Review Long COVID Is Not a Functional Neurologic Disorder 2024 Davenport, Tyson et al

[Andy]

Abstract

Long COVID is a common sequela of SARS-CoV-2 infection. Data from numerous scientific studies indicate that long COVID involves a complex interaction between pathophysiological processes. Long COVID may involve the development of new diagnosable health conditions and exacerbation of pre-existing health conditions. However, despite this rapidly accumulating body of evidence regarding the pathobiology of long COVID, psychogenic and functional interpretations of the illness presentation continue to be endorsed by some healthcare professionals, creating confusion and inappropriate diagnostic and therapeutic pathways for people living with long COVID.

The purpose of this perspective is to present a clinical and scientific rationale for why long COVID should not be considered as a functional neurologic disorder. It will begin by discussing the parallel historical development of pathobiological and psychosomatic/sociogenic diagnostic constructs arising from a common root in neurasthenia, which has resulted in the collective understandings of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) and functional neurologic disorder (FND), respectively. We will also review the case definition criteria for FND and the distinguishing clinical and neuroimaging findings in FND vs. long COVID. We conclude that considering long COVID as FND is inappropriate based on differentiating pathophysiologic mechanisms and distinguishing clinical findings.

Open access, https://www.mdpi.com/2075-4426/14/8/799

 

[Hutan]

Primary author Todd Davenport, senior author Sarah Tyson
(sorry for edits)

 

[Deanne NZ]

I am so happy to see this comprehensive paper. Great work from all involved. Thank you.

 

[Deanne NZ]

All together now, Long Covid is NOT FND!

 

[SNT Gatchaman]

All together now, Long Covid is NOT FND!

(Also FND is not even FND)

 

[Jonathan Edwards]

It's all on message but where is this mysterious "pathophysiology"? The case for this not being FND isn't to do with physical evidence. I am not sure that claiming so does more than dent the credibility of the authors and the case.

It disappoints me that everything is couched in this polarized argument which simply reveals that neither side knows what is going on, and that we should focus on real evidence.

 

[Hutan]

The fact that ME/CFS (and Long Covid) looks different to functional seizures and functional motor disorders doesn't prove that ME/CFS (and Long Covid) aren't also functional disorders. Functional seizures and functional motor disorders present very differently from each other too. The imagination of psychosomatic medicine proponents has no trouble scooping all of them up into the 'functional' bucket.

 

[Nightsong]

In my view the reason for it not being "FND" is much simpler: those conditions are mimics of neurological conditions - e.g. pseudoseizure mimicking epilepsy. The same with the others - pseudoparalysis, hysterical blindness etc - they mimic neurological disorders but with no lesion that would result in paralysis/blindness/etc & where reflexes or visual pathways are demonstrably intact so there is a marked disparity between reported symptoms and examination findings.

By contrast ME/CFS isn't a mimic of anything. Findings on examination don't "refute" the symptoms, and in addition there are novel features such as PEM.

 

[Trish]

I have done a quick skim read and can see a lot of thought and work has gone into this article and some useful points are made.

I have some concerns. If I'm misinterpreting the article please say so, I admit I haven't read every word. The neuroimaging part is over my head, for example.

I would want to see the argument framed as:

Neuraesthenia/hysteria/conversion disorder historically has morphed into several 'functional' disorders in some medical quarters, and for those clinicians still comes under the banner of psychosomatic/non organic disorders.

That is problematic because it stifles biomedical research and testing and leads to ineffective psychobehavioural treatments, gaslighting, lack of appropriate care etc etc.

Under that umbrella have variously been placed symptom patterns that can't currently be explained by any known pathology.

They have been subdivided according to symptoms and to some extent precipitating factors into

FND if there are specific neurological signs,

fibromyalgia or complex regional pain syndrome if predominantly pain,

PVFS, ME, CFS if predominantly fatigue with or without infectious trigger,

and MUS, SSD, PPS etc by those who want to lump them all together.

Since the biological basis of all of these conditions is not established, and attempts to treat them with psychobehavioural therapies have been ineffective and in the case of ME/CFS actively harmful, they all need to be regarded as probably having a physical basls, not as psychosomatic.

