Oral Levosimendan Misses Primary Endpoint in Phase III Trial of HFpEF-Related Pulmonary Hypertension

Prespecified subgroup analyses identified a substantial beneficial treatment effect in patients with greater disease burden, supported by clinically meaningful changes in predefined cardiac biomarker and pulmonary hemodynamic measures across the overall trial population. TNX-103 was generally safe and well tolerated, with serious adverse events and adjudicated clinical worsening events balanced across treatment arms. Based on the results of the LEVEL trial, Tenax intends to request a Type C Meeting with the U.S. Food and Drug Administration (FDA) to discuss revisions to the ongoing registrational development of TNX-103 for the treatment of PH-HFpEF.

“Having reviewed the available trial results, I am encouraged we have a clear path forward. The prespecified subgroup analyses demonstrate that there is a meaningful beneficial treatment effect of TNX-103 in patients with more disease burden, which is the patient population with the greatest unmet need. Our entry criteria enrolled a broad population, including too many patients with less severe disease. In a prespecified subgroup analysis of patients who walked less than the trial median of 333 meters at baseline, the improvement in the levosimendan group compared to placebo was 26.3 meters (95% confidence interval 6.0, 46.7 meters), demonstrating a strong result in a more characteristic HFpEF population that needs this therapy. This association is corroborated by a 49% decrease in NT-proBNP compared to placebo (p=0.0001, nominal) and a clinically meaningful reduction in right ventricular systolic pressure of 3.5 mmHg, compared to placebo (p=0.0045, nominal). Taken together, these data support TNX-103’s substantial cardiovascular and pulmonary biologic effects on reducing the severity of disease in patients with pulmonary hypertension due to HFpEF,” said Stuart Rich, MD, Chief Medical Officer of Tenax Therapeutics.

“Although the result in the overall population was not statistically significant, the LEVEL trial is an important advancement for patients with pulmonary hypertension due to HFpEF, a disease that still has no approved therapy. What stands out to me is the marked reduction in NT-proBNP, the most established marker of cardiac wall stress and prognosis in heart failure, which was reduced by an extraordinary 49% relative to placebo. The treatment effect on NT-proBNP is larger than any prior HFpEF trial, and it was accompanied by improvements in multiple echocardiographic measures of cardiac structure and function, spanning right ventricular systolic function, left atrial function, left ventricular mass and diastolic filling, and pulmonary pressure. In my opinion, a drug that lowers wall stress and modifies cardiac structure and function this robustly is likely to be disease-modifying. The improvement in exercise capacity was concentrated in patients with lower baseline walk distance and more advanced disease, which is consistent with a therapy that may improve long-term outcomes in those patients who need it most,” said Sanjiv Shah, MD, Director of the HFpEF Program at Northwestern University Feinberg School of Medicine and LEVEL’s Principal Investigator.

Key Results from LEVEL

LEVEL (NCT05983250) was a registrational Phase 3, double-blind, randomized, placebo-controlled clinical trial evaluating TNX-103 in patients with PH-HFpEF across 41 sites in the United States and Canada. 241 patients were randomized to TNX-103 1 mg twice daily, titrated to 1 mg three times daily starting at Week 5 as tolerated, versus placebo. The primary endpoint was change in 6-minute walk distance (6MWD) at Week 12, and a key secondary endpoint was change in Kansas City Cardiomyopathy Questionnaire total symptom score (KCCQ-TSS). N-terminal pro-B natriuretic peptide (NT-proBNP), a measure of cardiac wall stress, and right ventricular systolic pressure (RVSP) by echocardiography, were exploratory endpoints. Patients completing the double-blind period were eligible to enter an open-label extension of up to two years, which is ongoing.

• 6MWD at Week 12: least-squares means +14.0 meters (SE 7.7) on TNX-103 (n=116) versus +10.4 meters (SE 7.6) on placebo (n=120); least-squares mean difference 3.5 meters (SE 7.3), p=0.63. The primary endpoint was not met. Estimates are from a mixed model for repeated measures in all 241 randomized patients (120 TNX-103, 121 placebo), adjusted for baseline 6MWD and randomization strata with missing Week 12 values handled by the prespecified estimand strategies.

