50391 (30977 – 57160) | 41638 (30977 – 56886) | 51620.5 (36244.3 – 58551) | 0.45 |
| Race/ethnicity | 0.005 | |||||
| White, n (%) | 118 (15) | 25 (21) | 49 (42) | 28 (24) | 16 (14) | |
| Black, n (%) | 297 (37) | 91 (31) | 152 (51) | 43 (14) | 11 (4) | |
| Hispanic, n (%) | 282 (35) | 74 (26) | 137 (49) | 55 (20) | 16 (6) | |
| Asian, n (%) | 10 (1) | 2 (20) | 6 (60) | 2 (20) | 0 (0) | |
| Other, n (%) | 104 (13) | 34 (33) | 45 (43) | 22 (21) | 3 (3) | |
| Male sex, n (%) | 262 (32.3) | 95 (42) | 113 (29) | 40 (27) | 14 (30) | 0.003 |
| LDL-C, mg/dl | 203 (194 – 217) | 202 (193 – 216) | 203 (194 – 215) | 205 (197 – 221.3) | 203 (194 – 219.3) | 0.09 |
| HLD-C, mg/dl | 55±17.1 | 53.6±18.4 | 54.6±15.2 | 57.8±19.3 | 57.1±18.1 | 0.24 |
| TGs, mg/dl | 150 (114 – 210) | 152 (113.8 – 216.3) | 146 (115 – 192.5) | 151.5 (108.5 – 222.3) | 165 (121 – 217) | 0.42 |
| Total Cholesterol, mg/dl | 294 (278 – 312) | 294 (277 – 314.3) | 291 (277 – 307) | 301 (286 – 321.3) | 297 (283.8 – 310.3) | <0.001 |
| History of Hypertension, n (%) | 584 (72) | 137 (61) | 282 (73) | 124 (83) | 41 (89) | <0.001 |
| History of Heart Failure, n (%) | 91 (11) | 23 (10) | 38 (10) | 20 (13) | 10 (22) | 0.077 |
| History of Dementia, n (%) | 10 (1) | 0 (0) | 4 (1) | 4 (3) | 2 (4) | 0.015 |
| History of Asthma/COPD, n (%) | 205 (25) | 49 (22) | 104 (27) | 40 (27) | 12 (26) | 0.53 |
| History of Liver disease, n (%) | 53 (7) | 15 (7) | 22 (6) | 11 (7) | 5 (11) | 0.56 |
| History of Diabetes, n (%) | 166 (21) | 38 (17) | 80 (21) | 37 (25) | 11 (24) | 0.28 |
| History of CKD, n (%) | 120 (15) | 26 (12) | 60 (15) | 25 (17) | 9 (20) | 0.35 |
| On Statin, n (%) | 662 (82) | 177 (78) | 324 (83) | 124 (83) | 37 (80) | 0.47 |
| Mortality, n (%) | 109 (13) | 18 (8) | 45 (12) | 32 (21) | 14 (30) | <0.001 |
Table shows demographics, comorbidities, pertinent lab values and mortality for the overall population and each TAC severity group. CKD = Chronic kidney disease; COPD = Chronic obstructive pulmonary disease; HDL-c = High-density lipoprotein cholesterol; LDL-c = Low-density lipoprotein cholesterol; TAC = Thoracic artery calcification; TGs = Triglycerides.
All-cause mortality by CAC groups
Comparison between survivors and non-survivors is presented in Table 3. Patients without CAC (n=376) had an all-cause mortality of 7.7%. Patients who died (n=109) were older, with higher burden of comorbidities. Patient who died had lower use of statin (71.6 vs 83.4%; p<0.001). Higher CAC groups (n=435) were associated with increased mortality (p<0.001). Unadjusted cumulative risk for all-cause mortality per CAC group is shown in the Figure 2. Each increasing level of CAC severity had a higher graded crude all-cause mortality rate (no CAC=10.9 per 1,000-person/year; mild CAC= 21.6 per 1,000-person/year; moderate CAC= 32.8 per 1,000-person/year; and severe CAC= 56.8 per 1,000-person/year). The final Cox regression model was constructed with all-cause mortality adjusted for age, gender, statin use, history of diabetes and CKD based on CAC groups and is presented in Table 4. The patients in mild (HR = 1.7 [1.03 – 2.83]), moderate (HR = 2.13 [1.15 – 3.94]), and severe (HR = 3.49 [1.94 – 6.27]) had increased mortality risk when compared to the patients without coronary artery calcification. When ordinal CAC was entered in the model as a continuous variable, each point increase in CAC was associated with an 11.9% increase in all-cause mortality (HR=1.119 [1.055 – 1.188]).
