Comprehensive Prediction Model for Incident CHD

Dec 16, 2024
Font Size
A
A
A
Authors:
Naderian M, Norland K, Schaid DJ, Kullo IF.
Citation:
Development and Evaluation of a Comprehensive Prediction Model for Incident Coronary Heart Disease Using Genetic, Social, and Lifestyle–Psychological Factors: A Prospective Analysis of the UK Biobank. Ann Intern Med 2024;Dec 10:[Epub ahead of print].
Summary By:
Elizabeth A. Jackson, MD, FACC

Quick Takes

  • A polysocial score (PSSCHD) incorporating a broad array of self-reported social determinants of health (SDOH) and lifestyle–psychological factors was associated with coronary heart disease (CHD), the strength of association being similar to that of the polygenic risk score (PRSCHD) with CHD. PRSCHD and PSSCHD were not correlated, and their effects on CHD risk were independent and additive.
  • Adding SDOH, lifestyle–psychological factors, and genetic information to clinical risk calculators improved their performance.
  • Self-reported non-White participants had higher PSSCHD scores than self-reported White participants.
  • Regardless of genetic predisposition, unfavorable environmental and lifestyle–psychological factors captured in the PSSCHD increased the risk for CHD by approximately 1.5 times for a 1 SD increase in PRSCHD.

Study Questions:

Does the inclusion of a polygenic risk score (PRS), polysocial score (PSS), and lifestyle–psychological factors improve clinical risk calculators for coronary heart disease (CHD)?

Methods:

Data from the UK Biobank, a cohort study of >500,000 residents from England, Wales, and Scotland, aged 40-70 years, recruited between 2006 and 2010, were used for the present analysis. Individual clinical risk for CHD at the time of recruitment was estimated using three clinical risk calculators (pooled cohort equations [PCEs], Predicting Risk of cardiovascular disease EVENTs [PREVENT], and QRISK3). Inclusion of genetic risk used the PRS (Polygenic Score Catalog identification: PGS000018) for CHD (PRSCHD), while social determinants of health (SDoH) included 100 related covariates (PSSCHD). Machine-learning and time-to-event analyses and model performance indices were used to evaluate the addition of PRS and PSS to the three CHD models.

Results:

After excluding missing variables, the present study included 388,224 participants (aged 55.5 [SD, 8.1] years; 42.5% men; 94.9% White). Factors such as slow walking pace, snoring, financial difficulties, and lack of education had the highest positive coefficients, indicating an increased risk for CHD. Conversely, factors like brisk walking pace, having a high educational qualification, and not having tiredness or lethargy had the highest negative coefficients, suggesting a lower risk for CHD.

Persons in the fifth quintile for both PRSCHD and PSSCHD had the highest risk for CHD (hazard ratio [HR], 2.95; 95% confidence interval [CI], 2.27-3.81; p < 0.001), and conversely, those in the first quintile for both scores had the lowest risk for incident CHD (HR, 0.29; 95% CI, 0.18-0.46; p < 0.001). A weak correlation was seen between the clinical risk scores and PRSCHD. Non-White persons had higher PSSCHD than White persons. The effects of PRSCHD and PSSCHD on CHD were independent and additive. For each 1 SD increase in PSSCHD, the HR for incident CHD was 1.43 (95% CI, 1.38-1.49; p < 0.001), and for each 1 SD increase in PRSCHD, the HR was 1.59 (95% CI, 1.53-1.66; p < 0.001). At a 10-year CHD risk threshold of 7.5%, adding PRSCHD and PSSCHD to PCE reclassified 12% of participants, with 1.86 times higher CHD risk in the up- versus down-reclassified persons, and showed superior performance compared with PCE. Similar results were seen when incorporating PRSCHD and PSSCHD into PREVENT and QRISK3.

Regardless of genetic predisposition, unfavorable environmental and lifestyle–psychological factors captured in the PSSCHD increased the risk for CHD by approximately 1.5 times for a 1 SD increase in PRSCHD.

Conclusions:

The authors conclude that a PSSCHD is associated with incident CHD, and its joint modeling with PRSCHD improves the performance of clinical risk calculators.

Perspective:

This study supports the incorporation of social, genetic, and lifestyle factors into risk prediction and supports the use of policy and public health initiatives to improve modifiable factors that lower the risk for CHD. As the authors acknowledge, this cohort was predominantly White, suggesting that these factors need to be examined in other populations.

Clinical Topics: Arrhythmias and Clinical EP, Cardiovascular Care Team, Genetic Arrhythmic Conditions, Prevention

Keywords: Coronary Disease, Genetics, Life Style, Risk Assessment, Social Determinants of Health


< Back to Listings