Treating patients suffering from myalgic encephalopathy/chronic fatigue syndrome (ME/CFS) with sodium dichloroacetate, Comhaire 2018

[Jaybee00]

@ME/CFS

If you are continuing to track the response in your patients, are the responding patients continuing to respond or are you seeing signs of tachyphylaxis?

Thank you for participating in this forum.

 

[Trish]

Negating OBSERVATIONS is funny indeed. But it is OK as long as you are amused.

I have no idea what this means.

I made a genuine attempt to understand what might be going on in your mysterious formulae. I found something that seemed interesting to me. I shared it here. If you would like to repeat the calculations I made and tell me where I went wrong, please do. I'd be happy to share the particular sets of figures I used to do the calculations.

Once I found something that looked interesting to me, I speculated on what might be going on. I was also aware that in your initially published study of 10 patients, only 5 responded, and the other 5 turned out to have been misdiagnosed. I wondered whether that could be the case in your second study too. If that is not the case, that's fine - it was simply a hypothesis based on available evidence, namely the formula and the FSS questionnaire.

When I said it was fun, I meant in the sense that I find doing science and maths fun, I didn't mean to imply that I find your patients funny. I am a severely affected patient and have had ME for 28 years. I would never make fun of other patients, whether they have ME or depression or anything else.

 

[Hutan]

I wrote this last night and then fell asleep before posting. So, it's slightly out of context.

But when generating a formula like this based on the data
you need separate training and test sets

Yes. @ME/CFS, you also need to think a bit logically about what might make sense in terms of the parameters you use. It is highly unlikely that answers to 4 vague and highly correlated questions about fatigue are going to identify responders in any new group of patients. What you have there is just a statistical artefact. 'Area under the curve figures' don't make it more true.

You might as well have asked the participants about the number of consonants in their first name, their surname, the name of their favourite pet, the name of the street that they live on, the name of their favourite aunt and so on. With seven parameters that can be negatively or positively correlated with response together with a few other parameters such as age and length of illness, it's extremely likely that you can make an equation that appears to be great at identifying the responders in a small group. But when you apply that equation to a new group of patients, it doesn't work.

That is indeed what you found when you went from a sample size of 22 to 33 - the equation had to be significantly changed. And 33 is still a small group for making a predictive formula.

But really, at the stage of investigation that you are at, identifying characteristics of the responders is a diversion from the main question.

The first job is to show that there is a real response in at least some people. That is, a response over and above a placebo response and that that response is sustained over a much longer period than 30 days.

Edit - For that you need better, and preferably objective, measures of success and a blinded control treatment.

 

[Hutan]

Somatic reaction are controlled by the autonomous system, with its orthosympatic and the parasympathic components. The activity of these are regulated by the thalamus, which is part of the diencephalon. There is convincing evidence that the function of the diencephalon is deregulated, and the blood supply to this region of the brain commonly is abnormal. Blood supply and uptake of the Tecnecium isotope can be studied (NeuroSpect scan) and are related to cell function. So, abnormal cell metabolism will equally affect the autonomous system and deregulate the normal equilibrium between the 2 components, with so-called psychosomatic problems (tachycardia, orthostatic hypotension, irritable bowel syndrome) .

Even assuming there is a low blood supply to the diencephalon of people with ME/CFS, which is far from adequately proven, this doesn't actually explain a psychosomatic mechanism. There is no credible evidence for an unusual prevalence of pre-illness inadequate stress management in people with ME/CFS. The Dubbo study which followed a large number of people after several infective illnesses did not find any distinguishing psychological characteristics in those who went on to develop post-viral fatigue/ ME/CFS.

@ME/CFS, you have an unexplained leap from the ill-founded conjecture of 'inadequate stress management' to an (assumed on as-yet inadequate evidence) 'abnormal blood supply to a part of the brain'.

In the case of, for example, abrupt onset ME/CFS, how does 'overly authoritarian education' actually cause people to be going about their adult lives quite happily and productively for years with no unusually inadequate stress management and then suddenly develop and sustain low blood supply to the diencephalon?

