Impact of Intima-Media Thickness Progression in the Common Carotid Arteries on the Risk of Incident Cardiovascular Disease in the Suita Study.

BACKGROUND: No prospective study of the relationship between intima-media thickness (IMT) progression and incident cardiovascular disease (CVD) has been performed. METHODS AND
RESULTS: We studied 4724 participants (mean age: 59.7±11.9 years; without CVD at the baseline) who had carotid ultrasonographic measurement of IMT on both sides of the entire carotid artery area (ie, the entire scanned common carotid artery [CCA], carotid artery bulb, internal carotid artery, and external carotid artery areas for both sides) between April 1994 and August 2001. Carotid ultrasonographic follow-up was performed every 2 years between April 1994 and March 2005 in 2722 of these participants, newly revealing 193 CCA plaques (maximum IMT in the CCA >1.1 mm). We followed up for incident CVD until December 2013. Statistical analyses were performed using a Cox proportional hazards regression model, evaluated using C statistics, and net reclassification improvement. During the 59 909 person-years of follow-up, we observed 221 strokes and 154 coronary heart disease events. CCA plaque and maximum IMT in the whole carotid artery area >1.7 mm were risk factors for CVD. CCA plaque presented an increased risk of CVD based on C statistics and the reclassification improvement of the current risk prediction model. After adding the new incident CCA plaques, during the 23 702 person-years of follow-up, 69 strokes and 43 coronary heart disease events occurred. The adjusted hazard ratios for incident CCA plaque were 1.95 (95% confidence interval, 1.14-3.30) in CVD and 2.01 (95% confidence interval, 1.01-3.99) in stroke.
CONCLUSIONS: Maximum IMT in the CCA contributed significantly but modestly to the predictive power of incident CVD used in calculating traditional risk factors. This study provides the first demonstration that new progression of incident CCA plaque is a CVD risk.

Clinical Perspective

What Is New?

High mean intima–media thickness (IMT) and carotid plaque in the common carotid artery (CCA) and the entire carotid artery area were risk factors for cardiovascular disease (CVD) in a general Japanese population.

Maximum IMT in the CCA and the in the entire carotid artery area predicted CVD, but only maximum IMT in the CCA contributed significantly but modestly to the predictive power of the risk factors for CVD used in calculating the current risk prediction model and improved risk classification, providing the first such evidence in a non‐Western population.

This study provides the first evidence that new progression of incident carotid plaque (maximum IMT in the CCA >1.1 mm) is associated with an increased risk of later CVD.

What Are the Clinical Implications?

In medical examinations and outpatient clinics, carotid ultrasonography at the CCA can be easily measured both to screen for and to evaluate carotid plaque (maximum IMT in the CCA >1.1 mm) as a risk factor for future CVD.

We plan to develop a risk score for carotid atherosclerosis in the Suita Study cohort for use in medical examinations and outpatient clinics without the need for carotid ultrasonography, with the ultimate goal of preventing carotid atherosclerosis.

Introduction

The carotid intima–media thickness (IMT) is a noninvasive intermediate marker that can be used for the prediction of stroke1 and coronary heart disease (CHD).2, 3 The carotid IMT is also a surrogate marker for cardiovascular risk factors4 and atherosclerosis.5, 6 For preventive medicine, it would be useful to determine whether carotid IMT can be used to supplement traditional cardiovascular risks to increase the probability of predicting cardiovascular disease (CVD). The ARIC (Atherosclerosis Risk in Communities) study showed that adding the carotid IMT to traditional risk factors improved CHD risk prediction.2 In the Framingham Offspring Study, maximum internal carotid artery (ICA) IMT and mean common carotid artery (CCA) IMT both predicted CVD outcomes, but only maximum ICA IMT slightly but significantly improved the classification of risk of CVD.7 In contrast, carotid IMT did not consistently improve the risk classification of individuals in a general8 or hypertensive population9 in meta‐analyses, although carotid IMT was predictive of CVD events. These inconsistent results may have been due to the use of the mean or maximum IMT or to differences in age range, ethnicity, and background of participants, such as the presence of hypertension between studies.

A meta‐analysis of the PROG‐IMT Collaborative Project (individual progression of carotid intima media thickness as a surrogate of vascular risk) revealed a strong association between carotid IMT and CVD events, but the association between individual carotid IMT progression and CVD in the general population is still unproven.10 Almost all studies of the association between carotid IMT and incident CVD were conducted in Western countries; moreover, no prospective study on the association between carotid IMT progression and incident CVD has been performed. We conducted this study (1) to analyze the relationship between various carotid IMT values at baseline and incident CVD and (2) to evaluate the association between carotid IMT progression and the risk of CVD in a general urban Japanese population.

Methods

The data, analytical methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.

