The Phase 3 trial missed its primary endpoint. Then the data pointed somewhere else.
Tenax Therapeutics (Nasdaq: TENX) read out its registrational LEVEL study of oral levosimendan on August 10. Patients with pulmonary hypertension due to heart failure with preserved ejection fraction (PH-HFpEF) did not walk measurably farther than placebo at twelve weeks. The stock-focused headline was a miss. The trial itself told a more interesting story, and that story is now the company's entire path forward.
The drug's long road from the ICU to the clinic
Levosimendan is not new. The intravenous form has held market authorization in sixty countries for decades, used in hospitalised patients with acutely decompensated heart failure. It is not available in the United States or Canada. Tenax owns the global rights and spent years repurposing the molecule for a different problem: a calcium sensitiser and potassium-channel activator moved from the emergency room to a twice-daily pill.
TIMELINE: Levosimendan and the PH-HFpEF programme
─────────────────────────────────────────────────────────────
IV era HELP LEVEL LEVEL-2
60 countries Phase 2 Phase 3 Global Phase 3
[emoji] ──── [emoji] ──── ◉ NOW ──── [emoji]
Acute HF PH-HFpEF Readout Enriched design
Source: Tenax Therapeutics disclosures, 2026
The Phase 2 HELP study showed that intravenous levosimendan improved exercise capacity in PH-HFpEF patients. That gave the company the signal to run LEVEL, a double-blind, placebo-controlled trial across 41 sites in the United States and Canada. It randomised 241 patients to the pill, titrated up from one milligram twice daily to three times daily, against placebo.
Where the trial missed, and where it landed
The primary endpoint was change in six-minute walk distance at week 12. The treated group walked 14.0 meters farther than baseline; placebo walked 10.4. The gap was not statistically significant. The key secondary measure, a symptom score called KCCQ, also came up short. On those numbers the trial failed.
Then the biomarker data arrived.
NT-proBNP, the standard marker of cardiac wall stress, fell 49 percent more in the treatment arm than placebo, a difference the company and its principal investigator called the largest seen in any HFpEF trial to date. Right ventricular systolic pressure, measured by echocardiography, dropped 3.5 mmHg against placebo. Both were prespecified exploratory analyses. They were not multiplicity-adjusted, and on their own they do not establish efficacy.
NT-proBNP fell 49% more on TNX-103 than placebo (nominal p<0.0001). In patients who could walk less than 333 meters at baseline, six-minute walk distance improved 26.3 meters over placebo (95% CI 6.0, 46.7).
The functional benefit did show up, but only in the sickest patients. Those who walked less than the trial median of 333 meters at baseline gained 26.3 meters over placebo. A post-hoc quartile analysis made the pattern explicit: the sickest quarter improved 32.4 meters, the next 21.4, while the two fittest quarters actually did worse on the drug than on placebo, by 10.7 and 27.3 meters.
This is a familiar shape in heart-failure research. Enrol a broad, mildly affected population and the treatment effect gets diluted to nothing. Enrich for the people who actually need the therapy and a signal emerges. Tenax's own chief medical officer conceded the entry criteria had pulled in too many patients with less severe disease.
Safety held up, which matters for a repurposed drug
Adverse events were more common on the drug, at 86.7 percent versus 71.9 percent on placebo, and treatment-related events were roughly double. Serious adverse events were balanced, at 10.8 percent versus 10.7 percent. No new safety signals appeared.
For an investor, the risk profile is the part that does not show up in a headline miss. A repurposed molecule with a decades-long IV safety record, an oral formulation that tolerated the trial, and a clear biological signal on wall stress is a different asset than an experimental agent that failed outright.
Turning points
Three things changed because of LEVEL. First, the regulatory conversation: Tenax will request a Type C meeting with the FDA and seek parallel scientific advice from the European Medicines Agency. Second, the trial design: the next study, LEVEL-2, will enrich for the higher-burden population where the effect concentrated. Third, the proof standard: the FDA had already agreed that a single Phase 3 trial with a p-value of 0.01 could support an NDA for this indication.
A subgroup benefit on nominally significant exploratory endpoints is not an approval. The enrichment strategy still has to survive a larger, longer trial against placebo, and the FDA's Type C guidance is not a green light.
The market for PH-HFpEF is real. It is the most prevalent form of pulmonary hypertension globally, and no drug is approved for it. That is the investment thesis underneath the readout: not that levosimendan worked in everyone, but that it worked in the patients who matter, with a safety profile old enough to have a history.
The careful biotech read of this is the same one applied to every promising subgroup story. Clinical history is full of drugs that moved mountains in the sickest patients and collapsed in a confirmatory trial with a cleaner design. LEVEL-2 is the test that decides whether the biomarker signal was biology or noise.
What this means
Tenax now owns a falsifiable hypothesis and the funding path to test it. The company's mission, as its executives put it, is to bring the first approved therapy to PH-HFpEF. The LEVEL result did not deliver that. It delivered the more valuable thing in clinical development: a clear signal of which patients respond, and a regulator that has already told the company what a winning trial would look like.
The readout was a miss. The story is not over. It is, in the way these stories usually turn out, only just being designed properly.