Erythrocyte Deformability As a Potential Biomarker for Chronic Fatigue Syndrome, Davis et al (2018)

[Londinium]

Perfectly put. I am concerned about this claim: it would explain how they get such a small P value from the tiny sample, but that P value does not tell you how likely the finding is to hold up in the wider patient population. On the letter is what we need to know.

There is another issue. IIRC, statistical theory says that you need a sample of at least 20 patients per group to have confidence that the sample is representative of the wider population (which is why you see very few studies published with you with fewer than 20 individuals per group).

I was mulling this over and I think *theoretically* you could get the p values mentioned with an N=18 sample. I agree with the above that sampling hundreds of cells gives you more confidence of the deformability value for a given patient/control but does not provide greater confidence of the differences in average deformability *between* patients and controls. However, if for example all nine patients had deformability lower than all the controls with no overlap between the groups, the chance of getting that result if the null hypothesis was true would imply p = 7.6x10^-6 (based on the heavily oversimplified approach of treating each patient/control sampled as being a 50/50 coin toss between being in the 'high deformability' and 'low deformability' groups). So I could see how one could get a very low p-value if the results were massively clear-cut with no overlap. As you say, we need to see the results.

The wider question is your point around whether the sample is representative of the patient population: the severely ill study as benefits and drawbacks. The benefit is any 'underlying signal' might well be stronger in the severely ill than in mild/moderate ME/CFS patients; the drawback is that severely ill patients are likely to be much more deconditioned and on other medications meaning that any findings may be merely a result of that instead.

ETA: and the real biggie, for all ME/CFS research - we don't even know if severely ill patients have the same disease aetiology as other patients!

 

[chrisb]

This may well be a silly question, as I know nothing about the subject. It may be treated as rhetorical if necessary.

If, for the sake of argument, we accept that this is a valid finding and connected in some way with some of the symptoms of some ME patients, how could this fit into a relapsing /remitting condition? Presumably there is a constant production of new RBCs and destruction of old ones. Are the new cells produced already faulty or do they become faulty only when introduced to some other circulating cells? During remission does the proportion of normal cells to faulty cells increase, producing a lessening of symptoms?

 

[Trish]

Red blood cells are fascinating things. I've just been reminding myself about them from this article:
https://www.britannica.com/science/blood-biochemistry/Red-blood-cells-erythrocytes#ref62335
Here are a few key points:

Unlike most cells they don't have a nucleus or mitochondria. They are produced in bone marrow, contain mostly haemoglobin which transports oxygen from the lungs to the rest of the body and brings back carbon dioxide to the lungs. They need energy mainly for maintaining the balance of sodium and potassium within the cell against diffusion gradients and make the energy by glycolysis or the pentose pathway. Not having any DNA, they can't repair themselves and after about 120 days they disintegrate and the iron is carried back in the blood to the bone marrow for making more red blood cells. The biconcave shape of RBC's provides lots of surface for exchanging oxygen and CO2.

 

[Andy]

This may well be a silly question, as I know nothing about the subject. It may be treated as rhetorical if necessary.

If, for the sake of argument, we accept that this is a valid finding and connected in some way with some of the symptoms of some ME patients, how could this fit into a relapsing /remitting condition? Presumably there is a constant production of new RBCs and destruction of old ones. Are the new cells produced already faulty or do they become faulty only when introduced to some other circulating cells? During remission does the proportion of normal cells to faulty cells increase, producing a lessening of symptoms?

Or what they might be seeing here is the effect of the unidentified substance in the plasma(?) that causes cells from healthy controls to be affected once exposed to plasma from patients? Though how that fits relapsing/remitting itself I don't know.

 

[Trish]

The wider question is your point around whether the sample is representative of the patient population: the severely ill study as benefits and drawbacks. The benefit is any 'underlying signal' might well be stronger in the severely ill than in mild/moderate ME/CFS patients; the drawback is that severely ill patients are likely to be much more deconditioned and on other medications meaning that any findings may be merely a result of that instead.

