Directory UMM :Data Elmu:jurnal:A:Atherosclerosis:Vol150.Issue2.Jun2000:

(1)

Correlations between measures of atherosclerosis change using

carotid ultrasonography and coronary angiography

Wendy J. Mack

a,b,

_{*, Laurie LaBree}

a,b

_{, Chao-Ran Liu}

a,c

_{, Chi-Hua Liu}

a,c

_,

Robert H. Selzer

a,d

_{, Howard N. Hodis}

a,b,c

a_{Atherosclerosis Research Unit,}_Uni₆_{ersity of Southern California,}_{Los Angeles,}_CA,_USA

b_{Departments of Pre}₆_enti₆_{e Medicine,}_Uni₆_{ersity of Southern Califiornia,}₁₅₄₀_Alcazar,_CHP_-₂₁₈_,_{Los Angeles,}_CA₉₀₀₈₉_,_USA c_{Department of Medicine,}_Uni₆_{ersity of Southern California,}_{Los Angeles,}_CA,_USA

d_{CalTech Jet Propulsion Laboratory,}_Pasadena,_CA,_USA

Received 8 February 1999; received in revised form 7 July 1999; accepted 3 September 1999

Abstract

Few studies have examined the correlation between change in carotid artery intima-media thickness (IMT) and change in coronary artery disease. In the Cholesterol Lowering Atherosclerosis Study, current nonsmoking men with coronary artery disease were randomized to colestipol-niacin or placebo. Among 133 subjects with baseline and on-trial coronary angiography and carotid ultrasonography, colestipol-niacin treatment significantly reduced progression of atherosclerosis by both end point measures (2-year average change in percent diameter stenosis by coronary angiography and rate of change in carotid IMT). Significant correlations between change in common carotid artery IMT and quantitative coronary angiographic measures of change were evident over all coronary artery lesions, and in mild/moderate (B50% diameter stenosis), but not severe (]50% diameter

stenosis) coronary artery lesions. In mild/moderate lesions, correlations with change in common carotid IMT were: percent diameter stenosis (r=0.28, P=0.002), minimum lumen diameter (r= −0.28, P=0.002), and vessel edge roughness (r=0.25,

P=0.003). While measures obtained by carotid ultrasonography and coronary angiography are correlated, they each assess different aspects of atherosclerosis change. © 2000 Elsevier Science Ireland Ltd. All rights reserved.

Keywords:Atherosclerosis; Angiography; Ultrasonography; Coronary disease; Carotid arteries

www.elsevier.com/locate/atherosclerosis

1. Introduction

Assessment of carotid artery intima-media thickness (IMT) by B-mode ultrasound is increasingly utilized as a measure of atherosclerosis, both in epidemiological studies examining the course of atherosclerotic disease and correlations with CHD risk factors [1,2] as well as in clinical trials testing the efficacy of various anti-atherosclerosis therapies [3 – 9]. Determination of carotid artery IMT allows for direct, non-invasive, re-peated measurement of early, pre-intrusive atheroscle-rosis in asymptomatic as well as symptomatic individuals [10].

The association between carotid IMT and the extent of coronary artery disease has been addressed primarily through cross-sectional analyses. These reports either correlated carotid artery IMT and various measures of coronary artery disease obtained by coronary angiogra-phy or compared carotid artery IMT in persons with (typically defined by having a lesion of at least 50% diameter stenosis) and without established coronary artery disease [11 – 19]. No studies have examined the correlation between serial measures of both carotid artery IMT and coronary artery disease. Few quantita-tive imaging studies have compared changes in atherosclerosis in one vascular bed with that of another [20].

The Cholesterol Lowering Atherosclerosis Study (CLAS) was a serial arterial imaging clinical trial that tested the efficacy of LDL-lowering in reducing the rate of progression of atherosclerosis in the coronary, * Corresponding author. Tel.: +1-323-4421820; fax: +

1-323-4422993.

E-mail address:[email protected] (W.J. Mack)

(2)

carotid, and femoral arteries [21]. We have previously reported on the treatment benefit of colestipol-niacin on common carotid artery IMT among CLAS subjects and on the correlation of IMT measured in the common carotid artery with visual assessment of coronary artery lesion severity [3,4]. We have also reported that both common carotid IMT rate of change and change in coronary artery disease assessed by quantitative coro-nary angiography (QCA) are significantly predictive of clinical coronary events among the CLAS cohort [22,23]. We now report on the association between common carotid artery IMT measured by automated computerized edge tracking and various angiographic measures of coronary artery disease determined by QCA. We examined these associations at a single time point (at baseline) as well as serially (change in end point measures).

2. Methods

2.1. CLAS study design

CLAS was a randomized, double-blind, placebo-con-trolled serial arterial imaging clinical trial testing the efficacy of LDL-cholesterol lowering in reducing the rate of progression of atherosclerosis. Subjects were 188 non-smoking males, 40 – 59 years old, with prior coro-nary artery bypass graft surgery who were randomized to colestipol-niacin or placebo [21]. All subjects pro-vided written informed consent for participation in the trial, which was approved and monitored by the univer-sity IRB.

2.2. Coronary angiography and study end points

Coronary angiograms were obtained at baseline and 2 years using the percutaneous femoral technique with sufficient right and left anterior oblique views to clearly visualize all coronary artery lesions. After completion of the 2-year angiogram, quantitative coronary angiog-raphy (QCA) was performed blinded to treatment group assignment. A panel of expert angiographers identified the coronary artery lesions prior to evalua-tion by the QCA technician. Film pairs were processed in tandem with dual projectors to match frames for orientation and degree of contrast filling.

For all coronary artery lesions evaluable by QCA, percent diameter stenosis and minimum lumen diameter were measured. For all evaluable coronary artery seg-ments, defined from branch to branch, the percent diameter stenosis, minimum lumen diameter, average diameter, and vessel edge roughness were computed. The vessel edge roughness for a coronary artery seg-ment was defined as the root mean squared deviation of the measured vessel edge from a straight line fit through

the segment vessel edge profile. Three sequential frames exposed during end-diastole were digitized when possi-ble; otherwise three sequential frames from other phases of the cardiac cycle were used. The QCA mea-sures were averaged over the three frames for each lesion or segment evaluated [24] and were obtained in all evaluable native coronary arteries. For each subject, summary QCA measures were computed by averaging over all measured lesions or segments.