 

[benji]

I've had a message from someone trying to work out this formula in their own case, and was reminded that there is seemingly an error in the paper. So to save anyone else puzzlement, the number you calculate is not p, but logit (p) where,

logit (p) = ln {p/(1-p)}

So, e^logit (p) = p/(1 - p), and solve for p.

In Comhaire's Fig.2, logit (p) runs from 0 to 1. I asked him about this, but he didn't clarify. Presumably a typo; presumably it should be p, not logit (p) on the axis.

For anyone with brain fog or math-aversion, m.wolframalpha.com will do the calculations for you. There are apps for mobile devices.

Thank you! now I finally get it right. Couldn’t understand that I got a number outside the interval from 0 to 1. This helps.

THis comment is perfectly correct. The vertical axis shows (p) not logit (p). Thank you!

And getting it from author, is even better.

Quite complicated math/statistics, or perhaps a sign of that I have been away from math and physics for a long time. Maybe the latter. But I got it, and I got a good p value.

I recently heard from a patient of Kenny DeMeirleir that he had prescribed a dose of 2g DCA in the morning, for one very sick patient. That surprised me, after reading this.

 

[benji]

What I also wonder, are there any adverse effects? I can’t remember any side effects reported in the articles.

 

[ME/CFS]

Even assuming there is a low blood supply to the diencephalon of people with ME/CFS, which is far from adequately proven, this doesn't actually explain a psychosomatic mechanism. There is no credible evidence for an unusual prevalence of pre-illness inadequate stress management in people with ME/CFS. The Dubbo study which followed a large number of people after several infective illnesses did not find any distinguishing psychological characteristics in those who went on to develop post-viral fatigue/ ME/CFS.

@ME/CFS, you have an unexplained leap from the ill-founded conjecture of 'inadequate stress management' to an (assumed on as-yet inadequate evidence) 'abnormal blood supply to a part of the brain'.

In the case of, for example, abrupt onset ME/CFS, how does 'overly authoritarian education' actually cause people to be going about their adult lives quite happily and productively for years with no unusually inadequate stress management and then suddenly develop and sustain low blood supply to the diencephalon?

As you know I have published an article on a hypothesis. I have also written a book in which about 70-80 pages are devoted to explaining the pathogenesis of ME/CFS which I find plausible. Regretfully, the book is in Dutch. So, it is not possible to summarise all in one paragraph. It has been proven that stress causes direct DNA changes (epigenetics) and that a number of external factors play a role in whether or not such DNA changes do occur. Among the external factors nutrition, life style and exposure to environmental pollution may play a role. Stress also occurs during infant life and may have introduced epigenic schanges making e.g. the immune system of a particular person react in a disordered way to, e.g. infection. The impairment of blood flow and technecium inclusion in the brain is mostly located at the prefrontal region and the diencephalon. Within the forthcoming months we hope to conduct further studies on the effect of DCA on the NeuroSpect findings, and of repetitive transcranial magnetic stimulation (rTMS), as well as of hyperbaric oxygen therapy (HOT) on brain blood circulation. I am convinced of the organic changes occurring in ME/CFS patients, but the "first cause" of these remains speculative. Nonetheless several pieces of the puzzle are already fitting.

 

[ME/CFS]

@ME/CFS

If you are continuing to track the response in your patients, are the responding patients continuing to respond or are you seeing signs of tachyphylaxis?

Thank you for participating in this forum.

4 responding patients have been taking DCA treatment since 12 months, and continue to benefit.

 

[Wonko]

...and don't forget that epigenetic changes experienced by a mother can also influence their offsprings epigentics, the most well known example being obesity risk factors increasing if the mother's diet was poor before conception/during gestation.

So...maybe the unremembered stress/trauma is unremembered because it never happened, to the individual who has M.E., but instead I may have M.E. because my mother was stressed by being shouted at by a sargent when she was in the air force, which caused epigenetic changes which were passed on to me.

Just how far back can these changes go? Maybe I have M.E. because 40,000 years ago one of my ancestors got a bit stressed that pot noodles had not been invented yet?

 

[Alvin]

As you know I have published an article on a hypothesis. I have also written a book in which about 70-80 pages are devoted to explaining the pathogenesis of ME/CFS which I find plausible.

