Add Anxiety and Depression to CVD Risk Prediction Model?

Adding measures of anxiety and depression to the American Heart Association’s Predicting Risk of Cardiovascular Disease Events (PREVENT) model may have little additional effect on predicting the risk for new cases of cardiovascular disease (CVD) at the population level, according to new research.

Yet numerous studies have noted the associations between these mental health conditions and an increased risk for CVD, the authors wrote, so other tools, broader mental health conditions, or diagnostic interview data could be useful in future studies.

“The additive values of these mental health measures were lower than we expected. New CVD risk factors don’t always improve the existing CVD risk scores meaningfully, as shown by a previous study, which investigated the additive values of apolipoproteins,” said lead author Shinya Nakada, a PhD student at the University of Glasgow, Glasgow, Scotland, who researches mental health and CVD.

“However, new predictors with limited additive values may still offer practical advantages when it is reproducible, noninvasive, and less costly,” Nakada said. “The measures used in this study were drawn from real-world settings, such as electronic health records and short self-reports, since they are of lower cost and can be implemented relatively easily.”

The study was published online on January 13 in the Canadian Medical Association Journal.

Little Additional Effect

In 2024, PREVENT model was developed to predict the risk for an initial CVD event among the general adult population aged 30-79 years in the United States. The model incorporates risk predictors relevant to cardiovascular, kidney, and metabolic diseases, including factors related to obesity, diabetes, and antihypertensive and statin medications.

Because anxiety and depression are the most common mental health conditions worldwide, increasing in prevalence, and independently associated with CVD, Nakada and colleagues analyzed whether incorporating certain mental health measures into PREVENT model could detect high-risk groups that were previously overlooked.

The research team developed and internally validated prediction models of 10-year risk for incident CVD (comprising coronary artery disease, stroke, and heart failure) using cohort data from the UK Biobank. They included mental health predictors such as baseline depressive symptom score, self-reported anxiety and depression, and a record-based history of anxiety and depression diagnoses.

After randomly assigning more than 500,000 UK Biobank participants to a derivation set (60%) and a validation set (40%), the research team determined incremental predictive values based on C-indices, sensitivity, specificity, and net reclassification improvement indices, using a threshold of 10-year risk for incident CVD > 5%.

The derivation set of 195,000 participants had 15,787 CVD events, whereas the validation set of 130,000 participants had 10,639 CVD events. In both groups, participants with a CVD event were more likely to be older, men, and current smokers. They also were more likely to have higher systolic blood pressure, diabetes, and mental health conditions.

In the single-predictor models, all mental health predictors were associated with CVD. In the validation set, including all the mental health measures — except self-reported anxiety — resulted in a modest increase in the C-index and specificity, though sensitivity was unchanged.

Among the mental health predictors, the depressive symptom score had the largest, though still modest, improvements in C-index (difference of 0.005) and specificity (difference of 0.89%). As a result, the model including depressive symptom score had the largest overall (1.14) and nonevent (0.89) net reclassification indices.

Based on the findings, the models including depressive symptom score were more likely to distinguish those at higher risk for CVD from those at lower risk, the authors wrote. For every 1000 people at lower risk, about nine would no longer be incorrectly classified as high risk, which suggests relatively limited effects on risk classification.

“In our study, although the measures of anxiety and depression were associated with CVD, including them in PREVENT had little additional effect on the risk classification of CVD at the population level,” Nakada said. “So it may not be worthwhile, especially when it is costly.”

‘Extremely Debilitating Conditions’

However, new CVD risk predictors — such as anxiety and depression — could help identify targeted therapeutic pathways or actionable responses in primary care settings, the study authors wrote. For instance, the depressive symptom score may help identify previously undiagnosed depression and could be considered during CVD risk prevention visits, despite its limited contribution in the PREVENT model itself.

“Anxiety and depression are extremely debilitating conditions, and in some cases can substantially interfere with treatment for CVD,” said Scott Lear, PhD, professor of health sciences and the Pfizer/Heart and Stroke Foundation chair in cardiovascular prevention research at Simon Fraser University in Burnaby, British Columbia, Canada.

“In addition, some CVD risk factors likely to be affected by depression and anxiety, such as smoking and blood pressure, are already included in the PREVENT model, so this may water down the value of assessing anxiety and depression,” he said.

Lear, who wasn’t involved with this study, has researched various CVD risk predictors, including by ethnicity. In this study using UK Biobank data, participants were more likely to be White, affluent, and healthy than the general UK population — and the results may not be generalizable to other populations worldwide. Future studies could look at other populations and broader mental health measures, such as anxious feelings and insomnia, and associations with CVD risk prediction, the authors wrote.

“I wouldn’t want people to think that anxiety and depression don’t matter when considering CVD risk and treatment,” Lear said. “I’m sure many primary care physicians realize that.”

No independent funding for the study was reported. Nakada and Lear reported no relevant financial relationships.

Carolyn Crist is a health and medical journalist who reports on the latest studies for Medscape Medical News, MDedge, and WebMD.