Carotid plaque score is associated with 10-year major adverse cardiovascular events in low–intermediate risk patients referred to a general cardiology community clinic

Abstract

Aims

Atherosclerotic carotid plaque assessments have not been integrated into routine clinical practice due to the time-consuming nature of both imaging and measurements. Plaque score, Rotterdam method, is simple, quick, and only requires 4–6 B-mode ultrasound images. The aim was to assess the benefit of plaque score in a community cardiology clinic to identify patients at risk for major adverse cardiovascular events (MACE).

Methods and results

Patients ≥ 40 years presenting for risk assessment were given a carotid ultrasound. Exclusions included a history of vascular disease or MACE and being >75 years. Kaplan–Meier curves and hazard ratios were performed. The left and right common carotid artery (CCA), bulb, and internal carotid artery were given 1 point per segment if plaque was present (plaque scores 0–6). Administrative data holdings at ICES were used for 10-year event follow-up. Of 8472 patients, 60% were females (n = 5121). Plaque was more prevalent in males (64% vs. 53.9%; P < 0.0001). The 10-year MACE cumulative incidence estimate was 6.37% with 276 events (males 6.9% vs. females 6.0%; P = 0.004). Having both maximal CCA intima media thickness < 1.00 mm and plaque score = 0 was associated with less events. A plaque score < 2 was associated with a low 10-year event rate (4.1%) compared with 2–4 (8.7%) and 5–6 (20%).

Conclusion

A plaque score ≥ 2 can re-stratify low–intermediate risk patients to a higher risk for events. Plaque score may be used as a quick assessment in a cardiology office to guide treatment management of patients.

Graphical Abstract

Graphical Abstract.

Introduction

It is estimated that over 50% of men and 64% of women who die of sudden cardiac death did not have a previous manifestation of atherosclerosis and most were not considered high risk by the Framingham score.¹ Detection of carotid plaque by ultrasound for cardiac risk stratification is recommended by the European guidelines (Class IIa, level b),² recognizing the value of screening in low–intermediate risk patients to reduce major adverse cardiovascular events (MACE). The Canadian guidelines address neither the benefit nor disadvantages of plaque detection. Although statin therapy is indicated for ‘clinical atherosclerosis’, there are no guidelines for detecting subclinical disease. We have recently made recommendations on carotid plaque assessment, but outcome-based studies are lacking to support change in guidelines.³

Historical methods of cardiovascular (CV) risk assessment, such as the Framingham Risk Score (FRS) and Reynolds Risk Score, have utilized traditional risk factors and scoring algorithms to predict risk of outcomes.⁴ However, there are other clinical tools available that allow atherosclerosis identification, thereby propelling risk assessment beyond traditional risk factors.^5–7 Magnetic resonance imaging (MRI) has been established as an excellent modality for the characterization of plaque in the carotid arteries, but can be expensive and difficult to integrate into a general cardiology clinic.^8,9 Carotid ultrasound is a relatively inexpensive, rapid, and safe tool that provides a multitude of data points that have been shown to predict CV morbidity and mortality.^3,10–12 Whereas carotid intima media thickness (CIMT) adds little to CV risk prediction, especially in women,^13,14 carotid plaque is a more powerful predictor of CV events in both sexes.^15–18

There are several methods of quantifying plaque burden, including maximum plaque height,^19,20 total plaque area,^19,20 and total plaque volume.^21–23 Most have not been routinely implemented into general clinical practice due to the time-consuming nature of the imaging and measurements. Plaque score, using the Rotterdam method,²⁴ as described by the American Society of Echocardiography recommendations,²⁵ is both simple and quick to perform and can be obtained with as little as 4–6 B-mode ultrasound images.

The purpose of this study was to assess the benefit of using plaque score in a general cardiology community clinic setting to identify patients at risk for MACE and provide further evidence for atherosclerosis screening.

Methods

Human ethical considerations

This study was approved by Queen's University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board (#6022425, DMED-2105-17). All participants provided written consent. This study conforms to the ethical guidelines of the 1975 Declaration of Helsinki.