Thats great but does not prove you have figured out this disease. We like theories but we eye them all with a critical mind.

Regretfully, the book is in Dutch. So, it is not possible to summarise all in one paragraph.

So if it were in English you could summarize it in one paragraph?

It has been proven that stress causes direct DNA changes (epigenetics) and that a number of external factors play a role in whether or not such DNA changes do occur. Among the external factors nutrition, life style and exposure to environmental pollution may play a role. Stress also occurs during infant life and may have introduced epigenic schanges making e.g. the immune system of a particular person react in a disordered way to, e.g. infection.

Almost any "plausible" theory contains some truth.

I am convinced of the organic changes occurring in ME/CFS patients, but the "first cause" of these remains speculative. Nonetheless several pieces of the puzzle are already fitting.

If you design a theory based on pieces you already have then of course it will fit.

The lack of universally effectieve treatments is due to the fact that ME/CFS is not a disease, but that the name covers many different diseases involving several pathogenic processes

So your saying only some have pyruvate dehydrogenase kinase inhibition caused by psychosomatic trauma and you can determine who by a questionnaire that doesn't pass the laugh test but is treatable by your patent protected product?
So do the others not have pyruvate dehydrogenase kinase inhibition but somehow still registered as being so by Fluge/Mella?

Your house of cards is making little sense. I think what your doing is cherry picking information you like and forming a sciencey theory around it, defending it from patient observations and getting mad we are not towing your line.

 

[ME/CFS]

...and don't forget that epigenetic changes experienced by a mother can also influence their offsprings epigentics, the most well known example being obesity risk factors increasing if the mother's diet was poor before conception/during gestation.

So...maybe the unremembered stress/trauma is unremembered because it never happened, to the individual who has M.E., but instead I may have M.E. because my mother was stressed by being shouted at by a sargent when she was in the air force, which caused epigenetic changes which were passed on to me.

Just how far back can these changes go? Maybe I have M.E. because 40,000 years ago one of my ancestors got a bit stressed that pot noodles had not been invented yet?

It is Always nice if comments are based on knowledge , which is evidently not the case with this comment. It is not by rediculising that you can make a scientific argument.

 

[adambeyoncelowe]

I have no idea what this means.

I made a genuine attempt to understand what might be going on in your mysterious formulae. I found something that seemed interesting to me. I shared it here. If you would like to repeat the calculations I made and tell me where I went wrong, please do. I'd be happy to share the particular sets of figures I used to do the calculations.

Once I found something that looked interesting to me, I speculated on what might be going on. I was also aware that in your initially published study of 10 patients, only 5 responded, and the other 5 turned out to have been misdiagnosed. I wondered whether that could be the case in your second study too. If that is not the case, that's fine - it was simply a hypothesis based on available evidence, namely the formula and the FSS questionnaire.

When I said it was fun, I meant in the sense that I find doing science and maths fun, I didn't mean to imply that I find your patients funny. I am a severely affected patient and have had ME for 28 years. I would never make fun of other patients, whether they have ME or depression or anything else.

Could you post your numbers and working for me? I still can't wrap my head around it and keep getting numbers way off what they should be. It doesn't help that the explanations are all quite long and I'm in a major crash right now, post-NICE.

 

[Wonko]

It is Always nice if comments are based on knowledge , which is evidently not the case with this comment. It is not by rediculising that you can make a scientific argument.

I agree, it is nice when arguments are based on scientific reasoning. I further agree that it is nice when arguments are based on knowledge and not self serving speculation, whatever the reason behind such a position.

Maybe you should apply these considerations, that we both agree on, to your reasoning and posts, as, from my position, you seems to be lacking in both understanding and scientific reasoning. Maybe after you have done so you might understand the points I was using reductio ad absurdum to illustrate the flaws in your argument(s) by.

