Thread by @HZiauddeen

🧵 Long COVID Part 2: The concept of syndromes

This is about a major conceptual problem in the study of LongCOVID, which is one of the important sequelae of SARS-CoV-2 infection. This conceptual problem appears to be a major obstacle in understanding & studying LongCOVID.
(1/25)

From what we know so far, LongCOVID is a syndrome. It may turn out to be several different syndromes or disorders or diseases, as we learn more. Unfortunately, given how badly we are managing the COVID pandemic, we will learn more, at considerable cost to those affected.
(2/25)

A syndrome is a constellation of symptoms and signs that co-occur. The individual signs and symptoms can occur on their own or in combination with other signs and symptoms. What characterises a given syndrome is it's particular constellation.
(3/25)

Often the co-occurrence of a particular set of symptoms and signs indicates that a particular kind of dysfunction is occurring in an organ system(s) but that dysfunction can be caused by several different diseases. Here is a nice illustration.
(4/25)
niddk.nih.gov/health-informa…

Another possibility is that a syndrome is the final step in multiple different pathways e.g. the syndromes of depression and psychosis.

In either case we return to the same central point, it is the constellation of co-occurring signs and symptoms.
(5/25)

So if you are trying to define or study a syndrome, you are looking to see if there is a cluster of symptoms and signs that keeps coming up i.e. do we keep seeing B, N and L coming up together?

To examine this, you can examine instances of BNL in a clinic.
(6/25)

(One week)

This would be a way of describing and characterising what seems to be the BNL syndrome.
You could calculate the likelihood of BNL occurring by chance and compare it with what you find (observed vs expected).
(7/25)

(Be my Yoko Ono)

If the observed is much higher than the expected, then it's likely that BNL is not occurring just by chance.
One important point, given that B,N and L are occurring in the same body, they can't really be treated as independent as they are likely related.
(8/25)

(Pinch me)

What this means is that the calculation of expected is not straightforward (symptoms relating to a particular organ system are more likely to co-occur). This is an important point to bear in mind as these constellations are not random.
(9/25)

(The Big Bang)

So if we're looking to study BNL syndrome or LongCOVID, we need to be looking for this co-occurrence. Looking at whether people with BNL have more B than the rest of the population is of limited value, unless B is very specific.
(10/25)

(It's all been done)

By specific, I mean that just by having B you are much more likely to have BNL (perhaps B is a marker of a core part of the underlying dysfunction), like loss of smell in LongCOVID.

However, what you are really interested in is how common is the BNL combination?
(11/25)

(Enid)

This is a fundamental issue with a lot of LongCOVID research, it's comparing Bs, Ns & Ls separately between people who've had COVID & people who haven't (or are presumed not to have as they never had a PCR test, which may have been for several reasons).
(12/25)

(Brian Wilson)

This is a major conceptual problem as it does not acknowledge the co-occurrence and relatedness of B, N and L and one can conclude (as many have) that because B, N and L individually are no more common in the potential BNL group vs the controls, BNL doesn't exist.
(13/25)

(Jane)

Even if you're doing this, the selection of the control group and the symptoms you consider is absolutely crucial. If you compare the least specific symptoms with a poorly selected control group (e.g. people with other ongoing illnesses), your signals will be poor.
(14/25)

(A)

So far I've only talked about the relatively straightforward problem of looking at the BNL combination. There is the much more complicated business of ascertaining and measuring B, N and L or any other signs and symptoms.
(15/25)

(Maybe Katie)

For some signs and symptoms this may be relatively straightforward e.g. breathlessness, high blood pressure, paralysis. For others, the task is much more complex and this is particularly so with those that are more to do with subjective experience.
(16/25)

(This old apartment)

This is an area we have not done as well with in medicine, especially when it comes to subjective experiences that do not have accompanying clear evidence (clinical findings or investigations) or seem to have no 'objective evidence'.
(17/25)

(If I had $100000)

I'll use the example of concentration difficulties which was one of the symptoms in the paper discussed in Part 1 of this thread.

One needs to consider at least the following for the symptom of difficulty concentrating:
1. For how long? (duration)
(18/25)

(Call and answer)

Persistent difficulties with concentration lasting several weeks are unusual, especially if the person does not usually experience this.

2. How severe?
Subjective report- mild/moderate/severe (depends on what internal comparator the person has)
(19/25)

(Alcohol)

The change from previous baseline is crucial. It doesn't matter if 20% of the population have mild difficulty concentrating, the only meaningful comparison is within the same person.
(20/25)

(In the car)

Most of us have a decent sense of our own capabilities and are acutely aware when we are not able to function at the level we know we are capable of. People with fluctuating capabilities (e.g. with chronic illness) often know the range of their fluctuation.
(21/25)

(Leave)

3. What has been it's course?
Is it improving, static or worsening? Continuous or episodic? All of these are vital pieces of information.

This is to illustrate that it important and possible to evaluate symptoms that are about subjective experience.
(22/25)

(Go home)

And here is perhaps the most important point when it comes to symptoms that are about subjective experience: believe people.

Now to all the above, add in the following:
-a precipitating event (COVID infection)
-a persistence of symptom from acute infection
(23/25)

(Baby seat)

-evidence of ongoing organ system dysfunction
-plausible biological mechanisms for observed symptoms

Once you include the above, then it becomes even more ridiculous to just compare individual symptom prevalences.
(24/25)

(Helicopters)

When investigating a potential new syndrome the task is one of gathering evidence, not dismissing evidence till you no longer can.
At the very least you'll explore and try & help your patients' suffering, even if it turns out there's nothing new.
(25/25)

(Too little too late)

Part 1 if you're in a masochistic mood:

https://twitter.com/HZiauddeen/status/1487247705638584322?s=20&t=V2Hdl0K46YnvrThXIRzrRA