Words but can't get the pictures/charts etc.
Anavex: An Update To Their Phase 2b/3 Alzheimer's Trial Data
Jul. 30, 2024 7:16 PM ETAnavex Life Sciences Corp. (AVXL) Stock1 Comment
C. C. Abbott profile picture
C. C. Abbott
3.81K Followers
(15min)
Summary
On Sunday, July 28th, Anavex Life Sciences announced that data from their phase 2b/3 Alzheimer's Disease trial was presented at the 2024 Alzheimer's Association International Conference.
New data can be found in the announcement and the corresponding slide presentation.
I take a close look at the data and share my thoughts.
Firing Neurons
koto_feja/E+ via Getty Images
On Sunday, July 28, Anavex Life Sciences (NASDAQ:AVXL) announced that the results from their phase 2b/3 Alzheimer's Disease (AD) trial were presented at the 2024 Alzheimer's Association International Conference (AAIC).
I have previously covered AVXL's p2b/3 AD trial data as the company disclosed them. In this article, I'll provide an update, as there is new data being disclosed.
Before we dive into the data closely, the table below lists the headline and the sub-headlines from the July 28 PR:
Results from Anavex Life Sciences Landmark Phase IIb/III Trial of Blarcamesine Presented at The Alzheimer's Association Conference
Oral, once daily blarcamesine significantly slowed clinical decline for early Alzheimer's disease patients with a good comparative safety profile and no associated neuroimaging adverse events.
Clinical benefit of blarcamesine consistently observed for both 30 mg and 50 mg treatment groups.
Benefits of blarcamesine on both amyloid-beta and brain volume, two underlying pathological hallmarks of Alzheimer’s disease.
EMA submission expected in Q4.
I'll comment on the sub-headlines in the Discussion section below.
Let us now turn to the actual data, as disclosed by the company in their AAIC presentation slides.
Efficacy data: Co-primary, and key secondary endpoint
According to the company, this trial has two co-primary endpoints (ADAS-Cog 13, ADCS-ADL), one key secondary (CDR-SB) and one exploratory (CGI-I) efficacy endpoint.
Previously, the company stated* that the statistical-significance for the two co-primary endpoints were either both p<0.05 or one co-primary endpoint, plus the key secondary endpoint each has a p<0.025.
*On this topic, the company seems to have pivoted to a different view, i.e. "a sole cognitive measure can serve as the primary endpoint for early Alzheimer's trials."
The results of these endpoints are shown in slide 13 to 16 (see below).
Please note that the color, e.g. red, blue etc., boxes in the slides shown below are added by me for emphasis.
Co-primary endpoint data: ADAS-Cog 13 & ADCS-ADL
a slide
Slide 13 (AVXL 2024 AAIC presentation)
a slide
Slide 14 (AVXL 2024 AAIC presentation)
There are a few things to note here:
For ADCS-ADL, both 30 mg (p=0.354) & 50 mg (p=0.527) treatment groups did not show any statistically significant improvement over the placebo at 48 weeks, as shown on slide 14 above.
For ADAS-Cog 13, if p<0.025 is still the criteria, per AVXL's prior statement, then the 30 mg treatment group just missed the statistical-significance (p=0.026).
It's notable that the patient numbers at 48 weeks in these two measurements are not identical, and the % of patients (48 weeks/0 week) is highest in the placebo group (74-77%) and the lowest in 50 mg group (58-59%) (see table below).
Week 0
for both co-primary endpoints
patient number
Week 48
for ADAS-Cog 13
patient number at 48 weeks
(% of week 0 patient number)
Week 48
for ADCS-ADL
patient number at 48 weeks
(% of week 0 patient number)
30 mg 154 108 (70%) 109 (70%)
50 mg 144 83 (58%) 85 (59%)
placebo 164 122 (74%) 126 (77%)
(complied by author from slide 13 & 14)
Key secondary endpoint: CDR-SB
a slide
Slide 15 (AVXL AAIC 2024 presentation)
As mentioned before, if AVXL's prior statement pre statistical-significance of p<0.025 is still valid, then in the CDR-SB, 50 mg treatment group did not meet the statistical-significance (p=0.045).
