Naviaux et. al.: Metabolic features of chronic fatigue syndrome

[wdb]

I am uncertain abut this study. If it has identified a particular metabolic profile shift for ME/CFS then it may well be groundbreaking. However, despite the claim that they found a consistent signature they also say that most of the out of line findings were individual or idiosyncratic. I cannot quite follow that. What also worrisome is that if there is a consistent pattern you do not need multiple regression analysis - you just look at the scatter plots.

It's a relief I'm not alone feeling a little bit of scepticism, n as low as 18 and measuring 612 metabolites has got to carry a risk of false positives. I'd be keen to see some Scatter plots too to easy my mind, no sign of the raw data yet but hopefully soon.

 

[BurnA]

You could do a post exercise analysis - obviously with potential problems but you might get useful information from rather inactive healthy people starting an exercise programme. Otherwise other control groups with low activity would be an obvious thing to look at.

Would an analysis of mild patients be also useful, any impact from inactivity would be greatly lessened compared to a more severe patient ?

If the result from mild and severe patient showed a consistency I presume this would be telling ?

 

[JamBob]

Given the importance of age matching for these metabolomic studies, I'm wondering if it is even possible to find young healthy controls who are completely inactive. I was thinking of people with agoraphobia (who don't leave the house) but they might still be really active around their houses doing housework.

Is it even possible to find a non-diseased young person who stays in bed 24/7 for years on end?

 

[trishrhymes]

Given the importance of age matching for these metabolomic studies, I'm wondering if it is even possible to find young healthy controls who are completely inactive. I was thinking of people with agoraphobia (who don't leave the house) but they might still be really active around their houses doing housework.

Is it even possible to find a non-diseased young person who stays in bed 24/7 for years on end?

The most obvious suggestion is quadriplegia and other forms of paralysis, though I'm not sure I'd want to expect people who already have so much to cope with to take part in a study for others' benefit.

I think the space agency have done some studies getting healthy people to lie in bed all day for several weeks to see what happens to them physiologically to help plan for long space travel. Maybe if they are still ongoing they could provide samples.

 

[Nielk]

What about the opposite? Using other chronically ill patients suffering from other diseases that force them to be inactive - like those suffering from congestive heart failure? If their metabolomic signatures mimic those found in the study using ME patients, then we know it is not a unique signature for ME.

 

[Cheesus]

Would an analysis of mild patients be also useful, any impact from inactivity would be greatly lessened compared to a more severe patient ?

If the result from mild and severe patient showed a consistency I presume this would be telling ?

I'm not sure that would indicate a direction of causality. A milder hypometabolic state could be a result of a higher degree of inactivity, but it could also be the cause of a less severe form of ME.

 

[Cheesus]

What about the opposite? Using other chronically ill patients suffering from other diseases that force them to be inactive - like those suffering from congestive heart failure? If their metabolomic signatures mimic those found in the study using ME patients, then we know it is not a unique signature for ME.

That's a strong idea. You could match ME patients and a control group for activity levels and duration of disability then see where we stand.

 

[Tuha]

Is not elevated lactic acid a good sign that there is something wrong with our metabolisme. I am mild case and I am walking every day but lactic acid is still elevated and my muscles are tired all the time. In my case it cannot be a question of decondition.

 

[Sidereal]

It would be good to match patients to healthy sedentary controls or housebound patients with other illnesses like lupus or agoraphobia for something like average daily number of steps in future studies to control for activity levels to help clarify the chicken-or-the-egg issue.

But I think it's a common misconception that ME/CFS patients are as a rule extremely inactive compared to the general population. The more severe cases of course are very inactive by necessity and they will tend to be overrepresented on forums such as this one but there are many who are mildly affected, some are still working, walking / going about their daily lives and still ill and most certainly not deconditioned.

According to the NHS, the average person in Britain only takes 3000-4000 steps a day. I consider myself on the mild end of moderately affected and I hit that range on most days when I am not crashed with PEM. During good spells I can do 7k-10k per day. And yet I still feel sick all the time due to things that tend to be attributed to "inactivity" by genius doctors like brain fog, rapid fatiguability on exertion, muscle pain, muscle weakness, dizziness, POTS etc. These symptoms prevent you from meaningfully engaging with life or holding down a real job.

I know many people whose only daily physical activity consists of walking from the car to the office and picking up the remote control for the TV and they don't have any of the symptoms I struggle with so I'm gonna hazard a wild guess here that these metabolic abnormalities are more likely to the the cause of our inability to make sufficient energy than the consequence.

