Open Cervicocranial dysfunction, neuroinflammation and infection in ME/CFS compared to healthy subjects, Bragée & Bertilson [MEPRO study]

I will try to summarise a few points on anatomy and physiology, as I understand them.

1. CSF is produced by choroid plexus within fluid spaces called ventricles deep inside the brain and flows out of the brain at the fourth ventricle into the subdural space between the brain and skull. The CSF around the spinal cord is a dead end and not involved in any directional flow.

2. Therefore, cervical spine stenosis has no effect on CSF flow in the skull so does not affect intracranial pressure. It is irrelevant to any theory of raised intracranial pressure.

3. The same almost certainly applies to craniocervical instability since it does not interfere with circulation of CSF. Unless there is associated Chiari Malformation, in which case it is the Chiari that is the problem.

4. Chiari Malformation, if major, can lead to compression of the cerebellum on to the brainstem, which can block the outflow of CSF from the inner ventricles to the brain/skull space. This can lead to raised pressure within the brain but should not raise the pressure of CSF outside the brain. Such an effect is visible on an ordinary CT or MRI scan as an increase in size of ventricles relative to the brain/skull subdural CSF space. As far as I am aware nobody has said anything about such an appearance in any cases of ME/CFS.

5. Chiari Malformation comes in all degrees and over diagnosis of mild 'Chiari' is something that has been a concern for decades.

6. As far as I can establish the consensus view is that Chiari is not associated with EDS. The suggestion of an association comes from the Bolognese group in 2007. However, at that time the EDS criteria were even vaguer than they ae since 2017. Moreover, the report is based on a cohort of 'Chiari' cases attending a tertiary referral centre. It may be of note that they found that the combination of 'EDS' and Chiari was eight times more common in women. As far as I know EDS is not more common in women but 'hypermobility' as judged by Beighton score is much commoner in women because it does not have a different scale by gender and women have more mobile joints (elbows in particular). Everything points to the 'EDS' cases in this study being polygenic hypermobility and since the rate was only about 10%, which is normal, I doubt anything was found more than expected by chance. Which suggests that the consensus view is correct - there is no association with EDS.

7. The population based studies I have seen do not indicate any association between EDS and ME/CFS.

8. Idiopathic intracranial hypertension (IIH) is essentially unrelated to any of the above because the problem is failure to resorb CSF in the subdural brain/skull space. There is no blockage to flow. I do not know the detail of the imaging but I would not expect increase in ventricle size relative to subdural space. That may be why optic nerve oedema has been used to judge raised pressure in these cases.

9. Optic nerve oedema can occur with both an obstructive cause of high pressure and IIH. Note, however, that we cannot say that raised pressure may have been missed in CTD cases because the optic nerve is protected from oedema if the only evidence of raised pressure is optic nerve oedema. (And we have no reason to think CTD cases are relevant to ME.)

These are some of the reasons why I have said the anatomy and physiology do not add up.

One particular issue that worries me is that this group has chosen to send patients to a clinic in London that is highly controversial in terms of its interpretation of findings. There are fully competent neuroradiologists in Sweden, I am quite sure.

 

In the study plan document, they outline the only reason sending patients to London is due to lack of upright machine in Sweden (or anywhere closer for that matter). The London part is an extended study where only 10% of patients are included and according to the same document it would be performed in Stockholm if such a machine became available in time. Actually I think they can perform most of these measurements regardless in Sweden in the non-upright MRI. Anyway, I'd be curious to know why the London clinic is considered controversial.

Regarding the study itself, the observations are the more curious part to me. Actually the hypothesis outlined in the abstract is nothing more than figuring out whether hypermobility or craniocervical obstructions are overrepresented in ME/CFS. If the measurements were done correctly by a competent radiologist, it seems to confirm their hypothesis. 80% with obstruction in cervical spine sounds really high. 50% (h)EDS sounds really high as well. But obviously it leads to the question whether these findings indicate correlation or causation in terms of ME/CFS.

 

It is as bit more than that:To test the hypothesis that hypermobility and craniocervical obstructions is overrepresented in patients with Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and that a large portion of these patients may have a degree of intracranial hypertension explaining many of the symptoms of the ME/CFS syndrome.

I think that is intended to be read as implying that they think that hypermobility and craniocervical obstruction are potential causes of intracranial hypertension - which might explain symptoms. But, as indicated, I t don't think we have any reason to think these are potential causes of cranial hypertension in this context. If Chiari was involved it would have to be quite major and there would need to be an explanation for the absence of expanded ventricles.

The key problem for me is the assumption that somehow they are studying evidence for a relation between cranial hypertension and ME when they aren't.

