What Next With Cortexyme’s Atuzaginstat?

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!

Not All Patients Respond To Treatment With Atuzaginstat…

This is the undebatable fact. But what it means that “not to reach the level of statistically significance”? It means that for the dose cohort the probaility of the dosed patients to have results like those in placebo cohorts is lower than 5%. The dispertion about the mean for those 226 patients did not meet this criteria. It sounds worse than what it implies after better look at the data. But first, I have some remorks on the demographics data.

The ADAS-Cog Mean reads impossible number. I think somebody screwed up because ~24 ADAS-Cog is similar to ~22 MMSE. Then again if you have 50% from MMSE between 12 and 18 and you take as average of 15 and if you take average of 19 to 24 MMSE and then get that averaged then you get MMSE 18.25 which is more like ADAS-Cog=40. I used partialy this graphic calculator. If indeed the MMSE average is 18.25 then we deal with very low scores. It would be interesting to see data for every patient just like it was provided by $AVXL AD phase2a and $SAVA AD Open Label.

But then again if you divide 24 ADAS-Cog by 2.33 (70/30) and subtruct it from 30 you get 19.7 MMSE. I don’t think this works in the real world as the graphic calculator by Steve Blasis at al. indicates. This data point should not have been calculated that way but it should come from testing.

Right below the top plots we get the data for the demographics of the cohorts or groups. I would like to point out that only about 25% of all dosed patients benefit from the drug as after 48 weeks, only 64% remains in the cohort or study. This also can be indication that these patients are located more toward lower MMSE scores spectrum or that due to side effect they quit. The other aspect of these plots is that if we take the bars at the end points as Standard Deviations than the Effect Size Cohen’s d can be calculated and it is in the vicinity of ES=3.00 (huge). The low numbers for p-values for 80 mg cohort (p=.02) support this. If the bars are not SD than these number for ES is meangless. We have to have in mind that Donezepil has been approved on ES=.28.

The primary difference is that we have different regulatory environment now than decads ago. The prices of drugs have risen so much that they threaten the solvency of healthcare system in the USA. The regulator will not approve either a costly drug or one not providing full benefits. Yes, I know these rules are only for small biotechs, not the Big Pharma. We are right now witnesses to the fight over diminishing resources in public health and unlimited desire for higher drug prices from Big Pharma, Something has to give. The company was right to expact just 50% reduction in the rate of deterioration but missed the fact that just relatively few would experience the benefits. Depending on how effective against P. gingivalis Atuzaginstat is, and how benefitial to patients (there is a difference here) it is, it still can be a blockbuster. Meantime, Cortexyme will sharpen its skills in desiging better drugs against this bacterium. A lot depends on finding the reason behind the liver enzyme levels. If indeed this is due to immune resonce then much can be expected from the drug. Another aspect of this study is that maybe we need to block both Lysine and Arginine proteases to show better results.

My forecast of the Atuzaginstat results was looking at the best possible case. The assumption being that all the patients will respond to the treatment. Unfortunately, this was not the case. I hope that future therapetic arsenal will include Atuzaginstat as it seems to have a role to play in this game but all hinges on regulators.

Currently, only $AVXL ‘s Blarcamesine and Simufilam by $SAVA can rescue patients. Blarcamesine has improved and retained some patients for 3 years on ADCS-ADL (Daily Living) scale. Retarding the rate of decline over many years can still give very meaningfull results as the average length of dementia is 3 to 4 years and usually is cut short due to death of other reasons. Mostly, these are people who decline ever so slightly due to age and then have sudden onslaught of dementia placing them in care of others. If those who responded to Atuzaginstat are the more advanced in age and severity then Atuzaginstat might much better drug than it seems. I hope these findings will be released at CTAD 2021. The evaluation of AD drug is quite complex as only two drugs till now have the ability to rescue some. I have to mention here Simufilam by $SAVA seems after a year to level off and the next question is whether it will rolll over and reverse of its benefits will occur. This would leave only Blarcamesine as the only drug able to rescue patients. Nevertheless, these rescued patients are limited to some potential 44% of sporadic cases. The number one condition is MMSE score above 20. For that reason it might be possible that Atuzaginstat might find a niche for itself if it can work on patient with more advanced disease where the redardation of the disease progress by 50% might epidemiologically can make sense.

Let us go over the findings from GAIN trial

  1. Atuzaginstat lowers the virulence of P. gingivalis
  2. Roughly 50% of AD patients are subject to the virulence of P. gingivalis
  3. Side effects include liver enzyme increases but are mostly asymptomatic