 

[A.B.]

Would serum creatinine be an acceptable way to match controls to patients? As far as I know, its levels depend on muscle activity (and certain diseases which can be excluded via other tests).

PS: I strongly agree with Sideral. The deconditioning narrative is being inappropriately applied to all patients when only a small minority will be so severely affected as to be truly deconditioned. And chances are these people are not taking part in studies! In the study, Karnofsky scores were a little above 60 for males, a little above 50 for females. These were not patients that are bed bound.

 

[Lolinda]

I think it's a common misconception that ME/CFS patients are as a rule extremely inactive compared to the general population.

Adding to this: The strongest finding of the Naviaux study is the reduction in sphingolipids in CFS. Here the Naviaux study concludes that "The low sphingolipid profile in CFS ... is opposite to the increased sphingolipids observed in metabolic syndrome". Now, if CFS people had these markers because of inactivity, then what should we say about the typically fat and sedentary people with metabolic syndrome?

(just a first thought that came to my mind. Take it as that. This is no scientific argument whatsoever. For that one needed to look at the individual sphingolipids measured by Naviaux and measured by others in people with metabolic syndrome. And one had to find studies in which long term training / inactivity effects are examined on the same variables)

 

[ukxmrv]

I'd be interested to see PWME compared to a group who are TATT (who don't meet any of the ME or CFS criteria)

 

[JaimeS]

According to the NHS, the average person in Britain only takes 3000-4000 steps a day. I consider myself on the mild end of moderately affected and I hit that range on most days when I am not crashed with PEM. During good spells I can do 7k-10k per day. And yet I still feel sick all the time due to things that tend to be attributed to "inactivity" by genius doctors

Ditto @Sidereal -- I usually walk 3000-4000 steps per day, and on days when I'm more active I go up to 5000 or so. If I hit 10,000, chances are, I'll be sick, but it's good to know I walk as much as 'average'.

-J

 

[Mij]

I'm wondering if there will be differences between M.E. pain patients and non pain M.E patients.

 

[Gijs]

I am uncertain abut this study. If it has identified a particular metabolic profile shift for ME/CFS then it may well be groundbreaking. However, despite the claim that they found a consistent signature they also say that most of the out of line findings were individual or idiosyncratic. I cannot quite follow that. What also worrisome is that if there is a consistent pattern you do not need multiple regression analysis - you just look at the scatter plots. In the past disease mechanisms have not been found this way. You may need to look at three variables, like calcium,phosphate and alkaline phosphatase to identify hyperparathyroidism, but three is about the most and there are obvious reasons for the triad. I am inherently cautious about big data fishing trips. The other issue is whether this is just a reflection of inactivity - as several people have said.

There are various aspects of context like confident statements about ME/CFS being treatable with supplements that look out of line for a rigorous scientific approach when so far we have no hint of mechanism for the metabolic profile.

Autoimmune disease can set in motion just about any mechanism or sequence of events you like because the autoantibody can bind to any of 40,000 proteins, each responsible for a different function. So yes, an autoimmune process could lead to a shifted metabolic state just as in Addison's disease or Hashimoto's disease.

So Rituximab could be working by removing B cells that generate antibodies that drive the hypo metabolic state although these would seem not to be alarm signals in the usual sense. But the proposal about rituximab working via lipid rafts is a misunderstanding of the biology. The killing mechanism is quite irrelevant to the effect of having no B cells, whether in autoimmunity or cancer. It has nothing to do with any relation of ME to cancer if there is one. I think that can be disregarded.

I don't think you can say that Naviaux findings are all due to inactivity, it sounds to simplistic. If we take sphingolipids for example; we see in metabolic syndrome (inactive) patiënts that sphingolipids are usually high, not low like in ME/CFS patiënts. What worries me is that Hanson didn't find low sphingolipids in her patiënts. If i take this into account i can't believe that Naviaux has found a metabolic profile for CFS.

 

[CCC]

I'm wondering if there will be differences between M.E. pain patients and non pain M.E patients.

That's a very good question. I would expect so.

 

[Sidereal]

What worries me is that Hanson didn't find low sphingolipids in her patiënts.

That is worrying but I think we should wait for the published paper to see what Hanson has measured and what results she got before we declare doom and gloom. If I recall correctly, her sample size was tiny.

 

[Tuha]

Ditto @Sidereal -- I usually walk 3000-4000 steps per day, and on days when I'm more active I go up to 5000 or so. If I hit 10,000, chances are, I'll be sick, but it's good to know I walk as much as 'average'.