The reason I say the London clinic is controversial is that I have seen a document produced by radiologists and neurosurgeons, and I am afraid I forget where from, which indicates that in the UK at least upright MRI is not thought to provide useful information. I personally cannot see why it should since we are not talking about upward migration of the dens of the axis in these cases. (If there was upward migration one could see it on an ordinary lateral cervical spine x-ray.)

 

Right, and my understanding is that 10% will *also* be imaged in Sweden. I think a part of the goal is to compare how observations (measurements) change between supine and upright, and with dynamic v. static imaging. There’s already some literature on this but very little systematic data in this or adjacent patient populations.

Upright can matter a lot for Chiari Malformation/tonsillar ectopia/hindbrain herniation. They found a rate of 13.2% in the first study. It could be higher than that in patients who get upright MRIs.

Here is an example of measurements changing between supine and upright.

https://twitter.com/user/status/1137004631178059776


And for cervical stenosis, there can also be a positional component. Here’s one study: https://www.thespinejournalonline.com/article/S1529-9430(14)01771-9/fulltext#/article/S1529-9430(14)01771-9/fulltext.

There is a whole literature on weight-bearing, dynamic imaging in spine disorders, and an argument that this imaging method is more sensitive for a wide range of conditions, but I haven’t really delved into it.

A part of the problem, of course, is that just because some measurements change, it doesn’t mean they are abnormal...the references ranges for upright, dynamic imaging have not been well-established in many conditions. Or in people with hypermobility, where there is some argument about whether their abnormal measurements may be normal for them.

However, the calculation changes somewhat when you see an individual case, that person presents with clear symptoms, and when you have additional objective evidence of pathology (e.g., abnormal CSF flow on cine MRI, abnormal intracranial pressure readings). Clinical history, presentation and context matter a lot. Imaging measurements are just a first step.

 

I guess there are differences between standing and lying MRIs in healthy people too. Is there a study comparing EDS (the example given in the tweet) patients and healthy control's supine and upright MRIs?

 

Although I have no medical qualifications whatsoever, this feels inevitably correct to me. You take any system, biological, mechanical, whatever, with non-rigid joining structures, then they will align a bit differently according to the direction of the forces they are being subjected to. So if gravity is 90° different in two different scenarios, the stresses and alignments have to be influenced by that.

 

Since weight is big factor in IHH, wouldn't that make a difference in upright and lying down MRI?

 

I think it may be dangerous to assume that changes on sitting are actually due to being upright.

If we take the two MRI pictures above the most significant difference is that the CXA has dropped from 141 to 133. What that means is that in the sitting position the head is a bit more 'nodded' forward than in the supine position. If you think about it in the supine position your head tends to be quite far back and when you sit you nod forward a little to get your eyeline horizontal or slightly down. So the CXA change is not due to gravity, but to posture.

I am not sure what the other measurements mean but I am not convinced that there has been any drop of the cerebellum. If you look at the top of the cerebellum (the frondy blob) on the right hand picture it actually looks a little bit higher in relation to the midbrain to the left of it than on the left picture. It seems that the cerebellum has gone up! The bottom of the cerebellum looks to be at about the same level on both. On the right picture the cerebellar tonsil below the cerebellum (the Chiari( goes further down - but it is also longer. If it had just dropped down it would not be longer, just lower. So the immediate question is whether or not these MRI slices are exactly comparable, because a millimetre or two shift in slice position can make structures seem bigger or smaller. Everything about the cerebellum on the right seems a bit bigger - so maybe this slice is just little bit closer to the midline.

Perhaps more significantly, it does not look o me that the fourth ventricle is in any way compromised by any of this so why is anyone bothering to measure anything at all? The tip of cerebellar tonsil is miles away from anywhere that might be adversely affected unless the issue is Syringomyelia, which it is not.

 

I don't think it is thought that in obese people with IIH the fat actually presses on anything. By and large not much fat (if any) presses on our skull contents.

 

A lot of people have measurements of greater tonsillar herniation when upright v. supine. You’re saying that is largely a function of radiology error or misinterpretation?

Why do surgeons use cine MRI to determine if a Chiari Malformation/tonsillar herniation is actually causing a problem?

I’m going to see if I can chat about that vascular neurosurgeon about some of this, because he is observing in his patient population a lot of weird things that are not in the textbooks but that he has come to expect to see in people with EDS.

This is why I haven’t even attempted to touch this stuff (the IIH stuff). It is so complicated! The anatomy is really complicated. Fluid dynamics is really, really complicated.

