The echolucency of the carotid intima-media is related to increased cardiovascular risk factor levels, morbidity, and mortality. The aim of this study was to assess the effect of statins on the echolucency of the common carotid intima-media in a low-risk population.
Data from the Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin study were used. Ultrasound images from the far walls of the left and right common carotid arteries were used for evaluation of the echolucency of the carotid intima-media, measured by grayscale median (GSM). Low GSM values reflect echolucent structures, whereas high values reflect echogenic structures. The primary end point was the difference in the annual rate of change in GSM between rosuvastatin and placebo.
Two-year change in GSM did not significantly differ between rosuvastatin and placebo in the total population, with a mean difference in the rate of change in GSM of 1.13 (95% confidence interval, −1.00 to 3.25). The effect of rosuvastatin differed across quintiles of baseline GSM values ( P for interaction = .01). In the lowest quintile ( n = 175) (i.e., in those with the most echolucent intima-media), the difference in the rate of change in GSM between rosuvastatin and placebo was 4.18 (95% confidence interval, −0.23 to 8.58). Increases in GSM were significantly related to decreasing low-density lipoprotein cholesterol levels in the lowest quintile (β = 0.76; 95% confidence interval, 0.26 to 1.25).
Treatment with rosuvastatin did not affect the echolucency of the arterial wall in all low-risk individuals. However, a potential effect of rosuvastatin on the echolucency of the common carotid intima-media is most likely to be found in individuals with echolucent arterial walls at baseline.
Atherosclerosis is a chronic and progressive disease that may result in the formation of atherosclerotic plaque lesions in the arterial wall. Although atherosclerosis is a lifelong process, the disease remains asymptomatic for decades. However, once atherosclerosis becomes symptomatic, it often manifests as a major cardiovascular (CV) event such as myocardial infarction, stroke, or sudden death. Atherosclerotic plaques leading to CV events (i.e., vulnerable plaques) may not necessarily be the biggest or most stenotic plaques but primarily have vulnerable content. Autopsy and imaging studies have shown that vulnerable plaques are characterized by a large lipid-rich necrotic core, a thin fibrous cap, and a high macrophage content.
Plaque content can be measured noninvasively using B-mode ultrasound. On these images, echolucent plaques are lipid rich and hold less collagen, whereas echogenic plaques are more stable and have higher contents of fibrous tissue and calcification. Echolucent plaques have been associated with CV risk factors and with a higher risk for CV events than echogenic plaques in both symptomatic and asymptomatic individuals.
A disadvantage of plaque assessment is that it can be performed only in individuals with overt plaques and not in earlier stages of atherosclerosis. Carotid intima-media thickness (CIMT) is a measure of atherosclerosis that could be used in all stages of the disease process, and measurement of the echolucency of the carotid intima-media may be performed in individuals in whom plaques are absent. We recently showed in a sample of elderly individuals with a high prevalence of plaque that the echolucency of the common carotid intima-media was closely related to plaque echolucency. Also, an echolucent common carotid intima-media is associated with a threefold increased risk for all-cause mortality and an eightfold increased risk for CV mortality compared with an echogenic common carotid intima-media in elderly men. Therefore, examination of the echolucency of the carotid intima-media might be of use to determine CV risk.
Lipid-lowering therapies, especially statins, are widely used to reduce CV risk and events. Although scarce, data on the effect of statins on plaque echolucency have indicated a beneficial effect. However, it is unknown whether the echolucency of the common carotid intima-media could be decreased by statin therapy. We hypothesized that the echolucency of carotid intima-media with statin use decreases primarily in individuals with echolucent carotid intima-media. Therefore, we examined the effects of statin therapy on the rate of change in the echolucency of the common carotid intima-media in a post hoc analysis of the Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin (METEOR) study.