Regarding the patient numbers, at week 48, for CDR-SB, the placebo has the highest number & percentage (77%) of patients completing, and the 50 mg treatment group the lowest (58%) (see below).
Week 0
patient number
Week 48
for CDR-SB
patient number at 48 weeks
(% of week 0 patient number)
30 mg 154 107 (70%)
50 mg 144 84 (58%)
placebo 164 126 (77%)
(complied by author from slide 13 & 14)
I note that for these measurements, the treatment groups' patient numbers at week 48 are not identical, from one endpoint to the next, e.g. for 30 mg group, n=108 in ADAS-Cog 13, n=109 in ADCS-ADL and n=107 in CDR-SB. The same is true for 50 mg group.
It's not clear, at least to me, whether the discrepancy is due to typos, or some patients' data has been selectively excluded in different endpoint analysis.
Biomarker data
According to the PR:
The [efficacy] findings are supported by biomarkers from the A/T/N spectrum, including plasma Aß42/40-ratio and reduction of brain atrophy.
More details can be found in the presentation, from slide 18-21 (see below).
Please note that the color boxes are added by me for emphasis.
a slide
Slide 18 (AVXL 2024 AAIC presentation)
a slide
Slide 19 (AVXL 2024 AAIC presentation)
a slide
Slide 20 (AVXL 2024 AAIC presentation)
a slide
Slide 21 (AVXL 2024 AAIC presentation)
For the biomarker data, I note that only in the so-called "Brain Atrophy" (slide 18) and "Plasma Amyloid Beta 42/40" (slide 19), there are any statistically-significant [beneficial] differences observed, but not in plasma Nf-L (slide 20) nor plasma p-Tau (slide 21).
I also note that for slide 18, only 55-56% (or 165-166 out of 298) pooled treatment group and 61% (100 out of 164) placebo group data are being reported.
For slide 19, the percentage is even lower [than slide 18], i.e. 40% (119/298) from the pooled treatment group, and 48% (78/164) from the placebo group's data is reported.
As for slide 20 and 21, there are no two patient numbers alike.
In summary, I'm unsure how meaningful or reliable these biomarker data can be interpreted, when the company seems to be selectively including or excluding patients' data, without any explanation.
Safety data
According to the PR:
Oral, once daily blarcamesine...[showed] good comparative safety profile and no associated neuroimaging adverse events
The details of the safety data can be seen in slide 22 (see below).
a slide
Slide 22 (AVXL 2024 AAIC presentation)
(Red boxes added for emphasis)
The table below reproduces the highlighted parts of the results of the table above in slide 22.
Adverse Events Summary 30 mg 50 mg Placebo
Patients, n 167 168 168
TEAE leading to treatment and study discontinuation, n (%)
40
(24.0)
63
(35.7*)
12
(7.1)
Treatment titration AE ≥5%, n (%) 167 168 168
Dizziness
53
(31.7)
67
(39.9)
10
(6.0)
Confusional state
24
(14.4)
24
(14.3)
1
(0.6)
Treatment maintenance AE ≥5%, n (%) 148 153 161
Dizziness
28
(17.7*)
48
(31.4)
10
(6.0)
Confusional state 16 (10.1*) 24 (15.7)
4 (2.4*)
*These data points seem to have typos or are erroneous, e.g. TEAE leading to discontinuation-50 mg, the % data given as "35.7"* should be 37.5=(63/168)x100%; or in Dizziness at 30 mg-maintenance period, the % data should be 18.9% (28/148)x 100% [not 17.7*% as reported].
