Exercise, PEM or PENE, and general medical research

[anciendaze]

I'm still looking for references which would show either that a drop in anaerobic threshold lasting 24 hours or more after exercise was either known or not known in prior medical literature of any condition. (I already know about overtraining syndrome in athletes.) If this shows up in any pathological situation we have good reason to call for research into this aspect of the pathology of ME/CFS. If it does not, we have reason to ask why medical testing has just now gotten around to asking if assumptions about the benefits of exercise in all contexts are true.

For those wedded to biopsychosocial models, I have to ask how patients whose "perceptions" of the threshold at which exercise becomes destructive are screwed up can tell they are approaching these limits prior to waking up in the emergency department -- an experience I and some others here have had. Telling us we "have to push ourselves", but not, of course, "too hard" is not useful.

If I wanted generic medical advice, I could listen to Dr. Oz more cheaply. We have been paying for personalized expert advice, unless disability and poverty have forced society to pay our doctors. By and large, this is not what either we or society have been getting for the money.

We need research criteria for such thresholds, and we need proxies for these measurements patients can use in everyday life. We also need recognition that normal maintenance activities to keep food, drink and/or medication in us, and clean clothes on us, or to arrive at therapy sessions with some idea of where we are and what we are doing, consume energy from a fixed budget which many researchers are at some pains to avoid measuring. (Consider the way PACE dropped actigraphy.)

For progress to take place, we need someone to actually investigate patient reports of exercise intolerance, orthostatic intolerance, dysautonomia, weakness, dizziness, etc. At that point we can began to do something about more subtle problems like episodic cognitive impairment ("brain fog") and/or emotional lability. Current practice has the cart before the horse.

 

[alex3619]

It might interest people to know that i searched for this information on repeat exercise some time ago and found nothing much. There is information about repeat testing, yes, but its about repeat testing that is not on consecutive days, at least in anything I found. This was typically cardiac research.

 

[Sean]

Telling us we "have to push ourselves", but not, of course, "too hard" is not useful.

Ahh, the 'Goldilocks' form of advice: Not too much, not too little, just the right amount.

Nobel prize winning material right there.

 

[RustyJ]

Incidentally, I am perplexed by the quoted comment about the control arm of these studies. If significantly reduced anaerobic threshold 24 hours following exercise is common among seriously deconditioned, but otherwise healthy, people where is the documentation for this claim? This should not depend on two particular groups of researchers, it should already be part of normative physiological data. Neither should it come as a surprise, as this did.

Snell has done some groundwork on this. He used healthy but sedentary controls in this CPET study. Although a subjective test, there is correlation between this study and his subsequent objective studies for the me/cfs patient responses, if not the sedentary controls.

J Womens Health (Larchmt). 2010 Feb;19(2):239-44. doi: 10.1089/jwh.2009.1507.

Postexertional malaise in women with chronic fatigue syndrome.
VanNess JM1, Stevens SR, Bateman L, Stiles TL, Snell CR.

Abstract
OBJECTIVE:
Postexertional malaise (PEM) is a defining characteristic of chronic fatigue syndrome (CFS) that remains a source of some controversy. The purpose of this study was to explore the effects of an exercise challenge on CFS symptoms from a patient perspective.

METHODS:
This study included 25 female CFS patients and 23 age-matched sedentary controls. All participants underwent a maximal cardiopulmonary exercise test. Subjects completed a health and well-being survey (SF-36) 7 days postexercise. Subjects also provided, approximately 7 days after testing, written answers to open-ended questions pertaining to physical and cognitive responses to the test and length of recovery. SF-36 data were compared using multivariate analyses. Written questionnaire responses were used to determine recovery time as well as number and type of symptoms experienced.

RESULTS:
Written questionnaires revealed that within 24 hours of the test, 85% of controls indicated full recovery, in contrast to 0 CFS patients. The remaining 15% of controls recovered within 48 hours of the test. In contrast, only 1 CFS patient recovered within 48 hours. Symptoms reported after the exercise test included fatigue, light-headedness, muscular/joint pain, cognitive dysfunction, headache, nausea, physical weakness, trembling/instability, insomnia, and sore throat/glands. A significant multivariate effect for the SF-36 responses (p < 0.001) indicated lower functioning among the CFS patients, which was most pronounced for items measuring physiological function.

CONCLUSIONS:
The results of this study suggest that PEM is both a real and an incapacitating condition for women with CFS and that their responses to exercise are distinctively different from those of sedentary controls.

