Risk-enhancing factors and social determinants of health in risk assessment for atherosclerotic cardiovascular disease.

BACKGROUND

The Pooled Cohort Equations (PCEs) do not accurately estimate atherosclerotic cardiovascular disease (ASCVD) risk in certain populations. The 2018 AHA/ACC cholesterol guideline identified risk-enhancing factors as a supplement to PCEs-based risk assessment. However, the role of each risk-enhancing factor in ASCVD risk assessment has not been well quantified. Further, social determinants of health (SDOH) are not included in the PCEs nor considered as risk-enhancing factors in the US cholesterol guideline. We sought to evaluate ASCVD risk associated with each risk-enhancing factor and commonly collected SDOH including education, income, and employment status, and to assess if adding risk-enhancing factors and SDOH to the PCEs improve ASCVD risk prediction.

METHODS

We included individuals aged 40 to 75 years, without ASCVD or diabetes at baseline, and with low-density lipoprotein cholesterol 70-189 mg/dL from two contemporary prospective cohort studies (MESA and REGARDS) and from Kaiser Permanente Southern California (KPSC). The primary endpoint was incident ASCVD defined as nonfatal myocardial infarction, fatal coronary heart disease, or fatal or nonfatal stroke over a 10-year period (median follow-up 10 years). We used Cox proportional hazards models to estimate associations between risk-enhancing factors and SDOH with ASCVD. We also assessed changes in model performance after adding risk-enhancing factors and SDOH to the PCEs.

RESULTS

We included 13,863 adults (mean age 60.7 years) from the prospective cohorts and 307,931 adults (mean age 54.8 years) from KPSC. Risk-enhancing factors including hypercholesterolemia, hypertriglyceridemia, metabolic syndrome, and chronic kidney disease were associated with a higher ASCVD risk, independent of 10-year risk estimated by the PCEs. Low education, low income, and unemployment were also associated with higher ASCVD risk. While adding individual risk-enhancing factors or SDOH to the PCEs had limited impact on model performance, adding multiple risk-enhancing factors and SDOH simultaneously led to modest improvements in discrimination (C-index increased by up to 0.07), calibration (integrated Brier score reduced by up to 2.3%), and net reclassification improvement up to 41.4%.

CONCLUSIONS

These findings suggest including SDOH and risk-enhancing factors may improve ASCVD risk assessment.