Jonathan Edwards
Senior Member (Voting Rights)
Sasha asked me: "I have the feeling that you as a clinician are approaching the question of diagnosis in a way that's based on your training, and that the concepts used are perhaps very different to what we as patients might assume.
If that's the case, is there an accessible summary article or similar that we could read to get our heads around it?"
My reply:
No there isn't. Nobody has written on this, but this is the story.
What is a Syndrome like ME/CFS and how do we decide how to diagnose it?
I was discussing this with Professor John Martin at our presentation to UCL Division of Medicine this week. John is about as old as I am and like me thinks he is a wise old bird. I think he often gets things wrong. The best way to score points off him is to quote an even wiser and older bird, Robert Souhami. I pointed out that ME/CFS is a distinct syndrome and why - much as I do in my Qeios piece.
The art of making diagnoses is based on an intuitive skill we all use every day to seek out categories of happenings that usefully tell us about what causes what and so what is likely to happen next time. That may sound very philosophical (metaphysics to be precise) but it is worth trying to understand if we want to do reliable science and medicine. Moreover, it is important to realise that we all use this skill unconsciously and it works remarkably well, even in terms of having a repeating error-monitoring system built in, but it works by trial and error, so it constantly needs that monitoring.
When I say 'categories of happenings' that usually means what we call 'things' - tables, ospreys, broken legs - but since this is all about inferring and predicting cause and effect we are always dealing with categories of events or happenings. An osprey is what we call a category of events, including hatching from an egg, catching fish, and even lying alongside a ruler in such a way as to appear 39 inches long.
OK, so physicians in ancient times recognised categories of events that got called illnesses or diseases. They were, in effect, trying to pick out causal paths that would allow them to predict outcomes in other cases and maybe ways to change those outcomes. They got things wrong, with ideas about humours and vapours but they could probably predict prognosis quite well and identify simple causes and advise. No doubt they could work out that frostbite was due to cold.
(I strongly suspect the 'history of medicine' is as bogus as the 'history of navigation'. We now know from an early fifteenth century Turkish map that South American towns were known before Columbus. The Portuguese had been fishing up the North American coast for ages but kept it quiet. They must have known the earth was a sphere even if learned scholars were supposed to think it flat. The value of vaccination and the origin of smallpox in tiny particles in pus was known way back in the Middle East, maybe ancient Egypt. Jenner got to hear of it from an erudite lady traveller I believe.)
But as we started to understand things clearly, with the circulation of the blood and chemistry and microbiology, it became clear that each person’s illness involved a whole lot of events in both converging and diverging sequences. Person 1's illness might involve A and B interacting to cause P and Q, with P leading on to W and V.
A.B >> Q + (P >> W+V)
Person 2 might have an illness in which A and C interacted to produce P and R, with P causing W and X and R causing Z.
A.C >> (P>>W+X) + (R>>Z)
This sort of complexity is everywhere in modern illness, now that the main causes of ill health are no longer whooping cough, starvation, war and falling trees.
Making a diagnosis in this context, means trying to pick out what A is, or maybe what P is, in the context of W. What we mean by 'ME/CFS' is likely to be something like P. A syndrome is a pattern of history of symptoms plus or minus physical signs - a category of happenings - that we call a syndrome because our unconscious seeking out system tells us that it suspects there is some P hidden in the pathways behind it. A syndrome is a ‘real’ syndrome only if eventually we discover there was a P rather than just a spurious similarity between cases.
John Martin got muddled over this, as most doctors do. I used the example of diabetes, or excessive passing of water, as W. It turns out that it can be caused by both a P and a separate Q. With P, diabetes mellitus, you also get hungry and lose weight. With Q, insipidus, you don't. Polyphagia, polydipsia, weight loss and polyuria is the syndrome of diabetes mellitus. But it is further complicated by there being a quite separate B or C causing P (and no common A). B causes type 1 diabetes mellitus and C type 2.
