ME and PEM recovery via Cyclophosphamide (personal story)

[siobhanfirestone]

Thanks, @Hutan . You outlined the reasons why I think the data are still relevant really well – you saved me a post!

I think what’s difficult, @siobhanfirestone , is that you’re posting your story – which is fabulous; on a personal level it is so wonderful to hear of someone going from sick to apparently healthy and I and others are delighted for you – on a science forum, so people are responding in sciencey ways. I think all that cautious people are saying on here is that we are not comfortable saying anyone's remission is likely due to cyclo until a randomised, placebo-controlled trial is positive.

I think it’s worth looking at the last few publications from Rekeland et al. – i.e. the Norwegian Fluge and Mella group.

In 2020, Rekeland et al. published the results of the cyclophosphamide study. https://pubmed.ncbi.nlm.nih.gov/32411717/ They concluded:


Another paper by the same group, Rekeland et al. 2022, is also very interesting in this regard. This was not a treatment study. They just followed people with ME/CFS for 6 months using Fitbits and questionnaires. Their data are publicly available. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9484698/

Out of 14 patients who were moderate at baseline (SF36PF 30-59 in weeks 1-4):

• 3 reported SF36PF scores that were ≥20 points higher in the final month (weeks 21-24). The biggest jump was 45 points.
• The step count of 4 patients (i.e. 29%) increased by more than 2000 steps in this time period.
• One patient who started at 3900 steps per day finished at 10219 steps per day.
• Another started at 4453 and finished at 8661, all in 6 months of no intervention.

They conclude:


Then Rekeland et al. 2024 present promising data from a 6 year follow-up, and conclude https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11265720/:


What people are doing on here is just heeding the warnings of these excellent scientists, and learning from long experience of our own remissions and relapses, improvements and deteriorations, treatments tried, and the rituximab rollercoaster.

Edited formatting

Oh i wouldnt expect these doctors to do anything else, but I chose myself what I wanted. Thats not going to be something everybody can and should do, and I have purposfully with-held information so that its impossible to know how I went about this, this is not me saying that this drug is key - its me saying my own case study indicates IgG reduction is what caused remission during chemo.

 

[Jonathan Edwards]

but you cant have a randomsied placebo controlled trial with chemo, at all

Having agonised over trial data with patent lawyers to squeeze the last drop of true wisdom out of them I ended up being able to say in court that what really matter are not in fact placebo controlled trials but dose response studies. If you really want to know if a drug works you accept nothing until you have seen the sigmoid curve (rarely it is another shape).

A trial of cyclo is feasible I think. You give four or five dose levels of cyclo, blind to dose, together with either a synergistic agent like a monoclonal, with no detectable side-effects, or a placebo infusion. You then look for a dose response relation that tracks to a definable biological endpoint, like IgG level maybe, rather than just tracking to cyclo side-effects. In other words you should get a sigmoid curve following cyclo dose that shifts significantly to the left for patients receiving the active combination. I have done trials as complicated as that in terms of analysis and it is doable.

The downside of trials like that is that a lot of subjects get sub therapeutic doses of what may be harmful drugs. You get over that by offering the full dose to all if the data are positive.

 

[Jonathan Edwards]

Is there a reasonable less toxic alternative to cyclo with same moa? Could also be a combination of other drugs that effectively do the same thing?

If the mode of action is lowering IgG levels there are lots of drugs that will do that, mostly when used in combination because none is that good alone. Cyclo has a big disadvantage as an investigative drug because it poisons all cells and so you don't really know why it works if it does. Something like rituximab or daratumumab have very precise actions so tell us more.

 

[Jonathan Edwards]

I think the point about these things being a personal choice is well taken. If individuals choose to undergo experimental treatments that give us important clues that is not to be undervalued.
Where I begin to worry is when that personal informed choice gets blurred. We have seen parents taking children for experimental treatments for ME/CFS and Long Covid. When I started using rituximab perhaps the hardest decision I had to make was to refuse to treat a child with potentially fatal autoimmune liver disease. There was no way the child could understand what the risks were, or indeed that I had very little idea what the risks were. For adults I would express all my doubts and if they still said 'give me the treatment' I was happy.

 

[EndME]

Having agonised over trial data with patent lawyers to squeeze the last drop of true wisdom out of them I ended up being able to say in court that what really matter are not in fact placebo controlled trials but dose response studies. If you really want to know if a drug works you accept nothing until you have seen the sigmoid curve (rarely it is another shape).

A trial of cyclo is feasible I think. You give four or five dose levels of cyclo, blind to dose, together with either a synergistic agent like a monoclonal, with no detectable side-effects, or a placebo infusion. You then look for a dose response relation that tracks to a definable biological endpoint, like IgG level maybe, rather than just tracking to cyclo side-effects. In other words you should get a sigmoid curve following cyclo dose that shifts significantly to the left for patients receiving the active combination. I have done trials as complicated as that in terms of analysis and it is doable.

The downside of trials like that is that a lot of subjects get sub therapeutic doses of what may be harmful drugs. You get over that by offering the full dose to all if the data are positive.

Would the current non-existence of a biological endpoint that could be meaningful in ME/CFS not present a problem?

You might track IgG levels but what if people actually do respond to cyclo but do so in a completely different fashion and at a strange dose due to something weird for example something to do with T-cells that no one is able to anticipate?

