On October 20, Rapport Therapeutics reported that RAP-219 reduced Young Mania Rating Scale (YMRS) scores by 3.1 points more than placebo after three weeks in 253 hospitalized adults with acute mania (p=0.045). Twice as many patients on RAP-219 as on placebo stopped treatment, 24.6 percent against 11.8. Eight days later Rapport priced $432.5 million of stock, pre-funded warrants and convertible notes before fees, about $410.7 million net (Behavioral Wire’s sum of the company’s net figures).
The 3.1 points is below the smallest difference, 3.5 points, in the monotherapy trials behind four approved mania labels, though cross-trial comparisons cannot rank drugs. The money is not earmarked for mania: Rapport says the proceeds and existing cash fund its operations through 2029, across all its programs. And the Phase 3 it plans would not repeat this trial’s regimen, with a higher dose, countries outside the United States and a revised protocol.
The Phase 2 supports continued development. It does not yet establish the effect or tolerability the registrational trials will test.
What Phase 2 showed
The trial (NCT07046494) ran at 20 U.S. sites in adults with bipolar I disorder and a YMRS of at least 25. Patients were randomized 1:1:2 to RAP-219 0.75 mg once daily, started over a two- or four-day titration (the two arms pooled for efficacy), or placebo, for 21 days as inpatients, after washout of medications indicated for bipolar disorder and of all anxiolytics and hypnotics. Baseline YMRS averaged 36.9 on drug and 35.8 on placebo, and the least-squares mean reduction at week 3 was 15.4 points against 12.3. The key secondary endpoint, a clinician severity rating, favored RAP-219 by 0.4 points (p=0.024).
Approved labels give a rough yardstick. In the monotherapy trials behind the current cariprazine, asenapine and carbamazepine extended-release labels, placebo-subtracted YMRS differences at week 3 (day 21 for carbamazepine) ran from 3.5 to 8.0 points; quetiapine’s label reports 4.0 and 7.9 without saying which visit. The smallest, 3.5 points, appears in an asenapine trial and in carbamazepine’s first trial (asenapine’s add-on trial, a different design, reported 2.4), whose p-value and severity-rating difference also sit close to RAP-219’s. That resemblance is numerical, not evidence the drugs are equivalent. Populations, baseline severity, eras and analysis methods differ, and the labels report mean differences, not standardized effects, so the comparison places the result without ranking anything. RAP-219’s patients were sicker at baseline than in any of those trials, and its placebo arm fell 12.3 points, more than in most of them.
Rapport published no confidence interval. Working back from the rounded estimate and the p-value with a normal approximation, the 95 percent interval runs from roughly 0.1 to 6.1 points (Behavioral Wire arithmetic, approximate): compatible with an effect near zero and with one as large as cariprazine’s. The separation grew each week (0.9, 2.2, then 3.1 points) and all 11 YMRS items numerically favored RAP-219 (four reaching nominal significance in an exploratory analysis not adjusted for multiplicity), but those come from the same patients and scale and are not independent confirmation.
Why discontinuation and tolerability matter
Of 253 randomized patients, 31 of 126 on RAP-219 (24.6 percent) and 15 of 127 on placebo (11.8 percent) discontinued treatment. Adverse events explain few of them (four and one). Most of the gap is withdrawal of consent, 21 patients against 6, with stated reasons that include wanting to leave the inpatient unit (five on RAP-219, none on placebo) and worsening symptoms or wanting care elsewhere (three and one).
Rapport’s presentation says missing data for those who stopped were handled under a missing-at-random assumption. That assumption does more work on the active arm, where twice as many patients left, and it holds only if those who withdrew did not respond differently from those who stayed. The topline neither shows it fails nor shows it holds. In plain terms, the 3.1 points partly depends on what would have happened to the patients who left.
Lorazepam rescue was permitted, and giving it required an adverse event to be recorded whether or not the symptom was pre-existing or part of the illness. Rapport has not reported rescue use by arm, which complicates reading the agitation and insomnia events and could bear on the YMRS items for irritability and aggression.
Tolerability is the stronger half of the readout. Treatment-emergent adverse events were similar, 51.6 percent on drug and 52.8 percent on placebo; none was severe or serious; somnolence was 0.8 percent against 3.1, and no akathisia, extrapyramidal symptoms or orthostatic hypotension were reported. Two imbalances stand out: agitation, 16.7 percent against 9.4, mostly in the first week and mostly judged unrelated, and nausea, 5.6 percent against 1.6. The data are also interim, from a September 16 database lock. One RAP-219 patient stopped for suicidal ideation, judged unrelated. The 56-day post-treatment safety data and suicidality results have not been reported.