This is important because treatment is partly directed at P, which is a raised blood sugar, and partly at C, which is mostly a matter of obesity. One day we should manage to deal with B, which is autoimmune damage to insulin producing cells.
The long and the short of it is that for syndromes like ME/CFS we are trying to pick out a common element in the causal path P that may indicate directly how to treat both this and that person or may instead point us to some A, B or C that is what we want to treat, maybe different for different people.
I had to work all this out in order to get to grips with rheumatoid. My life work was to work out that rheumatoid factor antibodies are not P but A. There are alternative antibodies: B, C. We worked out that P is the binding of very small complexes of antibodies to a receptor on a subset of macrophages that only occur in places like joints and pericardium. Having worked that out we were suddenly able to string all the letters together and suggest that rituximab would be useful.
I think it may help to use the analogy of species because we have now worked out exactly what the problems and mistakes are in identifying species. People started out categorising species for hunting and eating. Presumably whales were fish and barnacles were shellfish then. There was probably a major effort to be more precise when people started breeding varieties of livestock and birds. As Darwin points out, the breeding of pigeons and songbirds is very instructive and I suspect was relevant to the attempts by people like Linnaeus to categorise ‘species’ precisely.
It turns out that what roughly we mean by species is a group of animals (or plants) whose existence was caused by the same DNA molecules some time in the distant past. Rather neatly what people mean by a disease (with the a in front) is all the cases of illness caused by identical bits of DNA – taking the role of A. Haemophilia is a disease due to a defective factor 8 gene. Tuberculosis is a disease due to Mycobacterium tuberculosis DNA. Things are now complicated because illnesses rarely have a single simple A cause now, so the idea of a disease is less and less helpful.
Moreover, for species it turns out that what we call species aren’t actually defined by common DNA. This is hugely complicated, but you can have converging and diverging paths here too. If you look hard enough you find that there is no precise, coherent concept of a species. It is, in a sense, not even a real syndrome. But there is a pretty good practical category and that is those individuals that, at this point in history, will tend to interbreed, and not with others, if left to their own devices. They don’t have to look alike, as Dachshunds and Great Danes prove. They can have inherited some DNA from different species in the past. The way they form actually has nothing to do with natural selection, so Darwin was wrong about The Origin of Species even if he was right about how biodiversity evolves.
The relevance of this to syndromes, though, is that taxonomists have gone through a process of refining species categories very similar to doctor’s refining of syndrome categories and for species much of it has involved replacing body similarities with DNA similarities. If you look at a barnacle carefully you can actually see its many legs. That is the first stage. Birds of prey were put together because they have hooked beaks, even if some are scavengers and some wasps’ nest burglars, not ‘birds of prey’. But recently the DNA has shown that falcons are not closely related to buzzards. They are closer to parrots and both falcons and parrots are the closest thing to songbirds – in terms of the causal pathway of descent. Condors were also classed as hawks because of things like beak shape even though they are not birds of prey. Early DNA studies suggested that in fact they were storks, not vultures. But later studies showed that they are in the hawk family after all but not particularly related to Asian vultures.
What I see as the quest for understanding ME/CFS is trying to find the commonality in causation of what looks like the same illness for many thousands of people – which is going to be a P, not an A or B. Like rheumatoid, we can then trace back to A and B. The practical question is what is the most reliable clue to P? Is it the end of the beak or the position of the toes or a way of flying? Is it ‘PEM’ or is it something close to that but not quite that. For rheumatoid we had detailed pathology in many organs nicely described by Dougal Gardner and Eric Bywaters. For ME/CFS it is more difficult, but human biology has moved on so we have new methods to draw on like GWAS screening.
Birdwatches still argue furiously in pubs whether the arctic redpoll is the same species as the lesser redpoll even though they look different, but they pretty much all agree that Caspian gulls are not herring gulls even if it takes an expert to tell them apart. Hopefully, it is now understood that ME/CS is not just chronic fatigue, but until we know what P is, and even if there is a single P, we cannot be sure why we can be sure of that.
