Tenax Highlights TNX-103 Subgroup Signal After LEVEL Phase 3 Primary Endpoint Miss

Tenax Therapeutics used an Aug. 29 late-breaking presentation at ESC Congress 2026 to highlight a stronger TNX-103 response in patients with lower baseline exercise capacity, three weeks after disclosing that the Phase 3 LEVEL trial failed its overall primary endpoint. The new presentation did not reverse the primary-endpoint miss, but it gave the company a more developed case for focusing future testing on patients with greater disease burden.

LEVEL evaluated oral levosimendan, known as TNX-103, in patients with pulmonary hypertension associated with heart failure with preserved ejection fraction, or PH-HFpEF. Tenax first reported on Aug. 10 that the study had not shown a statistically significant improvement in six-minute walk distance across the full population and had also missed its key Kansas City Cardiomyopathy Questionnaire symptom endpoint.

At ESC, Tenax presented additional completed analyses from the 241-patient trial, which enrolled participants at 41 sites in the United States and Canada. Sanjiv Shah of Northwestern University, LEVEL’s principal investigator, presented the data in a late-breaking pulmonary hypertension session.

Lower baseline walking distance marked the strongest treatment signal

The most prominent finding came from the 119 patients whose baseline six-minute walk distance was below the trial median of 333 meters. At Week 12, the least-squares mean change was an improvement of 23.7 meters with TNX-103 and a decline of 2.6 meters with placebo, producing a placebo-adjusted treatment difference of 26.3 meters. Tenax reported a p value of 0.0112 for that prespecified subgroup analysis.

Symptom scores moved in the same direction in that subgroup. Kansas City Cardiomyopathy Questionnaire total symptom score increased 9.8 points on TNX-103 and 4.1 points on placebo. Tenax noted that a five-point difference is generally regarded as clinically meaningful, although its Aug. 29 release did not present a separate multiplicity-adjusted significance test establishing the subgroup KCCQ result as a confirmatory endpoint.

The company also emphasized two physiologic measures in the same lower-walking-distance population. NT-proBNP, a marker of cardiac wall stress, was reported to fall 47% relative to placebo, with p below 0.0001. Right ventricular systolic pressure showed a placebo-adjusted reduction of 4.9 mmHg, with p equal to 0.009. Tenax explicitly labeled the subgroup analyses of NT-proBNP and RVSP as post hoc, so those findings are supportive signals rather than independently confirmed Phase 3 efficacy endpoints.

Tenax said the completed dataset showed an inverse relationship between baseline walking distance and treatment effect, with larger apparent benefits among participants who entered the study with more limited exercise capacity. Management plans to use that pattern to guide enrollment in the ongoing LEVEL-2 trial. The interpretation remains a company position based on subgroup and exploratory analyses after the full trial missed its primary endpoint.

The full Phase 3 population did not show a walking-distance benefit

The overall result remains the central limitation. In the full 241-patient population, TNX-103 improved six-minute walk distance by a least-squares mean 14.0 meters at Week 12, compared with 10.4 meters for placebo. The placebo-adjusted difference was only 3.5 meters, with p equal to 0.63, well short of statistical significance.

The key secondary endpoint also failed. Kansas City Cardiomyopathy Questionnaire total symptom score improved 6.6 points with TNX-103 and 6.5 points with placebo, a least-squares mean difference of 0.1 point. That result offered no evidence of a treatment effect on the prespecified symptom measure across the trial as a whole.

Other prespecified exploratory measures were more favorable. Tenax’s Aug. 10 topline report said NT-proBNP was reduced 49% relative to placebo in the overall population, with a nominal p value below 0.0001, and RVSP fell by a placebo-adjusted 3.5 mmHg, with a nominal p value of 0.0045. Those analyses were not adjusted to turn the failed primary trial into a positive confirmatory study, and Tenax itself stated at the time that the exploratory analyses did not establish efficacy.

Safety data also need to be read beyond the company’s description of TNX-103 as generally safe and well tolerated. Adverse events were reported in 86.7% of TNX-103 patients and 71.9% of placebo patients, while treatment-related adverse events occurred in 38.3% and 17.4%, respectively. Serious adverse events were almost identical at 10.8% on TNX-103 and 10.7% on placebo. The Aug. 29 presentation added that no significant arrhythmias were observed.

LEVEL therefore produced a mixed dataset rather than a conventional Phase 3 success. The trial did not demonstrate the planned overall improvement in exercise capacity or the key symptom endpoint, but several biomarker, hemodynamic and lower-capacity subgroup analyses pointed in a favorable direction. The distinction matters because a subgroup signal can help design another trial without carrying the same evidentiary weight as a positive primary endpoint in the randomized population.

LEVEL-2 now carries the next test of Tenax’s enrichment strategy

Tenax said it intends to enrich the LEVEL-2 population using what it learned from LEVEL. The current ClinicalTrials.gov record for LEVEL-2 describes a Phase 3, double-blind, randomized study expected to enroll about 540 patients, with participants assigned two-to-one to TNX-103 or placebo. Its listed primary endpoint is change in six-minute walk distance at 26 weeks, a longer efficacy window than the 12-week primary assessment in LEVEL.

The registry currently shows LEVEL-2 as an active study that began in March 2026. Because Tenax has now said the LEVEL findings will be used to enrich the study population, the eventual protocol may differ from the version presently displayed in the registry. The company has not yet publicly disclosed a final revised eligibility threshold based on the 333-meter LEVEL median.

Tenax also said after the Aug. 10 topline announcement that it planned to request a Type C meeting with the U.S. Food and Drug Administration and seek scientific consultation from the European Medicines Agency about changes to the registrational path. The Aug. 29 ESC presentation gives the company a more complete dataset to take into those discussions, but no new FDA agreement on the proposed enrichment strategy was announced with the presentation.

That leaves LEVEL-2 as the key prospective test. If Tenax narrows enrollment toward patients with more limited baseline exercise capacity, the next study will need to show that the subgroup pattern seen in LEVEL can be reproduced in a prospectively defined population. Until then, the 26.3-meter result is a potentially useful signal for trial design, not a substitute for the primary endpoint that LEVEL failed to meet.