Table 3.
Patient characteristics according to primary outcome.
| Status | |||
|---|---|---|---|
| Dead | Alive | P-value* | |
| n | 109 | 702 | |
| Age, years | 63 (57 – 73) | 59 (52 – 67) | <0.001 |
| Income, US dollars | 50153 (28921 – 58551) | 50832 (33932 – 57160) | 0.67 |
| Race/ethnicity | 0.11 | ||
| White, n (%) | 18 (17) | 100 (14) | |
| Black, n (%) | 49 (45) | 248 (35) | |
| Hispanic, n (%) | 29 (27) | 253 (36) | |
| Asian, n (%) | 0 (0) | 10 (1) | |
| Other, n (%) | 13 (12) | 91 (13) | |
| Male sex, n (%) | 38 (35) | 224 (32) | 0.35 |
| LDL-C, mg/dl | 207 (196 – 225) | 203 (194 – 215.3) | <0.001 |
| HLD-C, mg/dl | 55.6±19.4 | 54.9±16.8 | 0.98 |
| TGs, mg/dl | 148 (113.5 – 193) | 151 (113.8 – 211) | 0.95 |
| Total Cholesterol, mg/dl | 300 (284 – 317.5) | 293 (278 – 310) | <0.001 |
| History of Hypertension, n (%) | 92 (84) | 492 (70) | 0.002 |
| History of Heart Failure, n (%) | 28 (26) | 63 (9) | <0.001 |
| History of Dementia, n (%) | 4 (4) | 6 (1) | 0.002 |
| History of Asthma/COPD, n (%) | 36 (33) | 169 (24) | 0.046 |
| History of Liver disease, n (%) | 14 (13) | 39 (6) | 0.002 |
| History of Diabetes, n (%) | 43 (39) | 123 (18) | <0.001 |
| History of CKD, n (%) | 37 (34) | 83 (12) | <0.001 |
| On Statin, n (%) | 78 (72) | 584 (83) | <0.001 |
| CAC group | <0.001 | ||
| None, n (%) | 29 (27) | 347 (49) | |
| Mild, n (%) | 36 (33) | 229 (33) | |
| Moderate, n (%) | 20 (18) | 74 (11) | |
| Severe, n (%) | 24 (22) | 52 (7) | |
| TAC group | <0.001 | ||
| None, n (%) | 18 (17) | 208 (30) | |
| Mild, n (%) | 45 (41) | 344 (49) | |
| Moderate, n (%) | 32 (29) | 118 (17) | |
| Severe, n (%) | 14 (13) | 32 (5) | |
Table shows demographics, comorbidities, pertinent lab values and mortality according to all-cause mortality.
Univariable Cox regression.
CAC = Coronary artery calcification; CKD = Chronic kidney disease; COPD = Chronic obstructive pulmonary disease; HDL-c = High-density lipoprotein cholesterol; LDL-C = Low-density lipoprotein cholesterol; TAC = Thoracic artery calcification; TGs = Triglycerides.
Figure 2. Unadjusted cumulative risk for all-cause mortality by CAC (A) and TAC (B) severity groups.
Kaplan-Meier curves show cumulative risk for all-cause mortality stratified by CAC (A) or TAC (B) severity. CAC Log-rank p<0.001. TAC Log-rank p<0.001. CAC = Coronary artery calcification; TAC = Thoracic artery calcification.
Table 4.
