Long Time No Write…New Scoop on $AVXL Blarcamesine or Anavex 2-73 (Revised)

  • Do not Trade on this post as 4 billion years of evolution are against my judgment and God himself might be too.

I read this, and…

I read the latest paper on Anavex 2-73 and precision medicine, the link can be found here, top author Dr.Harald Hampel.

My interest is in the groups showing High Concentration.  Their performance is key to assessing the success of Blarcamesine (Anavex 2-73).  So I looked into Groups 1 & 2. I reached the following conclusions;

  1. Both groups consist of patients with MMSE scores above 20.  This can be controlled by better screening, testing, and future biomarkers.
  2. Group 1 is composed of subjects carrying the APOE 3 alleles.  This fact correlates (causation?) with decreasing odd of the late-onset Alzheimer disease (LOAD), and its severity. Yet unknown is the genotype, are they e3e3 only, or e2e3 or e3e4 possibly?
  3. Group 2 by extension includes Group 1 and Group A, which is made from those patients who are, MMSE>20, SIGMAR1 wild, High Concentration, and don’t carry APOE3 alleles. Are they e2e4 or e4e4, only?
  4. The performance of Group A can be inferred from the performance of Group 1 & 2.

See Illustration.

ANAVEX 2-73 Blarcamesine plot 4 26 2020

Before we discuss anything, on Table 1 in Study on APOE alleles in LOAD population ( link in text below) is given distribution of genotypes in LOAD. If e2 is neutral, e3 “good” and e4 “bad” , then 12% are poor prospects patients, 38% good prospects and 50% in varying degrees between.

What insight did I gain from this plot?

Regarding High Concentration Cohort.

  • Group 1 of 2 patients out of a total of 8 in the High Concentration cohort did not deteriorate. That is n=2 (@38% this should be 3)
  • What is the probability of patients belonging to this group?  Since I am not smarter than my computer; I don’t know much about statistics or probability, I just multiply probabilities of factors going into Group 1.  .80 (SIGMAR1 wild) * .37 (APOE3 alleles carriers among LOAD, see link below) = ~.30 or 30 percent of patients in a large population of LOAD.  See Study on APOE alleles in LOAD population.
  • Group 1 is just 2 patients which are a bit less than 30% of the High Concentration Cohort population, yet in line with the calculations.
  • If Group A is just the same as Group 1 but with at least one copy of APOE4, then the same calculations are, .8*.56=.44 (~50) which is 44% of the LOAD population. Out of 8, this is ~4 patients.
  • We are not accounting for 2 other patients within the Cohort.  Those two are quite different, one is moderate (patient 2008, -4 ADCS-ADL @ 57 weeks) and other extreme (patient 2002, -22 @ 57 weeks). If we assume that patient 2002 is SIGMAR1 mutated and e4e4 then .20*.13=3%, which is rare but not impossible. These patients can be assumed to be SIGMAR1 mutated, Low on MMSE, and any APOE3 genotype carriers. There is negative synergy in the two former and possible negative in the third, but not necessarily since e3 can not be taken from the equation here.
  • From the calculation (very crude) of the distribution of factors affecting Blarcamesine performance, it seems that the Cohort should be a good approximation of performance in Phase 2b/3.

Before I discuss the performance of Group A, I would like to described possible composition of Phase 2b/3 cohorts. If indeed these selection criteria are going to be used as the paper suggests then by selecting the patients to be SIGMAR1 Wild + MMSE > 20, and having the distribution of APOE alleles among them in accordance with the data in the referenced Brazilian study, the overall performance should be the similar to Group 2. The breakdown goes like this, 30% patients “stable”, 60% “delayed” and 10% ‘dropouts”. This is just a working hypothesis based on very crude thinking. The expectations here provided seem not violate, in my very poor understanding, the statistics of LOAD population. Sometimes such rudimentary analysis can be helpful as it sets boundaries for what is possible, but within these boundaries all sorts of phenomena can change the efficacy of the drug. In final reckoning the future will either void my model or validate it.

Regarding the performance of Group A

  • The difference between Group 1 and Group A is the absence of APOE3 alleles.
  • To access the impact of APOE3 and APOE4 on LOAD see the reference Brazilian study.
  • The meaning of Odds Ratio is that if events are statically independent the ratio is 1, above 1 there is indication of events not being independent , and below the opposite effect is found.
  • The ratio for LOAD and e4e4 is ~14, e3e4 ~2.33, and for e3e3 is 0.36. So the genotype has very significant impact on LOAD and its severity, creating great variance in the outcomes among patients in Group A.
  • At the worst, the plot of data for Group A providing information till 148 week of study (~3 years) implies that the top benefit is accrued till year 3, and then there is sharp decline. If one linearly extrapolate the last period slope in about just year 4 the outcome is the same as STANDARD of CARE.
  • Taking another shot at extrapolation, the curve fitting software produces quadratic equation curve making the patients on average reaching score about ADCS-ADL=20 at 5.5 years.
  • If Group A follows the characteristics of STANDARD of CARE curve, there should be transition to accelerated decline sometime after 57 weeks. Is the accumulation of amyloid plaque adding its destructive synergy to the disruption in homeostasis? If so, removal of plaque might be beneficial to these patients, yet till now, the removal by itself has not worked. Is e4 doing its destructive job?
  • The transitions seems to be taking place at 57/70 week mark. The fitted curve slop is y=-1.78 points/13 weeks. Since the point of departure for this regime is about 0 loss, at 5.5 years from treatment start the decline for the group could be -30, leaving the score ADCS-ADL=30 @ 5.5years.

 Remaining questions.

  • How effective is Blarcamesine in patients with APOE4+SIGMAR1 Wild vs. STD of Care?

The partial data to answer this question is now in the genotypes and performance of the remaining 4 patients in High Concentration Cohort. There are six genotypes, two of them, e4e4 and e3e4 having varying degree of detrimental effects on LOAD patients.


Leave at this, I think that I will not live to see the stock reach the promised land of $100, and certainly not the $1000, making us all millionaires. As it lasted it was fun.


Leave a Reply