Persistent fatigue induced by interferon-alpha: A novel, inflammation-based, proxy model of Chronic Fatigue Syndrome, 2018, Pariante et al

Thanks @Lucibee, @JohnTheJack, @Michiel Tack

Just wondering whether anyone has approached Pariente to ask what his view of the causes of ME is now. Those tweets seem to come from early 2017, and i’m not sure of the date of the review.

I think it would be useful to have on record now whether, in the light of his research paper, he still believes ME is caused by excessive rest.

I also think this is an entirely reasonable question in the light of his ( their) paper and its ( seeming) biomedical leanings after his 2012 review.

 

Let's not get too distracted by their garbage. We should be trying to support the Karl Morten team and help him get articles about his work into the press.

I heard the ME Association has a media expert working for them now. If Morten lands a big grant, this would be an opportunity to get articles into the press.

Wouldn't you love to read an article with statements such as

"Dr Morten has found that GET does not improve the metabolic abnormalities in patients."

"That the MRC has given a large grant to biomedical investigation marks a real shift in how the illness is viewed by society."

"We think this might be an illness that resembles diabetes, except that the problem is not with glucose not getting into the cell but with the utilization of glucose in the cell" (I made that up, I'm sure Morten will have interesting things to say).

 

>>It seems like "the data thugs" might be able to help us debunk some of the BPS work.

FWIW, I do not think the tools that my colleagues and I have developed will be much use here, for a number of reasons that are very boring and technical, plus the simpler one that I have no reason to think that the problems with ME/CFS research involve fabrication or falsification of data; my guess is that they will turn out to be much more about data dredging, or what the statistician Andrew Gelman calls "The Garden of Forking Paths".

But if anyone who knows a bit about means and standard deviations wants to play around with SPRITE, have at it: http://shiny.ieis.tue.nl/sprite/

 

Thanks @JohnTheJack. Do you, or anyone else, have a reference? It would be interesting to know his view now, expressed clearly.

 

@Sly Saint, thank you very much for your transcription of the interview with Pariante. I find it very difficult to take in radio (and also video) stuff. I really appreciate that you take the time to make transcriptions so I can take the information in at my slow rate.

 

Moderator note. Thread on measuring of fatigue has been moved here:
https://www.s4me.info/threads/measuring-fatigue-discussion-of-alternatives-to-questionnaires.7325/

I get the impression that a key problem we're facing is that we don't have a good measure of 'fatigue' yet people who've made their careers as experts in 'fatigue' want to try to overlook that. At this point, they know what outcomes they can 'improve', and so they can pre-specify them. Post-PACE all CBT/GET CFS trials will be doing all they can to focus on subjective self-report outcomes in a way that makes it as easy as possible for them to claim they've reached a clinically significant improvement.

Is anyone aware of any literature on that sort of long-term 'Garden of Forking Paths', with researchers choosing primary outcomes for later trials on the basis of what they can get as positive in earlier trials rather than on the basis of what is a the more useful outcome for providing patients with useful information about treatment efficacy? I get the impression that a lot of academics do not view this as a problem.

 

@Graham was a Maths teacher. Don't know if this would interest him.

 

Garden of forking paths:

http://www.stat.columbia.edu/~gelman/research/unpublished/p_hacking.pdf

 

All what follows is IMHO, obviously:

So my view would be no - personally I would just chalk this up as another scientific paper with mildly interesting findings (and to my mind they *are* interesting, even with some of the already-discussed flaws) and a massively overblown press release. Not much different to some other ME research institutes that tend to overhype, in part (understandably) to try and attract more funding. Why would we treat this paper to more forensic investigation than those other papers? It would easily be portrayed that Pariante is receiving 'vexatious' data requests solely because he's opined in the past that ME/CFS could have psychological factors (a view that recent appearances suggest he has quietly stepped back from). A treatment that other researchers who make somewhat overblown claims in the past have not received from ME/CFS patients.

This paper feels miles away from PACE, where data requests were very necessary to stop a poorly-evidenced treatment that might cause actual harm and which implicated ME patients would improve if they'd just get out of bed. We must be clear time and time again that the objection to PACE was it was poorly conducted, not because of the fear avoidance leanings of its authors. The PACE team have argued many times that the criticism of their work is because they believe in a psychological component to ME/CFS. If we start doing data requests on overhyped studies only because of some of the historic pro-psychological-factors statements of their authors, we play exactly into that narrative.

Here, even if a data request suggested the conclusion was statistically iffy, what would we have gained? I can see the headline: 'ME activists show immune system *not* linked to CFS - must be psychological then'.

So, my view: it's an interesting paper, bit iffy here and there, the news reporting was overhyped but at least was of the 'this is a real illness' variety, and it's a positive sign if even Kings - Kings! - feels the ground has shifted sufficiently that it now has to emphasise the non-psychological evidence.

 

Urk, thank you for the link - I think it's well beyond me to understand. But thank you for posting, it makes it much easier for others to follow a link rather than having to search, and then wonder which items are OK and which are kinda weirdo.