Multivariate regression analysis for CAC prediction of all-cause mortality.
| HR | 95% CI | ||
|---|---|---|---|
| Age, per year | 1.017 | 0.999 | 1.034 |
| Male sex | 1.167 | 0.777 | 1.753 |
| Statin therapy | 0.390 | 0.253 | 0.601 |
| Diabetes | 1.753 | 1.138 | 2.699 |
| CKD | 2.264 | 1.461 | 3.509 |
| CAC group | |||
| Mild | 1.711 | 1.033 | 2.834 |
| Moderate | 2.127 | 1.147 | 3.944 |
| Severe | 3.489 | 1.943 | 6.267 |
Table shows multivariable Cox proportional hazards regression model used to estimate the adjusted HRs of all-cause mortality for the CAC TAC groups. CAC = Coronary artery calcification; CKD = Chronic kidney disease.
All-cause mortality by TAC groups
As shown in Table 2, patients without TAC (n=226) had an all-cause mortality of 8.0%. Higher TAC groups (n=196) were associated with increased mortality (p<0.001). Unadjusted cumulative risk for all-cause mortality per TAC group is shown in the Figure 2. When different groups of TAC were compared, crude all-cause mortality rate increased as TAC score progressed (no TAC=11.1 per 1,000-person/year; mild TAC= 17.4 per 1,000-person/year; moderate TAC= 36.7 per 1,000-person/year; and severe TAC= 48.9 per 1,000-person/year). A Cox regression model to estimate HR for TAC groups for all-cause mortality was adjusted for age, gender, statin use, LDL-C level and history of hypertension based on TAC grouping and is presented in Table 5. We found an increased risk of death for moderate (HR = 2.34 [1.19 – 4.59]) and severe (HR = 3.02 [1.36 – 6.74]) TAC groups when compared to patients without TAC. Patients with mild TAC (n=389) had a trend towards higher mortality without reaching statistical significance (HR= 1.44 [0.80 – 2.59]). When entered as continuous variable, each point increase in TAC was associated with a 12.6% increase in all-cause mortality (HR=1.126 [1.051–1.206]).
Table 5.
Multivariate regression analysis for TAC prediction of all-cause mortality.
| HR | 95% CI | ||
|---|---|---|---|
| Age, per year | 1.016 | 0.997 | 1.036 |
| Male sex | 1.343 | 0.891 | 2.023 |
| Statin therapy | 0.441 | 0.285 | 0.682 |
| LDL-C, per mg/dl | 1.012 | 1.008 | 1.017 |
| Hypertension | 2.225 | 1.283 | 3.859 |
| TAC group | |||
| Mild | 1.440 | 0.801 | 2.589 |
| Moderate | 2.336 | 1.189 | 4.591 |
| Severe | 3.023 | 1.356 | 6.739 |
Table shows multivariable Cox proportional hazards regression model used to estimate the adjusted HRs of all-cause mortality for the TAC groups. LDL-C = Low-density lipoprotein cholesterol; TAC = Thoracic artery calcification.
Combination of CAC and TAC on risk estimation
As presented in Figure 3A, survival of patients with CAC= 0 differed by their TAC score (P-value log rank = 0.002). Using previously described Cox regression models, patients with CAC=0 but moderate or severe TAC (n=37) had significantly higher mortality (HR = 5.78 [1.66 – 20.11]) when compared to those with CAC/TAC=0 (n=183). For those with CAC=0 but mild TAC (n=156) there was a trend towards higher mortality (HR = 2.47 [0.91 – 6.60]).
Figure 3. Unadjusted cumulative risk for all-cause mortality by TAC severity for patients with CAC=0(A) and CAC severity for patients with TAC=0 (B).
Kaplan-Meier curves show cumulative risk for all-cause mortality stratified by TAC severity for patients with CAC=0(A) or CAC severity for patients with TAC=0 (B) severity. (A) Log-rank p=0.002. (B) Log-rank p<0.001. CAC = Coronary artery calcification; TAC = Thoracic artery calcification.