Other measures of change in coronary artery atherosclerosis which might be considered more clini-cally relevant were determined from the QCA data. A new lesion was defined as a lesion that was not iden-tified on the baseline angiogram, but was at least 20% diameter stenosis on the 2-year angiogram. A progress-ing lesion was defined as a lesion of at least 20% diameter stenosis on the baseline angiogram, which had showed an increase of at least 12% diameter stenosis on the 2-year angiogram. In a reproducibility study of measures obtained by our QCA methodology, 17 pa-tients (measuring 49 coronary artery segments and 55 coronary artery lesions) received two angiograms dur-ing the same examination, with repositiondur-ing between the two angiograms [25]. The standard deviation of the difference in lesion percent diameter stenosis, an esti-mate of the inherent measurement error, was 6%. We therefore chose twice this standard deviation as a cutoff for defining ‘true’ progression (or regression) of coro-nary artery lesions (i.e. lesion change beyond that ex-pected only by measurement error). A regressing lesion was defined as a lesion of at least 20% diameter stenosis at baseline, which had decreased by at least 12% diame-ter stenosis at 2 years. A new total occlusion was an occluded vessel at 2 years, which was not occluded at baseline. The number of each of these QCA-derived measures was computed for each subject. The number of changing lesions for each subject was computed as the difference between the number of progressing and regressing lesions.

2.3. Carotid ultrasonography and measurement of

common carotid IMT

Subjects were scanned with a Biodynamics Biosound ultrasound system using a 9-MHz probe. Carotid artery ultrasounds were performed prior to randomization and during the trial.

Ultrasound images were analyzed at end diastole with a Northgate 386/33-computer system equipped with a Data Translation DT 2862 digitizing and image processing board. The process of digitizing the ultra-sound image, tracking the lumen-intima and media-ad-ventitia echoes, and computing IMT by automated edge detection (PROSOUND software) has been previ-ously described [26]. In brief, the lumen-intima and media-adventitia echoes are located with an edge

(3)

track-ing algorithm in which: (1) an initial echo boundary is identified with a mouse; (2) this initial boundary is used to guide an edge-finding algorithm to locate a condi-tional set of edges, and (3) these condicondi-tional edges are tested for ‘edge strength’ and weak edges are elimi-nated. Using all acceptable edge pairs (lumen-intima to media-adventitia), the IMT is then computed as the average separation between the two edges.

An operator who selected one end-diastolic frame analyzed videotapes from a 5 to 10-s recording using maximum lumen-intima continuity and echo brightness as selection criteria. Average IMT was computed over a 1 cm length of the far wall of the common carotid artery ending approximately 1/2 centimeter proximal to the transition between the common and bulb regions. For all subjects, care was taken to match the longitudi-nal position of the alongitudi-nalyzed segment on serial images. In a short-term reproducibility study of our ultra-sound and IMT measurement methodology, eight sub-jects were scanned twice (1 week apart) by three ultrasonographers [26]. The inter-sonographer standard deviation of carotid artery IMT was 0.026 mm for scans performed on the same day. The intra-sonogra-pher standard deviation was 0.016 mm for replicate scans obtained on the same day, and 0.029 mm for replicate scans obtained 1 week apart.

2.4. Statistical analysis

QCA variables selected for analysis were considered standard angiographic measures of extent and change in coronary artery disease by virtue of their prior use as end points in coronary angiographic trials. Because of the different progression rates observed in native coro-nary arteries as compared with corocoro-nary artery bypass grafts [27], each QCA measure was evaluated over native artery lesions/segments only. Analyses were also limited to native artery lesions/segments that were hemodynamically unrelated to bypass grafts; these analyses included only lesions or segments in vessels that had not been bypassed. In addition, because the predictability of coronary events by change in coronary artery disease is related to initial lesion severity [22], each QCA measure was evaluated according to two categories of baseline severity: B50% diameter stenosis (mild/moderate) and ]50% diameter stenosis (severe). Both baseline and on-trial changes (2-year minus base-line) for these QCA measures were analyzed. Analyses of changes in coronary artery disease also included the QCA-derived measures of numbers of new lesions, pro-gressing lesions, repro-gressing lesions, changing lesions, and new total occlusions.

Two carotid IMT measures were analyzed: baseline IMT and rate of change over the 2-year study period. The IMT change rate was estimated by regressing each subject’s IMT data on time since randomization; the

resulting slope estimate was used as the subject’s IMT change rate (in mm/year).

Associations between QCA and IMT variables were evaluated using correlational analyses. Nonparametric Spearman rank correlations were used for all analyses because of the non-normal distributions of some of the QCA variables. To estimate the degree of association at a single point in time, baseline variables were analyzed using the entire sample (combined treatment groups). Associations between coronary angiographic versus carotid IMT changes were also estimated by correlating QCA and IMT change variables in a combined analy-sis. To further evaluate the common carotid IMT change rate as an indicator of atherosclerosis change relative to QCA end point measures, simple correla-tions of on-trial lipids were correlated with the common carotid IMT change rate and the average change in percent diameter stenosis of coronary artery lesions. Finally, the effect of randomized treatment on each atherosclerosis outcome measure (common carotid IMT change rate and change in percent diameter steno-sis of coronary artery lesions) was analyzed using t -tests for independent samples.

