Free Radicals Bad! Antioxidants Good! Not. (was Re: Body Saturation with Vitamins (was Re: Green Tea)
Steve Harris
news:T4Qb8.574$aM2.57...@newssvr21.news.prodigy.com...
> Steve Harris wrote:
> > COMMENT
> > Well, pick one out. Your reason for taking any of these more than once a
> > day is?
>
> I assume you are only speaking of vitamins and minerals. Not plant based
> extracts, medications, or other nutrients that are found in the body.
> Otherwise you would be quite mistaken, doc.
>
> One very obvious reason to take some multiple times is to increase the
> amount absorbed in order to save money.
COMMENT:
These are therapeutic uses, however-- you're using them basically like
drugs. Or else you're fiddling around with something really inefficient (a
lycopene capsule nerd who could instead be having a marinara dish with olive
oil).
> Animals don't abuse their health, oh wise one. They live in a static,
> stable environment with a healthy diet.
COMMENT:
So they do. And it's a great hypothesis that taking vitamins will protect
you from all those nasty things you do to your body. Alas, the evidence is
lacking. Atherosclerosis and Alzheimer's may be the two most interesting
topics in nutritional prevention, and the animal models for both are lousy
(for either expense or mechanistic reasons).
> >> Simple! To keep your blood levels of vitamin c within a certain
> >> quantitative range of possible values.
> >
> > COMMENT:
> > Which you want to do, why? You're treating a number, you fool. There
are
> > no prospective trials of vitamin C in humans for any chronic disease, so
> > how could you possibly know what any given number means?
>
> Was one of your parents a troll? You know damn well how damaging free
> radicals are and it is basic common sense that loading up with a lot of
> antioxidants will protect the body from some of this harm.
COMMENT:
This is no troll. I'm quite serious. Ever since Pearson and Shaw
popularized the idea that free radicals were The Cause of All Disease and
Aging in the early 80's, we've been on some kind of crazy bandwagon that
free radicals are entirely bad, and the less of them you have, the better
off you are. Well, it isn't true. For example a free-radical molecule called
nitric oxide is used by the body as a major signal molecule for all kinds of
things, including regulating blood flow. And that's not all-- this and other
radicals are utilized by the body to kill bacteria, and to activate the
inflammatory response. Which itself is a double-edged sword. Inflammation
isn't all bad; nature gave it to you for a reason. Besides being necessary
to fight bugs, inflammation activates healing and tissue growth as well.
The more you abuse your body, the MORE you need it, in some ways. Finding
the right balance is one of the key balancing acts that faces doctors every
day in the ICU with patients who have septic shock (SIRS), ARDS,
resuscitation encephalopathy, and so and so on. And in outpatient medicine
as well. The COX-2 inhibitor which might be preventing your Alzheimer's
disease is also preventing healing of ulcers in your stomach. And so on.
Time for an illustration. You might suppose that since antioxidants are
good for arteriosclerosis (everybody knows that, right), that adding a mix
of them to heart disease patients who have lipid problems would be good for
them, and keep them from having as many heart attacks. Well, if the lipid
problems are low HDLs, the answer is that antioxidants don't do that much
good. Statin-plus-niacin works far better. And if you should hedge by adding
antioxidants TO the statin-plus-niacin, you cut the beneficial effect in
half. Yep. Same study, beautifully controlled.
Why is this?? We don't know. My own guess is that HDL is partly made in
response to inflammation in the liver CAUSED by the near overdose levels of
nicotinic acid necessary to induce it, and if you damp down that
inflammatory response too much, you damp out niacin's ability to raise HDL.
Wups. But whatever the reason turns out to be, the point still stands. I
post the abstract below to illustrate the idea that you can NOT just
cookbook your way out by dumping antioxidant vitamins on every medical
problem.
N Engl J Med 2001 Nov 29;345(22):1583-92 (Comment in:N Engl J Med. 2001 Nov
29;345(22):1636-7.)
Simvastatin and niacin, antioxidant vitamins, or the combination for the
prevention of coronary disease.
Brown BG, Zhao XQ, Chait A, Fisher LD, Cheung MC, Morse JS, Dowdy AA, Marino
EK, Bolson EL, Alaupovic P, Frohlich J, Albers JJ.
Department of Medicine, University of Washington School of Medicine, Seattle
98195, USA.
