The recent breastfeeding/IQ study: I've read it, which is more than some journalists have done. Anyone want to hear what the study actually says?

[welliemum]

I'm treading on tamum's toes here a bit - she's a Proper Geneticist and I'm not.

However.. I do have a bit of practice at reading papers.

First of all, here's the abstract. The full article needs subscriber access unfortunately.

Main finding: In the populations they examined, they found 2 versions of a particular section of a gene. For children with the (commoner) C version, breastfed children had significantly higher IQ scores. For children with just the G version, breastfeeding didn't correlate with IQ.

Can we be sure this is a true effect?
The researchers were very careful on this point. They ruled out:

1. confounding by social class
2. confounding by mother's IQ
3. the gene affecting whether a baby breastfeeds
4. whether the mum's genetic profile affected the baby's growth and hence development

... and several other more technical points.

I think it's a very good study. The researchers have been very careful to exclude other explanations for their findings, and are only claiming results that are well-supported by evidence. No wild claims here.

IMO the study is convincing in a number of ways:

1. the gene cluster in question has a strong link to both intelligence and breastfeeding (via fatty acids) - ie biological plausibility
2. the study methods are sound as mentioned above
3. it can explain why previous studies on bf and intelligence have had conflicting findings (the effect would depend on each population's genetic profile and this wasn't measured).
4. the fact that they studied 2 populations half a world apart and found a consistent effect is very significant. (Repeatability is a big pitfall in this type of genetic study)
5. er, and another one which I now can't remember

It's a pity though, that they couldn't measure how long the babies were bf for. If they showed a great effect for longer bf, that would be additionally convincing; and more detail would also be hugely helpful for women unsure how long to bf for.

One of these populations was a birth cohort from NZ in the 70s. Only 57% were bf, and they would have had a very short duration of bf. No 6 month exclusive bf in those days. This suggests to me that even a little bf can have a big effect.

Tamum or anyone - what do you think?

Yes ben goldacre is great. Hooray for reason!

Welliemum - thanks so much for your summary of the new study. Really useful for interested non-scientist mums like me.

While I was googling to try to find out more I came across another study on this subject from 2006 (which I vaguely remember being reported at the time) which concluded that the difference in iq's between breastfed and non-breastfed children was accounted for by the fact that women with a higher iq are much more likely to breastfeed (ie it's hereditary), and that once you take that factor out there's no significant difference between the iq's of bf and ff children. I think they compared breastfed and non-breastfed siblings - see [http://www.bmj.com/cgi/content/full/333/7575/945].

I know the new (2007) study filtered out these confounding factors (sorry not sure if I'm using these terms quite correctly), and I'm really interested to know how the researchers on the 2006 study could have got it so wrong.

Do you (or any other science literate MN'ers) know anything about this, and whether the new study referred in any way to flaws in the 2006 one?

Hi kittock!

Just writing in great haste to say I've seen your post - but I don't have much access to the computer at non-work times like weekends....

The study you mention has been pretty much superseded by this one, but I agree, it would be very interesting to compare the studies and think about why their conclusions were different.

Will get onto this as soon as I can (but unlikely to be over the weekend) - can you bear the suspense?

PS I wouldn't say the other researchers were wrong, just that their method was less sensitive iykwim.

Thanks welliemum. Looking at some of the responses to the 2006 study on the BMJ website, a few have pointed out that breastfeeding was not well defined - infants were classed as breastfed regardless of the duration or exclusivity of breastfeeding - and durations were generally pretty short - so perhaps this could be an explanation as to why so little difference was found between the two groups.

Do you know what the duration of breastfeeding was in the new study?

Finally, have some computer time and am just reading the articles now.....

Back again.

The BMJ study measured intelligence in bf vs non-bf babies, and found a big difference - but the difference pretty much disappeared when corrected for mothers' cognitive ability.

Hard to say exactly without seeing their data, but the populations seem very different.

Why did one study show a significant effect of breastfeeding and the other didn't? My guesses would be:

1. The BMJ study was of a US population and was ethnically mixed - genetic study was of 2 homogeneous populations.
2. The BMJ study babies had a median duration of 3 months bf (that's not exclusive, any bf) and only a quarter were bf - compared to half of the babies in the genetic study (duration not known but I believe likely to be longer)
3. The BMJ study showed quite a small difference even before correcting for mothers' intelligence

... all of which put together suggest that any true effect of bf on intelligence could have been hidden by the small proportion of bf babies and the mixing of genetic types.

The main problem with the BMJ study though, is the "black box" approach. They measured intelligence and noted bf, but without any insight into the mechanism of how one might be influenced by the other.

We all know that intelligence testing is full of difficulty because of social/educational confounding, and we also know that bf has a strong social component. So when you see a result on paper, it's hard to judge what is truly happening.

What I suspect happened is that the genetic effects were obscured by mixing in the US population, so the study ended up measuring the effect of maternal intelligence on baby intelligence instead - or even (I am a cynic) the effect of maternal social class on baby social class.

The genetic study, on the other hand, has shown a different effect for babies with different genetic makeup, which is a very powerful finding because it takes out a lot of the social aspects. The researchers did explore other explanations for their finding (I talked about this in the OP), but it's very plausible that the effect they're reporting is a true one.

OK, must run, but thanks kittock - it was an interesting read!
PS you're right about confounding - and it's a huge issue in this sort of study so it's a really important concept.

Agh, I meant to say that the BMJ study found a significant difference, not a particularly big one.

Thanks Welliemum! Interesting study, esp the NZ angle
As someone who struggled to BF last time, ended up expressing for 7 weeks until milk ran dry you've given me a good reason to do the same again if we can't establish BF this time again.
Must admit I've always been a subscriber to the view that effects on things like intelligence must be more sociological than biological but this is v interesting.
BTW, you might like to meet my Mum one day - she's a sociologist specialising in early childhood educational development.

Yikes, buzzy, fingers crossed you don't need to go the expressing route this time. That sounds seriously hard.

I'd love to meet your mum! But only if she doesn't mind being cross-examined for hours....

She's actually in Qatar at the moment as part of a contract setting up ECE centres over there (entirely new concept it seems!) but I'm sure she'd love to meet you some time. I'll make sure to put your name on the list of people she needs to let know when LO arrives and who knows you might bump into her at my house one day over Xmas/NY when I'll be grateful for all visitors!!!

Thanks again Welliemum - it's really interesting.

Shame they didn't look at the duration of breastfeeding in the new study - would be great if there was more research on this.

I agree, kittock, it's frustrating.

It only takes a moment's thought to realise that bf is absolutely not a "yes/no" thing but a spectrum, with everything from giving baby just a little colostrum to bf for 4 (or more) years. And that's before you think about mix-feeding or age of introducing solids.

Not the fault of the people doing the studies as they were dealing with old data. But the people collecting data originally should have thought of this really.

Buzzy - wow, your Mum's job sounds so interesting! Excellent plan and I hope to meet her plus baby buzzy soon!