3. Results

3.1. Sample characteristics and treatment effects Of the 188 subjects randomized, 19 (10%) had neither an ultrasound or QCA end point, 23 (12%) had a QCA but no ultrasound end point, and 13 (7%) had an ultrasound but no QCA end point. The sample for analysis was composed of 133 subjects who had both baseline/2-year coronary angiograms evaluated by QCA and baseline and follow-up carotid ultrasonogra-phy (Table 1). Subjects were primarily middle-aged Table 1

Baseline sample characteristics (n=133)

Mean (SD) or % (n)

Age at randomization 54.1 (4.5) Total cholesterol (mmol/l) 6.3 (0.9) HDL-cholesterol (mmol/l) 1.1 (0.2) 4.4 (0.8) LDL-cholesterol (mmol/l)

Total triglycerides (mmol/l) 1.7 (1.0) 118.6 (18.4) Apolipoprotein A-I (mg/dl)

Apolipoprotein B (mg/dl) 125.5 (27.8) 121.4 (12.8) Systolic blood pressure (mmHg)

79.6 (8.6) Diastolic blood pressure (mmHg)

71% (94) Ex-smokera

Treatment group

Drug 50% (67)

50% (66) Placebo

a_{The CLAS design required all subjects to be current non-smokers;}

(4)

males who were on average normotensive with moder-ate hypercholesterolemia; 71% were ex-smokers. QCA measures were computed on an average of 5.7 native coronary artery lesions and 7.7 native coronary artery segments (defined from branch to branch) per subject. Carotid IMT was obtained from an average of 2.8 serial ultrasonograms per subject, with an average du-ration of 2 years from first to last carotid ultrasound. These 133 subjects did not differ from the 55 subjects not included in this analysis on age, years since bypass, smoking history, randomized treatment group, or base-line and on-trial levels of blood pressure, total choles-terol, HDL-cholescholes-terol, LDL-cholescholes-terol, and total triglycerides (data not shown).

Among these 133 subjects with both coronary angio-graphic and carotid IMT end point measures, the treat-ment effect on change in coronary artery lesion percent diameter stenosis was statistically significant (mean

(SEM) change in colestipol – niacin group=0.39 (0.75) percent diameter stenosis, in placebo group=2.87 (0.71) percent diameter stenosis, P=0.02). The treat-ment effect on common carotid IMT change rate was also statistically significant (mean (SEM) IMT change rate= −0.024 (0.004) mm/year in the colestipol – niacin group, and 0.020 (0.003) mm/year in the placebo group, PB0.0001). Treatment effect sizes (difference in treat-ment group means, divided by the pooled standard deviation) were 0.42 for change in percent diameter stenosis, and 1.62 for the IMT change rate.

3.2. Correlations between common carotid artery IMT

and QCA measures at baseline

Table 2 provides the Spearman correlations between a single baseline carotid IMT measure and QCA mea-sures obtained from the baseline coronary angiogram.

Table 2

Correlations between carotid IMT and QCA measuresa

QCA measure Correlation (95% CI)b_{of baseline QCA} Correlation (95% CI) of QCA change measure measure with baseline carotid IMT (mm) with carotid IMT change rate (mm/year)

Percent diameter stenosis(%S)

−0.04 (−0.22, 0.14) 0.21 (0.04, 0.38)§

All native lesions (n=129)

0.01 (−0.17, 0.19)

Native lesionsB50%S (n=125) 0.28 (0.10, 0.43)

−0.03 (−0.23, 0.19)

Native lesions]50%S (n=93) 0.00 (−0.21, 0.21)

−0.03 (−0.24, 0.18)

Unrelated native lesions (n=93)c _{0.33 (0.13, 0.50)}

All native segments (n=133) _{0.16 (}₋_{0.01, 0.33)} 0.28 (0.11, 0.44)¶

Native segmentsB50%S (n=133) 0.14 (−0.04, 0.31) 0.30 (0.13, 0.45)¶ −0.08 (−0.30, 0.14) 0.07 (−0.15, 0.28) Native segments]50%S (n=85)

0.03 (−0.16, 0.23)

Unrelated native segments (n=107)c _{0.25 (0.06, 0.42)}§

Minimum lumen diameter

−0.03 (−0.21, 0.14)

All native lesions (n=129) −0.22 (−0.38,−0.04)§

Native lesionsB50%S (n=125) −0.07 (−0.25, 0.11) −0.28 (−0.44,−0.11) −0.02 (−0.22, 0.20)

Native lesions]50%S (n=93) 0.01 (−0.20, 0.22)

Unrelated native lesions (n=93)c _{0.04 (}₋_{0.17, 0.25)} ₋_{0.20 (}₋_{0.40, 0.01)}§ −0.15 (−0.32, 0.02)

−0.09 (−0.26, 0.08) All native segments (n=133)

−0.07 (−0.25, 0.10)

Native segmentsB50%S (n=133) −0.17 (−0.34, 0.00)§

Native segments]50%S (n=84) 0.06 (−0.16, 0.27) −0.03 (−0.25, 0.20) 0.12 (−0.08, 0.31) −0.08 (−0.27, 0.12) Unrelated native segments (n=107)c

Vessel edge roughness

All native segments (n=133) 0.10 (−0.08, 0.27) 0.20 (0.02, 0.36)§

0.17 (0.00, 0.34)§ _{0.25 (0.08, 0.41)} Native segmentsB50%S (n=133)

−0.10 (−0.31, 0.12)

Native segments]50%S (n=84) −0.07 (−0.29, 0.15)

Unrelated native segments (n=107)c _{0.17 (}₋_{0.02, 0.36)} _{0.25 (0.06, 0.43)}

A6erage segment diameter

−0.04 (−0.22, 0.13) 0.01 (−0.17, 0.18) All native segments (n=133)

Native segmentsB50%S (n=133) −0.05 (−0.23, 0.12) −0.01 (−0.18, 0.17) 0.03 (−0.19, 0.25) 0.05 (−0.17, 0.27) Native segments]50%S (n=84)

0.12 (−0.08, 0.31)

Unrelated native segments (n=107)c _{0.13 (}₋_{0.07, 0.31)}

a_{Each QCA measure represents the value (baseline or 2-year change) averaged over all evaluable lesions}_/_{segments in native coronary arteries.} b_{95% CI}₌_{95% confidence interval.}

c_{Analysis includes only native lesions}_/_{segments hemodynamically unrelated to bypass graft (lesion-based analyses}_n₌_{93 subjects;}

segment-based analysesn=107 subjects).