Similarly, survival of patients with TAC= 0 differed by their CAC score (Figure 3B, P-value log rank <0.001). Patients with TAC = 0 but mild or moderate/severe CAC had significantly higher mortality (HR = 4.42 [1.45 – 13.42]) for mild (n=30) and (HR = 6.35 [1.64 – 24.55]) for moderate/severe (n=8) when compared to those with TAC/CAC=0 (n=183).
Discussion
To the best of our knowledge, this is the first study to show CAC and TAC as independent predictors of all-cause mortality in patients with LDL-C≥190 mg/dL, specifically using incidentally detected CAC and TAC from routine chest CTs. In particular, mild, moderate or severe CAC were associated with a doubled risk of all-cause mortality when compared to those without CAC. In addition, TAC was associated with increased mortality in moderate and severe groups.
Patients with LDL-C≥190 mg/dL are labeled as high-risk by current guidelines without needing further risk stratification with the recommendation of starting high-intensity statin therapy for adults and children> 8 years of age3. Our findings suggest a potential role for further risk stratification with CAC and TAC. Prior work by Sandesara et al. from the MESA cohort10, showed in 246 individuals with LDL-C≥190 mg/dL that the absence of CAC was associated with a lower risk of cardiovascular events (10-year risk 3.7% vs 20%; p=0.005). In their study, however, mortality prediction by CAC showed a trend without reaching statistical significance (10-year risk 4.5% vs 14%; p=0.053) likely due to the smaller size of their cohort and their lower overall mortality rates. In this cohort, CAC was absent in 37% of patients and in our study in 46%. As opposed to data from MESA with 43% not on statin therapy, in our cohort 18% were not on statin therapy.
A systematic review found 9 studies with a total of 1,176 participants, up to January 2019, evaluating CAC for patients with FH12. Prevalence of CAC=0 was 45% with a mean age of 47 years and LDL-C level of 158 mg/dL (135–205 mg/dL) with 97% of participants on statin at follow-up. Importantly, these estimates were derived from specialized registries, possibly not fully representing the true prevalence in FH. Recently, Gallo et al. studied 1,624 patients with a median follow-up of 2.7 years from two registries (REFERCHOL and SAFEHEART)13. In this cohort (mean age 48 years, 82.9% on statin therapy), CAC >100 was associated with a HR 32.05 (95% CI:10.08–101.94) of developing ASCVD when compared to CAC=0. Moreover, the addition of CAC to SAFEHEART-RE significantly improved prediction of ASCVD (AUC 0.884 vs 0.793; p<0.001) with a net reclassification improvement of 45.4%. Similar to their findings, CAC was 0 in 46.4% of the patients in our study.
Patients with the FH phenotype represented as LDL-C≥190 mg/dL have a 5–6-fold higher ASCVD risk than those with normal cholesterol1,2. In particular, LDL-C≥190 mg/dL in adults is used as an initial marker to consider FH17. It is known that in around 3% of these cases with severe hypercholesterolemia, the disease is caused by monogenic defects that impair the removal of plasma LDL that are characteristic of FH and patients with positive mutations have higher risk than those without (up to 24-fold), likely due to higher lifetime exposure to high cholesterol levels1. However, even in severe disease like FH, the risk of ASCVD is heterogeneous and depends not only on the gradient of blood cholesterol concentration but also on other factors such as age, lower HDL-C and smoking18. Moreover, the family history needed for common FH clinical diagnostic tools is many times unknown and genetic testing is frequently not available. Many individuals may not develop ASCVD despite the presence of severe hypercholesterolemia. Our paper adds to the growing evidence that imaging atherosclerosis through CAC and TAC possibly represents an easy risk stratification tool with potential widespread applicability. Presence of CAC and/or TAC may help identify individuals at higher risk that could benefit from more intensive but higher cost therapies such as ezetimibe, PCSK9 inhibitors, icosapent ethyl, bempedoic acid and inclisiran19. Importantly, median LDL-C achieved on follow-up was <100 mg/dl in only 50% and <70 mg/dl in 15% of the patients, showing a large opportunity for further therapy intensification and potential ASCVD/mortality prevention. In addition, CAC/TAC could be helpful for patients to encourage starting lipid lowering therapy and compliance20.