The symptom patterns of the different subgroups may overlap in some individuals, but the diagnostic criteria are distinct, so there is no reason to use the wrong label for any of them. So ME/CFS and LC are not FND because the diagnostic criteria are different.

I think the framing in this article of FND as the true inheritor of the neuraesthenia/conversion disorder/psychosomatic labelling and LC/ME/CFS as biomedical is a big oversimplification on the history and is unhelpful to people who are diagnosed with FND on the basis of positive signs. It feels uncomfortably to me like in order to insist LC is not FND the authors are suggesting FND is psychosomatic. Have I understood that correctly?

I'm thinking particularly of the diagram in the middle of the paper that starts with neuraesthenia on the left and the upper branch as biomedical ME/CFS/LC and the lower branch as psychosomatic FND. I notice the diagram leaves out the Oxford definition from 1990 that deliberately widened ME/CFS to all unexplained fatigue, and the subsequent 30 years of psychosomatic/psychobehavioural interpretation and treatment that has been so harmful. Edit: This is partly explained away in the text, but still a problematic diagram I think.

To summarise, why not just say LC is not FND because the symptoms are different.

 

[Hutan]

You make a lot of good points Trish, and yes, it looks as though the authors have put a lot of time into the paper, but I agree with Jonathan that it is unfortunately counter-productive. I don't think the paper should be used in advocacy.

To summarise, why not just say LC is not FND because the symptoms are different.

But that is essentially what the authors say. And, as per my comment above, that doesn't work as an argument, because the psychosomatic medicine proponents are happy to give any health condition or symptom that they believe isn't adequately explained by a physical cause a 'functional' label. Those conditions can have a very different physical presentations (e.g. seizures, tics, post-exertional malaise, muscle weakness). Unless conditions and symptoms have a clear physical cause, they are vulnerable to being labelled as psychosomatic.

There's a lot of reliance on there being 'rule-in' signs for FND that are usually absent in Long Covid and ME/CFS (and it is suggested that if the rule-in signs are present in a person with Long Covid or ME/CFS, then the person has both the post-infection condition and an FND). But, we know that the rule-in signs are not reliable. We know that not finding typical epileptic brain wave patterns doesn't mean that such patterns aren't there, just deeper in the brain than the sensor detects. It looks as though the paper has been shaped by Sarah Tyson's previous commitment to the idea that FND exists and can be (and should be) confidently identified.

The intensity and disablement of fatigue associated with long COVID is similar to other post-viral conditions, including post-treatment Lyme disease [26], chronic Epstein–Barr infection [27], and post-mononucleosis syndrome [28].

There is no consistency in how Long Covid is framed - at times it is recognised that Long Covid is a collection of health conditions with different causes (e.g. ME/CFS, observable lung damage caused by the infection). At other times, as in the quote above, it is treated as an ME/CFS equivalent.

There is a lack of polishing of the paper, which we have seen in other papers by Davenport. For example the first sentence in section 4.2 says

Typically, people with functional disorders often exhibit numerous multi-systemic and multi-organ concerns, including various neurologic and psychiatric manifestations such as sensory disturbance, motor weakness, balance difficulty, chronic dizziness, chronic vertigo, chronic pain, chronic fatigue, sleep impairment, urinary and gastrointestinal symptoms, and cognitive dysfunction (Table 1).

So, people with FND often have chronic dizziness.
But Table 1 reports that ME/CFS and Long Covid (here again treated as one condition) can involve dizziness, but can be distinguished from FND, because FND does not.



There are issues with references, for example, this statement talks about chronic conditions associated with infections

Long COVID has caused a renewed scholarly and clinical focus on complex chronic conditions associated with infections [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19].

Reference 18 is 'Multiple Chemical Sensitivity Syndrome: First symptoms and evolution of the clinical picture: Case-control study/epidemiological case-control study'.
Reference 19 is 'Chemical intolerance in primary care settings: Prevalence, comorbidity, and outcomes.'
Neither paper is directly related to post-infection diseases or specifically mentions the condition being caused by an infection, and both leave the door wide open to chemical intolerance being functional disorders e.g.