Study Population

The study population consisted of members of the Suita Study, an epidemiological study of CVD and cerebrovascular diseases that was based on a random sampling of 15 200 participants in Suita City, stratified by sex and age in 10‐year increments from 1989. For these participants, we set the baseline of the present study as the medical examination held between April 1994 and August 2001. Of the 6590 people examined during their visit to the National Cerebral and Cardiovascular Center (NCVC; located in Suita City), we excluded 352 individuals who had a past or present history of CVD, 1008 who had not undergone a medical examination that included ultrasonography, and 506 who could not be followed up to the end of 2013. A total of 4724 individuals had interpretable carotid ultrasonography images. Missing data (n=1008) were due to scheduling issues or unavailability of the ultrasonography device. The baseline characteristics of the current and excluded participants are shown in Table S1. Details of the original Suita Study design have been published elsewhere.11, 12 Supplemental methods are shown in the expanded methods (Data S1). All participants in our study provided written informed consent, and the NCVC institutional review board approved our study.

Risk Factors and Anthropomorphic Variables

Well‐trained physicians measured each participant's blood pressure (BP) 3 times using a mercury column sphygmomanometer, an appropriately sized cuff, and a standard protocol.11 Before the initial BP reading was obtained, participants were seated at rest for at least 5 minutes. BP values were taken as the average of the second and third measurements, which were recorded >1 minute apart. At the time of the baseline examination, each participant was classified into 1 of 3 categories of systolic and diastolic BP alone: normal (<120/80 mm Hg), systolic/diastolic prehypertension (120–139/80–89 mm Hg), and systolic/diastolic hypertension (≥140/90 mm Hg), respectively. Categories of body mass index, calculated as weight (kg) divided by height (m2), were defined as underweight (<18.5), normal weight (18.5 to <25), and overweight (≥25).13 The prevalence of obesity among all participants was <1.6%, and obese participants were included in the overweight category.

At the baseline examination, we performed routine blood tests that included serum total and high‐density lipoprotein cholesterol and glucose levels. An individual's non–high‐density lipoprotein cholesterol level was calculated by subtracting the high‐density lipoprotein cholesterol from the total cholesterol. Glucose categories were defined as diabetes mellitus (fasting plasma glucose levels [FPG] ≥126 mg/dL, non‐FPG ≥200 mg/dL, or use of diabetes mellitus medication), impaired fasting glucose (FPG 100–125 mg/dL and non‐FPG 140–199 mg/dL), and normal glucose tolerance (FPG <100 mg/dL and non‐FPG <140 mg/dL). The glomerular filtration rate (shown as GFR; mL/min per 1.73 m2) of each participant was calculated using the Modification of Diet in Renal Disease equation modified by the Japanese coefficient (0.881), as follows14:

Chronic kidney disease was defined as an estimated glomerular filtration rate <60 mL/min per 1.73 m2.

Physicians and nurses administered a questionnaire that assessed each participant's habits and illness status at the time. Smoking and drinking habits were classified as current, quit, or never. We defined excessive alcohol consumption as ≥48 g/d ethanol (≥2 goes/d for Japanese sake). The questionnaire asked the participant about his or her past and present history of stroke (cerebral infarction, intracerebral hemorrhage, and subarachnoid hemorrhage) and CHD (myocardial infarction, angina pectoris, and coronary intervention).

Carotid IMT Measures

Details of the carotid ultrasonic examination methods have been published previously.4, 15, 16 Briefly, for this examination, the participant was in a supine position on a bed. We used a high‐resolution B‐mode ultrasound system (SSA‐250A; Toshiba) with a 7.5‐MHz transducer (SMA‐736S; Toshiba). Carotid atherosclerosis was evaluated by high‐resolution ultrasonographic measurement with atherosclerotic indexes of IMT on both sides of the CCA, carotid artery bulb, ICA, and external carotid artery. All measurements were made at the time of scanning with an electronic caliper and were recorded as photocopies. The same well‐trained physician performed all examinations.

Mean IMT was defined as the mean of the IMT of the proximal and distal walls for both sides of the CCA on a longitudinal scan at a point 10 mm proximal from the beginning of the dilation of each carotid artery bulb. The procedure has been described in detail previously.16 The maximum IMT in the CCA (max‐CIMT) was defined as the maximum measurable IMT in the scanned CCA areas, and the maximum IMT in the entire area (max‐IMT) was defined as the maximum measurable IMT in the entire scanned CCA, bulb, ICA, and external carotid artery areas for both sides.

The intrareader reproducibility of the measurements was assessed for the IMT in the CCA of 50 participants just before the start of this study. IMT was examined twice with a 1‐month interval in a blinded manner to obtain the correlation coefficient between the first and second IMT measurements (r=0.87, P<0.001). A paired t test showed no significant difference between the 2 measurements. The mean difference and standard deviation between the 2 measurements of the IMT was 0.02±0.08 mm.4

Confirmation of Stroke and CHD

Medical records were reviewed by registered hospital physicians or research physicians who were blinded to the baseline data. Strokes were defined according to the US National Survey of Stroke criteria.17 For each stroke subtype (ie, cerebral infarction [thrombotic or embolic], intracerebral hemorrhage, and subarachnoid hemorrhage), a definite diagnosis was established based on the examination of computed tomographic scans, magnetic resonance images, or autopsies. Definite and probable myocardial infarctions were defined according to the criteria set out by the Monitoring Trends and Determinants in Cardiovascular Disease (MONICA) project.18 The criteria for a diagnosis of CHD included first‐ever acute myocardial infarction, sudden cardiac death within 24 hours after the onset of acute illness, or coronary artery disease followed by coronary artery bypass surgery or angioplasty. In the present study, we defined CVD as stroke or CHD.

