ME Epidemiology - prevalence, peak ages of onset and gender ratio

Interesting.

Sorry about my mistake. I was misguided because of a graph that was reported in Valdez et al. (2018), the study that tried to use ICD-codes in the US to estimate the prevalence.

They probably looked at table I which gives the demographics of all patients in the clinics studied and I made the same mistake in checking. I also didn't realize the Norwegian study was based on such good data. Thanks for pointing that out.

 

They don't.

https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-014-0167-5

 

I would think that being the largest effort of its kind, the IOM report would likely be the most accurate.

Which does not mean it is entirely accurate but likely the most. There is consistency around those numbers as well so if they are off it's likely to not be by much in %.

 

I think IOM is generally pretty good but there are some gotchas to watch for.

It reports an ME/CFS prevalence of 836,000 to 2.5M. But the supporting reference for that upper prevalence is a publication by Jason which includes this statement...

The CDC currently estimates there are up to 900,000 Americans with CFS and another 2.5 million with CFS-like illness.​

CDC has defined "CFS-like" illness to be be 6 months of fatigue but not meeting the 4 out of the other 8 symptoms required by Fukuda. It's certainly not ME. So IMO, that 2.5M upper prevalence for ME/CFS appears to be an error in the report.

 

I agree. I never fully understood where those prevalence figures came from. The article of Jason that is referenced is just a short overview of CFS prevalence in the CFIDS Chronicle.

There are some (rather misleading) studies such as the CDC Georgia study and Wessely's prevalence study in the UK that reported a much higher prevalence of 2,5%. But given that there are approximately 300 million Americans, that would lead to an estimated higher bound of 7,5 million. So that can't be where the 2,5 million came from. I also don't see how it could refer to the CFS-like patients in the 1999 Chicago study. Anyone else has a clue from which study the 2,5 million figure comes?

Even the lower bound of 836.000 is strange. It's the figure the Chicago study reported. But that was in 1999, there are approximately 500.000 Americans more now. Also this prevalence study gave a figure of 0,42% which is higher than the 0,23% (Reyes et al. 2003) or 0,19% (Nacul et al. 2011)study reported, so why should it be used as a lower boundary?

In my view, they should have said that ME/CFS has an estimated prevalence of 0,2%-0,4% and then calculated how many Americans that would be for the current US population (the prevalences are for adult populations only, If I'm not mistaken).

 

Isn't 2.5 million equivalent to the 0.82% (or thereabouts) from the 'CFS-like' figure?

 

Where does the 0,82% figure come from?

 

Yes - In the IOM report, the source of the 2.5M is CDC's report of "CFS-like" illness.

The Georgia study used the overly broad Reeves criteria that included "unwellness" and more mental illness and as a result inflated prevalence 10 fold over CDC's own previous estimates.
The Wessely study from 1997 reported 2.6% for Fukuda but only 0.5% when they excluded comorbid psychiatric illness.

The Nacul study reported 0.19% for Fukuda but only 0.11% for Canadian so perhaps the lower range should be even lower. In reality, we have no idea because much of the epi research is so flawed. In the meantime and at least in the US, the 0.42% estimate from Jason's study has tended to be the most accepted estimate.

 

Ok, but where does that figure come from, which study is it based on?

 

I can't remember. I thought it was due to a doubling of one of other numbers as a guesstimate.

 

The prevalence rates are estimates from studies in adults so 2.5M would equate to about .99% of the current adult US population. There are few studies in children but the prevalence estimates are much lower, particularly in children under 10-12.

 

To my knowledge, there isn't one. I dont know of any study that reported a prevalence of 2.5M (.99% by today's population) in adults and I think the estimates of "CFS-like" illness were much higher than 2.5M. So even if "CFS-like" illness is accepted as a valid thing, that number is still strange.

The upper limit in the IOM appears to be based simply on the CDC statement that there are 2.5M people with CFS-like illness. I can't see anything else in the cited reference that would support the 2.5M prevalence estimate - even ignoring that its for CFS-like illness and not ME

Back to the original point - I think the IOM stats are generally good and accurately reflect the underlying studies but there's a few places like this that appear to be screwed up.

 

According to figure 1 they do.

 

Maybe the way to think about the incidence in women is to consider that there’s a dip in incidence in your 20s, rather than two peaks, one in your teens and one in your thirties....
So perhaps there is something protective going on in your 20s, when you are ‘in your prime’? And if you are genetically susceptible, then you will mostly have succumbed by your forties....?

 

I agree that the ratio is not necessarily convincing, but I am thinking that doctors expect it in women and not in men, so more women get diagnosed (and computers learn to expect this so predict that outcome).

When we know the vast majority of people are undiagnosed (and in other studies, people are overdiagnosed but oddly enough probably still not diagnosing many cases, as they may exclude people with nurological signs for example), it's hard to generalize demographics.

 

I think my story is a pretty common one. I first got sick at 19 but it took 10 years for it to become so serious I was diagnosed with ME. So officially my onset was at age 29, but in reality it was (probably) at 19. So I was pretty sick through all of my 20s, but since I managed to stay in school and didn’t have a diagnosis on paper I was healthy.

I don’t think we can say much about the second age peak for sure until people are more accurately and promptly diagnosed.

 

I was thinking, if this finding would turn out to be true, perhaps it tells us something about the heterogeneity of ME/CFS.

If a predominance of 80% of women is rather unusual in medicine, it might make it unlikely that ME/CFS is a broad collection of different illnesses. Cause it would be quite a coincidence that practically all those illnesses have that unusual predominance of 80% of women. That would be like multiplying an unlikely chance to get a much lower chance. You want to make that multiplication as little as possible because the chance becomes improbable very quickly. If a ratio of 4/5 women is really unusual in medicine, it would be statistically more reasonable to assume that there are few rather than many of these illnesses making up ME/CFS. And if some of those underlying illnesses have a lower female/male ratio, than it would require the other illnesses making up ME/CFS to have an even larger predominance of women than that 80%, which also seems unlikely.

So perhaps the fact that almost 80% of ME/CFS patients are women forms a weak and uncertain but nonetheless an argument against large heterogeneity of illnesses within the ME/CFS label.

PS: Sorry if this doesn't make any sense at all. It's really hot in Belgium right now...