To me, these double-digit % differences in TEAE [Treatment Emergent Adverse Event], between treatment groups vs. placebo, which are reported in the "TEAE leading to treatment and study discontinuation" as well as in "Dizziness" and "Confusional state" are negative and noteworthy, especially when the latter two persisted in the treatment maintenance period.
In terms of "TEAE leading to treatment and study discontinuation", the company offered further explanation on slide 23 and 24 (see below).
Summary: Safety Population
• TEAEs tend to occur in first 24 weeks and related to titration schedule
• AEs including dizziness:
• Mostly Grade 1 or 2 (mild)
• Transient (approx. 7-11 days)
• Manageable by adjusting titration and dosing time
(Slide 23; bold emphasis added)
• Early discontinuations due to TEAE (BLUE) before Week 24 might be related to up-titration of blarcamesine to the target doses coupled with administration early in the morning
• These events can be addressed by adjusting titration schedule to slower titration and nighttime dosing, as has been positively observed in the blarcamesine compassionate use program
• The low dropouts for non-TEAE reasons, ‘Other’ (yellow) are consistent across blarcamesine and placebo groups, which suggests that there are no dropouts due to lack of efficacy in the blarcamesine group
• There is no evidence that early discontinuations introduced a bias in favor of blarcamesine.
a chart
AVXL 2024 AAIC presentation
(Slide 24; bold added for emphasis)
According to these explanations, the company claims that "TEAEs tend to occur in the first 24 weeks" and "can be addressed by adjusting the titration schedule to slower titration and nighttime dosing".
The company also claims that "The low dropouts for non-TEAE reasons, ‘Other’ (yellow)...suggests that there are no dropouts due to lack of efficacy in the blarcamesine group".
The former is not impossible, although unlikely in my opinion, judging by the fact that the TEAE data were similar in both periods, with even some new items being observed in the maintenance period, e.g. Urinary Tract Infection, Fall, Depression and Disorientation.
As for the latter, in my opinion, the opposite [to the company's suggestion] is more likely to be true, namely the similar [low] dropout rates due to other [than TEAE] reasons that suggest that a similar effect or the lack of it were experienced by the patients in all three groups.
Early AD Endpoints guidance
In the PR, the company claims that:
As specified in the March 2024 FDA Guidance for Early AD, a sole cognitive measure can serve as the primary endpoint for early Alzheimer’s trials...
The prespecified key secondary composite endpoint CDR-SB, also recommended as an alternative primary endpoint for early AD in the new FDA guidance
For any reader who is interested in AVXL's AD program, I highly encourage them to check out this FDA document for themselves.
The reason for this is, that, in my opinion, one would come away with totally different conclusions than what was claimed or implied by the company.
For example, in the March 2024 FDA guidance, one does not find any statement stating "a sole cognitive measure can serve as the primary endpoint for early AD trials", nor does it recommend using "CDR-SB as an alternative primary endpoint for early AD" anywhere.
In fact, this document lists none of the measurement endpoints used in AVXL's AD trial specifically, e.g. ADAS-Cog, ADCS-ADL, or CDR-SB.
The extracts below are some statements which, in my opinion, are relevant to the FDA's guidance on early AD.
Historically, studies to support approval for drugs in the overt dementia stages of AD (Stages 4 through 6) have used an approach, which required the assessment of both cognitive and functional (or global) measures as co-primary endpoints (page 4).
FDA may consider other approaches, including endpoints based on cognitive assessments or surrogate endpoints, which may allow for shorter trial durations as a basis for approval in the earliest stages of AD (i.e., Stages 1, 2, and early 3) (page 5).
Sponsors considering these issues [endpoints for stage 1 AD] should meet with FDA early in development, to discuss the evidence that would be needed to support a marketing application (page 7).
(Bold added for emphasis by author)
If I recall correctly, AVXL has yet to meet with the FDA to discuss their AD program in general or this AD p2b/3 data in particular.
Perhaps this [have yet to meet/discuss with the FDA about the endpoints] is the reason why the company plans to file an application only to EMA and not the FDA in Q4 2024.
