Trial Report Plasma cell targeting with the anti-CD38 antibody daratumumab in ME/CFS -a clinical pilot study, 2025, Fluge et al

[EndME]

The responses in the ritux pilot do not look like the dara ones - someone posted them in one of these threads the other day. They are not anything like as sustained or as large.

That did not seem very convincing. What @Murph posted a while ago and @Evergreen discussed a while ago, does not suggest that substantial differences. The results are somewhat more consistent and slightly better (which might also be a necessary criterium for F&M to pursue things in the first place) but I don't think they are of a different kind per se. There exists large and sustained improvements in RTX samples (for example there are some with sustained improvements after 6 years, even if they are the small minority) so it's a bit a matter of sample sizes, recruitment factors etc.

In phase 2 ritux failed in its primary outcomes but they went ahead because of secondary stuff iirc.

Yes and here we see plenty of talk about NK cells with that being only a post-hoc analysis, which are the type of things that are bound to show up in small trials. Certain issues will always remain.

If F&M had included controls, We would at least have some information on whether sustained 'responses' like in the dara pilot occur in ten pwME who recieve a placebo. And I think that would be valuble whatever it said.

Yes, 10 controls would probably have given a quite bit more information, but of course would have taken longer and required more funding. More useful data is always better, no arguing that. If the argument is for a 20 person controlled trial, does it make more sense to do a 60 person controlled trial straight away and how easy is it to get funding for that if you have no other data?

 

[Utsikt]

As a layperson wrt to trials, but with experience from project management, I thought the idea was to build more and more reliable information as you go, because there is always a tradeoff between the cost of the trial and the reliability of the data.

Most ideas will not pan out, so testing every idea with a full scale trial is simply too reaource intensive. So you do a small and possibly uncontrolled trial to see if there is at least a chance that a larger trial will be positive. In product design this is similar to the concept of MVP - minimum viable product.

Arguing after the fact that we’d have been further along if they did a more comprehensive trial from the beginning is a non-starter, because at the time they made the decision on what kind of trial to do, they didn’t have enough information to justify the use of resources a larger trial would require.

 

[EndME]

As a layperson wrt to trials, but with experience from project management, I thought the idea was to build more and more reliable information as you go, because there is always a tradeoff between the cost of the trial and the reliability of the data.

Most ideas will not pan out, so testing every idea with a full scale trial is simply too reaource intensive. So you do a small and possibly uncontrolled trial to see if there is at least a chance that a larger trial will be positive. In product design this is similar to the concept of MVP - minimum viable product.

Arguing after the fact that we’d have been further along if they did a more comprehensive trial from the beginning is a non-starter, because at the time they made the decision on what kind of trial to do, they didn’t have enough information to justify the use of resources a larger trial would require.

Agree. My inital question was more so about how to run trials and how one can build more and more reliable data if it turns out Dara is negative, which would quite strongly suggest that one currently can't get much reliable information in open label trials of ME/CFS (it would also suggest a much higher rate of natural medium term recovery or something of that sort than is typically suspected under certain recruitment conditions etc). What alterations could be made in the future to hopefully get more reliable data? It would be nice to learn something, just as something was learned after RTX. For example does one place more focus on having clinical cohorts of patients that are regularity checked, focus on improving activity tracking (some devices might be iffy) etc..

 

[V.R.T.]

I maintain that pilot trials should have controls, even if that isn't 'how its done'. I think there would be less failed phase 2s if they did.

 

[Utsikt]

I maintain that pilot trials should have controls, even if that isn't 'how its done'. I think there would be less failed phase 2s if they did.

And a lot less phase 1s because they’d require far more resources to do.

Minimising the amount of failed p2s isn’t inherently positive.

 

[Yann04]

Most ideas will not pan out, so testing every idea with a full scale trial is simply too reaource intensive. So you do a small and possibly uncontrolled trial to see if there is at least a chance that a larger trial will be positive. In product design this is similar to the concept of MVP - minimum viable product.

Agreed. The problem is a very large chunk of people doing pilots aren’t exactly looking to find out “the rigorous truth” but to get more funding / create advertisment-hype. When a certain group has every single one of their pilots a success. The signal provided is essentially zero, unless they have some objective data they publish and people like this forum can compare between trials and flag the one that “stands out”. But in most cases it’s just a couple questionnaires.

Obviously this is not the case for F&M. But it is for much of the rest of the field.

 

[Utsikt]

Agreed. The problem is a very large chunk of people doing pilots aren’t exactly looking to find out “the rigorous truth” but to get more funding / create advertisment-hype. When a certain group has every single one of their pilots a success. The signal provided is essentially zero, unless they have some objective data they publish and people like this forum can compare between trials and flag the one that “stands out”. But in most cases it’s just a couple questionnaires.

Obviously this is not the case for F&M. But it is for much of the rest of the field.

Absolutely.

A good scientist would still start with a p1 in most cases due to resource constraints (meaning no controls in some cases) and to limit the risk of adverse events (meaning as few participants as possible).