Tenax Therapeutics, Inc. has presented additional Phase 3 LEVEL data suggesting that TNX-103, its oral formulation of levosimendan, produced its clearest clinical benefit in patients with pulmonary hypertension associated with heart failure with preserved ejection fraction who entered the study with greater functional impairment. The August 29 presentation at the European Society of Cardiology Congress in Munich showed a 26.3-meter placebo-adjusted improvement in six-minute walk distance among 119 participants whose baseline walking distance was below the trial median of 333 meters. The same subgroup showed substantial changes in cardiac-wall-stress and pulmonary-pressure measures, potentially giving Tenax a way to refine the population enrolled in its next Phase 3 study.
The results need careful interpretation because LEVEL itself was not a positive Phase 3 trial. In the full 241-patient population, TNX-103 failed to achieve statistical significance on its primary endpoint of six-minute walk distance and also missed the key secondary endpoint measuring symptoms through the Kansas City Cardiomyopathy Questionnaire total symptom score. Tenax is therefore not presenting a conventional confirmatory success; it is using prespecified subgroup and biomarker analyses to argue that a drug effect may have been diluted by enrolling patients whose disease was not severe enough for a functional benefit to become measurable over 12 weeks.
Why did the overall Phase 3 LEVEL trial fail despite the stronger subgroup results?
LEVEL randomized 241 patients across 41 sites in the United States and Canada to oral TNX-103 or placebo. Patients initially received 1 mg twice daily, with titration to three times daily beginning at week five when tolerated, while the primary efficacy analysis examined change in six-minute walk distance at week 12. In the overall population, least-squares mean walking distance improved by 14.0 meters with TNX-103 and 10.4 meters with placebo, leaving a placebo-adjusted difference of only 3.5 meters and a p-value of 0.63.
The symptom endpoint was similarly neutral. Kansas City Cardiomyopathy Questionnaire total symptom scores improved by 6.6 points with TNX-103 and 6.5 points with placebo, producing essentially no separation between the two arms. Those results explain why Tenax cannot simply characterize LEVEL as positive based on the newer European Society of Cardiology presentation.
The more interesting pattern appeared when baseline exercise capacity was examined. Among patients walking less than 333 meters when they entered the study, TNX-103 recipients improved by 23.7 meters while placebo recipients declined by 2.6 meters, producing the 26.3-meter treatment difference highlighted at the August 29 congress presentation. The company also reported that treatment effect diminished progressively as baseline walking ability increased, suggesting that patients capable of walking relatively long distances at enrollment may have introduced a ceiling effect that reduced the study’s ability to distinguish drug from placebo.
What do NT-proBNP and pulmonary-pressure results say about TNX-103’s biological activity?
Tenax’s strongest argument for continuing development may come from measurements that do not depend on how far a patient chooses or is able to walk on a particular day. Across the overall LEVEL population, TNX-103 produced a 49% placebo-adjusted reduction in NT-proBNP, a biomarker associated with cardiac wall stress, alongside a 3.5 mmHg placebo-adjusted reduction in right ventricular systolic pressure. These were exploratory endpoints rather than the study’s primary measures, so they cannot reverse the statistical failure of the primary analysis, but they provide evidence that the drug was producing measurable cardiovascular effects.
In the subgroup below the 333-meter baseline walking threshold, the August 29 analysis showed a 47% reduction in NT-proBNP compared with placebo and a 4.9 mmHg placebo-adjusted reduction in right ventricular systolic pressure. Kansas City Cardiomyopathy Questionnaire symptom scores improved by 9.8 points with TNX-103 compared with 4.1 points on placebo in this group.
The consistency matters because PH-HFpEF is a complex syndrome in which pulmonary vascular pressure, cardiac filling abnormalities, exercise intolerance and multiple comorbidities intersect. A treatment capable of reducing cardiac stress and pulmonary pressure while also improving exercise capacity in the more impaired population provides a biologically coherent signal, even if LEVEL did not establish efficacy in the population originally randomized.
Why is PH-HFpEF such a difficult drug-development target?
Heart failure with preserved ejection fraction is already biologically heterogeneous, and adding pulmonary hypertension creates another layer of complexity. Patients can have varying degrees of left-heart disease, pulmonary vascular remodeling, obesity, atrial fibrillation, renal dysfunction and other conditions that influence symptoms and exercise performance.
Tenax noted that LEVEL participants were generally older, symptomatic and receiving substantial guideline-directed medical therapy, including mineralocorticoid receptor antagonists, diuretics and sodium-glucose cotransporter-2 inhibitors. The company’s analyses suggest TNX-103’s effect became more apparent as disease burden increased.
This presents a familiar biotechnology dilemma. Broad inclusion criteria can make a drug applicable to a larger future market, but they can also increase biological heterogeneity and dilute an effect concentrated in a particular disease phenotype. Enriching the next trial could improve the probability of statistical success while simultaneously narrowing the eventual label.
How could Tenax redesign LEVEL-2 around the new findings?
Tenax said it intends to enrich the LEVEL-2 population using evidence generated by LEVEL and to do so within established U.S. Food and Drug Administration principles for prospectively defined trial populations. The company had already indicated after the initial August 10 topline results that it planned regulatory discussions concerning revisions to the ongoing registrational program.
A straightforward approach would be to impose a lower maximum baseline six-minute walk distance, thereby concentrating enrollment among patients resembling the subgroup in which the 26.3-meter difference appeared. Other markers of disease burden could also potentially be incorporated if agreed with regulators.
The trade-off is important. Every enrichment criterion that increases trial sensitivity may reduce the number of patients represented by the future indication. The optimal design is therefore not necessarily the narrowest population showing the largest retrospective effect, but one that can be prospectively defined, clinically justified and enrolled at sufficient scale.
What makes levosimendan different from conventional PH-HFpEF development programs?
Levosimendan is a calcium sensitizer and potassium-channel activator that has been used intravenously for acute decompensated heart failure in numerous countries, although it is not approved in the United States or Canada. Tenax owns global rights and is developing oral TNX-103 specifically for PH-HFpEF.
That existing pharmacology gives the program a different risk profile from a completely novel molecule. The fundamental uncertainty is less about whether levosimendan has cardiovascular biological activity and more about whether chronic oral dosing can generate a sufficiently large and reproducible clinical benefit in a defined PH-HFpEF population.
LEVEL has made that question more complicated rather than answering it. The full population failed, but the trial also generated a gradient suggesting that baseline disease severity may determine whether functional benefit can be detected.
What should investors and clinicians watch next in the TNX-103 program?
The most important development will be the finalized LEVEL-2 protocol and the outcome of Tenax’s discussions with the U.S. Food and Drug Administration. A prospectively enriched Phase 3 study that reproduces the subgroup effect would materially strengthen the program because the analysis would no longer depend on interpreting slices of a failed primary trial.
Failure to reproduce the effect would have the opposite implication. Biomarker changes can support biological plausibility, but an approvable therapy ultimately requires robust clinical evidence within the population specified before the confirmatory trial begins.
The August 29 presentation therefore does not erase LEVEL’s miss. It does something more useful for a development-stage biotechnology company: it provides a testable hypothesis about why the trial failed and identifies a patient population in which the next study can determine whether that explanation was correct.
