This looks like the typical meta-analysis, which looks at
the PUBLISHED randomized trials.  Any competent statistician
can tell you what is wrong with that.

This isn't my posting but I thought was
worth passing on as there are separate threads
on the same topic in different Usenet forums.
Geoff's comments are on target with references.
His posting is long
If I had my way, his posting would be picked up
by the wire services and given the same play as
the previous press articles on the Lancet "paper".

The quote is below
-------------------------------------------------

"geoff" <shaman7uk@yahoo.com> wrote in message
In my opinion this is one of the most biased and
unsubstantiated reports on antioxidants I've ever
read. If you look at the actual results of this
supposed comprehensive analysis of research you will
see that the only really significant finding in a
considerable reduction in gastrointestinal cancer risk
with selenium supplementation. Overall, it shows that
antioxidant supplements reduce the risk of oesophageal
cancer, have little effect on pancreatic or
oesophageal cancer, and slightly increase the risk of
gastric cancer. Overall, the clear trend is towards
protection, not harm. I believe this is an
underestimation of the prevention power of
antioxidants because this claimed comprehensive
analysis of research excludes some very well designed
positive studies, such as a trial of 864 people with a
history of colorectal adenomas, by the National Cancer
Institute (1). The participants were given either 25mg
of betacarotene and/or both 100mg of vitamin C and
400mg of vitamin E, versus placebo. While there was
approximately a halving of recurrence of colorectal
adenomas in those who took either the betacarotene or
vitamin C and E or both, there was a modest increase
in cancer recurrence among those who only took
betacarotene supplements and both smoked and drank
alcohol every day.  Why was this trial excluded?
Perhaps it didn't give the results the researchers
wanted.

The final table in the Lancet study, which is the only
one showing a small negative overall effect on
mortality (the difference between 1 in 14 cancer
patients on antioxidants, versus 1 in 15 cancer
patients), was arrived at by removing any positive
studies on the grounds of 'low methodological
quality', leaving only 7 studies out of the original
167 studies! Of these studies, one is quoted as
showing a massive increased risk. Without this study
there is no such effect.  However, this study actual
showed the exact opposite. The study in question,
Correa et al (2), published in the Journal of the
National Cancer Institute, gave people with gastric
cancer either beta-carotene, vitamin C or
anti-Helicobacter Pylori treatment (gastric cancer is
increasingly being thought to be initiated by H.Pylori
infection, not antioxidant deficiency). All three
interventions produced highly significantly
improvements, causing substantial regression of
gastric cancer. The authors conclude "dietary
supplementation with antioxidant micronutrients may
interfere with the precancerous process, mostly by
increasing the rate of regression of cancer precursor
lesions, and may be an effective strategy to prevent
gastric carcinoma." (see abstract below).

So, how could this study bias the results towards
increased mortality? For the simple reason that six
people out of 368 treated with antioxidants died, many
of whom were smokers, compared to none out of 117
people treated with anti- H.Pylori treatment died! The
most logical explanation for this finding is that, by
virtue of participating in this trial, these patients
were excluded from taking anti- H.Pylori treatment,
which is highly recommended for gastric cancer. It is
highly unlikely that the antioxidants had anything to
do with it. The authors of this study make no
reference to the possibility of antioxidants
increasing mortality risk, instead concluding that
both beta-carotene and vitamin C reduce risk.

A review of the Lancet study (also published in the
Lancet) by David Forman and Douglas Altman of the
Centre for Epidemiology and Biostatistics says "The
mortality analysis in this review does not offer
convincing proof of hazard." In my opinion this is the
most atrocious piece of biased number crunching, and
I'm surprised that the Lancet published it. The
funding source for this trial should be seriously
investigated, just to check it is not as biased as the
rhetoric. Drug companies have a lot to gain by
discrediting nutritional treatments and I have no
doubt that there is an orchestrated campaign under way
to do just this. I certainly won't be stopping my
daily antioxidant supplement, although I wouldn't
advise heavy smokers to supplement beta-carotene on
its own. I would advise people wanting to reduce their
cancer risk to supplement 50 to 150mcg of selenium,
together with other antioxidant nutrients."

Patrick Holford.

References

1 Baron, J et al., 'Neoplastic and antineoplastic
effects of beta-carotene on volorectal adenoma', J
Natl  Cancer Inst. 95, 10, pp. 717-22 (2003).

2 Correa P et al., 'Chemoprevention of gastric
dysplasia:randomised trial of antioxidant supplements
and anti-helicobacter pylori therapapy', J Natl Cancer
Inst. 2000 Dec 6;92(23):1881-8.

ABSTRACT OF THE CRITICAL STUDY

J Natl Cancer Inst. 2000 Dec 6;92(23):1881-8.
Chemoprevention of gastric dysplasia: randomized trial
of antioxidant supplements and anti-helicobacter
pylori therapy.
Correa P, Fontham ET, Bravo JC, Bravo LE, Ruiz B,
Zarama G, Realpe JL, Malcom GT, Li D, Johnson WD, Mera
R.Department of Pathology, Louisiana State University
Health Sciences Center, New Orleans, LA 70112-1393,
USA.

BACKGROUND: Previous research has identified a high
risk of gastric carcinoma as well as a high prevalence
of cancer precursor lesions in rural populations
living in the province of Narino, Colombia, in the
Andes Mountains. METHODS: A randomized, controlled
chemoprevention trial was conducted in subjects with
confirmed histologic diagnoses of multifocal
nonmetaplastic atrophy and/or intestinal metaplasia,
two precancerous lesions. Individuals were assigned to
receive anti-Helicobacter pylori triple therapy and/or
dietary supplementation with ascorbic acid,
beta-carotene, or their corresponding placebos.
Gastric biopsy specimens taken at baseline were
compared with those taken at 72 months. Relative risks
of progression, no change, and regression from
multifocal nonmetaplastic atrophy and intestinal
metaplasia were analyzed with multivariate polytomous
logistic regression models to estimate treatment
effects. All statistical tests were two-sided.
RESULTS: All three basic interventions resulted in
statistically significant increases in the rates of
regression: Relative risks were 4.8 (95% confidence
interval [CI] = 1.6-14.2) for anti-H. pylori
treatment, 5. 1 (95% CI = 1.7-15.0) for beta-carotene
treatment, and 5.0 (95% CI = 1.7-14.4) for ascorbic
acid treatment in subjects with atrophy. Corresponding
relative risks of regression in subjects with
intestinal metaplasia were 3.1 (95% CI = 1.0-9.3), 3.4
(95% CI = 1.1-9.8), and 3.3 (95% CI = 1.1-9.5).
Combinations of treatments did not statistically
significantly increase the regression rates. Curing
the H. pylori infection (which occurred in 74% of the
treated subjects) produced a marked and statistically
significant increase in the rate of regression of the
precursor lesions (relative risks = 8.7 [95% CI =
2.7-28.2] for subjects with atrophy and 5.4 [95% CI =
1.7-17.6] for subjects with intestinal metaplasia).
CONCLUSIONS: In the very high-risk population studied,
effective anti-H. pylori treatment and dietary
supplementation with antioxidant micronutrients may
interfere with the precancerous process, mostly by
increasing the rate of regression of cancer precursor
lesions, and may be an effective strategy to prevent
gastric carcinoma

 1. Baron, J et al., 'Neoplastic and antineoplastic
effects of beta-carotene on volorectal adenoma',
Journal of the National  Cancer Institute 95, 10, pp.
717-22 (2003).