Longeveron's Stem Cell Therapy Misses Its Main Goal in a Pediatric Heart Trial — But a Survival Signal Emerges

On September 16, 2026, biotech company Longeveron announced topline results from ELPIS II, a Phase 2b trial testing its allogeneic mesenchymal stem cell candidate laromestrocel in infants born with hypoplastic left heart syndrome (HLHS), a rare and life-threatening congenital heart defect. The trial did not meet its primary efficacy endpoint — an important, honest headline. But buried inside the topline data is a real, if statistically inconclusive, safety and survival signal that is worth understanding on its own terms, without either dismissing it or overselling it.

Background: A Defect With No Good Options

HLHS occurs in roughly 1 in 3,846 babies born in the United States, according to the CDC, and results in a heart too underdeveloped on its left side to pump blood to the body. Without intervention, it is fatal. The standard of care is a grueling sequence of three open-heart surgeries performed between birth and about age three — the Norwood procedure, the Glenn shunt, and finally the Fontan procedure — that reroute blood flow to let the right ventricle do the work of two chambers. Critically, the CDC is explicit that these operations are "not a cure": children require lifelong cardiac monitoring, and many eventually face heart failure or the need for a transplant. That unmet need is why a cell therapy aimed at strengthening heart function during this surgical sequence has drawn serious attention.

The Trial: Cell Therapy Alongside Stage 2 Surgery

ELPIS II enrolled 40 infants with HLHS and administered laromestrocel — an allogeneic (donor-derived), off-the-shelf mesenchymal stem cell product — as an adjunct treatment during the Glenn (Stage 2) surgery, the second of the three staged operations. The idea is that supporting the overworked single ventricle earlier in the surgical sequence, when the heart is under the most strain, could improve its long-term function and reduce the odds of later heart failure or transplant.

The Result: Primary Endpoint Not Met

The trial's primary endpoint measured the change in right ventricular ejection fraction (RVEF) — essentially, how well the single functioning ventricle pumps blood — at 12 months. Here, the results were unambiguously negative: the least-squares mean difference between the treated and control groups was -0.7 percentage points (95% CI: -7.3 to 5.9; P=0.8336), nowhere close to statistical significance. By this primary measure alone, laromestrocel did not demonstrate a benefit.

The Signal Worth Watching: Survival and Safety

Where the data get more interesting is in the trial's exploratory, secondary analyses — figures that should be read as hypothesis-generating rather than proven, but are genuinely worth reporting. In the as-treated analysis, there were zero deaths in the laromestrocel arm over 12 months, compared with one death in the control arm. Longer-term transplant-free survival also favored the treatment group (one event among 17 treated infants versus two events among 21 controls), and major adverse cardiovascular events were roughly 31% lower in the treated group, though this difference did not reach statistical significance given the small sample size. On safety, the news was unambiguously reassuring: the company reported no new safety signals, with treatment-emergent adverse events actually somewhat less frequent in the treated group (94.1%) than in controls (100%), and no adverse events were judged to be related to the therapy itself.

Why the Primary Endpoint Still Matters

It would be a disservice to readers to lead with the survival numbers and bury the miss — the company itself has acknowledged that regulators have indicated RVEF alone would not be sufficient to support approval, meaning this specific readout was always going to need a broader package of evidence. A single missed primary endpoint in a 40-patient trial doesn't rule out a real effect on survival, but it also doesn't establish one; larger, adequately powered trials with survival or hospitalization as a pre-specified primary endpoint would be needed to know whether the encouraging directional signal here reflects a true treatment effect or is simply the play of chance in a small dataset.

What's Next

Longeveron says it plans further discussions with the FDA on potential next steps, drawing on both the HLHS program's Orphan Drug, Fast Track, and Rare Pediatric Disease designations. The company's broader laromestrocel program also holds Regenerative Medicine Advanced Therapy (RMAT) and Fast Track designations for a separate Alzheimer's disease indication, underscoring that the cell product itself is a company-wide platform still under active, multi-indication development. As of this report, these are topline results from a company press release rather than a peer-reviewed publication, and the full dataset has not yet been published or independently reviewed.

Bottom Line

This is genuinely mixed news, reported honestly: ELPIS II missed its primary efficacy endpoint, and laromestrocel has not been shown to improve heart pump function in infants with HLHS. At the same time, a 40-patient safety dataset showing zero treatment-arm deaths, a favorable transplant-free survival trend, and no new safety concerns is a real, if statistically fragile, positive signal in a disease where families currently have no therapeutic options beyond staged surgery. Neither the disappointment nor the promise should be erased from the story — the accurate takeaway is that this specific trial did not succeed on its own terms, while leaving open a survival question that a larger, purpose-built trial would need to answer.