Talk:Antioxidant

From Wikipedia, the free encyclopedia

	Antioxidant is a featured article; it (or a previous version of it) has been identified as one of the best articles produced by the Wikipedia community. If you can update or improve it, please do.
Article Milestones Date Process Result October 19, 2006 Peer review Reviewed March 9, 2007 Featured article candidate Promoted Current Status: Featured article

	MCB Portal This article is within the scope of the Molecular and Cellular Biology WikiProject. To participate, visit the WikiProject for more information. The current monthly improvement drive is Signal transduction.
FA	This article has been rated as FA-Class on the assessment scale.
Mid	This article is on a subject of mid-importance within molecular and cellular biology.
Article Grading: The article has been rated for quality and/or importance but has no comments yet. If appropriate, please review the article and then leave comments here to identify the strengths and weaknesses of the article and what work it will need.

	This article is within the scope of WikiProject Medicine. Please visit the project page for details or ask questions at the doctor's mess.
FA	rated as FA-Class on the assessment scale
High	rated as high-importance on the assessment scale

	Antioxidant is part of WikiProject Pharmacology, a project to improve all Pharmacology-related articles. If you would like to help improve this and other pharmacology articles, please join the project. All interested editors are welcome.
FA	This article has been rated as FA-Class on the quality scale.

The following users are actively contributing to this topic and may be able to help with questions about verification and sources: TimVickers (talk • contribs • email)

Archives September 2005 - March 2007

Contents 1 JAMA citation 2 Work towards consensus within guidelines

[edit] JAMA citation

These harmful effects may also be seen in non-smokers, as a recent meta-analysis including data from approximately 180,000 patients showed that β-carotene, vitamin A or vitamin E supplementation is associated with increased mortality.[132] However, this meta-analysis found no significant effect from vitamin C supplementation on mortality.

This sites the JAMA study, which has been criticized ( http://oregonstate.edu/dept/ncs/newsarch/2007/Feb07/vitaminstudy.html ) for methodological problems and a study selection bias. The statement should probably be softened to reflect this criticism (both from LPI and others). This is doubly true given how recently that particular report was published. --Nachtrabe 21:40, 20 March 2007 (UTC)

This criticism has not been published in a peer-reviewed journal, so I think it carries comparatively little weight. If there are indeed serious methodological problems of this paper then when the journal publishes corrections or commentary papers they should certainly be added to the article. However, criticisms that appear to have been published on the researcher's university news website do not really pass the quality guidelines. TimVickers 00:50, 21 March 2007 (UTC)

Given that the paper in question was published three weeks ago, making authoritative statements from it when there are potential methodological problems is a bit overzealous. Research is an ongoing process, and IMHO the study should be considered valid only after there has been some time for the peer review process to go into effect. This is particularly true when so many doubts have been raised by different groups. From NPI Center: International Experts Dispute Conclusions Of Antioxidant Review, Meta-Analysis On Antioxidants Provides Muddled Conclusions. --Nachtrabe 19:31, 21 March 2007 (UTC)

Peer-review occurs before papers are published and I do not find it surprising that the "International Alliance of Dietary Supplement-Food Associations" is critical of a study suggesting many of their products are harmful. This review is a serious study published in a top-quality journal, it's conclusions deserve to be highlighted in this article. Moreover, it is not like this is a result out of line with other recent studies. The last major meta-analysis previous to this one (Link) that looked at people with colon cancer also saw an increase in mortality. TimVickers 19:48, 21 March 2007 (UTC)

Here's some criticism of the flawed study, including a lot references: Another Flawed Attack Against Antioxidants Regulations 18:01, 23 March 2007 (UTC) The average age of the subjects was 62! You have to use antioxidants before the damage is done. Preventing damage is easier than reversing it. Regulations 18:06, 23 March 2007 (UTC)

This is not a peer-reviewed study, so it can't be given as much weight as higher quality publications. See the quality guidelines. TimVickers 18:27, 23 March 2007 (UTC)