Follow‐up Survey

For the detection of CHD and stroke occurrences, each participant's health status was checked during clinical visits to the NCVC every 2 years. Yearly questionnaires by mail or telephone were also completed by all participants. In addition, to complete our surveillance for fatal strokes and CHD, we conducted a systematic search for death certificates. All data were checked against medical records to confirm the incidence of CVD. When informed consent could not be obtained for a medical records survey (19.5%), we identified possible strokes or CHD using information from (1) questionnaires for verifying stroke and CHD at the health examination and/or (2) death certificates bearing a diagnosis of probable stroke or CHD. The end point of the follow‐up period for each participant was whichever of the following occurred first: (1) the date of the first diagnosis of CHD or stroke event; (2) the date of death; (3) the date of leaving Suita City; or (4) December 31, 2013. For each participant, the person‐years of follow‐up were calculated from the date of baseline survey to the first end point.

For all participants who attended the baseline carotid ultrasonographic survey between April 1994 and August 2001, carotid ultrasonographic follow‐up was performed every 2 years between April 1998 and March 2005. After the baseline survey, we performed carotid ultrasonography an average of 2.1±0.9 times (range: 1–6 times) using the same protocol. During the follow‐up for carotid ultrasonography measurement, we excluded the following participants: those with plaques at baseline (n=1156 for max‐CIMT >1.1 mm, n=1088 for max‐IMT >1.7 mm); those lost to follow‐up (n=804 for max‐CIMT >1.1 mm, n=822 for max‐IMT >1.7 mm); and those with incident CVD during echocardiography examination at follow‐up (n=42 for max‐CIMT >1.1 mm, n=46 for max‐IMT >1.7 mm). Finally, we could follow up participants (n=2722 for max‐CIMT >1.1, n=2768 for max‐IMT >1.7 mm) and observed new developing plaques (n=193 for max‐CIMT >1.1 mm, n=153 for max‐IMT >1.7 mm; Figure). The end point of the plaque follow‐up period for each participant was whichever of the following occurred first: (1) the date of the first developing plaque; (2) the date of the last carotid ultrasound examination; or (3) March 31, 2005 (censored). Subsequently, after the follow‐up of developing plaques, we set the end point of plaque follow‐up as the baseline of the CVD follow‐up. The end point of the follow‐up period for incident CVD was whichever of the following occurred first: (1) the date of the first diagnosis of a CHD or stroke event; (2) the date of death; (3) the date of leaving Suita City; or (4) December 31, 2013. For each participant, the person‐years of follow‐up for incident CVD were calculated from the end point of the plaque follow‐up and the incident CVD.

Figure 1

A schematic flow of 2 prospective study designs for carotid ultrasonography in the Suita Study. (A) Carotid IMT at the baseline survey and at follow‐up for incident cardiovascular disease. (B) Progression of carotid plaques (max‐CIMT >1.1 mm or max‐IMT >1.7 mm) and follow‐up for incident CVD. CVD indicates cardiovascular disease; IMT, intima–media thickness; max‐CIMT, maximum intima–media thickness in the common carotid arteries; max‐IMT, maximum intima–media thickness in the entire carotid artery area.

Statistical Analysis

We used ANOVA and χ2 tests to compare the mean values and frequencies according to incident CVD. We examined the associations between the quartiles of carotid IMT indexes and the risk of incident CVD and its subtypes using a sex‐ and age‐adjusted and multivariable‐adjusted Cox proportional hazards regression model. Multivariate analyses used age, sex, body mass index (overweight, normal weight, underweight), systolic/diastolic normal BP, prehypertension, and hypertension, total cholesterol (<160, 160–239, 240–279, >280 mg/dL), high‐density lipoprotein cholesterol (<35, 35–49, 50–59, >60 mg/dL), glomerular filtration rate (>60, 45–59, <45 mL/min per 1.73 m2), antihypertensive and antihyperlipidemic drug use, diabetes mellitus, impairment of fasting glucose, current smoking, and excessive drinking.

The current risk prediction model was directly calculated by employing each variable of the Suita Risk Score19 in the Cox regression model using the current data. After adding carotid plaques (max‐CIMT >1.1 mm or max‐IMT >1.7 mm) to the current risk model, we conducted statistical analysis to assess whether max‐CIMT or max‐IMT significantly improved 3 metrics used to determine CVD prediction: the C statistic, the net reclassification index, and the integrated discrimination index for the current risk model. Discrimination was assessed by Harrell C statistics, which were interpretable as the area under a receiver operating characteristic curve for comparing predictions in 2 outcome groups. A value of 1.0 indicates perfect predictive discrimination, whereas a value of 0.5 indicates no ability to discriminate. The 95% confidence intervals (95% CIs) of the C statistic were estimated using 200 bootstrap samples, then the standard error for the difference in the C statistics between each model was estimated from the bootstrap samples and used to calculate a z score and a P value for the difference. To compare the calibration that is the ability to estimate the accuracy of the model's prediction, we used the log‐likelihood ratio test and the Bayesian information criterion. Lower values for the Bayesian information criterion suggest better calibration of the model. We also examined the associations between the progression of carotid plaques per 5 years and the risk of incident CVD and its subtypes using sex‐ and age‐adjusted and multivariable‐adjusted Cox proportional hazards regression models.

