Rituximab and placebo response

This post and two following have been copied from this thread:
Concerns about craniocervical instability surgery in ME/CFS
______________________________

The trouble with 'objective' is that it has layers of meaning. Placebos probably do not alter things like haemoglobin or CRP level. However, although going back to work is 'objective' in one sense it is not the same sense and I think we have to accept that, as in the rituximab trial, people got well enough to get back to work having had a treatment that has no effect in itself. That may be a 'placebo effect' in the broad sense of all the reasons why people get better that are not specifically due to a treatment but it may also be in the narrower sense. We just do not know.

Having worked in clinical trials for years and rituximab specifically I think I can be pretty sure, having looked at the Norwegian results, that the drug has no effect. The reasons are complex and they are not just about the lack of any difference in the two phase 3 groups. The detailed dynamics of the open phase 2 trial makes it pretty clear that the apparent benefits from treatment were spurious. The apparent effects in the open phase 2 were not just in a tiny subgroup that might not have shown up in phase 3. The idea that a subgroup benefits doesn't fit.

 

Could you please expand a bit on this and explain in detail? I'd really like to understand.

 

There are various factors that make me say that having looked at the results in detail. But I will try to pick out the main points.

It is worth remembering that the first blinded study gave a negative result at the primary end point. It was only because it was noticed that there was an apparently significant difference at 6 months that it was thought a real benefit might have been missed.

In this first blinded study there was no pattern to the 'response' time curves. I spent ages looking for consistent curves and found nothing. When you have real responses they follow a pattern - either immediate or slow or biphasic or whatever. There was nothing to see in the first study.

After this was published I pointed out that the primary endpoint should probably have been at 6 months because that is when improvement is clear in autoimmune diseases. From this point on an unblinded follow up study was done and in this study there were consistent patterns of 'response' over 6 months, with relapse around the time of B cell return. Whether this was due to suggestibility of patients or doctors or both I don't know but it had not occurred in the blinded study.

Then in the last blinded study there were similar 'typical responses' but there were slightly more in the placebo group, so they were spurious.

Peter White tried to persuade us that the results PACE must be genuine because there could not be a major or prolonged placebo response in ME/CFS. The rituximab studies show he was wrong and that is an important part of being able to say why the PACE results are no better than any other unblinded study. You can get major placebo responses in ME that last for years. That may be an uncomfortable fact but we have to accept realities. Otherwise we are in the same fairyland as the BPS people.

If there was a real autoimmune subset they should have shown up in the first trial. The time curves for different measures in at least some cases should have been internally consistent with all the different measures going down together and up together. I could not find that anywhere in the data. It worried me from the start but I thought that the chance of an effect was so important that it was worth doing a further trial.

 

I suspect this is flat out wrong. Simply because we do not appreciate a certain mechanism(s), does not mean we slap on a voodoo label and walk away.

IMHO, it is better to simply admit we do not know which organic factors are at play, and that more research is desperately needed.

 

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

Why do patterns of response suddenly appear when patient and doctor start expecting them to appear and not before? If you look at the time profiles of the repeated treatments in the open study the changes in scores are not trivial. They are major.

Recognising that a placebo effect may occur is not voodoo. It is responsible science.

 

I cannot.

Your question is full of bias.

Not that it acts for years in objectifiable manners. Placebo is not responsible science - it's a placeholder.

 

Totally agree with Duncan.

 

Was the difference significant, or down in the noise?

ETA: What are the known side effects of Rituximab?

 

But nonetheless placebo responses must be real. Otherwise responsible scientists the world over have been getting it badly wrong all these years by controlling for it; running controlled trials, a significant component of which is to control for the placebo effect. Controlling for the effects of a real phenomenon is real science. Failing to control for placebo responses adequately, and then erroneously believing their magic cure works because of that failing - that is irresponsible science.

 

Define placebo effect, and then tell me it's parameters, eg, is it just subjective reports, do time limitations come into play, etc. THEN tell me what causes it. Then assure me that doctors don't just embrace it whenever they are too lazy to actually look for a causal mechanism...This is very little different from white coat syndrome or subconscious-at-play arguments - or prayer.

They've taken a phenomenon they can't explain, and made it so plastic as to expand or contract as needed. This is not science. At it's heart it's dangerous shorthand for "We Can't Explain Something." Psychology was built on this sort of sandy terrain.

I appreciate there is a phenomenon that sometimes occurs that we call Placebo. But I will not grant it exhaustive elasticity because it's easy to do so.

 

You were very selective in your quoting me. The answer to your questions is in the rest of that post. We often know the real existence of something long before we understand its underlying mechanisms. ME for instance.

 

But we don't lend it unrestricted domain or characteristics simply because it's convenient. Show me a text book that claims placebo lasts for years, for instance.

 

Aside from the appeal to authority fallacy, responsible "scientists" have been making a mess of things quite reliably for many, many years. Yes, they get some things right, collectively. Yes, they get some things wrong, collectively.

I can wheel out the psych model again if you wish. But "regular" medicine too.

 

No the difference was small, but in favour of the control group.

There are all sort of known side effects of rituximab but in my experience the most serious and common is pneumonitis - which in an adverse context can be fatal. Otherwise there is the long term reduction in antibody levels in some cases.

 

I would but after a few months of reading it mysteriously disappeared, just as if it had never existed.

My understanding of placebo is simple. It seems apparent that the body primes itself in response to certain stimuli, such as reving up some parts of the metabolism when you taste something sweet, even if what produced the sweet taste contains no sugar, or even calories, the body reacts as if it does. Presumably because over most of the world, throughout most of evolutionary time, a sweet taste meant sugar,

I don't view a placebo response as any different, a certain stimuli is perceived, by any means including societal, so the body reacts as if things are going to get better. Sometimes this has a temporary positive benefit, or at least effect, but it will always disappear once the body 'realises' that the promised 'treat' does not, if fact, exist.

In some cases this reving up in expectation is likely to have a net negative effect, such cases including when the reving up used resources that could only be spared if the 'treat' actually existed, when it doesn't.

In other cases there may be no obvious downside.

 

Cool. Maybe this is accurate. Maybe not. It's theory. But MY understanding is it cannot fix broken bones and it cannot last for years, ie, there are limits, usually understood to be subjective reports with time limits. This should not be a get-out-of-jail-free card.

My main problem is not that placebo doesn't happen, it's that we don't know why it does when it does with only some people. Worse, to apply it wholesale - as in it can last years - feels unscientifically hazardous to me.

 

I wouldn't. Nor would I rest my argument on such an ill-defined, amorphous concept. I'd simply say it is unclear why people got well when indications are they shouldn't have, so more research is needed. As I said, placebo is little more than a medical placeholder, and I find it inadequate and demeaning to apply to patients - not to mention a disservice to them - as liberally as is the custom today.

Great for doctors, though.

 

I would.

FYI, because many of us have been on the receiving end of a medical lie, and it matters that that be exposed.