Johnson & Johnson now has two positive placebo-controlled Phase 3 trials of CAPLYTA (lumateperone) in acute bipolar mania. The second, Study 452, released on November 10, showed a smaller effect than the first but pointed the same way on every outcome the company disclosed. J&J says the two trials “will support the planned U.S. regulatory submission.”

CAPLYTA is not approved for mania. It is approved for schizophrenia, for depressive episodes of bipolar I or II disorder (alone or added to lithium or valproate), and, since November 2025, as an add-on to antidepressants in major depressive disorder. A mania label would extend it from the depressive pole of bipolar I disorder to the acute manic episode, which is a separate indication with its own evidence.

Two trials built as duplicates

The registry records for the two trials are near copies. Both enrolled inpatients aged 18 to 75, hospitalized for a current manic episode with or without mixed features and scoring at least 20 on the Young Mania Rating Scale (YMRS). Both randomized patients to 42 mg once daily or placebo for three weeks. Study 451 enrolled 354 people at 42 sites and Study 452 enrolled 351 at 35, most of the sites in the U.S. and the rest in Europe and, for Study 451, India.

At week 3Study 451 (Sept. 21)Study 452 (Nov. 10)
YMRS reduction beyond placebo4.8 points3.7 points
Standardized effect size−0.69−0.43
Response on CAPLYTA45.8%43.9%
Response on placebo20.9%23.7%
Severity rating (CGI-S) vs placebo−0.5−0.41
First significant separationDay 3Day 3

Response means at least a 50% drop in YMRS score. All primary, response and CGI-S results in both trials were statistically significant (p=.0003 or lower).

Study 452 reproduced the first trial’s direction, its day-3 onset, its response pattern and its key secondary result. The point difference is about 23% smaller and the standardized effect about 38% smaller (Behavioral Wire arithmetic). Response on the drug barely moved. The response gap narrowed from 25 to 20 percentage points because placebo response rose 2.8 points and response on the drug fell 1.9.

Why the standardized effect fell more than the point difference cannot be told from what J&J has published. Neither of its two press releases gives baseline scores, standard deviations, the share of patients with mixed features or discontinuation rates by arm; both say only that discontinuation was “low,” and the registry records post no results. Both trials have also been presented as posters at medical meetings. The disclosures reviewed here are the two releases, the registry records and a published summary of the Study 451 poster. Two trials are not one randomized comparison. Patient mix, placebo response and site composition can all differ, and these disclosures do not separate them. The result supports a replicated effect with a smaller observed difference from placebo. It does not show that the drug worked less well the second time.

Against drugs that already have a mania label

The FDA labels for cariprazine (Vraylar), asenapine (Saphris) and extended-release carbamazepine (Equetro) report placebo-adjusted YMRS differences of 3.5 to 8.0 points in the monotherapy trials behind their mania indications, from baselines of roughly 27 to 33. Study 451 falls inside that range and Study 452 near its low end. On the standardized scale, a 2011 Lancet network meta-analysis of 68 trials put the ten drugs that beat placebo at three weeks between −0.20 (valproate, ziprasidone) and −0.56 (haloperidol). Study 452’s −0.43 matches the analysis’s estimate for olanzapine; Study 451’s −0.69 is above its highest.

None of this is head to head. The label trials were run years apart and differ in baseline severity and design, and J&J has not released baselines for its own trials, so severity cannot be matched. A single trial’s estimate also carries wider uncertainty than a pooled one. The comparison supports one narrow point: both results sit in the range of drugs that already hold the indication, which lithium, valproate and many atypical antipsychotics already cover.

That crowded field is the commercial question. J&J’s argument is continuity: its release says the results reinforce CAPLYTA’s “potential to address both depressive and manic episodes of bipolar disorder.” J&J has reported no trial comparing lumateperone with lithium, valproate or another antipsychotic in mania, so nothing yet shows it works better or is better tolerated than those options.

On safety, each release lists the treatment-related events reported by at least 5% of patients on the drug and at least twice as often as on placebo. In Study 452 they were dizziness, dry mouth and sleepiness or sedation, with no percentages given. In Study 451 they were dry mouth (7.9% vs 3.4% on placebo) and nausea (7.9% vs 2.3%). A published summary of the Study 451 poster, presented at Psych Congress in September, put stops for adverse events below 2% on the drug and below 3% on placebo, which does not give discontinuation for all causes. Neither release reports weight, movement-disorder or metabolic results, and three-week trials cannot show long-term effects.

What the second trial changes

Two positive, adequate and well-controlled trials are the traditional way to meet FDA’s substantial-evidence standard for a new indication, and with Study 452 J&J now has two. It has not said when it will file, has not described the application, and has not said whether it has reviewed the package with the FDA. A planned submission is not an accepted one, and acceptance is not approval.

The asset is the one J&J bought. It acquired Intra-Cellular Therapies for $132 per share in cash, a total equity value of about $14.6 billion, and closed the deal on April 2, 2025. When it announced the deal in January 2025, J&J noted that Phase 3 trials in bipolar mania were underway and described CAPLYTA as having “$5 billion+ potential in peak year sales,” a figure it labeled non-risk-adjusted and inclusive of partner sales. It did not say how much of that figure depends on mania. Study 452 is the second of those trials to report. It confirms the first result rather than opening a new indication.

The drug is already selling without mania. On October 13, J&J said CAPLYTA had reached $1 billion in year-to-date sales, crediting uptake as an add-on treatment in major depressive disorder. J&J has not said what a mania label would add to that.

The second trial removed the most obvious risk to a mania filing: that the first result would not repeat. What it leaves open is what the topline omits, including baseline severity, dropout by arm, weight and movement side effects in hospitalized manic patients, and a filing date.