We used data from the METEOR study. METEOR was a 2-year, randomized, double-blind, placebo-controlled trial in 984 individuals. The rationale, design, and main findings of METEOR have been reported in detail previously. The main inclusion criteria were age 45 to 70 years (men) or 55 to 70 years (women) and screening low-density lipoprotein (LDL) cholesterol 120 to <190 mg/dL (3.1–4.9 mmol/L) for those with only age as a coronary heart disease risk factor and 120 to <160 mg/dL (3.1–<4.1 mmol L) for individuals with two or more coronary heart disease risk factors, a 10-year Framingham risk of <10%, at least one maximum CIMT measurement > 1.2 mm, and no measurement ≥ 3.5 mm from two separate ultrasound examinations. The main objective of METEOR was to assess the impact of rosuvastatin 40 mg/d versus placebo on the 2-year rate of change in CIMT. Eligible potential participants were randomized to either rosuvastatin or placebo in blocks of seven (five rosuvastatin, two placebo). The reason for this was that the statistical analysis of the primary end point data involved a two-stage design. In stage 1, a between-group comparison was made. If this analysis showed a statistically significant difference that favored rosuvastatin, the stage 2 analysis was performed, in which a within-group comparison for the rosuvastatin-treated group only was made. This analysis was performed to assess whether the active treatment caused a negative rate of change in CIMT rather than simply no change from baseline.
Carotid Artery Ultrasound Evaluation
Sonographers performed carotid ultrasound examinations 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 examination, images were taken from the common carotid artery, carotid bifurcation, and internal carotid artery, each providing near-wall and far-wall data for the right and left carotid arteries. For the right carotid artery, measurements were performed at 60°, 90°, 120°, 150°, and 180° and for the left carotid artery at 300°, 270°, 240°, 210°, and 180°. Images of these 12 artery segments were obtained from five different angles. Each examination potentially provided a maximum of 60 images. Sonographers were trained to provide images showing the maximum thickness of a particular site. Hence, when a plaque was located at these predefined locations, it was included in the maximum CIMT measurement. The Acuson Sequoia 512, Acuson Sequoia 256, SONOLINE Antares (Siemens Medical Solutions USA, Inc., Mountain View, CA), and HDI 5000 (Phillips Medical Systems, Andover, MA) were the ultrasound devices used in METEOR. These devices all have electrocardiographic gating and use high-resolution linear-array probes at about 7.5 MHz. All ultrasound measurements were made on end-diastolic images at the top of the R wave of the electrocardiogram.
For the present analysis, images from the left and right far wall of a single angle (left, 240° or 270°; right, 90° or 120°) from the common carotid artery were evaluated, resulting in a maximum of 14 measurements (2 × 7) per participant over the whole study duration. The choice for these carotid segments and angles was based on the results from an earlier study that showed that measurements from the far-wall measurements from the common carotid artery on one carotid angle were predictive of CV events.
The primary end point was the annualized rate of change in the grayscale median (GSM) of the common CIMT on the basis of all scans performed during the 2-year study period from two carotid artery sites (the far walls of the right and left common carotid arteries). Low GSM values represent an echolucent carotid intima-media, whereas high GSM values represent an echogenic carotid intima-media ( Figure 1 ).
Far-wall images from the common carotid artery were digitized and imported into Artery Measurement Software, automated software for the analysis of GSM. A maximal 10-mm segment with good image quality was chosen, and the Artery Measurement Software program automatically identified the borders of the CIMT of the far wall and the inner diameter of the vessel and calculated CIMT and the diameter from about 100 discrete measurements through the 10-mm-long segment. This automated analysis could be manually corrected if not found appropriate at visual inspection. Next, a region of interest was manually placed around the intima-media segment that was evaluated for CIMT, and the program calculated the GSM from analysis of the individual pixels within the region of interest on a scale ranging from 0 (black) to 256 (white). The ultrasound images were originally acquired for evaluation of CIMT, for which adjustment of ultrasound settings is of importance to obtain the clearest view of the intima and media interfaces. Hence, the transducer and ultrasound settings of the ultrasound images were not standardized. Yet images were calibrated by using blood as the reference for black and the adventitia as the reference for white, and gain settings for measurements within an individual were similar throughout the study. Nevertheless, we evaluated the gain settings of ultrasound images that were taken using the Acuson Sequoia 512 and Acuson Sequoia 256 ultrasound devices to obtain more insight into any potential effect of ultrasound settings. These machines were chosen because ultrasound gain settings could only be retrieved from stored B-mode ultrasound images acquired by these ultrasound devices. Even though ultrasound gain setting theoretically can have a wide range, ultrasound gain settings in METEOR participants were clustered and had a median of 2 dB and an interquartile range of 0 to 7 dB, suggesting a relatively small impact of variability in ultrasound gain settings.
