This One HFpEF Group May Benefit From Diuretic After All

There is one high-risk phenotype of heart failure with preserved ejection fraction (HFpEF) in particular that may benefit from spironolactone therapy, according to an analysis of the TOPCAT trial.

Patients in the troubled trial could be separated into three HFpEF phenogroups as identified by Julio Chirinos, MD, PhD, of the Hospital of the University of Pennsylvania in Philadelphia, and colleagues in JACC: Heart Failure:

• Younger (mean age 61 years) with mild symptoms and normal left ventricular (LV) geometry (n=1,214)
• Older (mean age 77 years) with stiff arteries, small ventricles, and atrial fibrillation (n=1,329)
• Intermediate-age (mean age 66 years) with obesity, diabetes, and LV hypertrophy (n=899)

Phenotype 1 was at lowest combined risk of cardiovascular death, heart failure hospitalization, and aborted cardiac arrest as compared with people categorized as phenotype 2 (HR 2.17, 95% CI 1.76-2.68) and phenotype 3 (HR 3.44, 95% CI 2.79-4.24).

Phenotype 3 was the sole subgroup to benefit from spironolactone over placebo in TOPCAT in terms of this primary endpoint (HR 0.75, 95% CI 0.59-0.95). The other two phenotypes had no significant reduction in events.

“These findings suggest distinct underlying mechanisms across clinically-identifiable phenogroups of HFpEF, which may benefit from different targeted interventions,” Chirinos and colleagues concluded.

Results were unchanged after excluding people from Eastern Europe, they noted.

Phenotypes 2 and 3 can be said to be “genuine high-risk HFpEF,” according to them, whereas phenotype 1 is “a low-risk group which may not represent genuine HFpEF and may be confounded by lung disease.”

TOPCAT was a large international trial of spironolactone in people over age 50 with symptomatic HFpEF and LV ejection fractions 45% and above, a group that to date has no treatment proven to improve outcomes. The inexpensive drug did not appear to improve clinical outcomes when the trial was first reported in 2013.

Notably, however, the trial’s conduct was questioned after regional variation suggested that participants enrolled in Russia and Georgia did not actually have HFpEF and that many may not have actually even received spironolactone.

“Strikingly, phenogroup 1 had a paucity of obvious HF signs/symptoms and very normal biomarker concentrations and was disproportionately representative of patients enrolled in Eastern Europe, essentially confirming the long-held suspicion that many TOPCAT patients enrolled in this region did not have HFpEF or if they did it was quite mild,” according to an editorial by Tariq Ahmad, MD, MPH, of Yale University School of Medicine, and colleagues.

Supporting the theory that phenogroup 1 was confounded by lung disease was the fact it included the most smokers and the highest levels of MMP-9, a marker of respiratory tract remodeling that is elevated in individuals with asthma and chronic obstructive pulmonary disease, the study authors noted.

Chirinos and colleagues had revisited the TOPCAT database and performed new biomarker analyses on frozen plasma samples saved from the trial. A clustering statistical technique led the investigators to their finding of three HFpEF phenogroups.

Other studies have already proposed various subtypes of HFpEF associated with different clinical outcomes — one identifying six phenogroups and another three phenogroups, Chirinos’ group noted.

A separate TOPCAT analysis recently proposed another three phenogroups of HFpEF using a different methodology, Ahmad’s group pointed out.

Thus, given that one analytic approach can yield different results from another, it’s imperative that researchers understand all the statistical tools before proceeding to clinical application of the data, they advocated.

“Clearly something is amiss in our definition of HFpEF but attempts to redefine the syndrome need more than just the trait of novelty: applying ‘big data’ to the problem of HF has been unable to go beyond confirming what clinicians already know, that HFpEF is a false construct comprised of many diseases,” Ahmad’s group said.

More clinical trials in HFpEF are not needed, they said. Instead, the basics of how heart failure is diagnosed and defined — currently using inexact descriptors of disease such as ejection fraction and New York Heart Association class — need to be revisited, they continued.

“We believe that it is at the intersection of care delivery and the electronic health record where implementation of advanced analytics could fulfil [sic] their dual promise of improving clinical outcomes and redefining HF,” according to the editorialists.