PARAGON-HF Frustrates Med Therapy Hopes for Preserved-EF Heart Failure

September 01, 2019

PARIS — Hopes for an evidence-based drug therapy for heart failure with preserved ejection fraction (HFpEF) were dashed once again by the results of PARAGON-HF, which saw no significant advantage to sacubitril/valsartan (Entresto, Novartis) for its primary clinical outcome.

Patients treated with the first-of-its-kind renin-angiotensin system (RAS) inhibitor, compared to those who received standard valsartan, had 13% fewer heart failure (HF) hospitalizations or cardiovascular (CV) deaths over an average of about 3 years, but the difference missed significance at a P value of .059.

Still, PARAGON-HF trialists and some observers point to secondary findings that seemed encouraging for sacubitril/valsartan in some patients. For example, more of those treated with the novel drug than those getting valsartan showed improved New York Heart Association (NYHA) functional class.

And the primary-endpoint reduction in the patients receiving the drug reached significance in women, who made up about 52% of entered patients, and in those with a below-median left ventricular ejection fraction (LVEF) of 45% to 57%, which is classified as mid-range LVEF. 

"The treatment effect was  heterogeneous," said Scott D. Solomon, MD, Brigham and Women's Hospital, Boston, Massachusetts, at a media briefing on PARAGON-HF here at the European Society of Cardiology (ESC) Congress 2019.

In women and in patients with mid-range LVEF, "the absolute risk reduction was similar to that of patients with an ejection fraction under 40%," he said.

The findings suggest that patients with mid-range LVEF "would indeed benefit from sacubitril/valsartan. Both the relative risk reduction and the absolute risk reduction were quite high in those patients," Solomon said.

"I'm not speaking for regulators or for the guidelines, but I suspect that in this group, there's at least some rationale for treating patients with this therapy."

Solomon is lead author on the trial's September 1 publication in the New England Journal of Medicine, timed to coincide with his presentation of the trial here at the ESC Congress.

Focus on Subgroups

"Although the overall effect of sacubitril/valsartan was modest, the subgroup analysis to me is compelling, and suggests that it is possible to reduce heart failure hospitalization in a key group, particularly the HFpEF patient with a structural abnormality on echo, including at least some modest abnormality of left ventricular function," said Stuart J. Connolly, MD, McMaster University, Hamilton, Ontario, Canada, an invited discussant following Solomon's formal presentation of the trial here.

"The idea that there is actually a true benefit here is supported by several secondary endpoints," he said. The improvement in NYHA class, quality-of-life measures, and less worsening of renal function in the sacubitril/valsartan group "are all supportive that this is a real effect."

At the media briefing, several panelists agreed. "To me, PARAGON-HF was positive in those patients that had an LVEF of, say, 50%," said Deepak L. Bhatt, MD, Brigham and Women's Hospital, who was not a PARAGON-HF investigator.

"The way I interpret the trial, the drug worked in those patients that had an ejection fraction that was low, but not what we conventionally call reduced ejection fraction."

Table 1. Rate Ratio for Clinical Endpoints in PARAGON-HF: Sacubitril/Valsartan vs Valsartan

Endpoints Rate Ratio (95% CI)
Primary: Total HF hospitalizations or CV death 0.87 (0.75 - 1.01) (P = .059)
HF hospitalizations 0.85 (0.72 - 1.00)
CV death 0.95 (0.79 - 1.16)
Death from any cause 0.97 (0.84 - 1.13)
Median follow-up, 35 months.


Also at the briefing, P. Gabriel Steg, MD, Hôpital Bichat, Paris, France, pointed out, "The unmet need in this population is absolutely huge."

That should justify some perspective about the primary endpoint's degree of significance, observed Steg, who was not affiliated with PARAGON-HF. It can be "misleading" to make a big distinction between a P value of .059 and one that edges below .05 into significance, he said. The trial "I think is a very important advance in the field."

Currently No Evidence-Based Med for HFpEF

Indeed, optimism about the trial's encouraging secondary endpoints stems in part from the field's hunger for a "positive" randomized drug trial in patients with HFpEF. No medication has reduced mortality in such patients, although prior trials have pointed toward some important clinical benefits.

The benefits in those trials — including TOPCAT for the aldosterone inhibitor spironolactone and CHARM-Preserved for candesartan, an angiotensin receptor blocker (ARB) — extended to patients with mid-range LVEF, said PARAGON-HF investigator John J.V. McMurray, MB, MD, University of Glasgow, United Kingdom, at the briefing.

"You do see that drugs that work in patients with heart failure with reduced ejection fraction seem to have some benefit, perhaps a bit less, but benefit when patients have a below-normal ejection fraction."

Against that background, it seemed plausible to Connolly and panelists at the briefing that the nonsignificant primary-endpoint result for sacubitril/valsartan in the trial, given the secondary findings going in the same direction, reflected a trend toward real benefit from the drug, especially in the mid-range LVEF group.

Table 2. Adjusted Rate Ratio for Primary Endpoint in PARAGON-HF by Subgroups Sex and LVEF Category

Subgroup Rate Ratio (95% CI)
Women (vs men) 0.73 (0.59 - 0.90)
LVEF below 57% median 0.78 (0.64 - 0.95)


But the same thinking didn't necessarily apply to women, who made up 52% of the trial population and was the other subgroup showing a significant benefit from sacubitril/valsartan, in this case compared to men.

