Influence of Subclinical Atherosclerosis Burden and Progression on Mortality
Original Research
Abstract
Background
Atherosclerosis is a dynamic process. There is little evidence regarding whether quantification of atherosclerosis extent and progression, particularly in the carotid artery, in asymptomatic individuals predicts all-cause mortality.
Objectives
This study sought to evaluate the independent predictive value (beyond cardiovascular risk factors) of subclinical atherosclerosis burden and progression and all-cause mortality.
Methods
A population of 5,716 asymptomatic U.S. adults (mean age 68.9 years, 56.7% female) enrolled between 2008 and 2009 in the BioImage (A Clinical Study of Burden of Atherosclerotic Disease in an At Risk Population) study underwent examination by vascular ultrasound to quantify carotid plaque burden (cPB) (the sum of right and left carotid plaque areas) and by computed tomography for coronary artery calcium (CAC). Follow-up carotid vascular ultrasound was performed on 732 participants a median of 8.9 years after the baseline exam. All participants were followed up for all-cause mortality, the primary outcome. Trend HRs are the per-tertile increase in each variable.
Results
Over a median 12.4 years’ follow-up, 901 (16%) participants died. After adjustment for cardiovascular risk factors and background medication, baseline cPB and CAC score were both significantly associated with all-cause mortality (fully adjusted trend HR: 1.23; 95% CI: 1.16-1.32; and HR: 1.15; 95% CI: 1.08-1.23), respectively (both P < 0.001), thus providing additional prognostic value. cPB performed better than CAC score. In participants with a second vascular ultrasound evaluation, median cPB progressed from 29.2 to 91.3 mm3. cPB progression was significantly associated with all-cause mortality after adjusting for cardiovascular risk factors and baseline cPB (HR: 1.03; 95% CI: 1.01-1.04 per absolute 10-mm3 change; P = 0.01).
Conclusions
Subclinical atherosclerosis burden (cPB and CAC) in asymptomatic individuals was independently associated with all-cause mortality. Moreover, atherosclerosis progression was independently associated with all-cause mortality.
Introduction
Cardiovascular disease (CVD) is the leading cause of death worldwide,1 and the clinical manifestations of the disease are usually preceded by many years by the underlying process of atherosclerosis. Current tools for cardiovascular risk prediction like the Framingham risk score (FRS) are based on conventional cardiovascular risk factors (CVRFs)2; however, while these equations work at the population level, their accuracy at the individual level is limited, especially for long-term risk prediction.3
The visualization and quantification of subclinical atherosclerosis by noninvasive vascular imaging is gaining an expanding role in cardiovascular risk assessment.4 Previous studies have revealed that atherosclerosis is very frequent even in apparently healthy middle-aged populations5,6; moreover, CVD risk prediction based on CVRFs is improved by additionally considering the presence of subclinical atherosclerosis in the carotid or femoral arteries assessed by vascular ultrasound (VUS) or the presence of coronary artery calcium (CAC) detected by noncontrast computed tomography (CT).7-10 Risk classification can be further improved by considering not merely the presence of subclinical atherosclerosis but its extent (atherosclerotic burden), quantified either as the CAC score or as VUS-detected carotid plaque burden (cPB), both of which correlate with long-term CV events.4,11 Atherosclerosis is a dynamic process, and plaque burden can remain stable, progress, or even regress over time, and plaque progression has been shown to increase substantially with the increasing presence of CVRFs.12 These changes in atherosclerosis burden can modify the predictive capacity of cross-sectional imaging. It is therefore plausible that the risk-predictive capacity of atherosclerosis imaging could be further improved by quantifying atherosclerosis progression. However, this possibility in relation to the cPB has not been tested before.
The CAC score has been extensively used to improve risk prediction over CVRFs alone4,13; however, it is less than optimal for detecting early atherosclerosis or for longitudinal evaluations. Plaque calcification occurs at late stages of plaque formation,14 and CAC analysis thus will not detect early, noncalcified plaques. Moreover, although CAC progression has been recently shown to correlate with all-cause mortality,15 it may reflect plaque stabilization rather than disease progression, thus potentially reducing prognostic value, especially in patients who have initiated lipid-lowering therapy.14,16 VUS examination of easily accessible peripheral arteries (commonly the carotids and femorals) has theoretical advantages over CAC analysis because it can identify early noncalcified plaques and monitor their progression.5,12,17,18 While some studies have demonstrated that VUS-detected atherosclerotic burden in peripheral arteries provides incremental predictive value for ischemic events over CVRFs alone,4,19 and others have shown that the percentage of carotid stenosis by VUS predicts mortality,20,21 there is a paucity of data regarding the independent predictive value of VUS-detected subclinical cPB and its progression and all-cause mortality.
The aim of this study was to evaluate the association with all-cause mortality of atherosclerotic burden (cPB by VUS and CAC by noncontrast CT) and cPB progression in a population of asymptomatic individuals.
Study design and participants
The BioImage (A Clinical Study of Burden of Atherosclerotic Disease in an At Risk Population; NCT00738725) study is a prospective study that, between January 2008 and June 2009, enrolled 7,687 asymptomatic members of the Humana Health System (men 55-80 years of age and women 60-80 years of age) residing in the Chicago, Illinois, or Fort Lauderdale, Florida, metropolitan areas. Methodological aspects have been described in detail.22 Eligibility criteria included freedom from previous CVD (myocardial infarction, stroke, angina, heart failure, or arterial revascularization) and from active cancer treatment or any medical condition precluding long-term participation. A total of 6,102 participants entered the imaging arm of the study and were assessed by carotid VUS and CAC scoring. Of these participants, 732 had a second carotid VUS performed a median of 8.9 years (range: 8.4-9.7 years) later to assess cPB progression. Participants were followed until October 2021. Deaths were identified from Social Security and National Death Index searches.
