Re: Woman and statins - A good debate

MarilynMann wrote:
On May 27, 8:07 pm, David Rind <d...@xxxxxxxxxxxxxxxxxxxxx> wrote:

Sorry, didn't answer the other half of this. Are you talking about the
study published a couple of weeks ago in BMJ? My recollection is that it
showed no overall effect of aspirin on cognition.



Yes, that is the study I am talking about. The aspirin group
performed better than the placebo group in category fluency but not in
the other tests. In the subsets of women who were current smokers or
who had hyperlipidemia, the aspirin group experienced significantly
less cognitive decline than the placebo group. Intuitively, those
results make sense if women at CVD risk are also at risk of dementia.
It is also worth noting that the dose was only 100 mg every other
day. Obviously, more research is needed, but those results didn't
seem like proof of no effect to me. It is possible the result would
have been more impressive at, say, 81 mg/day and/or starting at a
younger age.

Marilyn

I'd have to look more carefully at the actual study, but this seems like the sort of issue of subgroup analysis I was mentioning.

Overall the treatment had no effect on cognition, but on one subset measure of cognition (fluency) it was beneficial, and in a couple of subsets of patients (smokers, hyperlipidemia) it was beneficial.

Even though the latter subset results may be biologically plausible in some way, you need to be very suspicious that all those subset results could be due entirely to chance. It's just the nature of what happens when lots of statistical tests are performed on lots of subsets.

--
David Rind
drind@xxxxxxxxxxxxxxxxxxxxx
.

Quantcast