All images were read in a blinded batch fashion after each participant had completed study treatment. A batch consisted of all images collected for a single participant and was read in random order by a single reader to minimize the impact of between-reader variability and temporal drift on estimates of change in CIMT within participants. The intraclass correlation coefficient for the mean GSM values was 0.79 on the basis of the two duplicate scans at baseline; the coefficient of variation was 18%, and the mean absolute difference in GSM was 0.41 ± 27. Reproducibility findings based on duplicate end-of-study scans showed similar results.
Data were analyzed according to an intention-to-treat principle in all individuals with at least one additional GSM reading after the baseline measurement. In the rosuvastatin group, 624 participants had ultrasound scans at 0 months, 618 participants at 6 months, 594 participants at 12 months, 554 participants at 18 months, and 538 participants at 24 months, compared with 252, 252, 240, 222, and 211 participants, respectively, in the placebo group. All participants had duplicate scans at baseline. At 24 months, 514 participants in the rosuvastatin group and 202 participants in the placebo group had duplicate scans. Because we hypothesized that the effect of rosuvastatin on the echolucency of carotid intima-media may be primarily found in individuals with an echolucent carotid intima-media, we tested the multiplicative interaction between quintiles of baseline GSM, treatment allocation, and rate of change in GSM. Data were analyzed in quintiles of baseline GSM value in case of statistical significance.
A multilevel, repeated-measures, linear mixed-effects model was used to study the rate of change in GSM. This model was similar to the main CIMT analyses of the METEOR trial. The levels used for the data were defined by (1) the individual 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, 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 date of ultrasound examination. Random effects within the model were intercept and slope for individual participants and sites within participants. Because GSM may depend on ultrasound gain settings, sensitivity analyses were performed with correction for gain setting in individuals with ultrasound gain settings documented ( n = 325). To assess the linearity of changes in GSM values across the study measurements, time-squared terms were included in the model. Random effects within the model were intercept and slope for individual participants and sites within participants.
The associations between percentage change in GSM and percentage change in LDL cholesterol, percentage change in high-density lipoprotein (HDL) cholesterol, and percentage change in C-reactive protein (CRP) were assessed using linear regression analyses. The model was specified with percentage change in GSM as the dependent variable and percentage changes in LDL cholesterol, HDL cholesterol, and CRP as independent variables, adjusted for age, sex, and treatment allocation.
Statistical analyses were performed with R statistical software (R Project for Statistical Computing, Vienna, Austria), and a two-sided significance level of .05 was used for statistical inferences.
Table 1 shows that the baseline characteristics of the study participants were well balanced over the treatment arms.