 

[Marco]

I don't feel its necessary to go beyond the brain to explain PEM as post concussion syndrome (PCS) patients suffer very similar symptoms to ME/CFS even down to 'alcohol intolerance'. Exercise intolerance in PCS is posited as resulting from the brain injury leading to an 'uncoupling' between the the autonomic and cardiovascular systems. At least with PCS we know pretty unambiguously what the initial trigger was.

I can't copy and paste from this unfortunately :

http://natajournals.org/doi/pdf/10.4085/1062-6050-48.5.02

The research doesn't however, as yet, look at delayed symptom exacerbation following exercise or repeat testing although longer term follow up is recommended and patients' personal experiences do seem to be somewhat at odds with the expecatations of occupational health/rehabilitation professionals :

http://neurotalk.psychcentral.com/thread158982.html

 

[Valentijn]

In the exercise study @Mark mentioned, it's not at all clear which alleles are associated with ability to improve VO2max performance. Basically they say that 47% of ability to improve is genetic, and that 11 of the SNPs they looked at capture 23% of the variance, so about half of the genetic aspect. 7 of those are tested by 23andMe.

For two of those, prevalence rates in 21 ME patients from this forum are quite a bit different from either the general population or the 23andMe data from 21 "controls". rs7386139 on DEPTOR has 2% prevalence in the ME patients, versus 21% in the general population and 10% in the controls. DEPTOR is involved in regulating mTORC1 and mTORC2, which are involved in cellular metabolism. rs7324557, which isn't on a gene but is nearest to MIPEP, has 38% prevalence of the minor allele in the ME patients, 30% in the general population, and 29% in the controls.

rs6481619 on SVIL shows very little variation between ME patients, the general population, and controls.

While ME patients have a much different minor allele prevalence from the general population on rs2457571 on SLC22A3 (higher), and rs11574 on ID3 (higher), and rs10497520 on TTN (lower), the controls have nearly identical values to the ME patients. Hence those differences might be partially due to ethnicity, but also might reflect different abilities to increase VO2max.

ME patients have a slightly higher prevalence of rs2792022 on BTAF1 (33%) than the general population (30%), but a much lower prevalence than the controls do (40%).

Here's the image of the results. SNPs which are underlined and red are missense mutations. Red boxes indicate results present in 1-2.5% of the population, and yellow boxes indicate results present in 5-10% of the population. Black boxes are "no calls" due to the newer 23andMe chip not testing for those SNPs. Green boxes highlight results containing minor alleles which have a prevalence of 10% or greater. Click on the image to zoom in.

This study might have found something useful, but due to withholding allele data, it's impossible to guess how it might apply to us.

 

[MeSci]

I can't copy and paste from this unfortunately :

http://natajournals.org/doi/pdf/10.4085/1062-6050-48.5.02

You can copy and paste from pdfs in Adobe Reader by clicking 'edit', selecting 'take snapshot' (or similar) and selecting the text, table, graphic, etc. you want - it copies it automatically when you release the mouse button and you can then paste it where you want (e.g. here). I just tried it with a bit of text in this pdf and it worked (see example which I have tried to remove but can't!). The quality/size of the image depends on the resolution in your window, e.g. if you enlarge it on your screen it comes out larger in the image so the legibility is better.

I found this out after a lot of trial and error recently!

HTH.

 

[Marco]

You can copy and paste from pdfs in Adobe Reader by clicking 'edit', selecting 'take snapshot' (or similar) and selecting the text, table, graphic, etc. you want - it copies it automatically when you release the mouse button and you can then paste it where you want (e.g. here). I just tried it with a bit of text in this pdf and it worked (see example which I have tried to remove but can't!). The quality/size of the image depends on the resolution in your window, e.g. if you enlarge it on your screen it comes out larger in the image so the legibility is better.

I found this out after a lot of trial and error recently!

HTH.

Many thanks for the tip but its not opening in Adobe for me and all I can cobble togther is this :

Symptoms from concussion typically resolve within 7 to 10 days however, up to 33% of individuals with concussionshave postconcussion syndrome (PCS), with up to 30% continuing to meet criteria for PCS 6 months post injury. Age and sex may play roles in the diagnosis of PCS because women but not girls report higher symptom scores than males on a commonly used PCS diagnostic, the Rivermead Post
Concussion Symptoms Questionnaire.