§_P_B_0.05.

_PB0.01.

(5)

QCA measures represented in Table 2 were obtained from lesions/segments in native arteries only. Except for a positive correlation with the average vessel edge roughness in native coronary artery segments of less than 50% diameter stenosis, there were no significant correlations between common carotid IMT and base-line QCA measures. Results were similar when analyses were conducted over all native lesions (or segments) or limited to native lesions (segments) hemodynamically unrelated to bypass grafts. Correlations of common carotid IMT at baseline with numbers of coronary artery lesions evident at baseline (at least 20% diameter stenosis) were: all native coronary artery lesions (r= 0.11); severe native coronary artery lesions of at least 50% diameter stenosis (r=0.02); mild/moderate native coronary artery lesions of 50% diameter stenosis or less (r=0.15) (all correlations not significantly different from zero).

3.3. Correlations between serial measures of common

carotid artery IMT and QCA

QCA measures of change in coronary artery disease were correlated with the rate of change of common carotid IMT (slope of IMT versus time; Table 2). These correlations were evident when assessed in all native lesions, in native lesions hemodynamically unrelated to bypass grafts, and in mild/moderate lesions. QCA mea-sures assessing change in severe lesions (]50% diame-ter stenosis) were not related to the rate of change of carotid IMT. The associations reached statistical signifi-cance for QCA measures of percent diameter stenosis (measured both over coronary artery lesions and ments), minimum lumen diameter (in lesions and seg-ments), and vessel edge roughness. Change in vessel edge roughness was positively correlated with common carotid IMT rate of change, with the correlation most

evident in mild/moderate lesions in native vessels (r= 0.25, P=0.003). In contrast to these measures of change in coronary artery disease, the change in the average lumen diameter (averaged over all coronary artery segments) was not correlated with the carotid IMT rate of change.

3.4. Correlations between serial measures of common

carotid artery IMT and QCA deri6ed measures

The common carotid IMT change rate was not corre-lated to the number of new coronary artery lesions (r=0.02, 95% CI= −0.15, 0.20) or new total occlu-sions of the native coronary arteries (r=0.04, 95% CI= −0.13, 0.21). Carotid IMT change was non-significantly correlated with the number of progressing (r=0.17, 95% CI= −0.01, 0.33) and regressing (r= −0.15, 95% CI= −0.32, 0.03) native coronary artery lesions. The summary variable of total number of changing lesions, which was the number of progressing lesions minus the number of regressing lesions, was positively and significantly correlated with the common carotid IMT change rate (r=0.24, 95% CI=0.07, 0.40, P=0.005).

3.5. Associations between common carotid IMT, QCA,

and lipids

Correlations between the average lipid levels mea-sured on-trial with the common carotid IMT change rate and change in percent diameter stenosis of coro-nary artery lesions are displayed in Table 3. For each lipid, the row displays the simple (unadjusted) correla-tion with change in carotid IMT (or coronary artery percent diameter stenosis). All correlations are in the expected direction and are statistically significant. Sim-ple correlations of on-trial lipids with change in coro-nary artery lesions ranged from 0.19 (triglycerides) to 0.29 (LDL-cholesterol). Lipid correlations with the common carotid IMT change rate ranged from 0.18 (triglycerides) to 0.44 (LDL-cholesterol).

4. Discussion

Data from CLAS indicate that, over a 2-year study period, the rate of change of common carotid artery IMT as a noninvasive measure of early preintrusive atherosclerosis is positively correlated with changes in coronary artery atherosclerosis determined by a variety of QCA measures. The relationships with common carotid artery IMT rate of change were most apparent in mild/moderate, but not severe coronary artery le-sions. Previous studies evaluating a single measure of IMT have shown that carotid artery IMT is higher on average in hypertensive compared to normotensive sub-Table 3

Correlations between common carotid IMT change rate, change in coronary artery lesion percent diameter stenosis, and on-trial lipidsa

Change in coronary artery Common carotid IMT

change rate lesions: percent diameter stenosis

On-trial le6els of

0.26 (0.10, 0.42)‡

0.39 (0.24, 0.53)§

Total cholesterol

0.29 (0.12, 0.43)§

LDL- 0.44 (0.30, 0.57)§

cholesterol

HDL- −0.35 (−0.49,−0.19)§ ₋_{0.26 (}₋_0.41,₋_0.09)‡

cholesterol

0.19 (0.02, 0.35)†

0.18 (0.01, 0.34)†

Triglycerides

a_{Numbers in table are correlations (95% confidence intervals).} †_P_B_0.05.

‡_P_B_0.01. §_P_B_0.001.

(6)

jects [28], in persons with carotid artery lesions [29], in persons undergoing coronary angiography for chest pain [15], in persons with coronary heart disease and stroke [14], and in asymptomatic persons with a posi-tive exercise stress test [30].

Generalizability and interpretation of these results are possibly limited by the fact that all subjects had undergone coronary artery bypass surgery. Changes in disease of bypass grafts may not be representative of atherosclerosis changes in the unbypassed native coro-nary arteries. For this reason, we limited analyses to native coronary arteries that were unrelated to bypass grafts; results from these analyses were consistent with the broader analyses using all native coronary arteries (Table 2).

Except for coronary vessel edge roughness, we found no correlation between baseline QCA measures and common carotid artery IMT. These results are in con-trast to previous reports, which have shown associa-tions between cross-sectional measures of carotid artery IMT and percent diameter stenosis used as a coronary angiographic measure of atherosclerosis [11 – 19]. While the majority of previous studies have used more global visual measures of extent of coronary artery disease, such as the number of diseased vessels, only one study has previously used QCA measures of lesion severity and segment size [19].

