BioVie (Nasdaq: BIVI) reported positive Phase II data for bezisterim in early Parkinson's disease, with the investigational oral anti-inflammatory demonstrating statistically significant improvements over placebo on a novel composite clinical endpoint and multiple inflammatory biomarkers. The findings are likely to prompt discussion over endpoint selection and subgroup interpretation as the company prepares for Phase III.
The SUNRISE-PD trial (NCT06757010) enrolled 57 treatment-naïve patients with early Parkinson's disease, randomized 1:1 to receive bezisterim 20 mg or placebo twice daily for 12 weeks. The study met its prespecified EPNIC-15 composite endpoint, which integrates motor and non-motor clinical measures, with statistically significant improvement over placebo (p=0.0006). A predefined subgroup of patients with higher baseline platelet counts, representing approximately half the study population, also demonstrated statistically significant improvements across all clinical measures, including MDS-UPDRS Total.
Biomarker findings were broadly consistent with the clinical results. Bezisterim significantly reduced a composite neuroinflammation score (p=0.0018), lowered neurofilament light chain (NfL), and improved multiple biomarkers associated with neurodegeneration and inflammation. Across a proteomic analysis of 380 proteins, 283 shifted in the direction of reduced inflammatory activity. Trends toward improvement were also observed for Alzheimer's disease biomarkers Aβ42 and pTau-217, although these did not reach statistical significance.
Interpretation of the findings requires caution. The trial was exploratory, enrolled 57 participants, and used EPNIC-15 — a composite endpoint developed by the company that combines multiple motor and non-motor measures rather than a conventional Parkinson's efficacy endpoint such as MDS-UPDRS Total or Part III alone. While the biomarker-enriched subgroup was predefined, analyses within that subgroup involve approximately half the study population. In addition, BioVie noted that reported p-values were nominal and were not adjusted for multiplicity, increasing the possibility of false-positive findings across multiple endpoints.