Diagnosis of PCS requires a history of traumatic brain injury plus three or more symptoms :

(headache; dizziness; fatigue; irritability; insomnia; difficulty in concentration or memory; intolerance of stress, emotion, or alcohol).

 

[MeSci]

Many thanks for the tip but its not opening in Adobe for me and all I can cobble togther is this :

Diagnosis of PCS requires a history of traumatic brain injury plus three or more symptoms :

Wonder if it overlaps with traumatic pituitary damage, which (unsurprisingly) affects the neuroendocrine system and HPA and seems hard to distinguish from ME. I often wonder whether all these different syndromes are actually the same or overlapping illnesses. I don't think 'conventional' medicine has very good ways of classifying illness.

 

[Marco]

Wonder if it overlaps with traumatic pituitary damage, which (unsurprisingly) affects the neuroendocrine system and HPA and seems hard to distinguish from ME. I often wonder whether all these different syndromes are actually the same or overlapping illnesses. I don't think 'conventional' medicine has very good ways of classifying illness.

There are a number of 'syndromes' which all appear to follow a similar pattern of onset and have very overlapping symptoms. Amongst these are ME/CFS, Gulf War Illness, Complex Regional Pain Syndrome and Post Concussion Syndrome. In all cases, following some 'trigger' a minority of those affected by the original trigger go on to develop a chronic multisystem/multisymptom illness that appear to be 'out of proportion' to the initial trigger.

In the case of Post Concussion Sydrome the trigger is easily identified but according to neurologists any acute brain trauma should have resolved within around 11 days. What remains has been proposed to be chronic immune activation in the brain or neuroinflammation that doesn't result in obvious tissue damage. A similar mechanism has been proposed for complex regional pain syndrome and of course we have had preliminary evidence of neuroinflammation in ME/CFS.

I would tend to agree that we may be looking at a shared mechanism (neuroinflammation) with different triggers all leading to the same place. The reason we end up with different diagnoses depends on the individual pattern of symptoms and the circumstances at onset.

Why only a minority go on to develop these syndromes is the key question. I would suggest that what may appear to be a 'sudden onset' may be the 'last straw' that uncovers an existing predisposition to neuroinflammation.

All speculation of course!

 

[anciendaze]

@Marco

Thanks for the link to Post Concussion Syndrome, this should parallel work on some kinds of stroke in which there are changes in perfusion which trigger damaging response.

Incidentally, I had no trouble copying the pdf to local storage, then opening it with a pdf reader, though I have not tried Adobe. I have been able to copy and paste, albeit with some formatting trouble:

Postconcussion syndrome is the nonresolution of symp-
toms after a concussion. Definitions of PCS given by the
American Psychiatric Association
9
and the World Health
Organization
10
rely heavily on self-reported symptoms as
the basis for diagnosis.

While you see no reason "to look beyond the brain" I would like to investigate on-going inflammation and endothelial dysfunction, which produce particularly dramatic effects in the brain, but affect the entire body. I think we can agree on one point: physical effects which produce loss of consciousness should be investigated prior to delicate questions about motivation and will power.

 

[Marco]

@Marco

Thanks for the link to Post Concussion Syndrome, this should parallel work on some kinds of stroke in which there are changes in perfusion which trigger damaging response.

Incidentally, I had no trouble copying the pdf to local storage, then opening it with a pdf reader, though I have not tried Adobe. I have been able to copy and paste, albeit with some formatting trouble:

Thanks. I'll try saving the document locally and then opening via Adobe. At the moment I've the link saved as a favourite and it prefers to open as you say in a generic .pdf reader.


While you see no reason "to look beyond the brain" I would like to investigate on-going inflammation and endothelial dysfunction, which produce particularly dramatic effects in the brain, but affect the entire body. I think we can agree on one point: physical effects which produce loss of consciousness should be investigated prior to delicate questions about motivation and will power.[/quote]



I've diverted the thread enough but suffice to say while I feel neuroinflammation can explain most if not all ME/CFS symptoms the 'predisposing factor' that causes only a minority to develop these syndromes could very well be systemic. One example reported recently is a SNP in a TNF-a modulating gene that appeared to be strongly associated with 'sickness behaviour' - fatigue, mood etc in primary biliary cirrhosis patients.

 

[Mark]

For two of those, prevalence rates in 21 ME patients from this forum are quite a bit different from either the general population or the 23andMe data from 21 "controls". rs7386139 on DEPTOR has 2% prevalence in the ME patients, versus 21% in the general population and 10% in the controls...