Inclusion criteria for the CLAS study limited subjects to those with coronary artery disease severe enough to warrant bypass graft surgery. The resulting reduction in variability in coronary artery disease relative to a popu-lation of subjects presenting for coronary angiography was a possible factor in reducing the observed correla-tions between baseline measures of coronary artery disease and carotid artery IMT. Previous cross-sec-tional correlation studies have included subjects who had undergone coronary angiography for clinical indi-cations [11 – 19]. The resulting samples included persons with coronary artery disease (with angiographic evi-dence of a coronary artery lesion of a pre-defined severity) and without coronary artery disease (subjects with lesions less than the pre-defined severity). Al-though these studies were therefore more likely than the present study to detect cross-sectional associations be-tween coronary artery disease and carotid artery IMT, the definition of coronary artery disease was typically (but not always) limited to lesions of at least 50% diameter stenosis. In other words, in these cross-sec-tional and case-control studies that used such a defini-tion, individuals with lesions of less than 50% diameter stenosis were considered to have no coronary artery disease (‘normal’ angiograms), and were used as the comparison group with subjects with more severe le-sions. It is now well recognized that lesions of less than 50% diameter stenosis do not represent normal coro-nary anatomy. To the contrary, progression of this

class of lesions is most closely related to the risk of clinical coronary events [22].

The present report from CLAS is one of only a few to evaluate the relationship between change in atherosclerosis, determined by IMT in the carotid artery and by QCA in the coronary arteries. In contrast to our current findings, a clinical trial of pravastatin reported no significant correlations between the change in carotid IMT and QCA measures of change in per-cent diameter stenosis, change in average lumen diame-ter, and change in minimum lumen diameter [31]. All QCA measures were obtained over all evaluable coro-nary artery segments, without regard to severity of disease of the segment. Comparison of results of these two studies is hampered by the fact that the prior study did not report the correlation coefficients [31]. How-ever, in CLAS, QCA change measures evaluated over all coronary segments showed lower correlations with rate of change of common carotid artery IMT than QCA change measures evaluated over segments with lesions (Table 2). In addition, significant correlations were noted only in lesions of mild/moderate severity and not in severe lesions of greater than 50% diameter stenosis.

The correlations between the rate of change of com-mon carotid artery IMT and QCA measures of coro-nary artery atherosclerosis change ranged from 0.17 to 0.33. These correlations are of a comparable magnitude to those reported with cross-sectional data between the carotid and coronary arteries and within carotid artery sites. Representative of the several studies using visual assessment of coronary artery disease [11 – 18], a corre-lation of 0.34 was reported between a summary mea-sure of maximal IMT over several carotid artery sites and the presence of a coronary artery lesion of at least 50% diameter stenosis [32]. Studies evaluating changes in carotid artery lesions, rather than IMT, in relation to progression of coronary artery disease have found larger correlations of around 0.41 [20]. In addition, the histological age-adjusted correlations between carotid and coronary artery atherosclerosis in a large autopsy study ranged from 0.29 to 0.36 [33]. In agreement with these histological correlations was our finding that common carotid IMT and coronary artery vessel edge roughness were significantly correlated (r=0.25), since computer measures of angiographic edge roughness correlate significantly with vessel wall cholesterol con-tent (r=0.70, PB0.001) [34]. Further, a previous re-port from the CLAS study demonstrated that carotid artery IMT is significantly correlated with carotid an-giographic vessel edge roughness (r=0.31, PB0.05) [3]. It is also notable that the magnitude of the correla-tions between common carotid artery IMT and QCA measures of coronary artery atherosclerosis change re-ported here are completely within the range of correla-tions observed in CLAS for lipid and lipoprotein risk factors and coronary artery atherosclerosis (Table 3).

(7)

Using data from the same CLAS cohort, we have reported that the rate of change of IMT in the common carotid artery is significantly associated with the risk of incident clinical coronary events over a 10-year follow-up period [23]. In several epidemiologic studies, a single measure of carotid IMT was related to the risk of subsequent cardiovascular events (including fatal and nonfatal myocardial infarction and stroke) among male and female subjects free of cardiovascular disease at their IMT measurement [35 – 38]. In the CLAS cohort, for each 0.03 mm increase per year in common carotid artery IMT, the relative risk (RR) for nonfatal MI or coronary death was 2.2 (95% CI=1.4, 3.6), and the RR for nonfatal MI, coronary death, or coronary artery revascularization was 3.1 (95% CI=2.1, 4.5) [23]. In the same post-trial follow-up of CLAS subjects using QCA measures of percent diameter stenosis and mini-mum lumen diameter, changes in mild/moderate lesions (B50% diameter stenosis), but not severe lesions (] 50% diameter stenosis), was predictive of subsequent clinical coronary events [22]. These results are of partic-ular interest given our current finding that the correla-tion between changes in common carotid IMT and coronary artery disease is most pronounced with pro-gression of mild/moderate lesions, the very lesions pre-dictive of clinical coronary events. Current understanding is that mild/moderate lesions lead to clinical coronary events through plaque rupture with sudden vascular occlusion [39].

When considered multivariately, both carotid IMT rate of change and QCA change in percent diameter stenosis were independently associated with the risk of coronary events [23]. The positive but weak correlations between change in common carotid artery IMT and coronary angiographic measures of atherosclerosis also suggest that the two methods, while capturing certain common aspects of the atherosclerotic process, sepa-rately reflect unique aspects of atherosclerosis progres-sion. While visualization and quantification of the intima-media complex has been proposed to be a mea-sure of vessel wall changes early in the atherosclerotic process [10], coronary angiography can only indirectly measure the degree of atherosclerosis by assessing resid-ual lumen size. Coronary angiography can therefore only quantify atherosclerosis at the relatively later stage of intrusive lesion formation. This is consistent with observations of higher correlations between angio-graphic measures of coronary artery disease and carotid artery lesions as compared to carotid artery IMT [20]. In conjunction with the greater precision and reliability of IMT measures compared to QCA measures, it also explains why the carotid IMT change rate was a more sensitive measure of the effect of colestipol – niacin treatment (treatment effect size=1.62) than was change in percent diameter stenosis (treatment effect size= 0.42). Taken together, these data indicate that common

