$AVXL Blarcaminase Improves APOE e4 gene Carriers in Alzheimer’s

Do not trade this blog as 4 billion years of evolution is against me and possibly even God laughs at it, He has a great sense of humor!

Do you remember these slides from CTAD 2017?
Slide 23 CTAD 2017

slide 29 CTAD 2017

I decided to plot all patients in High Concentration Cohort from CTAD 2017 for the first 57 weeks.  From slide 29, we have the starting MMSE scores for 6 patients, all those who improved or barely declined.  The remaining patients declined very quickly but their initial scores are not known.

  • I assumed that the mean MMSE score for all patients in the Cohort was 21 MMSE.
  • The mean for the other 3 patients turned out to be 18.
  • We know that 65 years old patients diagnosed on average 3 years after first symptoms should have on average MMSE 0 scores after 8.7 years, after declining from MMSE 20. Avg. Δ=-2.4 MMSE
  • The 90 years old patients decline rapidly from 20 MMSE, and in 3.4 years we assume they have MMSE scores 0. Avg. Δ=-5.8 MMSE
  • 75% of patients in High Concentration Cohort are APOE ε4 (rapidly declining)

So let’s see the graph:

Patients 57 weeks in High Concentration Cohorts 2018


  • Three classes of patients are clearly visible.
  • We can assume that the rapidly declining patients can be carries affected by all the detrimental conditions to improvements while dosing with Blarcaminase. 
  • The 3 patients are just 30% of the Cohort.  From CTAD 2019 presentation we know that 75% ~6 patients are carriers of APOE e4.  In the general population ~50%.
  • These numbers suggest that Blarcamenise helps even patients carrying the APOE e4 gene, the most destructive form of the gene connected with a specific form of amyloid plaque. 
  • The number of 30% of rapid decliners is in line with SIGMAR1 mutation prevalence in the general population. Yet $AVXL did not have SIGMAR1 info on 11 patients, as of 2019 this information is only provided for the 21 patients who are currently enrolled in the Phase2a extension. I bet that as of 2019 the 3 patients are dropouts, so it is highly likely they had the SIGMAR1 mutation. 
  • There is a disparity between the 75% of patients, either 6 or 7 carriers of APOE e4 of 9, and 60% of patients who are helped by Blarcamesine.  If indeed the SIGMAR1 mutation is the harbinger of the worst outcome so that the 3 patients could be carries of all the “bad” genes that makes at least 4 remaining patients be APOE e4 carriers, then Blarcamesine lifts this disadvantage.  
  • The only conclusion that I can, in light of the point above, make is that the plaque theory of Alzheimer’s is not identifying the prime cause, so homeostasis might be a better candidate for it. 

Beer Fund

$1.00 is about one bottle so .....please contribute..


CTAD 2019 and Balrcamesine Again. $AVXL

Do not trade on this post as 4 billion years of evolution is against me and God might not agree with me either. LOL


It Pays off Being Suspicious. Only the Paranoid Survive…

I suspected that the curve for deterioration of the general progress of Alzheimer’s disease might likely contain a region of acceleration. I was for a short while looking for it on the internet but quit it all too soon.  Fortunately, I looked today at iHub and gentleman going under the moniker of amstock82 shared a paper illustrating the general trendline for Alzheimer’s patients.  Here is a link to the paper:

Rate of decline in Alzheimer’s disease measured by a dementia severity rating scale

again the illustration will tell more:

MMSE general Placebo deteriration.png

  • First 2 years: y=-2.15*x
  • After year 2 to year 5 y=-3.88x
  • After year 5 y=-1.6*x

I have included this new info in a graph which old version can be found at this link:


Nothing like Blarcamesine..

