Neurocrine Biosciences has achieved its hoped-for profile in a phase 2 schizophrenia trial, delivering its targeted level of efficacy with a lower rate of gastrointestinal adverse events than Bristol Myers Squibb’s KarXT. But the biotech only saw efficacy at the lowest dose—the three higher doses all failed.
San Diego-based Neurocrine told investors ahead of the readout that it was looking to see an eight-point difference between treatment and placebo scores on the Positive and Negative Syndrome Scale (PANSS). The biotech based the target on the spectrum of PANSS scores seen in other studies, such as a phase 3 trial of BMS’ KarXT that reported an 8.4% improvement over placebo.
Neurocrine saw a statistically significant 7.5-point improvement over placebo in patients who received NBI-1117568, a muscarinic M4 selective agonist, for six weeks. KarXT, which BMS snagged in its $14 billion Karuna Therapeutics buyout, hits M1 and M4. Neurocrine sees benefits to selective M4 agonism.
A narrow look at the phase 2 data supports that position. The 20-mg, once-a-day dose delivered efficacy that is in the same ballpark as competitors with a lower rate of adverse events. Five percent of patients on 20 mg of NBI-1117568 reported constipation, compared to 21% of recipients of KarXT in one of Karuna’s phase 3 studies.
The problems, for Neurocrine, start when the results of the other cohorts are factored in. Participants in the other three cohorts received 40 mg or 60 mg of NBI-1117568 once a day or 30 mg of the molecule twice a day. The placebo-adjusted PANSS reductions in those cohorts ranged from 1.9 to 5.0, well below the bar Neurocrine was targeting going into the readout.
Neurocrine’s failure to move the needle at the higher doses spooked investors, who sent shares in the biotech down 16% to $128 in premarket trading. Management is keeping the faith, though. Neurocrine plans to move the once-daily 20 mg dose into a phase 3 study early next year. The biotech also plans to expand into additional indications.
The decision to advance into phase 3 is built on the PANSS data, which includes statistically significant improvements starting after three weeks, and secondary endpoints that tracked wins on other schizophrenia scales.
If the 20 mg data accurately reflect safety and efficacy, NBI-1117568 could be an effective schizophrenia drug that is more tolerable than rivals and has more convenient dosing. The big question is whether the 20 mg data or the three failed arms are more representative of the effects of the molecule.
William Blair analysts conceded the readout was not a “home run,” but said the trial “warrants further development.”
“While we are positive on the effect size demonstrated by NBI-‘568, we anticipate weakness given the placebo-normalized PANSS score improvement of 7.5 is notably below others in the space and below what investors had been anticipating based on these trials (9-plus points) and what management had suggested at prior investor conferences of around 8 points,“ the analysts said in an Aug. 28 note. “We also note the lack of dose response, suggesting no room to improve with differing dose schedules.“