carotid artery IMT incorporates additional, indepen-dent information on prediction of coronary events be-yond the angiographic measurement of lumen narrowing. This is perhaps illustrated by the relation-ship between common carotid artery IMT and coro-nary vessel edge roughness which may represent atherosclerotic burden and the propensity for plaque rupture and clinical coronary events. These results sug-gest that use of both IMT and QCA change end points may be particularly useful multiple end points for clini-cal trials of anti-atherosclerosis interventions with the ultimate goal of reducing coronary event risk. This is of particular importance since the ability to solely use coronary angiographic end points to test anti-atherosclerosis therapies has become a great challenge with the current use of aggressive mechanical interven-tion for cardiac symptoms. Intervening revasculariza-tion procedures have profound effects on follow-up coronary angiograms, and in most cases render the follow-up angiographic data useless as a comparison to a pre-intervention angiogram. Even the effects of coro-nary angioplasty, once thought to be limited to the intervened coronary artery segment, may effect remote vessel segments through vascular remodeling, thereby altering the true angiographic effects of an anti-atherosclerosis medical therapy [40].

Higher correlations with measures of coronary artery disease have been noted for IMT measured in the carotid bulb compared to common carotid IMT [19]. Since atherosclerosis in the carotid arteries generally develops at the bifurcation well before the common carotid artery, a greater progression of IMT at the bifurcation would be expected to more highly correlate with disease in the coronary arteries. However, for research and clinical purposes, distinct advantages of common carotid IMT over IMT measured at the carotid artery bifurcation lie in its greater reproducibil-ity and abilreproducibil-ity to more consistently visualize the intima-media complex on ultrasound.

In conclusion, our results indicate that the noninva-sive measurement of rate of change of the common carotid IMT by ultrasonography is positively correlated with the change in coronary artery atherosclerosis de-termined by angiography. Our results also suggest that while measures obtained by the two arterial imaging methods are correlated, they also each assess different aspects of atherosclerosis, and thus may be useful as multiple end point measures in clinical trials.

Acknowledgements

This work was supported by NIH grant numbers RO1-HL49885 and RO3-HL54532.

(8)

References

[1] The ARIC Investigators. The Atherosclerosis Risk in Communi-ties (ARIC) Study: design and objectives. Am J Epidemiol 1989;129:687 – 702.

[2] O’Leary DH, Polak JF, Wolfson SK, et al. Use of sonography to evaluate carotid atherosclerosis in the elderly: the Cardiovascular Health Study. Stroke 1991;22:1155 – 63.

[3] Blankenhorn DH, Selzer RH, Crawford DW, Barth JD, Liu C-R, Liu C-H, Mack WJ, Alaupovic P. Beneficial effects of colestipol-niacin therapy on the common carotid artery: two-and four-year reduction of intima-media thickness measured by ultrasound. Circulation 1993;88:20 – 8.

[4] Mack WJ, Selzer RH, Hodis HN, Erickson JK, Liu C-R, Liu C-H, Crawford DW, Blankenhorn DH. One-year reduction and longitudinal analysis of carotid intima-media thickness associ-ated with colestipol/niacin therapy. Stroke 1993;24:1779 – 83. [5] Crouse JR, III, Byington RP, Bond MG, Espeland MA, Craven

TE, Sprinkle JW, McGovern ME, Furberg CD. Pravastatin, Lipids, and Atherosclerosis in the Carotid Arteries (PLAC-II). Am J Cardiol 1995; 75:455 – 459.

[6] Furberg CD, Adams Jr, HP, Applegate WB, Byington RP, Espeland MA, Hartwell T, Hunninghake DB, Lefkowitz DS, Probstfield J, Riley WA, Young B. Effect of lovastatin on early carotid atherosclerosis and cardiovascular events. Circulation 1994;90:1679 – 1687.

[7] Salonen R, Nyyssonen K, Porkkala E, Rummukainen J, Belder R, Park J-S, Salonen JT. Kuopio Atherosclerosis Prevention Study (KAPS): a population-based primary preventive trial of the effect of LDL lowering on atherosclerotic progression in carotid and femoral arteries. Circulation 1995;92:1758 – 64. [8] Hodis HN, Mack WJ, LaBree L, et al. Reduction in carotid

arterial wall thickness using lovastatin and dietary therapy: a randomized, controlled clinical trial. Ann Intern Med 1996;124:548 – 56.

[9] Suurkula M, Agewall S, Fagerberg B, Wendelhag I, Wikstrand J. Multiple risk intervention in high-risk hypertensive patients: a 3-year ultrasound study of intima-media thickness and plaques in the carotid artery. Arterioscler Thromb Vasc Biol 1996;16:462 – 70.

[10] Pignoli P, Tremoli E, Oreste P, Paoletti R. Intimal plus medial thickness of the arterial wall: a direct measurement with ultra-sound imaging. Circulation 1986;74:1399 – 406.

[11] Craven TE, Ryu JE, Espeland MA, Kahl FR, McKinney WM, Toole JF, McMahan MR, Thompson CJ, Heiss G, Crouse JR III. Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis: a case-control study. Circulation 1990;82:1230 – 1242.

[12] Salonen R, Salonen JT. Determinants of carotid intima-media thickness: a population-based ultrasonography study in Eastern Finnish men. J Int Med 1991;229:225 – 31.

[13] Wofford JL, Kahl FR, Howard GR, McKinney WM, Toole JF, Crouse JR, III. Relation of extent of extracranial carotid artery atherosclerosis as measured by B-mode ultrasound to the extent of coronary atherosclerosis. Arterioscler Thromb 1991;11:1786 – 1794.

[14] O’Leary DH, Polak JF, Kronmal RA, Kittner SJ, Bond MG, Wolfson SK, Jr., Bommer W, Price TR, Gardin JM, Savage PJ. Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study. Stroke 1992;23:1752 – 1760.

[15] Geroulakos G, O’Gorman DJ, Kalokiki E, Sheridan DJ, Nico-laides AN. The carotid intima-media thickness as a marker of the presence of severe symptomatic coronary artery disease. Eur Heart J 1994;15:781 – 5.

