Methylene Blue Phase 3 failed?

Alzheimer's, cardiovascular, and other chronic diseases; biomarkers, lifestyle, supplements, drugs, and health care.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

Harrison, thank you for replying!

Yes, post hoc certainly has that meaning of putting together a plan after the prespecified one has ended.
Yet, I think alzforum doubled down and called it a prespecified post hoc. That really got me confused.
I am OK with one or the other, though when you put it together like that it seems like an oxymoron.

This is very true about complexities arising later in the article that obscure the results.
This is why I wanted to start on page 1 and work through it page by page in sequence.
I am still not sure what they were doing with the placebos. The article does not appear
to correspond to the dosing arms in the trials on clinicaltrials.gov. It is extremely unclear to me
what they did with the placebo group, the 158 mg and even the 228 mg group.
Another MB trial would be such a blessing.

alzforum is talking of a new AD clinical trials patient registry. Such a registry would be a great help
for a methylene blue redo. With AD trials, there never seems to be a time when they are
not arguing that the placebo arm did not decline enough or the active arm had some issue.
If they could draw patients from well characterized AD patients with highly defined rates of
decline and clear genetics (e.g. APOE 4), there would be much less to argue about.
For example, the error bars for the placebo group might be essentially zero.
There would also be a very good measure of what the initial rate of decline was for those entering into the
treatment arm.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

This is extremely exciting!

taurx is ready to launch another LMTX study that should begin any day.
They are using an 8 mg per day dose!
Consistent with such dosing would be that they view all phase III trials (the Alzheimer's and the FTDbv) as being successful.

It is somewhat disappointing that the trial is only expected to last 6 months. How could a clear treatment effect be established for AD in such a short trial? I am also unsure about how they could have a treated placebo arm? Wonder if they could add in a placebo with methylene blue? 8 mg/day LMTX could have a very powerful treatment effect while discoloring urine. A 10 mg per day dose of MB would likely have only a minimal treatment effect while also discoloring the urine. Perhaps this could be a creative way of solving the blinding problem.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

The Phase 3 trials with LMTX were even more confusing than I first realized.
In the first phase 3 trial (Figures A2 and A3), it was found that there was no significant difference
within the add-on arms and the monotherapy arms. The 4 mg bid monotherapy subgroup performed
essentially at the same level as the 75 and 125 mg bid monotherapy subgroups, as did the corresponding add-on
subgroups. This was known before the second phase 3 datalocked.

Now look at the second Phase 3 trial for LMTX that was recently published in the primary analysis in Table 2a which shows Comparison A as Change from Baseline for the entire 4 mg bid subgroup versus the 100 mg twice daily monotherapy.

This is incredibly confusing!

If they knew that the 79 patients in the 4 mg LMTM mono arm had an equivalent treatment effect as the other LMTM mono arm, why were they included in Comparison A as if they were the non-treatment group 4 mg add-ons? The two subgroups represented in "388" did not have a homogenous treatment response and this was disclosed in the first phase 3 article.

The column should only have had the 4 mg bid subgroup who did have a homogenous response. So this would be the second column of Comparison B ("309") as a substitute for the first column in Comparison A. Now the magnitude of benefit increases in each of the rows of the table and the p-values should be even lower.

Even stranger in Comparison B they actually do untangle the 4 mg bid group into those on add on versus the 4 mg bid who were on monotherapy. Comparison C then looks at the 100 mg mono and 100 mg add-on subgroups.

The most sensible thing that they could have done would have been to put the 79 4 mg mono and 76 100 mg mono patients into one group of 155 patients and the 309 on 4 mg bid add-on and the 297 100 mg bid add-on into another group of 606 patients. This would put patients with homogenous expected responses into their respective subgroups, intensify the differences on the different tests and lower the p-values.

For example, the width of the 95% Confidence Interval for the ADAS-cog for comparison A for the 4 mg bid total group with 388 patients was less than 2 points. Yet, for the 100 mg bid monotherapy subgroup of 76 patients it was more than 4 points. Or with the MMSE: there was nearly a 3 point 95% Confidence Interval for the 100 mg mono in Comparison A. So basically, an error band of 3 points on MMSE and they still nailed a significant p-value.