The BioImage study was approved by the Institutional Review Board, and all participants provided written informed consent and Health Insurance Portability and Accountability Act authorization.
Baseline examinations
All participants had an in-person baseline examination and interview. Diabetes mellitus was defined as current medication with oral hypoglycemic agents or insulin or self-reported diagnosis. Hypertension was defined as systolic blood pressure (SBP) ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or current use of antihypertensive medication. Current smoking status was self-reported. A nonfasting venous blood sample was processed for routine biochemistry tests, including determination of plasma lipid levels.
Carotid VUS
Plaque burden in both carotid arteries was assessed at baseline using a high-resolution, 2-dimensional L9-3 linear array transducer using a Philips iU22 ultrasound system (Philips Healthcare), with scanning in longitudinal and cross-sectional orientations from the proximal common carotid artery to the distal internal carotid artery on each side. Ultrasound scans were read by a core laboratory at the Department of Vascular Surgery, Rigshospitalet, University of Copenhagen, Denmark. cPB was obtained as the sum of plaque areas from all images in the cross-sectional sweeps from the right and left carotid arteries, providing a quantitative metric of total plaque area (mm2) along the length of the visualized bilateral carotids using Philips QLAB quantification software (QLAB-VPQ, version 13, Philips). Interobserver agreement has been previously described with an intraclass correlation coefficient of 0.823.23 In addition, from December 2008, all consecutively enrolled participants were also examined by carotid VUS using a real-time VL13-5 3D linear array transducer (Philips Healthcare), ensuring that the third dimension is acquired in a standardized way over a period of about 3 seconds, and total plaque volume (mm3) was calculated as the sum of plaque volumes in the same longitudinal segments from both carotids, as previously described.22,24 Participants undergoing follow-up carotid VUS were examined with the 3-dimensional transducer, and plaque progression was defined as any increase in cPB from baseline to the follow-up VUS scan. Absence of plaque progression was defined as a decrease in cPB from baseline to follow-up (plaque regression) or cPB = 0 at both baseline and follow-up (no carotid atherosclerosis). Baseline and follow-up 3-dimensional images of participants from the progression cohort were analyzed in a blinded fashion at Centro Nacional de Investigaciones Cardiovasculares (Madrid, Spain) using the Carotid Model CM2020 software version 2.1 (Philips Research). Interobserver variability was assessed in the first 56 cases by 2 independent readers (1 from each site) with an intraclass correlation agreement of 0.92 (95% CI: 0.86-0.96). Intraobserver concordance has been previously reported, with an intraclass correlation coefficient of 0.998 (95% CI: 0.996-0.999).24
CAC quantification
CAC was measured by the Agatston method from noncontrast multidetector CT scans of the coronary arteries obtained using a Philips Brilliance 64-slice CT scanner (Philips Healthcare) with prospectively electrocardiographically gated acquisition. All operators and core laboratory readers were blinded to clinical data and results from other imaging modalities.
Statistical analysis
Baseline characteristics were expressed as mean ± SD for continuous variables or median (Q1-Q3) if marked skewness existed, and as number and percentage for categorical variables. For each imaging modality, participants were classified as either having no atherosclerosis or by tertile of increasing CAC score or cPB.
Associations between baseline characteristics and cPB or CAC score tertiles were analyzed using trend tests across groups. Predictors of extent of plaque progression were assessed with a multivariate linear regression analysis.
All-cause mortality was plotted against cPB and CAC score tertiles using the Kaplan-Meier method, and mortality rates were compared among groups by the log-rank test. Associations between cPB or CAC score (categorized as tertiles or as continuous log-transformed variables) and all-cause mortality were assessed by Cox proportional hazards regression. Trend HRs were calculated as per tertile increase in cPB or CAC score. Potential confounders were adjusted for in 2 models: model 1, including age, sex, and race; and model 2, including age, sex, race, diabetes mellitus, current smoking, body mass index, SBP, antihypertensive medication, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol, and use of lipid-lowering or hypertensive drugs. To evaluate the incremental risk predictive value of CAC score or cPB relative to model 1 and model 2, we assessed each model for overall fit (likelihood ratio test), discrimination (Harrell’s C-index), and calibration (Hosmer-Lemeshow test). To examine the impact of cPB progression on all-cause mortality, participants in the subcohort with follow-up VUS data were classified as showing no progression between baseline and follow-up (no atherosclerosis if cPB = 0 at both time points, or regression when cPB decreased from the first to the second examination) or progression if cPB increased. All-cause mortality rates were plotted using the Kaplan-Meier method and compared between groups by the log-rank test. Cox proportional hazards analysis was used to assess the association between progression as a continuous variable and all-cause mortality, with adjustment for CVRFs at the time of the second VUS and baseline (continuous log-transformed) cPB. All study participants were followed until time of death or close of study (October 2021), whichever came first. All analyses were performed with Stata version 17 (StataCorp).
Results
A total of 7,687 participants were enrolled, of whom 6,102 underwent imaging. Of these, 386 were excluded due to missing data on carotid VUS, CAC imaging, or any of the variables required to calculate the FRS, yielding a final population of 5,716 study participants, 732 of whom underwent a second carotid VUS examination a median 8.9 years later to assess cPB progression. Study participant flow is depicted in Figure 1. Participants in the excluded population had a higher cardiovascular risk due to a higher percentage of males and diabetes and slightly higher body mass index, total cholesterol, and LDL-C (Supplemental Table 1).