[16] Adams MR, Nakagomi A, Keech A, et al. Carotid intima-media thickness is only weakly correlated with the extent and severity of coronary artery disease. Circulation 1995;92:2127 – 34. [17] Crouse JR III, Craven TE, Hagaman AP, Bond MG.

Associa-tion of coronary disease with segment-specific intimal-medial thickening of the extracranial carotid artery. Circulation 1995;92:1141 – 1147.

[18] Visona A, Pesavento R, Lusiani L, et al. Intimal medial thicken-ing of common carotid artery as indicator of coronary artery disease. Angiology 1996;47:61 – 6.

[19] Hulthe J, Wikstrand J, Emanuelsson H, Wiklund O, de Feyter PJ, Wendelhag I. Atherosclerotic changes in the carotid artery bulb as measured by B-mode ultrasound are associated with the extent of coronary atherosclerosis. Stroke 1997;28:1189 – 94. [20] Tanaka H, Nishino M, Ishida M, Fukunaga R, Sueyoshi K.

Progression of carotid atherosclerosis in Japanese patients with coronary artery disease. Stroke 1992;23:946 – 51.

[21] Blankenhorn DH, Johnson RL, Nessim SA, Azen SP, Sanmarco ME, Selzer RH. The Cholesterol Lowering Atherosclerosis Study (CLAS): design, methods, and baseline results. Control Clin Trials 1987;8:354 – 87.

[22] Azen SP, Mack WJ, Cashin-Hemphill L, et al. Progression of coronary artery disease predicts clinical coronary events: long-term follow-up from the Cholesterol Lowering Atherosclerosis Study. Circulation 1996;93:34 – 41.

[23] Hodis HN, Mack WJ, LaBree L, Selzer RH, Liu C-R, Liu C-H, Azen SP. The role of carotid artery intima-media thickness in predicting clinical coronary events. Ann Intern Med 1998;128:262 – 9.

[24] Selzer RH, Hagerty C, Azen SP, Siebes M, Lee P, Shircore A, Blankenhorn DH. Precision and reproducibility of quantitative coronary angiography with applications to controlled clinical trials: a sampling study. J Clin Invest 1989;83:520 – 6.

[25] Blankenhorn DH, Azen SP, Kramsch DM, et al. Coronary angiographic changes with lovastatin therapy: the Monitored Atherosclerosis Regression Study (MARS). Ann Intern Med 1993;119:969 – 76.

[26] Selzer RH, Hodis HN, Kwong-Fu H, Mack WJ, Lee PL, Liu C-R, Liu C-H. Evaluation of computerized edge tracking for quantifying intima-media thickness of the common carotid artery from B-mode ultrasound images. Atherosclerosis 1994;111:1 – 11.

[27] Blankenhorn DH, Selzer RH, Mack WJ, et al. Evaluation of colestipol/niacin therapy with computer-derived coronary end point measures: a comparison of different measures of treatment effect. Circulation 1992;86:1701 – 9.

[28] Suurkula M, Agewall S, Fagerberg B, Wendelhag I, Widgren B, Wikstrand J. Ultrasound evaluation of atherosclerotic manifesta-tions in the carotid artery in high-risk hypertensive patients. Arterioscler Thromb 1994;14:1297 – 304.

[29] Wendelhag I, Wiklund O, Wikstrand J. Atherosclerotic changes in the femoral and carotid arteries in familial hypercholes-terolemia: ultrasonographic assessment of intima-media thick-ness and plaque occurrence. Arterioscler Thromb 1993;13:1404 – 11.

[30] Nagai Y, Metter J, Earley CJ, Kemper MK, Becker LC, Lakatta EG, Fleg JL. Increased carotid artery intimal-media thickness in asymptomatic older subjects with exercise-induced myocardial ischemia. Circulation 1998;98:1504 – 9.

[31] deGroot E, Jukema JW, van Swijndregt ADM, et al. B-mode ultrasound assessment of pravastatin treatment effect on carotid and femoral artery walls and its correlations with coronary arteriographic findings: a report of the Regression Growth Eval-uation Statin Study (REGRESS). J Am Coll Cardiol 1998;31:1561 – 7.

[32] Crouse JR, Toole JF, McKinney WM, et al. Risk factors for extracranial carotid artery atherosclerosis. Stroke 1987;18:990 – 6.

(9)

[33] Sternby NH. Atherosclerosis in a defined population: an autopsy survey in Malmo, Sweden. Acta Pathol Microbiol Scand 1968;194:1 – 216.

[34] Crawford DW, Brooks SH, Selzer RH, Barndt R Jr., Becken-bach ES, Blankenhorn DH. Computer densitometry for angio-graphic assessment of arterial cholesterol content and gross pathology in human atherosclerosis. J Lab Clin Med 1977;89:378 – 92.

[35] Salonen JT, Salonen R. Ultrasound B-mode imaging in observa-tional studies of atherosclerotic progression. Circulation 1993;87:II56 – 65.

[36] Chambless LE, Heiss G, Folsom AR, Rosamond W, Szklo M, Sharrett AR, Clegg LX. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) Study 1987 – 1993. Am J Epidemiol 147;1997:483 – 94.

[37] Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarction: the Rotterdam Study. Circulation 1997;96:1432 – 7.

[38] O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, Wolfson SK. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. New Engl J Med 1999;340:14 – 22.

[39] Fuster V, Badimon L, Badimon JJ, Chesebro JH. The pathogen-esis of coronary artery disease and the acute coronary syn-dromes. New Engl J Med 1992;326:242 – 309.

[40] Shircore AM, Mack WJ, Selzer RH, Lee PL, Azen SP, Alau-povic P, Hodis HN. Compensatory vascular changes of remote coronary segments in response to lesion progression as observed by sequential angiography from a controlled clinical trial. Circu-lation 1995;92:2411 – 8.

(1)

3.2. Correlations between common carotid artery IMT and QCA measures at baseline

Table 2 provides the Spearman correlations between a single baseline carotid IMT measure and QCA mea-sures obtained from the baseline coronary angiogram.