Sure; we get it. When you have that much money (/alpha) nothing matters anymore. You don't even need to pretend to try.
Why bother with decimal points when the differences are so overwhelmingly large?

Going down the "388" column in Comparison A compared to the "309" Column in Comparison B , it can be seen that
all of these numbers are within the error bounds. Going down the "76" column most of these are within the error bounds of "79" Column, though there appears to be several problems with missing minus signs for the last few rows.

By doing all these unnecessary comparisons they split their alpha and split their sample sizes and this resulted in less insight into the nature of the treatment benefits of LMTX. I am not totally sure that you would be allowed to combine the subgroups in this way even if it were pre-specified only because you thought it would be to your advantage, though if it were allowable then it is not easy to understand why it wasn't done.

They would have saved a fair amount of paper and the article would have been much easier to follow if they simply had one primary comparison: monotherapy versus add-on. The Figures and Tables would then have only needed 1 Comparison column.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

I promised that I would stay away from the MOA, though here goes.

What is a gene on chromosome 19 around 45 Meg?
No, not APOE!

The other one, Mark4 (microtubule affinity regulating kinase 4 isoform)?
This is possibly the gene that confers the benefit of MB.
Wonder if there are SNPs that might enhance/interfere with MB?
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

What is especially exciting is that there are a bunch of plausible ways to amplify the MB effect.

Lower APOE?
https://www.sciencedaily.com/releases/2 ... 131707.htm

Other research speaks of light therapy, taurine, keto diet and others as enhancing MB.
This is going to be great when we have MB/LMTX on the market!

taurx would then have a pile of money and they could start doing a great deal of combination trials
and trials with epsilon 4 carriers etc. . It is so difficult when research funding is a bottomless pit which
typically produces nothing with revenue generating potential and even what does produce revenue
usually has no disease modifying effects. As soon as the MB story firms up, there could be a substantial
number of people who will want to hop on board the bandwagon. AD does cost the global economy
$2 billion per day. At $2 billion per day, there's no time to lose.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

The phase 3 trials with LMTX illustrate for me that clinical trials can often be poorly conceived from a strategic point of view.
It is all the more surprising that this was true with the taurx studies as a major backer of the company is a gaming company.
People who are competitive at gambling, think very very carefully about what bets they make and how they can maximize their advantage.

The phase 3 trials did not substantially show this quality.
The ideal way that the trials could have been structured was for the first phase 3 in AD to have a year or two head start in recruiting over the other phase 3 in AD with the second phase 3 finishing a year or two after the first. In a trial with an adaptive design, information could have flowed from the first to the second trial about what was found about the monotherapy versus add-on. As it was only the Statistical Analysis Plan was changed before datalock and not the patient randomization. While this is a post hoc observation, when spending that much money on a clinical trial it is important to try and think of everything. I have not previously thought that adding such suggestions would be helpful as I was sure that they must give a great deal of thought into what they are doing. Perhaps this should not assumed.

In that light here are some comments/suggestions for the soon to be launched trial with LMTX.
Using a placebo dosing of MB/LMTX that turned the urine blue but had minimal therapeutic effect would be a great step forward
to reduce much of the confusion that has arisen in previous trials in the series. Hopefully, the upcoming trial will have found such a
way out of confounding the treatment effect with the placebo effect. {Taking genetic samples for sequencing (now or later) would also be something to consider.}

I also think enriching the trial for epsilon 4s would be helpful. This is not just self-serving (This is of course the APOE4 Forum). epsilon 4s are more likely to actually have Alzheimer's. epsilon 4 offer the opportunity to have AD clinical trials with more homogeneous pathology and disease trajectories. It is difficult to understand why in the reported phase 3s, they appear to have actually enriched against epsilon 4s.

Taking advantage of existing AD Registries would be wise. If you could draw patients from such a registry you would have measures that showed the decline of patients through time and you could use the overall decline of the patients within the registry as a measure against the treatment effect found in a trial.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

Here's another one!
Good to have them all together for reference.
J11
Contributor
Contributor
Posts: 3351
Joined: Sat May 17, 2014 4:04 pm

Re: Methylene Blue Phase 3 failed?

Post by J11 »

Now that the Lucidity trial has reported I will bump up this earlier thread on methylene blue for reference.
Post Reply