Patient population

The Kingston Heart Clinic (KHC, Ontario, Canada) is a community-based referral clinic for general cardiology that has been providing clinical and cardiac testing for >25 years. Consecutive males aged 40–70 years and females aged 40–75 years, referred for cardiovascular risk assessment between 16 January 2006 and 31 December 2018, were approached for research. Patients received either a formal carotid Doppler or a focussed carotid screen.^15,19 Exclusions were: known vascular disease, history of known coronary interventions [coronary artery bypass graft (CABG) or percutaneous coronary intervention (PCI)], history of CV events [myocardial infarction (MI), stroke, or transient ischaemic attack], or other non-cardiac vascular procedures [carotid endarterectomy, carotid stenting, abdominal aortic aneurysm repair or stenting, any lower extremity bypass or percutaneous transluminal angioplasty]. These patients are already considered high risk by FRS or the Canadian lipid guidelines.²⁶ Baseline demographics were collected from the clinic visit and patients who had both IMT and plaque score data were included.

Carotid ultrasound imaging

Carotid images were obtained using either a 9L probe on a GE VIVID E9 or a 10L probe on a GE VIVID 7 Dimension or an 11L probe on a GE E90 ultrasound machine (GE Healthcare, Mississauga, Canada). One long axis still image from the common carotid artery (CCA) and one of the bulb/internal common carotid artery (ICA) were obtained for the left and right sides to determine intima media thickness (IMT) and plaque score. Maximal CCA IMT was measured using the GE automated edge detection tool or the manual callipers if plaque was present. If plaque was present in the CCA, it was used as the maximal CCA IMT. Carotid plaque was assessed as either present or absent, defined as CIMT > 1.5 mm, using the Atherosclerosis Risk in Communities Study (ARIC) definition.^11,27 Plaque score was calculated using the Rotterdam method,²⁴ in which each the CCA, carotid bulb and internal carotid artery (ICA) on each side receives a single point if plaque is present on either wall of the segment. The minimal score is ‘0’ and the maximal score is ‘6’; the score only requires the presence/absence of plaque and does not consider the amount or number of plaques. For example, a patient with any plaque present in the left carotid bulb, right carotid bulb, and right CCA would yield a plaque score of three (Figure 1).

Figure 1.

Plaque score using the Rotterdam method. The score is calculated from the CCA, bulb, and ICA segments (the ECA is not used in this score). If plaque is present in one segment (regardless of number of plaques or wall location), 1 point is assigned. (A) Illustration and (B) ultrasound example of a patient's left side with a plaque score = 1 (plaque in bulb). Corresponding (C) illustration and (B) ultrasound of the patient's right side with a plaque score = 2 (plaque in bulb and CCA). The total plaque score for this patient was 3. BioRender.

As carotid IMT is clinically defined as being normal if <1.00 mm (ARIC study)²⁸ and the absence of plaque¹⁷ has been associated with lower rates of MACE, we assessed maximal CCA IMT < 1.00 mm combined with a plaque score = 0. Carotid maximum plaque height and total plaque area in the bulb/ICA for subgroups were measure as previously described.^19,20

General protocol: Perform a short axis sweep of the carotid artery starting at the CCA, moving towards the bulb/bifurcation, and following the ICA to identify any presence of plaque. If plaque is present, the orientation of the transducer is changed to long axis to locate the plaque in the three segments (CCA, ICA, bulb) and images are captured. One point is assigned per segment where plaque is present to calculate the left side plaque score. This is repeated on the right side. The sum of the plaque scores for the two sides gives the patient plaque score. The method takes <10 min to perform.

Follow-up for MACE

If patients had multiple carotid examinations throughout the accrual window, the first examination was considered the index date. MACE was defined as a composite of CV death [death date and cause of death matching one of the outcome codes, determined from Office of the Registrar General (ORG)], OR non-fatal acute coronary syndrome (MI, STEMI or NSTEMI, or unstable angina) OR non-fatal stroke OR unplanned coronary revascularization (CABG or PCI associated with inpatient hospital admission through emergency department or via ambulance). End of follow-up period was 31 December 2018, as cause of death was not available beyond that date, limiting the use of valid composite outcomes with CV death. As some patients had multiple events, only the first CV event was used. Time to event was calculated as the number of days between the index date and the earliest of the date of first event or end of follow-up.