Table 2

Correlations between carotid IMT and QCA measuresa

QCA measure Correlation (95% CI)b_{of baseline QCA} Correlation (95% CI) of QCA change measure measure with baseline carotid IMT (mm) with carotid IMT change rate (mm/year)

Percent diameter stenosis(%S)

−0.04 (−0.22, 0.14) 0.21 (0.04, 0.38)§ All native lesions (n=129)

0.01 (−0.17, 0.19)

Native lesionsB50%S (n=125) 0.28 (0.10, 0.43)

−0.03 (−0.23, 0.19)

Native lesions]50%S (n=93) 0.00 (−0.21, 0.21)

−0.03 (−0.24, 0.18)

Unrelated native lesions (n=93)c _{0.33 (0.13, 0.50)}

All native segments (n=133) _{0.16 (}₋_{0.01, 0.33)} 0.28 (0.11, 0.44)¶ Native segmentsB50%S (n=133) 0.14 (−0.04, 0.31) 0.30 (0.13, 0.45)¶

−0.08 (−0.30, 0.14) 0.07 (−0.15, 0.28) Native segments]50%S (n=85)

0.03 (−0.16, 0.23)

Unrelated native segments (n=107)c _{0.25 (0.06, 0.42)}§

Minimum lumen diameter

−0.03 (−0.21, 0.14)

All native lesions (n=129) −0.22 (−0.38,−0.04)§

Native lesionsB50%S (n=125) −0.07 (−0.25, 0.11) −0.28 (−0.44,−0.11) −0.02 (−0.22, 0.20)

Native lesions]50%S (n=93) 0.01 (−0.20, 0.22)

Unrelated native lesions (n=93)c _{0.04 (}₋_{0.17, 0.25)} ₋_{0.20 (}₋_{0.40, 0.01)}§

−0.15 (−0.32, 0.02)

−0.09 (−0.26, 0.08) All native segments (n=133)

−0.07 (−0.25, 0.10)

Native segmentsB50%S (n=133) −0.17 (−0.34, 0.00)§

Native segments]50%S (n=84) 0.06 (−0.16, 0.27) −0.03 (−0.25, 0.20) 0.12 (−0.08, 0.31) −0.08 (−0.27, 0.12) Unrelated native segments (n=107)c

Vessel edge roughness

All native segments (n=133) 0.10 (−0.08, 0.27) 0.20 (0.02, 0.36)§

0.17 (0.00, 0.34)§ _{0.25 (0.08, 0.41)}

Native segmentsB50%S (n=133)

−0.10 (−0.31, 0.12)

Native segments]50%S (n=84) −0.07 (−0.29, 0.15)

Unrelated native segments (n=107)c _{0.17 (}₋_{0.02, 0.36)} _{0.25 (0.06, 0.43)}

A6erage segment diameter

−0.04 (−0.22, 0.13) 0.01 (−0.17, 0.18) All native segments (n=133)

Native segmentsB50%S (n=133) −0.05 (−0.23, 0.12) −0.01 (−0.18, 0.17)

0.03 (−0.19, 0.25) 0.05 (−0.17, 0.27) Native segments]50%S (n=84)

0.12 (−0.08, 0.31)

Unrelated native segments (n=107)c _{0.13 (}₋_{0.07, 0.31)}

a_{Each QCA measure represents the value (baseline or 2-year change) averaged over all evaluable lesions}_/_{segments in native coronary arteries.} b_{95% CI}₌_{95% confidence interval.}

c_{Analysis includes only native lesions}_/_{segments hemodynamically unrelated to bypass graft (lesion-based analyses}_n₌_{93 subjects;} segment-based analysesn=107 subjects).

§_P_B_0.05. _PB0.01. ¶_P_B_0.001.

(2)

3.3. Correlations between serial measures of common carotid artery IMT and QCA

3.4. Correlations between serial measures of common carotid artery IMT and QCA deri6ed measures

3.5. Associations between common carotid IMT, QCA, and lipids

4. Discussion

Correlations between common carotid IMT change rate, change in coronary artery lesion percent diameter stenosis, and on-trial lipidsa Change in coronary artery Common carotid IMT

change rate lesions: percent diameter stenosis

On-trial le6els of

0.26 (0.10, 0.42)‡ 0.39 (0.24, 0.53)§

Total cholesterol

0.29 (0.12, 0.43)§ LDL- 0.44 (0.30, 0.57)§

cholesterol

HDL- −0.35 (−0.49,−0.19)§ ₋_{0.26 (}₋_0.41,₋_0.09)‡ cholesterol

0.19 (0.02, 0.35)† 0.18 (0.01, 0.34)†

Triglycerides

a_{Numbers in table are correlations (95% confidence intervals).} †_P_B_0.05.

‡_P_B_0.01. §_P_B_0.001.

(3)

class of lesions is most closely related to the risk of clinical coronary events [22].

(4)

pre-dictive of clinical coronary events. Current

understanding is that mild/moderate lesions lead to clinical coronary events through plaque rupture with sudden vascular occlusion [39].

carotid artery IMT incorporates additional, indepen-dent information on prediction of coronary events

be-yond the angiographic measurement of lumen

narrowing. This is perhaps illustrated by the relation-ship between common carotid artery IMT and coro-nary vessel edge roughness which may represent atherosclerotic burden and the propensity for plaque rupture and clinical coronary events. These results sug-gest that use of both IMT and QCA change end points may be particularly useful multiple end points for clini-cal trials of anti-atherosclerosis interventions with the ultimate goal of reducing coronary event risk. This is of particular importance since the ability to solely use coronary angiographic end points to test anti-atherosclerosis therapies has become a great challenge with the current use of aggressive mechanical interven-tion for cardiac symptoms. Intervening revasculariza-tion procedures have profound effects on follow-up coronary angiograms, and in most cases render the follow-up angiographic data useless as a comparison to a pre-intervention angiogram. Even the effects of coro-nary angioplasty, once thought to be limited to the intervened coronary artery segment, may effect remote vessel segments through vascular remodeling, thereby altering the true angiographic effects of an anti-atherosclerosis medical therapy [40].