Despite its relatively high prevalence in nonclinical samples, the diagnosis and etiology of chemical intolerance remain controversial and understudied. Skeptics and proponents often frame the debate in a dualistic manner, claiming that chemical intolerance is either completely psychogenic or completely toxicogenic; however, accumulating data suggest that a more nuanced, multifactorial psychobiological process underlies the condition.....
Although some investigators label these patients as having somatoform spectrum disorder,11,39,40the mean age of onset is typically older than 30 years,12 a finding that is inconsistent with the diagnostic criterion requirement for the earlier age of onset in the Diagnostic and Statistical Manual of Mental Disorders (4th Edition) (DSM-IV) for classic somatization disorder.41 Apart from etiology, elevated levels of subjective mental distress (eg, somatization, anxiety, depression) are a major factor in increasing health care use across diagnoses in the general population.42,43 Persons with chemical intolerance score higher on scales measuring somatization, anxiety, and depression,5,17although the scores on these psychiatric dimensions account for only a small portion of variance in total chemical intolerance scores outside clinical settings.6,44

It is difficult to understand why these two references were included.

 

[Trish]

But that is essentially what the authors say.

I think I expressed that point too baldly. I was trying to make the point that the symptoms are different but none have evidence of being psychosomatic.

I got stuck on that diagram which seems to me to be wrong historically. Maybe in the USA it's different, so they view ME/CFS as having been decided to be biomedical decades ago, whereas FND has continued to be stuck in the psychosomatic field as the inheritor of the conversion disorder neurasthenia label. In the UK the problem is much more that both ME/CFS and FND have been stuck in the psych category and continue to be so as we see with Miller's awful article a few days ago. And it's clearly still a big problem in the USA, with the likes of Walitt and his effort preference nonsense.

I hope one of the authors will explain where I'm misunderstanding. I can see the intention is good, and some excellent people are involved in the article, so I'm a bit bemused.

 

[poetinsf]

Is it really necessary to prove that LC is not an FND? It seems to me that they should prove that LC is an FND and they failed in that regard since FND treatments doesn't work for LC. Treatment, after all, is the ultimate testable prediction that an hypothesis can make.

 

[Jonathan Edwards]

Yes but FND treatment very likely doesn't work for FND. So it is not a discriminatory test.

 

[Yann04]

Is it really necessary to prove that LC is not an FND? It seems to me that they should prove that LC is an FND and they failed in that regard since FND treatments doesn't work for LC. Treatment, after all, is the ultimate testable prediction that an hypothesis can make.

Useful to have a study for patients to have on hand as long covid patients are getting “FND’d” en masse.

 

[Hutan]

Useful to have a study for patients to have on hand as long covid patients are getting “FND’d” en masse.

Yes, but it works against us if the holes in the paper are big enough to drive a truck through. If a patient takes this paper to their GP and says 'look, I don't have FND! because I get dizziness and Table 1 says that people with FND don't', or 'look, I don't have FND! because I have Raynaud's', 'look, I don't have FND! because I don't walk funny and I don't have seizures' I don't think the GP will be convinced.

Is it really necessary to prove that LC is not an FND?

This is a question closer to the mark. The literature is not yet at a point where we can be certain about any physical marker of ME/CFS-like LC. But we could usefully answer the question 'does classifying LC as an FND help the patients at all?'.