 

WHOA. I just found this really interesting study on whiplash and Chiari: https://www.tandfonline.com/doi/abs/10.3109/02699052.2010.490512

I have heard several clinicians speculate that there seems to be a post-trauma or post-infectious trigger in people who are later observed to have Chiari Malformation but were previously asymptomatic. This study also illustrates the important of upright scans.




Primary objective: Chiari malformation is defined as herniation of the cerebellar tonsils through the foramen magnum, also known as cerebellar tonsillar ectopia (CTE). CTE may become symptomatic following whiplash trauma. The purpose of the present study was to assess the frequency of CTE in traumatic vs non-traumatic populations.

Study design: Case-control.

Methods and procedures: Cervical MRI scans for 1200 neck pain patients were reviewed; 600 trauma (cases) and 600 non-trauma (controls). Half of the groups were scanned in a recumbent position and half were scanned in an upright position. Two radiologists interpreted the scans for the level of the cerebellar tonsils.

Main outcomes and results: A total of 1195 of 1200 scans were read. CTE was found in 5.7% and 5.3% in the recumbent and upright non-trauma groups vs 9.8% and 23.3% in the recumbent and upright trauma groups (p = 0.0001).

Conclusions: The results described in the present investigation are first to demonstrate a neuroradiographic difference between neck pain patients with and without a recent history of whiplash trauma. The results of prior research on psychosocial causes of chronic pain following whiplash are likely confounded because of a failure to account for a possible neuropathologic basis for the symptoms.

 

I have seen that sort of misinterpretation so often I think it is highly likely, especially when the majority of the radiological and neurosurgical world do not seem to buy in to this at all.

I have no idea but from what I have seen of the stuff on measurements so far maybe because they think they can tell what matters but actually have no idea. Again, that is something I have seen many times.

To be fair, you are right. Understanding fluid compartments is complex. But the main problem is that people think they can learn about it from books and equations, without having a practical non-verbal model in their head of what is actually going on.

When I started out in synovial physiology in 1980 the received wisdom was that the fluid was secreted as an ultra filtrate from synovial venules and absorbed by the lymphatics. Sounds OK until you point out that the lymphatics are in the synovium as well - in the same place as the venues. So the fluid is secreted and absorbed in the same place. How does that make sense? It turns out that there is a beautifully cunning process that has nothing much to do with bulk fluid flow but ensures that the volume in the synovial cavity (not the synovium) is constant. Something pretty similar is likely to apply to CSF. But the Swedish hypothesis does not even work if you forget the niceties.

 

Yes, fully agree. I was just being a bit of a pedant and saying there must be some change, not that such changes would necessarily be significant ones.

That I could well believe.

I've not been following the technicalities of this discussion, so in this sense I am coming to this cold. Yes, I do think these two slices could well be displaced from each other. I'm guessing the change in CXA is unlikely to be significantly affected by the two slices being displaced a mm or two.

But there are things that look not just to have changed in shape, but in x-sectional area quite significantly. As well as some artefacts that have changed from light to dark, presumably something present in one that was a void in the other? I suppose this could be due to changed stresses moving stuff around, but a much simpler explanation would be the slices not being exactly synchronised. Also guessing that if the structures are approximately symmetrical, then it might not be that easy to tell if a slice is slightly left or right of centre.

e.g. :



I would guess the slightly larger 'diameter' cerebellum is again, as you suggest, indicative of the two slices being slightly displaced from each other. If the cerebellum is roughly spherical, then the two slices could be at slightly different chords.

How do you ensure adequately accurate synchronisation between slices from two completely different MRI scans anyway? Presumably with an external frame and some fancy alignment technology. Nonetheless, unless the supine and upright scans were done in the same session, with the frame not detached in between, I suspect discrepancies would creep in. And how do you ensure the two slices are exactly parallel to each other, given they are not taken from the same MRI run? Unless the head is aligned to precisely the same angle relative to the MRI, in the two completely separate MRIs, then the two slices could diverge from each other across the width of scanned slices; so could be 'aligned' on one side and not the other.

I've gone on about this somewhat, because unless you can eliminate these potential sources of discrepancy between the two slices, then all bets are presumably off as to what else any discrepancies might be due to.

 

If some forms of chiari and CCI are only present in an upright position, wouldn't that mean that staying in a supine position for extended periods would improve symptoms? For example, if one was to stay in bed all day, wouldn't that eliminate or greatly improve the symptoms? I know that for me this is not the case at all. On the contrary, if I stay in bed for too long I feel worse. And what about severe patients who are bed-ridden and are almost never upright, if the pathology only presents itself in an upright position, then why are they so sick, despite almost never being upright?