( n = 702)
( n = 282)
|Age (y)||57 ± 6.2||57 ± 6.0|
|Men||421 (60%)||167 (59%)|
|Caucasian||659 (94%)||268 (95%)|
|Body mass index (kg/m 2 )||27.1 ± 4.0||27.5 ± 4.0|
|Smoking (previous month)||22 (3%)||16 (6%)|
|Family history of premature coronary heart disease ∗||65 (9%)||31 (11%)|
|Hypertension (blood pressure ≥140/90 mm Hg or taking antihypertensive medication)||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)|
|GSM||84 ± 29||84 ± 31|
|Maximum CIMT (mm)||1.00 ± 0.23||1.00 ± 0.22|
Rate of Change in GSM
The annual rate of change in GSM was −1.18 (95% confidence interval [CI], −2.32 to −0.05) in the rosuvastatin group and −2.31 (95% CI, −4.10 to −0.52) in the placebo group (i.e., grayscale in both groups became more echolucent over time). The difference between groups was 1.13 (95% CI, −1.00 to 3.25) and was not statistically significant. Because the interaction term between quintile of baseline GSM and treatment allocation was statistically significant ( P = .01; i.e., the effect of rosuvastatin on the rate of change in GSM differed by baseline level of GSM), further analyses were based on quintiles of baseline GSM ( n = 175). Table 2 shows the findings by quintiles of baseline GSM. Differences in the rate of change in GSM between treatment groups were not statistically significant within any of the quintiles. However, the effect of rosuvastatin to retard echolucency of the carotid intima-media was possibly most pronounced in those with low baseline GSM values (difference between rosuvastatin and placebo, 4.18; P = .06 in the first quintile).
|Quintile||Rate of change in GSM (95% CI)||P value, rosuvastatin vs placebo|
|Rosuvastatin||Placebo||Difference between groups|
|1 (≤59.8)||9.25 (6.86 to 11.64)||5.08 (1.30 to 8.85)||4.18 (−0.23 to 8.58)||.06|
|2 (>59.8 to ≤72.75)||2.17 (0.09 to 4.26)||4.04 (0.74 to 7.33)||−1.87 (−6.18 to 2.45)||.40|
|3 (>72.75 to ≤86.9)||−0.70 (−2.83 to 1.43)||−1.25 (−4.62 to 2.12)||0.55 (−3.29 to 4.39)||.78|
|4 (>86.9 to ≤106)||−3.99 (−6.18 to −1.79)||−5.52 (−8.99 to −2.04)||1.53 (−2.51 to 5.56)||.46|
|5 (>106)||−12.90 (−15.45 to −10.35)||−14.24 (−18.28 to −10.21)||1.34 (−3.45 to 6.13)||.58|
The analyses adjusted for ultrasound gain settings may have lacked statistical power; however, the results confirm that a potential effect of rosuvastatin on the echolucency of the arterial wall is most likely to be found in individuals with echolucent arterial walls at baseline ( Supplementary Table 1 ).
The test for linearity of change in GSM between baseline and week 104 indicated that the rate of change was linear over the 2-year period ( P = .07 for the rosuvastatin group and P = .08 for the placebo group).
Relationship between Change in GSM and Change in LDL Cholesterol
The relationship between percentage change in GSM and percentage change in LDL cholesterol is shown in Table 3 . The LDL cholesterol–lowering effect of rosuvastatin versus placebo was highly significant and ranged from −45% to −54% (−71.5 to −85.1 mg/dL) across quintiles of baseline GSM. Linear regression analyses showed that percentage increase in GSM was positively associated with the percentage decrease in LDL cholesterol in the lowest quintile of baseline GSM (β = 0.76; 95% CI, 0.26 to 1.25). The relations in the other quintiles of baseline GSM were modest and not statistically significant.
|Quintile||Percentage change in LDL cholesterol, mean||Percentage change in LDL cholesterol|
|Mean difference (95% CI) ∗||P||β (95% CI) †||P|
|1 (≤59.8)||−47.71 (−54.80 to −40.61)||<.01||0.76 (0.26 to 1.25)||<.01|
|2 (>59.8 to ≤72.75)||−50.79 (−59.15 to −42.43)||<.01||0.18 (−0.06 to 0.42)||.15|
|3 (>72.75 to ≤86.9)||−54.47 (−61.49 to −47.78)||<.01||0.11 (−0.15 to 0.38)||.40|
|4 (>86.9 to ≤106)||−44.87 (−51.44 to −38.30)||<.01||0.15 (−0.08 to 0.37)||.19|
|5 (>106)||−51.58 (−58.89 to −44.27)||<.01||−0.07 (−0.25 to 0.10)||.42|