Rituximab and placebo response

[rvallee]

So how do you explain the results of the unblinded rituximab study versus the blinded ones?

I think on this one it's safe to say that it simply shows the problems with self-reports from arbitrary questionnaires, particularly that they are detached from objective reality. Maybe a competent bespoke questionnaire would have more accurate results but the typical ones used have little relevance.

In fact, it even shows that it's completely unnecessary to go through the trouble of making up an elaborate system to convince people, since questionnaires don't even capture much that has to do with the illness itself, they rarely ask the right questions and often have misleading options. So the millions wasted on PACE were even more wasted than initially thought as it could have been done for 1/10th the price and still gotten the same illusionary benefits. But then it would only have been 1/10th as convincing, sunk cost and all.

Asking the wrong questions generally leads to useless answers. That probably explains 90%+ of the usual "placebo" response in ME research, which isn't even placebo so much as an uncertainty baked into the questionnaires themselves. Like asking people to guesstimate their weight after not weighing themselves for a period of time. Some will get it pretty close. Others not. But asking the same a bit later will have the same general variations but some who were close would be less so and vice versa. It's just noise.

 

[Jonathan Edwards]

I think on this one it's safe to say that it simply shows the problems with self-reports from arbitrary questionnaires, particularly that they are detached from objective reality.

I don't think it is as simple as that. What I am referring to is the fact that in the unblinded study raw data you see major swings in self-reported measures, considerably greater than I have seen in the PACE data, and these swings repeatedly go up and down in exact synchrony with the supposed time scale of improvement and relapse expected in a response to rituximab. The weird thing is that these people relapse when their B cells returned, just as would have been predicted from a real response and then they improve again with another dose. Whereas in the initial blinded study there was no pattern to any of the data, once it was thought that there might genuinely be an effect and the time frame from autoimmune disease was understood, the data followed expectations with great regularity. And the synchronised ups and downs went on for up to three years.

If the proportion of people showing this sort of pattern had had a genuine response to drug then it is very hard to see why it did not show up in the blinded phase 3 trial as a difference between treatment and control.

I have no vested interest in thinking that there is placebo response in ME and it is certainly not 'convenient' for me as an explanation. But it seems to be the reality and I have not yet seen any argument as to why it should not be the reality. Maybe this is bias based on both patient and investigator expectation and therefore not pure classic placebo response but if you look at the graphs it is very hard to see how it can be anything other than expectation bias.

 

[Trish]

Have the Rituximab trial authors published anything learned from the blood samples they took from patients in the trials? Could there be anything there that gives any clues?

 

[Jonathan Edwards]

Have the Rituximab trial authors published anything learned from the blood samples they took from patients in the trials? Could there be anything there that gives any clues?

I don't remember how much was published with the trials. There was a paper on autoantibodies with Carmen Scheibenbogen but I found it hard to interpret. I have probably forgotten anything that was published simply because I am pretty sure they found nothing much, other than of course that B cells went away and came back.

 

[duncan]

So @Jonathan Edwards , it would be fair to say you believe ME/CFS is not a disease and ME/CFS sufferers can have remissions/improvements that last for years due simply to the placebo effect?

 

[Jonathan Edwards]

So @Jonathan Edwards , it would be fair to say you believe ME/CFS is not a disease and ME/CFS sufferers can have remissions/improvements that last for years due simply to the placebo effect?

The certainty is all yours Duncan.

 

[duncan]

The certainty is all yours Duncan.

Well, in your capacity as a medical doctor, do you know of any diseases that have a history of a swathe of patients with remissions/improvements that last for years attributed to the placebo effect? Is there reasonable precedent here?

 

[Jonathan Edwards]

I am not sure that comparison with other diseases has proved terribly helpful in working out what is going on in ME on any front. I think there may be more uncertainties here than we understand. But if the evidence points to a placebo response I think that has to be taken reasonably at face value.

 

[duncan]

I am not sure that comparison with other diseases has proved terribly helpful in working out what is going on in ME on any front.

We couldn't agree more here.

I think there may be more uncertainties here than we understand.

Again agreed.