Outcome data sources

Outcome data were made available through a data sharing agreement between the ICES (Institute for Clinical Evaluative Sciences) and the Ontario Ministry of Health and Long-Term Care (MOHLTC). These datasets were linked using unique encoded identifiers and analysed at ICES. ICES is an independent, non-profit research institute whose legal status under Ontario's health information privacy law allows it to collect and analyse health care and demographic data, without consent, for health system evaluation and improvement. The ICES data holdings are a comprehensive provincial registry capturing all diagnostic and clinical activity unique to our region (Ontario, Canada, population 15.5 M). Access to this registry was possible because the OHIP is a single payer system that provides nearly the entire Ontario population with universal insurance coverage for physician services and hospital care. Clinic patient data were deterministically linked to the ICES data holdings using each patient's unique Ontario Health Insurance Plan (OHIP) number, date of birth, and sex. The OHIP database contains billing information of about 94% of Ontario's physicians.²⁹ All outcomes were identified through a variety of administrative data holdings: Discharge Abstract Database (DAD) from Canadian Institute for Health Information (CIHI); National Ambulatory Care Reporting System (NACRS) from CIHI; Same Day Surgery (SDS) from CIHI; Office of the Registrar General Database (vital statistics—cause of death); and Registered Persons Database (RPDB—death date), using the International Classification of Diseases 10th Revision (ICD-10-CA) codes and OHIP billing codes. These administrative databases have been previously validated for these outcomes.^30,31

Statistical analyses

Statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Continuous variables were compared using the independent t-test and Kruskal–Wallis test (non-parametric). Categorical variables were compared using the χ² for homogeneity. Gray's Test was used to determine differences in MACE cumulative incidence between sexes at any time points over 10 years accounting for the competing risk of non-CV death. Receiver operating characteristic curves were used to assess the area under the curve (AUC) and the Youden's index was used to determine the carotid variables’ optimal threshold value for predicting 1-year, 5-year, and 10-year MACE. Contingency tables were used to calculate test sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Kaplan–Meier curves and log rank test were used to assess event-free survival over 10 years. Plaque score risks for MACE were evaluated using Cox proportional hazard models. Unadjusted and adjusted models were performed {adjusting for age [continuous], sex, estimated glomerular filtration rate [eGFR] [stages G1–G5], body mass index [BMI underweight (<18.5), normal (18.5–25), overweight (25–30), and obese (≥30)], hypertension, diabetes, dyslipidaemia, and current smoking}. The proportional hazards assumption was evaluated using Martingale residuals, specifically the Kolmogorov Supremum test was based on 1000 re-samples of 20 simulated residual patterns each. McNemar's test was used to compare differences in sensitivity and predictive accuracy of cut-points for maximal CCA IMT and plaque score. P < 0.05 was considered significant.

Results

Population characteristics

A total of 8472 patients were included. Reasons for referral were asymmetric brachial pressures, carotid bruit, diabetes, dizziness, dyslipidaemia, follow-up on carotid disease/dissection, hypertension, left arm or facial numbness, pain on side of neck at carotid location, presyncope, pulsatile noise in ears or neck, spasm on left side of face, retinal haemorrhage, smoker, subclavian bruit, syncope, tinnitus, transient cerebral ischaemic attack, assessment of coronary ischaemia, and ‘other’ cardiovascular risk stratification (see Supplementary data online, Table S1).

A majority of patients was females (60.4%) and the population mean age was 58.7 ± 8.43 years (Table 1). Traditional cardiac risk factors, such as diabetes, hypertension, dyslipidaemia, and smoking, were more prevalent among males than females. Males had a higher maximal CCA IMT (1.12 ± 0.49 mm) compared with females (0.99 ± 0.41 mm); P < 0.0001. Presence of plaque was more prevalent in males (64%) than females (54%); P < 0.0001. A large portion of patients with a plaque score ≥ 2 (59%) was not on aspirin nor a statin at baseline (see Supplementary data online, Table S2).