Perampanel, a non-competitive AMPA antagonist approved for epilepsy, carries a boxed warning for serious psychiatric and behavioral reactions. Rapport describes RAP-219 as a negative allosteric modulator selective for AMPA receptors carrying the protein TARP gamma-8 and argues that selectivity improves tolerability. That is a company hypothesis, untested in a psychiatric population at the 1.25 mg dose planned for Phase 3.
Why Phase 3 is a different test
Before this trial, RAP-219’s patient data were all in epilepsy. Its Phase 2a was open-label and uncontrolled: 30 adults took 0.75 mg for five days and then 1.25 mg through eight weeks, and Rapport reported dizziness (26.7 percent) and falls and somnolence (10 percent each) among common events, with 10 percent stopping for adverse events (company-reported topline figures, with no breakdown by dose). That record informs safety, drug levels and target engagement. It says nothing about efficacy in mania. It means 1.25 mg has human exposure but not controlled psychiatric evidence, and the mania trial’s clean profile was at 0.75 mg.
Rapport’s presentation lists three changes for the mania Phase 3: a higher dose, 1.25 mg, which it says (on a slide labeled modeled) reaches receptor occupancy faster; ex-US geographies with, in its words, strong historical signal in neuropsychiatric trials; and protocol optimization based on the Phase 2. It expects an FDA meeting in the first quarter of 2027 and two Phase 3 starts in the second half. Each change is defensible, and a single 0.75 mg dose gave no dose-response information. Together they make a different test, and a larger Phase 3 effect would carry an attribution problem: higher dose, different sites, revised protocol, no way to apportion it (Behavioral Wire inference). Rapport’s license with Tenacia, which holds rights in China, Hong Kong, Macau and Taiwan, requires Tenacia to run that territory’s part of the planned trials and share their costs.
What the financing changes
The $432.5 million gross (Behavioral Wire’s sum) was three instruments: an underwritten offering of shares and pre-funded warrants at $21.00, a $30.0 million private placement at the same price, and $172.5 million of 2.25 percent convertible notes due 2032. Net proceeds were about $243.9 million from the equity and placement and $166.8 million from the notes, $410.7 million in all, so debt is 41 percent (Behavioral Wire arithmetic). The new stock and warrants equal about 25.8 percent of the 47,955,079 shares outstanding at June 30, and full conversion of the notes would add roughly 13 points (Behavioral Wire arithmetic).
The proceeds are not earmarked for mania. Rapport lists ongoing and planned RAP-219 trials, other research, working capital and general purposes, and has not disclosed the cost of the two planned mania Phase 3 trials.
Rapport held $436.1 million in cash and short-term investments at June 30 and estimates about $395.0 million at September 30 (preliminary, unaudited), or about $805.7 million with the net proceeds (Behavioral Wire arithmetic). Before the raise it said cash funded operations into the second half of 2029. It now says through the end of 2029. Cash roughly doubled while the stated runway moved by at most two quarters, which suggests the raise supports a larger planned spending base than the earlier statement implied, though the earlier statement may simply have been conservative. For scale, spending the $805.7 million evenly through 2029 would average about $62 million a quarter, close to the $60.8 million of operating expenses Rapport reported for the second quarter.
What remains unresolved
Rapport’s chief executive called RAP-219 the potential “first new mechanism of action for bipolar mania in decades.” Selective AMPA modulation would differ from the mechanisms in current approved mania labels, which attribute efficacy, with hedges, to dopamine and serotonin receptor activity or say the mechanism is unknown. But Rapport is not alone. Bristol Myers Squibb’s KarXT, a muscarinic drug, has two placebo-controlled Phase 3 mania trials with registry-estimated primary completion in November 2026, and BMS’s October 29 presentation places the data in 2027; no results have been announced. Biohaven reported that its potassium-channel activator BHV-7000 “did not statistically differentiate” from the comparator in a Phase 2/3 mania trial, a reminder that a new mechanism can fail here. Johnson & Johnson reported a positive Phase 3 for lumateperone, not a new mechanism (4.8 points, company-reported topline). On timing, a muscarinic drug is ahead of RAP-219, whose Phase 3 is expected to start in the second half of 2027.
That is where the financing leaves RAP-219: a proof of concept at the low end of label-reported effects, from a trial in which a quarter of the active arm stopped treatment and one dose was tested; an interim tolerability profile that is Rapport’s main argument for the drug; and a financing that keeps the program funded through 2029 while the registrational mania trials change the regimen.