-J

For me it´s similar. I can walk up to 10 000 if I dont have a bad day. It also depends on if I relax during the day and I dont sit too much behind the computer. In generally I prefere to walk than to sit behind the computer. SO in my case I cannot be deconditioned.

 

[user9876]

I am uncertain abut this study. If it has identified a particular metabolic profile shift for ME/CFS then it may well be groundbreaking. However, despite the claim that they found a consistent signature they also say that most of the out of line findings were individual or idiosyncratic. I cannot quite follow that. What also worrisome is that if there is a consistent pattern you do not need multiple regression analysis - you just look at the scatter plots. In the past disease mechanisms have not been found this way. You may need to look at three variables, like calcium,phosphate and alkaline phosphatase to identify hyperparathyroidism, but three is about the most and there are obvious reasons for the triad. I am inherently cautious about big data fishing trips. The other issue is whether this is just a reflection of inactivity - as several people have said.

There are various aspects of context like confident statements about ME/CFS being treatable with supplements that look out of line for a rigorous scientific approach when so far we have no hint of mechanism for the metabolic profile.

I've wondered about medicine looking at single variables in the past. I've done quite a bit of work on classification where you need to take multiple variables to get a decent signal. Sometimes it is because of noise. But I've been looking a bit at anomaly detection recently and I think multivariate data here is good because you have one or multiple normal clusters from a multivariate vector then anomalies may deviate from these in many ways.

I've not read the paper yet but it did occur to be what they are showing is the disruption of certain dynamic processes with ME. Given Chaos theory should we expect a consistent disruption from the normal equilibrium. I was wondering if their results suggested certain processes are pushed in a given direction but each person is different.

The big thing to me then is if they can characterise normal states as well as states (abnormal or normal) seen in other diseases.

But the big question seems to me to be what causes the disruption and hence how to change it. Could supplements really change things back. I could see they could perhaps help shift things in the right direction. But is this a new equlibra that could be moved back (I don't think I believe this) or is it happening because something is signalling to the body to act differently. I was assuming that would be something like an auto-antibody. In which case to stop the response we need to stop the signal hence Rituximab?

 

[Jonathan Edwards]

I don't think you can say that Naviaux findings are all due to inactivity, it sounds to simplistic. If we take sphingolipids for example; we see in metabolic syndrome (inactive) patiënts that sphingolipids are usually high, not low like in ME/CFS patiënts.

It is important to exclude simple explanations before moving on to more speculative ones. I would guess that sphingolipids in plasma derive from damaged cell membranes. I doubt they are synthesised and exported to be used in other cells. People with metabolic syndrome may be relatively inactive but they have lots of reasons for damaging cells. Many are overweight and so will have more daily muscle damage simply from carrying that weight around. Many have high glucose levels, which damages cells. Hypertension also damages cells. And of course they have abnormal lipid handling anyway. So there are lots of reasons for them not having low sphingolipid even if inactivity lowers sphingolipid if all other factors are equal. (Much the same applies to femoral osteopenia, which occurs with inactivity but not in obese people who load the femur at high rates.)

I've wondered about medicine looking at single variables in the past. I've done quite a bit of work on classification where you need to take multiple variables to get a decent signal. Sometimes it is because of noise. But I've been looking a bit at anomaly detection recently and I think multivariate data here is good because you have one or multiple normal clusters from a multivariate vector then anomalies may deviate from these in many ways.

You may be right. But there is something that does not quite fit for me here. If this is a question of a heterogeneous mix of overlapping pathway shifts then it is extremely unlikely that any replicable formula based on MVA will identify 94-96% of cases correctly. As you know you can always produce a formula that will identify a group this precisely within a cohort but if things are very heterogeneous then almost by definition this is going to be a non-replicable formula.

In conditions like RA we have never got near a formula that identifies 95% of cases correctly. In at least 10% of cases all physicians will say that are not really sure if it is RA anyway.

I would be much more impressed if a formula of three factors picked out 60% of ME/CFS patients and 10% of controls, for instance. The trouble with the current trend for presenting statistically pre-digested data is that we never see the original scatter plots.

Of course hidden behind the 96% score may be exactly such a 60/10 split for some major variables and it may pan out as something of key importance. Maybe we need a friendly FOI request for the raw data!

More to the point a rapid re-run with another blinded controlled cohort like the UK Biobank could turn this from a maybe into a racing certainty.