Thoracic and abdominal aortic calcification has recently been gaining further interest as a risk stratification tool beyond CAC8,9,21. In the study by Han et al. using data from the CAC consortium, a TAC>300 improved risk prediction for cardiovascular mortality when added to the ASCVD risk score and CAC8. Interestingly, authors found the highest reclassification for cardiovascular death in the high ASCVD risk. Patients with LDL-C≥190 mg/dL and diabetes were included in the high-risk group, however lipid data was not available for all patients and they included only 7% diabetics and only 11% non-White population. In our study, TAC provided significant risk estimation for all-cause mortality in patients with LDL-C≥190 mg/dL. Additionally, the study by Santos et al. showed that TAC predicted all-cause mortality, independently of risk factors and CAC9.
The importance of consistent reporting and standardization of CAC evaluation in non-contrast non-cardiac chest CT scans is highlighted in the creation of the Society of Cardiac Computed Tomography (SCCT) CAC-DRS reporting system consensus document22 and the Class I recommendation for interpreting CAC on all chest CTs23. Our results confirm the importance of this assessment even in patients with LDL-C≥190 mg/dL. Furthermore, ordinal calcium scoring is easy to learn, reproducible and does not require additional software to be calculated. In this way its use could be widely applied across different healthcare and socioeconomical settings. It is important to remark however that even that the overall burden of coronary calcification is identified in non-gated scans, it is possible that a small amount of coronary calcification can be missed, emphasizing the role of dedicated gated coronary calcium scoring scans. Moreover, it is possible that in asymptomatic high-risk patients, CT coronary angiography may provide extra value over coronary calcium scoring24. Further studies are needed to explore this concept.
Lastly, current recommendations indicate that patients with LDL-C≥190 mg/dL should be started on high-intensity statin even if CAC/TAC=0. Lifetime exposure to high levels of blood cholesterol, safety and low cost of statins together with the possibility of atherosclerosis in non-imaged vasculature support this concept in particular in the young and middle aged. Further studies on this particular group with LDL-C≥190 mg/dL but CAC/TAC=0 are needed to further characterize this cohort.
Our study has several limitations. First, the retrospective and observational design of our study may be prone to potential unmeasured confounding factors such as duration of hyperlipidemia and duration and intensity of statin therapy. It is known that statin therapy can increase the amount of calcified plaque and CAC score. As described by Osei et al.25 using data from the CAC Consortium, in statin users, both the Agatston and CAC volume have prognostic utility for CHD and CVD risk, suggesting that CAC also predicts CHD and CVD risk in this group. The improvement in predictive value compared to risk factor models alone was similar in statin and non-statin groups in Osei et al. study. Second, we retrospectively studied patients with a clinical indication for a chest CT without contrast, preselecting a higher risk group with inherent referral bias, enriched in comorbidities. However, despite selecting a group at high risk of all-cause mortality, the incidence of CAC was similar to that reported by other cohorts10,12. Third, we used the ordinal coronary calcification score for quantification of aortic calcification although this score was designed for the coronary arteries. Even that this concept also applies to the Agatston score, Han et al. also found applicability for aortic calcifications8. Lastly, smoking history/status was not available.
Conclusion
Patients with LDL-C≥190 mg/dL represent a heterogeneous group. CAC and TAC, even when detected incidentally on routine chest CT, are independent predictors for all-cause mortality.
Clinical Perspective.