Results

A total of 4724 participants with a mean age (±SD) of 59.7±11.9 years and no history of CVD at the baseline were followed for an average of 12.7 years; 2566 (54.3%) were women. The baseline characteristics of the study participants by incident CVD during follow‐up are summarized in Table 1. On average, the participants with CVD were older and had higher systolic and diastolic BP, body mass index, and prevalence of antihypertensive drug use, diabetes mellitus, and current smoking compared with participants without CVD during follow‐up.

Table 1

Baseline Characteristics of the 4724 Participants Without Evidence of CVD on Carotid Ultrasonography, According to CVD Status at Follow‐up

Characteristic	No CVD at Follow‐Up (n=4349)	CVD at Follow‐up (n=375)
Duration of follow‐up, y	13.1	8.1
Age, y	59.0±11.8	66.6±9.6
Sex (male), %	44.5	59.5
Body mass index, kg/m2	22.6±3.1	23.1±3.2
Systolic BP, mm Hg	126.3±19.4	137.4±19.4
Diastolic BP, mm Hg	78.2±10.8	81.8±10.5
Total cholesterol, mg/dL	209.1±33.2	215.9±32.3
HDL cholesterol, mg/dL	60.4±15.7	56.7±16.0
GFR, mL/min/1.73 m2	80.9±19.6	75.5±31.0
Antihypertensive drug use, %	14.6	27.2
Antihyperlipidemic drug use, %	5.2	5.3
Diabetes mellitus, %	4.0	10.1
Current smoking, %	23.2	29.1
Excessive drinking, %	8.2	9.3
IMT
Mean CCA thickness, mm	0.86±0.13	0.96±0.15
Maximum CCA thickness, mm	1.04±0.38	1.25±0.44
Maximum carotid artery thickness, mm	1.41±0.68	1.84±0.81
CCA thickness >1.1 mm, indicating plaque, %	19.6	44.5

Data are shown as mean±SD except as noted. To convert the values for cholesterol to mmol/L, multiply by 0.02586. BP indicates blood pressure; CCA, common carotid artery; CVD, cardiovascular disease; GFR, glomerular filtration rate; HDL, high‐density lipoprotein; intima–media thickness.

During the 59 909 person‐years of follow‐up, 375 incident CVD events (221 incident stroke and 154 incident CHD events) occurred. Table 2 provides the age‐ and sex‐adjusted and multivariable‐adjusted hazard ratios (HRs) for CVD and its subtypes according to quartiles of various IMT values. Compared with the first quartile of mean IMT, the age‐ and sex‐adjusted HRs for CVD and stroke indicated that the third and fourth quartiles of mean IMT values presented an increased risk of CVD and stroke and that those HRs for the fourth quartile of mean IMT values presented an increased risk of CHD; however, after multivariable adjustment, only the fourth quartile of mean IMT values presented an increased risk of CVD.

Table 2

Age‐ and Sex‐Adjusted and Multivariable‐Adjusted HRs for CVD and Subtype According to Quartiles of Various IMTs, Plaques, and per 1‐SD Increase in IMT

Mean IMT in the CCA	≤0.75 mm	0.76–0.85 mm	0.86–0.95 mm	>0.95 mm	>0.95 mm vs ≤0.95 mm	Per 1‐SD Increase
Participants at baseline, n	1042	1224	1337	1021	…	…
Person‐years	15 446	16 423	16 890	11 151	…	…
CVD, n	26	59	123	167	…	…
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	1.33 (0.83–2.13)	1.92 (1.21–3.04)	3.08 (1.90–4.97)	1.95 (1.56–2.44)	1.50 (1.35–1.67)
Multivariable‐adjusted HR (95% CI)	1 (Ref)	1.07 (0.67–1.72)	1.37 (0.86–2.18)	1.93 (1.18–3.13)	1.58 (1.26–1.98)	1.31 (1.17–1.48)
Stroke, n	14	34	78	95
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	1.30 (0.69–2.45)	1.96 (1.06–3.62)	2.74 (1.44–5.22)	1.84 (1.38–2.46)	1.40 (1.21–1.62)
Multivariable‐adjusted HR (95% CI)	1 (Ref)	1.06 (0.56–2.00)	1.47 (0.79–2.74)	1.88 (0.98–3.61)	1.55 (1.16–2.07)	1.25 (1.07–1.46)
CHD, n	12	25	45	72
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	1.37 (0.67–2.79)	1.88 (0.94–3.79)	3.86 (1.88–7.95)	2.25 (1.58–3.20)	1.68 (1.44–1.97)
Multivariable‐adjusted HR (95% CI)	1 (Ref)	1.05 (0.52–2.15)	1.19 (0.59–2.42)	2.04 (0.98–4.25)	1.71 (1.20–2.44)	1.45 (1.22–1.72)