 

[Trish]

Here's the information you need to do the calculation:

Read and circle a number.
Strongly Disagree → Strongly Agree

1. My motivation is lower when I am fatigued. 1 2 3 4 5 6 7
2. Exercise brings on my fatigue. 1 2 3 4 5 6 7
3. I am easily fatigued. 1 2 3 4 5 6 7
4. Fatigue interferes with my physical functioning. 1 2 3 4 5 6 7
5. Fatigue causes frequent problems for me. 1 2 3 4 5 6 7
6. My fatigue prevents sustained physical functioning. 1 2 3 4 5 6 7
7. Fatigue interferes with carrying out certain duties and responsibilities. 1 2 3 4 5 6 7
8. Fatigue is among my most disabling symptoms. 1 2 3 4 5 6 7
9. Fatigue interferes with my work, family, or social life. 1 2 3 4 5 6 7

Based on 33 "cases" the formula is presently:
logit(p)= 7.09 + 2.27(item 3 before treatment) + 3.39(item 4 before) - 2.57(item 7 before) -1.97(item9 before) -2.52(total FSS before).

logit (p) = ln {p/(1-p)}

So, e^logit (p) = p/(1 - p), and solve for p.

Rearranging that equation gives:
p = e^logit (p) / ( 1 + e^logit (p) )
This will always give an answer between 0 and 1.

Here's a couple of examples of my calculations:
FSS scores: 4,7,5,7,4,7,6,5,5
Total 50, FSS = 50/9 = 5.556,
Logit(p) = 7.09 + 2.27 x 5 + 3.39 x 7 - 2.57 x 6 -1.97 x 5 -2.52 x 50/9 = 2.9
p = e^2.9 / ( 1 + e^2.9 ) = 0.95 rounded to 2 dp.

FSS scores: 7,4,5,4,7,3,6,7,7
Total 50, FSS = 50/9 = 5.556,
logit(p)= 7.09 + 2.27 x 5 + 3.39 x 4 - 2.57 x 6 -1.97 x 7 -2.52 x 50/9 = -11.21
p = e^-11.21 / ( 1 + e^-11.21 ) = 0.000013 = 0.00 to 2 dp.

So my first example, which I filled in on the basis of someone whose symptoms were mainly physical ME type symptoms, has a 95% chance of recovery.
My second example, which I filled in on the basis of someone whose symptoms had a more depressive pattern had a 0% chance of recovery.

As I said in my earlier post, I may have misinterpreted the formula.

 

[Wonko]

It is Always nice if comments are based on knowledge , which is evidently not the case with this comment. It is not by rediculising that you can make a scientific argument.

BTW As someone who's suffered from M.E. for over 31 years, speaking to someone who seemingly hasn't, who also appears to have severely limited understanding about M.E. (based entirely upon your own statements about your "research" and your interpretation of your results), there exists the possibility that I may find your choice of user name insensitive and offensive.

I can only assume this was your intention when choosing it.

 

[adambeyoncelowe]

Here's the information you need to do the calculation:







Rearranging that equation gives:
p = e^logit (p) / ( 1 + e^logit (p) )
This will always give an answer between 0 and 1.

Here's a couple of examples of my calculations:
FSS scores: 4,7,5,7,4,7,6,5,5
Total 50, FSS = 50/9 = 5.556,
Logit(p) = 7.09 + 2.27 x 5 + 3.39 x 7 - 2.57 x 6 -1.97 x 5 -2.52 x 50/9 = 2.9
p = e^2.9 / ( 1 + e^2.9 ) = 0.95 rounded to 2 dp.

FSS scores: 7,4,5,4,7,3,6,7,7
Total 50, FSS = 50/9 = 5.556,
logit(p)= 7.09 + 2.27 x 5 + 3.39 x 4 - 2.57 x 6 -1.97 x 7 -2.52 x 50/9 = -11.21
p = e^-11.21 / ( 1 + e^-11.21 ) = 0.000013 = 0.00 to 2 dp.

So my first example, which I filled in on the basis of someone whose symptoms were mainly physical ME type symptoms, has a 95% chance of recovery.
My second example, which I filled in on the basis of someone whose symptoms had a more depressive pattern had a 0% chance of recovery.

As I said in my earlier post, I may have misinterpreted the formula.

Oops! I wasn't dividing my total FSS number by 9! D'oh!

I've done the equation and it looks like I have an almost 100% chance of recovery (I got 0.9998).

