Background
Ultrasound protocols to measure carotid intima-media thickness (CIMT) vary considerably with regard to carotid sites and angles that are assessed. Measurements from the carotid bifurcation and internal carotid artery are thought to be affected by large numbers of missing data. Actual published quantification of completeness rates and the relation with cardiovascular risk factors, however, is scarce. Also, it is currently unknown whether extensive ultrasound protocols including assessment of the carotid bifurcation and internal carotid artery add information in detecting rate of change in CIMT induced by drug therapy. These issues were addressed in this study using data from Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin (METEOR).
Methods
In METEOR, carotid ultrasound examinations were performed twice before randomization, once each at 6, 12, and 18 months after randomization, and twice after 24 months of study treatment. B-mode ultrasound images were obtained from the near and far walls of the left and right common carotid artery, bifurcation, and internal carotid artery at five predefined angles. Completeness of CIMT data was assessed by carotid site and by angle. A site was considered complete when any of the five angles was measured. The relation between completeness at baseline and cardiovascular risk factors was assessed using logistic regression analyses. Ultrasound protocols with a reduced number of carotid sites and angles were retrospectively constructed, and differences in the rate of change in maximum CIMT between ultrasound protocols were compared.
Results
At each visit, CIMT measurements from all 12 carotid sites were available for >94% of the participants. Incompleteness was the highest for near wall of the internal carotid artery and for the extreme angles (60° and 300°). Of 12 risk factors examined, higher body mass index was related to incompleteness. Ultrasound protocols with a reduced number of angles resulted in similar estimates for the differences in rate of change in maximum CIMT. However, reductions in the number of sites gave results in the same direction but with different magnitudes and larger standard errors.
Conclusions
High levels of complete data can be obtained with extensive ultrasound protocols that include measurement from the carotid bifurcation and internal carotid artery. A high body mass index contributes to incompleteness of CIMT measurements. Extensive ultrasound protocols are required to obtain the highest precision to observe a treatment effect and to fully cover the degree of atherosclerotic burden.
Carotid intima-media thickness (CIMT) is commonly used in observational studies to examine determinants and consequences of atherosclerosis. In addition, changes in CIMT over time are increasingly being used in clinical trials as an alternative end point for cardiovascular morbidity and mortality to study the efficacy of pharmacologic interventions. The main advantage of the use of CIMT as a primary outcome rather than hard end points, such as cardiovascular morbidity and mortality, is the reduction in costs due to a smaller sample size and shorter duration of follow-up. This may result in earlier detection of the presence or absence of effects of drug therapies on atherosclerotic progression.
At present, little uniformity exists on ultrasound protocols across studies. Some studies have used extensive protocols, including measurements from the near and far walls of the common carotid artery, carotid bifurcation, and internal carotid artery at multiple angles of interrogation, whereas others have restricted their protocols by measuring only the far wall of the common carotid artery. The thickest sites of the carotid artery are typically found in the internal carotid artery and in the carotid bifurcation. Difficulties in visualizing these sites in general, and near wall sites in particular, are thought to give rise to large amounts of missing data. One of the main arguments to restrict ultrasound protocols to measurements of the far wall of common carotid artery is that common CIMT data collection is nearly always complete in this segment. However, data on the completeness of CIMT measurements are often not reported, and if reported, only overall completeness rates are provided. Also, cross-sectional observed differences in absolute thickness across carotid segments do not necessarily result in a difference in observed rate of change over time. A thorough evaluation of whether observed rates of change in CIMT differ across ultrasound protocols that vary in the numbers of segments, sites, and angles assessed is lacking.
In contrast to the general view in the field, a recent analysis demonstrated that high levels of complete CIMT data could be obtained with extensive ultrasound protocols that also included measurements of the near wall of the carotid bifurcation and internal carotid artery. The investigators found that body mass index (BMI) and waist circumference were the main predictors of incomplete data.
Because there is still no agreement on the optimal ultrasound protocol in terms of high completeness rates and the evidence on this issue is limited, we studied the completeness of CIMT data and its relation with cardiovascular risk factors using data from a randomized controlled trial that used an extensive ultrasound protocol. Because completeness alone should not be the main argument to use an extensive or a restricted ultrasound protocol, we subsequently studied the differences between extensive and restricted ultrasound protocols in relation to rate of change in CIMT.