BACKGROUND: Both lipid-modifying therapy and antioxidant vitamins are
thought to have benefit in patients with coronary disease. We studied
simvastatin-niacin and antioxidant-vitamin therapy, alone and together, for
cardiovascular protection in patients with coronary disease and low plasma
levels of HDL. METHODS: In a three-year, double-blind trial, 160 patients
with coronary disease, low HDL cholesterol levels and normal LDL cholesterol
levels were randomly assigned to receive one of four regimens: simvastatin
plus niacin, vitamins, simvastatin-niacin plus antioxidants; or placebos.
The end points were arteriographic evidence of a change in coronary stenosis
and the occurrence of a first cardiovascular event (death, myocardial
infarction, stroke, or revascularization). RESULTS: The mean levels of LDL
and HDL cholesterol were unaltered in the antioxidant group and the placebo
group; these levels changed substantially (by -42 percent and +26 percent,
respectively) in the simvastatin-niacin group. The protective increase in
HDL2 with simvastatin plus niacin was attenuated by concurrent therapy with
antioxidants. The average stenosis progressed by 3.9 percent with placebos,
1.8 percent with antioxidants (P=0.16 for the comparison with the placebo
group), and 0.7 percent with simvastatin-niacin plus antioxidants (P=0.004)
and regressed by 0.4 percent with simvastatin-niacin alone (P<0.001). The
frequency of the clinical end point was 24 percent with placebos; 3 percent
with simvastatin-niacin alone; 21 percent in the antioxidant-therapy group;
and 14 percent in the simvastatin-niacin-plus-antioxidants group.
CONCLUSIONS: Simvastatin plus niacin provides marked clinical and
angiographically measurable benefits in patients with coronary disease and
low HDL levels. The use of antioxidant vitamins in this setting must be
questioned.
Publication Types:
Clinical Trial
Randomized Controlled Trial
> > No research published in JAMA says you should take vitamin C pills more
> > than once a day. It'll be a red letter day when anybody writing in JAMA
> > suggests taking vitamin C pills at all.
>
> Cuz they're all ultra-conservative wimps like you? We need more
theoretical
> nutrition researchers. Like in physics. Not ones who continually spout
crap
> like "There is no published evidence backing up that claim."
COMMENT:
Theoretical nutrition researchers never would have done the trial above,
because they would already have decided the results were in the bag. Oops,
they weren't.
The THEORY is very complicated. If you look at the cellular level at what
vitamin C does, it isn't all nice, and the studies don't all show that the
more of it you have the better off you are. I'll quote just one for
illustration.
Brain Res 2001 Mar 23;895(1-2):66-72 Related Articles, Books, LinkOut
Oxidative stress induced by ascorbate causes neuronal damage in an in vitro
system.
Song JH, Shin SH, Ross GM.
Department of Physiology, Botterell Hall, Queen's University, Kingston,
Ontario, Canada, K7L 3N6.
Of particular physiological interest, ascorbate, the ionized form of
ascorbic acid, possesses strong reducing properties. However, it has been
shown to induce oxidative stress and lead to apoptosis under certain
experimental conditions. Ascorbate in the brain is released during hypoxia,
including stroke, and is subsequently oxidized in plasma. The oxidized
product (dehydroascorbate) is transported into neurons via a glucose
transporter (GLUT) during a reperfusion period. The dehydroascorbate taken
up by cells is reduced to ascorbate by both enzymatic and non-enzymatic
processes, and the ascorbate is stored in cells. This reduction process
causes an oxidative stress, due to coupling of redox reactions, which can
induce cellular damage and trigger apoptosis. Ascorbate treatment decreased
cellular glutathione (GSH) content, and increased the rates of lipid
peroxide production in rat cortical slices. Wortmannin, a specific inhibitor
of phosphatidylinositol (PI)-3-kinase (a key enzyme in GLUT translocation),
prevented the ascorbate induced-decrease of GSH content, and suppressed
ascorbate-induced lipid peroxide production. However, wortmannin was
ineffective in reducing hydrogen peroxide (H(2)O(2))-induced oxidative
stress. The oxidative stress caused ceramide accumulation, which was
proportionally changed with lipid peroxides when the cortical slices were
treated with ascorbate. These differential effects support the hypothesis
that GLUT efficiently transports the dehydroascorbate into neurons, causing
oxidative stress.
PMID: 11259761 [PubMed - indexed for MEDLINE]
Now, this is just cells in a jar, but you were talking about THEORY, weren't
you? How theoretical do you really want to be?
SBH
--
I welcome Email from strangers with the minimal cleverness to fix my address
(it's an open-book test). I strongly recommend recipients of unsolicited
bulk Email ad spam use "http://combat.uxn.com" to get the true corporate
name of the last ISP address on the viewsource header, then forward message
& headers to "abuse@[offendingISP]."
