Discussion in 'BioMedical ME/CFS Research' started by John Mac, Mar 25, 2019.
I had the impression that the illness responds to stimulants that temporarily appear to work with an even stronger pull towards a sickness state. Much like trying harder to function normally will soon result in a relapse that puts an end to trying harder for a while. That may be why it's essentially untreatable.
Someone else mentioned this, and I have experiened it as well: that after a brief but intense flu, I feel much better and I think it may be more than than merely not having the flu. Is it possible that this event temporarily resets a malfunctioning system?
I've wondered that too. And if it's about faulty homoeostasis, that would explain why some interventions work for a while, then stop. I'm thinking sleep meds, primarily, which many patients rotate to maintain efficacy.
Always interesting to see how ME/CFS is defined. Not sure about 'historically considered a neuropsychiatric disorder'; perhaps by some. Reference 3 is the Edwards, McGrath 2016 paper - the inference being that that paper establishes, or reports the establishment of a pathology that extends beyond a neuropsychiatric one.
Yes, that's the Rituximab work being referenced there in . This paper seems to have taken quite a while to get published.
Received: December 12, 2017; Accepted: February 27, 2019; Published: March 25, 2019
I don't think lower cortisol levels has satisfactorily been demonstrated. Importantly, there's the confounding effect of people with ME/CFS tending to wake later.
Reference 13 is
Tomas C., Newton J. and Watson S. (2013) 'A review of hypothalamic-pituitary-adrenal axis function in chronic fatigue syndrome', ISRN neuroscience, 2013.
Perhaps not adequately critical of the problems that seem to be common in studies of cortisol and ME/CFS?
Reference 14 is
Vangeel E., Van Den Eede F., Hompes T., Izzi B., Del Favero J., Moorkens G., et al. (2015) 'Chronic fatigue syndrome and DNA hypomethylation of the glucocorticoid receptor gene promoter 1F region: associations with HPA Axis hypofunction and childhood trauma', Psychosomatic medicine, 77(8), pp. 853–862.
That's got a number of buzzwords that I associate with studies that write the outcome and then fit data to match it. Would be interesting to look at that.
Well, I may be overly sceptical, but theoretical models based on reported findings that are wrong may not take us very far.
I think they are saying that we might not see statistically valid increases or decreases in particular substances across a population of ME/CFS patients or even consistently in a single patient over time. But there might be some global perturbation in systems regulation. So, ok, this might be an explanation of why the cortisol and other substances involved in the HPA axis don't seem to be very different on average in people with ME/CFS compared to controls while the idea of a problem in the HPA axis that so many researchers are attached to is still correct. Hmm. But I'm still listening.
(Well, actually I'm going to take a break now. But I'm interested to see what they come up with later in the paper.)
They started out with 80 patients, but only 42 had usable data for the autonomic analysis; compared with only 9 controls.
Only 15 patients had complete data for 'autonomic, cytokine and HPA Axis data' - and none of the controls did.
Given that the hypothesis seemed to be about systems-wide assessments, the lack of control data seems to be a problem. Only the autonomic variables could be compared between patients and healthy controls.
AUTONOMIC - mesurements while supine, Valsalva manoeuvre, standing - 10 am
Continuous heart rate
Blood pressure (beat to beat)
Impedence cardiography (cardiac output, stroke volume, ejection fraction and end-diastolic volume)
Day 1 - 1 sample
(dexamethasone at 11 pm)
Day 2 - 5 samples
SERUM INFLAMMATORY MARKERS - 10 am
Day 1 - 1 sample
SYMPTOM ASSESSMENT MEASURES
Cognitive failures questionnaire
Fatigue Impact Scale - questionnaire
RESULTS - AUTONOMIC FUNCTION IN PATIENTS AND HEALTHY CONTROLS
Here's the node diagram:
The analysis technique used is a mutual information algorithm in ARACNE.
I haven't read all the explanation of the analysis technique and even if I had, I doubt I would have fully understood it. But I'm assuming that the links indicate that if you know one variable e.g SBPa (Systolic Blood Pressure active stand) then that means you know something about the variables that are linked to it (I'm thinking of the strength of the link as being like the R squared value for a regression). So, in the case of the CFS network, when you know SBPa, you know something about heart rate, heart rate variability, and systolic blood pressure during the Valsalva manoeuvre. But, in the case of the healthy control network, when you know SBPa, you don't know anything about any other variable.
One of the strongest links in the healthy controls was between diastolic blood pressure during active stand and heart rate variability; whereas for the CFS patients, heart rate variability was linked to systolic blood pressure during active stand. Maybe there's something useful there? Or not.