Multivariable‐adjusted for age, sex, body mass index (overweight, normal weight, underweight), BP (normal, prehypertension, hypertension), total cholesterol (<160, 160–239, 240–279, ≥280 mg/dL), high‐density lipoprotein cholesterol (<35, 35–50, 50–59, ≥60 mg/dL), glomerular filtration rate (≥60, 45–59, <45 mL/min/1.73 m2), antihypertensive and antihyperlipidemic drug use, diabetes mellitus, impairment of fasting glucose, current smoking, and excessive drinking. The cutoff values of plaques for mean IMT at CCA, max‐CIMT, and max‐IMT were >0.95, 1.1, and 1.7 mm, respectively. CCA indicates common carotid artery; CHD, coronary heart disease; CI, confidence interval; CVD, cardiovascular disease; HR, hazard ratio; IMT, intima–media thickness; max‐CIMT, maximum intima–media thickness in the common carotid arteries; max‐IMT, maximum intima–media thickness in the entire carotid artery area; Ref, reference.

Regarding the quartiles of max‐CIMT, an increased risk of incident CVD was indicated by the third and fourth quartiles, an increased risk of stroke was indicated by the fourth quartile, and an increased risk of CHD was indicated by the second to fourth quartiles of max‐CIMT. After multivariable adjustment, the third and fourth max‐CIMT quartiles presented an increased risk of CVD and CHD, and the fourth Max‐CIMT quartile presented an increased risk of stroke. The third and fourth quartiles of max‐IMT showed increased risks of incident CVD and CHD. Mean CIMT >0.95 mm, max‐CIMT >1.1 mm, and max‐IMT >1.7 mm were all significantly associated with the risk of CVD in the multiple adjustment models.

Table 3 provides the various multivariable‐adjusted models values (mean IMT >0.95 mm, max‐CIMT >1.1 mm, and max‐IMT >1.7 mm) associated with CVD, stroke, and CHD and the difference of C statistics. When those cutoff points were added to the current risk prediction model, the differences of C statistics for CVD were all significantly increased. Moreover, when max‐CIMT >1.1 mm and max‐IMT >1.7 mm were added to the current risk prediction model, the difference of C statistics for stroke and CHD increased.

Table 3

Differences of C Statistic Values for the Current Risk Prediction Model of CVD, Stroke, and CHD by Various IMT Plaques

	C Statistic Differencea	P Values	z Score
CVD
Risk model	0
Risk model+mean IMT in the CCA >0.95 mm	0.0192 (0.0081–0.0304)	0.001	3.37
Risk model+max‐ CIMT >1.1 mm	0.0178 (0.0065–0.0291)	0.002	3.08
Risk model+max‐ IMT >1.7 mm	0.0233 (0.0112–0.0355)	<0.001	3.76
Stroke
Risk model	0
Risk model+mean IMT in the CCA >0.95 mm	0.0118 (−0.0009 to 0.0246)	0.069	1.82
Risk model+max‐CIMT >1.1 mm	0.0115 (0.0024–0.0206)	0.014	2.47
Risk model+max‐IMT >1.7 mm	0.0112 (0.0019–0.2050)	0.019	2.35
CHD
Risk model	0
Risk model+mean IMT in the CCA >0.95 mm	0.0107 (−0.0014 to 0.0229)	0.084	1.73
Risk model+max‐CIMT >1.1 mm	0.0120 (0.0010–0.0231)	0.032	2.14
Risk model+max‐IMT >1.7 mm	0.0199 (0.0051–0.0347)	0.008	2.65

CCA indicates common carotid artery; CHD, coronary heart disease; CVD, cardiovascular disease; IMT, intima–media thickness; max‐CIMT, maximum intima–media thickness in the common carotid arteries; max‐IMT, maximum intima–media thickness in the entire carotid artery area.

C statistic differences and the standard error between risk score model with and without each of the 3 IMT plaques were estimated from the bootstrap samples and used to calculate a z score and a P value for the difference.

Table 4 shows the reclassification of the current categories in the risk prediction model after the addition of carotid plaques. According to the current risk prediction model, which were calculated in these cases for the risk of CVD by max‐CIMT >1.1 mm and max‐IMT >1.7 mm, low risk indicated a risk of <6%, intermediate risk was 6% to 20%, and high risk was >20%. The net reclassification indexes for the current risk prediction model in max‐CIMT >1.1 mm and max‐IMT >1.7 mm were 4.8% (18/375) and 4.3% (16/375) for the participants with events, 1.1% (50/4349) and 1.6% (71/4349) for the participants without CVD events, and 6.0±2.0% and 5.9±2.1% overall, respectively. Further statistical analyses are shown in Table S2.

Table 4

Reclassification of the Current Risk Prediction Model Categories After Addition of Carotid Atherosclerosis

Original Risk Category	Current Risk Model Reclassification
	Low Risk	Intermediate Risk	High Risk
Max‐CIMT >1.1 mm
Participants with CVD events
Low risk, n	65	10
Intermediate risk, n	10	192	26
High risk, n		8	64
Participants without CVD events
Low risk, n	2204	99
Intermediate risk, n	174	1559	85
High risk, n		60	168
NRI±SE, %	6.0±2.0%, P=0.003
Max‐IMT >1.7 mm
Participants with CVD events
Low risk, n	63	12
Intermediate risk, n	11	191	26
High risk, n		11	61
Participants without CVD events
Low risk, n	2205	98
Intermediate risk, n	196	1539	83
High risk, n		56	172
NRI±SE, %	5.9±2.1%, P=0.005

CCA indicates common carotid artery; CVD, cardiovascular disease; IMT, intima–media thickness; max‐CIMT, maximum intima–media thickness in the common carotid arteries; max‐IMT, maximum intima–media thickness in the entire carotid artery area; NRI, net reclassification index.

