Acadia's Narrow Miss: A Phase 3 Bet Wrapped in a Commercial Beat

A $342M quarter and a 2041 patent moat can't fully distract from a RADIANT readout that missed—and a Phase 3 optimization plan management won't fully disclose.

ACAD · Earnings Call · 2026-11-04

The quarter that wasn't supposed to be the story

ACADIA's Q3 had everything a commercial biotech story could want: record DAYBUE sales of $140M, up 39%; NUPLAZID at $202M, up 12% adjusted; total revenue of $342M, up 22%; raised DAYBUE guidance; a patent allowance extending U.S. protection to March 2041; and EU approval with a German launch already underway. CEO Catherine Owen Adams called it “another strong quarter for Acadia, one that demonstrates the strength of our commercial foundation” — Catherine Owen Adams, Chief Executive Officer · 2026-11-04. The stock is not acting like it. ACAD is down 15% over the last 90 days and nearly 30% from its late-August peak, a drawdown that lines up with the RADIANT Phase 2 readout. The tape is telling you the market came for the pipeline and got a miss.

0.0603: the number management wants to reframe

The RADIANT Phase 2 in Alzheimer's disease psychosis missed its primary endpoint—SAPS-HND—with a p-value of 0.0603. That is the fact. The framing is that the study achieved both goals by selecting the 60mg dose and optimizing Phase 3.

While it narrowly missed statistical significance on the primary endpoint, SAPS-HND with a p-value of 0.0603. The 60 mg dose demonstrated a clear and consistent signal of activity.
Sanjeev Pathak, Senior Vice President and Head of Clinical Development · 2026-11-04

The 60mg dose showed efficacy across four psychosis scales, with CGI-ADP nominally significant at p=0.0077; effect sizes were 0.26 on the primary endpoint and 0.37 on CGI-ADP. For context, the company had powered RADIANT for a 0.4 effect size. In August, then-R&D head Liz Thompson said “we are 80% powered for a moderate effect size, 0.4 effect size on our SAPS-H+D.” — Elizabeth Thompson, Clinical Development/Medical Affairs Executive · 2026-08-04 She also acknowledged “a lower level beyond which you start worrying about whether you'd be able to replicate the effect.” — Elizabeth Thompson, Clinical Development/Medical Affairs Executive · 2026-08-04 A 0.26 primary effect size is below the powering assumption. Management's answer is that the 60mg dose, the clean safety profile, and a redesigned Phase 3 can close the gap. Three months earlier, Thompson had been even more direct about what would count as success: “we'll be pleased with an effect size that's in line with what we're powered for, which is a 0.4 or moderate effect size.” — Elizabeth Thompson, Head of Research and Development · 2026-05-06 The Phase 2 came in below that bar. That is why the framing matters.

The redesign is where the argument gets slippery. Acadia will drop the 30mg arm, expand the use of central raters, refine site selection, and make statistical modifications that incorporate baseline disease severity. But Sanjeev Pathak was explicit: “Our goal of optimization is to potentially increase the effect size. And what we are going to do is statistical modifications. I wanted to clarify that we are not changing the inclusion or exclusion criteria.” — Sanjeev Pathak, Senior Vice President and Head of Clinical Development · 2026-11-04 That matters because the prior narrative—and several analyst questions—centered on enriching for sicker patients. If you aren't changing enrollment, you are relying on the analysis plan to rescue the effect size. And Acadia won't show that plan: “we want to keep some elements of our statistical methodology and our analysis confidential to ensure the integrity of the trial.” — Sanjeev Pathak, Senior Vice President and Head of Clinical Development · 2026-11-04 Investors are being asked to underwrite an undisclosed statistical bet, with the FDA meeting not until Q1 2027. That is the real overhang.

The commercial buffer is real—and better than the pipeline story

The good news is that Acadia has something to fall back on. DAYBUE is no longer a single-formulation launch. DAYBUE STIX now accounts for more than 50% of U.S. patients, and the growth is coming from three buckets: switchers, new patients, and restarts of patients who previously discontinued. Tom Garner said “over 80% of those patients are yet to retry DAYBUE STIX.” — Tom Garner, Chief Commercial Officer · 2026-11-04 That is a rare thing in rare disease: a second act for a drug that already had a discontinuation problem. The patent allowance extends U.S. protection for both formulations to March 2041, adding nearly five years of runway and turning DAYBUE from a 2028 story into a 2040s annuity. Europe adds another leg—EC approval, Germany launched in early October, reimbursement dossiers for key markets before year-end—and Japan's Phase 3 is complete with a 2028 launch target and a partner search underway.

NUPLAZID is steadier. The expanded field force has doubled reach to more than 10,000 physicians, and the DTC campaign reached 1.3 million qualified Parkinson's patients, with roughly 1 in 10 visiting a neurologist and more than 1,500 starts. But guidance was narrowed, not raised, to $760-775M, with management pointing to the timing of the field-force impact. That is a small tell: the investment is working, but not fast enough to beat the prior range.

The financials support the commercial thesis. Total revenue reached $308M in Q2 2026, up 16% year over year, and Q3 came in at $342M. Operating income was $46M in Q2 2026, a swing from the losses that defined the prior decade. Cash stands at $976M, which CFO Mark Schneyer called 'a genuine strategic asset.' R&D was $76M in Q3, down from $88M a year ago, while SG&A rose to $159M on the field-force and marketing investments. The company is funding the pipeline without dilution, and price-to-revenue sits at 3.8x, a fraction of its historical peak.

What changed—and what to watch

Three things changed this quarter. First, the patent cliff moved from 2038 to 2041, which is a genuine strategic upgrade for DAYBUE. Second, the commercial mix shifted from centers of excellence to community—nearly 80% of Q3 DAYBUE referrals came from community practices—which validates the field-force expansion but also means the company is now fighting for a more dispersed prescriber base. Third, the pipeline narrative shifted from 'RADIANT will tell us if remlifanserin works' to 'trust our Phase 3 optimization.' That is a less falsifiable story, and the market is right to discount it until the FDA meeting.

The next catalysts are clear: CTAD data later this month, the FDA meeting in Q1 2027, Lewy body dementia Phase 2 results from ILLUMIRA in Q4 2027, and ACP 211 Phase 2 in major depressive disorder in H2 2027. The LBDP program may actually be the more interesting near-term read: pimavanserin's prior LBDP data were striking, and remlifanserin's clean safety profile in RADIANT—no motor adverse effects, no cognitive hit, no QTc signal—matters more in a frailer population. Fast Track designation helps. But the same statistical-optimization playbook will be applied there, and investors should expect the same opacity.

One more thing: gene therapy is gone from the conversation. In prior calls, analysts asked repeatedly whether gene therapy would compress DAYBUE's Rett opportunity. This quarter, with STIX uptake strong and the patent extended to 2041, nobody asked—and management didn't raise it. That is what a de-risked commercial story looks like. The irony is that the commercial story is now the safe part; the pipeline is the volatile part. Acadia spent years trying to become a commercial company. It succeeded. The market is now asking whether the pipeline can justify the next leg.