Prevalence of subclinical atherosclerosis at baseline
The mean participant age was 68.9 years, and 56.7% were female; baseline characteristics are summarized in Supplemental Table 1. At baseline, 87.7% of participants had subclinical atherosclerosis; 57.6% had multiterritorial disease (some cPB in 1 or both carotids and CAC score>0), 19.7% had disease only in the carotids, and 10.5% had only CAC (Figure 2). When stratified by FRS groups, the prevalence of multiterritorial atherosclerosis was 47% in the low-risk group, increasing to 69% in the intermediate-risk group and 81% in the high-risk group. While the prevalence of isolated CAC remained similar in all risk groups (approximately 10%), isolated carotid disease was significantly more prevalent in the low-risk group (25%) and decreased with increasing FRS. To better evaluate the factors associated with subclinical atherosclerosis at baseline, the population was divided into tertiles of cPB and CAC score (Tables 1 and 2). Increasing tertiles of cPB and CAC score were associated with older age; male sex; White race; and CVRFs including smoking, diabetes, higher SBP, higher FRS, and lipid-lowering and antihypertensive therapy. The distributions of baseline cPB and CAC score are shown in Supplemental Figure 1.
Vascular Ultrasound–Carotid Plaque Burden (mm2) | P Value (Trend) | ||||
---|---|---|---|---|---|
No Disease (n = 1,305) | Tertile 1 (n = 1,471) | Tertile 2 (n = 1,471) | Tertile 3 (n = 1,469) | ||
Age, y | 67.4 ± 5.7 | 68.2 ± 5.9 | 69.4 ± 6.0 | 70.3 ± 5.9 | <0.001 |
Sex | <0.001 | ||||
Female | 870 (66.7) | 955 (64.9) | 814 (55.3) | 602 (41.0) | |
Male | 435 (33.3) | 516 (35.1) | 657 (44.7) | 867 (59.0) | |
White | 834 (63.9) | 1,090 (74.1) | 1153 (78.4) | 1156 (78.7) | <0.001 |
Diabetes mellitus | 184 (14.1) | 203 (13.8) | 220 (15.0) | 263 (17.9) | 0.003 |
Current smoker | 54 (4.1) | 107 (7.3) | 112 (7.6) | 210 (14.3) | <0.001 |
BMI, kg/m2 | 29.5 ± 5.8 | 29.1 ± 5.6 | 29.1 ± 5.6 | 28.5 ± 5.1 | <0.001 |
LDL cholesterol, mg/dL | 114.1 ± 32.5 | 114.4 ± 32.7 | 114.8 ± 33.1 | 113.6 ± 34.5 | 0.75 |
HDL cholesterol, mg/dL | 58.1 ± 15.2 | 57.3 ± 15.1 | 55.4 ± 14.9 | 53.7 ± 15.0 | <0.001 |
Total cholesterol, mg/dL | 202.4 ± 37.6 | 203.4 ± 38.1 | 202.5 ± 37.0 | 199.4 ± 38.5 | 0.02 |
Triglycerides, mg/dL | 151.3 ± 71.4 | 158.5 ± 72.9 | 161.7 ± 71.4 | 160.7 ± 74.2 | <0.001 |
SBP, mm Hg | 136.7 ± 18.2 | 137.7 ± 17.8 | 140.2 ± 19.0 | 142.8 ± 18.6 | <0.001 |
DBP, mm Hg | 79.2 ± 9.4 | 78.1 ± 8.7 | 77.9 ± 9.0 | 77.6 ± 9.1 | <0.001 |
Lipid-lowering therapy | 300 (23.0) | 427 (29.0) | 458 (31.1) | 480/1,469 (32.7) | <0.001 |
Antihypertensive therapy | 291 (22.3) | 340 (23.1) | 428 (29.1) | 494/1,469 (33.6) | <0.001 |
Framingham 10-y risk, % | 5.8 (3.0-10.6) | 6.4 (3.5-11.9) | 9.1 (4.4-15.9) | 12.4 (7.1-18.5) | <0.001 |
Framingham 10-y risk group | <0.001 | ||||
Low risk (≤10%) | 952 (73.0) | 1,003 (68.2) | 799 (54.3) | 558 (38.0) | |
Intermediate risk (10%-20%) | 294 (22.5) | 365 (24.8) | 483 (32.8) | 613 (41.7) | |
High risk (>20%) | 59 (4.5) | 103 (7.0) | 189 (12.8) | 298 (20.3) | |
Deaths | 122 (9.3) | 174 (11.8) | 250 (17.0) | 355 (24.2) | <0.001 |
Mean follow-up time, y | 12.3 | 12.1 | 11.7 | 11.3 |
CAC Score | P Value (Trend) | ||||
---|---|---|---|---|---|
No Disease (n = 1,829) | Tertile 1 (n = 1,289) | Tertile 2 (n = 1,300) | Tertile 3 (n = 1,298) | ||
Age, y | 67.1 ± 5.6 | 68.5 ± 5.9 | 69.7 ± 5.9 | 70.8 ± 5.8 | <0.001 |
Sex | <0.001 | ||||
Female | 1,306 (71.4) | 774 (60.0) | 707 (54.4) | 454 (35.0) | |
Male | 523 (28.6) | 515 (40.0) | 593 (45.6) | 844 (65.0) | |
White | 1,185 (64.8) | 924 (71.7) | 1,024 (78.8) | 1,100 (84.7) | <0.001 |
Diabetes mellitus | 216 (11.8) | 172 (13.3) | 213 (16.4) | 269 (20.7) | <0.001 |
Current smoker | 117 (6.4) | 107 (8.3) | 110 (8.5) | 149 (11.5) | <0.001 |
BMI, kg/m2 | 28.8 ± 5.6 | 29.1 ± 5.4 | 29.2 ± 5.4 | 29.1 ± 5.6 | 0.11 |
