Below tables and figure are from the second phase 3 trial of LMTM with the adjusted statistical plan that read out positive.

I wanted to instill a sense of finality with my original post for this thread: LMTM has already been proven effective in Alzheimer's.

A phase 2 trial and 2 phase 3 trials have provided definitive evidence.

Here's the second phase 3.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5734125/Here's the first phase 3. Figures show the same pattern of homogeneous addons and monotherapy subgroups.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5164296/Yet, I still wanted to crunch some numbers to understand the results better.

What put me off was that Comparison A in Table 2a compared all those on 4 mg bid with those on 100 mg bid monotherapy.

It occurred to me that this was the wrong comparison.

The 4 mg bid group was composed of those on monotherapy and those on add-on therapy.

So, this is a mixture of those who had a response and those who did not have a response.

Here the 4 mg group is supposed to be the "placebo" or non-treatment group.

I am not sure why this comparison was chosen, as the first phase 3 that had already been unlocked

had shown that the composite add on and monotherapy groups had heterogeneous treatment responses.

This can all be seen readily by looking at the figure below.

The 4 mg and 100 mg monotherapy treatment arms had almost the same treatment response.

The 4 mg and 100 mg add-on treatment arms also had almost the same treatment response.

What I found interesting was that I could work back the calculations in Table 2a.

For example, with ADAS-Cog Comparison A, the population standard deviation for the total 4 mg group was = 0.49

The individual SD1(per) = 9.7; likewise for the 100 mg monotherapy arm: SD (pop)= 1.11, SD2(per)= 9.9

When you plug all the values into the equation:

z= D/ [{{(SD1(pop)^^2)/n1} + {(SD2(pop)^^2)/n2}}^^0.5]

You find z= 2.59.

This is the exact result for the p-value presented 0.0047.

After gaining confidence that the expected calculations gave the correct results, I thought it would be interesting to

make what would seem the obvious comparison.

Namely Comparison D: compare the combinations of 4 mg bid and 100 mg bid add on with

4 mg bid and 100 mg bid monotherapy.

When you refer back to the figure below, this combination would seem to be the most rational.

Both pairs in the comparison appear to be very homogeneous.

When I crunched through the numbers using similar formulae as above I found a

z value =

**5.1**.

This corresponds to a p-value = 0.000 000 2

That is, 2 chances in 10 million.

I did the same calculation for the first phase 3 trial (combining the add ons into one group and then the

monotherapy into another group and then comparing them). This time the z- value was ~7.

The p-value was ~0.000 000 000 01.

1 chance in 100 billion.

The grand z value for both phase 3 trials combined was ~8.

The standard normal values aren't reported for 8 SD.

My best guess is that this p-value might be perhaps 1 in 10 trillion.

What is of additional interest is that we now know that the add-on treatment arms were not a true

non-treatment comparison. Methylene blue actually benefited those on add-on treatment, thus the

differences calculated above are actually underestimates.

When you complete your calculations and you realize that the 1 in 10 trillion p-value is quite probably

a highly conservative estimate, it is probably about time that the 44 million people coping with dementing

illness should be be given access to this treatment to help them.

Phase 3 Comparisons 1 Figure.PNG

[attachment=0]Phase 3 Comparisons 2.PNG

Phase 3 Comparisons 1.PNG

You do not have the required permissions to view the files attached to this post.