Table 1.

Population characteristics

Variable	All	Male	Female	P-value
Sex, n (%)	8472 (100)	3351 (39.6)	5121 (60.4)	<0.0001
Age (years), mean ± SD	58.7 ± 8.43	57.15 ± 7.98	59.7 ± 8.56	<0.0001
Median (Q1–Q3)	59 (52–65)	58 (51–64)	60 (53–67)	<0.0001
Cardiac risk factors
Current smoker, n (%)	1606 (19.0)	728 (21.7)	878 (17.1)	<0.0001
Hypertension, n (%)	3626 (42.8)	1538 (45.9)	2088 (40.8)	<0.0001
Dyslipidaemia, n (%)	2866 (33.8)	1342 (40.0)	1524 (29.8)	<0.0001
Diabetes, n (%)	1161 (13.7)	529 (15.8)	632 (12.3)	<0.0001
0 cardiac risk factor	1560 (18.4)	571 (17.0)	989 (19.3)	0.0003
1 cardiac risk factor	2507 (29.6)	958 (28.6)	1549 (30.2)	0.0003
2+ cardiac risk factors	3963 (46.8)	1661 (49.6)	2302 (45.0)	0.0003
Cardiac medications, n (%)
Aspirin	1853 (21.9)	982 (29.3)	871 (17.0)	<0.0001
Statins	1449 (17.1)	752 (22.4)	697 (13.6)	<0.0001
ACE inhibitor	972 (11.5)	522 (15.6)	450 (8.8)	<0.0001
ARB	472 (5.6)	203 (6.1)	269 (5.3)	<0.0001
Carotid variables
Maximal CCA IMT, mean ± SD (mm)	1.04 ± 0.44	1.12 ± 0.49	0.99 ± 0.41	<0.0001
Plaque presence, n (%)	4902 (57.9)	2143 (64.0)	2759 (53.9)	<0.0001

ACE, angiotensin-converting-enzyme inhibitors; ARB, angiotensin receptor blockers; SD, standard deviation.

Ten-year MACE

A total of 276 first CV events occurred, with a 10-year cumulative incidence of 6.37% (Table 2). Males had higher early rates of MACE and non-fatal MI compared with females (Gray's Test P = 0.004 for both), though the difference in 10-year risk between sexes was not statistically significant at 10 years.

Table 2.

Population 10-year CV event counts and cause specific cumulative incidence estimates (1-survival estimates)

Event	All n = 8472	Male n = 3351	Female n = 5121	Gray's P-value
MACE (composite), n (%)	276 (6.37)	141 (6.90)	135 (6.16)	0.004
Cardiovascular death, n (%)	13 (0.25)	6 (0.28)	7 (0.24)	0.76
Non-fatal myocardial infarction, n (%)	136 (3.26)	74 (3.47)	62 (3.30)	0.004
Non-fatal stroke, n (%)	92 (2.36)	41 (2.38)	51 (2.36)	0.77
Emergency revascularization, n (%)	35 (0.78)	20 (1.00)	15 (0.63)	0.11

MACE, major adverse cardiovascular events.

Normal maximal CCA IMT and absence of plaque

Having either CCA IMT < 1.00 mm (Figure 2A) or a plaque score = 0 (Figure 2B) was associated with less events (P < 0.0001). Having both a maximum CCA IMT < 1.00 mm and a plaque score = 0 was also associated with lower events, which was consistent among males (P = 0.0008, Figure 2C) and females (P < 0.0001, Figure 2D).

Figure 2.

Kaplan–Meier curves for 10-year MACE. (A) By maximal CCA IMT < 1.00 mm (normal carotid IMT) vs. ≥1.00 mm, (B) by plaque score = 0 (no plaque) vs. 1–6, (C) by combined CCA IMT < 1 mm and plaque score = 0 vs. the rest in males, and (D) by maximal CCA IMT < 1.00 mm and plaque score = 0 vs. the rest in females.

