Commentary
Birth cohorts are in some ways a dream tool for identifying risk factors for any illness, with risk factors giving clues about what might cause the illness.
Method
This study is based on a well-established birth cohort of 16,500 babies born in one week in Great Britain, with particpants followed up at ages 5, 10, 16 and 29-30 - mostly with questionnaires covering physical, psychological and social factors, but with some medical data too. "CFS/ME was identified by self report questionnaire completed at 30 years. Participants were asked whether they had ever had CFS/ME, whether they currently had CFS/ME, and at which age their condition developed".
The self-report is the major weakness of this study. I don't think it renders the findings meaningless, but it does mean they need to be treated wtih a degree of caution. Interestingly, a recent UK study found that 31% of CFS cases identified by GPS did not meet CFS Fukuda criteria, according to an exhaustive questionnaire. (Presumably a good proportion of these cases would have had their diagnosis confirmed by a consultant.). I'm assuming that self-report would be less accurate than this, with maybe 40%-50% of cases not meeting Fukuda.
Prevalence rates
91 people out of 11,261 reported
ever having CFS/ME, giving a 'lifetime' rate of 0.8%. The prevalence rate, ie those who
currently had CFS/ME was 0.4%, which is in line with other estimates - one reason why I think the self-report data isn't meaningless.
"Lifetime Prevalence"
Of course, ME/CFS only really started to get mainstream recognition in the eighties and it's no surprise that almost all reported cases in all 3 Birth Cohorts are from the eighties or later, regardless of which cohort you look at. For this 1970 cohort, there were only 2 reported onset prior to age 16 (ie before 1986); range 14-29: median 24, mean 22.9, SD 4.6 So effectively the 'lifetime prevalence' rates really means 'prevalence for all cases since the 80s'.
Debunking theories of ME
The main aim of the study was to test numerous theories about what caused CFS against some data: risk factors examined included birth weight, atopy (allergies), obesity, childhood behavioural problems, childhood psychological problems and maternal psychological problems. Almost all of these were ruled out (or rather there was no evidence to support them), with only 2 possible risk factors identified:
Childhood sedetary behaviour is a risk... or is it?
High exercise levels had been suggested as a risk factor for ME but this study found evidence of the opposite: sedentary childhood behaviour (or rather never/rarely taking part in sports in spare time at age 10) was a risk factor for CFS. However, only 16% of CFS patients had been sedentary, compared with 8% of the main sample
so sedentary behaviour could only explain 8% of all CFS cases. This is fairly typcial of cohort findings: even where a risk factor is identified, it usually only explains a small proportion of CFS cases ie is a bit of a side show.
However, the sedentary finding is rather shaky:
- it's only just significant, with p=0.04. Arguably, the authors should have made some adjustment for multiple comparisons (more here) so the threshold for significance should probably have been lower, eg. p=0.01, not 0.05.
- If 40% of self-reported cases were misdiagnosed, that leaves far fewer 'true' cases, and the smaller numbers would almost certainly push the p value above 0.05 - and so to non-significance
- When different authors tried to replicate this finding in the 1958 cohort they couldn't: childhood sedentary behaviour was not a risk factor for CFS.
None of which stopped newspaper headlines like "COUCH POTATOES MORE AT RISK OF ME" (info from Dolphin).
So that just leaves:
Childhood chronic illness is a risk factor for CFS
Higher risk of CFS/ME was associated with having a limiting longstanding condition in childhood (odds ratio 2.3, 1.4 to 3.9)
Bizarrely, the authors didn't discuss this finding, even though it was more significant than the exercise one with p=0.003. Yet a link to prior childhood illness could be telling.
Does this finding fall apart like the sedentary one above? Good question, reader. First, as the p value is smaller it's more likely to survive the loss of those misdiagnosed cases. I built a wee model and played with the numbers and it seems this finding is robust. It's still significant even if
1. Only half of all cases are real CFS
2. Up to 60% of the misdiagnosed cases are really due to other undiagnosed physical illnesses and 45% of those cases had chronic childhood illness
ask me for more info if you're really interested, the main point is that you can shake this finding fairly hard and it holds good, though it will break if pushed hard enough.
One note of caution on this finding (actually, two, as I'm not sure it's been replicated in other cohorts):
What's meant by chronic childhood illness?
"The presence of a longstanding condition in the child was recorded from both medical examination and maternal report at 10 years, with maternal report of whether the condition considerably limited the childs daily life."
Seems fine, but I was surprised that as many as 14% of total participants reported a longstanding condition. I was also suprised that there was no correlation between CFS cases and time off school at age 10 (for health or emotional reasons). So I'm just a little suspicicious of how bad the reported longstanding condition really was.
Conclusion
There's some evidence that childhood ill health is a risk factor for CFS, but no evidence for childhood or maternal psychopathology or a host of other factors being risk factors - and sedentary childhood behaviour doesn't cut it either.
Thanks to those who commented on proto-versions of this, and especially to Dolphin who gave the tyres a good kicking.