So after updating the graph, it looks like this:

Blarcamesine CTAD 2019 12-7-2019

  • Blue Line: Actual data till week 148 (~3 years) Blarcamesine High Concentration Cohort Phase 2a; extrapolated to year 9 with the rate of decline the same as between year 2 and 3; y=-1.3*x  (236%) (previously y=-.55*x).  This is an assumption which is in line with the general trend line for Alzheimer’s found in the above paper as after year 2 the rate increases from y=-2.15  to y-3.8 (173%) The prevalence of APOE e4 gene among High Concentration cohort (75% of carriers) has not been addressed or adjusted.  If we follow the adjustment numbers for the decline rate due to additional 27% of patients with APOE e4, from the previous post (link provided here), then the rate becomes
    y=-(1.3-.3525)*x=-.97*x and y=-.97*x is 172% of y=-.55*x (wow, what a coincident (?))
  • Yellow Line: Quadratic equation fitted to 3-year data for the above cohort. The worst-case possible with ever-increasing acceleration in the decline of patients.
  • Green Line: ADNI Synthetic Placebo Trend Line y=-2.2*+20, actual data till year 2, afterward extrapolated till year 9. This should not be the case as far as we know from the above paper.  48% APOE e4 of carriers.
  • Red Line:  General Trend Line, from the above paper.  Described above.  30% have no APOE e4 and 45% at least one copy of the gene. Comparable to ADNI synthetic Placebo.  The first 2 years for both are identical.  It is expected for the GREEN Line to follow Red Line.


The above-mentioned post was looking at the worst case.  Now let us look at the best case, but still not adjusted for APOE e4 gene prevalence in the High Cohort, at least not yet:

Blarcamesine CTAD 2019 advantage 12 7 2019

  • The Blue Bar:  MMSE points over General Tred Line for Alzheimer’s.  Observe that the difference between Blarcamesine patients and placebo grows up to the sixth year, and even there it holds.
  • The most important Green Bars illustrate the expected delay Blarcamesine gives to an average patient at the current rate of decline.  An average of 65 year old patient expects to live with Alzheimer’s 8.3 years, for 90-year-old this is 3.4 years.  In reality, at about 10 MMSE independent life is impossible but Blacamesine can buy about at least 6 years of life.


  • The CTAD 2019 presentation dwells only on the period of 2 years because data from ADNI had been collected only for that period, and I suspect that the regulators recognize as Synthetic Placebo Arm only the data from ADNI due to its versatility and statistical significance.  ADNI was set up by the industry with this purpose in mind.  Most trials might never extend for more than 2 years so there is valid logic for limitation on the duration of collection.
  • Some people raised questions as to why $AVXL used only 2 years of data, suspecting that Blarcamesine did not hold much efficacy as treatment advanced.  The prevalence of APOE e4 gene in the Cohort might lower the comparative performance of Blarcamesine over ADNI Synthetic Placebo Arm. In the very imperfect study conducted by me, the APOE e4 could have contributed about 60% of the absolute decline rate these two years.  Is this correct; I don’t know.
  • The ADNI Synthetic Placebo Arm for 2 years behaved as the above reference would call for. The graph of general Alzheimer’s progress was done with over 500 subjects. I don’t think there are problems with being not statistically significant.  Yet this data might not be valid to create a synthetic placebo arm since placebo effect exists.  Is this applicable to dementia?  I am not an expert.

Just don’t, It isn’t worth it!

Only $3 Bills Accepted



First Impressions from CTAD 2019 Presentation. $AVXL Blarcamesine

Don’t trade this post
as 4 billion years of evolution and
possibly God is against my judgment



The beginning of the presentation dwells on advancement in clinical studies beyond the slow process of adding sources of information to process which might have not changed for decades towards a revolution.  It seems that now due to the availability of data placebo arm might be eliminated and the evaluation of efficacy and safety might be dynamic (adaptive) that is done in real-time i.e. responding to changes in the data.

I will leave this to more knowledgeable people and let’s graph some adjusted lines on paper for fun.

Before I do this, let me show you what I got in response to my blog (the previous post on Blarcamesine):

Tweet aboy blog AVXL

I am not petulant, and I have to give partial credit to this gentleman.  Since my training is in engineering I can say that I am not a scientist.  I have been told that an engineer needs to get to +/- 20% of the real-world numbers in his model of reality.   If you want to do something in this field of the human endeavor, you model something, or use some heuristics (rules of thumb), and then validate model just test what you build and if everything works out you call it a success if not you go back to your model and start thinking harder.

I know that I am not exactly right but I am not entirely wrong either.  A number of people who were called geniuses were lionized in the public minds. When you read their biographies the success is the result of much efforts and some of these attempts might not be rewarded as the public tends to think.   If somebody tells you that you are smart, run away because humans have their limitations and could be easily duped.  However, limited my thinking is it is really fun to get on a limb and theorize, extrapolate, play with numbers and make yourself look plain stupid.  At least you don’t watch TV.