Current guidelines recommend measuring CAC scoring for stratification of atherosclerotic cardiovascular disease (ASCVD) risk only in patients with borderline to intermediate risk score by the pooled cohort equation with low density lipoprotein cholesterol (LDL-C) of 70–190 mg/dL. It remains unknown if CAC or TAC, detected on routine chest CT, can provide further risk stratification in patients with LDL-C≥190 mg/dL. From a multi-site medical center, we retrospectively identified all patients with age≥ 40 years, without established ASCVD and LDL-C≥190 mg/dL who had non-gated non-contrast chest CT within 3 years of LDL-C measurement. To the best of our knowledge, this is the first study to show CAC and TAC as independent predictors of all-cause mortality in patients with LDL-C≥190 mg/dL, specifically using incidentally detected CAC and TAC from routine chest CTs. In particular, mild, moderate or severe CAC were associated with a doubled risk of all-cause mortality when compared to those without CAC. Similarly, moderate and severe TAC were associated with increased mortality.. Moreover, the analysis of TAC in patients with CAC=0 and CAC in patients with TAC=0 provided further risk stratification showing the importance of reporting both. Many individuals may not develop ASCVD despite the presence of severe hypercholesterolemia. Our paper adds to the growing evidence that imaging atherosclerosis through CAC and TAC possibly represents an easy risk stratification tool with potential widespread applicability. It also highlights the importance of consistent reporting and standardization of CAC/TAC evaluation in non-contrast non-cardiac chest CT scans. Presence of CAC and/or TAC may help identify individuals at higher risk that could benefit from more intensive but higher cost therapies.
Acknowledgments
Disclosures:
Francesco Castagna is supported by a grant from the National Institute for Health (NIH) (T32HL144456) and the National Center for Advancing Translational Science (NCATS) Clinical and Translational Science Award at Einstein-Montefiore (UL1TR001073). Lili Zhang is supported by a grant from the New York Academy of Medicine. Dr Blaha has received Grants from NIH, FDA, AHA, Amgen, Novo Nordisk, Bayer; participated on Advisory Boards for Amgen, Novartis, Novo Nordisk, Bayer, Roche, 89Bio, Kaleido, Inozyme and works as a Consultant for Kowa and emocha. Carlos J. Rodriguez is supported by grants from the NIH (R01 HL04199, 75N92019D00011, 1U01HL146204-01, 5R01HL144707) and the American Heart Association (5P50HL120163-04) and has participated on Advisory Boards for Amgen and has worked as a Consultant for Merck. Leandro Slipczuk has worked as a Consultant for Amgen and participated on the Advisory Boards for Esperion and Regeneron.
Non-standard Abbreviations and Acronyms.
- ASCVD
atherosclerotic cardiovascular disease
- CAC
Coronary artery calcification
- CKD
Chronic kidney disease
- COPD
Chronic obstructive pulmonary disease
- CT
Computerized tomography
- HDL-C
High-density lipoprotein cholesterol
- LDL-C
Low-density lipoprotein cholesterol
- TAC
Thoracic artery calcification
- TGs
Triglycerides
Footnotes
Presented in part at the American College of Cardiology Scientific Sessions in Washington DC, April 2 – 4, 2022.
References