Maximal CCA and plaque score threshold values for predicting MACE

CCA IMT had moderate discriminatory power at 1 year (AUC = 0.669, CI = 0.605–0.733), 5 years (AUC = 0.643, CI = 0.604–0.682), and 10 years (AUC = 0.674, CI = 0.635–0.712). AUC was higher in females than in males at all evaluated time points (Table 3). The optimal cut-off for 1 year was 1.31 mm for overall, 1.37 mm for females, and 1.00 mm for males. The percentage of correct prediction was highest for females (92.2%). The optimal threshold and AUC increased with age groups. At 5 years, the optimal threshold was nearly unchanged: overall ≥ 1.30 mm, ≥1.04 mm in males, and ≥1.38 mm in females. The 10-year optimal thresholds were lower, but similar between overall (≥0.91 mm), males (≥0.92 mm), and females (≥0.91 mm).

Table 3.

Test characteristics for CCA IMT and plaque score in predicting 1-year MACE

Group	n	MACE Events	Cut-off	PPV (%)	NPV (%)	Sens (%)	Spec (%)	LR+	LR−	AUC (95% CI)	Correct prediction (%)
Max CCA IMT (mm)
All	8013	70	1.31	2.49	99.4	35.7	87.7	2.89	0.73	0.669 (605–0.733)	87.2
40–49 years	1231	1–5a	0.79	0.66	100.0	100.0	38.7	1.63	0.00	0.652 (0.396–0.908)	39.0
50–59 years	2860	20–24a	1.04	1.41	99.6	55.0	72.9	2.03	0.62	0.671 (0.563–0.779)	72.8
<60 years	3998	25	1.04	1.36	99.6	52.0	77.0	2.26	0.62	0.680 (0.581–0.779)	76.8
≥60 years	4115	45	1.31	2.9	99.2	44.4	82.6	2.56	0.67	0.637 (0.548–0.725)	82.2
Females	4804	35	1.37	4.08	99.6	42.9	92.6	5.79	0.62	0.713 (0.622–0.803)	92.2
<60 years	2309	8	1.04	1.48	99.9	62.5	85.6	4.33	0.44	0.759 (0.565–0.952)	85.5
≥60 years	2495	27	1.37	4.17	99.3	44.4	88.8	3.97	0.63	0.645 (0.529–0.760)	88.3
Males	3209	35	1.00	1.47	99.3	68.6	49.4	1.35	0.64	0.598 (0.506–0.690)	49.6
<60 years	1806	17	0.80	1.15	100.0	100.0	18.4	1.23	0.00	0.583 (0.462–0.704)	19.2
≥60 years	1403	18	1.31	2.29	99.1	44.4	75.3	1.80	0.74	0.607 (0.468–0.745)	74.9
Plaque score
All	8013	70	2	1.68	99.6	74.3	61.8	1.94	0.42	0.738 (0.677–0.798)	61.9
40–49 years	1231	1–5a	2	1.46	99.8	60.0	83.5	3.64	0.48	0.689 (0.401–0.976)	83.4
50–59 years	2860	20–24a	2	1.56	99.7	70.0	69.0	2.26	0.43	0.718 (0.600–0.836)	69.0
<60 years	3998	25	2	1.54	99.7	68.0	73.4	2.56	0.44	0.721 (0.613–0.828)	73.4
≥60 years	4115	45	3	2.75	99.4	64.4	73.4	2.42	0.48	0.727 (0.651–0.804)	73.3
Females	4804	35	3	2.61	99.6	54.3	85.1	3.65	0.54	0.730 (0.636–0.824)	84.9
<60 years	2309	8	2	1	99.8	62.5	78.5	2.91	0.48	0.678 (0.457–0.898)	78.4
≥60	2495	27	3	3.01	99.5	63.0	77.8	2.84	0.48	0.712 (0.603–0.821)	77.7
Males	3209	35	2	1.94	99.6	80.0	55.4	1.79	0.36	0.732 (0.653–0.812)	55.6
<60 years	1806	17	2	1.99	99.6	70.6	67.0	2.14	0.44	0.715 (0.591–0.839)	67.0
≥60 years	1403	18	3	2.45	99.3	66.7	65.6	1.94	0.51	0.745 (0.638–0.852)	65.6

AUC, area under the curve; CCA IMT, common carotid artery intima media thickness; CI, confidence interval.