It's not a study at all. I'm not asking that that article be cited. I am saying that it has a lot of references in it that can be used in this article. It's got links to scientists saying that the study was flawed. Their opinions are noteworthy on whether a study is flawed or not. You don't need a "peer-reviewed study" claiming that experiment is flawed. But more importantly, the article is chocked full of complete citation references to studies that can be cited in this article that prove the benefit of antioxidants. Regulations 18:47, 23 March 2007 (UTC)

I don't think we should be trying to "prove" anything in this article, that is completely the wrong approach. What we should try to do is lay out as clearly as possible the data on the health effects of antioxidants. I would be very happy for you to expand the section on "Disease prevention" with high-quality peer-reviewed articles. TimVickers 18:53, 23 March 2007 (UTC)

Read closer. I didn't say cite studies "to prove." I said cite the studies "that prove." Regulations 18:57, 23 March 2007 (UTC)

[edit] Work towards consensus within guidelines

OK, people do not all agree with this peer-reviewed meta-analysis that has published in a major medical journal. The two guidelines that apply here are WP:Attribution and WP:Undue weight. I see the section below as giving undue weight to non-peer reviewed opinion articles over a major piece of high-quality research (as defined by the Attribution guidelines). However, these opinions should be noted, so what about the revised formulation below as a compromise? TimVickers 18:37, 23 March 2007 (UTC)

Original

These harmful effects may also be seen in non-smokers, as a recent meta-analysis including data from approximately 180,000 patients showed that β-carotene, vitamin A or vitamin E supplementation is associated with increased mortality.[132] However, this meta-analysis found no significant effect from vitamin C supplementation on mortality. This study has been criticized as being flawed. Jeffrey Blumberg, Director of the Antioxidants Research Laboratory at Tufts University in Boston, Massachusetts, says: "One of the major premises of doing such a meta-analysis is that the studies should be comparable…here, they looked at primary prevention, treatment, old people, young people, smokers, nonsmokers. Only when they used their own criteria of what was good and what was bad were they able to show an increase in all-cause mortality."^[1] Meir Stampfer, Professor of Nutrition and Epidemiology at the Harvard School of Public Health says: "This study does not advance our understanding, and could easily lead to misinterpretation of the data."^[2]

Compromise wording

These harmful effects may also be seen in non-smokers, as a recent meta-analysis including data from approximately 180,000 patients showed that β-carotene, vitamin A or vitamin E supplementation is associated with increased mortality.[132] However, this meta-analysis found no significant effect from vitamin C supplementation on mortality. The conclusions and methodology of this study have been controversial,^[3][4] but its conclusions are consistent with a previous meta-analysis that looked at the effects of antioxidants on people with colon cancer.(Link)

It's not only the conclusions that are controversial, but the methodology of the study. It was a flawed study. And, respected scientists are asserting that it's a flawed study. Those people should be named. Regulations 18:50, 23 March 2007 (UTC)

If you read the guidelines, you will see that it would be inappropriate to give non-peer reviewed opinions more weight than a piece of peer-reviewed research that is of high quality (as defined by Wikipedia). Taking your comments into account, I have revised the proposed new version. Comments? TimVickers 18:59, 23 March 2007 (UTC)

You're still not addressing the fact that scientists are claiming the study was flawed. Noting that they claim this is not giving them more weight than the peer-reviewed research. What makes you think it's giving them more weight? Those scientists who claim this are their peers too, you know? Regulations 19:33, 23 March 2007 (UTC)

Notice that you're giving more weight to that study than it deserves, because there is a whole list of studies that disagree, as can be found in that article I showed you. The weight of the evidence is in favor of anti-oxidants. Regulations 19:45, 23 March 2007 (UTC)

As I'm sure you are aware, the function of meta-analyses is to look at all the clinical trials in an area and bring this together into an overall picture. If you go to [PubMed] and search for "Antioxidants" with the limit set for "type of article - Meta-analysis" you get a list of all the recent meta-analyses on this subject. Of these analyses, none saw protective effects of dietary antioxidants and several, as this section of the Wikipedia article comments, saw harmful effects. This is the data. To reflect this accurately in one sentence - "There is no reliable evidence that antioxidant supplementation is benificial in healthy adults, and some evidence that it is harmful."

