Interesting article from Leonard Jason. Here's an excerpt: https://undark.org/2024/03/28/opinion-define-long-covid/ In 2023, the National Institutes of Health’s RECOVER consortium developed an influential long Covid case definition using a survey of nearly 10,000 participants. These researchers found that a yes-or-no question regarding “loss of or change in smell or taste” was the best discriminator between those infected and those not infected with SARS-CoV-2. This symptom also accounted for two-thirds of the score determining whether a person would be diagnosed with long Covid. Among those who were ultimately classified as positive for long Covid, 41 percent had this symptom. In contrast, when questions about frequency and severity of symptoms are included, the results are very different. A study I conducted with a colleague at DePaul University’s Center for Community Research found that only 12.6 percent of those with long Covid had lost or experienced changes in smell or taste at least half of the time and with at least moderate severity. In other words, two-thirds of RECOVER’s case definition, which has been widely cited, relied on a relatively infrequent and less severe long Covid symptom. Establishing a case definition with an emphasis on one such symptom neglects the possible importance and impact of other critical long Covid symptoms. It’s worth examining how symptom questions were asked for ME/CFS, another area of post-viral research. For many years, ME/CFS researchers only assessed the occurrence of symptoms. But in the mid-1990s, scientists realized that the somatic symptoms in the criteria for ME/CFS were very common in the general population, so only measuring whether or not a symptom occurred was too imprecise. Furthermore, one of the most prevalent mental health disorders, major depressive disorder, could not be differentiated from ME/CFS using just measures of occurrence. However, when measures of frequency and severity were introduced, it was possible to differentiate ME/CFS from major depressive disorder with 100 percent accuracy. As some continue to believe that long Covid is a psychologically-based illness, assessing symptoms with frequency and severity ratings in the case definition is essential to reduce possible stigma.
This is all my greatest fear concerning Long Covid research - that any progress will be undone by a failure to define and stratify what is a highly heterogenous condition. If for an observational trial, scientists take blood from a set of patients that ranges from those with post-COVID ME/CFS and post-COVID POTS to those who ‘just’ have post-viral fatigue or anosmia, it’s difficult to see how any meaningful conclusions or associations can be drawn - that category of Long Covid as a standalone entity has limited meaning. And how can findings be replicated if the samples are just looking at quite radically different conditions? Rather than acknowledging the complexities inherent in researching such a heterogenous condition, psychologisers will inevitably use a failure to find associations or to replicate findings as a way to dismiss the biological basis of the condition, which is exactly what Kleinschnitz did recently - he failed to replicate the cortisol, IL-6 and other findings, and immediately concluded that the condition was psychological (which represents such a slovenly academic approach, a complete failure to engage with the subject matter in an analytical rather than partisan manner). I wish far more studies would try and focus specifically on, say post-covid ME/CFS patients or post-covid POTS/post-covid MCAS and so on.
He's obviously right. Doing any study on the basis of "was any symptom on a list of 200 symptoms present 12 weeks after a Covid infection" will and was never going to lead to any fruitful research. However, all of this has been known for years and similar publications critising the vague and inconsistent nomenclature on Long Covid where already published years ago (see for instance https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)02444-2/fulltext). Unfortunately, that never prevented any of this meaningless research to be done or published. Regarding your fears it does seem that some things have become better in the past year and more researchers are focused on certain phenotypes, but the bulk of Long-Covid research is still inherently useless research where apples are compared to oranges, i.e. someone with a slight cough for a couple of weeks, to someone with PICS, to someone with anosmia, to someone with a stroke post-Covid, to someone that has ME/CFS-like symptoms for years. The question is has this understanding come to late with funding already drying up? How many meaningful Long Covid studies are there? Certainly not more than a handful and all of those focused on certain phenotypes/symptomology. The problem with the Kleinschnitz study is that many LC advocates and researchers seemed to advocate that cortisol resembled some sort of marker, whilst the opposite was actually the case. I don't even see that being a cohort problem, but much rather a problem of biomedical researchers misinterpreting their own study results.