^aSmall cell suppressed (<6 patients) for patient identification privacy.

Plaque score had acceptable discriminatory power for prediction of 1-year MACE in the overall population (AUC = 0.738, CI = 0.677–0.798). This was significantly different in sensitivity and predictive accuracy compared with maximal CCA IMT (P < 0.0001). Discriminatory power was similar in each sex (Table 3). The optimal plaque score cut-off was = 2 with a high negative predictive value (99.6%), but poor positive predictive value (1.7%). The optimal plaque score cut-off for males was 2 vs. 3 for females. The percentage of correct prediction was highest for females (84.9%). The optimal threshold and AUC increased with age groups. Overall thresholds were the same for 5- and 10-year MACE, though discriminatory power declined with increasing time scale (5-year AUC = 0.674, CI = 0.637–0.711; 10-year AUC = 0.628, CI = 0.588–0.668). Although females had a cut-off = 3 for 1-year MACE, it was = 2 at 5 years and 10 years. Males had a threshold = 2 at any time point.

Increasing plaque score was associated with increasing 10-year risk (Figure 3). A plaque score of 0 or 1 was associated with an extremely low 10-year event rate of 4.1% (1-survival estimates). The 10-year risk for plaque scores 2–4 was 8.7% and 20% for plaque scores 5–6. Kaplan–Meier curves demonstrated lower estimated event-free survival for patients with plaque scores 0–1 vs. 2–6 (P < 0.0001), which was consistent among males (P < 0.0001) and females (P < 0.0001).

Figure 3.

Kaplan–Meier curves for 10-year MACE. (A) By plaque score, (B) by plaque score < 2 vs. ≥2, (C) by plaque score < 2 vs. ≥2 in males, (D) by plaque score < 2 vs. ≥2 in females. Increased risk for MACE is observed with increasing plaque score.

Hazard risk for MACE by plaque score

Unadjusted risk for MACE increased with plaque score. Having a plaque score = 1 was not statistically different than a plaque score = 0 (HR = 1.08), but risk for MACE increased with plaque score ≥ 2 (HR = 2.79), with the highest risk with a plaque score = 6 (HR = 9.86) (Table 4). After adjusting for traditional cardiac risk factors, a plaque score ≥ 2 was still associated with higher risk for MACE (HR = 2.05) with the highest risk with plaque score = 6 (HR = 5.75).

Table 4.

Hazard ratios for MACE

Plaque score	Unadjusted hazard ratio (95% CI)	P-value	Adjusteda hazard ratio (95% CI)	P-value
0	Reference		Reference
1	1.08 (0.69–1.68)	0.74	0.94 (0.60–1.48)	0.785
2	2.08 (1.45–2.99)	<0.0001	1.65 (1.13–2.40)	0.010
3	3.09 (2.11–4.52)	<0.0001	2.23 (1.47–3.36)	0.0001
4	3.15 (2.05–4.85)	<0.0001	2.16 (1.36–3.42)	0.001
5	5.06 (2.84–9.04)	<0.0001	2.99 (1.61–5.54)	0.0005
6	9.86 (5.32–18.3)	<0.0001	5.75 (2.94–11.2)	<0.0001
0	Reference		Reference
1–6	2.29 (1.72–3.04)	<0.0001	1.63 (1.20–2.22)	0.002
0–1	Reference		Reference
2–6	2.79 (2.16–3.61)	<0.0001	2.05 (1.55–2.72)	<0.0001
0–2	Reference		Reference
3–6	2.83 (2.20–3.65)	<0.0001	2.01 (1.53–2.66)	<0.0001

CI, confidence interval.

^aModel adjusted for age, sex, eGFR, BMI, hypertension, diabetes, dyslipidaemia, and smoking.

