Adding psychological risk exposure to improve cardiovascular risk prediction: a commentary

This invited commentary refers to ‘Does the psychological risk exposure increase carotid plaque detection among the Chinese midlife population? A latent class and regression analysis’, by X. Gou et al., https://doi.org/10.1093/eurjcn/zvaf061. The development of plaque in the carotid arteries often indicates the presence of more widespread atherosclerosis, increasing the risk not only of stroke but also of other cardiovascular events such as myocardial infarction.1–3 Therefore, early prediction and identification of carotid plaque are important to the management of atherosclerosis and in the context of overall cardiovascular health.4 To date, the most useful predictors of carotid plaque development encompass a combination of well-established traditional risk factors, including non-modifiable elements such as age, sex,5,6 and family history of cardiovascular disease (CVD),6,7 and modifiable factors including hypertension, dyslipidaemia, smoking, and diabetes mellitus.8–10 In addition to these established risk factors, there is also evidence that psychological health variables such as anxiety, depression, and chronic stress are associated with adverse physiological changes such as inflammation, endothelial dysfunction, and carotid plaque development.11–16 The analysis by Gou et al.17 is potentially important since it is one of the rare occasions where psychological risk is assessed in addition to traditional cardiovascular risk factors in a midlife population, often before development of heart disease.17 The American Heart Association (AHA) has acknowledged that it is currently difficult to ascertain how best to combine measures of psychological health.18 The authors’ approach in using latent class analysis is a potential practical solution since it permits the use of multiple indicators (anxiety, depression, stress) to identify homogeneous subgroups within a heterogeneous population.19 The authors demonstrated that the composite psychological risk exposure was an independent contributor to carotid plaque detection, over and above traditional risk factors. Importantly, the authors also took a nuanced approach in analysing the separate contributions of each psychological risk factor to development of carotid plaque, finding a significant association only between perceived stress and carotid plaque in males. These findings may be explained if we recognize that most CVD risk factors interact synergistically in accumulating total CVD risk and may produce a greater combined effect than if each factor worked independently.20 The findings of Gou et al.17 imply a need to also consider psychological risk factors in screening for CVD and potentially in CVD risk calculation tools. Current CVD primary prevention risk assessment algorithms from Europe [Systematic COronary Risk Evaluation (SCORE) model,21 SCORE2,22 SCORE2-OP23] and North America [Predicting Risk of CVD EVENTs (PREVENT)]24 and New Zealand (PREDICT25) include only traditional risk factors. Similarly, the Australian (AusCVDRisk26) algorithm uses traditional risk factors but does permit reclassification to a higher risk category if the individual has a history of severe mental illness. The European Society of Cardiology (ESC) guidelines on CVD prevention have specifically recognized the potential contribution of anxiety and depression in the context of primary prevention.27 The ESC guidelines also recognize that psychosocial stress is strongly associated with socioeconomic and behavioural risk factors that are often included in risk algorithms including poor medication adherence and cigarette smoking.27 Additionally, the AHA has acknowledged the importance of psychological health in CVD risk, onset, and recurrence18 and there is a proposal to specifically include psychological health in Life’s Critical 9 model for cardiovascular health.28 There have been attempts to improve upon established risk algorithms by incorporating non-traditional risk factors. An Australian study added clinically assessed depression to the landmark Framingham Risk Equation model29 and found that this improved accuracy for predicting 10-year coronary heart disease in a sample of women, but over-estimated the actual number of events that occurred.30 More recently, polysocial risk scores have been developed using social determinants of health and these have added significant discriminatory power over models comprising traditional risk factors only.31,32 These social determinants, however, have focused mainly on social disadvantage and have not specifically included anxiety and depression, but have included psychological distress.31 Another promising approach has been the development of a comprehensive prediction model developed via machine learning, using genetic susceptibility (polygenic risk score), polysocial risk score, and lifestyle–psychological factors including anxiety and depression.33 Recent developments in machine learning have facilitated the analysis of many diverse risk factors and their interactions leading to potentially better discrimination and calibration of risk predictions than traditional algorithms.34 The authors acknowledge several limitations to their analyses that may limit the generalizability of their findings. The sample size was relatively small for a prospective cohort study in the general population with close to 80% being female. It would be interesting to determine whether psychosocial risk score remained equally predictive in a predominantly male sample. The primary outcome of the study was limited to binary assessments of presence or absence of carotid plaque by ultrasonography, rather than calculation of a carotid plaque score that may be more promising in atherosclerotic screening.1 Although ultrasound screening may have the advantage of being low cost and radiation free, the gold standard for assessment of carotid plaque development is now recognized as magnetic resonance imaging.35,36 The potential for reverse causality and bidirectional associations of psychological health with CVD should also be acknowledged. For example, the burden of annual medical expenditures associated with CVD may promote financial and psychological distress.37 The study by Gou and colleagues has focused on primary prevention, however, it could be argued that tracking psychological health is important throughout the entire cardiovascular journey.28,38 Recurrent CVD risk algorithms39–41 could also be potentially enhanced by the addition of psychological and other social determinants of health. A comprehensive approach in CVD risk assessment aligns well with the use of multifactorial psychological interventions that are proving to be useful in treating adults with CVD.42 The validation of psychological health factors and their potential utility and interaction with traditional risk factors remains a worthwhile research pursuit. Michael R. Le Grande (Writing—original draft [lead], Writing—review & editing [equal]), Barbara M. Murphy (Writing—review & editing [equal]), and Alun C. Jackson (Writing—review & editing [equal]) The authors received no financial support for the research. No new data were generated or analysed in support of this commentary.