Author + information
- ↵⁎Reprint requests and correspondence:
Dr. Michiel L. Bots, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, HP Stratenum 6.131, Heidelberglaan 100, Utrecht 3584 CX, the Netherlands
For >3 decades, the measurement of carotid intima-media thickness (CIMT) has been applied in cardiovascular research (1). Yet, there are still various questions regarding the CIMT measurement (2,3). Two are addressed in the report by Baldassarre et al. (4) in this issue of the Journal: “What is the ‘best’ CIMT measure for increased cardiovascular risk?” and “Does a CIMT measurement help improve the prediction of cardiovascular events?.”
Baldassarre et al. (4) present the results of a large cohort study (IMPROVE [Carotid Intima Media thickness (IMT) and IMT-PROgression as Predictors of Vascular Events in a High-Risk European Population]) among 3,703 asymptomatic subjects, age 54 to 79 years, who had at least 3 vascular risk factors. IMPROVE recruited from 7 centers in Finland, France, Italy, the Netherlands, and Sweden. CIMT was measured at the common and internal carotid segments and at the carotid bifurcation, at both the near and far wall. From these measurements, several CIMT measures were constructed. The interadventitial lumen diameter of the common carotid artery was also measured. Mean Framingham risk score (10-year absolute risk to develop a cardiovascular event) of the cohort was 22%. Subjects were followed up for an average of 36.2 months, during which 215 cardiovascular events occurred.
What is the “best” CIMT measure for increased cardiovascular risk?
Over the past decades, a large number of studies have reported on the relation of increased CIMT and increased cardiovascular risk. Most studies had only common CIMT measurements. Some had also measurements from the other segments and expressing the CIMT measure as mean-maximum CIMT. Because the studies used different ultrasound approaches (scanning protocols, ultrasound equipment, and quantification methods) that affect the value of the CIMT measure, a direct comparison between various CIMT measures in their ability to reflect increased cardiovascular risk could not be done validly (5). The IMPROVE study makes this possible, however. By using 1 common ultrasound protocol, the authors found that for all CIMT measures (common, bifurcation, internal, mean and maximum, and aggregated CIMT estimates), an increased value is related to an increased risk of cardiovascular disease: the increased risk per 1 SD increase in CIMT varies between 22% and 43% (Table 3, model 4) (4). Importantly, there are no strikingly significant differences between the various CIMT measures in their ability to reflect risk. Thus, a common CIMT measure seems as good as a more elaborate CIMT measure. This finding supports the current view that there is no carotid artery segment (i.e., comparison among the various segments) that has clearly demonstrated a more significant association with cardiovascular disease (6). Of course, one may argue that the current sample (215 events) is too small to detect significant differences in the magnitude of the relation with cardiovascular event between the different CIMT measures. Thus, just because no difference is found does not convincingly prove that no difference exists. Unless results from larger studies or results from pooled analyses of existing studies show otherwise, there seems to be no compelling evidence to suggest that combined measurements or measurement of a specific segment is superior.
Does a CIMT measurement help improve the prediction of cardiovascular events?
Adding a measurement of CIMT to cardiovascular risk factors has been proposed as a strategy to improve prediction of cardiovascular risk in an individual. So far, several studies reported on this issue, but the evidence seems inconsistent (3). Some reported improvement (6) whereas others did not (7,8) or only in subgroups or for specific cardiovascular events (9). Despite the inconsistency, private initiatives proposing to have an individual's CIMT measured for adequate cardiovascular risk assessment are rapidly emerging. Therefore, more evidence on this issue as provided by IMPROVE is needed.
Baldassarre et al. (4) provide net reclassification improvement estimates for various CIMT measures by using the Framingham risk factors as a starting point for risk stratification (Table 7 in their paper) based on the entire cohort. When only the common CIMT is added to Framingham risk variables, >80% of the population remains in the original risk category. Approximately 5% of the individuals were correctly reclassified. For the more elaborate CIMT measures, >70% does not change the risk group. The combined CIMT measure (mean-maximum) leads to correctly reclassifying 10.5% of the individuals. Adding an interadventitial diameter measurement to the mean-maximum CIMT measure leads to a correct reclassification in 12% of the individuals. Table 7 (4) clearly also shows that a considerable proportion of the reclassified individuals (36% to 38%) was incorrectly moved to another risk category.
