Thanks for posting this; it's very interesting, particularly since I've
a husband I'd like to keep alive and healthy:) I've only looked at the
abstract because I'm too mean to subscribe to Wiley, but the
significant bit seems to be:

"Overall, there was a moderate nonsignificant reduction in prostate
cancer rate associated with the supplementation (hazard ratio = 0.88;
95% CI = 0.60-1.29). However, the effect differed significantly between
men with normal baseline PSA (< 3 microg/L) and those with elevated PSA
(p = 0.009). Among men with normal PSA, there was a marked
statistically significant reduction in the rate of prostate cancer for
men receiving the supplements (hazard ratio = 0.52; 95% CI =
0.29-0.92). In men with elevated PSA at baseline, the supplementation
was associated with an increased incidence of prostate cancer of
borderline statistical significance (hazard ratio = 1.54; 95% CI =
0.87-2.72).

It seems to say that the supplementation is good if you are not at high
risk for prostate cancer, but marginally increases the risk if you are
already at higher risk.

The article you posted also seems to imply that the supplementation,
together with vitamin C supplementation, or at least increased serum
levels of vitamin C, would also increase the risk (marginally).

If you have read the whole paper, I'd be interested to know if that's
the case.

The article you posted also stated that:
"Supplementation was associated with a 48% reduction in the incidence
of
prostate cancer among men with a baseline PSA level below 3
micrograms/L, the authors report in the August 20th International
Journal of Cancer.

In contrast, men with a baseline PSA level of 3 micrograms/L or greater

who took supplements experienced a 54% increase in the incidence of
prostate cancer."

I've read elsewhere that the normal level of PSA is up to 4
micrograms/L rather than the 3 quoted in the abstract, which would be
significant for those on the borderline.

Cheers,

Cathy

Reading these studies is so frustrating.  I skimmed the full
article.  I'm not a statistician or a doctor but it struck me
that the statistics would be extraordinarily difficult to
interpret even if you were a statistician or a doctor.

If I understood it, five supplements were used.  We don't
know if one of those supplements accounted for all of the
benefit or all of the damage, or if it was some combination,
or if changes in the dosage would have led to a different
outcome.

Mean levels of five measures of PCa: PSA, Free PSA, % free
PSA, IGF-I and IGF-II were taken before and after the trials
and the _mean_ levels did not differ between the treatment
and placebo arms.  However (again, if I am reading this
correctly), the number of men who came out of the trial
with a PSA > 4.0 and who had subsequent biopsies showing
cancer, was statistically greater in the treatment arm
for men with an initial PSA > 3.0 while being stastistically
less for the others.

Do we know what all this means?

It seems to me that statistical studies like this can
sometimes produce useful information.  It may even be that
this particular study is useful to the right scientist who
can make sense of it or put it in context with other
studies.

However what we really need is a testable theory about
the actual operation of the supplements.  What are their
actual physiological and metabolical effects?  Until then
we're throwing rubber balls against an object hidden in
fog and trying to guess what the object looks like by
the angles of the bounces.  There could be a wall in the
fog, or there could be an elephant, or perhaps there's a
sofa.  Guessing based on the bounces is pretty hard.

   Alan