Financials
According to AVXL's Q2 2024 report (on page 8), the company has a cash runway of $139M, and a 6-months net loss of $19.2M (or a quarterly net loss of $9.6M), as of March 31, 2024.
Per AVXL, the operation is funded for 4 years (slide 5).
a chart
Page 8 (AVXL Q2 2024 report)
On Monday, July 29, the company filed a prospectus for the sale of their stock to raise an additional $150M.
Discussion
Firstly, according to the company's prior PR concerning the statistical significance of co-primary endpoints, either both p<0.05 or one co-primary endpoint, plus the key secondary endpoint each has a p<0.025. By this standard, the present co-primary endpoints data is mixed and inconsistent or failed.
Namely, in ADCS-ADL, both treatment groups (30 mg & 50 mg) failed to reach a statistical significance (p>0.05); in ADAS-Cog 13, only the 50 mg group is successful (p<0.025); in CDR-SB, only the 30 mg group is successful (p<0.025).
Secondly, biomarker data also seems to be mixed and inconsistent or failed, reporting far fewer [than for the efficacy endpoint] patients' data, and with patient numbers (reported) varying widely in different biomarkers for no apparent justification or reason.
Thirdly, on the safety data, I find the double-digit % differences between treatment groups vs. placebo in TEAE related discontinuation rate; for "Dizziness"; and for "Confusional state" troubling.
I disagree with the company's views that these safety concerns can be addressed by a slower titration schedule, or by evening dosing.
I also disagree with the company's view that low/similar dropout rates in the maintenance period for other [than TEAE] reasons, across all three groups, suggest that they are not due to the drug's lack of efficacy.
In my opinion, the low/similar drop-out rates across all three groups for other [than TEAE] reasons in the maintenance period reasons, do suggest the drug's lack of efficacy.
In other words, patients who were in the treatment groups were not more inclined [than the placebo group] because of the treatment effect, to continue in the trial, even if they could tolerate the treatment.
Conclusion
In my previous article, I stated that I've sold my AVXL stake, exited the AD space altogether as an investor and gave the stock a HOLD rating.
After looking at/analyzing the data for this article, I think I'll continue to stay away from the AD space in general, from AVXL in particular, and will revise my rating to a SELL.
The change of rating is due to the reasons mentioned above, e.g. the efficacy data is mixed/inconsistent, the biomarker data is mixed and problematic due to selected reporting, and safety data are negative or less positive than the company is claiming or suggesting, in my opinion.
I find it doubtful that these mixed, inconsistent or failed data will be strong enough to support a filing in early AD, let alone to secure an approval, wherever AVXL ends up filing their application.
To the best of my knowledge, AVXL has not filed for any other indication (slide 14), e.g. Rett syndrome, Parkinson's Disease anywhere, besides the current guidance of planning to file with the EMA in Q4 2024 for early AD.
The above can account for my bearishness for AVXL's near-term prospect.
As for its long-term prospects as an investment opportunity, I'm bearish still. Why? Readers may wonder, if the company seems determined to continue.
My answer: As an investor, I'm unable to trust the management at their words, e.g. their claims regarding what the FDA guidance says about the required endpoint data in early AD vs. my own understanding of the document.
In my opinion, this [being unable to trust the management at their words] is not just a red-flag, or weakness/setback in the R&D process which is common or unavoidable, but a deal-breaker for anyone whose investment thesis is based on the fundamentals and not on opportunistic trading.
Risks
Risks here includes but are not limited to; irrational market response to the upside for macro or company-specific reasons, if/when the FDA agrees with AVXL pre the filing with the current data, if/when AVXL successfully file with the EMA and if/when such a filing results in an approval, and if/when AVXL reports positive material news or trial results etc.
Thanks for reading. Hope that this article is clear and helpful to your research.
This article was written by
C. C. Abbott
John k9uwa
AI