• Among patients with an observed Week 12 walk test, the mean change from baseline was +17.7 meters (SD 40.2) on TNX-103 (n=116) versus +9.8 meters (SD 54.7) on placebo (n=120), an unadjusted difference of +8.0 meters.

• KCCQ-TSS at Week 12: least-squares mean +6.6 (SE 2.4) on TNX-103 (n=114) versus +6.5 (SE 2.3) on placebo (n=120); least-squares mean difference 0.1 points (SE 2.2).

Prespecified Analyses of Primary Endpoint

• Baseline 6MWD below the trial median of 333 meters (n=119): least-squares mean treatment difference +26.3 meters (95% confidence interval 6.0, 46.7), nominal p=0.0112. On average, patients on TNX-103 in this sub-group improved 26.7 meters while patients on placebo declined 2.6 meters.

• Patients aged 71 years and older: least squares mean treatment difference +27.1 meters (95% confidence interval 9.9, 44.4), nominal p=0.0021. In the oldest tertile (>74 years): +37.6 meters (95% confidence interval 14.7, 60.5), nominal p=0.0013. Mean age was 71.8 years among patients walking less than 333 meters at baseline, versus 66.5 years among those at or above that threshold.

Prespecified Exploratory Biomarker and Hemodynamic Analyses

• NT-proBNP (overall population): geometric least-squares mean ratio of 0.51 versus placebo (95% confidence interval 0.43, 0.60), nominal p<0.0001, corresponding to a 49% reduction compared to placebo. On average, the change from baseline was −39.1 pmol/L (SD 78.6) on TNX-103 (n=116) versus +13.7 pmol/L (SD 81.0) on placebo (n=120).

• RVSP by echocardiography (overall population): placebo-controlled reduction of 3.5 mmHg on TNX-103 (95% confidence interval -6.01, -1.10), nominal p=0.0045. In patients with a baseline 6MWD below 333 meters the reduction was 4.9 mmHg (nominal p=0.009).

Post hoc Analysis of Primary Endpoint

• By baseline 6MWD quartile, the treatment difference diminished consistently as baseline exercise capacity increased: +32.4 meter improvement in the first quartile (≤264 meters), +21.4 meters in the second (>264–333 meters), −10.7 meters in the third (>333–384 meters) and −27.3 meters in the fourth (>384 meters).

Safety

TNX-103 was generally safe and well-tolerated in the LEVEL trial, and no new safety signals were observed.

• Adverse events were reported in 86.7% of patients on TNX-103 versus 71.9% on placebo. Treatment-related adverse events were reported in 38.3% of patients in the treatment group versus 17.4% on placebo.

• Serious adverse events were balanced across arms, at 10.8% on TNX-103 and 10.7% on placebo. Adverse events of special interest occurred in 9.2% versus 5.8%.

The biomarker and hemodynamic analyses described above were prespecified in the statistical analysis plan; the quartile analysis was post hoc. Nominal p-values are not adjusted for multiplicity and these analyses do not establish efficacy. Multivariable analyses are ongoing.

“We believe the data from LEVEL provide new information that will guide us toward approval. Our mission is to bring the first approved therapy to patients with PH-HFpEF. The data from LEVEL provide us with a key insight into the population that exhibits the treatment effect of TNX-103. These trial results also confirm we had the appropriate dosing in HFpEF patients, illustrate a very favorable overall pharmacokinetic profile of oral levosimendan, and demonstrate the safety of oral levosimendan. Based on the results from LEVEL, we are moving quickly to strengthen the development plan for TNX-103 to generate data we believe will support regulatory submissions,” said Chris Giordano, President and Chief Executive Officer of Tenax Therapeutics.

Tenax intends to request a Type C meeting with the FDA to present the complete LEVEL dataset together with the Company's recommendations, and in parallel to seek scientific consultation from the European Medicines Agency. The Company intends to enrich the study population to ensure a robust treatment effect, as demonstrated in the subgroup findings observed in LEVEL. We will discuss an enrichment strategy in the Type C meeting, along with the FDA’s previous agreement that a single Phase 3 trial with a p-value of 0.01 would be sufficient for an NDA submission for TNX-103 for the treatment of PH-HFpEF.

Full results from LEVEL will be presented in a Late-Breaking Clinical Science session at the European Society of Cardiology (ESC) Congress 2026, being held August 28-31 in Munich, Germany.