If that's the case, is there an accessible summary article or similar that we could read to get our heads around it?"
My reply:
No there isn't. Nobody has written on this, but this is the story.
What is a Syndrome like ME/CFS and how do we decide how to diagnose it?
I was discussing this with Professor John Martin at our presentation to UCL Division of Medicine this week. John is about as old as I am and like me thinks he is a wise old bird. I think he often gets things wrong. The best way to score points off him is to quote an even wiser and older bird, Robert Souhami. I pointed out that ME/CFS is a distinct syndrome and why - much as I do in my Qeios piece.
The art of making diagnoses is based on an intuitive skill we all use every day to seek out categories of happenings that usefully tell us about what causes what and so what is likely to happen next time. That may sound very philosophical (metaphysics to be precise) but it is worth trying to understand if we want to do reliable science and medicine. Moreover, it is important to realise that we all use this skill unconsciously and it works remarkably well, even in terms of having a repeating error-monitoring system built in, but it works by trial and error, so it constantly needs that monitoring.
When I say 'categories of happenings' that usually means what we call 'things' - tables, ospreys, broken legs - but since this is all about inferring and predicting cause and effect we are always dealing with categories of events or happenings. An osprey is what we call a category of events, including hatching from an egg, catching fish, and even lying alongside a ruler in such a way as to appear 39 inches long.
OK, so physicians in ancient times recognised categories of events that got called illnesses or diseases. They were, in effect, trying to pick out causal paths that would allow them to predict outcomes in other cases and maybe ways to change those outcomes. They got things wrong, with ideas about humours and vapours but they could probably predict prognosis quite well and identify simple causes and advise. No doubt they could work out that frostbite was due to cold.
(I strongly suspect the 'history of medicine' is as bogus as the 'history of navigation'. We now know from an early fifteenth century Turkish map that South American towns were known before Columbus. The Portuguese had been fishing up the North American coast for ages but kept it quiet. They must have known the earth was a sphere even if learned scholars were supposed to think it flat. The value of vaccination and the origin of smallpox in tiny particles in pus was known way back in the Middle East, maybe ancient Egypt. Jenner got to hear of it from an erudite lady traveller I believe.)
But as we started to understand things clearly, with the circulation of the blood and chemistry and microbiology, it became clear that each person’s illness involved a whole lot of events in both converging and diverging sequences. Person 1's illness might involve A and B interacting to cause P and Q, with P leading on to W and V.
A.B >> Q + (P >> W+V)
Person 2 might have an illness in which A and C interacted to produce P and R, with P causing W and X and R causing Z.
A.C >> (P>>W+X) + (R>>Z)
This sort of complexity is everywhere in modern illness, now that the main causes of ill health are no longer whooping cough, starvation, war and falling trees.
Making a diagnosis in this context, means trying to pick out what A is, or maybe what P is, in the context of W. What we mean by 'ME/CFS' is likely to be something like P. A syndrome is a pattern of history of symptoms plus or minus physical signs - a category of happenings - that we call a syndrome because our unconscious seeking out system tells us that it suspects there is some P hidden in the pathways behind it. A syndrome is a ‘real’ syndrome only if eventually we discover there was a P rather than just a spurious similarity between cases.
John Martin got muddled over this, as most doctors do. I used the example of diabetes, or excessive passing of water, as W. It turns out that it can be caused by both a P and a separate Q. With P, diabetes mellitus, you also get hungry and lose weight. With Q, insipidus, you don't. Polyphagia, polydipsia, weight loss and polyuria is the syndrome of diabetes mellitus. But it is further complicated by there being a quite separate B or C causing P (and no common A). B causes type 1 diabetes mellitus and C type 2.
This is important because treatment is partly directed at P, which is a raised blood sugar, and partly at C, which is mostly a matter of obesity. One day we should manage to deal with B, which is autoimmune damage to insulin producing cells.