It isn't clear from the abstract whether they used blood from the severely ill study. I suspect not, as they only collected samples for that study once so as not to cause too much disruption to the patients, and the samples would have to have been frozen some time ago.
I'm guessing they might have to use fresh blood for this study so the cells would not be deformed by the freezing process.

 

[Hutan]

If

we accept that this is a valid finding and connected in some way with some of the symptoms of some ME patients, how could this fit into a relapsing /remitting condition?

This paper, 'The effect of sepsis on the Erythrocyte' is really interesting, although there's a lot going on and most of it is well beyond me.

This was demonstrated in two related in vitro studies where human whole blood was challenged with Escherichia coli (E. coli) endotoxin. Todd et al. [88,89] reported that endotoxin increased RBC intracellular Ca2+ (discussed in greater detail in Section 5) and membrane viscosity, while decreasing RBC deformability (a marker of sepsis as discussed earlier). Furthermore, removal of leukocytes or treatment with oxygen free radical scavengers attenuated the deleterious effects of E. coli endotoxin on the RBC, linking RBC exposure to leukocyte ROS with RBC pathophysiology.

Additionally, Weiss [90,91] found activated human neutrophils were capable of releasing superoxide and damaging RBCs and that activated neutrophils in contact with RBCs, but independent of phagocytosis, were able to generate reactive oxygen species ...

In vitro experiments have found that oxygen free radicals alter proteolytic susceptibility and RBC rheology, by degrading the integral membrane protein band 3 (anion channel) and the cytoskeleton protein, spectrin, while decreasing RBC deformability

Basically, (I think) if there are toxins or even if the body just thinks it's fighting something, there's going to be a lot of oxidation going on that makes the red blood cells less deformable . And it seems that those changes to the red blood cells are permanent for the individual cell - so damaged red blood cells are going to have a harder time moving around the body. So, while you have a lot of damaged red blood cells, you probably aren't feeling great.

But it looks like there are lots of ways that things related to red blood cells could improve quickly once the threat or perceived threat has gone - it doesn't have to be weeks until the damaged red blood cells are replaced. Here's just one mechanism:

A hallmark of sepsis in both experimental models [14,15,114] and clinical patients [16,22] is a high degree of capillary stopped flow (Figure 1). While the in vivo mechanism of capillary stopped flow has not been completely elaborated, in vivo and in vitro evidence suggests that several factors may be involved: (1) RBC adhesion to endothelial cells, mediated by LPS and intermittent flow conditions [115], increased RBC intracellular Ca2+ and redistribution of phosphatidylserine (PS) to the outer RBC membrane leaflet [116], interaction with endothelial chemokine ligand [117] and matrix thrombospondin [118]; (2) decreased RBC deformability [3]; (3) increased platelet adhesion [18]; and (4) increased coagulation [28]. While sepsis induced decreased RBC deformability may be a factor in capillary stopped-flow [3], the septic spleen will likely trap the most rigid RBCs, as its smaller and shorter interendothelial slit dimensions [38,39] makes it more sensitive to RBCs with abnormal morphology and decreased deformability than skeletal muscle capillaries, which have a diameter of approximately 5–6 µm.

Correction of the calcium ion issues and repletion of antioxidant levels - and therefore less capillary stopped flow and therefore less areas of hypoxic tissue sound like things that could potentially get back to normal in a timescale that fits with PEM.

There's lots of potential in this reduced deformability of blood cells idea I think.

 

[Mij]

Nanotracker

This device was in the news recently too.

 

[alex3619]

IIRC, statistical theory says that you need a sample of at least 20 patients per group to have confidence that the sample is representative of the wider population

The problem is these estimates are arbitrary, and as you say a measure of likely confidence. However its either right or wrong, statistics usually cannot determine that, no matter the group size. Statistics can only say, at best, that presuming there are no confounds or undetermined biases, and presuming there is a random distribution, an outcome with a very tiny p value is something we can have very great confidence in. Never certainty.

This is a lot like the closed world assumption in computer science. You know that additional factors, not taken into account, can create bugs or errors, but you have to presume that until some other factor is shown to be important enough then the software will do. Other things are outside the specification of the software. Now studies, by their methodology etc., have a defined scope. There can always be confounds, or at least we should presume so.