That's interesting!

 

[wdb]

I was looking for some data on how often drugs from promising pilot studies actually turn out to be tolerable effective treatments, it's pretty low, the the table below suggests a large majority fail a phases 2 or 3, and until phase 2 is complete the drug is not presumed to have any therapeutic effect whatsoever.

 

[Trish]

Oops! I wasn't dividing my total FSS number by 9! D'oh!

I've done the equation and it looks like I have an almost 100% chance of recovery (I got 0.9998).

That's interesting!

Until Dr Comhaire checks my version of the calculations, I wouldn't take my version as the correct one. The formula is confusing in specifying ''Total FSS''. If you put the total in the formula, rather than the average, it will always give a p value around zero, so I assumed he meant average across the 9 scores.

But the more important health warnings are:

1. This formula has only been applied on the data it was derived from. Dr. Comhaire has provided us with NO evidence that he has tested it on a completely new and separate cohort of patients. It is, therefore, not a predictive formula. It is a retrospective formula that only applies to his current cohort.

His previous formula based on part of this cohort was completely different to this one, and gave wildly different probabilities, so proved to be a complete dud as a predictive formula.

2. This trial is an open label trial using an inappropriate subjective outcome measure based on fatigue, which is not the defining symptom of ME/CFS. (Remind you of anything?)

3. And, as @wdb has just wisely reminded us, stage 1 trials usually fail to result in successful stage 3 trials. (think Rituximab).
This trial is, as Dr Comhaire himself has pointed out, a preliminary proof of concept trial.

Personally I think it is irresponsible to publish so called 'predictive' formulae of likely success when they are, as his first formula demonstrated, nothing of the kind.

Again, to quote Dr. Comhaire himself, nobody should experiment with this formulation unless they are part of a properly set up and ethically approved medical trial done under medical supervision.

 

[adambeyoncelowe]

sodium dichloroacetate

Until Dr Comhaire checks my version of the calculations, I wouldn't take my version as the correct one. The formula is confusing in specifying ''Total FSS''. If you put the total in the formula, rather than the average, it will always give a p value around zero, so I assumed he meant average across the 9 scores.

But the more important health warnings are:

1. This formula has only been applied on the data it was derived from. Dr. Comhaire has provided us with NO evidence that he has tested it on a completely new and separate cohort of patients. It is, therefore, not a predictive formula. It is a retrospective formula that only applies to his current cohort.

His previous formula based on part of this cohort was completely different to this one, and gave wildly different probabilities, so proved to be a complete dud as a predictive formula.

2. This trial is an open label trial using an inappropriate subjective outcome measure based on fatigue, which is not the defining symptom of ME/CFS. (Remind you of anything?)

3. And, as @wdb has just wisely reminded us, stage 1 trials usually fail to result in successful stage 3 trials. (think Rituximab).
This trial is, as Dr Comhaire himself has pointed out, a preliminary proof of concept trial.

Personally I think it is irresponsible to publish so called 'predictive' formulae of likely success when they are, as his first formula demonstrated, nothing of the kind.

Again, to quote Dr. Comhaire himself, nobody should experiment with this formulation unless they are part of a properly set up and ethically approved medical trial done under medical supervision.

Absolutely. I would never dream of trying something like sodium dichloroacetate. It sounds like nasty stuff. Some of the side effects: 'dichloroacetate kills brain, breast and lung cancer cells'; and causes 'nerve damage resulting in weakness and numbness, liver damage and low blood sugar' (according to this article).

Your way (of calculating the numbers) seems to work. I tried it about four times and got different numbers every time (all of which were massively out). Then I did it your way and got a number that was in range.

When I say the result is 'interesting', I mean that I'm sceptical such a treatment (well, any treatment!) would have almost 100% likelihood of success. It seems exceedingly unlikely.

 

[ME/CFS]

Could you post your numbers and working for me? I still can't wrap my head around it and keep getting numbers way off what they should be. It doesn't help that the explanations are all quite long and I'm in a major crash right now, post-NICE.

The value coming out of the formula is lpgit(p) which should be converted to (p), being the statistical probability of being a "reponder".