Given the big difference in the numbers in each group (42 to 9), I'm not sure there's much basis for drawing conclusions. The different sample sizes resulted in different thresholds for each group. And although the healthy control's network appears more connected, there's a lot of weak links between those nodes that presumably could drop out if you changed the threshold.
I think this network approach has potential. But I think the small number of controls really affects the utility of this particular study.
(If someone understands the technique used, please explain.)
RESULTS - 'THE RELATIONSHIP BETWEEN DIFFERENT REGULATORY SYSTEMS IN CFS'
I've put the heading of this section of the results in quotes, as I think it's a very brave call to be talking about regulatory systems on the basis of this analysis.
Here's the network - remember, there is only one for the CFS patients. (There is no data for anything other than autonomic variables in the healthy controls. It seems that something went wrong with this study.)
Some of the links make sense - e.g EF (the ejection fraction), the efficiency of the heart, is related to CFQ, the Cognitive Failures Questionnaire (although for all we know, a low ejection fraction might be related to good cognitive performance, rather than bad cognitive performance). The Null, Dex 10, Dex 100 and LPS are in vitro glucocorticoid receptor response to 0, 10 and 100% dexamethasone solutions and lipopolysaccharide respectively.
So, moving on to the discussion to see what the authors made of all this.
DISCUSSION - AUTONOMIC COMPARISON BETWEEN CFS/ME AND CONTROLS
The authors suggest that the network built from the data of 9 controls is
In contrast, the autonomic network looks quite different for the CFS patients
That all makes sense given how common orthostatic problems are in people with ME/CFS.
The dissociation between systolic and diastolic blood pressure might reflect a narrowing of pulse pressure.
So, they are calling for larger studies in order to target therapeutic interventions. Probably fair enough although I don't know if they need to be network studies. It's a shame they don't report the actual measurements made - are some variables different to healthy controls? Maybe these were reported in other papers from the same study.
Tagging @mariovitali - may be of interest to him
DISCUSSION - The relationship between the ANS, immune system and HPA axis in CFS/ME
And there's the obligatory, 'more studies are needed':
The cytokines that survived the process were IL1b, IFNg, IL12, IL17.
There's a paragraph there about TH1 and TH2 profiles and the paraventricular nucleus that needs unpicking; it's beyond me now.
Regarding the measure of cognitive function, they note its relationship to ejection fraction, but also note that ejection fraction hasn't been found to be abnormal in people with ME/CFS.
So they are suggesting that a number of the cytokines are acting on autonomic function.
Limitations - they acknowledge the problems created by the small sample sizes and lack of a healthy combined network for comparison.
My conclusion. The study has obvious limitations in sample size and controls, and the selection of thresholds of significance with different cohort sizes seems a bit murky. I'd like to see actual data alongside the networks (Edit - some of the data is in supplementary tables). But there seems to be good intent, there's minimal BPS-type speculation and further refined, larger studies of this type might be useful.
I'll take Trish's advice about how, if your posts are more than 50% of the last 10 posts in a thread, it's time to take a breather.
Looking at the node diagram it is the lack of interconnectedness for CFS that is striking. To me the linear nature perhaps represents a lack of feedback adjustment potential ?
Note - i m not overly familiar with these so the comment is based on an assumption of how you would read this
Pity. The introduction is a mix of blather trawled from just about anywhere. It is a pity that they use our paper, which says we do not have any clear pathology yet, to argue that there is good evidence for more pathology. Julia Newton really ought to have either vetted this or take her name off.
Just to give a hypothesis, always keeping in mind that i do not know if this makes sense medically : During viral infection the body needs to boost thermogenesis to generate fever. Fever is necessary because viruses do not replicate as easily in higher temperatures. Boosting thermogenesis is achieved by boosting metabolism.It may be the case that because of this metabolic boost we are not in a hypometabolic state for some time, hence we feel better (Note : i also experienced this when i had symptoms).
Regarding the paper, i agree with @Hutan there seem to be some issues with the approach but i must say that i am happy that researchers are looking at ME through different perspectives and types of analysis. In doing so however we need quality data (GIGO : Garbage-In, Garbage-out), large enough sample sizes and algorithms suited for our purpose.
Network Analysis is a great tool to identify important areas of further research. We need to move from simply recognising the different "faces" of ME/CFS to hypothesise on the cause. Network Analysis may have some answers.
? I tend to wake early!
Separate names with a comma.