Jay Tanzman
Steve Harris wrote:
> N Engl J Med 2001 Nov 29;345(22):1583-92 (Comment in:N Engl J Med. 2001 Nov
> 29;345(22):1636-7.)
> Simvastatin and niacin, antioxidant vitamins, or the combination for the
> prevention of coronary disease.
>
> Brown BG, Zhao XQ, Chait A, Fisher LD, Cheung MC, Morse JS, Dowdy AA, Marino
> EK, Bolson EL, Alaupovic P, Frohlich J, Albers JJ.
> BACKGROUND: Both lipid-modifying therapy and antioxidant vitamins are
> thought to have benefit in patients with coronary disease. We studied
> simvastatin-niacin and antioxidant-vitamin therapy, alone and together, for
> cardiovascular protection in patients with coronary disease and low plasma
> levels of HDL. METHODS: In a three-year, double-blind trial, 160 patients
> with coronary disease, low HDL cholesterol levels and normal LDL cholesterol
> levels were randomly assigned to receive one of four regimens: simvastatin
> plus niacin, vitamins, simvastatin-niacin plus antioxidants; or placebos. ...
I wouldn't call a single trial with 40 persons per treatment arm conclusive, but
this is certainly interesting.
-Jay
Steve Harris
news:3C7140B9...@sph.llu.edu...
These are groups of 40, which means that any given comparison (trial) of one
treatment vs another contains 80 patients. Sure, the absolute number goes
into the pot, but it's balanced by the size of the effect. That's what we
have p values for, don't you know. An N of 80 is certainly large enough to
have a decent p value. And a randomized blinded controlled prospective study
of 80 is worth any number of retrospective uncontrolled epidemiologic
regressions, no matter how large.
Yes, the finding of the study will need to be confirmed independently before
we take it as gospel. There's always the possiblility of sampling bias, or
error, or fraud.
Jay Tanzman
In a small trial (eg, 40 x 40), the p-value can be small due to confounding. In
a trial of this size, simple randomization cannot be relied on to control for
confounding. (For this reason, incidentally, a small RCT may be more confounded
than a large _prospective_ observational study which has been controlled in the
analysis.) I tried to get the full text article so I could read the details of
the randomization procedure (if they were stated) and the analysis, but, as
usual, the NEJM site doesn't like my computer's cookies, or something, so I'll
have to get the paper tomorrow from the library.
> Yes, the finding of the study will need to be confirmed independently before
> we take it as gospel. There's always the possiblility of sampling bias, or
> error, or fraud.
Yes, basically that is what I was referring to.
-Jay
Steve Harris
> In a small trial (eg, 40 x 40), the p-value can be small due to
confounding.
Comment: yes, but the odds are against it. That is precisely what the p
value tells you: the odds of your accidentally "confounding" (I'm not sure
I'd use this word here) your study by sampling error with respect to causal
agents, when you randomize groups. It doesn't protect against some hidden
bias if you didn't sample blindly, and it doesn't guarantee that your result
is inductively valid across all groups (much less the one you "think" you
sampled). But it does protect you against randomization sampling errors.
That's the point.
>In
> a trial of this size, simple randomization cannot be relied on to control
for
> confounding.
Why not? You can rely on it to the limit of your p value. It's never 100%,
but it's more than 95% and often more than 99%.
(For this reason, incidentally, a small RCT may be more confounded
> than a large _prospective_ observational study which has been controlled
in the
> analysis.)
Why is that? You can never know if you're controlled post-hoc for all the
confounders. On the other hand, you can pretty much guarantee (to the limit
of p) that you've randomized all the ones you don't know about, IF you
randomize. Sample variances should be caused by variables you don't know
about and can't control. That's the whole IDEA of randomizing. I presume
that people publishing in NEJM will have pre-stratified for anything already
strongly known to influence outcome in such a study. And/or will do post-hoc
control tests for things they didn't, but knew about and were suspicious of.
Jay Tanzman
Steve Harris wrote:
>
> "Jay Tanzman" <jtan...@sph.llu.edu> wrote in message
> news:3C7155BF...@sph.llu.edu...
>
> > In a small trial (eg, 40 x 40), the p-value can be small due to
> confounding.
>
> Comment: yes, but the odds are against it. That is precisely what the p
> value tells you: the odds of your accidentally "confounding" (I'm not sure
> I'd use this word here) your study by sampling error with respect to causal
> agents, when you randomize groups. It doesn't protect against some hidden
> bias if you didn't sample blindly, and it doesn't guarantee that your result
> is inductively valid across all groups (much less the one you "think" you
> sampled). But it does protect you against randomization sampling errors.
