Paleo Phil wrote:
>>> [SF is a strong predictor of CHD mortality in middle aged American
>> Indians.]
>>
>> Are these Indians by any chance overweight and drinking too much?
> 
> The abstract does say "independent of other established CHD risk factors,"

Ok, there are two other factors which I think are relevant with
respect to my hypothesis... 1) BMI (i.e. being overweight), and
2) alcohol consumption.  

Now 1) is, I believe, considered an "established risk factor".
But what do they mean by independent?  At most that they 
corrected for the expected increase in CHD risk resulting
from excessive BMI.  But that says nothing about whether or
not BMI and SFA consumption interact to increase CHD risk.
To oversimplify a bit, they may have corrected for an additive
effect, but I'm postulating a multiplicative effect, so that
if excess BMI = 0 multiplied by SFA consumption 50% results
in 0% increase CHD risk, but if excess BMI = 1... you get
the idea.

And, I could be wrong, but I think that 2) is NOT considered
an established CHD risk factor... even though anecdotally we
"know" that people who drink too much are more likely to die
of heart disease, as far as I know this has not been
scientifically established.

The problem with almost all scientific studies of the health
effects of diet is that there are so many complex interactions
in diet that it's extremely difficult to isolate variables to
a useful extent.  For example, if as I think, no major studies
have shown increased CHD risk from alcohol consumption, it is
could be because the control subjects were drinking enough sugar
water (sodas) to reach virtually the same levels of high-glycemic
carbs in their diet, or because the test subjects were advised
to limit their SFA intake in order to isolate that "established
risk factor".

And the reality is much more complex than that still... we 
already know that certain foods can be very healthful or very
harmful depending on whether they're accompanied by certain
co-factors, i.e. extra amounts of a certain vitamin, etc.
It is extremely difficult to control for all these factors,
and impossible if you are starting with bad assumptions.
And so far nutritional science is still full of bad 
assumptions as we all know.

:j