 

well, in the PACE trial, when they wanted to obtain a "normal range," they used the formula for a normally distributed population even though they knew they were working with a set of values skewed toward the healthy end. And they knew specifically that this method would provided a distorted and excessively broad "normal range" that would overlap with their definitions for disability on the same measures. I using a statistical calculation you know is the wrong one for your dataset but that makes it easier for your to claim success is going well beyond "data-dredging" into data manipulation and perhaps "falsification." These are not statistical "normal ranges," no matter what the PACE authors say.

 

Data requests could justifiably come from researchers, especially those who suspect methodological flaws. It should probably not come from patients. At this point this paper is not leading to dangerous therapies. If however this paper is used as a justification for further papers, heading even deeper into psychobabble territory, then those papers might warrant deeper scrutiny. Right now this paper serves as weak evidence for issues following infections. It says nothing useful about ME. It does not even address ME except in name, which is a flaw that can be criticised even without data.

The issue on hype is more serious than the usual. This is being pushed by the SMC, and in this area they have a history of excessive public relations hype, such as with PACE.

The initial XMRV paper had better quality evidence relating to ME cause than this one. We know how that turned out. With respect to interferon research the only substantive new thing appears to be that some patients might, just might, be more sensitive to interferon. Why is that a surprise? It happens with many other signalling factors. This is minor news right now.

Psychobabble works by constructing persuasive rhetoric. This paper could fit in with that. That would indeed be something ME activism should be interested in, especially if it develops further.

In the long run this latest iteration of interferon findings in ME might be more important with future studies if they move to higher quality methodology. I am not particularly hopeful though. Its also one small lead out of hundreds, and most leads go nowhere after further research.

 

On cytokines and supposed CFS patients, if you ignore the cohort issue (which would be a mistake) it seems they did not even test for cytokine expression shift based on duration of illness. Or did I miss it?

 

By weird coincidence, I just happened to stumble upon that letter myself today, as I was reading one from Wessely that came before it: http://sci-hub.tw/https://www.thela...6(88)90028-1/fulltext?version=printerFriendly

When this happened it seemed amazing... now I'm about to post about it it seems likely to be really boring to everyone else. I'm going for it anyway!

 

Changing which format of the chalder fatigue scale you use to try and achieve a reportable p- number?

 

It would be interesting to see how informed consent was for this trial.

Given what is known about interferon , what risks were explained to participants?

Were they all aware that potentially there was a ?% chance that they would not recover?

 

When the trial started, it was the standard treatment...

 

Moderator note: This post has been copied across to a new thread here:
Measuring fatigue. Discussion of alternatives to questionnaires

I agree that this is what needs focussing on.

There is even doubt that, as a symptom, the target should be called fatigue.

So I am increasingly of the view that what we want to measure is what Esperanza has called motor fatigue. And the need is to measure a specific pattern of motor fatigue X, that is characteristic of ME, or maybe patterns X and Y and even Z that are patterns characteristic of subsets of ME. Whether the symptom is best called fatigue does not matter so much. X or Y would be objective indicators of whatever symptomatology was responsible for the impact on active life in ME.

It would be the equivalent of respiratory function tests for asthma. Nobody assesses asthma by asking how wheezy someone is. They measure air volumes shifted. And to confirm that the problem is truly asthma, which is reversible from episode to episode, patients are asked to measure their own volumes on a daily basis.

This is of course where the accelerometers come in, but they need to be used intelligently, not just as a crude index of the number of wiggles. Off hand I can think of about a dozen PWME whose movement patterns I am familiar with. They will differ from those of healthy people in a variety of specific respects even if the total number of wiggles is not different. I am pretty sure this can be measured, but it requires intelligence in interpretation.

I do not know of literature on Garden Paths as a problem but the reality of the phenomenon is very familiar to any rheumatologist. Around 1970 the first new drug to have an immediate effect on joint pain and swelling was developed using what we now consider standard trial methods - ibuprofen. Jason et al. devised a scoring system that came to be known as DAS, based on counting numbers of swollen and tender joints. It was a useful system for anti-inflammatory drugs because it measured the rather limited benefits they gave. But it became embedded in scoring systems for all anti-rheumatics, to the extent that you could not make an application to the FDA for a new drug license without using it. This was a serious problem for the designers of collagenase inhibitors, which were intended to have no effect on symptoms but to protect tissue from damage. Another scoring system was the ACR improvement grading. ACR20 was for many years considered the end point, because a 20% improvement across various variables was considered good going. But when we started seeing ACR70 grades on a regular basis there was argument about the relevance of ACR20.

I guess what this illustrates is that picking outcome measures that look good is a regular part of treatment development and to some extent makes sense and is legitimate - but it can also distract away from more important goals. And if claims are made based on limited outcomes that are overblown, and negative findings suppressed, then things are going badly wrong.

I think a fatigue measure could be generated from accelerometers with good software. You would probably need ankle and wrist monitors at least, maybe both sides.