4.3. Refutative Evidence from Neuroimaging
Despite its relatively recent recognition, the literature describing significant structural brain abnormalities in long COVID is already extensive. This literature suggests FND cannot commonly explain long COVID, because it indicates neurologic signs, symptoms, and disability may be caused by structural changes in the brain. Such abnormalities have been demonstrated using various imaging modalities and range from changes in gray matter thickness and volume, to macro- and microstructural white matter changes and evidence of metabolic and neuroinflammatory derangement. It is also important to note that despite clearly distinctive systemic immunological abnormalities [120,121,122] and abnormal neuroimmune profiles [121,123] in long COVID, routine clinical structural magnetic resonance imaging (MRI) sequences often return normal findings [123].
Douaud et al. [124] examined structural brain changes in a biobank cohort based in the United Kingdom, before and after SARS-CoV-2 infection. Compared to uninfected controls, the authors found significant gray matter thickness reduction following infection, with a reduction in global brain size. Hosp et al. [125] used an MRI diffusion microstructure imaging technique to evaluate subtle changes in both gray and white matter integrity. Compared to recovered infected patients, those with ongoing symptoms demonstrated widespread changes in microstructure, which correlated with evaluations of cognitive dysfunction. Wu et al. [126] used another diffusion tensor imaging (DTI) technique to evaluate the perivascular space and glymphatic system. These authors calculated flow in the perivascular spaces alongside medullary veins, which lie orthogonal to the projection and association nerve fibers in the periventricular deep white matter. They reported reduction in the indices for glymphatic function in people living with long COVID even following a mild acute infection. Another small cohort study compared recovered, brain fog positive, and brain fog negative patients with long COVID [127]. Dynamic contrast-enhanced MRI (DCE-MRI) showed significant whole brain leakage, indicating increased blood–brain barrier (BBB) permeability, in only the ‘brain fog’ sub-group [127]. Chaganti et al. combined techniques in a longitudinal study of 14 patients with long COVID-related cognitive impairment [128]. DCE-MRI and DTI revealed impairments in the integrity of BBB and white matter microstructure [128]. Simultaneously, MR spectroscopy demonstrated reduced glutamate/glutamine in these areas, leading the authors to suggest that white matter injury may result from glutamatergic excitotoxicity, secondary to reduced BBB integrity associated with neuroinflammation [128]. VanElzakker et al. [129] used positron emission tomography (PET) with a tracer for activated microglia ([11C]PBR28) to report evidence of significantly increased neuroinflammation in many brain regions in LC. Peluso et al. [130] used a novel PET tracer ([18F]F-AraG) to tag activated T cells. Following SARS-CoV-2 infection, activated T cells were found in multiple organs including the bowel and bone marrow, but notably had trafficked into central nervous system (CNS) sites such as the brainstem and spinal cord, where they should be absent. This finding was more exaggerated in patients with long COVID signs and symptoms [130]. Biopsy-accessible tissues such as colon tissue demonstrated residual viral components, and the authors speculated they also might be present in the CNS [130].
A 2021 review article canvased the literature of neuroimaging in FND [131]. The highlighted modalities were functional MRI (fMRI), using both resting-state and task-based paradigms; high-resolution structural MRI evaluation of gray matter; DTI of white matter microstructure; MR spectroscopy; CT/MR positron emission tomography; and near-infrared spectroscopy. The authors note that neuroimaging in FND is early in its development, with few replicated studies, and with confounding factors in terms of clinical heterogeneity and co-morbidities. They conclude by encouraging a multimodal neuroimaging approach to advance the field. Most fMRI studies using blood oxygen level-dependent (BOLD) techniques have shown abnormalities in specific brain regions, yet data have been inconsistent [132]. Very recently, Schneider et al. [133] have attempted to further define the variability of BOLD signal in FND, with particular emphasis on the somatomotor, limbic, and salience networks. However, when structural abnormalities have been found in gray [134] or white matter [67], it remains unclear whether they are a cause, consequence, or comorbidity [132,135].
While neuroimaging in FND is an evolving field, there are already replicated findings in long COVID that point toward a coherent structural pathophysiology. Aspects highlighted in the literature to date involve neuroinflammation with microglial activation, a dysfunctional blood–brain barrier, white matter microstructural changes, as well as reduction in gray matter volume. Systemic dysfunction, such as orthostatic intolerance with reduced cerebral blood flow, is also shown to be a key contributor to symptoms [136,137,138,139,140]. The detail of how these findings are driven from specific and potentially correctable upstream causes is enthusiastically anticipated by patients, clinicians, and researchers alike.

That section on neuroimaging is a nice review of the LC brain imaging field, but the uncertainties noted as relating to FND apply equally to those in the assortment of conditions that constitute Long Covid. Specifically this sums up the situation for both labels beautifully:

The authors note that neuroimaging in FND is early in its development, with few replicated studies, and with confounding factors in terms of clinical heterogeneity and co-morbidities.