Thanks for following up on that @Valentijn; very disappointing that the full allele data isn't public.

I think the work you're doing on SNPs etc is really interesting, and that whole line of inquiry could be really useful in terms of hypothesis generation. However, I think it's important to get the stats of variations in their proper context, which means things like p values etc; I don't know all that stats stuff myself (beyond the basic concept of p values) but it would be well worth while I think for you to get well on top of that, maybe take an online course if you're not already on top of it.

That thought was prompted by something that confuses me in the above quote. How can anything have 2% prevalence in a sample of 21 patients? If only one of the 21 had it, that would be 5%. If two had it, that would be 10% (same as the controls). Maybe that's explained by some technical detail (eg rs7386139 on DEPTOR is made up of 2 different things so you're saying one patient had one of the two and so it's about 2.4%). But even so, the difference between 2% and 10% sounds really large, at first, and possibly significant, but then you realise that 10% of 21 is just two people and so there's really no statistical significance to that.

So anyway, basically what I'm thinking is: I love what you're trying to do, and I'm really grateful you've had a look at it for this study I mentioned (I was half-hoping you might), but now that you've done so I'm thinking that a little extra stats info on these percentages would go a long way towards making the figures you're calculating more useful.

 

[anciendaze]

@Mark

You've hit the kind of point that bothers me. It might be that 21 patients had 42 alleles, so if only one allele were present this would be about 2% incidence. It would still not be very significant. It might also be something entirely different. This is a problem with a great deal of statistical information quoted widely. It might actually be valid, but those reading this don't have the context or background necessary to understand it. It is also possible for numbers to be blunders or misprints. Either way the result becomes meaningless in the context where people use it to make decisions.

 

[Valentijn]

Thanks for following up on that @Valentijn; very disappointing that the full allele data isn't public.

It's a bit odd, because they say it's in the supplementary data. But if it is, they've left out a lot of data essential for decoding it.

I think the work you're doing on SNPs etc is really interesting, and that whole line of inquiry could be really useful in terms of hypothesis generation. However, I think it's important to get the stats of variations in their proper context, which means things like p values etc; I don't know all that stats stuff myself (beyond the basic concept of p values) but it would be well worth while I think for you to get well on top of that, maybe take an online course if you're not already on top of it.

I did take half of a free statistics course from Coursera, focused on exactly for that sort of thing, before getting floored by ME. We're currently working on getting a proper database set up for the 23andMe data, which is finally done more or less, and the next step is getting R talking to that database, so I can easily determine p values, generate scatterplot graphs, and maybe even determine how much each variable is contributing to things.

The problem is that the current PC is about 15 years old, and doesn't have the memory, CPU speed, hard drive size and speed, or stability needed to do anything, especially in a sensible time frame. We were hoping to replace it last week, but had to get a new refrigerator instead. I need to store antibiotics for a few days each week at 5 C or colder, and the old one couldn't get under 8 C at the coldest setting when closed overnight! So long story short, I'm still plugging away with Excel on my laptop for now, and doing anything fancy is off the table for a while. But at least the Mystery of the Spoiling Meat has been solved

That thought was prompted by something that confuses me in the above quote. How can anything have 2% prevalence in a sample of 21 patients?

Minor allele frequency (MAF) is based on how often the minor allele appears in a sample. Because there are 21 patients, they typically have 42 alleles for any SNP. Exceptions are for the Y chromosome, mitochondrial DNA, and men's X chromosome - in each case there's just one allele at most.

Sometimes studies will also look at the effect of genotype (the combination of the two alleles) in addition to the MAF, and provide calculations based on that, but MAF itself is pretty much always included as well, as that's the easiest way to spot a significant difference. In fact, if there's no significant MAF variation between controls and patients, yet genotypes show a significant variation, it usually looks like a suspicious use of over-analyzing data in desperation to generate some sort of correlation - and they often end up with absurd results like saying the heterozygous version is a bigger risk factor than either homozygous version of an SNP

So anyway, basically what I'm thinking is: I love what you're trying to do, and I'm really grateful you've had a look at it for this study I mentioned (I was half-hoping you might), but now that you've done so I'm thinking that a little extra stats info on these percentages would go a long way towards making the figures you're calculating more useful.

Yeah, me too ... that's the main goal of getting a database up and running Though I also want to re-take the statistics course which focuses on using R, now that my brain is more functional.

