ReachMD
Be part of the knowledge.™Validated risk prediction tool for incident heart failure for use in primary care setting
onlinejacc.org
The prevalence of heart failure (HF) continues to increase globally, due to the growing aging population, the increasing burden of HF risk factors such as obesity and diabetes, and advances in secondary prevention treatments for coronary heart disease [1-3]. Considering the significant morbidity, mortality and cost of care associated with HF, it is crucial to focus on prevention.
To date, no validated race- and sex-specific risk prediction tools for incident HF are available to guide personalized risk-based decision making [4]. Existing HF risk prediction models have limitations, such as being based on nonrepresentative, noncontemporary or restricted higher-risk sample, with limited ethnic diversity [5]. Moreover, many risk models have not extensively been externally validated, thus have limited applicability and implementation in a general population without prevalent CV disease.
This study used pooled individual data from five diverse population-based cohorts (ARIC, CARDIA, CHS, FOF and MESA: pooled cohorts: PC), to develop equations to estimate the 10-year risks of incident HF using routinely available clinical data in a broad general population (30 to 79 years of age) of white and black adults without prevalent CVD. Subsequently, the Pooled Cohort equations to Prevent HF (PCP-HF) were assessed for external validity in two additional large population-based cohorts (PREVEND – white participants, Jackson Heart study – black participants).
HF risk factors considered for development of the PCK-HF model were: age, systolic blood pressure (SBP), antihypertension medication usage, body mass index (BMI), total cholesterol (TC) and HDL-C levels, current smoking status, fasting glucose, diabetes medication usage, and electrocardiogram measurement of QRS duration.
Main results
- During a 10-year follow-up, 1339 HF events occurred in the derivation cohort.
- The PCP-HF had good-to-excellent discrimination in the derivation and internal validation (PC) and in the external validation cohorts (PREVEND and JHS).
- The Cs-statistic in the white men and women, and black men and women in de PC internal validation sample, were 0.79 (95%CI: 0.76-0.81), 0.71 (95%CI: 0.63-0.80), 0.85 (95%CI: 0.82-0.88) and 0.78 (95%CI: 0.71-0.85) for the respective sex-race groups.
- External validation demonstrated good discrimination (C-statistic ranged from 0.71 to 0.88 in black and white men and women) and strong calibration.
- Sex- and race-specific 10-year PCP-HF risk estimates for a 50-year-old with a high-risk profile (current smoker, SBP: 150 mmHg, fasting glucose: 126 mg/dL, treatment for hypertension and diabetes, BMI: 35 kg/m², TC: 250 mg/dL, HDL: 30 mg/dL, and QRS: 120 ms) ranged from the lowest risk in white women (14%), to highest in black men (24%) and women (26%). White men showed 22% risk. In 50-year old participants with an optimal or intermediate risk factor profile, all groups showed <3% risk.
- A web-based PCP-HF risk tool has been developed, to estimate 10-year risk of HF for adults aged 30 to 79 years without prevalent CVD.
Conclusion
This study describes development and validation of sex- and race-specific 10-year risk prediction models for incident HF, using clinical variables that are routinely available in primary care setting. The PCP-HF tool was derived from a cohort representative of the general US population, without prevalent CVD. Its clinical utility of the PCP-HF tool in risk assessment was confirmed in a general primary prevention population. Thus, the PCP-HF score can serve as an initial inexpensive screening tool to identify individuals in the primary care setting who are at higher risk for development of incident HF and who may benefit from additional screening measures to refine and personalize risk assessment.
Editorial comment
Cleland et al. [6] note that everyone may be destined to develop HF, unless they die from something else first, and unless the aging process itself can be reversed or stopped. Postponing HF however, is a realistic and worthwhile goal and indeed treating risk factors has increased the average age at which HF occurs, but age will eventually take its toll.
The authors note that procrastination and prevention differ fundamentally in resource requirements and economic consequences. ‘Only if there is “compression” of morbidity and mortality (i.e. fewer disability-adjusted life-years [DALY]), will postponing the onset of disease reduce costs.’ Strategies to reduce DALY due to HF include enhancing health education and awareness in the general population. Another strategy is to identify and target individuals who are at high risk of developing HF and who might benefit from additional intervention beyond routine management of risk factors. Risk prediction models to do so should not only predict individuals at high risk, but should also detect a large proportion of those who develop HF. Another strategy might be to manage HF better once it develops, by detecting and treating it before it leads to substantial disability or by developing more effective interventions.
About Khan et al.’s efforts to develop a model to predict incident HF they note that the study population comprised of cohorts that enrolled people between 1985 and 2000, which may no longer reflect contemporary practice. Moreover, based on the observation that the decile of patients at highest risk shows an incidence of only 12%, they conclude that the very high-risk groups in the study must apply to very, very few people, ‘which renders a strategy of targeted intervention based on this model unlikely to have an impact on the population incidence of HF.’
A stronger criticism of Cleland and colleagues of the analysis is that case-ascertainment was heterogeneous, was often subjective and lacked sensitivity. Part of the problem is that there is no generally agreed definition of HF, but rather the diagnosis is formed based on a clinical interpretation of subjective criteria that are difficult to verify and validate in retrospect. Cleland et al. propose it is time to create a first universal definition of HF. ‘A universal definition of HF based on objective and verifiable measurements (biomarkers and imaging) that are deranged early in the course of disease and dissociated from subjective criteria will enable earlier intervention.’ Trials that evaluate the effect of interventions based on such a universal definition are underway.
References
1. Benjamin EJ, Virani SS, Callaway CW, et al. Heart disease and stroke statistics-2018 update: a report from the American Heart Association. Circulation 2018;137:e67–492.
2. Gerber Y, Weston SA, Redfield MM, et al. A contemporary appraisal of the heart failure epidemic in Olmsted County, Minnesota, 2000 to 2010. JAMA Intern Med 2015;175:996–1004.
3. Tsao CW, Lyass A, Enserro D, et al. Temporal trends in the incidence of and mortality associated with heart failure with preserved and reduced ejection fraction. J Am Coll Cardiol HF 2018;6:678–85.
4. Karmali KN, Lloyd-Jones DM. Implementing cardiovascular risk prediction in clinical practice: the future is now. J Am Heart Assoc 2017;6:e006019.
5. Echouffo-Tcheugui JB, Greene SJ, Papadimitriou L, et al. Population risk prediction models for incident heart failure: a systematic review. Circ Heart Fail 2015;8:438–47.
6. Cleland J, Pellicori P and Clark AL. Prevention or Procrastination for Heart Failure? Why We Need a Universal Definition of Heart Failure. J Am Coll Cardiol. 2019: 73(19). DOI: 10.1016/j.jacc.2019.03.471
Facebook Comments