The novel phosphodiesterase 9 (PDE9) inhibitor tovinontrine failed to improve a heart failure biomarker in cases with reduced and preserved ejection fraction (HFrEF, HFpEF) in two parallel phase II trials, albeit with some signal of benefit from a low dose.

In the CYCLE-2 PEF trial, N-terminal pro-B-type natriuretic peptide (NT-proBNP) didn’t improve as much with a 50-mg twice-daily oral dose as with placebo over 12 weeks, a 9.4% difference favoring placebo (P = .147).

Change in NT-proBNP trended the right direction with the 50-mg dose in the CYCLE-1 REF trial but still missed that primary endpoint at 12 weeks (-4.4% vs placebo, P = .45), James E. Udelson, MD, of Tufts University in Boston, reported at the Heart Failure Society of America (HFSA) 2025 Annual Scientific Meeting in Phoenix.

However, the HFrEF population showed an inverse dose response with a trend for benefit in two lower-dose groups (-14.1% and -10.8% vs placebo for 2.5 and 25 mg twice daily, P = .07 and .05, respectively).

“Maybe we were increasing cyclic GMP [cGMP] too much,” Udelson said at the late-breaking abstract session where he presented both trials. “And maybe there’s a threshold of myocardial [cGMP] elevation above which counter-regulatory mechanisms limit downstream signaling, such as downregulation of protein kinase G.”

In a post hoc analysis dividing the data by the increase in cGMP, each dosage group did much better when the plasma cGMP rose less than 20%. Placebo-corrected change from baseline in NT-proBNP at week 12 ranged from -14.8% to -18.5% across all three dose groups with cGMP increase less than 20% but from -3.8% to -10.1% among those with greater cGMP increases.

The higher doses had been selected to achieve levels of PDE9 inhibition that maximally increased plasma cGMP “because we usually think that the more inhibition the better,” Udelson said. “In the sacubitril/valsartan study, [cGMP] went up in the early studies and good clinical things happened. So the thinking here was, if we can amplify that signal even more by inhibiting the breakdown, that may enhance the favorable effect of sacubitril/valsartan along that very distinct pathway.”

Beyond unintended consequences, another possibility is that, while NT-proBNP has “led us in the right direction many times,” it might have not been the right choice in this case, said session study discussant Lynne Stevenson, MD, of Vanderbilt University in Nashville.

“If it can be affected by some of these other feedback loops, you could still get benefit to the myocardium, but [that] may not necessarily always show up in that marker,” she said.

Udelson agreed that measuring structural and functional change at 6 or 9 months would have been ideal, but not without having the safety data yet to support it.

The CYCLE-1 REF trial enrolled 529 patients across 19 countries who had HF with an EF of 40% or less (mean, 32%). It randomly assigned patients to 2.5, 25, or 50 mg of tovinontrine or placebo twice daily for 12 weeks.

CYCLE-2 PEF included 277 people in 13 countries who had HF with an EF between 40% and 60% (mean, 50%). Patients were randomly assigned to 50 mg tovinontrine twice daily or placebo for 12 weeks.

Baseline NT-proBNP had to be at least 600 pg/mL at baseline (or 1000 if in atrial fibrillation) in CYCLE-1 REF and at least 300 pg/mL (or 500 in atrial fibrillation) in CYCLE-2 PEF. Both trials required New York Heart Association class 2 or 3 symptoms, with around 40% being in class 3 in both.

Udelson characterized both trial cohorts as “very well treated.” In CYCLE-1, over 90% of patients were on sacubitril/valsartan or another renin angiotensin system (RAS) inhibitor, about 80% on SGLT2 inhibitors or mineralocorticoid receptor antagonists, 95% on beta-blockers and 60% on four guideline-directed medical therapy drugs.

In CYCLE-2 PEF, about one third were on sacubitril/valsartan and most on some other RAS inhibitor, almost 75% on an SGLT2 inhibitor, and half on a mineralocorticoid receptor antagonist.

No differential effects were seen by sacubitril/valsartan background therapy.

Udelson noted that the 40% increase in plasma cGMP with the drug in HFpEF was higher than in the HFrEF study “probably because PDE9 is upregulated more in HFpEF than in HFrEF. And yet this did not lead to any change in NT-proBNP.”

He called the safety findings good, with similar rate of serious adverse events as with placebo, minimal changes in blood pressure, and only small, transient and reversible renal impacts.

The HFrEF “results were convincing enough — though hypothesis generating, of course — that we’re going to go right ahead and conduct another trial with a lower dose range because it looks so promising,” Udelson said. The drug will move to phase IIb with the CYCLE-3 trial in HFrEF starting sometime in 2027, “and then hopefully move on to HFpEF after that,” he said.

Stevenson said she will be interested to see if the preferential benefit in HFrEF is real as trials move forward. “I’m still struggling with the differences between HFpEF and HFrEF,” she said.

The CYCLE trials were funded by Cardurion Pharmaceuticals. 

All investigators or institutions received funding for trial activities from Cardurion Pharmaceuticals or were employees and shareholders of the company. 

Stevenson disclosed paid relationships with the NHLBI and uncompensated relationships with Cytokinetics and Endotronix/Edwards LifeSciences. She reported a spouse or immediate household member having paid relationships with Abbott Medical, Biotronik, Boston Scientific, Johnson & Johnson, and Novartis-EP. 

Crystal Phend is an award-winning medical journalist with decades of experience reporting on clinical research and healthcare developments across specialties. When not walking the halls at a medical conference, she can be found at a keyboard in upstate New York.