Insights from a Systematic Examination of a Novel Heart Failure Risk Prediction Tool: The Pooled Cohort Equations to Prevent Heart Failure

Abstract

IntroductionEarly identification of individuals at risk for heart failure (HF) in the community would facilitate targeted preventive interventions. We recently developed a new tool to predict incident HF in the general population: the Pooled Cohort equations to Prevent Heart Failure (PCP-HF). The PCP-HF tool was derived and validated using individual-level data from seven diverse US population-based cohorts. A better understanding of the intrinsic properties of the multivariable PCP-HF tool would be useful to clinicians to allow recognition of risk factor profiles resulting in high or low predicted risk and to clarify the inherent limitations of the tool.ObjectiveThis study aimed to examine the PCP-HF tool systematically to determine the impact of varying risk factor patterns on predicted HF risk in different age, sex, and race groups.MethodsThe PCP-HF tool integrates age and clinical risk factors into a multivariable sex- and race-specific risk model to estimate absolute 10-year risk of incident HF among adults aged 30-79 years. We created spreadsheets for each race-sex group, incorporating the coefficients for each risk factor and interaction term based on the published formulas to create 10-year predicted HF risk score estimates for a hypothetical white man, black man, white woman, and black woman. We varied individual risk factors and combinations of risk factor levels using the PCP-HF risk tool and examined the range of predicted 10-year HF risks.ResultsThe 10-year predicted HF risk was substantially higher with older age when other risk factors were held constant at age-adjusted national mean values for each race-sex group (Figure). Risk for HF was significantly higher in current smokers and among those requiring treatment for hypertension or diabetes. Predicted 10-year HF risk in a hypothetical 60-year old varied from 1.9% to 31.9% in a white male, 1.7% to 31.7% in a black male, 1.4% to 27.9% in a white female, and 1.3% to 33.0% in a black female. There were significant race-sex interactions with different risk factors which impacted the predicted 10-year risk of HF.ConclusionsThis analysis provides context of specific risk factor patterns that identify individuals who are likely to be at high risk of HF. Our analysis highlights the importance of a race and sex-specific multivariable risk prediction model for HF to guide the clinician-patient discussion. However, younger adults tend to have low 10-year predicted HF risk in spite of high-risk factor levels. Early identification of individuals at risk for heart failure (HF) in the community would facilitate targeted preventive interventions. We recently developed a new tool to predict incident HF in the general population: the Pooled Cohort equations to Prevent Heart Failure (PCP-HF). The PCP-HF tool was derived and validated using individual-level data from seven diverse US population-based cohorts. A better understanding of the intrinsic properties of the multivariable PCP-HF tool would be useful to clinicians to allow recognition of risk factor profiles resulting in high or low predicted risk and to clarify the inherent limitations of the tool. This study aimed to examine the PCP-HF tool systematically to determine the impact of varying risk factor patterns on predicted HF risk in different age, sex, and race groups. The PCP-HF tool integrates age and clinical risk factors into a multivariable sex- and race-specific risk model to estimate absolute 10-year risk of incident HF among adults aged 30-79 years. We created spreadsheets for each race-sex group, incorporating the coefficients for each risk factor and interaction term based on the published formulas to create 10-year predicted HF risk score estimates for a hypothetical white man, black man, white woman, and black woman. We varied individual risk factors and combinations of risk factor levels using the PCP-HF risk tool and examined the range of predicted 10-year HF risks. The 10-year predicted HF risk was substantially higher with older age when other risk factors were held constant at age-adjusted national mean values for each race-sex group (Figure). Risk for HF was significantly higher in current smokers and among those requiring treatment for hypertension or diabetes. Predicted 10-year HF risk in a hypothetical 60-year old varied from 1.9% to 31.9% in a white male, 1.7% to 31.7% in a black male, 1.4% to 27.9% in a white female, and 1.3% to 33.0% in a black female. There were significant race-sex interactions with different risk factors which impacted the predicted 10-year risk of HF. This analysis provides context of specific risk factor patterns that identify individuals who are likely to be at high risk of HF. Our analysis highlights the importance of a race and sex-specific multivariable risk prediction model for HF to guide the clinician-patient discussion. However, younger adults tend to have low 10-year predicted HF risk in spite of high-risk factor levels.