- 1.Khera A v., Won HH, Peloso GM, Lawson KS, Bartz TM, Deng X, van Leeuwen EM, Natarajan P, Emdin CA, Bick AG et al. Diagnostic Yield and Clinical Utility of Sequencing Familial Hypercholesterolemia Genes in Patients With Severe Hypercholesterolemia. J Am Coll Cardiol. 2016;67:2578–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Perak AM, Ning H, de Ferranti SD, Gooding HC, Wilkins JT, Lloyd-Jones DM. Long-term risk of atherosclerotic cardiovascular disease in US adults with the familial hypercholesterolemia phenotype. Circulation. 2016;134:9–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Grundy SM, Stone NJ, Chair V, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, Braun LT, de Ferranti S, Faiella-Tommasino J et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/ APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol: Executive Summary Circulation. Circulation. 2019;139:e1082–e1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Orimoloye OA, Budoff MJ, Dardari ZA, Mirbolouk M, Iftekhar Uddin SM, Berman DS, Rozanski A, Shaw LJ, Rumberger JA, Nasir K et al. Race/ethnicity and the prognostic implications of coronary artery calcium for all-cause and cardiovascular disease mortality: The coronary artery calcium consortium. J Am Heart Assoc. 2018;7:e010471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shemesh J, Henschke CI, Shaham D, Yip R, Farooqi AO, Cham MD, McCauley DI, Chen M, Smith JP, Libby DM et al. Ordinal scoring of coronary artery calcifications on low-dose CT scans of the chest is predictive of death from cardiovascular disease. Radiology. 2010; 257:541–8. [DOI] [PubMed] [Google Scholar]
- 6.Hughes-Austin JM, Dominguez A, Allison MA, Wassel CL, Rifkin DE, Morgan CG, Daniels MR, Ikram U, Knox JB, Wright CM et al. Relationship of Coronary Calcium on Standard Chest CT Scans With Mortality. JACC Cardiovasc Imaging. 2016;9:152–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Slipczuk L, Castagna F, Schonberger A, Novogrodsky E, Sekerak R, Dey D, Jorde UP, Levsky JM, Garcia MJ. Coronary artery calcification and epicardial adipose tissue as independent predictors of mortality in COVID-19. Int J Cardiovasc Imaging. 2021;37:3093–3100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Han D, Kuronuma K, Rozanski A, Budoff MJ, Miedema MD, Nasir K, Shaw LJ, Rumberger JA, Gransar H, Blumenthal RS et al. Implication of thoracic aortic calcification over coronary calcium score regarding the 2018 ACC/AHA Multisociety cholesterol guideline: results from the CAC Consortium. Am J Prev Cariol. 2021;8:100232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Santos RD, Rumberger JA, Budoff MJ, Shaw LJ, Orakzai SH, Berman D, Raggi P, Blumenthal RS, Nasir K. Thoracic aorta calcification detected by electron beam tomography predicts all-cause mortality. Atherosclerosis. 2010;209:131–5. [DOI] [PubMed] [Google Scholar]
- 10.Sandesara PB, Mehta A, O’Neal WT, Kelli HM, Sathiyakumar V, Martin SS, Blaha MJ, Blumenthal RS, Sperling LS. Clinical significance of zero coronary artery calcium in individuals with LDL cholesterol ≥190 mg/dL: The Multi-Ethnic Study of Atherosclerosis. Atherosclerosis. 2020;292:224–229. [DOI] [PubMed] [Google Scholar]
- 11.Miname MH, Bittencourt MS, Moraes SR, Alves RIM, Silva PRS, Jannes CE, Pereira AC, Krieger JE, Nasir K, Santos RD. Coronary Artery Calcium and Cardiovascular Events in Patients With Familial Hypercholesterolemia Receiving Standard Lipid-Lowering Therapy. JACC Cardiovasc Imaging. 2019;12:1797–1804. [DOI] [PubMed] [Google Scholar]
- 12.Mszar R, Grandhi GR, Valero-Elizondo J, Virani SS, Blankstein R, Blaha M, Mata P, Miname MH, al Rasadi K, Krumholz HM et al. Absence of Coronary Artery Calcification in Middle-Aged Familial Hypercholesterolemia Patients Without Atherosclerotic Cardiovascular Disease. JACC Cardiovasc Imaging. 2020;13:1090–1092. [DOI] [PubMed] [Google Scholar]
- 13.Gallo A, Pérez de Isla L, Charrière S, Vimont A, Alonso R, Muñiz-Grijalvo O, Díaz-Díaz JL, Zambón D, Moulin P, Bruckert E et al. The Added Value of Coronary Calcium Score in Predicting Cardiovascular Events in Familial Hypercholesterolemia. JACC Cardiovasc Imaging. 2021;14:2414–2424. [DOI] [PubMed] [Google Scholar]
