Abstract P010: Do the Pooled Cohort Equations Accurately Predict Cardiovascular Disease Risk in Older Adults?

Abstract

Background: The ACC/AHA Pooled Cohort Equations (PCE), derived using data from adults ages 40-79, have not been evaluated for accuracy among older adults. Methods: We evaluated 2,667 adults aged ≥75 years without known ASCVD in the NIH pooled cohorts (Framingham, Framingham offspring, MESA, & CHS), stratified by number of major risk factors [smoking, diabetes, lipids (LDL≥130 mg/dL or on lipid-lowering therapy), and BP (≥140/90 mmHg or on BP medication)]. Observed vs. predicted 5-year ASCVD event rates were compared across strata. We also evaluated 5-year PCE model performance overall. Results: At the group level, the PCE somewhat overestimated risk across all strata in Kaplan Meier analysis: observed versus expected was 7.0% vs. 11.0%, (p=0.005) for those with no risk factors (N=429), 12.5% vs. 15.4%, (p<0.001) for those with 1 risk factor (N=1,179), 14.7% vs. 18.4 % (p=0.003) for those with 2 risk factors (N=908), and 24.2% vs. 25.1 % (p=0.54) for those with 3+ risk factors (N=151). The Figure shows the performance of the PCE by decile of predicted risk. In the figure, risk-overestimation appears to be driven by over-estimation of risk in the highest deciles of risk, rather than those at low predicted risk. Conclusion: The PCE can be used to risk-stratify older adults. However, it may over-predict risk, especially in those at the top quartile of risk.