Methods
Study Population
In this post hoc analysis, data from the Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin (METEOR) trial. The rationale, design, and main outcomes of this study have been described in detail elsewhere. Briefly, METEOR was a randomized, double-blind, placebo-controlled trial among 984 individuals at low risk for cardiovascular disease on the basis of 10-year Framingham risk of <10% but with modest increases in CIMT as evidenced by at least one maximum CIMT measurement ≥1.2 mm and no measurement ≥3.5 mm from two separate ultrasound examinations in any of the 12 carotid segments assessed (described below). The main objective of METEOR was to assess the impact of rosuvastatin 40 mg/day versus placebo on the rate of change in CIMT over 2 years of follow-up.
B-Mode Ultrasound Method
Ultrasound examinations were performed twice before randomization; once each at 6, 12, and 18 months after randomization; and then twice at the end of 24 months of study treatment. At each visit, ultrasound images were obtained from six well-defined arterial segments from both the left and right carotid artery ( Figure 1 ), as detailed elsewhere. These segments were defined as the near and far walls of the common carotid artery extending from 10 to 20 mm proximal to the tip of the flow divider; the near and far walls of the carotid bifurcation as assessed from the tip of the flow divider, extending 10 mm proximal to the tip of the flow divider; and the near and far walls of the internal carotid artery as assessed in the 10 mm distal to the tip of the flow divider. There was no minimal acceptable length for a measurement, and a measurement was accepted as long as the reader had enough confidence to mark the CIMT. In METEOR, all images were measured in end-diastole at the top of the R wave of the electrocardiogram. The head of each participant was rotated 45° to the other side to better expose each carotid artery for interrogation. The Meijer carotid arc was used to image the artery at five prespecified angles in steps of 30° (from 60° to 180° on the right carotid artery and from 300° to 180° on the left carotid artery). This resulted in a maximum of 60 sets of CIMT measurements that could be provided at each examination.
The equipment used in METEOR for CIMT imaging was the Acuson Sequoia 512, Acuson Sequoia 256, and SONOLINE Antares (all Siemens Medical Solutions USA, Inc., Mountain View, CA) and the HDI 5000 (Phillips Medical Systems, Andover, MA). These devices all have electrocardiographic gating, and all used high-resolution linear-array probes around 7.5 MHz.
These systems were all approved by a quality control center, and the performance of these systems was monitored using phantoms. The real-time captured images were all stored on S-VHS videotape and were sent to one of two specialized ultrasound reading centers in Europe (University Medical Centre, Utrecht, The Netherlands) and the United States (Wake Forest University School of Medicine, Winston-Salem, NC) for CIMT measurement. All images were digitized and read at these two at reading centers.
All ultrasound scans were manually read with Image Pro software (Microsoft Corporation, Redmond, WA). Images were read by 16 readers who completed a uniform training program that ensured standardized settings across reading stations and core laboratories. Readers traced trailing edges on the near wall boundaries and leading edges on the far wall boundaries and were free to set as many anchors as needed to trace the interface, with straight lines connecting the anchor points. All images were read in a blinded batch fashion after each participant had finished. A batch consisted of all images collected from a single participant and was read by a single reader over an interval of a few days with visits arranged in random order to minimize the impact of between-reader variability and temporal drift on estimates of change in CIMT within participants.
Laboratory Analyses
As previously described, blood samples were taken for analysis of serum lipids, lipoproteins, and C-reactive protein. The samples were sent immediately for analysis to one of two central standardized laboratories in Europe (Covance Central Laboratory Services, Geneva, Switzerland) and in the United States (Covance Central Laboratory Services, Indianapolis, IN). Baseline samples were used for the present analysis.
Variations in Ultrasound Protocols
We simulated eight different ultrasound protocols that may be considered once designing a CIMT trial by selectively excluding up to 83% of the specific carotid segments and carotid walls. The most extensive ultrasound protocol was the original METEOR protocol, which included assessment of the near and far walls of the left and right common carotid artery, carotid bifurcation, and internal carotid artery. In the most restricted ultrasound protocol, only measurements from the far wall of the left and right carotid artery were performed. Each of the eight ultrasound protocols was applied to settings in which five, three, or one circumferential carotid angle was measured. Angle-specific measurements in ultrasound protocols with assessment of five carotid angles were performed at 60°, 90°, 120°, 150°, and 180° on the right carotid artery and for the left carotid artery at 300°, 270°, 240°, 210°, and 180°. Ultrasound protocols with three circumferential CIMT measurements for each carotid site were performed at 90°, 120°, 150° for the right carotid artery and at 270°, 240°, and 210° for the left carotid artery. CIMT measurements were performed at 120° and 240° of the right and left carotid artery, respectively, in protocols in which only one carotid angle was evaluated for each carotid segment. The primary end point for this analysis was the annualized rate of change in the mean of the maximum CIMT measurements based on all scans performed during the 2-year study period from each of the 12 carotid artery sites. The maximum CIMT in each of the 12 sites is defined as the largest measurement derived from the assessed interrogation angles, each one 30° different from the adjacent angle (i.e., the number of CIMT values used for the model fit was maximally 12 for each visit, irrespective of the number of angles assessed). When a carotid plaque was present, it was included in this measurement.