One thing I like to keep considering is that a statistically significant bias is an estimate of the likelihood a result is due to chance. If its due to high bias or powerful confounds then its not due to chance. So it can be statistically significant, but still wrong.

 

[Londinium]

It isn't clear from the abstract whether they used blood from the severely ill study. I suspect not, as they only collected samples for that study once so as not to cause too much disruption to the patients, and the samples would have to have been frozen some time ago.
I'm guessing they might have to use fresh blood for this study so the cells would not be deformed by the freezing process.

Good point, I had assumed it was from that study due to the number of participants, but that's quite possibly wrong.

 

[Jim001]

This may well be a silly question, as I know nothing about the subject. It may be treated as rhetorical if necessary.

If, for the sake of argument, we accept that this is a valid finding and connected in some way with some of the symptoms of some ME patients, how could this fit into a relapsing /remitting condition? Presumably there is a constant production of new RBCs and destruction of old ones. Are the new cells produced already faulty or do they become faulty only when introduced to some other circulating cells? During remission does the proportion of normal cells to faulty cells increase, producing a lessening of symptoms?

I know very little about the molecular biology of such things, but I have wondered if the poor RBC deformability could be the product of Naviaux's Cell Danger Response. IIRC, part of the CDR involves a stiffening of the cell membrane. If this is the case, and it could be behind the RBC low deformability, then maybe it is something that is variable with the engagement of the CDR, and therefore could normalize suddenly during remissions.

ETA: Trish's comment reminded me that RBC's dont have mitochondria. I'm not sure if this would preclude them from the affects of CDR in other cells.

 

[Kalliope]

#MEAction: Study shows red blood cells less deformable in ME patients - possible biomarker
Author: Ron Davis

This paper documents that red blood cells are less deformable in ME/CFS patients compared to healthy controls. It potentially could be a biomarker, and we are proceeding to design new devices that will make a clear distinction between patients and healthy controls. These devices will be hand-held and easy to use by doctors in their offices, or in clinical testing labs.

Past work has looked primarily at the shape of red blood cells, which is difficult to quantitate. Our approach will give a clear quantitative number. It measures the ability of red blood cells to deform while squeezing into a capillary, something that blood cells must do for healthy flow. We measure hundreds of cells from each patient, so, because of this, even though the number of patients is low, we get a very statistically significant distinction between patient and healthy cells’ deformability. We are putting our energy into developing the new devices as soon as possible.

 

[Hutan]

From the MEAction article which reproduces a press release by Ron Davis:

Original Post on Blood Journal here.
This research has been accepted for publication in Clinical Hemorheology and Microcirculation.

I wish researchers spruiking their work would mention the sample size.

This paper documents that red blood cells are less deformable in ME/CFS patients compared to healthy controls.

This statement sounds a lot more conclusive than it really is. And Ron really should get that statement about the large number of cells in each patient sample counteracting the small sample size to produce a very statistically significant distinction deleted. It decreases my confidence that the stats have been done right and in Ron's knowledge of stats. I expect it was just a momentary slip and whoever did the p calculation got it right, but that statement creates doubt.

 

[Andy]

From the MEAction article which reproduces a press release by Ron Davis:


I wish researchers spruiking their work would mention the sample size.

This statement sounds a lot more conclusive than it really is. And Ron really should get that statement about the large number of cells in each patient sample counteracting the small sample size to produce a very statistically significant distinction deleted. It decreases my confidence that the stats have been done right and in Ron's knowledge of stats. I expect it was just a momentary slip and whoever did the p calculation got it right, but that statement creates doubt.

I agree, especially with what you say about the sample size, which is, if I recall correctly, 9 patients and 9 controls. Really, all they can be certain of is that they looked at lots of cells from 9 patients, and that they can be pretty certain of the results, in regard to those 9 patients - how many times have we criticised various papers here, both psych and biomedical, for having small sample sizes, this should be no different. For Ron to make such definite statements he really needs to be sampling far far more people.