 

[siobhanfirestone]

Having agonised over trial data with patent lawyers to squeeze the last drop of true wisdom out of them I ended up being able to say in court that what really matter are not in fact placebo controlled trials but dose response studies. If you really want to know if a drug works you accept nothing until you have seen the sigmoid curve (rarely it is another shape).

A trial of cyclo is feasible I think. You give four or five dose levels of cyclo, blind to dose, together with either a synergistic agent like a monoclonal, with no detectable side-effects, or a placebo infusion. You then look for a dose response relation that tracks to a definable biological endpoint, like IgG level maybe, rather than just tracking to cyclo side-effects. In other words you should get a sigmoid curve following cyclo dose that shifts significantly to the left for patients receiving the active combination. I have done trials as complicated as that in terms of analysis and it is doable.

The downside of trials like that is that a lot of subjects get sub therapeutic doses of what may be harmful drugs. You get over that by offering the full dose to all if the data are positive.

as plasma cells grow back relatively quickly, like within a month or so, would that not put patients who did a smaller bit of chemo to no avail, hypothetically, at risk of having to do it all over again at whatever dose was found to be correct, and therefore increassing thier risk of side effects?

 

[Karen Kirke]

scepticism is fine and expected, but you cant have a randomsied placebo controlled trial with chemo, at all. trust me on this, when you spend days vomiting after. they use the ritux failed data to do the same thing

that 36 questionnarire thing doesnt capture my sickness at all that was 100% PEM and brain fog driven, so in all honetsy im not sure how useful it is, i think they should develop a seperate questionnaire for it tbh

I've edited my post to add the possibility of a trial like the one @Jonathan Edwards described in #83, that looks at different doses of the same drug.

The SF36PF has been very useful in demonstrating just how physically disabling ME/CFS is (e.g. Komaroff 1996), correlates well with steps per day (see Rekeland 2022 and van Campen 2020), and helps us compare the severity of different studies' patients, so I wouldn't chuck it just yet.

 

[Jonathan Edwards]

Would the current non-existence of a biological endpoint that could be meaningful in ME/CFS not present a problem?

No, as long as you define your primary postulated mediator endpoint in advance - say IgG level, you have a hard statical framework. You would need to combine cyclo with something with a precise action such as rituximab. If you picked the wrong mode of action then presumably the combination drug would make no difference. You are then left with a dose response study for cyclo. That might still work because nobody would know the optimal dose or its relation to the degree of side effects. But the advantage of a combination would be that you could look at response relation to a defined variable that cut across even the cyclo side-effect profile - belt and braces if you like.

If the dose response to cyclo is not sigmoid you are a bit stuck but it probably will be, because you are unlikely to have any paradoxical high dose negative effect to give a bell shape.

 

[Jonathan Edwards]

at risk of having to do it all over again at whatever dose was found to be correct, and therefore increassing thier risk of side effects?

Yes, you have to accept that. But no trial is perfect.

 

[siobhanfirestone]

I've edited my post to add the possibility of a trial like the one @Jonathan Edwards described in #83, that looks at different doses of the same drug.

The SF36PF has been very useful in demonstrating just how physically disabling ME/CFS is (e.g. Komaroff 1996), correlates well with steps per day (see Rekeland 2022 and van Campen 2020), and helps us compare the severity of different studies' patients, so I wouldn't chuck it just yet.

Maybe not cut it out but combine with something ? I really feel like they need to sort of track PEM better

 

[siobhanfirestone]

No, as long as you define your primary postulated mediator endpoint in advance - say IgG level, you have a hard statical framework. You would need to combine cyclo with something with a precise action such as rituximab. If you picked the wrong mode of action then presumably the combination drug would make no difference. You are then left with a dose response study for cyclo. That might still work because nobody would know the optimal dose or its relation to the degree of side effects. But the advantage of a combination would be that you could look at response relation to a defined variable that cut across even the cyclo side-effect profile - belt and braces if you like.

If the dose response to cyclo is not sigmoid you are a bit stuck but it probably will be, because you are unlikely to have any paradoxical high dose negative effect to give a bell shape.

Is it possible that natural variations in starting IgG would also have to play a factor in who was given a certain dose - that might be the same for Dara as well?

 

[EndME]

No, as long as you define your primary postulated mediator endpoint in advance - say IgG level, you have a hard statical framework. You would need to combine cyclo with something with a precise action such as rituximab. If you picked the wrong mode of action then presumably the combination drug would make no difference. You are then left with a dose response study for cyclo. That might still work because nobody would know the optimal dose or its relation to the degree of side effects. But the advantage of a combination would be that you could look at response relation to a defined variable that cut across even the cyclo side-effect profile - belt and braces if you like.

If the dose response to cyclo is not sigmoid you are a bit stuck but it probably will be, because you are unlikely to have any paradoxical high dose negative effect to give a bell shape.

Thanks for explaining. I was indeed thinking about something more bell shaped, but I suppose not knowing anything about clinical trials might make such thoughts rather irrelevant.

 

[Jonathan Edwards]

Is it possible that natural variations in starting IgG would also have to play a factor in who was given a certain dose - that might be the same for Dara as well?

I doubt it. Most things in pharmacodynamics are logarithmic - proportional effects. We interpret antibody levels in tenfold steps. Normal IgG variation is probably not very important.

 

[siobhanfirestone]

I doubt it. Most things in pharmacodynamics are logarithmic - proportional effects. We interpret antibody levels in tenfold steps. Normal IgG variation is probably not very important.

Understood - i hope there is this level of analysis in the daratumumab study but I think as a series of case studies it’s still a proof of concept trial