 

[MeSci]

If I wanted generic medical advice, I could listen to Dr. Oz more cheaply. We have been paying for personalized expert advice, unless disability and poverty have forced society to pay our doctors.

Agree with your sentiments, of course, but the stuff about how doctors are paid sounds US-specific, and we need doctors of all nations to get this.

In the UK, society pays for the vast majority of medical advice - it's the default situation via the NHS. Doctors in the UK are particularly in need of education on ME (and a lot else besides!).

 

[anciendaze]

Agree with your sentiments, of course, but the stuff about how doctors are paid sounds US-specific, and we need doctors of all nations to get this.

In the UK, society pays for the vast majority of medical advice - it's the default situation via the NHS. Doctors in the UK are particularly in need of education on ME (and a lot else besides!).

The ideal, even in the UK, is that doctors are providing expert personal advice tailored to the patient. You can find plenty of complaints about pay not matching this requirement. In reality many such interactions are simply replays of prerecorded material by an actor in a white coat with an MD. The system is not set up to reward thinking.

I recall a disagreement I had with a U.S. doctor over the case in which a transplant patient died of rabies. My complaint was that it took almost a year to figure this out. The medical defense was that this was so improbable as to make this delay reasonable. My objection that if we were only going by probability, when we have new information constantly available from an actual patient under treatment right in front of us, we could dispense with a large part of the medical establishment, was considered outrageous. Doctors in that case knew something was very wrong for many months, yet did not revisit initial assumptions until far too late. Fortunately, other transplant recipients survived long enough to be vaccinated for rabies.

Here's the sequence of reasoning I would expect: this patient has a serious infection => this patient has a viral infection => what serious viral infections are possible? It does not take a year to work through this and order tests. Unbiased random testing might have found the answer more quickly. Expert opinion actually delayed discovery, because most doctors are not aware that rabies can take a year to manifest itself.

I think the problem here is that most people do not seem to realize that use of probability involves the idea of missing information, and when available information changes, so does the most probable answer. In medical circles there is a strong tendency to conflate probability with peer group consensus. If you are wrong, but go with the herd, you will survive, even if the patient does not. This is far worse reasoning than even common textbook misconceptions about probability found among the general public, and I have no trouble pulling up examples from history in which it condemned many patients to death.

 

[MeSci]

My objection that if we were only going by probability, when we have new information constantly available from an actual patient under treatment right in front of us, we could dispense with a large part of the medical establishment, was considered outrageous. Doctors in that case knew something was very wrong for many months, yet did not revisit initial assumptions until far too late.

After a terrible experience with paramedics who came when I called 999 (our equivalent of the US 911) in desperation, having had no help from my GP, I lodged a complaint about them not listening to me when I insisted that my illness was not psychosomatic (among other things). I was seriously ill with hyponatraemia (I only discovered this 3 years later when admitted to hospital with the same thing).

Their reply was apologetic but predictably unsatisfactory. I did some searching for info about paramedic training and was shocked to find that they were taught to stick with their initial diagnosis and ignore further evidence!

I fervently hope that this aspect of their training has been changed since then.

 

[anciendaze]

More information concerning the case of rabies acquired via transplant. As I said, most MDs do not realize rabies infection in humans can take a year to produce symptoms. I checked an old (14th) edition of the Merck Manual:

In man, the incubation period varies from 10 days to > one yr.

The information is not exactly confidential or missing from common medical literature, just absent from the consciousness of the average doctor.

The deep question here is how such ignorance has any impact on measures of probability.

 

[anciendaze]

...I did some searching for info about paramedic training and was shocked to find that they were taught to stick with their initial diagnosis and ignore further evidence!

I fervently hope that this aspect of their training has been changed since then.

The problem of ignorance with paramedics is even worse than with full MDs, though arrogance is less. You have to wonder what they would have done if they had started trying things. I believe the reasoning here is that they have already made whatever mistakes they are going to make with the current diagnosis, and the best thing to do is hope the patient survives until better judgment can be exercised. You really don't want to even think about what poorly-trained medics will do when they get a wild idea in time-critical situations like combat emergency medicine.

There was a terrible joke in the book M.A.S.H. which illustrates how bad this can get. "We were afraid of what he would do if we tried to stop him" (from pissing on a heater while half asleep.) At least this funny example didn't kill anyone, and did not involve explosives or white phosphorous. (From another story, not in M.A.S.H., "Couldn't we at least put out his fires?")