Statistical Analyses
Completeness rates were evaluated at the level of the carotid angle for the left and right carotid artery separately (five angles each), and carotid site (12 sites). Completeness of the angle-specific measurements is expressed as the average completeness on the six carotid sites that were examined on both the left and right carotid artery. Site-specific measurements were considered available when at least one CIMT measurement at any of the five angles was performed. This definition was used because in METEOR, the availability of at least one of five angle-specific measurements was required to have a site-specific CIMT value. All availability rates are presented as percentages. The numerator for completeness was the number of available CIMT measurements, and the denominator for completeness at a specific visit consisted of all those participants who indeed underwent carotid ultrasound examinations at that visit. Subgroup analyses were performed to examine differences in completeness for participants with BMIs ≥ 30 kg/m 2 and to study differences between European participants and participants from the United States.
Logistic regression analyses were used to study the relation between completeness of CIMT measurements at the site level at two baseline measurements and cardiovascular risk factors. Risk factors included were age, sex, BMI, alcohol use, family history of coronary heart disease (in a first-degree male relative aged < 55 years or in a first-degree female relative aged < 65 years), hypertension (blood pressure ≥ 140/90 mm Hg or taking antihypertensive medication), smoking during the previous month, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, C-reactive protein, and maximum CIMT. The levels of the risk factors were restricted to baseline measurements. Analyses were repeated with adjustments for age, sex, and BMI, because these variables were considered to confound the univariate associations. Results of the adjusted and unadjusted analyses are presented as odds ratios (ORs) and corresponding 95% confidence intervals (CIs).
The rate of change in CIMT in each of the ultrasound protocols considered was evaluated by a multilevel, repeated-measures, linear mixed-effects model using the lme4 package available in R statistical software. This analysis was similar to the analysis performed in the original METEOR study. The levels used were defined by (1) the participant and (2) the carotid artery site within the participant. The repeated measure was time. The model was specified in terms of fixed effects for carotid artery site, age, sex, scan reader, ultrasound machine, treatment group, time, and the interaction between treatment group and time. Time as a continuous variable was the interval from the date of randomization to the date of CIMT measurement. Random effects within the model were intercept and slope for individual participants and for sites within participants.
Results
The baseline characteristics of the study participants by treatment allocation are presented in Table 1 . The mean age of the participants was 57 ± 6 years, the majority were men (60%), and the mean BMI was 27 ± 4.0 kg/m 2 . Figures 2 A to 2E provide example longitudinal B-mode images of the carotid artery showing clear interfaces for measurements of CIMT in the near and far walls of the common carotid artery, carotid bifurcation, and internal carotid artery.
| Variable | Rosuvastatin ( n = 702) | Placebo ( n = 282) |
|---|---|---|
| Age (y) | 57 ± 6.2 | 57 ± 6.0 |
| Men | 421 (60%) | 167 (59%) |
| Caucasian | 659 (94%) | 268 (95%) |
| BMI (kg/m 2 ) | 27.1 ± 4.0 | 27.5 ± 4.0 |
| Smoking (during the previous month) | 22 (3%) | 16 (6%) |
| Family history of premature CHD ∗ | 65 (9%) | 31 (11%) |
| Hypertension † | 138 (20%) | 58 (21%) |
| Total cholesterol (mg/dL) | 229 ± 28.7 | 230 ± 27.7 |
| LDL cholesterol (mg/dL) ‡ | 155 ± 24.1 | 154 ± 24.2 |
| HDL cholesterol (mg/dL) ‡ | 50 ± 9.0 | 49 ± 9.2 |
| Triglycerides (mg/dL) | 126 ± 64.3 | 134 ± 67.8 |
| CRP (mg/L) | 1.40 (0.80–2.90) | 1.60 (0.80–3.20) |
| Mean of the maximum of all 12 CIMT sites (mm) | 1.15 ± 0.19 | 1.17 ± 0.20 |
| Mean CIMT of the CCA § (mm) | 0.76 ± 0.12 | 0.76 ± 0.12 |
∗ Defined as CHD in a first-degree male relative aged < 55 years or in a first-degree female relative aged < 65 years.