LDL cholesterol, mg/dL | 117.6 ± 32.5 | 114.5 ± 33.7 | 113.6 ± 32.8 | 109.7 ± 33.6 | <0.001 |
HDL cholesterol, mg/dL | 58.7 ± 15.1 | 55.8 ± 15.0 | 55.8 ± 15.4 | 52.8 ± 14.5 | <0.001 |
Total cholesterol, mg/dL | 206.6 ± 37.0 | 202.5 ± 38.3 | 201.6 ± 37.7 | 195.0 ± 37.7 | <0.001 |
Triglycerides, mg/dL | 151.5 ± 70.1 | 160.6 ± 73.0 | 161.0 ± 72.7 | 162.6 ± 75.0 | <0.001 |
SBP, mm Hg | 137.3 ± 18.6 | 139.0 ± 18.5 | 140.3 ± 18.4 | 142.2 ± 18.4 | <0.001 |
DBP, mm Hg | 78.3 ± 9.0 | 78.4 ± 8.9 | 77.8 ± 9.1 | 78.1 ± 9.2 | 0.18 |
Lipid-lowering drug | 398 (21.8) | 371 (28.8) | 435 (33.5) | 461 (35.5) | <0.001 |
Antihypertensive agent | 399 (21.8) | 332 (25.8) | 377 (29.0) | 445 (34.3) | <0.001 |
Framingham 10-y risk, % | 5.4 (3.0-9.7) | 7.6 (4.1-13.8) | 9.5 (5.0-15.3) | 13.6 (7.7-19.6) | <0.001 |
Framingham 10-y risk group | <0.001 | ||||
Low risk (≤10%) | 1,391 (76.1) | 798 (61.9) | 684 (52.6) | 439 (33.8) | |
Intermediate risk (10%-20%) | 355 (19.4) | 380 (29.5) | 465 (35.8) | 555 (42.8) | |
High risk (>20%) | 83 (4.5) | 111 (8.6) | 151 (11.6) | 304 (23.4) | |
Deaths | 184 (10.1) | 175 (13.6) | 221 (17.0) | 321 (24.7) | <0.001 |
Mean follow-up time, y | 12.2 | 12.0 | 11.8 | 11.3 |
Association between subclinical atherosclerosis and all-cause mortality
Over a median follow-up of 12.4 years (Q1-Q3: 12.2-12.9 years), 901 (11.7%) participants died. Higher mortality rates were observed with increasing tertiles of cPB (log-rank P < 0.001) (Figure 3A, Table 1) and CAC score (log-rank P < 0.001) (Figure 3B, Table 2). Baseline cPB and CAC score both remained significantly associated with higher all-cause mortality after adjustment for the variables in model 1 (age, sex, and race) and model 2 (model 1 plus risk factors and medications; see Methods), with trend fully adjusted HRs of 1.23 (95% CI: 1.16-1.32) and 1.15 (95% CI: 1.08-1.23), respectively (both P < 0.001) (Figure 4, Table 3). Similar evidence was obtained when cPB and CAC score were considered as continuous log-transformed variables. Moreover, the addition of cPB and CAC score, both separately and in combination, significantly improved the performance of models 1 and 2 (Supplemental Table 2). The improvement was greater with cPB than with CAC score.
Continuous Variablea | No Atherosclerosis | Tertile 1 | Tertile 2 | Tertile 3 | Trend HRb | P Value (Trend) | |
---|---|---|---|---|---|---|---|
Carotid plaque burden | |||||||
Model 1c | 1.06 (1.04-1.08) | 1.0 (ref) | 1.19 (0.94-1.50) | 1.53 (1.23-1.91) | 2.01 (1.62-2.49) | 1.27 (1.19-1.36) | <0.001 |
Model 2d | 1.05 (1.03-1.07) | 1.0 (ref) | 1.17 (0.92-1.47) | 1.49 (1.19-1.86) | 1.84 (1.48-2.28) | 1.23 (1.16-1.32) | <0.001 |
Calcium score | |||||||
Model 1c | 1.04 (1.03-1.06) | 1.0 (ref) | 1.16 (0.94-1.43) | 1.30 (1.06-1.59) | 1.69 (1.38-2.06) | 1.19 (1.11-1.27) | <0.001 |
Model 2d | 1.04 (1.02-1.05) | 1.0 (ref) | 1.11 (0.90-1.36) | 1.22 (0.99-1.49) | 1.52 (1.24-1.86) | 1.15 (1.08-1.23) | <0.001 |
Association between plaque progression and all-cause mortality
A subgroup of 732 (12.8%) participants underwent follow-up carotid VUS 8.9 years (range 8.4-9.7 years) after the baseline exam. The baseline characteristics of this subgroup are compared with the rest of the study population in Supplemental Table 3. Within the follow-up VUS subgroup, 571 (78.0%) participants showed progression of cPB, 63 (8.6%) showed cPB regression, and 98 (13.4%) remained free of carotid atherosclerosis. The number of deaths and 3-year risk were 33 (5.4%), 2 (2.0%), and 1 (1.6%), respectively. Baseline characteristics of the follow-up VUS subgroup stratified by progression profile are summarized in Table 4. Significant predictors of progression were baseline cPB, male sex, central obesity, serum triglyceride level, and age (Supplemental Table 4). Supplemental Figure 2 is a plot of follow-up vs baseline cPB in all participants.