My interpretation of findings

  1. Atuzaginstat works as intended. It lowers the virulence of P. gingivalis.
  2. Stoping the virulence is not enough to rescue the patients, it just lowers the rate at which the disease progresses.
  3. Other drugs like $AVXL ‘s Blarcamesine should be used in a cocktail to provide therapeutic boost to CNS cells.
  4. P. gingivalis seems to be an oportunistic infection for AD patients at rate of 50% but bearing more weight on the progress of the disease at 37% of patients. Atuzaginstat has found its niche. Biomarkers are ready.
  5. P. gingivalis infection affects also blood vessels, heart, intestinal tract and liver. Till a drug directly killing the bug is found Atuzaginstat is the drug to go to lower the damage done by the infection. The statistical links are already known, for example between heart disease and Pg infection. The disease is chronic and depends on immune system phenotype. The drug can have effect on general health of peopel with the infection but without dementia.
  6. The side effect might be connected with the immune system robustly attacking the infected cells (clearing pool of infected cells) in liver and gut. As in any system of regulation this kind of “failure” should rather be investigated to decisively classify either as sign of success or real failure.
  7. Periodontal disease is the best known chronical condition leading to loss of teeth and general health. The population intended to treat here is very large so blockbuster status has not been lost yet. The concurrent periodontal study should reveal the extend of benefits going to patients on Atuzaginstat with periodontal disease. Follow these developments.
One-Time
Monthly
Yearly

Make a one-time donation

Make a monthly donation

Make a yearly donation

Choose an amount

$5.00
$15.00
$100.00
$5.00
$15.00
$100.00
$5.00
$15.00
$100.00

Or enter a custom amount

$

Your contribution is appreciated.

Your contribution is appreciated.

Your contribution is appreciated.

DonateDonate monthlyDonate yearly

What To Expect from $CRTX and Atuzaginstat at CTAD 2021 $AVXL $SAVA

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!

An Illustration….

X-axis is in weeks of dosing, y-axis is in MMSE points and the number 30 represents the highest score.

HowI Got There….

  • I took the best four patients who never declined from the group of the Superperformers from phase 2a and calculated their baseline average and then their changes in time (weeks) on MMSE scale. (Blarcamesine)($AVXL). From the way this group is defined by biomarkers it should be composed of ~44% of AD population. This should model the best possible outcome for Blarcamesine patient. [1]
  • I did similar thing with Simufilam data ($SAVA) but I only limited it to those 64% patients who never declined and divided the ADAS-Cog11 numbers by crude proportionality ratio of 2.33 between the 70 point ADAS-Cog11 and 30 MMSE scales. I anchored the Simufilam data at 22.5 MMSE which is the same as Blarcamesine average baseline. This should model the best outcome for a Simufilam patient. 32 patients out of 50 at the mark of 9 months.

From the previous trials of Blarcamesine, Simufilam, Donezepil and many others, one can see that the responce to various drugs usually is very rapid and then we either have rollover and reassumption of decline and deterioration in cognition scores or like in the case of Blarcamesine rapid improvement followed by slower rate of the same. Simufilam is the most puzzling as it starts slowly then accelerates to only prepare to rollover into deterioration or if things turn for the better to stagnate. Let’s see an illustration in ADAS-Cog11 sores so negative scores are indicating improvement.

Of course, this is the average for all the 50 patients.

$CRTX shall soon provide us witht the data on its GAIN trial. The composition of the baseline population is much diffrent from those present in Blarcamesine Phase 2b/3 and Simufilam Open Label. It is 50% between 12 and 18 MMSE and 50% between 19 and 24. It is difficult to rescue those lower scoring on MMSE scale. So the success of the trial can be best defined by the extend by which the number of the patients are helped/rescued out the total number of the patients. In case of Blarcamesine the number now has been theoretically assumed to be 44%. For Sumufilam the number was at 6 month 90-something% but now has decreased 64%. At the 12 months either the improvement scores don’t go up any more or some patients start crossing from those help/rescued to those who have actually declined. As $CRTX claimes that 90% AD sufferers are also carriers of P. gingivalis in their brains, and this being behind the onset of the disease then Atuzaginstat should affect almost all these patients, but that should not to be expacted to be to the same degree. The other aspect of the antoicipated rescue of AD in the GAIN trial is that 50% should see at the least improvements similar to Blarcamesine 44%. That is their MMSE scores should go up about 3.5 points from baseline. The other 50% scores should be lower, by let’s say 50%, but the further decline should be about 10% of the overall population.

The above number should give ~2.2 MMSE average score and Effect Size Cohen’s d about 2.0 (huge).

I hope the results will be even better. The most important part is the precentage of the patients who were stopped from decline as this validates the P ginvivalis etiology of AD. Let’s wait for November 11th to find out. Good luck to all.

One-Time
Monthly
Yearly

Make a one-time donation

Make a monthly donation

Make a yearly donation

Choose an amount

$5.00
$15.00
$100.00
$5.00
$15.00
$100.00
$5.00
$15.00
$100.00

Or enter a custom amount

$

Your contribution is appreciated.

Your contribution is appreciated.

Your contribution is appreciated.

DonateDonate monthlyDonate yearly

[1] https://piotrpeterblog.com/2021/09/30/can-blarcamesine-catch-up-to-simufilam-and-surpass-it-avxl-sava-ctrx/

[2] https://piotrpeterblog.com/2021/09/27/why-sava-simufilam-might-not-have-any-advantage-over-avxl-blarcamesine-sava-avxl-crtx-snpx/

Why You Should Count On Cortexyme To Deliver The Alzheimer’s Drug! $CRTX

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!

