LB-102 Shows Phase 2 Efficacy and Short-Term Safety in Schizophrenia
Key Takeaways
- LB-102 50 mg and 75 mg were superior to placebo on week 4 PANSS total score change, and 100 mg was nominally significant.
- CGI-S and PANSS positive symptom measures improved across doses, several Marder factor scores moved in the same direction, and negative symptom results were more mixed.
- Short-term tolerability included more treatment-emergent adverse events, weight gain and prolactin increases with LB-102, low extrapyramidal symptom burden, and no notable mean QTc signal.
NOVA1 was a US multicenter, double-blind, placebo-controlled, phase 2 randomized clinical trial conducted at 25 inpatient sites from December 2023 to August 2024. Eligible participants were adults aged 18 to 55 years with schizophrenia and an acute exacerbation requiring hospitalization or continued hospitalization. Baseline PANSS totals were about 94 across groups, eligibility required scores of 80 to 120 and CGI-S scores of at least 4, and the mean age was 39.1 years. Researchers randomized 359 participants in a 3:3:3:1 ratio to placebo, LB-102 50 mg, 75 mg, or 100 mg, with group sizes of 108, 107, 108, and 36. Participants underwent a 28-day inpatient treatment period, a 5-day inpatient stabilization period, and follow-up about 2 weeks later, and the primary end point was change from baseline to week 4 in PANSS total score.
At week 4, least-squares mean PANSS total score changes were -9.3 with placebo, -14.3 with 50 mg, -14.0 with 75 mg, and -16.1 with 100 mg. Placebo-adjusted differences were -5.0 for 50 mg, -4.7 for 75 mg, and -6.8 for 100 mg. The 95% confidence intervals were -8.0 to -2.1, -7.7 to -1.7, and -11.1 to -2.6, with P values of less than .001, .002, and nominal .002. Improvement was evident by week 1 and persisted through week 4, with Hedges g values of 0.61, 0.41, and 0.83 across ascending doses. All doses improved CGI-S and PANSS Positive Symptoms scores, several Marder factors moved in the same direction, and negative symptom findings were more mixed. The prespecified 20% PANSS responder analysis favored 75 mg and 100 mg, while the post hoc floor-adjusted analysis favored all LB-102 doses.
Safety results in the JAMA Psychiatry report included TEAE rates of 55.6%, 69.2%, 57.4%, and 75.0% across placebo through 100 mg. Common events included insomnia, headache, anxiety, agitation, and weight increase, which was the only TEAE occurring at least twice placebo frequency. TEAEs led to early withdrawal in 2, 2, 3, and 3 participants, serious TEAEs in 2, 1, 1, and 1, one placebo death, and no QTc increase reached 500 ms. Mean day 28 weight change was 2.0 kg versus 4.3, 3.3, and 2.9, 7% gain occurred in 10.4%, 25.7%, 18.7%, and 8.8%, and blood pressure, pulse, and temperature showed no clinically meaningful changes. Mean prolactin change was 1.3 ng/mL versus 59.9, 51.4, and 51.2; EPS stayed low, SAS, BARS, and AIMS were stable, and QTc changed by 1.2, 4.9, 4.3, and 5.4 ms.
The authors noted that the 4-week inpatient design was not intended to establish long-term durability. They also noted that the inpatient setting may limit generalizability to outpatient populations and that the trial did not include an active comparator. The discussion added that placebo response may have reflected inpatient stabilization or design-related factors, and that negative symptom change in a short acute trial may reflect pseudospecificity. Within those limits, LB-102 showed short-term symptom benefit, with tolerability defined chiefly by weight and prolactin changes rather than EPS or QTc findings.