But if the evidence points to a placebo response I think that has to be taken reasonably at face value.

No. That is inference, and it belongs imo more in the psych community than a discussion of ME/CFS as an organic disease. This theory, this line of thinking, leads over a cliff. You couple that with your stated inclination that ME/CFS is NOT a disease, and where does that leave us?

 

[Jonathan Edwards]

It leaves me giving up trying to respond to a string of non sequiturs, Duncan.

 

[duncan]

It's a non sequitur to suggest that a belief that ME/CFS is not a disease, coupled with a belief that ME/CFS patients can enjoy improvements after medication that last for years, but due solely to the placebo effect, might present potential problems?

If so, I'm guilty. But you've managed to not address my observations and concerns. You also have not provided any sort of relevant precedent for the years-lasting placebo theory.

 

[Marky]

I don't remember how much was published with the trials. There was a paper on autoantibodies with Carmen Scheibenbogen but I found it hard to interpret. I have probably forgotten anything that was published simply because I am pretty sure they found nothing much, other than of course that B cells went away and came back.

They did this one on pyruvate dehydrogenase function https://insight.jci.org/articles/view/89376

 

[Barry]

I don't think it is as simple as that. What I am referring to is the fact that in the unblinded study raw data you see major swings in self-reported measures, considerably greater than I have seen in the PACE data, and these swings repeatedly go up and down in exact synchrony with the supposed time scale of improvement and relapse expected in a response to rituximab. The weird thing is that these people relapse when their B cells returned, just as would have been predicted from a real response and then they improve again with another dose. Whereas in the initial blinded study there was no pattern to any of the data, once it was thought that there might genuinely be an effect and the time frame from autoimmune disease was understood, the data followed expectations with great regularity. And the synchronised ups and downs went on for up to three years.

If the proportion of people showing this sort of pattern had had a genuine response to drug then it is very hard to see why it did not show up in the blinded phase 3 trial as a difference between treatment and control.

I have no vested interest in thinking that there is placebo response in ME and it is certainly not 'convenient' for me as an explanation. But it seems to be the reality and I have not yet seen any argument as to why it should not be the reality. Maybe this is bias based on both patient and investigator expectation and therefore not pure classic placebo response but if you look at the graphs it is very hard to see how it can be anything other than expectation bias.

What self-reported outcome measures are you speaking of here? Given we are talking about placebo responses and self reported outcomes, is it possible that in the unblinded study the trial design (or things outside of the design) inadvertently boosted expectations, without the investigators realising they were doing so? I imagine that in an unblinded study it can very difficult spot all the signals that participants might pick up on?

 

[Saz94]

I have just been informed, by somebody who knows more about the trial than me, that Fluge and Mella changed the dosing approach between Phase 2 and Phase 3 of the trial. That's worth consideration as a reason why the phases got different results.

 

[Forbin]

It's probably worth remembering that, between the trial and placebo groups combined, 70% of the patients did not respond (105 out of 151).

26% of those who received rituximab (20 of 77), and 35% of those who received placebo (26 of 74) responded, making the combined response rate 30%.

I wonder if the response rate was lower in the rituxumab group simply because the side effects of the drug itself impeded the manifestation of any kind of placebo effect. On the other hand, you might think that an "illusory" rituximab response rate might be higher, as the drug's side effects would signify to the patients that that they were indeed getting the drug.

I guess one interpretation might be that 30% of ME patients are experiencing placebo triggered effects to the point of self-reporting improvement during a double-blinded study. I'm not sure how surprising this is. How does this compare to the placebo response rate in other diseases, where the nature of the disease mechanism is better understood?


[ ETA: Looking for information on "typical" placebo response rates, the figure of 30% does seem to come up going back to the 1950's.

Beecher reviewed 15 articles describing the treatment of 1,052 patients and estimated that across all the conditions studied, the average placebo effect was 32%. These results were confirmed by Haas a few years later based on 1,400 cases from 96 articles. He also found an average of around 30%, but he found a lot of variation depending on what was being studied. The improvement in the symptom of pain varied from 15% to 60%.
https://sciencebasedmedicine.org/placebo-are-you-there/

A 2015 article on Placebo-Associated Blood Pressure Response found:

Results: At the end of the titration phase, 58 patients who were treated with placebo (31%) achieved a goal diastolic blood pressure lower than 90 mm Hg and 57 (30%) achieved success at 1 year.

