Are Prognostic Models Accurate at Predicting Mortality, Morbidity in Patients With Chronic Heart Failure?

In 2017, heart failure (HF) contributed to 1 in 8 deaths. This inability of the heart to function at its optimal level led to an estimated $30.7 billion in annual healthcare costs in 2012, including healthcare services, medicines, and missed work.¹ Wanting a more complete model through which they could predict mortality and morbidity in patients with chronic HF, the authors of a recent study in JAMA Cardiology used data from several trials to validate their prognostic, predictive model derived from the PARADIGM-HF trial.²

“Predicting risk of death or hospitalizations in patients with HF can allow physicians and patients to make important decisions regarding appropriateness and timing of treatments or the need for end-of-life care,” they noted.

These authors used complete baseline data on 8011 of 8399 patients from the Prospective Comparison of ARNI With ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM-HF) trial to fashion their prognostic model, PREDICT-HF (PARADIGM-HF Risk of Events and Death in the Contemporary Treatment of Heart Failure). To validate their model, they used the Aliskiren Trial to Minimize Outcomes in Patients with Heart Failure Trial (ATMOSPHERE) study and the Swedish Heart Failure Registry (SwedeHF).

Three outcomes were studied:

1. Composite of cardiovascular (CV) death or HF hospitalization at 1 and 2 years (primary outcome)
2. CV death
3. All-cause mortality

The mean (SD) patient age was 64 (11.4) years, 78.2% (6567/8399) were men, and 70.5% (5919/8399) had New York Heart Association class II disease. Patients were excluded if they had left ventricular ejection fraction of 50% or above. During the mean follow-up of 27 months, 1546 patients died while 2031 suffered CV death or were hospitalized for HF.

Several characteristics were common among the patients with adverse outcomes across PREDICT-HF, ATMOSPHERE, and SwedeHF: older age, male sex, higher resting heart rate, evidence of more advanced HF, more comorbidity, worse renal function, and higher natriuretic peptide levels. And among typical lab variables, bilirubin, uric acid, and albumin were common across the 3 outcomes studied. However, N-terminal pro brain natriuretic peptide was the most powerful predictor of them all.

At the 1- and 2-year marks, respectively, the C statistic was as follows for the 3 measured outcomes:

1. Composite endpoint: 0.74 (95% CI, 0.71-0.76) and 0.71 (95% CI, 0.70-0.73)
2. CV death: 0.73 (95% CI, 0.71-0.75) and 0.71 (95% CI, 0.69-0.73)
3. All-cause mortality: 0.71 (95% CI, 0.69-0.74) and 0.70 (95% CI, 0.67-0.74)

ATMOSPHERE validation produced similar results:

1. Composite endpoint: 0.71 (95% CI, 0.69-0.72) and 0.70 (95% CI, 0.68-0.71)
2. CV death: 0.71 (95% CI, 0.69-0.74) and 0.70 (95% CI, 0.69-0.72)
3. All-cause mortality: 0.71 (95% CI, 0.69-0.74) and 0.70 (95% CI, 0.68-0.72)

“Our model, using contemporary patients from around the world, receiving contemporary levels of guideline-recommended therapies and using natriuretic peptides, and that has been externally validated in a real-world cohort, may be considered the strongest of currently available models,” they concluded. “Clinically, the model can be used to predict patient outcomes and has potential wider uses in the organization and targeting of services and therapies in patients with HF-REF [heart failure with reduced ejection fraction.

References

1. Heart failure. CDC website. cdc.gov/heartdisease/heart_failure.htm. Accessed February 3, 2020.

2. Simpson J, Jhund PS, Lund LH, et al. Prognostic models derived in PARADIGM-HF and validated in ATMOSPHERE and the Swedish Heart Failure Registry to predict mortality and morbidity in chronic heart failure [published online January 29, 2020]. JAMA Cardiol. doi: 10.1001/jamacardio.2019.5850.