- Since there is no individual data on each patient performance and their membership in the efficacy biomarker group as per page 12 & 13, we extrapolate individual performance in linear fashion ( doubtful yet the only tool available ) to 148 weeks. For some data is given up to 109 weeks: patients 1014, 2006, 2010 & 1013. Other we simple assume that they continue the pattern of progress of disease up to 57 weeks through to 148 week: 1007, 2008. 2001, 2002, 1002 & 2001. Patients 2001, 1002 & 2002 decline so rapidly that by 148 week they could be among the dropouts. By 148 weeks patient 2001 is considered “ALMOST DEAD” (MMSE 5 (?)).
- The decline of -2.0 MMSE points over 148 weeks for the cohort from CTAD 2018 presentation is a mean of three groups of patients. Those who continuously improve: 1014, 2006 & 2010 ( “CURED” ). Those who “SLOWLY DECLINE”: 1013, 1007 & 2008. Patients 1007 and 2008 are hypothetical since their condition is not known after 57 weeks. The same applies to patient 2001, the only one in “ALMOST DEAD” group.
- The shifting membership in the group “SLOWLY DECLINING” makes the total of patients from 6 to 7. For 6 patients current in the biomarker selection is 66% of the High Concentration Cohort population, for 7 patients it is 77% of population.
Following is the illustration of that those assumptions:
To verify this complete this (27*3 + 16.6*3 + 5)/7 = 19.4 very close to ~19 MMSE (if drop is about 2.0 points per 148 weeks).
- It seems that by the power of addition, multiplication and division out of 9 patients 7 might be in the trial by 148 weeks. Are they all fulfilling the requirements of those four criteria on page 12 & 13 CTAD Presentation? Are there any more dropouts? 7 out of 9 is 77% of the total population that would be miraculously in line with the 80% of general population carrying the genes variants most responsive to ANAVEX 2-73.
- With time the composition of trial population changes from one description to another. This plot illustrates this:
- The same for general population of Alzheimer sufferers:
- Benefits for the group No Decline/”CURED” is self-evident. Group Decline -2.0/Year/”SLOWLY DECLINING” declines slower than octogenarians on placebo (-1.6/Year vs. -2.6/Year placebo) (see my earlier posts). This group accrues benefits as well. The only variable which can be controlled in clinical practice is the early detection of onset of the disease, and here ANAVEX 2-73 seems to offer greater benefits given at higher MMSE scores. This points to 66% of Alzheimer patients benefiting in meaningful way from ANAVEX 2-73.
- Caveat here is that I am not familiar with sophisticated statistical tools. As data becomes available I try with simple assumptions and tools illustrate the data’s implications in more accessible way. No doubt that I oversimplify things but I hope at least my not too far from what the data says at the moment.
All patients number are from CTAD 2017 November presentation.
I dismissed any talk about possibility of success of monoclonal antibodies in targeting Amyloid Plaque in Alzheimer. Mea Culpa Maxima! Sometimes one has to make about-face, and acknowledge that BIIB037 does prove something. In Alzheimer model, from this reference Ref 1 3D model of Alzheimer in vitro , the Amyloid Plaque appears after incubation and causes the inflammation (due to action of innate immune system) to proceed with destruction of the neurons and atrocities which nurture them. From the results of trial one might think that Amyloid Plaque once cleared can make the disease either stop or slow down. At least that is what number of companies claim and try to prove with their work. If indeed BIIB037 can stop decline then it seems that the homeostasis disturbance does not by itself calls for inflammation, and this stage the culprit here might be Amyloid Plaque which alone draws immune system in.
