Biogen's diranersen missed its Phase 2 primary endpoint, yet the tau data was strong enough to greenlight Phase 3. Here's what the CELIA trial actually showed, and why the science and the stock told different stories.

Key Highlights

  • CELIA Phase 2 trial failed its primary dose-response endpoint on CDR-SB at Week 76
  • Diranersen delivered robust, sustained tau reductions across all doses tested
  • Cognitive slowing was strongest at the lowest 60 mg dose, complicating the dose-response thesis
  • Biogen confirms Phase 3 registrational development
  • Analyst opinion divided between scientific optimism and Capital allocation concern

Biogen (Nasdaq:BIIB) announced on Thursday that its Phase 2 CELIA study of diranersen did not meet its primary endpoint. Markets opened with optimism, with shares briefly surging in pre-market on the strength of the biomarker data, before the primary endpoint miss registered fully. As the session progressed, BIIB reversed course sharply, snapping an eight-day winning streak and falling toward its largest single-day decline since January. The catalyst was a dataset that defied easy categorisation: robust reductions in tau pathology across all doses, slowing of cognitive decline particularly at the lowest 60 mg dose, and a safety profile consistent with earlier studies, yet a dose-response relationship on the headline CDR-SB measure that was never statistically established. Biogen confirmed it will nonetheless advance diranersen to registrational Phase 3 development. The science gave management enough confidence to proceed. The market, at least on Thursday, disagreed.

What the Data Actually Says

The absence of a clean dose-response curve is the central analytical problem CELIA leaves unresolved. When a lower dose outperforms higher ones on cognitive outcomes, researchers cannot rule out that the benefit is a statistical artefact rather than a true treatment effect. That uncertainty is legitimate, and the market's reaction reflected it.

What the data does establish is mechanistic. The trial enrolled 416 patients with mild cognitive Impairment or mild Alzheimer's dementia over a 76-week placebo-controlled period, evaluating three intrathecal doses. Diranersen produced measurable, durable reductions in tau pathology across every dose arm studied, as confirmed through both cerebrospinal fluid tau biomarkers and positron emission tomography imaging. For a protein whose accumulation has been linked to neurodegeneration and cognitive decline across decades of Alzheimer's research, that signal carries weight independent of the dose-response question. William Blair noted that meeting the primary endpoint was never viewed as a prerequisite for advancing development, expressing cautious optimism at the overall dataset.

The safety profile added to that case. Adverse event incidence was broadly comparable across dose groups, with serious adverse events elevated only at the highest dose. For a disease space where amyloid-targeting therapies carry documented risks of brain swelling and microbleeds, a tau-directed agent with a cleaner tolerability profile represents a meaningful differentiator if efficacy holds in Phase 3.

A Strategic Bet With Real Costs

Biogen's decision to advance is not without risk, and some analysts have said so directly. The concern raised by Cantor Fitzgerald centres not on the science but on resource allocation. Biogen has spent the past several years rebuilding institutional credibility after the Aduhelm controversy, a drug that received accelerated FDA approval in 2021 over near-unanimous advisory committee objection, was priced at $56,000 annually, lost Medicare coverage for most patients, and was ultimately discontinued in early 2024. Against that backdrop, committing Phase 3 capital to a programme with unresolved mechanistic questions carries reputational as well as Financial Risk.

The counter-argument is structural. Diranersen, co-developed with Ionis Pharmaceuticals, would be the first approved therapy targeting tau if it succeeds. No drug currently on the market addresses that pathological axis. Biogen's other Alzheimer's asset, Leqembi, targets amyloid. A successful diranersen programme would give Biogen two mechanistically distinct positions in the disease, with commercial optionality that justifies a measured Phase 3 commitment.

BMO Capital Markets noted that the lack of dose response creates uncertainty about how consistent the observed benefit is. That is a fair characterisation. It is also, structurally, what Phase 3 is designed to resolve.

Conclusion

CELIA produced a dataset that resists a simple verdict. The primary endpoint was not met, and the dose-response paradox introduces genuine scientific uncertainty that only a well-powered Phase 3 can address. But the tau reduction data is real, the cognitive signal is present, and the safety profile is manageable. Biogen is making a calculated decision, not a desperate one. Whether that calculation proves correct will depend on trial design, regulatory engagement, and a Phase 3 readout that remains several years away. For now, the science provides just enough foundation to justify the next step.