Rituximab and placebo response

[Jonathan Edwards]

Yes, they thought they were responding to the drug, hence reporting biases...

I don't follow that. If they thought they were better, and that this was due to a drug, then that counts as a placebo effect, doesn't it?
I would use 'reporting bias' to cover other sorts of effects, like wanting to please the investigators.

 

[Snow Leopard]

I don't follow that. If they thought they were better, and that this was due to a drug, then that counts as a placebo effect, doesn't it?

It depends if you think that reporting biases are part of the placebo effect. I don't consider people optimistically reporting they are healthier than they actually are as a placebo effect. I only consider actual changes in health or symptoms as a potential placebo effect.

Many of us have had experiences where we have tried some sort of 'treatment' and we interpret and exaggerate small changes in health, that could be due to natural variation as much larger than they really are. In hindsight, when the treatment fails to deliver any sort of long term improvement, we realise we had been overly optimistic in our assessment of our health.

 

[Jonathan Edwards]

Had some treatment been used during that time, it would likely have been at least correlated with remission (though incorrectly as recovery).

I think we have to take into account the fact that with repeated rituximab infusions a number of patients had saw tooth 'response' charts coinciding with the theorised period of benefit of about 6 months corresponding to the period of B cell depletion. The likelihood of chance correlation looked to me to be negligible.

 

[Jonathan Edwards]

I don't consider people optimistically reporting they are healthier than they actually are as a placebo effect. I only consider actual changes in health or symptoms as a potential placebo effect.

The trouble is that I don't think there is any way to validate that distinction. There is no 'god's eye view' of how healthy someone actually is in a condition characterised only by symptomatology.

I actually think it doesn't really matter for the discussion, which as I see it is about how we judge the reliability of evidence of treatment effects.

 

[duncan]

Many of us have had experiences where we have tried some sort of 'treatment' and we interpret and exaggerate small changes in health, that could be due to natural variation as much larger than they really are.

I think it was Dr. Bell that reported a similar thing about some of his ME/CFS patients. He observed some patients actually thought they were better, but when pressed about their health, they proved to be still sick.

 

[Robert 1973]

I can understand that PWME find it hard to believe that ME might respond long term to placebo but we have pretty good evidence that at least a subset of patients do. Maybe there are two illnesses, one that respond and one that does not.

Before I started following research more closely, I used to assume that this was the case. I used to assume that CBT (and various other therapies) probably helped some people, but that those people had a different illness to me. But the evidence from PACE changed my view because it tells us that CBT/GET doesn’t significantly help anybody diagnosed with ME/CFS – certainly not long term – even using loose diagnostic criteria.

If one was to infer from the Rituximab trials that a sub-group of people diagnosed with ME/CFS can respond positively long term to a placebo, then one would also have conclude that CBT/GET is not only ineffective but also a very bad placebo.

Also, if one accepts that placebos do not directly affect long term pathology, if a sub-group of patients respond positively long term to a placebo, does that imply that their degree of incapacity is at least partly affected by their thoughts? What are the other possible explanations?

As with the recent meta-epidemiological study on blinding (discussed here), I find the data very puzzling.

(FWIW, as I’ve said before, I suspect that are more than 2 different illnesses within the ME/CFS diagnosis, even using stricter criteria. It may be that a majority have the same thing, but even then I suspect that there will be various different as yet unknown illnesses within the minority.)

 

[Snow Leopard]

I don't think there is any way to validate that distinction. There is no 'god's eye view' of how healthy someone actually is in a condition characterised only by symptomatology.

Now you are unnecessarily leaping into philosophy. I am not asking for an ultimate distinction, simply reasonable quality of evidence. This is why we keep mentioning objective outcomes. Yes, physical activity levels can be influenced by all sorts of factors, but the fact remains that if participants are able to keep this up for a long time, then their underlying health must be able to sustain it, hence it provides valuable additional perspective. Most importantly, evidence from clinical trials has shown that true activity levels can deviate from self-reported activity levels and reported symptomatology.

A randomised blinded trial doesn't necessarily mean that there aren't indicators such as side effects that subtly reveal to participants that they are in fact taking the active drug. As far as I know, Fluge 2011 did not ask participants at the end of the trial whether they knew if they were taking the active drug or the placebo... However two participants in the placebo group ended up withdrawing (pregnancy and alternative therapy), hinting that they knew they were not receiving the active treatment.

 

[adambeyoncelowe]

As I understand it, the measures were subjective but the trial was blinded? So what we may have is patients reporting improvement despite no improvement in both arms--perhaps with the lack of serious side effects in the sham arm leading to better results in that group.

It's also possible that patients really, really wanted this work--for themselves but also for their community--and so potentially felt a duty to report improvement even where there was none.

The other possibility that springs to mind is that all the 'responders' were a subset that would have improved anyway, and only the pattern of improvement was influenced by expectations while the improvement itself was due to the natural illness course.

I'm referring to studies where about a third of patients get a little or somewhat better, about a third stay the same, and about a third get worse.

Factor in the Bell findings, where patients report being recovered despite being hugely impaired still, and it's not too much of a stretch.

Of course, the patients on placebo weren't just taking salt water, either. It would be interesting to compare the results of the rituximab placebo with something that is truly inert, but that might be impossible.

 

[Jonathan Edwards]

then one would also have conclude that CBT/GET is not only ineffective but also a very bad placebo.

Yes that was the conclusion I put in my report to NICE. And not very surprising I would have thought. If CBT had been delivered universally by people with great charisma it might have done better. There was I think an interesting comparison of CBT at King's and in Holland that suggested that UK practitioners may not be that good at the magic.

There are of course all sorts of confounders relating to who volunteers for any given study.