cPB Regression/No Atherosclerosis (n = 161) | cPB Progression (n = 571) | P Value | |
---|---|---|---|
Demographics | |||
Age, y | 67.3 ± 5.5 | 69.2 ± 5.4 | <0.001 |
Sex | 0.66 | ||
Female | 99 (61.5) | 362 (63.4) | |
Male | 62 (38.5) | 209 (36.6) | |
White | 124 (77.0) | 501 (87.7) | <0.001 |
Medical history | |||
Current smoker | 8 (5.0) | 28 (4.9) | 0.97 |
Obesity | 59 (36.6) | 180 (31.5) | 0.22 |
Central obesity | 113 (70.2) | 424 (74.3) | 0.30 |
Dyslipidemia | 77 (47.8) | 295 (51.7) | 0.39 |
Hypertension | 80 (49.7) | 308 (53.9) | 0.34 |
Diabetes mellitus | 22 (13.7) | 60 (10.5) | 0.26 |
BMI, kg/m2 | 28.1 ± 5.4 | 28.3 ± 4.9 | 0.71 |
SBP, mm Hg | 134.4 ± 18.4 | 135.9 ± 17.7 | 0.37 |
DBP, mm Hg | 76.8 ± 8.6 | 76.1 ± 8.6 | 0.32 |
VUS imaging and CAC score | |||
cPB at baseline, mm3 | 0.0 (0.0-29.1) | 42.2 (10.2-112.5) | <0.001 |
cPB at follow-up, mm3 | 0.0 (0.0-20.4) | 132.3 (56.7-264.7) | <0.001 |
Calcium score | 2.5 (0.0-92.5) | 52.0 (0.0-229.0) | <0.001 |
Biochemistry | |||
eGFR, mL/min/1.73 m2 | 71.6 ± 12.4 | 71.9 ± 12.9 | 0.82 |
Creatinine, mg/dL | 0.9 (0.8-1.1) | 0.9 (0.8-1.0) | 0.16 |
Total cholesterol, mg/dL | 196.4 ± 40.6 | 203.6 ± 36.3 | 0.03 |
LDL cholesterol, mg/dL | 107.1 ± 30.6 | 113.1 ± 31.4 | 0.03 |
HDL cholesterol, mg/dL | 57.1 ± 17.6 | 57.1 ± 14.7 | 1.00 |
Triglycerides, mg/dL | 153.0 (106.5-190.5) | 150.0 (107.0-216.0) | 0.24 |
Glucose, mg/dL | 98.0 (90.0-110.0) | 97.0 (90.0-107.0) | 0.23 |
Uric acid, mg/dL | 5.1 ± 1.5 | 5.2 ± 1.5 | 0.29 |
GGT, μ/L | 22.0 (16.0-31.0) | 22.0 (16.0-31.0) | 0.59 |
Pharmacological treatment | |||
Lipid-lowering drug | 49 (30.4) | 192 (33.6) | 0.45 |
Antihypertensive agent | 34 (21.1) | 142 (24.9) | 0.33 |
Antidiabetic | 18 (11.2) | 42 (7.4) | 0.12 |
10-y CVD risk scores | |||
Framingham 10-y risk, % | 5.8 (3.0-11.8) | 6.5 (3.5-13.1) | 0.14 |
Framingham 10-y risk group | 0.62 | ||
Low risk (≤10%) | 108 (67.5) | 366 (64.7) | |
Intermediate risk (10%-20%) | 42 (26.3) | 152 (26.9) | |
High risk (>20%) | 10 (6.3) | 48 (8.5) | |
Follow-up | |||
Deaths | 3 (1.9) | 33 (5.8) | 0.04 |
Mean follow-up time, y | 12.3 | 12.3 |
Individuals with cPB progression had significantly higher all-cause mortality compared with individuals with cPB regression or absence of disease (log-rank P = 0.04) (Figure 5, Table 4). The association between cPB progression (as a continuous measurement, not as a dichotomous one) and all-cause mortality remained significant after adjustment for baseline cPB and the variables in model 1 (age, sex, and race) and model 2 (model 1 plus risk factors and medications) (HR: 1.03; 95% CI: 1.01-1.04 per absolute 10-mm3 change; P = 0.01) (Table 5, Central Illustration).
HR (95% CI) | P Value | |
---|---|---|
Model 1 | ||
cPB progression trajectory | ||
Regressorsa/absence of diseaseb | Reference group | |
Progressors | 2.73 (0.79-9.46) | 0.11 |
Absolute change in cPB (per 10 mm3) | 1.03 (1.01-1.04) | 0.01 |
Model 2 | ||
cPB progression trajectory | ||
Regressorsa/absence of diseaseb | Reference group | |
Progressors | 2.93 (0.84-1.0.22) | 0.09 |
Absolute change in cPB (per 10 mm3) | 1.03 (1.01-1.04) | 0.01 |
Discussion
In a contemporary prospective cohort of 5,716 asymptomatic adults in the United States assessed by carotid VUS and CAC scoring, the presence of subclinical atherosclerosis at baseline detected with either imaging modality was significantly associated with all-cause mortality, and these associations remained after multivariate adjustment. While both imaging modalities provided additional predictive value beyond CVRFs alone, carotid VUS performed significantly better than CAC. Moreover, VUS-detected cPB progression independently predicted all-cause mortality even after adjustment for CVRFs, background medication, and baseline cPB.