After the new incident plaques in the CCA (max‐CIMT >1.1 mm) and the entire CA (max‐IMT >1.7 mm), we observed 112 and 117 incident CVD events (69 and 79 strokes and 43 and 38 CHDs) during 23 702 and 24 062 person‐years of follow‐up, respectively. Table 5 shows that after the new incident plaques (max‐CIMT >1.1 mm), the multivariable‐adjusted HRs for CVD and stroke were 1.95 (95% CI, 1.14–3.32) and 2.01 (95% CI, 1.01–3.99), respectively.

Table 5

HRs for Incident CVD, Stroke, and CHD by IMT Progression of Max‐CIMT (>1.1 mm) and Max‐IMT (>1.7 mm)

	Max‐CIMT			Max‐IMT
	≤1.1 mm	>1.1 mm	P Value	≤1.7 mm	>1.7 mm	P Value
Person‐years	21 771	1931		22 501	1561
CVD, n	90	22		106	11
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	2.00 (1.20–3.35)	0.008	1 (Ref)	1.15 (0.57–2.32)	0.705
Multivariable‐adjusted HR (95% CI)	1 (Ref)	1.95 (1.14–3.32)	0.014	1 (Ref)	1.08 (0.53–2.22)	0.837
Stroke, cases	58	11		72	7
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	2.04 (1.06–3.95)	0.034	1 (Ref)	1.20 (0.50–2.87)	0.676
Multivariable‐adjusted HR (95% CI)	1 (Ref)	2.01 (1.01–3.99)	0.047	1 (Ref)	NA
CHD, cases	32	11		34	4
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	2.05 (0.90–4.66)	0.088	1 (Ref)	1.03 (0.31–3.39)	0.967
Multivariable‐adjusted HR (95% CI)	1 (Ref)	1.80 (0.74–4.35)	0.195	1 (Ref)	NA

CHD indicates coronary heart disease; CI, confidence interval; CVD, cardiovascular disease; HR, hazard ratio; IMT, intima–media thickness; max‐CIMT, maximum intima–media thickness in the common carotid arteries; max‐IMT, maximum intima–media thickness in the entire carotid artery area; NA, not assessed (incalculable value given the small sample size); Ref, reference.

As shown in Table 6, compared with the first quartile of 5‐year progression of max‐CIMT, the multivariable‐adjusted HRs for incident CVD and stroke were 2.80 (95% CI, 1.54–5.11) and 2.30 (95% CI, 1.14–4.63) in the fourth quartile, respectively. The multivariable‐adjusted HRs for max‐CIMT 1 mm per 5 years were 2.89 (95% CI, 1.40–5.95) for CVD and 3.06 (95% CI, 1.19–7.87) for stroke. There was no association between progression of max‐IMT and CVD (data not shown). Supplemental results are expressed in Data S1.

Table 6

HRs of Incident CVD, Stroke, and CHD by the Quartiles of Progression of Max‐CIMT for 5 Years (Highest vs Lowest Quartiles) and Progression of Max‐CIMT per 1 mm at for 5 Years

	Quartiles of 5‐Year Progression of Max‐CIMT		5‐Year Progression of Max‐CIMT per 1 mm
	Lowest Quartile (< −0.12 mm/5 y)	Highest Quartile (>0.02 mm/5 y)	HR (95% CI)	P Value
CVD
Person‐years	5691	5124		…
Cases, n	19	39	…	…
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	2.80 (1.55–5.08)	2.92 (1.48–5.78)	0.002
Multivariable‐adjusted HR (95% CI)	1 (Ref)	2.80 (1.54–5.11)	2.89 (1.40–5.95)	0.004
Stroke
Cases, n	15	22
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	2.32 (1.16–4.64)	3.00 (1.24–7.24)	0.015
Multivariable‐adjusted HR (95% CI)	1 (Ref)	2.30 (1.14–4.63)	3.06 (1.19–7.87)	0.020
CHD
Cases, n	4	17
Age‐ and sex‐adjusted HR (95% CI)	1 (Ref)	5.14 (1.46–10.06)	2.81 (0.94–8.36)	0.064
Multivariable‐adjusted HR (95% CI)	1 (Ref)	NA	2.61 (0.79–8.63)	0.114

CHD indicates coronary heart disease; CI, confidence interval; CVD, cardiovascular disease; HR, hazard ratio; max‐CIMT, maximum intima–media thickness in the common carotid arteries; NA, not assessed (incalculable value given the small sample size; Ref, reference.