At now back to the main program:

CTAD 2019 Presentation page 13 APOEe4 gene occurance

  • The carries of APOE e4 deteriorate quicker but they also get Alzheimer’s when they are younger.
  • In a simple search, I did not find any clear number to assess the amount of acceleration in decline vs. noncarriers.
  • The difference in the prevalence of APOE e4 gene in ADNI Placebo vs. High Concentration A2-73 is -27%.
  • The ADNI Placebo Arm fidelity to Phase 2a conditions is limited by the data gathered by the organization. Only 104 weeks of data available and mismatch on APOE e4 carriers prevalence.

Let us go to the next slide:

CTAD 2019 104 week a2-73 data to adjustent for APOE e4

  • There are trend lines for Low & Medium and ADNI Placebo. Since there is 0.47 point difference between these at 104 weeks we can make assumptions about the bias due to the impact of the prevalence of APOE e4.  (30% and 48% respectively)
  • Simple grammar school math gives us (y=+.3525*x+20) bias to be applied to High Concentration trend line (y= -.5500*x+20)

Let’s see the adjusted Blaracamesine trend line.  (this is just a trend line, nothing more):

CTAD 2019 adjusted for APOe e4 gene

More info can be found in the previous post about Blarcamesine and $AVXL regarding the colors of different trend lines and how they were generated.

  • The top continuous black line depicts adjusted for APOE e4 bias the High Concentration trend line.
  • The trend line runs so close to the line for MMSE=20 that the portion of that line was removed.
  • After 6 years the drop would be -1.2 MMSE points.
  • The previous unadjusted line would be -3.2 MMSE points.
  • The unadjusted but extrapolation attempting to account for accelerated deterioration would be at -6.7 MMSE points.
  • APOE e4 is present in 15% of the population and increases the probability to have Alzheimer’s disease 12 times.
  • By attempting to apply the bias to the High Concentration trendline the comparison with ADNI Placebo is more realistic. After 6 years there is a projected 1000% greater deterioration in Placebo Arm than Blarcamesine patients of identical groups of patients.

In the process of making this exercise, no patient was harmed and if I were wrong only I would look stupid.  This is in the realm of possibilities but not a certainty.

One word here, I realize that extrapolations in the physical world are rather dumb.  Most biological systems are comprised of elements and under the influence of so many factors that highly non-linear behavior is possible, and each of us has experienced this with our health 9this wisdom comes with age).  The word non-linear means in vernacular that you wake up one day and you feel sick or health all of sudden.  Yet, $AVXL did extrapolate the ADNI trend line to 148 weeks.  LOL. I bet that they are just mocking me since I have been doing it for a time.

I hope that this gives you insight into the possible factors involved in the evaluation of trial results.  I know that the real curves for the deterioration of Alzheimer’s patients are much more complex as they involve dropout as well as MMSE scores decline.  Now I am waiting for my 15 minutes of fame! LOL. With a beer in my hand.

Beer Fund

$1.00 is about one bottle so .....please contribute..







There is Nothing like Blarcamesine in the Alzheimer’s Drug Search $AVXL

Do not trade on this post since 4 billion years of evolution and God himself might be against my conclusions!


Seekingalpha.com had on 12 04 2019 this press release.

Seeking alpha pree release 12-4-2019

The only problem with this press release is that it is warmed up info already given by $AVXL when it for the first time revealed the 148 weeks data adjusted for Precision Medicine a.k.a. genetic markers.  The following graph can be seen on page 20 of ANAVEX Corporate Presentation from October 2019, the first publication even predates this.