> That's the point.
Yes, the p-value does protect against this, but I've analyzed datasets from
small trials on more than one occasion where the statistical significance of the
result was dependent on one or two subjects with valid, but extreme outcomes.
In a large trial, this can't happen.
> >In
> > a trial of this size, simple randomization cannot be relied on to control
> for
> > confounding.
>
> Why not? You can rely on it to the limit of your p value. It's never 100%,
> but it's more than 95% and often more than 99%.
>
> (For this reason, incidentally, a small RCT may be more confounded
> > than a large _prospective_ observational study which has been controlled
> in the
> > analysis.)
>
> Why is that? You can never know if you're controlled post-hoc for all the
> confounders. On the other hand, you can pretty much guarantee (to the limit
> of p) that you've randomized all the ones you don't know about, IF you
> randomize.
Randomization can only be relied on to control unknown confounders (or
prognostic factors, as trialists like to say) when the number of these is small
compared with the sample size. That is why, in a small trial, you have to
randomize in blocks to force balance among known or suspected prognostic
factors.
> Sample variances should be caused by variables you don't know
> about and can't control. That's the whole IDEA of randomizing. I presume
> that people publishing in NEJM will have pre-stratified for anything already
> strongly known to influence outcome in such a study. And/or will do post-hoc
> control tests for things they didn't, but knew about and were suspicious of.
I wouldn't assume that, which is why I tried to get the full text.
-Jay
Herman Rubin
>> Steve Harris wrote:
>> -Jay
I am putting on my professional hat. Decision making under
uncertainty is not well served by the use of ancient rules,
now known not to be what they were claimed to be.
I suggest that people in medicine consider giving up their
statistical religion, and start learning what is involved
in decision making under uncertainty. The p value does not
mean what almost all who use it think it means, and from the
standpoint of making a decision, I would not use that aspect
of the data. Even when it comes in, it is not the p value
but its derivative or local difference which is what should
be used.
One can do better with blinded controlled prospective
studies, but they are short term and leave out too much.
However, epidemiological studies almost always ignore lots
of relevant variables. As Oliver Wendell Holmes replied to
a question, the way to live to a ripe old age is to choose
your grandparents carefully. It IS possible to use these,
and to combine them with controlled studies. It is not
easy.
>Yes, the finding of the study will need to be confirmed independently before
>we take it as gospel. There's always the possiblility of sampling bias, or
>error, or fraud.
Or bad luck; this cannot be avoided when randomness is
involved. But from the randomized trials above, the
likelihood function, which contains all the information,
should be provided so that individuals can decide which
course of action to take. Another caveat: even a well
designed contolled experiment cannot be THAT controlled,
and using the other available variables may modify the
conclusions.
Off topic, I tried to email you about another posting,
and the various attempts (plural) at fixing your address
did not work.
>SBH
>--
>I welcome Email from strangers with the minimal cleverness to fix my address
>(it's an open-book test). I strongly recommend recipients of unsolicited
>bulk Email ad spam use "http://combat.uxn.com" to get the true corporate
>name of the last ISP address on the viewsource header, then forward message
>& headers to "abuse@[offendingISP]."
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
hru...@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558
Kev
<< For example a free-radical molecule called
nitric oxide is used by the body as a major signal molecule for all kinds of
things, including regulating blood flow.>>
Nitric oxide is not considered a free radical. In fact, anti-oxidants increase
nitric oxide by quenching superoxide which destroys nitric oxide. Nitric oxide
can be bad if its in the wrong places, however.
Wow, youve managed to find some negative studies on free radicals.
Congratulations, thats hard to do even with the millions of clinical studies
available. You seem to be arguing that anti-oxidants are often bad, and that
free radicals are often good. You are really ignorant, and youre not helping
anyone. You should do a search of the literature and look at how many diseases
are caused by free radical damage and can be prevented and occationally treated
with antioxidants.
Your one study shows that it was ineffective in one short-term case. Youre
ignoring mountains of positive evidence. Your other study was an in-vitro
study. Its well known that any one anti oxidant can cause the generation of
free radicals in specific ways, thats why its important to take many types of
anti-oxidants to cascade the free radicals down to less potent forms. One type
of anti-oxidant, the spin trap, does not do this and fully traps superoxide
until it is elimated from the body. Do a seach on PBN and see the amazing
things it does.
[[Kev]]