Prevalence of subclinical atherosclerosis detected by CAC and VUS and its association with FRS
The prevalence of subclinical atherosclerosis detected with either modality at enrollment in this population was very high (near 80%), as previously reported by Baber et al.4 This prevalence is substantially higher than described in other cohorts,5,25 likely reflecting the older age and higher risk profile of the BioImage study population, as well as the greater sensitivity of the modalities used to detect atherosclerosis. Most of the BioImage participants, particularly those with an intermediate or high FRS, already had multiterritorial subclinical atherosclerosis at baseline affecting the carotid and coronary arteries. VUS detected disease in 72%, 83%, and 91% of participants with a low, intermediate, and high FRS, respectively, whereas CAC scoring detected disease in 58%, 80%, and 87% of participants in these categories. Interestingly, although the percentage of patients with isolated CAC was similar in each FRS category (∼10%), the percentage of patients with isolated carotid disease was substantially higher in the low-FRS group (25%) and decreased with increasing FRS category. This suggests that VUS is a more sensitive detector of subclinical atherosclerosis in individuals assigned a low cardiovascular risk based on CVRFs (in whom conventional scores may underestimate risk) and before the onset of calcification of coronary lesions.
Association between imaging-detected subclinical atherosclerosis and all-cause mortality
Subclinical atherosclerosis detected by carotid VUS or CAC scoring was positively associated with all-cause mortality at a median follow-up of 12.4 years. The positive association between increasing cPB or CAC score tertile and all-cause mortality was also observed when atherosclerosis burden was considered as a continuous variable. In the multivariate analysis, the association between atherosclerosis burden detected with either modality and all-cause mortality remained statistically significant after adjustment for CVRFs and background medication, highlighting the incremental prognostic value of direct atherosclerosis imaging in asymptomatic individuals. In our cohort, from the fully adjusted model we know that only age and cigarette smoking had a stronger prediction of mortality risk than carotid plaque burden. Moreover, while both imaging modalities improved the performance of the multivariate model incorporating CVRFs and background medication (model 2), carotid VUS performed significantly better than CAC scoring. It would thus appear that, while both imaging techniques improve risk prediction and reclassification based on traditional risk factors, carotid VUS more accurately refines the risk of all-cause mortality. There have been few direct comparisons of risk prediction with CAC scoring vs carotid vascular disease metrics such as carotid intima-media thickening (cIMT), maximal plaque thickness, or plaque burden. A previous study of the BioImage cohort showed that the addition of CAC score or cPB to traditional CVRFs produced a similar improvement in the prediction of a composite outcome of cardiovascular death, myocardial infarction, or ischemic stroke at a median follow-up of 2.7 years.4 However, given the short follow-up, this earlier study lacked sufficient statistical power to assess differences in all-cause mortality prediction. In a meta-analysis of 25 studies assessing the added value of flow-mediated dilatation, CAC score, cIMT, or carotid plaque (using a variety of indicators such as presence of plaque, sum of all plaque areas, or sum of segments with plaque) for the stratification of CVD risk, CAC scoring performed slightly better than the other parameters, although it should be noted that the population characteristics differed between studies.13 Two key considerations may underlie our findings. First, volumetric measurement of cPB is a more accurate than other atherosclerosis measures such as cIMT, which is no longer recommended for cardiovascular risk assessment because it instead reflects hypertensive or age-related changes.26 Indeed, Nicolaides et al19 demonstrated that VUS-detected carotid plaque area provided better prediction of future atherosclerotic CVD events than cIMT or carotid plaque thickness. Moreover, the novel volumetric measurement used in this study has a very intra- and interobserver concordance, with intraclass correlation coefficients higher than 0.9. The second consideration is the particular characteristics of the BioImage study population, in which 20% of participants with zero CAC had evidence of subclinical carotid atherosclerosis that may help to redefine mortality risk, whereas only 11% had isolated CAC.
Association between VUS-detected carotid atherosclerosis progression and all-cause mortality
The wide availability of VUS, which is safe and inexpensive, makes it an ideal technique for monitoring atherosclerosis progression. In our study, 732 participants underwent a second carotid VUS examination after a median interval of ∼9 years. Approximately two-thirds of these participants showed some degree of cPB progression, which was significantly related to older age, male sex, central obesity, serum triglyceride levels, and a higher cPB at baseline. The remaining third of participants showed cPB regression or remained free of carotid atherosclerosis. In a very recent study of the PESA (Progression of Early Subclinical Atherosclerosis) cohort (baseline age 40-55 years, 36% female), progression—defined as a ≥100% increase in carotid and femoral plaque burden from baseline to 6 years—occurred in 32.7% of individuals and was associated with older age, male sex, smoking, LDL-C, and SBP.27 While the association with age and male sex was similarly predominant in our study, baseline cPB, central obesity, and serum triglycerides had a greater predictive capacity than both LDL-C and SBP. This likely reflects the significantly older age profile of our population (mean age at enrollment 67 vs 47 years), the higher proportion of participants receiving antihypertensive and lipid-lowering therapy (22% vs 7%) and the differential effect of SBP and LDL-C according to age.27 Moreover, not just the baseline risk factors, but also the longitudinal changes in risk factor control throughout life may have also contributed. Interestingly, cPB progression between VUS evaluations in our study population was significantly associated with all-cause mortality even after adjusting for baseline cPB and for CVRFs and background medication at the time of the second carotid VUS examination. Differences in mortality between progressors and individuals with regression or free of disease did not reach statistical significance probably due to limited statistical power (just 3 deaths in the regression/absence of disease category).