Also, if one accepts that placebos do not directly affect long term pathology, if a sub-group of patients respond positively long term to a placebo, does that imply that their degree of incapacity is at least partly affected by their thoughts? What are the other possible explanations?

It might seem like that. On the other hand I am not sure that these things have to 'be affected by thoughts' to respond to a placebo. The thoughts may be by products.

 

[Jonathan Edwards]

Now you are unnecessarily leaping into philosophy. I am not asking for an ultimate distinction, simply reasonable quality of evidence.

I disagree. I am talking about whether or not we can make valid distinctions on a very pragmatic level. I don't think so. To me feeling optimistic despite no change in 'real health status' is the paradigmatic placebo situation. People with osteoarthritis get placebo responses to injections that make them feel their pain is less but the x-ray stays the same.

Activity levels tell us how someone has done, not how much they can do. They may be a good guide and a lot better than a questionnaire but I don't think we can consider them an objective measure of true health status in the way an x-ray might (with the caveat that the x-ray may not actually matter to symptoms anyway).

I don't think there are any hard and fast rules for any of this and all possibilities need considering.

 

[Jonathan Edwards]

As I understand it, the measures were subjective but the trial was blinded? So what we may have is patients reporting improvement despite no improvement in both arms--perhaps with the lack of serious side effects in the sham arm leading to better results in that group.

That was the phase 3 trial, Adam. I was using the unblinded continuation of phase 2 as an indicator of a placebo response. There are no controls n this context so 'response' is in terms of relation to states at other points in time. If you have a sufficiently long pharmacodynamic monitoring period then you start to get information of a sort that does not come up in simple comparator trials. It gets complicated.

 

[adambeyoncelowe]

That was the phase 3 trial, Adam. I was using the unblinded continuation of phase 2 as an indicator of a placebo response. There are no controls n this context so 'response' is in terms of relation to states at other points in time. If you have a sufficiently long pharmacodynamic monitoring period then you start to get information of a sort that does not come up in simple comparator trials. It gets complicated.

Gotcha. Thanks for the clarification.

 

[Saz94]

Yes that was the conclusion I put in my report to NICE. And not very surprising I would have thought. If CBT had been delivered universally by people with great charisma it might have done better. There was I think an interesting comparison of CBT at King's and in Holland that suggested that UK practitioners may not be that good at the magic.

I hope I have misunderstood you here. It sounds like you are saying that you told NICE that PWME need a more persuasive version of CBT?!

 

[spinoza577]

I hope I have misunderstood you here. It sounds like you are saying that you told NICE that PWME need a more persuasive version of CBT?!

I read it as trick to convey that nonsense has happened.

 

[voner]

Jonathan,

I have a question that stems from the Norwegian documentary on the rituximab studies that was posted in this forum:

https://www.s4me.info/threads/left-out-norwegian-documentary-on-me.13393/

in this documentary one of the featured people is a woman who is obviously pretty close to housebound and sick and she ends up being able to run a long road race and evidently feeling quite healthy after getting rituximab infusion’s in the study. how do we view her in terms of a placebo? do we say that she has experienced a placebo effect? Or do we say that she’s just an anecdotal story (n=1) and thus one can say nothing about her dramatic turn around? Or??

 

[Trish]

in this documentary one of the featured people is a woman who is obviously pretty close to housebound and sick and she ends up being able to run a long road race and evidently feeling quite healthy after getting rituximab infusion’s in the study. how do we view her in terms of a placebo? do we say that she has experienced a placebo effect? Or do we say that she’s just an anecdotal story (n=1) and thus one can say nothing about her dramatic turn around? Or??

Given that the study was done, and the film was made, over a period of several years, it's not altogether surprising that a few patients might go into remission over that period, and this person who went into remission happened to be in the Rituximab group. Since just as many people in the placebo group improved too over that period, we can't say this case demonstrates that Rituximab works.

 

[Jonathan Edwards]

Or do we say that she’s just an anecdotal story (n=1) and thus one can say nothing about her dramatic turn around?

Yes, an individual case is hard to make much of. It is the recurrent patterns that I think are more informative. The time course of change after treatment is something I always find more useful than just before and after. The extension of the phase 2 study seems to me most interesting in this respect.

 

[AliceLily]

I haven't been able to read all this thread so I may have missed finer detail and getting the wrong end of the stick. I apologize if I have.

I don't know how anyone with ME can respond to a placebo. Any placebo like experience will be the natural fluctuation of the ME itself where one has an easing of symptoms unrelated to any placebo that was given.

This is where we can end up getting the wrong interpretation of a situation.

Having experienced severe ME for many years I got only to familiar with how severe PEM could get and it's nature. Day after day, for many years I was on the rollercoaster of PEM. It's a miserable lot of suffering in those years.

Now at moderate level I can get away with doing more because of the many years of resting my way out of severe ME and the constant rest routine that I have to apply everyday to manage my ME and stay at this moderate level.

There are so many stages/levels to ME and fluctuations within each.

 

[Marky]

To me the phase 3 trial placebos told me that we cant successfully study this disease without objective markers like steps or school/work.

When you have ME in moderate/severe, very small increases in function can completely change how you perceive the level of suffering and function

You have to take into account that many people before starting the trial has been completely stuck in rooms or their houses

When they get out and feel part of something that may influence their health for the better, this renewed state of mind with hope may sneak into the questionnaires

I suspect one might even downplay the PEM episodes

I don`t think its impossible that something like this occurred in the placebo group

Now I personally don`t think there is a possibility for placebo to cure ME

In the lightning process trial on adolescents, many reported improvement BUT, there was no improvement on objective markers

I think it might be the same false "effects" we are seeing here

 

[anniekim]

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.

Can anyone expand further on this? Does anyone know why the dosing was changed for the Phase 3 of the trial? Many thanks.