From an article from Harvard Medical School on The Power of the Placebo Effect:

More recently, however, experts have concluded that reacting to a placebo is not proof that a certain treatment doesn't work, but rather that another, non-pharmacological mechanism may be present.

]

 

[Sly Saint]

I'm sure this will have been covered somewhere : ie the use of saline-infusion as the placebo given that this is provided as treatment for ME and CFS in some places?

Also;
can anyone shed any light on this Julia Newton research (funded by AfME) from 2016/2017

• Full title
  The effect of intravenous bolus's of fluid upon cardiac and brain function in chronic fatigue syndrome (CFS): a proof of concept study

https://www.hra.nhs.uk/planning-and...ies/research-summaries/action-for-me-cardiac/

I have been unable to find any details.

 

[ProudActivist]

I'm sure this will have been covered somewhere : ie the use of saline-infusion as the placebo given that this is provided as treatment for ME and CFS in some places?

Also;
can anyone shed any light on this Julia Newton research (funded by AfME) from 2016/2017

• Full title
  The effect of intravenous bolus's of fluid upon cardiac and brain function in chronic fatigue syndrome (CFS): a proof of concept study

https://www.hra.nhs.uk/planning-and...ies/research-summaries/action-for-me-cardiac/

I have been unable to find any details.

I would like to see the results of this too.

 

[Guest 2176]

I'm sure this will have been covered somewhere : ie the use of saline-infusion as the placebo given that this is provided as treatment for ME and CFS in some places?

Also;
can anyone shed any light on this Julia Newton research (funded by AfME) from 2016/2017

• Full title
  The effect of intravenous bolus's of fluid upon cardiac and brain function in chronic fatigue syndrome (CFS): a proof of concept study

https://www.hra.nhs.uk/planning-and...ies/research-summaries/action-for-me-cardiac/

I have been unable to find any details.

even though i think saline produces really intense changes in ME/CFS symptoms, I also thinik the effect is somewhat short lasting, so idk if I'd think it was responsible for the improvements

on the other hand there is such a thing as an active placebo. as in, something does something but we also fill in the blanks with our hopes and optimism or current state of mind to the extent that the original signal is obscured somewhat. with saline this might look like: the psychological benefit of returning to a physiological state of "normal" even for a short time, periodically, thus raising ones overall morale

throw in some antihistamines and steroids and i could see why at the very least self reported fatigue would change a lot

however, the thread was started with what i believe was an inappropriate analogy to cci surgery as placebo... and nobody answered why that would be an appropriate analogy if the rituximab study didn't track step counts or any objective measures, as some of the ppl who have recovered from ME due to cci surgery have.

 

[Jonathan Edwards]

and nobody answered why that would be an appropriate analogy if the rituximab study didn't track step counts or any objective measures, as some of the ppl who have recovered from ME due to cci surgery have.

The problem is that for CCI we do not even have controls, so it makes no difference whether measures are objective or subjective. Moreover, in this context step counts are subjective measures. Placebos are very likely to be able to improve step counts. What they do not improve are things like blood tests.

 

[Saz94]

The problem is that for CCI we do not even have controls, so it makes no difference whether measures are objective or subjective. Moreover, in this context step counts are subjective measures. Placebos are very likely to be able to improve step counts. What they do not improve are things like blood tests.

In a PWME, a placebo is not going to produce a persistent improvement in step count. Unless the PWME was currently doing less exercise than they actually could. Which is unusual - most of us are doing the max thay we can without making ourselves more ill. If someone increased their activity because a placebo made them think they were better, then PEM would come to bite them on the ass after a while, and they'd go back to how they were before or even worse. If placebos could improve our ability to do physical activity, then GET/CBT wouldn't be dangerous for us.

Didn't we criticise the PACE trial for not using measures of physical activity.