To our knowledge, this is the first report showing an association between cPB progression and all-cause mortality. Previous studies have reported associations between imaging-assessed plaque progression in the coronary tree and the incidence of acute coronary events. In the PROSPECT (Providing Regional Observations to Study Predictors of Events in the Coronary Tree) study,28 nonculprit lesions detected by intravascular ultrasound doubled in size between the baseline assessment and the diagnosis of acute coronary syndrome. In a cohort of 449 patients monitored by serial coronary artery CT, plaque progression was an independent predictor of acute coronary syndrome.29 Eghtedari et al15 recently showed that, in 3,260 individuals referred by their primary physician for CAC measurement, annualized CAC progression of >20 U/y independently predicted all-cause mortality. In addition, Sabeti et al30 demonstrated that progression of carotid stenosis within a 6- to 9-month interval by VUS predicted the occurrence of major adverse cardiovascular events. Our study supports these previous works that have demonstrated the prognostic relevance of the progression of atherosclerosis and adds that cPB progression (even without causing stenosis) is associated with greater all-cause mortality. These results suggest that noninvasive atherosclerosis monitoring has potential to improve primary prevention by reinforcing lifestyle recommendations, establishing finer control of CVRFs, and facilitating closer follow-up of patients showing disease progression with the ultimate goal of improving their survival. However, selecting the best imaging modality is not straightforward because plaque characteristics can change in response to lipid-lowering therapy. Several imaging studies have demonstrated that intensive statin therapy reduces total plaque burden and progression by decreasing necrotic core volume but typically increases fibrous cap thickness and the degree of calcification.31-33 The consistent increase in plaque calcification observed with statin therapy33,34 makes plaque volume a more useful prognostic indicator than parameters that record calcium volume, such as CAC scoring. In addition, VUS is an ideal method for this purpose due to its wide availability, patient safety, and low cost. Currently, it requires some training, but it might change in the future with the use of automatic AI measurements.35 Further research comparing the predictive accuracy of cPB and CAC progression would be needed to confirm which technique may provide more clinically useful information.
Study limitations
First, all-cause mortality was identified from Social Security and National Death Index searches, and we do not have reliable access to the cause of death or the incidence of other cardiovascular events. While we duly acknowledge this limitation, all-cause mortality remains the main hard event in all studies of mortality risk and has unquestionable relevance to the study of predictive factors and the evaluation of potential preventive measures. Second, the BioImage study participants were mostly White and about half were women, and the study population was older than others examined in previous primary prevention cohorts, so the results may not be fully generalizable. Third, baseline characteristics of the 385 (6.3%) participants excluded due to missing values significantly differed from the analysis population because of a higher cardiovascular risk. This fact has probably reduced the statistical power to demonstrate an association between atherosclerosis and mortality but reinforces the potential prognostic value of imaging in a population that was mostly low-intermediate risk. Finally, the participant subcohort available for assessment of atherosclerosis progression was relatively small, due to difficulties in locating participants or death before contact. The number of deaths between the second VUS and the end of follow-up was limited, and the death of some patients before the second VUS could be performed introduces a competing risk, as death precludes progression. However, even though this effect would tend to reduce the statistical power and dilute the association, we found a statistically significant association between progression and mortality adjusted for baseline plaque burden at enrollment.
Conclusions
In a population of asymptomatic individuals without previous clinical CVD, the identification and quantification of subclinical atherosclerosis by CAC scoring and carotid VUS was independently associated with all-cause mortality, improving prediction based on CVRFs and background medication. VUS performed better than CAC, adding value even beyond CAC scoring. Furthermore, carotid atherosclerosis progression assessed by carotid VUS independently predicted all-cause mortality adjusted for baseline plaque burden.
Funding Support and Author Disclosures
The High-Risk Plaque BioImage initiative was a precompetitive industry collaboration funded by BG Medicine, Abbott Vascular, AstraZeneca, Merck & Co., Philips, and Takeda. These funders had no role in the present study. Dr Ibanez was supported by the European Commission (grant numbers H2020-HEALTH 945118 and ERC-CoG 819775), La Caixa Foundation (HR22-00533, DIREQT2Heart), the Spanish Ministry of Science and Innovation (PID2022-140176OB-I00), and the Red Madrileña de Nanomedicina en Imagen Molecular–Comunidad de Madrid (2022/BMD-7403 RENIM-CM). Dr García-Álvarez was supported by the Instituto de Salud Carlos III (PI20/00742 and PI23/01341) and La Fundació la Marató TV3 (202314-30/31). Dr Devesa is an Alfonso Martin Escudero fellow and was scientifically supported by la La Caixa Foundation. The CNIC is supported by the Instituto de Salud Carlos III (ISCIII), the Ministerio de Ciencia, Innovación y Universidades (MICIU) and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (grant CEX2020-001041-S funded by MICIU/AEI/10.13039/501100011033). Dr Sánchez-González is an employee of Philips. Dr Muntendam is CEO of SQ innovation. Philips and SQ innovation had no role in the present study. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Abbreviations and Acronyms
CAC | coronary artery calcium |
cIMT | carotid intima-medial thickening |
cPB | carotid plaque burden |
CT | computed tomography |
CVD | cardiovascular disease |
CVRF | cardiovascular risk factor |
FRS | Framingham risk score |
LDL-C | low-density lipoprotein cholesterol |
SBP | systolic blood pressure |
VUS | vascular ultrasound |
References
1. "Heart Disease and Stroke Statistics-2022 Update: a report from the American Heart Association". Circulation . 2022;145:E153-E639.
2. "2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines". J Am Coll Cardiol . 2014;63:25 Pt B: 2935-2959. https://doi.org/10.1016/j.jacc.2013.11.005.
3. "Incidence of atherosclerotic cardiovascular disease in young adults at low short-term but high long-term risk". J Am Coll Cardiol . 2023;81:623-632.