Added value of plaque score to traditional cardiac risk factors

Patients were stratified by number of cardiac risk factors (age ≥ 60 years, current smoker, hypertension, dyslipidaemia, and diabetes) and assessed for survival probability for MACE. Increased risk for MACE was observed with increasing number of risk factors (Figure 4). Plaque score was able to separate high-risk patients for MACE in each category (having 0, 1, and ≥2 risk factors).

Figure 4.

Kaplan–Meier curves for 10-year MACE. (A) By number of traditional cardiac risk factors in patients with 0, 1, 2 or more risk factors, (B) by plaque score < 2 vs. ≥2 in patients presenting with no risk factors, (C) by plaque score < 2 vs. ≥2 in patients presenting with 1 risk factor, (D) by plaque score < 2 vs. ≥2 in patients presenting with 2 or more cardiac risk factors by plaque score. Risk factors: ≥60 years, current smoker, hypertension, dyslipidaemia, and diabetes.

Framingham score was not available for all patients as most of the patients presenting to the clinic did not have cholesterol levels available. The majority of patients was deemed intermediate risk by FRS. In 2094 patients, there was no significant difference between low-, intermediate-, and high-risk by FRS for probability of MACE, but adding plaque score separated the intermediate FRS risk patients (Figure 5).

Figure 5.

Kaplan–Meier curves for 10-year MACE by Framingham Score categories. (A) Patients categorized as high-risk, intermediate-risk, and low-risk by FRS categories. (B) Intermediate-risk FRS patients by plaque score 0-1 and 2-6.

Observer variability

The observer variability between two raters for plaque score was assessed in 100 patients. The Fleiss-Cohen weighted Kappa (equivalent to the ICC) was 0.86 (95% CI 0.81–0.91), indicating excellent agreement between the two raters.

Plaque score compared with maximum plaque height and total plaque area

Although the focus of this study was plaque score, we compared it to two other methods, maximum plaque height²⁰ and total plaque area²⁰ in a subset of patients. Maximum plaque height in the bulb/ICA was measured in 2987 patients and this method gave similar results to plaque score for predicting 1-year MACE [AUC = 0.746 (95% CI: 0.664–0.828) vs. AUC = 0.738 (95% CI: 0.677–0.798); respectively]. Total plaque area was assessed in 7347 patients and had similar predictive capability for 1-year MACE [AUC = 0.736 (CI 0.668–0.804) vs. AUC = 0.738 (CI 0.677–0.789); respectively]. As these methods often require offline software and additional calculations, we focussed on the simpler method, plaque score, which can be performed quickly at the bedside.

Discussion

This community-based study was initiated to determine if carotid atherosclerosis detection by plaque score was associated with MACE and could improve CV risk stratification beyond traditional cardiac risk factors. We found that hazard risk for MACE increased with plaque score and at the optimal threshold ≥ 2 the HR = 2.05. Low–intermediate risk patients by Framingham can be further stratified as higher risk for events if plaque score ≥ 2 and this was consistent in both males and females. To our knowledge, this is the largest study that has assessed the association between plaque score and MACE in a community-based general cardiology clinic. The strength of this study lies in the large number of patients and the use of high-quality health administrative data for outcome ascertainment, which has been extensively validated and used for research. Additionally, our population included a large proportion of female patients (60%), of which are historically underrepresented in CV observational studies and trials.³²

The World Health Organization states that non-communicable diseases kill 41 million people yearly, equivalent to 74% of all deaths globally; and that CV disease accounts for most of these deaths.³³ Carotid plaque can add important information on CV risk by detecting the presence of atherosclerosis, but the Canadian lipid guidelines have not been updated to reflect advances in ultrasound.³⁴ Interestingly, detection of carotid plaque by ultrasound for cardiac risk stratification is already recommended by the European guidelines.²

The initial Rotterdam study published in 2002 assessed patient risk for stroke and cerebral infarction using plaque score and found that increased plaque score was associated with increased risk, irrespective of plaque location.²⁴ In 2004, this crude method for atherosclerosis plaque assessment was found to be associated with incident MI.³⁵ We have now shown that this method can be extended to composite MACE (CV death, MI, stroke, and emergent revascularization) in low–intermediate risk patients referred for CV risk assessment.