Jitters of The Biotech Investors and Their Cure

In most PRs companies give some measure of improvement of patients, with the number of subject and the p-values. The certainty of the improvemnts are judged mostely throught the prism of p-values as it is a number with a threshold once attained a semblence of certainty is holly assumed. Alzheimer’s trials are the widowmakers of too many hopes to count. Many a times a promising phase 2a drug turns into disappointment in phase 3 trials. Companies, therefore make their phase 2a trial just as light on subjects count, to cut costs, as proverbial trial baloon. These results are short on certainty and make the investors to look for clues in the bevavior of the management which can be faked or genuine. The hopes connected with $SAVA’s drug Simufilam in the eyes of investors carry higher certainty than those from phase 2a by companies like $AVXL or $CTRX. I have tried to evaluate their chances to deliver the Alzheimer Drug. Maybe, such a drug does not exist but everybody believes that their favorite CNS company has the best horse in the race. A sign of pure lack of certainty, is to hold them all.

Power Of A Trial And One Patient’s Story

The management of $CRTX is fond of saying, or rather all they can say about the trial, that it is powered to detect 50% decrease in the rate of decline among the Alzheimer’s patients treated with Atuzaginistat. When I parse this assertion, it occures to me that this is the least of benefits to patients the management expects to detect with p-values smaller than .05 but this is my interpretation. The hint is “it is better than BIIB drug”. I might be wrong on this one, at least in the second part.

I have already looked into the APOE4 allele and its connection with allegedly increased virulences of P. ginvivalis in brains of some Alzheimer’s patients. $CRTX and the papers on the interacton between the bacterium and the APOE4 protein give the most plausible explanation of the accute character of dementia in individuals carring the APOE4 allele. They are the most quickly declining patients and are the most likely to drop out from trials.

However, you would try to discount the 28 days and n=9 phase 2a and short on duration and subjects number trial, there is very interesting effect of Atuzaginistat on the launguge abilities of the 6 patients who were in the dosed cohort. Let’s see the slide from the $CRTX AAIC 2021 Symposium.

The Winterlight Assesment of Propositions and Conjuctions achieved p-values smaller than .05 just after 15 days and the p-values decreased to less than .001. Would that suffiece to engender certainty in investors? We have here the one from the three (p-values). Missing is the good n number and the measure of efficacy as Winterlight is a black box. I had very few tools to assess the efficacy of Winterlight, because I did not know anything about the inner-workings of the test. Nevertheless my understanding was that there was a consistent separation between the two cohorts, both in magnitude of change and the dispertion around the means for both groups. Ergo, low p-values and high Effect Size, which I calculated to be about 3.8. This is way beyond huge. Fortunately the company published something to giving us insight into the Winterlight Assessment. Let’s see next slide.

The inriched lauguage examplifies the speach pattern of an elderly patients treated with Atuzaginistat over period of 28 days. The patern present in many trials is to exhibited immediate therapetic effect followed by either decline or slower movement toward improved cognitive scores, the latter has happened in the case of Blarcamesine by $AVXL. A similar early effect has taken place with Atuzaginistat where in just 28 days the patient experienced such change in speach pattern that the only conclusion to be drawn here is that we deal with deep recovery of severly deteriorated language and cognition patient, not just 50% decline in the rate of deterioration. The recruitment requirments listed the patients as being between MMSE-2 14 and 25. From the speech pattern one would expect the patient to be closer to the lover limit of recruitment.

As a person myself struggling with disability, especially with language, I recognize here somethig akin to phase change in physics in the language pattern of that patient. For those who have never suffered mental deterioration this might seem trivial or constitute marginal improvement. I consider these two slides anecdotal evidence of very strong therapetic signal. No doubt, that this is the strongest example $CRTX mangement could choose. Nevrtheless, for me it is a very strong argument to expect better results than just 50% decline in the rate of deterioration.

For investors to reach the comfort zone, the troika of n number, p-values and clinicaly meanigful effect has to be present in the data. Otherwise, tea leaves reading of mangement maneuvers, statements, or comings and goings issues. Sometimes, for some, all it takes is to present anecdotal evidence by just few patients to dispell the statistical uncertainty. In case of $CRTX we have it all save the respectfull duration, p-values on known measures and n numbers in the trial. For certainty, this is just incomplete picture. For me the language improvements or rather the phase change in cognition associated with the language of those dosed presents a very persuading argument to make me lean towards expactations of large therapeutic effect. $CRTX second market (periodontal disease) might be so large that the Alzheimer’s is just a fraction of it. Nevertheless, the seond market might not be of the same urgency as Alzheimer’s it can make up for it in numbers, so the bet is kind of asymetric but inherently safer than many others.

One-Time
Monthly
Yearly

Make a one-time donation

Make a monthly donation

Make a yearly donation

Choose an amount

$5.00
$15.00
$100.00
$5.00
$15.00
$100.00
$5.00
$15.00
$100.00

Or enter a custom amount

$

Your contribution is appreciated.

Your contribution is appreciated.

Your contribution is appreciated.

DonateDonate monthlyDonate yearly