In a recent issue of JASE , we evaluated the use of extensive or restricted ultrasound protocols for the measurement of carotid intima-media thickness (CIMT). Using data from the Measuring Effects on Intima-Media Thickness: An Evaluation of Rosuvastatin (METEOR) study, we evaluated the completeness rates and rate of change in CIMT over time using different ultrasound protocols. In METEOR, CIMT images were taken from the near and far walls of the common carotid artery, the carotid bifurcation, and the internal carotid artery of the left and right carotid arteries at five angles of interrogation ( Figure 1 ). In contrast to the general belief, we showed that high completeness rates can be obtained with extensive ultrasound protocols that include measurements from the carotid bifurcation and internal carotid artery. We concluded that these extensive protocols are required to obtain the highest precision to observe a treatment effect and to fully cover the degree of atherosclerotic burden.
Polak and Pencina furthered the discussion with a corresponding editorial in the same issue. They postulated that a relatively simple CIMT ultrasound protocol should not be discouraged and that the benefits of multiple-angle CIMT protocols do not necessarily outweigh their costs. Their appreciated and balanced response was directed mainly toward the feasibility of an extensive ultrasound protocol in terms of the duration of image acquisition, costs, and implementation in clinical practice. Their arguments were strongly focused on the multiple-angle approach, whereas our conclusions were based mainly on a multisegment approach in trials (e.g., common carotid artery, carotid bifurcation, and internal carotid artery). Because the ideal ultrasound protocol remains to be established, an open discussion of this topic is of importance. Hence, we would like to elaborate further on this issue in this report.
Completeness of Carotid Intima-Media Thickness Information
Near-wall measurements and measurements from the carotid bifurcation and internal carotid artery are commonly thought to be difficult to obtain and therefore to give rise to missing data. Although this may have been true in the past, recent data show a great improvement in data availability with extensive protocols. For example, in METEOR, we showed that measurements from all carotid segments were complete in >94% of participants. Although it should be acknowledged that METEOR was a trial in a selected group of participants in whom visualization of CIMT was possible, high completeness rates were also obtained in the Rating Atherosclerotic Disease Change by Imaging With a New CETP Inhibitor I and II trials, two “all comers” studies among patients with familial hypercholesterolemia and high-risk patients with mixed dyslipidemia, respectively. In addition, high completeness rates for all carotid segments also have been reported in recently conducted population-based cohort studies. Hence, our data and those of others refute the general idea that CIMT measurements in the carotid bifurcation and internal carotid segments are much more difficult to obtain than in the common carotid segment.
Precision and Magnitude of Estimated Rates of Change
In our report, we showed that the direction, magnitude, and precision of the rate of change in CIMT over time are similar with five-angle, three-angle, and one-angle protocols. This is not surprising, given that the maximum CIMT (i.e., the largest measurement derived from the assessed interrogation angles) was used for model fit. In other words, under the assumption that at least one angle is measured, the number of maximum CIMT values entered into the statistical model is equal with five-angle, three-angle, and one-angle protocols that evaluate the same number of carotid segments. Nevertheless, because the maximum CIMT value in five-angle, three-angle, and one-angle protocols is derived from five, three, and one carotid angle, respectively, it may be that selecting the maximum CIMT value out of five possible angles causes more heterogeneity (i.e., decreases precision) compared to a three-angle or a one-angle protocol. A multiple-angle protocol, however, may still be worthwhile for a number of reasons. First, Tajik et al. showed that the distribution of atherosclerosis is asymmetric, so selectively choosing one angle is likely to ignore the asymmetric nature of the disease. Second, additional analyses of the rate of change in maximum CIMT in 12 carotid segments indicate that the rate of change differs by angle, information that cannot be captured when only one angle is measured ( Table 1 ). Third, Dogan et al. recently showed that ultrasound protocols including measurements from two or more angles provide a better balance among high reproducibility, large progression rates, and large and precise intervention effects compared to single-angle protocols. This finding was consistent across a mixture of trial populations and encourages the use of a multiple-angle approach over a single-angle approach.
| Carotid angle (left-right) ∗ | Rate of change in maximum CIMT (mm/y) | P | ||
|---|---|---|---|---|
| Rosuvastatin | Placebo | Rosuvastatin versus placebo | ||
| 180°-180° | −0.0003 ± 0.0015 | 0.0105 ± 0.0024 | −0.0108 ± 0.0028 | <.001 |
| 210°-150° | −0.0012 ± 0.0014 | 0.0075 ± 0.0022 | −0.0087 ± 0.0026 | .001 |
| 240°-120° | −0.0026 ± 0.0013 | 0.0147 ± 0.0021 | −0.0173 ± 0.0024 | <.001 |
| 270°-90° | 0.0006 ± 0.0014 | 0.0141 ± 0.0022 | −0.0134 ± 0.0026 | <.001 |
| 300°-60° | −0.0025 ± 0.0020 | 0.0090 ± 0.0032 | −0.0115 ± 0.0039 | .003 |
∗ Circumferential carotid angles measured using the Meijer’s arc.
Other important conclusions that can be drawn from Table 5 and Figure 3 of the original publication are that (1) the rate of change in CIMT becomes more precise when near-wall measurements are performed, and (2) protocols using a multisegment approach including measurements from the carotid bifurcation and internal carotid artery consistently show larger rates of change compared to protocols measuring the common carotid artery alone. We acknowledge that there are trials showing a beneficial response of an intervention on the rate of change in CIMT measured using a single-angle, far-wall common CIMT measurement. However, there are also examples of trials that would have shown no effect of the intervention using only the common CIMT, while the multisegment approach showed a beneficial effect. Thus, a quick, precise, and comprehensive estimation of the rate of change in CIMT is most likely to be obtained using extensive protocols including near-wall measurements and measurements from the carotid bifurcation and internal carotid artery.
Precision and Magnitude of Estimated Rates of Change
In our report, we showed that the direction, magnitude, and precision of the rate of change in CIMT over time are similar with five-angle, three-angle, and one-angle protocols. This is not surprising, given that the maximum CIMT (i.e., the largest measurement derived from the assessed interrogation angles) was used for model fit. In other words, under the assumption that at least one angle is measured, the number of maximum CIMT values entered into the statistical model is equal with five-angle, three-angle, and one-angle protocols that evaluate the same number of carotid segments. Nevertheless, because the maximum CIMT value in five-angle, three-angle, and one-angle protocols is derived from five, three, and one carotid angle, respectively, it may be that selecting the maximum CIMT value out of five possible angles causes more heterogeneity (i.e., decreases precision) compared to a three-angle or a one-angle protocol. A multiple-angle protocol, however, may still be worthwhile for a number of reasons. First, Tajik et al. showed that the distribution of atherosclerosis is asymmetric, so selectively choosing one angle is likely to ignore the asymmetric nature of the disease. Second, additional analyses of the rate of change in maximum CIMT in 12 carotid segments indicate that the rate of change differs by angle, information that cannot be captured when only one angle is measured ( Table 1 ). Third, Dogan et al. recently showed that ultrasound protocols including measurements from two or more angles provide a better balance among high reproducibility, large progression rates, and large and precise intervention effects compared to single-angle protocols. This finding was consistent across a mixture of trial populations and encourages the use of a multiple-angle approach over a single-angle approach.