Discussion

Our findings demonstrated that carotid artery IMT, measured noninvasively with the use of carotid artery ultrasonography, is an independent predictor of new‐onset CVD events in a general Japanese population without a history of CVD at baseline. We observed that the max‐CIMT and max‐IMT, but not mean IMT, slightly but significantly improved classification for the prediction of incident CVD; this is the first such finding in a non‐Western country. In addition, the new progression of incident carotid plaque (max‐CIMT >1.1 mm) was shown to be associated with increased risks of CVD and stroke. To the best of our knowledge, this study is the first to reveal that new progression of incident carotid plaque is significantly associated with increased risk of incident CVD.

In the ARIC study, adding plaque and IMT to CHD risk prediction based on traditional risk factors improved CHD prediction, but the definition of plaque and IMT were not simple2; the IMT was defined as the mean of both the right and left sides of the distal CCA, the carotid artery bifurcation, and the proximal ICA, and plaque was defined as meeting 2 of the following 3 abnormal criteria: wall thickness (IMT >1.5 mm), shape, and wall texture. This measurement method seems to be complicated by screening of high risk for the development of CHD.

In the Framingham Offspring Study, the mean CCA IMT and the maximum ICA IMT were independent predictors of CVD events, but only the maximum ICA IMT (>1.5 mm) significantly improved the classification of CVD risk prediction.7 It can be said that the measurement method in the Framingham Offspring Study is simpler than that in the ARIC study. In the Framingham Offspring Study, Cox models were used by comparing the respective C statistics for the Framingham Risk Score20 before and after the addition of the variables for ICA IMT. In our present study, ICA IMT could not be evaluated in ≈10% of the participants. Compared with Western patients, the carotid bifurcation in Japanese patients is higher by ≈1 cervical vertebra.21 Consequently, we could not measure ICA IMT by carotid ultrasonography for ≈10% of our participants and could not use ICA IMT; instead, we used max‐IMT. Our analyses revealed that max‐CIMT >1.1 mm could be used to predict the incidence CVD, which significantly improved the classification of CVD risk prediction. Among the previous relevant Japanese population studies, there is a single study of the association between carotid IMT ≥1.1 mm (the highest quartile) and incident stroke.22 The US guidelines emphasize CVD risk as a high‐risk indicator when CCA IMT is >75th percentile.23, 24

To our knowledge, our study is the first to show that the progression of max‐CIMT >1.1 mm can be used to predict incident CVD, stroke, and CHD after that. In the PROG‐IMT Collaborative Project, there was no evidence of an association between individual IMT progression and the risk of subsequent CVD events,10 but the mean follow‐up periods for the scan interval of carotid ultrasonography and the mean CVD follow‐up periods after the second scan were only 3.7 and 6.9 years on average, respectively, which are shorter durations than that of our follow‐up study, at 5.1 years for the scan interval and 10.3 years for incident CVD. Our follow‐up period was not long enough; we believe that, empirically, the minimum follow‐up period for incident CVD is >10 years.

Compared with Western patients, Japanese patients have higher prevalence of hypertension, smoking, and excessive alcohol use,25 which are risk factors for incident stroke and CHD. The slope of the association between BP and CVD is steeper among Asian patients than Australasian patients.26 Systolic BP and smoking are more strongly associated with IMT in Asian patients.27 These characteristic backgrounds may have accentuated the association between the progression of plaque and incident CVD in our study.

Pathologically, IMT represents mainly hypertensive medial hypertrophy and thickening of smooth muscles in the media, which seem to be related to changes in local shear stress and represent partial arterial remodeling at the early stages of atherosclerosis.28 In contrast, carotid plaque is largely associated with traditional risk factors, namely, endothelial dysfunction, oxidative stress, smooth muscle cell proliferation, and an intimal process (with deposition of cholesterol, inflammation, and cell infiltration) at a later stage of atherogenesis.29, 30, 31, 32 Epidemiologically, mean IMT is more predictive of stroke than of CHD, and the reverse is true for carotid plaque,28, 29, 33 which is compatible with our data (Table 2).

This study has several strengths. First, this study was the first to investigate whether the development of carotid plaque is associated with incident stroke and CHD. The previous studies reported no association between the development of carotid plaque and incident stroke and CHD. Our present study covered a relatively long period and a relatively large sample size. Second, the same cohort of participants was examined by carotid ultrasonography every 2 years during follow‐up. Third, the same followed cohort members were randomly selected from the population registry of Suita City and stratified into groups by sex and 10‐year age increments in a premeditated manner. This cohort is representative of Japan's urban population, which composes >70% of the country's population. Fourth, the Suita Risk Score for CHD has just been adopted in the Japan Atherosclerosis Society “Guidelines for the Diagnosis and Prevention of Atherosclerotic Cardiovascular Diseases” in Japan, 2017 version.34 We use this guideline to evaluate the relationship between atherosclerosis and CVD prevention in Japan. The Suita Study is representative of CVD cohort studies in Japan. Fifth, the average number of follow‐up years for incident CVD after the progression of carotid plaque exceeded 10 years of follow‐up. The mean follow‐up period in the PROG‐IMT Collaborative Project was 6.9 years.10