For both, the studies are typically small, and all sorts of findings are reported. When I read that summary, I'm not seeing convincingly replicated findings for ME/CFS-like LC that would allow a patient to go for a scan and know with some level of confidence that they have ME/CFS-like LC. I'd be very happy to be convinced otherwise.

 

[Trish]

Perhaps it would be more helpful to frame the argument as Long Covid and ME/CFS are not psychosomatic disorders.

 

[alktipping]

FND exists because neurologists cannot admit how little they know about the human brain and of course its reliance on the body.

 

[poetinsf]

Yes but FND treatment very likely doesn't work for FND. So it is not a discriminatory test.

If FND treatments - CBT, psychotherapy and what have you - equally doesn't work on FND patients and non-FND patients, then maybe FND does not exist? We can't see the underlying mechanism for it after all, so there is nothing that ascertain the existence of FND if the treatment is worthless.

On personal note though, I know FND exists because I've suffered one or two and CBT worked for me.

 

[Hutan]

I think the paper usefully makes the point that FND equals hysteria and conversion disorder and mental illness, a

a dissociative neurological symptom disorder, defined as a mental health condition involving a loss of connection between thoughts, memories, feelings, surroundings, behavior, and identity

That may have some people who have gladly accepted their FND label thinking again.

But, it notes that it has been said that a label of FND can be helpful, and does not really critically evaluate that:

Proponents suggest that just the label of FND may be helpful for some patients who live with troublesome but medically unexplained symptoms and signs [76,77].

There is a lot that could have been said about the negative impacts of labelling Long Covid and ME/CFS as FNDs (suicide, despair, withdrawal of state and family support, ineligibility for income insurance payouts, redirection of research effort to unhelpful areas, treatments that don't help and may harm....), but I didn't see that in the paper.

For example, ME/CFS should not be considered as a functional disorder because more specific case definition criteria best explain its constellation of symptoms, signs, and pathophysiology.

That doesn't really work, because, for example psychogenic non-epileptic seizures has a more specific case definition than FNDs in general, but it is still seen by psychogenic paradigm proponents as an FND.

For example, the UK NICE attempted to classify ME/CFS as a functional disorder in 2017 [78]. This action was met with significant opposition from the ME community [79]. The dispute lasted over two years with the subsequent guideline removing the reference to ME/CFS as FND [78]. Both individual clinicians (generally from the fields of neurology and psychiatry) and prominent national medical organizations in European countries [80,81] persist in classifying ME/CFS as FND despite compelling evidence to the contrary.

This perpetuates the stereotype of the militant ME/CFS advocates rabidly opposed to a mental health diagnosis that have forced NICE into submission, while neurology and psychiatry clinicians and prominent national medical organisations favour an FND label. It might have been okay if the paper then went on to offer the promised 'compelling evidence', but it doesn't.

 

[Sean]

"The authors note that neuroimaging in FND is early in its development, with few replicated studies, and with confounding factors in terms of clinical heterogeneity and co-morbidities."

And boatloads of superficial, ideologically driven, subjective interpretations.

If FND treatments - CBT, psychotherapy and what have you - equally doesn't work on FND patients and non-FND patients, then maybe FND does not exist? We can't see the underlying mechanism for it after all, so there is nothing that ascertain the existence of FND if the treatment is worthless.

I remain unconvinced there is even such a thing as FND, certainly not in any sense that makes it a reliable and useful entity.

The concept, definition, and diagnostic process are a hot subjective-soaked methodological and ethical mess, and has been since it was first conjured up by Freud (as hysterical conversion), and there is currently no means to treat it. So what is even the point of it, certainly at the clinical, policy, and medico-legal levels?

It remains an empirical, intellectual, and ethical desert, and I predict it always will.

There is a lot that could have been said about the negative impacts of labelling Long Covid and ME/CFS as FNDs (suicide, despair, withdrawal of state and family support, ineligibility for income insurance payouts, redirection of research effort to unhelpful areas, treatments that don't help and may harm....), but I didn't see that in the paper.

Yes, what about the very real costs of the FND labelling exercise?