† Blood pressure ≥ 140/90 mm Hg or taking antihypertensive medication.
‡ To convert HDL and LDL cholesterol to millimoles per liter, multiply by 0.0259.
§ Average of the mean CIMT of the four CCA sites: near and far walls of the right and left CCA.
Completeness at the Site Level
Table 2 shows the rates of completeness that were obtained on each of the 12 carotid sites. The percentage of CIMT measurements at the baseline examinations was 94% for the near wall of the right internal carotid artery and 96% for the near wall of the left internal carotid artery. Completeness on the other carotid artery sites, including the carotid bifurcation, was >99%.
| Weeks after randomization | LCF | LCN | LBF | LBN | LIF | LIN | RCF | RCN | RBF | RBN | RIF | RIN |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| −4 | >99.9 | 99.9 | >99.9 | 99.7 | 99.7 | 96.2 | >99.9 | >99.9 | 99.8 | 99.9 | 99.1 | 93.9 |
| −2 | >99.9 | 99.9 | 99.8 | 99.4 | 99.4 | 96.2 | 99.9 | 99.9 | 99.9 | >99.9 | 99.4 | 94.2 |
| 26 | >99.9 | 99.9 | 99.8 | 99.7 | 99.5 | 96.7 | 99.9 | 99.9 | >99.9 | 99.9 | 99.2 | 95.0 |
| 52 | 99.9 | 99.6 | 99.8 | 99.4 | 99.2 | 97.5 | >99.9 | >99.9 | >99.9 | 99.9 | 99.5 | 94.8 |
| 78 | 99.9 | 99.9 | 99.9 | 99.2 | 98.8 | 96.0 | >99.9 | 99.7 | >99.9 | 99.9 | 99.2 | 95.0 |
| 104 | >99.9 | 99.7 | 99.6 | 99.6 | 99.5 | 96.8 | >99.9 | 99.9 | >99.9 | 99.7 | 99.7 | 95.2 |
| 105 | >99.9 | 99.7 | 99.9 | 99.4 | 98.9 | 97.2 | >99.9 | >99.9 | 99.9 | 99.9 | 99.3 | 95.8 |
The number of participants who withdrew from the study, who did not undergo ultrasound examinations and hence had no CIMT data, was nine (1%) for the 6-month visit, 52 (6%) for the 12-month visit, 110 (13%) for the 18-month visit, 133 (15%) for the first 24-month visit, and 160 (18%) for the second 24-month visit. Of those with ultrasound examinations after 2 years of follow-up, 95% of the near wall of the right internal carotid artery measurements and 97% of the near wall of the left internal carotid artery measurements were complete. Complete CIMT data on the other carotid sites was available in 99% of the subjects who underwent end-of-study ultrasound examinations.
Completeness rates for the near wall of the left and right internal carotid artery were lower than those of the far wall of the internal carotid artery. No major differences (≤0.5%) were present between the near and far walls of the right or left common carotid artery or carotid bifurcation during all visits.
Analyzing participants from Europe and the United States separately showed a similar completeness rate of >99% for all carotid sites, except for the near wall of the internal carotid artery. Completeness for the near wall of the right internal carotid artery over time was on average 97% for European participants and 93% for participants from the United States. For the left internal carotid artery, 97% of the measurements were available for European participants, and 96% of the measurements were available for participants from the United States.
Availability of CIMT data in participants with BMIs ≥ 30 kg/m 2 was 90% for the near wall of the right internal carotid artery and 95% for the near wall of the left internal carotid artery. Completeness rates for all other carotid sites were similar to the analyses on all subjects.
Completeness at the Angle Level
The availability of CIMT measurements was the highest at 240° of the left carotid artery and at 120° of the right carotid artery ( Table 3 ). At these angles, 93% of the CIMT measurements were available at baseline for both carotid arteries, with 94% available CIMT measurement at the end of follow-up for the left carotid artery and 93% for the right carotid artery. The lowest number of available measurements was found at the most extreme angles, which were 300° for the left carotid artery and 60° for the right carotid artery. At the beginning of the study, 38% of the CIMT measurements at 300° were available for the left carotid artery, and 50% of the measurements at 60° were available for the right carotid artery. Furthermore, completeness rates were generally lower in the left carotid artery than in the right carotid artery. The average completeness rates over all ultrasound examinations for the 180°, 210°, and 270° angles of the left carotid artery were 83%, 89%, and 87%, respectively. For the right carotid artery, the completeness rates for the 180°, 150°, and 90° angles were 83%, 89%, and 89%, respectively.