Potential clinical implication of the net reclassification improvement results
When assessing the reclassification statistics for clinical usefulness, it is important to actually see what happens to the individuals (who is shifted and in which direction [lower or higher risk]) and what are the observed risks within those groups. Table 8 provides this crucial information for the model where mean-maximum CIMT and interadventitial diameter were added (4). From that table, we find that 30% of the studied population has a Framingham risk score of ≥20%. One could argue whether this group of patients should actually undergo further risk stratification with CIMT measurements because they already qualify for pharmacological treatment, certainly in the presence of ≥3 established cardiovascular risk factors. A CIMT measurement in this group is only relevant when one would consider refraining from pharmacological treatment when the estimated absolute CVD risk of the individual becomes <20%. The IMPROVE authors clearly state that they do not favor less intensive treatment in this group and thus implicitly tell us that further risk stratification with CIMT in the high-risk group should not be conducted. Others (10) also indicated that less intensive intervention should not be advised according to ultrasound imaging data for individuals otherwise estimated at intermediate-to-high vascular risk on the basis of traditional vascular risk factors.
Based on the results in Table 8 (4), one could also argue whether measurement of CIMT would be worthwhile in the low-risk group. No individual moved from the low- to the high-risk category.
A considerable group of individuals (14.5%) shifted from the intermediate-risk group to the high-risk category. These individuals had a very high extrapolated 10-year risk of developing vascular events. In addition, the CIMT result led to a downward classification in 1 of 5 individuals. Unfortunately, the percentage of intermediate-risk individuals who were correctly reclassified and its variance estimate is not provided in the paper.
Where does IMPROVE bring us?
IMPROVE also confirmed that in populations with considerable medication use, an increased CIMT is related to increased CVD risk (4). Furthermore, it expands the knowledge base by showing that for such relations, common CIMT measures are as good as more elaborate CIMT measures.
IMPROVE found that for risk classification, the more elaborate CIMT measures are preferred above common CIMT measures (4). This finding is concordant with some (7–11) but not all studies (6). The Atherosclerosis Risk In Communities studies found no difference between the common and more elaborate CIMT measures in reclassification ability, a result compatible with IMPROVE (6). IMPROVE seems to show that CIMT measurements for improving prediction of cardiovascular risk should be restricted to intermediate-risk groups only, a finding that is congruent with most of the other studies.
IMPROVE was conducted among individuals with ≥3 cardiovascular risk factors (4). Its results highlight the dilemma of less intensive (pharmacological) treatment in those whose estimated risk, as based on traditional risk factors, is lowered as a consequence of the CIMT result. This applies to individuals going from high-risk to intermediate-risk and for individuals with intermediate-risk going to low-risk and calls on the concept of treating risk rather than risk factor levels. If we are not willing to treat risk, then the added value of a marker in the improvement of risk prediction should be based on upward classification only.
A number of issues remain. First, clinical trials are needed to determine whether, in the intermediate-risk group, a strategy of making therapeutic decisions by using CIMT in addition to established risk factors compared with using established risk factors alone in cardiovascular disease risk prediction will prevent cardiovascular events. In such trials, willingness to refrain from pharmacological treatment when down classification occurs must be taken into account. Second, cost-effectiveness studies need to be performed to study whether the estimated added value of CIMT measurement translates into health benefits against which costs, in particular for the intermediate-risk group. Finally, most individual studies lack power for valid and precise specific analyses by age, sex, ethnicity, risk factor groups, and cardiovascular event. Such studies are needed to identify those individuals in whom the benefits may be largest. Pooling of cohorts worldwide followed by individual participant data analyses may address these issues (12).
Dr. Bots has received, over the years, grants from AstraZeneca, Dutch Heart Foundation, Organon, Pfizer, Servier, The Netherlands Organisation for Health Research and Development, and TNO-Zeist; has received consultancy fees from AstraZeneca, Boehringer Ingelheim, Organon, Pfizer, Servier, Schering-Plough, and Unilever; and runs the Vascular Imaging Center in Utrecht, a core laboratory for CIMT measurements in national and international observational and intervention studies.
Dr. den Ruijter has reported that she has no relationships relevant to the contents of this paper to disclose.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- O'Leary D.H.,
- Bots M.L.
- Peters S.A.,
- den Ruijter H.M.,
- Bots M.L.,
- Moons K.G.
- Baldassarre D.,
- Hamsten A.,
- Veglia F.,
- et al.,
- IMPROVE Study Group
- Simon A.,
- Megnien J.L.,
- Chironi G.
- Nambi V.,
- Chambless L.,
- He M.,
- et al.
- Lorenz M.W.,
- Schaefer C.,
- Steinmetz H.,
- Sitzer M.
- Elias-Smale S.E.,
- Kavousi M.,
- Verwoert G.C.,
- et al.
- Rundek T.,
- Salameh M.J.