The long and the short of it is that for syndromes like ME/CFS we are trying to pick out a common element in the causal path P that may indicate directly how to treat both this and that person or may instead point us to some A, B or C that is what we want to treat, maybe different for different people.
I had to work all this out in order to get to grips with rheumatoid. My life work was to work out that rheumatoid factor antibodies are not P but A. There are alternative antibodies: B, C. We worked out that P is the binding of very small complexes of antibodies to a receptor on a subset of macrophages that only occur in places like joints and pericardium. Having worked that out we were suddenly able to string all the letters together and suggest that rituximab would be useful.
I think it may help to use the analogy of species because we have now worked out exactly what the problems and mistakes are in identifying species. People started out categorising species for hunting and eating. Presumably whales were fish and barnacles were shellfish then. There was probably a major effort to be more precise when people started breeding varieties of livestock and birds. As Darwin points out, the breeding of pigeons and songbirds is very instructive and I suspect was relevant to the attempts by people like Linnaeus to categorise ‘species’ precisely.
It turns out that what roughly we mean by species is a group of animals (or plants) whose existence was caused by the same DNA molecules some time in the distant past. Rather neatly what people mean by a disease (with the a in front) is all the cases of illness caused by identical bits of DNA – taking the role of A. Haemophilia is a disease due to a defective factor 8 gene. Tuberculosis is a disease due to Mycobacterium tuberculosis DNA. Things are now complicated because illnesses rarely have a single simple A cause now, so the idea of a disease is less and less helpful.
Moreover, for species it turns out that what we call species aren’t actually defined by common DNA. This is hugely complicated, but you can have converging and diverging paths here too. If you look hard enough you find that there is no precise, coherent concept of a species. It is, in a sense, not even a real syndrome. But there is a pretty good practical category and that is those individuals that, at this point in history, will tend to interbreed, and not with others, if left to their own devices. They don’t have to look alike, as Dachshunds and Great Danes prove. They can have inherited some DNA from different species in the past. The way they form actually has nothing to do with natural selection, so Darwin was wrong about The Origin of Species even if he was right about how biodiversity evolves.
The relevance of this to syndromes, though, is that taxonomists have gone through a process of refining species categories very similar to doctor’s refining of syndrome categories and for species much of it has involved replacing body similarities with DNA similarities. If you look at a barnacle carefully you can actually see its many legs. That is the first stage. Birds of prey were put together because they have hooked beaks, even if some are scavengers and some wasps’ nest burglars, not ‘birds of prey’. But recently the DNA has shown that falcons are not closely related to buzzards. They are closer to parrots and both falcons and parrots are the closest thing to songbirds – in terms of the causal pathway of descent. Condors were also classed as hawks because of things like beak shape even though they are not birds of prey. Early DNA studies suggested that in fact they were storks, not vultures. But later studies showed that they are in the hawk family after all but not particularly related to Asian vultures.
What I see as the quest for understanding ME/CFS is trying to find the commonality in causation of what looks like the same illness for many thousands of people – which is going to be a P, not an A or B. Like rheumatoid, we can then trace back to A and B. The practical question is what is the most reliable clue to P? Is it the end of the beak or the position of the toes or a way of flying? Is it ‘PEM’ or is it something close to that but not quite that. For rheumatoid we had detailed pathology in many organs nicely described by Dougal Gardner and Eric Bywaters. For ME/CFS it is more difficult, but human biology has moved on so we have new methods to draw on like GWAS screening.
Birdwatches still argue furiously in pubs whether the arctic redpoll is the same species as the lesser redpoll even though they look different, but they pretty much all agree that Caspian gulls are not herring gulls even if it takes an expert to tell them apart. Hopefully, it is now understood that ME/CS is not just chronic fatigue, but until we know what P is, and even if there is a single P, we cannot be sure why we can be sure of that.
Last edited:
