ME/CFS Epidemiology - sex ratios, female predominance

Copied post

Unless they provide a histogram, and account for participation biases... There may still be a lack of sample size to make any conclusions about an earlier age peak too.

 

Just found this: https://ourworldindata.org/gender-ratio#sex-ratio-at-birth

Interesting.

 

It may not be relevant to what they are saying here, but one x chromosome is switched off in every cell in a woman. this is the barr body (?) that they used in sport to determine if someone was female or not.

This switching off is one reason why one of identical twins can have a genetic disease while the other does not.

 

Does anyone know if the high female:male ratio is only in adults or in both adults and children?

 

I think both?

 

The way it felt for me as male adolescent with (maybe) ME/CFS was that I wasn't allowed to be ill. There was a lot of pressure to go to school and my symptoms were reinterpreted to fit the belief system of the adults. I was not considered a credible witness to my own condition (and indeed, I was very confused by what was happening, but still not as confused as everyone else). I was expected to push through, and if that didn't work or I refused to do it, then that was a behavioural/psychological problem. A teacher convinced himseld that I had a secret drug addiction. These may reflect more the dysfunctional in my family and school environment than society's attitudes.

Even though I clearly was ill (in retrospect), I didn't see myself as having an illness but kept trying to make the aforementioned approach given to me by the adults work. Until at some point, after many crashes, each causing more demoralization and despair, I rebelled against it. Then I was no longer pressured as much, but instead abandoned, as a "bad person".

I see a lot of similarities between this dynamic I exerienced and the PACE trial author approach to ME/CFS.

 

I think it's more acceptable for a woman to not work full-time than for a man.

And indeed what I've seen in the ME community is very few men who can't work full-time ending up with a more serious relationship (as defined by a composite of starting to live together/getting engaged/getting married/having children) while it's not rare for women who can't work full-time to end up with a more serious relationship (as defined by a composite of starting to live together/getting engaged/getting married/having children).

Working full-time is very challenging with ME/CFS. I don't think it's easier than working part-time (or not at all) with some extra domestic chores.

 

Copied post

In DecodeME we see a female/male ratio of 5-to-1 (https://openresearch.nihr.ac.uk/articles/3-20), and across all ICD10 G93.3 codes in England we see a ratio of 4-to-1 (https://www.medrxiv.org/content/10.1101/2024.01.31.24302070v1). As these are for UK populations and count >17,000 cases, I do think personally that these are the most relevant and reliable estimates.

 

Copied post

Bakken: this is the lowest of studies I trust. DecodeME is likely better as bigger and broader recruitment (doesn't need GP referral) at 83% female (Mail in rate sample return was VERY high, reducing bias risk.)

Japan/S Korea prevalence. As before, prevalence studies tend to have small samples, wide confidence intervals, so are unreliable for sex ratios. And most of the studies I've seen from these countries (not necessarily these ones) are of questionable quality.
Nacul: not ideal either: no doctor ME diagnosis for many cases (questionnaire only of cases selected by medical codes with a GP reviewing records only).

Insurance claim databases have a more than iffy record on diagnosis (including having the highest rates of incidence in those aged 65+, suggesting it is can be used as a bucket code).

It is possible the true ratio is as 3:1, still pretty high, but I think it could well be 5:1 (I think DecodeMe is 6:1). However, we are revisiting old ground and I will leave it there.

 

Copied post
I tend to agree with @Ravn about not needing a lot of background on ME/CFS, and also with @Hutan 's post (on the epidemiology thread) on sex ratio.

Just as a technical point. My main concern about internet recruitment originally was that the GWAS might pick up an allele linked to behaviour that included answering internet calls. The defence would be that we do not know of any such alleles and if they had a major influence they might have been picked up in other studies.

But we do have one genetic difference that is well known to be associated with behaviours relating to health care and that is the XX, XY difference. Rates of attending GPs and outpatients are confounded by female-related health issues in long adults, which is where differences are regularly reported. But we know that women are more likely to deliberately take up healthy diets (regardless of whether they stick to them!). My guess is that a call for volunteers for an ME/CFS study is likely, for a complicated combination of reasons, to be overweighted towards women responders by a factor of 1.5-1.7. I have no data but that means that for me the most likely interpretation of a 5:1 response is that it reflects a 3:1 prevalence.

Maybe, as Rain says, the safest thing, not to raise hares about methodology, is just not to quote a figure. It isn't relevant beyond there being a female predominance.

 

I've reread that paper. It's interesting and has some notable ME authors: Lucinda Bateman, Charles Lapp and Peter Rowe.

However, there are issues with the data that cast doubt on the accuracy of diagnosis and so the finding of 59% female (sex ratio 1.4) for ME codes as shown in Table 8.

The consensus on the discussion thread is that the CFS data isn't worth much because its codes includes chronic fatigue. The authors were unable to create a robust model using this data, which is why they looked separately at just ME using diagnostic codes G93.3, and ICDM-9 code 380.71, which was also when the data was collected.

The big anomaly about this ME data is the prevalence rate by age shown in fig 2. Prevalence should increase with age, with increases between age bands broadly reflecting age incidence. I've shown it below with added straight lines to help show where incidence increases:



What we see is fairly steady increases implying a steady rates of incidence - with a big increase over the age of 70 ) e.g for females prevalence increase from around 160/100k aged 50-59 to double that for ages 80-89.

Physicians normally say that diagnoses over age 60 are rare, not least because of the difficulty ruling out other potential causes.

For comparison, here is incidence (not prevalence, so you see peaks and troughs directly) for the recent 11k EMEA pan-european study that asked PwME about age of onset (Bakken/Hilland only have age at diagnosis). The pattern is similar to Bakkken/Hilland with a peak in adolescence (note, this is only seen in a few individual countries in the EMEA study) and a substantial increase midlife peaking in late 30s/early 40s. Then a sharp drop with very few new cases after age 60 (cropped by my screenshot but check the original)



And these studies are backed up by reports from people like Charles Shepherd and Tony Komaroff who have said that the most common age of onset is midlife. By contrast, the medical claims data has no midlife peak but a big geriatric surge. This suggests that the codes measured are not accurately picking up ME - or even a cohort that is mostly ME.

This might be because awareness/understanding of ME is poor among US doctors (e.g. the Jason prevalence study finding that a large majority of specialist-diagnosed cases it found had not already been diagnosed). The claims database covers both physician and hospital diagnoses (the split isn't given): physician diagnoses are usually less reliable, and I don't know how much specialist diagnosis there is in the US. Above all, I can't see a reason why we should assume this study is right, and many others are wrong.

I'm done for the day/this study

But when I have the energy, I will post on the big hospital-only G93.3 studies, which seem to offer the best sources for robust figures and hopefully an area where we can find agreement:
Bakken 2014
Hilland 2022 (an update of the Bakken findings)
Samms 2024 pre-print with 100k NHS hospital G93.3 cases.

 

Thanks to @Hutan for splitting off this thread on sex ratios/female predominance.

I was going to focus on the other big studies, but wanted to make a few points in response to yours above.

I'm glad we agree that the prevalence data (and implied incidence) looks wrong. The question is whether there are good explanations for it.

That's an interesting point. Even those 65 are quite likely to have got their diagnosis in the 1980s or earlier when ME was the only diagnosis available. That could boost relative rates for ME in the over 65 given that a significant proportion of younger people may well have got a CFS diagnoses instead.

The question mark over that is that any diagnosis was very hard to get back in the 80s and earlier. It certainly was in the UK, with only a few specialists known to be willing to make it, e.g. Peter Behan in Scotland. And the UK seemed to be well ahead of most places in ME diagnosis, with a few US exceptions like Dan Peterson. It's hard to see there being enough ME diagnoses back then to make a difference.

Also, we would still expect to see a jump in prevalence in midlife, when most people would be in the "CFS" era. The recent EMEA study - which recruited via social media and had few respondents over age 70 still saw strong midlife peaks in countries across Europe. (This is a similar age range to this study with the older cases removed.)

I'm not sure that follows. If doctors are misusing codes for diagnosing older folk, I don't think we can conclude their diagnoses of younger folk are reliable.

Many things are not inconceivable. The question here is are they and the findings of this study likely?

Let me summarise my concerns about this study:
1. The prevalence data is very different from what we would expect and it's not clear if this can be explained.
2. The US has poor reputation for diagnosing ME/CFS. There are the studies finding most community-recruited cases (eg Jason) had not been previously diagnosed. There seem to be very few specialist clinics. And #MEAction have campaigned hard in the US on the need to make diagnosis easier, at least when this data was gathered (different from the UK as I will post later).
3. As the paper notes, this data comes from many different medical providers and ther eis no way of knowing how the different organisations use the codes.
4. The data covers both phsysician codes (notoriously unreliable) and hospital ones. There seem to be very few specialist clinics in the US, so I don't know how good the hospital diagnoses are.

I will come back to these issues when looking at other relevant studies (which no doubt have their own weaknesses). But you may be pleased to hear I won't return to this study .

 

NHS Hospital Episode Statistics Samms & Ponting preprint (thread)
80% Female/3.9:1, 100,000 cases diagnosed with the G93.3 code.

This is the biggest study for sex ratios using codes that do not include chronic fatigue.

Data quality/diagnosis
The dataset reaches back to 1989, but outpatient data is only available from 2004 (initially on a trial basis), and this will provide almost all the cases. There will be negligible numbers of A&E cases. And sadly very few inpatient admissions (many of which might be wrongly-coded as eating disorders et cetera).

There's likely to be some reasonable consistency of diagnosis in that almost all of them will be through the new CFS/ME clinics set up from the late 90s onwards (note that some of these are also fatigue cinics, but fatigue would not be included under a G 93.3 diagnosis). The clinics were primarily set up to offer CBT and graded exercise, but diagnosis is a substantial part of their work. They only consider cases referred from GPs, so are unlikely to include any mild cases. Although the clinics are independent, historically they have networked and the majority of the 43-odd clinics are members of BACME.

However, diagnosis will be far from perfect because CFS/ME clinics are underresourced and can't always get the exclusionary tests they want.

Results
The study found an overall diagnosis rate of 0.16% for people referred to hospital, with a sex ration of 3.9 (80% female).

It also found that diagnostic rates amongst white people were far higher than those for non-whites, even though that the slim evidence we have suggests that true rates are higher amongst black and ethnic minority groups (e.g Dinos 2009). This is a big diagnostic discrepancy by ethnicity than for other common illnesses.

There's also an almost tenfold difference in diagnostic rates amongst white people between the different ICB local health boards (e.g. 0086% to 0.82% for women). Again, nobody believes there is a true difference in prevalence by region, so this again points to an issue with diagnosis.

Cornwall and the Isles of Scilly has the highest diagnostic rate
The highest diagnostic rate by a long way is for Cornwall and the Isles of Scilly Isle at 0.6% for white people (based on over 3,000 diagnosed cases). This makes a lot of sense, because for a very long time, Cornwall has been particularly good for CFS/ME – at least at identifying it (while offering the same ineffective treatments). Now-retired consultant Professor Tony pinching championed the illness for years, not only for Cornwall but for neighbouring counties (which also have higher diagnostic rates). He played a big role in training doctors in the area, including as Dean of the Peninsula School of Medicine. Cornwall has also been amongst the best at patient engagement.

Sex ratio remains high when diagnostic rates are very high
One argument is that the sex ratio is biased because of diagnostic choices and are not representative of the wider patient population. The results for Cornwall (based on over 3,000 cases) indicate that even where diagnostic rates are very high, the sex ratio is much the same: Cornwall has a sex ratio of 3.9, the average for the study.

 

I've often thought that the sex difference is a big assumption that we do really need to test before we use it as a clue to the level of using it to point to where to look for the mechanisms or checking 'fit' of any models.

due to all sorts of things that might be little to do with the pwme themselves ie cultural and about the people they may or may not see and their assumptions or as you say things like GWS or other alternative diagnoses, or just inappropriate dismissal or misdiagnoses.

and because of this, and because there are bodily differences between males and females, this leads to us needing to be aware of it being possible 'assumptions' on types and symptoms and prognoses/the way certain things affect other things in the illness (as well as common comorbidities people might have or that ME/CFS might make people more suscptible to) are possibly slightly red-herring things. Because that one is a chicken-and-egg - if you don't describe heart attacks as they appear for certain women then those take longer to pick up and vice versa.

I'd be pretty intrigued to have a thread of male experiences of these things, ie if there are a number of men here who might input directly too on this (so we can compare notes), to see if there are common themes/trapdoors that might make diagnosis take longer, need more luck or culmination of certain factors (like a good GP), different common 'other initial diagnoses' or assume its just behavioural/will go away, and so on

One example is I know as soon as I thought of it then realised how many I've had say it to me over the years that I've come across quite a lot of men who've had things described as 'had glandular fever once and since then I've never really been the same...' type thing. Only two of those were people I'd later met with ME (and were saying this retrospectively as a cause), so I wonder whether those people I came across always just described it as due to GF? or eventually recovered enough they just accepted it?

I also know I'm engaging a cliche by saying there is a suggestion that men apparently are less likely to chinwag about their lives than women - and from my experience with women there are a lot who just do smalltalk anyway or might hear 'fatigue' and think all the mis-assumptions. Plus men don't only have male friends and colleagues and family.

But if someone isn't getting medical support that knows ME/CFS but happens to know someone else who has it (not the classic 'my mate's mate who had it a year and recovered') be it a colleague, relative, friend or 'one or two removed' from this then that often is a helpful source of suggestion in the past at that stage where someone is stuck in the wilderness wondering what is happenning to them.

I'm curious is this any more or less likely to have happened for males too? If it is then I guess there is the roll-over effect because there are fewer men out there with it so fewer people to think to mention it.

 

Trouble is, if I felt the way I do these days when I was forty I would be sure I had ME/CFS. But I am pretty sure it is just being over the hill and full of worn out joints, blunted nerves and wasting muscles. A diagnostic requirement of ME/CFS is that the symptoms are not explained by something known. Being 75 is something known!