ANAVEX A2-73 Phase 2a 148 weeks MMSE score 20p Corporate Presentation 2019 OCt

Following trend lines are shown:

  • Blarcamesine High Concentration adjusted for genetic markers MMSE scores ~ y=-1.1*(x/52) (for this graph) (x in weeks)
  • Blarcamesine Low and Medium Concentration adjusted for genetic markers MMSE scores ~ y=-4.4*(x/52)
  • Blarcamesine High Concentration decline accelerates with the duration of the dosing, by data points, not the trend. The rate of which between 96 weeks and 148 weeks (52 weeks difference) is y=-1.3*(x/52) (x is weeks).
  • Better fit to Blarcamesine High Concentration decline data points can be given by quadratic equation y=-.21632*x^2+.025*x+20 (x in years).
  • Ultimately, it is entirely possible that ADNI Placebo might have an accelerated decline curve as well since Blarcamesine Low Medium Concentration might be affected by the drug.
  • At this time we have no further information on ADNI Placebo arm than the assumed linear decline of y=-4.4*x+20 (x in years) (2 days before the presentation)

These numbers are about the same as in the Press Release.  Our goal at first is to present the data extrapolated to 6 years. For both ADNI placebo and Blarcamesine.  The case of acceleration in the deterioration rate after 96 weeks can be presented by extrapolating the rate between years 2 and 3 into next year ( year 4).

MMSE SCORE for A2-73 woth Placebo by ADNI CTAD 2019as of 12-06-2019 copy

  • What can be seen here:
  • ADNI Placebo (yellow line) for the first 2 years declines -4.4 points. The extrapolation (black line) into an additional 4 years gives the final score of  6.8. (y=-4.4*x+20)
  • The extrapolation of Blarcamesine y=-1.1*x+20 gives a final score at 16.7 (magenta line)
  • Extrapolating Blarcamesine decline till year 6 and using the rate of decline from the last data points as specified on the graph from the corporate presentation, gives a score of 13.2.

Word of caution; the decline of Alzheimer’s patients over time can be accelerating. 6 years of placebo Alzheimer

(the legend says “probability” but it should say “population fitting given decline”). One can see that with time number of patients is dwindling.  This image gives the idea about the rapid decline in the number of patients whose decline can be keeping the average for the group higher (MMSE score).  At this point, 2 days before the presentation we have no data to present any conclusion on this. The assumption then is that the decline is linear but the decline of Blarcamesine patients follows the worst-case scenario:

Blarcamesine patient declining with accelerating rate and ADNI patient with a linear steady rate.


AVXL CTAD 2019 12-05-2019

  • We assuming that decline starts from MMSE score 20
  • Blarcamesine High Concentration Adjusted data points (4 data points in blue) are well fitted by the yellow curve (the quadratic equation)
  • The ADNI Placebo is showing a steady rate linear decline in scores. In the real world, that might not be the case.

To evaluate the efficacy of Blarcamesine vs. ADNI Placebo we looked at

  • The advantage in retained MMSE points by Blarcamesine patients over ADNI Placebo
  • The delay in years in the score for Blarcamesine patients over ADNI Placebo
  • The relative performance of Blarcamesine in terms of declining ADNI Placebo score.

AVXL CTAD 2019 Efficacy 12-5-2019

  • The Green Bar: the MMSE score difference grows till it reaches almost 6 points at 5 years into treatment
  • The Yellow Bar:  Years need for average Blarcamesine patient to reach deterioration of ADNI Placebo patient; starts at 3.5 years and steadily declines as shown.
  • The Blue Bar: The MMSE score difference between arms as a percentage of the current score of the ADNI Placebo patient; Steadily increasing.

In general, the age of diagnosis determines the average life expectancy after diagnosis; For a 65-year-old patient that can be 8.3 years, for a 90-year-old patient this is 3.4 years.  Assuming that at the score of 10 MMSE points patient is hospitalized; Blarcamesine buys additional 3 years out of an institution for a 65-year-old patient.  That is notwithstanding the benefit of retaining a person’s cognitive abilities.


  •  In the first 2 years, Blarcamesine patients are expected to decline 4 times slower than placebo patients.
  • If the extrapolations have any predictive value; Blarcamesine keeps the patients, in 6 years, from reaching the point of being classified as “Severe” and institutionalized.
  • Presented data on Blarcamesine Advantage is the WORST CASE SCENARIO; ADNI Placebo arm is declining in a linear fashion, that might no be the case in the real world.
  • Performance of Blarcamesine over Standard of Care is about six times greater than that of Donepezil.

Is Global Demand for $TSLA Cars to Go Down, Radically?

You should not trade this blog as 4 billion years of evolution is against my judgment. 