With those questionnaires, if the subject wants a certain outcome, they'll adjust their judgement of severity as appropriate. The answers for a questionnaire where a negative diagnosis results in job promotion and a positive diagnosis results in job loss, the same person will fill out the form quite differently, without even outright lying, just adjusting their judgement of the severity or frequency of a symptom. Studies based on questionnaires are of questionable value.
'A pre-determined reduction in functioning should not be required to have long Covid' I disagree with this and think it is one of the big problems with Long Covid case definitions in research: they selected people with symptoms that had relatively little impact on their functioning. What is of most interest for scientific research is not some lingering symptoms that will eventually go away but chronic illness following infection.
I might have misunderstood, but the article is a bit difficult to follow for me: first it suggests that one should measure not only the presence but also the severity of symptoms and then it argues that a reduction in functioning should not be required in case definitions. I would argue that requiring a reduction in functioning is useful for the same reason that measuring the severity of symptoms is useful: to differentiate Long Covid from lingering symptoms that have little impact or severity.
I interpreted this to mean that currently people shouldn't be denied a diagnosis if their symptomology isn't sufficiently severe or impacts their daily life sufficiently to prevent further stigma from happening. From what I understood is that he believes that to develop a case definition one should be using questionnaires that look at severity and impact on daily life and then see what those results end up yielding, rather than pre-determining which severity of functioning would be insufficient based on questionnaires that didn't account for this (For example if questionnaires were to reveal the extremely unlikely fact that not only is anosmia a symptom that is overly common in comparison to the general population but that the questionnaires would also reveal it to be effecting daily life a lot, then a diagnosis could be fruitful even if it doesn't impact the daily life of said person). But I agree, it's somewhat hard to understand what he means in the second part of the piece and maybe I also completely misunderstood it. I certainly agree that the problem in LC studies has been that they have typically merely recorded symptom presence rather than anything else, which is meaningless. My guess is that if questionnaires would account for severity and the impact on daily life then things such as brain fog and PEM would dominate a case definition and then a person could be given a diagnosis based on this case definiton (rather than what RECOVER did) and I think that's the point he's trying to make. I think he's mainly trying to criticise the RECOVER trial that prioritised looking at symptom presence rather then severity/impact on daily life.
I think for studies there might be something else that is optimal which nobody is looking at. That is to include patients that meet both sets of ways of looking at the problem, i.e. what he suggests (large impact on life and not self-resolvent) and what RECOVER suggests (higher prevalence in LC than in the general population). That is taking somebody that alongside all of his other symptoms developed a symptom that has a high prevalence in LC whilst it has a very low prevalence in the general population and that is also someone who has a symptom that is very severe and disabling. For example someone that after a Covid infection has developed anosmia and PEM and has had severe PEM for a very long time (more than 1.5 years). Anosmia is really the only thing that comes to my mind when I think of something that is highly selective of LC but some other symptoms that have a similar flair might exist (maybe something like internal tremors as suggested by Iwasaki could fit the bill and then you for example recruit patients that have internal tremors but that also have very severe brain fog that effects their daily functioning).
I think he is probably right that measuring the severity and frequency of a symptom is better than just a yes/no question. But overall I prefer the points threshold method, with extra weighting for some symptoms, that is used by the RECOVER research criteria to his approach to ME/CFS where every symptom is essential. There are lots of symptoms in ME/CFS which a lot of patients have but you can’t make them a compulsory part of a definition so they don’t get included in definitions he proposes. While you can get a more accurate criteria by using more of the data which is possible with a points threshold method, particularly one that uses weightings like the RECOVER criteria do. Also a problem can be that not everyone recognises they have a symptom: someone might have had, say, sleep problems before becoming ill and not recognise it as a new symptom particularly if they have to have it at least half the time and to a moderate severity. So they say they don’t have it and get excluded. Symptom counting allows for flexibility where it doesn’t all depend on a patient saying they have a particular symptom. As I understand it, Leonard Jason is ideologically against symptom counting and the related method of using a points threshold. So I’ve lost faith that he is the best person to come up with criteria. I would also make a distinction between research and clinical criteria with a higher number i.e. a more demanding threshold for research criteria.