- 14.Mortensen MB, Caínzos-Achirica M, Steffensen FH, Bøtker HE, Jensen JM, Sand NPR, Maeng M, Bruun JM, Blaha MJ, Sørensen HT et al. Association of Coronary Plaque With Low-Density Lipoprotein Cholesterol Levels and Rates of Cardiovascular Disease Events Among Symptomatic Adults. JAMA Netw Open. 2022;5:e2148139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dzaye O, Dardari ZA, Cainzos-Achirica M, Blankstein R, Agatston AS, Duebgen M, Yeboah J, Szklo M, Budoff MJ, Lima JAC et al. Warranty Period of a Calcium Score of Zero: Comprehensive Analysis From MESA. JACC Cardiovasc Imaging. 2021;14:990–1002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bassler JF, Marascuilo LA, McSweeney M. Nonparametric and Distribution-Free Methods for the Social Sciences. J Am Stat Assoc. 1978;73. [Google Scholar]
- 17.McGowan MP, Hosseini Dehkordi SH, Moriarty PM, Duell PB. Diagnosis and treatment of heterozygous familial hypercholesterolemia. J Am Heart Assoc. 2019;8:e013225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Santos RD, Gidding SS, Hegele RA, Cuchel MA, Barter PJ, Watts GF, Baum SJ, Catapano AL, Chapman MJ, Defesche JC et al. Defining severe familial hypercholesterolaemia and the implications for clinical management: a consensus statement from the International Atherosclerosis Society Severe Familial Hypercholesterolemia Panel. Lancet Diabetes Endocrinol. 2016;4:850–61. [DOI] [PubMed] [Google Scholar]
- 19.Mszar R, Nasir K, Santos RD. Coronary Artery Calcification in Familial Hypercholesterolemia: An Opportunity for Risk Assessment and Shared Decision Making with the Power of Zero? Circulation. 2020;142:1402–1407. [DOI] [PubMed] [Google Scholar]
- 20.Rozanski A, Gransar H, Shaw LJ, Kim J, Miranda-Peats L, Wong ND, Rana JS, Orakzai R, Hayes SW, Friedman JD et al. Impact of coronary artery calcium scanning on coronary risk factors and downstream testing: The EISNER (Early Identification of Subclinical Atherosclerosis by Noninvasive Imaging Research) prospective randomized trial. J Am Coll Cardiol. 2011;57:1622–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jurgens P, Carr JJ, Terry J, Rana JS, Jacobs DR, Duprez D. ASSOCIATION OF CORONARY ARTERY CALCIUM AND ABDOMINAL AORTA CALCIUM WITH INCIDENT FATAL AND NON-FATAL CARDIOVASCULAR DISEASE EVENTS: THE CORONARY ARTERY RISK DEVELOPMENT IN YOUNG ADULTS STUDY. J Am Coll Cardiol. 2021;10:e023037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hecht HS, Blaha MJ, Kazerooni EA, Cury RC, Budoff M, Leipsic J, Shaw L. CAC-DRS: Coronary Artery Calcium Data and Reporting System. An expert consensus document of the Society of Cardiovascular Computed Tomography (SCCT). J Cardiovasc Comput Tomogr. 2018;12:185–191. [DOI] [PubMed] [Google Scholar]
- 23.Hecht HS, Cronin P, Blaha MJ, Budoff MJ, Kazerooni EA, Narula J, Yankelevitz D, Abbara S. 2016 SCCT/STR guidelines for coronary artery calcium scoring of noncontrast noncardiac chest CT scans: A report of the Society of Cardiovascular Computed Tomography and Society of Thoracic Radiology. J Cardiovasc Comput Tomogr. 2017;11:74–84. [DOI] [PubMed] [Google Scholar]
- 24.Min JK, Labounty TM, Gomez MJ, Achenbach S, Al-Mallah M, Budoff MJ, Cademartiri F, Callister TQ, Chang HJ, Cheng V et al. Incremental prognostic value of coronary computed tomographic angiography over coronary artery calcium score for risk prediction of major adverse cardiac events in asymptomatic diabetic individuals. Atherosclerosis. 2014;232:298–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Osei AD, Mirbolouk M, Berman D, Budoff MJ, Miedema MD, Rozanski A, Rumberger JA, Shaw L, al Rifai M, Dzaye O et al. Prognostic value of coronary artery calcium score, area, and density among individuals on statin therapy vs. non-users: The coronary artery calcium consortium. Atherosclerosis. 2021;316:79–83. [DOI] [PMC free article] [PubMed] [Google Scholar]