4. "Prevalence, impact, and predictive value of detecting subclinical coronary and carotid atherosclerosis in asymptomatic adults: the BioImage study". J Am Coll Cardiol . 2015;65:1065-1074.
5. "Prevalence, vascular distribution, and multiterritorial extent of subclinical atherosclerosis in a middle-aged cohort". Circulation . 2015;131:2104-2113.
6. "Prevalence of subclinical coronary artery atherosclerosis in the general population". Circulation . 2021;144:916-929.
7. "Coronary artery calcification compared with carotid intima-media thickness in the prediction of cardiovascular disease incidence: the Multi-Ethnic Study of Atherosclerosis (MESA)". Arch Intern Med . 2008;168:1333-1339.
8. "Coronary calcium as a predictor of coronary events in four racial or ethnic groups". N Engl J Med . 2008;358:1336-1345.
9. "Carotid and femoral ultrasound morphology screening and cardiovascular events in low risk subjects: a 10-year follow-up study (the CAFES-CAVE study(1))". Atherosclerosis . 2001;156:379-387.
10. "Carotid intima-media thickness and presence or absence of plaque improves prediction of coronary heart disease risk: the ARIC (Atherosclerosis Risk In Communities) study". J Am Coll Cardiol . 2010;55:1600-1607.
11. "A simple disease-guided approach to personalize ACC/AHA-recommended statin allocation in elderly people: the BioImage study". J Am Coll Cardiol . 2016;68:881-891.
12. "Short-term progression of multiterritorial subclinical atherosclerosis". J Am Coll Cardiol . 2020;75:1617-1627.
13. "Improvements in risk stratification for the occurrence of cardiovascular disease by imaging subclinical atherosclerosis: a systematic review". Heart . 2012;98:177-184.
14. "Calcium deposition within coronary atherosclerotic lesion: Implications for plaque stability". Atherosclerosis . 2020;306:85-95.
15. "Coronary artery calcium progression and all-cause mortality". Coron Artery Dis . 2023;34:244-249.
16. "Progression of CAC score and risk of incident CVD". JACC Cardiovasc Imaging . 2016;9:1420-1429.
17. "Accurate quantification of atherosclerotic plaque volume by 3D vascular ultrasound using the volumetric linear array method". Atherosclerosis . 2016;248:230-237.
18. "Subclinical atherosclerosis burden by 3D ultrasound in mid-life: the PESA study". J Am Coll Cardiol . 2017;70:301-313.
19. "Arterial ultrasound testing to predict atherosclerotic cardiovascular events". J Am Coll Cardiol . 2022;79:1969-1982.
20. "Carotid lumen diameter is associated with all-cause mortality in the general population". J Am Heart Assoc . 2020;9:e015630.
21. "Diameter reduction determined through carotid ultrasound associated with cardiovascular and all-cause mortality: a single-center experience of 38 201 consecutive patients in Taiwan". J Am Heart Assoc . 2021;10:e023689.
22. "The BioImage study: novel approaches to risk assessment in the primary prevention of atherosclerotic cardiovascular disease-study design and objectives". Am Heart J . 2010;160:49-57.e1.
23. "Carotid plaque burden as a measure of subclinical atherosclerosis". JACC Cardiovasc Imaging . 2012;5:681-689.
24. "New 3-dimensional volumetric ultrasound method for accurate quantification of atherosclerotic plaque volume". JACC Cardiovasc Imaging . 2022;15:1124-1135.
25. "Association of ankle-brachial index and plaques in the carotid and femoral arteries with cardiovascular events and total mortality in a population-based study with 13 years of follow-up". Eur Heart J . 2006;27:2580-2587.
26. "2021 ESC guidelines on cardiovascular disease prevention in clinical practice". Eur Heart J . 2021;42:3227-3337.
27. "Determinants of progression and regression of subclinical atherosclerosis over 6 years". J Am Coll Cardiol . 2023;82:2069-2083.
28. "A prospective natural-history study of coronary atherosclerosis". New England Journal of Medicine . 2011;364:226-235.
29. "Plaque characterization by coronary computed tomography angiography and the likelihood of acute coronary events in mid-term follow-up". J Am Coll Cardiol . 2015;66:337-346.
30. "Progression of carotid stenosis detected by duplex ultrasonography predicts adverse outcomes in cardiovascular high-risk patients". Stroke . 2007;38:2887-2894.
31. "Serial coronary CT angiography-verified changes in plaque characteristics as an end point evaluation of effect of statin intervention". JACC Cardiovasc Img . 2010;3:691-698.
32. "Impact of statin therapy on plaque characteristics as assessed by serial OCT, grayscale and integrated backscatter-IVUS". JACC Cardiovasc Img . 2012;5:169-177.
33. "Effect of intensive versus moderate lipid-lowering therapy on epicardial adipose tissue in hyperlipidemic post-menopausal women: a substudy of the belles trial (beyond endorsed lipid lowering with ebt scanning)". J Am Coll Cardiol . 2013;61:1956-1961.
34. "Effect of statin treatment on coronary plaque progression - a serial coronary CT angiography study". Atherosclerosis . 2013;231:198-204.
35. "Personalized intervention based on early detection of atherosclerosis: JACC State-of-the-Art Review". J Am Coll Cardiol . 2024;83:21: 2112-2127.
Footnotes
Listen to this manuscript's audio summary by Editor Emeritus Dr Valentin Fuster on www.jacc.org/journal/jacc.
David Maron, MD, served as Guest Associate Editor for this paper. Javed Butler, MD, MPH, MBA, served as Guest Editor-in-Chief for this paper.
Review and acceptance occurred under Dr Valentin Fuster’s term as Editor-in-Chief.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.