- The data on MMSE score for patients selected with the four criteria in CTAD 2018 presentation (page 12 & 13) for ANAVEX 2-73 Δ=-1.8 over 2 years 10 months (148 weeks) (no info on dropouts ) (Blue)
- The same for BIIB037 over 3 years and 4 years. (no info on dropouts ) (Black)
- The MMSE scores for remaining in the study patients over 4 years (average) (Red). Ref 2: Alzheimer’s Natural Course
This is for the case of relative start from the same level about 21 MMSE (Mild to Moderate)
This is relative performance but it can not be compared as each of those caries different profile of so-called dropouts. These are patients who decline so rapidly that they stop reporting to investigators or are placed in institutions. The dropout dynamic can be seen for line #3 in this post Illustration of dropout rate in Alzheimer studies.
There is one thing that is salient in article about BIIB037 and that is the “Adjusted Mean”. Well adjusted mean might correct small distortion due to uncontrollable variable and might not alter the meaning of the data or can “overwhelm” the initial data with its correction. There is not further explanation to that so I still might be right not to show some contrition on my parts.
The most nagging question for my fellow $AVXL investors is whether ANAVEX 2-73 will get approved and is there enough efficacy to get to that point?
Today (10-26-2018) at CTAD Barcelona 2018 $AVXL released information on the results of dosing patients for 148 weeks ( 2 years 10 months ). The data provided can be found on pages 17 (MMSE scores) and 16 (ADCS-ADL scores). Here, again company changed the population of the disclosure and made it to be determined by four factors given on page 12 and 13 (we call it from now on as New Cohort). Relevent is also inclusion of APOE ε4 allele incidences within the High Concentration cohort as these patient account for 40-60% AD patients and also this mutation is present in cases of extremely aggressive Alzheimer. 75% of patients present with this mutation there strengthens the case for ANAVEX 2-73as it is more representative of Alzheimer population. The criteria covered on page 12 and 13 do not give us the exact number of patient present in thus created New Cohort but if we assume that it is representative of the population to be treated then they are to be expected in Phase 2b/3 as the same criteria are applied for inclusion in the study. Also another unknown is the number of dropouts. Let’s leave these speculations alone and concentrate on the data from the presentation.
Pages 12 & 13 tell you that these are the expected results from Phase2b/3.
Let’s look at ADCS-ADL scores. The plot indicates that there was a decline of -2.0 points (average) on 80 point scale, over 3 year period. When we compare this with a plot of ADCS-ADL scores from previous post of mine, it seems that the New Cohort haven’t deteriorated from 109 weeks but again the question is which patients were included in 148 week data. The data from 109 weeks implied -2.2 points decline (average) over 109 weeks and included six patients all improving MMSE scores or slowly declining. The New Cohort stands at -2.0/2.83 years (-.70/year). Nevertheless this is what one can expect from culling the patient population by the way of genetic markers for efficacy. The Low/Medium population exhibited declines on par with placebo for ADCS-ADL scores (stand about -8.0/year ADCS-ADL).
Now time for MMSE scores, again the population has been altered to be defined with criteria for Phase 2b/3 and registered MMSE decline of -2.0 points over 148 weeks, this roughly translates into -0.7 MMSE points per year vs. natural course of -3.8/year. The Low/Medium population retained some of their scores better than average natural course patients (-2.0/y vs. -3.8/y)
If the New Cohort is representative of general Alzheimer population then (80%x80%=)64% of that population can expect that at the least in 3 years of extended meaningful lives without acute declines in cognitive and daily living abilities and lead lives almost unaffected. Also that some 40% of that population (80%x50%=) can even be improved. It is worth to note that on some level people in advanced age have natural level of declining MMSE scores. The ages of those in the trial were 55-85 so no meaningful inference can be made. As my previous posts suggested by the time of 3 years from diagnosis there remains just about ~38% people in the study (the rest deteriorated significantly) and the number of those who did not decline is about 2% of the starting population. The ADCS-ADL and MMSE data for the New Cohort suggests that more or less there are “equal weights” above and below the average line (since due to ceiling effect on improved scores number of improving scores must be greater than those declining which can go deeper below average), it might mean that my calculations on possible outcome of Phase 2b/3 were in essence correct (however crude they are) but the final check will be Phase2b/3. See Previous post predicting Phase 2b/3 results.