Thank you. I’ve reviewed the short report titled “Blarcamesine Data Bloated with BS (Bad Statistics)”, dated July 13, 2025, authored by JC, who discloses a short position in Anavex Life Sciences (AVXL). Below is a critical evaluation of the report’s conclusions and whether they appear well-founded or biased, especially in light of the peer-reviewed paper by Macfarlane et al. (2025) and Anavex’s own public claims.
?
✅ Summary of Key Conclusions in the Short Report
1. Efficacy Data
• Primary Claim: The pivotal Phase 2b/3 trial (ANAVEX2-73-AD-004) failed to meet one of its two co-primary endpoints (ADCS-ADL), which undermines any regulatory claim of full efficacy.
• Assessment: This is factually correct. The trial met the ADAS-Cog13 cognitive endpoint but failed the functional endpoint (ADCS-ADL, p=0.357). EMA guidelines require success on both cognition and function to establish clinical relevance, making the trial a partial success at best.
2. Statistical Robustness
• Claim: The statistically significant result on ADAS-Cog13 is fragile and vulnerable to small changes in assumptions about missing data.
• Assessment: The analysis appears methodologically sound. The tipping point analysis demonstrates that small changes in assumptions (e.g., dropout bias) can render the positive result non-significant, especially under the EMA’s stricter p<0.025 threshold. That undermines the reliability of the finding.
3. Transparency Concerns
• Claim: There are signs of p-hacking, selective endpoint reporting, and unpublished Statistical Analysis Plan (SAP).
• Assessment: The failure to release a SAP and unexplained improvements in p-values over time (from 0.033 to 0.008 for ADAS-Cog13) are legitimate red flags. The attempt to reframe the CDR-SB (a secondary endpoint) as a substitute for a failed co-primary endpoint is not acceptable under standard regulatory practices.
4. Biomarker Evidence
• Claim: The drug failed to show meaningful changes in tau or neurodegeneration biomarkers (p-Tau, Nf-L), which severely weakens any claim of disease modification.
• Assessment: This conclusion is strongly supported. Lack of effect on tau and Nf-L contrasts with the biomarker signatures of approved agents like lecanemab and donanemab, both of which show strong, consistent effects on A/T/N markers.
5. Safety and Tolerability
• Claim: While the drug avoided ARIA, it caused high rates of neurological side effects, leading to significantly higher dropout rates (32% vs. 7% placebo), which could bias results.
• Assessment: The tolerability issue is clearly documented and aligns with standard concerns about differential dropout and functional unblinding, especially in trials with subjective outcome measures.
6. Regulatory Outlook
• Claim: Approval by EMA is highly unlikely given the data quality, partial efficacy, and weak biomarker evidence.
• Assessment: This conclusion is consistent with EMA’s precedent, particularly the March 2025 negative opinion on donanemab, which had stronger efficacy and biomarker data than Blarcamesine.
?
⚖️ Evaluation of Potential Bias or Incompetence
Bias
• Author Disclosure: The author is short AVXL stock, which introduces clear financial incentive.
• Tone: The language (“BS = Bad Statistics”) and framing are provocative and at times rhetorically charged, suggesting bias.
• Balance: Despite this tone, the core statistical, clinical, and regulatory critiques are well-argued, substantiated, and aligned with current EMA standards.
Incompetence
• No signs of technical incompetence are apparent. The report demonstrates a deep understanding of:
• Clinical trial design
• Regulatory guidance (especially EMA CPMP/EWP/553/95 Rev. 2)
• Statistical methodology (MMRM, tipping point, missing data assumptions)
• Biomarker relevance and validation standards
?
🔍 Cross-Check with Macfarlane et al. (2025) and Anavex Statements
• Macfarlane et al. (2025) reports that the co-primary ADCS-ADL endpoint failed, yet the paper and Anavex public communications have:
• Emphasized the CDR-SB and ADAS-Cog13 findings
• Downplayed the functional endpoint failure
• Promoted SIGMAR1 subgroup results without formal interaction testing or stratified design — clearly exploratory by EMA standards
The short report is more aligned with regulatory norms than the company’s communications and is stricter in its evidentiary interpretation, but not unfairly so.
?
📌 Final Summary
Are the Short Report’s Conclusions Likely Correct and Well-Informed?
✅ Yes, overall:
• The report is technically sound, uses appropriate regulatory standards, and exposes genuine weaknesses in the clinical trial data and company disclosures.
• While bias is present, the core arguments are strong, evidence-based, and reinforced by publicly available trial outcomes and EMA guidance.
• The recommendation for rejection by EMA is reasonable, not sensationalist, given the data shortcomings, poor tolerability, and lack of biomarker support.
?
Would you like a side-by-side matrix comparing the claims in the short report vs. the peer-reviewed article or company statements?
Market Data 
Markets 
AI