 

I guess it is possible lying flat might improve symptoms in some patients (and some ME/CFS patients with CCI have reported feeling better lying down — though this could also be due to POTS).

But I am not sure how common it is for CCI patients to have craniocervical measurements which only become pathological in the upright position.

In this video at 15:40 Dr Henderson details a patient who has a CXA of 141º on a supine MRI (which is borderline) and no apparent Chiari; but on an upright MRI his CXA now measures as 133º (which is pathological), and in addition a Chiari now becomes apparent.

Just for reference: normal CXA = 160º to 145º, borderline = 144º to 136º, and pathological = 135º or less.

So this patient was borderline CXA lying down, and with no Chiari; but had a pathological CXA and a Chiari when upright. This is one of the reasons why Dr Henderson and Dr Gilete believe upright MRIs are important, because they think that you may miss pathological cases if you do not use upright MRIs.

The other reason they think upright MRIs are important is because they allow images to be taken with the head in flexion (head facing downwards) and extension (head facing upwards), as well as in the neutral head position. Whereas with supine MRI machines, you can normally only take pictures with the head in neutral.

These flexion and extension images allow you to calculate the translational BAI (which is the change in BAI value as the head moves from flexion to neutral to extension). Any change in the BAI value greater than 2 mm as the head moves from flexion to extension is considered pathological.

However, Dr Bolognese takes a different view, and he does not normally use upright MRIs, preferring supine. Dr B says that supines have higher resolution (as they use stronger 3T magnetic fields rather than the 1T of uprights), and he says uprights can also be prone to motion blur artifacts (as the head is not constrained in upright MRI machines as it is in supines).

Dr Bolognese says he occasionally orders uprights for certain patients, but mostly he relies on supine MRIs, plus the use of the traction test for CCI. Which I guess means that Dr B will normally not measure the translational BAI.

It's not clear what the advantages and disadvantages are in the approaches used by Dr B with his supine MRIs plus neck traction, and Dr H and Dr G with their upright MRIs.

 

Which is interesting, because as i posted in another thread, Bolognese co authored a paper that discussed the differences between upright and supine measurements. I think his reasoning for changing is that the supine imaging is clearer. Not only do most upright MRIs have Lower teslas, often especially symptomatic patient will have a hard time being still , leading to a fuzzier image. Also I think that he thinks traction is important to diagnosis.

As to whether or not the differences in upright mri are due to being more in flexion, i have seen cases where the flexion doesn't even decrease the cxa much, or where extension decreases it, which is odd but I assume is due to horizontal instability. So I assume that while flexion can decrease the cxa one cannot conclude that a decreased cxa in upright position is due to flexion.

To really measure that it would be nice to have some kind of study of supine MRI with flexion and extension versus upright with flexion and extension.

But the other thing is that, if acute cxa is a proxy for brainstem compression, which is shown by modelling in the Henderson et al paper I posted in another thread , does it really matter whether the measurements are all perfectly neutral ans able to be aligned? What matters is that the cxa is showing pathology and whether that happens in slight flexion or not, it is a problem.

 

Unfortunately, I think that there are a lot of paradoxical things happening at the same time. Thinking solely of my case: I did have intracranial hypertension. I did feel worse when I was laying completely flat. However, I also felt worse when I was standing upright. There were two main reasons for this: one, I had diverticuli in my lumbar spine and was very slowly leaking spinal fluid. I also had a tight filum causing tethered cord syndrome, which meant when I walked around, I pulled on my spinal cord, further reducing blood flow to my spine and triggering dysautonomia symptoms. The ideal position for me would have been sitting up in bed, on an incline, feet slightly raised and bent. Totally flat? Bad. Upright? Also bad.

A lot of EDS patients report similar experiences and complex interactions between different physical problems.

Certain aspects of my case are more clear now that so many of my symptoms are gone. I do have a tendency to cycle between high and low pressure (due to recurrent spinal fluid leaks and rebound hypertension). I'm still not sure what is causing it, but I have to think that might have been going on the whole time, in the background.

I look a lot like an EDS patient on the inside. I'm just not hypermobile and have zero skin/wound healing/bone fusing, etc. issues.

 

I did not know he authored such a paper; if you happen to have link to it handy, I'd be interested to see it.

 

This is the one. I can't link the PDF but Google scholae gives option to download

 

You may disagree with the validity of the various measurements used --cxa, translational Bai, Bai, etc... But ample evidence has been provided that these measurements are different enough in upright position as to justify upright imaging in a study. The images will be provided and im sure third parties can calculate the angles etc themselves if the interpretations of the radiologists are seen as the problem.