Give a guy a computer and ability to make colorful pictures and graphs and he sure is going to commit a crime against reason. This is a prime example.

I used some sources to make educated guesses to the number of all models of $TSLA car sold in the 3 major markets; the US, EU, and China.

From data on one place in Colorado ( have a history for good correlation to months numbers) I estimated all models in the US and multiply by 2.24 the number of avg 2 months in the quarter in order to account for those fabulous surges in sales at the end of the quarter, especially this one. So that is how I got Oct 19, Nov 19 and Dec 19.

I kept China months numbers steady for Nov 19 (comparing to know Oct 19) and applied scaling for the last month of the quarter (1.61 multiplier).

Europe was a bit more complicated.  I have data for NL, NO, and SP, (Oct, Nov 19 (till 25th)), but I looked at the patterns in other quarters for the ratio between these countries and the missing rest of Europe. so the average of 2 front months was multiplied by 3.44 to get Dec 19. The number from Oct 19 was scaled by 1.55 from know deliveries to NL to give an estimate for Nov 19.

Since I am not that bright I just press a few keys and this is a plot of global demand all models of $TSLA cars by 3 main markets.

Toysla All MOdels 11 25 2019 sales

I am also lazy so I pressed another button and I got this trendline (second-degree polynomial) for all models demand 3 main markets.

Toysla World demand 2019 11 25 polynomial trend line

Toysla Projection 2020 all models all world


I only pressed buttons, blame Gig Oil.   It seems that the trend is right here and $TSLA’s demand is starting in even very continuous fashion declining (as smoothed over quarter’s 3 months).  The information is in the monthly data. I want to find out what the final numbers for the Q4 2019 are to update.

The clues point to the descending trend in global demand.  All markets seem to be poised to roll over.  Crucial to this being true are the numbers for the last month of Q4 2019.  Each market peaked and then started rolling over after going organic. The trendline equation parameters are given their current values by the weight of a sudden drop in the last 3 months.  If my estimates are off than we probably have to stay content with steady demand or even rising, but that would imply going from the bottomless pit to blue heaven.

There is a sudden drop in the world number for Oct 2019. A similar situation was seen in Q1 2019 with demand in the US. The explanation was that tax breaks and the backlog of backorders were exhausted or ended.  $TSLAQ has begun to appreciate the resourcefulness of Elon Musk in “keeping up appearances”, both in accounting and demand. So I heard voices claiming that the Oct, Nov 2019 drop is a “honey trap” for the shorts.  Having lived once in a dictatorship of the proletariat reading between the lines can save you a lot of grief.  There is the expectation of disappointment when Elon conjures up ~70,000 cars sold in Dec 2019.  Mr. Musk sealed all leaks and is baiting potential longs with numbers, both in the accounting of his business and the future orders of the already notorious #Cybertruck.

Model 3 has died (dying a slow death), long live Model Y.  Oh! Wait it is a reincarnation of Model 3. If indeed the numbers for reservation are in tens of thousands then #Cybertruck has been sorely needed.  $TSLA Q3 2019 results and the reservation numbers for Cybertruck seems sounding like a Wall Street’s siren song to Longs and a $TSLA’s swan song to $TSLAQ.  Let’s see how the snowfall in the US affected global sales.

$TSLAQ community had many more clues on production and sales. Now, it seems that the expectations for the quick demise of $TSLA have subsided the information has been a bit sparse.  That is why the clues and the methods rely on correlations from previous quarters.  Are they false?  I don’t know till it will be too late to explain to anybody how Elon Musked my numbers.  In this way, Elon will disgrace another Short.

By my reckoning, only God knows things and we all are kept in the dark.  Since I am humble enough to celebrate my ignorance I think I shall be rewarded.  Please, spare me standing on the street corner and begging.

Sincerely yours, Beer Fund

Beer Fund

$1.00 is about one bottle so .....please contribute..




$AVXL Some other musings on CTAD 2019 upcoming presentation. Blarcamesine

Four billion years of evolution is against my judgment so don’t trade on this.