 

[Forbin]

FWIW, it seems that one way RBCs can become more rigid is when small molecules (ligands) bind to a protein in the cell membrane called glycophorin A (GPA), which, in turn, interacts with other proteins in the cell membrane, reducing deformability.

When tracer antibodies to human glycophorin A attach to blood cell GPA the "ligated RBCs travel 33% slower than control RBCs".

Ligation of Glycophorin A Generates Reactive Oxygen Species Leading to Decreased Red Blood Cell Function (2016)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4718526/

Our results obtained in vitro using anti-GPA antibodies were validated using complement-altered RBCs isolated from control and septic patients. Our results suggest that during inflammatory conditions, circulating RBCs significantly contribute to capillary flow dysfunctions, and constitute an important but overlooked source of intravascular ROS and ATP, both critical mediators responsible for endothelial cell activation, microcirculation impairment, platelet activation, as well as long-term dysregulated adaptive and innate immune responses.

Our results using RBCs isolated from SLE [systemic lupus erythematosus] [4] and trauma patients demonstrate that binding nascent complement fragments C3b and C4b to human RBC plasma membranes has significant negative effects on RBC function, namely β-spectrin phosphorylation and decreased membrane deformability [5,6].

 

[OverTheHills]

This may be a very dumb question but could there be a link with livedo reticularis?

Wikipedia says:
Livedo reticularis is a common skin finding consisting of a mottled reticulated vascular pattern that appears as a lace-like purplish discoloration of the skin.[1] The discoloration is caused by swelling of the venules owing to obstruction of capillaries by small blood clots. The blood clots in the small blood vessels can be a secondary effect of a condition that increases a person's risk of forming blood clots, including a wide array of pathological and nonpathological conditions

I have had this since I was a teenager, and of course doctors checked a few things out and labelled it as idiopathic.

 

[inox]

I must admit I am confused about where it is exactly published and what kind of publication it is. But aside from that it is rather pleasing I must say to see the team publish.

Lots of confusing statements about that online, but I think Janet Dafoe (Ron Davis' wife) can be trusted to have the correct information

 

[Marco]

This may be a very dumb question but could there be a link with livedo reticularis?

I have had this since I was a teenager, and of course doctors checked a few things out and labelled it as idiopathic.

This is something I've developed over the last 10 years.

 

[chrisb]

This may be a very dumb question but could there be a link with livedo reticularis?

And @Marco . This is interesting. I thought I was the only one with this. The only factor I found was that it is a possible side effect of Amantadine, which may, or may not have contributed to the severity of my condition. It is mainly apparent on the backs of my hands at certain times. It had to be self diagnosed. I have mentioned it to a GP but it was not thought significant.

 

[Marco]

And @Marco . This is interesting. I thought I was the only one with this. The only factor I found was that it is a possible side effect of Amantadine, which may, or may not have contributed to the severity of my condition. It is mainly apparent on the backs of my hands at certain times. It had to be self diagnosed. I have mentioned it to a GP but it was not thought significant.

Hi Chris

The only prescription drug I've been on for any extended period of time has been Tagamet and that was decades ago although I was a heavy smoker until recently which probably didn't help.

It seems to restricted to my arms and upper body but I've a number of other issues which might plausibly be linked to small blood vessels (occasional parathesias; heat intolerance; thinning hair and rapidly deteriorating eyesight).

 

[OverTheHills]

@Marco, @chrisb
Mine is on my legs. In my case it isn't a drug reaction - I wasn't taking any medications when this developed as a teenager. Coincidentally teenage was when my allergy issues started. I have been +'ve ANA for years now, although I don't test positive for antiphospholipid etc so its idiopathic again.

Perhaps I could tell myself a handwavy story about unclassified autoimmune disease which began as a teenager and an increasingly dysregulated immune system producing an ever increasing number of issues such Livedo, ME, migraines, what looks like MCAS. And wait for Prof Edwards to pour cold water on it.

It seems Livedo is not a widespread issue for pwME though, would we expect that if this paper was on the track of a biomarker?