| Weeks after randomization | |||||||
|---|---|---|---|---|---|---|---|
| Angle | −4 | −2 | 26 | 52 | 78 | 104 | 105 |
| Left carotid artery | |||||||
| 180° | 82.7 | 81.9 | 83.0 | 83.7 | 82.6 | 82.9 | 83.7 |
| 210° | 88.3 | 88.7 | 88.8 | 89.9 | 89.4 | 89.1 | 89.8 |
| 240° | 93.2 | 93.7 | 92.9 | 93.6 | 92.3 | 93.6 | 94.2 |
| 270° | 87.6 | 88.3 | 87.6 | 88.3 | 87.6 | 86.4 | 86.5 |
| 300° | 37.5 | 39.0 | 38.5 | 37.3 | 33.9 | 31.2 | 33.1 |
| Right carotid artery | |||||||
| 180° | 82.8 | 82.9 | 84.4 | 84.4 | 82.1 | 83.2 | 84.2 |
| 150° | 89.1 | 88.7 | 89.7 | 89.3 | 89.7 | 90.5 | 89.4 |
| 120° | 92.7 | 93.8 | 93.5 | 92.7 | 93.2 | 93.1 | 93.1 |
| 90° | 90.0 | 90.3 | 91.3 | 82.7 | 89.8 | 89.2 | 90.7 |
| 60° | 49.5 | 49.8 | 50.7 | 54.0 | 42.4 | 39.5 | 40.7 |
Determinants of Completeness
The completeness rate at baseline for carotid sites was highly right skewed. The majority of participants (84%) did have CIMT data for all sites at baseline. Three participants (0.3%) had no data at five sites, three (0.3%) had no data at four sites, eight (0.9%) had no data at three sites, 38 (4%) had no data at two sites, and 82 (9%) had no data at one site.
Table 4 shows the determinants of arterial-site completeness at the prerandomization visits. Univariate analyses showed that male sex, higher BMI, alcohol use, higher levels of triglycerides, and higher C-reactive protein levels were associated with a lower number of available CIMT measurements. When adjusted for age and sex, a higher BMI (OR, 0.88; 95% CI, 0.84–0.92), alcohol use (OR, 0.68; 95% CI, 0.46–0.99), and a higher level of triglycerides (OR, 0.73; 95% CI, 0.58–0.91) remained associated with a lower number of CIMT measurements. When adjusted for age, sex, and BMI, the association between the number of sites with CIMT measurements and alcohol use and triglyceride level, respectively, were substantially attenuated, with 95% CIs that included 1.
| Variable | Unadjusted | Age and sex adjusted ‡ | Age, sex, and BMI adjusted § |
|---|---|---|---|
| Age | 0.98 (0.95–1.01) | 1.00 (0.97–1.04) | 0.98 (0.95–1.01) |
| Male gender | 0.65 (0.45–0.94) | 0.65 (0.42–1.00) | 0.69 (0.47–1.01) |
| BMI | 0.87 (0.84–0.91) | 0.88 (0.84–0.92) | NA |
| Alcohol use | 0.64 (0.44-0.93) | 0.68 (0.46–0.99) | 0.68 (0.46–1.01) |
| Family history of CHD ∗ | 1.42 (0.80–2.55) | 1.30 (0.72–2.36) | 1.29 (0.69–2.40) |
| Hypertension † | 1.53 (0.99–2.36) | 1.30 (0.72–2.36) | 1.13 (0.71–1.81) |
| Smoking | 1.66 (0.70–3.94) | 1.37 (0.57–3.35) | 1.89 (0.76–4.67) |
| LDL cholesterol | 1.05 (0.78–1.41) | 1.05 (0.78–1.41) | 1.03 (0.76–1.41) |
| HDL cholesterol | 1.28 (0.58–2.83) | 1.71 (0.74–3.95) | 1.05 (0.44–2.51) |
| Triglycerides | 0.75 (0.60–0.94) | 0.73 (0.58–0.91) | 0.86 (0.67–1.10) |
| CRP | 0.73 (0.55–0.98) | 0.78 (0.58–1.06) | 1.01 (0.70–1.45) |
| Maximum CIMT | 0.86 (0.61–1.21) | 0.87 (0.61–1.22) | 0.86 (0.61–1.23) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