"It is at this point speculative as to why we saw this finding — although we don't discount it, because it was very dramatic," said Solomon. In contrast, "We clearly understand the ejection fraction finding because there's so much biologic plausibility there."

Secondary Endpoints: Other Perspectives

"These findings do not support the routine clinical use of sacubitril-valsartan in patients with HFpEF" Gregg C. Fonarow, MD, University of California, Los Angeles, who was not part of the trial, told | Medscape Cardiology by email.

On the other hand, "the findings for the subgroup of patients with LVEF 45-57% are intriguing. This raises that possibility for further consideration of use of sacubitril/valsartan in patients with heart failure with ejection fraction in the 41% to 55% range, although further study is needed," said Fonarow.

Indeed, that the primary-endpoint difference was nonsignificant "tempers our enthusiasm to embrace any subgroup analyses, even if prespecified," Clyde W. Yancy, MD, MSc, Northwestern University, Chicago, Illinois, who was not involved with PARAGON-HF, told | Medscape Cardiology.

"Under these circumstances, the subgroup analyses are truly exploratory and are insufficient to drive clinical practice," he said by email.

Table 3. Rate Ratio for Secondary Parameters in PARAGON-HF, Sacubitril/Valsartan vs Valsartan

Parameter Rate Ratio (95% CI)
Improvement in NYHA class 1.45 (1.13 - 1.86)
>5 point improvement in KCCQ clinical summary score 1.30 (1.04 - 1.61)
Worsening renal function 0.50 (0.33 - 0.77)
KCCQ = Kansas City Cardiomyopathy Questionnaire


"It will be fascinating to see how clinicians interpret these data. As doctors, we treat patients who don't feel well, and when patients mirror scenarios for which there are suggestive data, for example, women with HFpEF or those with lower-range ejection fraction, perhaps there will be use, off-label, of the ARNI compound," Yancy said, referring to sacubitril/valsartan by the abbreviation for its drug class, an angiotensin-receptor/neprilysin inhibitor.

PARAGON-HF randomly assigned 4822 patients with HFpEF, which it defined as HF with an ejection fraction of at least 45%, and in NYHA class II to IV, to receive sacubitril/valsartan or the ARB valsartan for up to 57 months.

It was preceded by a single-blind run-in phase in which all entered patients received valsartan at half the target dose, followed by sacubitril/valsartan at half the target dose. Those with "no unacceptable side effects" on either drug proceeded to the double-blind randomized phase.

Fonarow said the trial's entry criteria, the "voluntary nature of trial enrollment," and dropouts from the run-in phase "resulted in a patient population at lower event risk than those in clinical practice."

But Solomon said that enrollment in the study required patients to have documented structural heart disease, that is, either left atrial enlargement or left ventricular hypertrophy, along with elevated natriuretic peptides. "So we were pretty confident that they in fact did have heart failure. I think we got the right patients."

Similar proportions of patients discontinued their assigned medication because of adverse events, not including death: 15.4% vs 16.2% for the valsartan group.

The Way Forward Studying HFpEF

"Overall these data suggest that HFpEF is not just heterogeneous with respect to phenotype, but also heterogeneous with respect to treatment," Solomon told | Medscape Cardiology. "And I think what we're learning is that one size might not fit all in HFpEF."

Increasingly there are calls to explore therapies in specific HFpEF phenotypes, rather than in the broad spectrum of patients with HFpEF defined only by LVEF that isn't reduced.

"I think you're going to see more and more of that in the future as we get better at phenotyping these patients," Solomon said.

Yancy proposed that as the next line of research in HFpEF. "HFpEF is not one disease," he noted.

"Our next investigational efforts for HFpEF should evaluate an 'approach,' not a therapy. It is the heterogeneity that must be dissected now, before therapy can be determined," he said. "Parse the phenotypes and then test therapies accordingly."  

That means most HFpEF cases may not ultimately be treatable with one type of intervention, Yancy said.

Fonarow took another view. Many will see PARAGON-HF results as "reflecting the heterogeneity of HFpEF, and that it is unreasonable to expect any medical therapy will be found to provide benefit unless deep phenotyping has occurred and a specific phenotype is precisely targeted. Yet, it remains to be seen whether this type of approach will yield success," he said.

In heart failure with reduced ejection fraction, he noted, "there is substantial heterogeneity in etiologies, clinical presentation, comorbidities, echocardiographic findings, and biomarkers, yet multiple medical therapies have been shown to substantially reduce all-cause mortality."

PARAGON-HF was sponsored by Novartis. Disclosures for Solomon, McMurray, and the other authors can be found at Yancy has previously disclosed an unspecified relationship with Abbott Laboratories. Fonarow discloses consulting and research involvement for Novartis and consulting for Abbott, Amgen, Bayer, CHF Solutions, Janssen, and Medtronic. Connolly has previously disclosed receiving consultant fees/honoraria from Sanofi-Aventis, Bristol-Myers Squibb, Boehringer-Ingelheim, Bayer, and Portola and research grants from Bayer and Portola. 

European Society of Cardiology (ESC) Congress 2019.  Presented September 1, 2019.

N Engl J Med. Published online September 1, 2019. Abstract

Follow Steve Stiles on Twitter: @SteveStiles2. For more from, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.