There were also several study limitations. As mentioned, we did not use ICA IMT in this study because we could not measure all IMTs at the ICA, but we measured max‐IMT in the entire area as much as possible. Second, our data were obtained in a Japanese urban population, and although our findings might be useful for Asian communities, they might not be helpful for individuals of other ethnicities and racial backgrounds. Nonetheless, we observed that the progression of carotid plaque is associated with future CVD events. Our data showed that carotid plaque was more predictive of CHD than of stroke, as in previous studies.28, 29, 33 Further studies are needed to elucidate whether the progression of incident plaque at the CCA is associated with an increased risk of incident CVD. Third, the IMT values could not be divided into quartiles precisely because of the discrete values of IMT. For the evaluation of the relationship of quantity reactivity, we also examined the association of risks per 1 mm and 1 SD of IMT values with CVD. Fourth, we could not recruit both ≈23% of the attendees at medical examinations during the baseline survey (Table S1) and the carotid ultrasonographic follow‐up after the baseline survey (Tables S3 and S4). We found similar but slightly significant differences in clinical background between the study participants and nonparticipants at both the baseline and carotid ultrasonographic follow‐up. The main reasons for exclusion from this study were scheduling issues and unavailability of the ultrasonographic device by 1 specific physician rather than health problems. The participants in this study were randomly extracted at medical examination visits, but those who were not subjects for carotid ultrasonography were encouraged to be examined every 2 years during follow‐up. It is possible that high‐risk participants may have dropped out during follow‐up of cardiovascular events; therefore, it is possible that the association may be underestimated. Fifth, we do not have any standardized tools such as an arc when we measure carotid IMT. Although measurement errors cannot be avoided, a specific well‐trained physician performed the examinations and good reproducibility was obtained. Sixth, defining max‐CIMT values >75th percentile as plaque may be premature even though evidence shows that CVD risk increases with >75th percentile values. Incidental findings that may require further evaluation should be described such as a soft plaque, calcification finding, and meandering of the carotid artery. Seventh, we did not use the Suita Risk Score in this study because the risk score is based on a risk model for CHD events. Consequently, we directly conducted analyses by using each variable of the Suita Risk Score19 in the Cox regression model with the current data. Finally, we did not examine the types of antihypertensive and statin agents used by the participants. We put antihyperlipidemic drug use in the statistical models, and we did not observe an association with carotid atherosclerosis. When divided into 2 groups, whether controlled by taking antihyperlipidemic drugs or not, incident CVD and CHD were higher in the uncontrolled group (Table S5). Most antihyperlipidemic drugs used were statins, but they were not classified by type of statin agent, and the current study started at the same time as statin use in the early 1990s. Because the rate of statin use rose during the follow‐up periods (Table S6), it is necessary to analyze the time dependency of statins, but that is a subject for future research.

Conclusions

High mean CIMT and carotid plaque in the CCA and entire carotid artery area were risk factors for CVD in a general Japanese population. Among the carotid ultrasonographic indexes, max‐CIMT >1.1 mm was revealed to be associated with increased risks of CVD, stroke, and CHD by the C statistic improvement and the reclassification of the current risk prediction model. New progression of incident carotid plaque (max‐CIMT >1.1 mm) was shown to be associated with an increased risk of CVD. In medical examinations and outpatient clinics, carotid ultrasonography at the CCA is easily measured at screening, and thus carotid plaque (max‐CIMT >1.1 mm) could be evaluated to help prevent increased risk of future CVD. We plan to develop a risk score for carotid atherosclerosis using the Suita Study population shortly, with the ultimate goal of using this score in medical examinations and outpatient clinics as a preventive screening measure for carotid atherosclerosis without the requirement of carotid ultrasonography.

Sources of Funding

This study was supported by grants‐in‐aid from the Ministry of Health, Labor, and Welfare of Japan: Challenging Exploratory Research (grant no. 17K19873) and a Grant‐in‐Aid for Scientific Research B (no. 16H05252), and the Intramural Research Fund of the National Cerebral and Cardiovascular Center (22‐1‐2) in Japan.

Disclosures

None.

Data S1.

Table S1. Baseline Characteristics of Current Study Participants, Those Excluded From This Study, and Those With Previous Cardiovascular Disease

Table S2. Comparison of Risk Prediction Models With Maximum Intima–Media Thickness in the Common Carotid Arteries >1.1 mm or in the Entire Carotid Artery Area >1.7 mm in the Current Study

Table S3. Baseline Characteristics of Participants Followed for Maximum Intima–Media Thickness in the Common Carotid Arteries (Max‐CIMT), Those Lost to Follow‐up, and Those With Incident Cardiovascular Disease During Follow‐up of Max‐CIMT

Table S4. Baseline Characteristics of Participants Followed for maximum intima–media thickness in the entire of carotid artery area (Max‐IMT), Those Lost to Follow‐up, and Those With Incident Cardiovascular Disease During Follow‐up of Max‐IMT

Table S5. Cross‐Sectional Study of the Risks of Plaques and Prospective Study of the Risks of Cardiovascular Disease, Stroke, and Coronary Heart Disease According to the Non‐HDLC Categories

Table S6. Characteristics of the Participants at Baseline and End Point Surveys During Carotid Echo and Changes of Characteristics by 5‐Year Follow‐up Periods

Click here for additional data file.