Synthetic Placebo Arm and Statistically Significant…

Alzheimer’s Disease Neuroimaging Initiative ADNI also was given data on placebo arms by a number of companies in the industry.  This data can be used to create a model of a statistically significant synthetic (virtual) placebo arm. The recruitment of patients can be limited to only those in the dosing arms since the placebo effect working in one trial is identical to workings in another.  Namely, the placebo bias is identical from trial to trial.  Many Alzheimer’s trials involve hundreds if not thousands of subjects so doing away with the onerous and expensive placebo arm might be a great help to many companies.

We might not believe it initially but the placebo effect has been even traced to genetic markers.  I have checked myself for it and I am a highly susceptible person to the placebo effect.  I have to confess that it is true that I always feel better after seeing a physician or getting any medical help, or even expecting it.

If you want to see your genetics material in terms of trait relevant to health, personality, intelligence and physical abilities you can transfer the DNA results from one of those ancestry sites to Genomelink .  These markers have varying degrees of reliability but sure it is fun.

Of course, the final thing here is the selection of those patients whose genetic make up allows them to be responders.  This is certainly in the scope of ADNI activities.

I hoped that Dr. Missling would also use this opportunity to present data on the plaque deposits changes in the brains of the responders.  I don’t think or remember that such data had been collected at the beginning of the phase2a.  It would be very helpful as FDA considers, it might have changed since that amyloid deposits are the only valid biomarker of the disease. Proving that these deposits were reversed in those who responded versus those who did not is from the standpoint of the regulatory body hard evidence of undeniable progress.  I would appreciate it if anybody would remember this hence my memory barely serves me.  Please, leave a response to the post.

Another aspect of this is that statistical significance is related to the background noise.  I am here revoking the image of statistics as looking for a signal in background noise.  It is like making out a word in a very loud night club.  LOL.  Similarly, I can compare the placebo arm in Alzheimer’s to very low-level noise as the outcome for patients leads to quick deterioration in a very predictable way.  Any word spoken over this din can be clearly discerned.  This nature of Alzheimer’s progression moves the bar low on validating trials with even small samples of patients who respond to medication aa statistically significant study.

I have a dog in this hunt, not only financial but my reputation if it ever existed.  I went on the limb many times making wild guesses about the Anavex 2-73 and $AVXL.  Some were grossly wrong, but now I face the ultimate moment as data on Blarcamasine is to be validated (hopefully) and the extent to which the drug is to help the patients can be at least given initial numbers.

So without much ado. Trial population CTAD 2018

That is the composition of the population of High Concentration Cohort when corrected selection for SIGMAR1 and COMT genotype will look like.  You can judge me upon how well this model fits the upcoming data from the LB20 presentation at CTAD 2019.

Beer Fund

$1.00 is about one bottle so .....please contribute..



$TSLA estimate future sales as defined by extrapolated trends in sales. $TSLAQ

Four billion years of evolution is against my judgment so do not trade on it.


Three main markets for $TSLA are the US, Europe, and China. See the graph for the number of cars sold in these markets per month. Data for China since January 2019 only.Cars per month TESLA all Models sep 18 to Oct 19

Model 3 was introduced in 2018 and in Europe and China has not been sold till about March 2019.  Now, let’s see the sum of all models sold in the main markets.

Cars sold per month all main markets TESLA sep 18 to oct 19

The equation of the line fitted to the data is bearly rising.  y=14.90*x + 28350.

If we assume that a quarter consists of three months with the number of cars defined by a point on the line, then we can estimate the future quarter sales based on simple extrapolation into the future.  Since this is an extrapolation of line fitting a set of data points for any 3 months the middle point is the quarter estimate.


The decline of sales in the USA accelerated after April 2019 so the trendline is based on this period.

Sold per Month All Models TESLA Apr 19 to Oct 19


It is obvious that since about March 2019 the sales entered into definitive decline.  y=-796*x+19420.


Cars sold per month TESLA all model Europe

The trendline suggests


rising sales. y=325*x + 7050.



Cars sold per month TESLA China FEB 19 to OCT 19

The trendline suggests declining sales. y=-172*x + 4610.


The sum of all main markets extrapolated from trends documented until October 2019 for the next 14 months.

$TSLA avg sales per month 3 main markets from oct 18 for 14 months extrapolated

This suggests that the sales keeping with trends would be about 70% of current sales barring any sudden event after 14 months that is Q4 2020.

This never happens.

Beer Fund

$1.00 is about one bottle so .....please contribute..