The role of carotid ultrasonography for the primary prevention of cardiovascular disease in apparently healthy persons

This editorial refers to ‘Carotid ultrasound and systematic coronary risk assessment 2 in the prediction of cardiovascular events’, by X. Bao et al., https://doi.org/10.1093/eurjpc/zwad139.

Primary prevention of atherosclerotic cardiovascular disease (ASCVD) is feasible and safe through lifestyle adaptations, related to smoking of tobacco, dietary habits and physical activity pattern, and optimal control of elevated blood pressure, dyslipidaemias, and dysglycaemias.

The target population for the primary prevention of ASCVD consists largely of asymptomatic and apparently healthy individuals, representing the majority of the adult community.

Primary prevention of ASCVD should consist of a population strategy including primordial prevention, complemented by a high-risk strategy targeting adult and elderly persons who are at high cardiovascular disease (CVD) risk. The high-risk strategy implies that the intensity of the preventive actions is in accordance with the total CVD risk of the individual aiming at a cost-effective application of resources and at an acceptable number needed to treat for high-intensity preventive strategies.

Given the multifactorial origin of ASCVD, the estimation of the total CVD risk should include multiple CVD risk factors; the clinical estimation of the combined synergetic effects of these factors is unreliable. Models have been developed for that purpose. The most appropriate model to be used in clinical practice is the one that is based on results from cohort studies in populations that are reasonably similar in terms of CVD risk to the patient population in which one works.

For Europe, the systematic coronary risk assessment 2 (SCORE2) and the SCORE2 O.P. models that were recently developed^1,2 have been recommended in the 2021 ESC guidelines on CVD prevention in clinical practice.³ These models have been calibrated to four clusters of countries that are grouped on the basis of national CVD mortality statistics (low, moderate, high, and very high CVD risk countries). These SCORE2 charts do not apply to patients with documented ASCVD or with other high-risk conditions such as diabetes mellitus, chronic kidney disease, familial hypercholesterolaemia, or other genetic or rare lipid or blood pressure disorders.

Using these models, the asymptomatic apparently healthy population aged 40 years and older can be stratified in subgroups of total CV risk. In the 2021 ESC Guidelines on CVD prevention, age-specific cut-off values have been suggested to define these subgroups. These risk estimation and stratification procedures have advantages but also limitations; their performances in terms of risk prediction remain modest. It is therefore recommended to consider also ‘risk modifiers’ in CVD risk estimation, but the exact way of integrating additional information on top of the estimates of existing models is not that evident. In many studies in which ‘risk modifiers’ were added to the results of existing risk estimation systems, separately or as a multimarker, no clinically relevant improvement in risk stratification was found.⁴

Improvements of risk estimation models are still a challenge and the subject of new explorations. One approach affects the artificial dichotomy of primary vs. secondary prevention of ASCVD. The term of ‘apparently healthy’ people implies that no apparent disease on history and/or examination in primary care practice is found. Developments in the field of imaging have blurred this traditional subdivision in the sense that many apparently healthy middle-aged and elderly persons may have subclinical ASCVD. But the presence of asymptomatic subclinical atherosclerosis on imaging is a diagnostic documentation of existing disease and not a risk factor per se.

Among the imaging techniques that have been introduced to detect subclinical atherosclerosis, carotid artery ultrasonography has been recommended and has the advantage on other imaging techniques of being noninvasive, safe, unexpensive, and non-radioactive. It started with the measurement of the intima-media thickness of the carotid artery. Conflicting results have been reported on the incremental value of carotid intima media thickness (CIMT) measurements in estimating CVD risk. This may be attributed to differences across studies in defining CIMT and to differences in measurement techniques. Based on results from a meta-analysis of 12 cohorts, Den Ruijter et al.⁵ concluded that the addition of common CIMT measurements to the Framingham risk score was associated with only a small improvement in 10-year risk prediction and this improvement was unlikely to be of clinical importance.

From another meta-analysis of 11 cohort studies, it was concluded that the presence of carotid plaque was better than CIMT in predicting CVD events;⁶ carotid plaque also correlates with the presence and extent of coronary artery calcium,⁷ which is recognized among imaging techniques as the one with the best reclassification ability.

Carotid plaque is defined as the presence of a focal wall thickening that is >50% greater than the surrounding vessel wall, or as a focal region with an CIMT measurement >1.5 mm that protrudes into the lumen.⁸ The presence of carotid plaques reclassifies CVD risk and may be considered as a ‘risk modifier’ in patients at intermediate CVD risk.

In this issue of the EJPC, results are presented from the CV cohort substudy of the Malmö Diet and Cancer study.⁹ In this report, results are presented on whether and how subclinical carotid atherosclerosis affects the performance of the SCORE2 risk estimation model in 4588 non-diabetic and apparently healthy participants aged 46–68 years. Plaques were present in 33% of the study population. Adding plaques or CIMT to SCORE2 significantly improved performance of predicted risk, documented with C-statistics, net reclassification improvement, and integrated discrimination improvement. The SCORE2 overpredicted the 10-year CVD risk in those without carotid plaques while under-predicted the risk in those with carotid plaques.

The CVD risk factors that are included in most of the existing risk estimation systems only explain a minor proportion of the carotid plaque burden.¹⁰ In the Malmö study, there was evidence of a minimal but statistically significant mediation of plaque in the association between SCORE2 and CVD. The estimated proportion of risk of CVD mediated by plaque was 5.50%.⁹ Other factors are associated with the development of carotid atherosclerosis such as markers of inflammation and thrombosis, psychosocial factors, or polygenic risk scores.

There is a dissociation between CVD risk estimates derived from existing systems and the presence of subclinical carotid atherosclerosis. At least a quarter of apparently healthy adults have subclinical carotid atherosclerosis; the presence of plaques is significantly associated with the incidence of ASCVD events independent of traditional CVD risk factors and should be combined with them aiming at a comprehensive risk assessment that may help to guide CVD prevention strategies.

Finally, given improvements in ultrasound techniques, it becomes possible to look not only at the presence of plaque but also at plaque properties such as location, number, density, thickness, and echogenicity; these properties may contain information that could improve CVD risk prediction even more.¹¹ The interest in CVD risk prediction is also shifting from 10-year risk estimation to lifetime risk; the added value of carotid plaque detection could become more cost-effective using a longer time span even in persons in whom the total CVD risk is considered as low to moderate according to existing models. This is supported by the observation in the Malmö study⁹ where the SCORE2 estimates were predictive of ASCVD events after more than 10 years of follow-up when the information of carotid atherosclerosis was taken into account; but this needs further explorations in long-term cohort studies.

Funding

The author received no financial support for the research, authorship and/or publication of this article.

References

Author notes