It is well established that plaque burden is a better predictor than CIMT in relation to CV disease and risk for CV events.¹⁷ In our population, the AUC for predicting MACE was slightly higher for plaque score compared with maximal CCA IMT (0.738 vs. 0.669). The smaller difference in AUC observed between max CCA IMT and plaque score compared with the literature may be due to the difference in measuring methods. Where carotid IMT is usually measured as a mean at the CCA distal far wall, it excludes any plaque.¹² Our measurement included presence of plaque in the CCA, in which plaque height was used as the max CCA IMT if present. A plaque score ≥ 2 was optimal in risk prediction regardless of IMT.

Other groups have developed variations on how to measure plaque score by including the near CCA as an additional segment (four segments per side) and calculating the plaque score by summing the maximum plaque thickness (in mm) of each segment.^36,37 The Rotterdam method is simpler in that it only requires a quick determination of plaque height (≥1.5 mm) to visually assess plaque presence in three segments per side and can be performed in seconds as no additional calculations are necessary.

The Multi-Ethnic Study of Atherosclerosis (MESA) is a large US diverse population study (>6000 patients) that has assessed patients for carotid plaque, CCA IMT, and coronary calcium scoring (CAC).^6,38–40 The authors demonstrated that presence of plaque was associated with a HR = 1.61 (95% CI: 1.17–2.21; P = 0.003) for cardiovascular events.⁴⁰ In comparison, we found that in our population, a plaque score ≥ 1 (representing presence of plaque) was associated with a HR = 1.63 (CI 1.20–2.22; P = 0.002). Although these finding are very similar, plaque score has the ability to further categorize risk for events by using a scale beyond just plaque presence (Table 4). The ARIC study found that adding both the CIMT and presence of plaque to the traditional cardiac risk factors had the most improvement in AUC for predicting CV events.¹¹

Carotid plaque score is a focused technique that can easily be applied in the clinic by non-specialist physicians, offering rapid and useful data for CV risk stratification beyond traditional risk factors. It is clear that atherosclerosis detection and screening must become a critical component of assessing CV risk factors.

Limitations

A limitation is that patients who were not already on a statin and aspirin found to have plaque 2–6 were recommended to be prescribed aspirin 81 mg daily and a statin to reduce LDL cholesterol by 50% or <2.00 mmol/L, whichever was the greatest. Patients with a plaque score of 0 had no recommendations. Patients with a plaque score of 1 and a plaque area ≥ 25 mm² had the same recommendation as plaque scores 2–6 (see Supplementary data online, Table S2). As >70% of patients were referred for testing only and were not seen by a KHC MD, we have no knowledge if these recommendations were followed. Our results may underestimate risk for MACE if changes in medication were initiated. Limitations to plaque score may include poor imaging of the plaque due to depth of vessel in some patients or plaque being out of plane due to 2D imaging. We recognize that there is a selection bias. Patients are referred to a cardiologist or vascular imaging because the referring physician has either cardiac or vascular concerns. This MACE data cannot be applied to the general population. However, the outcome data can be applied to any outpatient community cardiology practice.

Conclusion

We have demonstrated that a plaque score of 0 or 1 is associated with an extremely low 10-year CV event rate. A plaque score < 2 had a powerful negative predictive value (99.6%) for MACE. Plaque score increased with risk for MACE and a plaque score ≥ 2 helped identify patients at higher risk. Carotid plaque score, by ultrasound, may be used in a cardiology office to guide management of low–intermediate risk patients, including decision on initiation of statin therapy and the need for further testing. This study supports atherosclerotic screening.

Supplementary data

Supplementary data are available at European Heart Journal - Cardiovascular Imaging online.

Funding

This study was supported by ICES, funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC), and funded by Queen's University Department of Medicine and the Kingston Heart Clinic.

Data availability

The data underlying this article cannot be shared publicly due to ICES privacy policy. The data will be shared on reasonable request to the corresponding author.