As to the question of undue weight the first sentence of this policy reads "NPOV says that the article should fairly represent all significant viewpoints that have been published by a verifiable source, and should do so in proportion to the prominence of each." Meta-analyses in peer-reviewed scientific journals are the highest-quality source on the health effects of drugs there is, therefore we should give these far more weight than the opinion articles on news websites and dietary supplement industry press releases that you have cited. Please understand that this is not my personal opinion, I am just trying to reflect the policy of Wikipedia in this article's discussion of the various sources on this topic. TimVickers 21:03, 23 March 2007 (UTC)

For more information from non-nutracutical industry sources. See the following links:

• The Cochrane Collaboration
• JAMA News release
• MedScape Medical News

I've added a note to the discussion of this review pointing out that the effect they detected was not seen in the French SU.VI.MAX trial. This might deal with our problem over giving undue weight to opinion articles, as this trial was peer-reviewed and discussed in a previous section. TimVickers 01:16, 28 March 2007 (UTC)

After reading this Discussion first, I revised the paragraph on the JAMA meta-analysis to include a summary by the editor of Nutraceuticals World. Although an industry publication, this online news magazine is recognized as a high standard and presents a reasonable view. The content of my revision was simply to point out the JAMA report's obvious weaknesses which are apparent to anyone -- clinical scientist, statistician or layperson. My revision follows between the lines:

___ These harmful effects may also be seen in non-smokers, as a recent meta-analysis including data from 68 separate clinical trials of 232,606 patients showed that β-carotene, vitamin A or vitamin E supplementation was associated with increased mortality^[5], However, this meta-analysis found no significant effect from vitamin C supplementation on mortality, and its conclusions have been questioned due to the wide heterogeneity of patients already ill with varied diseases studied in different trial designs, treatment dosages and durations^[6]. ___

I think this is a fair view. Based on the discussion above -- which shows good debate in need of a conclusion -- my revision is a reasonable compromise, neither refuting the higher-quality JAMA citation nor being too strong in opposition with this one.

Also, I made other syntax changes that section needs, and feel they should be restored. --Paul144 02:04, 28 March 2007 (UTC)

I'm not sure that's a fair assessment and it is a non-peer reviewed article with an obvious bias problem. From the paper itself - The present review follows the Cochrane Collaboration method and is based on the principles of our peer-reviewed protocol and review on antioxidant supplements for gastrointestinal cancer prevention. We included all primary and secondary prevention trials in adults randomized to receive beta carotene, vitamin A, vitamin C, vitamin E, or selenium vs placebo or no intervention. Parallel-group randomized trials and the first period of crossover randomized trials were included.17 Trials including general or healthy populations were classified as primary prevention. Trials including participants with specific disease were classified as secondary prevention. We excluded tertiary prevention (treatment) trials, like trials on acute, infectious, or malignant diseases except nonmelanoma skin cancer."

The meta-analyses seem evenly split between seeing no effect on mortality and seeing a harmful effect, as I noted above. Look for example at PMID: 16842454, PMID: 12804424 PMID: 12814711 and PMID: 15537682. This section does need to be rewritten and expanded, but these other meta-analyses should also be added. TimVickers 02:29, 28 March 2007 (UTC)

How about:

"These harmful effects may also be seen in non-smokers, as a recent meta-analysis including data from approximately 180,000 patients showed that β-carotene, vitamin A or vitamin E supplementation is associated with increased mortality but saw no significant effect from vitamin C.^[7] These results are consistent with some previous meta-analyses that suggested that Vitamin E supplementation increased mortality, PMID: 15537682 and that antioxidant supplements increased the risk of colon cancer.PMID: 16842454 However, the results of this meta-analysis are inconsistent with other studies such as the SU.VI.MAX trial, which suggested that antioxidants have no effect on cause-all mortality,^[8]PMID: 12804424"

That's better, but I feel fair-minded doubt has to be stated about the original meta-analysis in JAMA conducted on such a diversity of clinical trial groups with different experimental designs. The statement I used should be considered: "its conclusions have been questioned due to the wide heterogeneity of patients already ill with varied diseases studied in different trial designs, treatment dosages and durations".

The JAMA paper states 68 randomized trials with 232,606 participants (385 publications), a set of data making conspicuous the variability inherent in this meta-analysis, even raising questions whether it was a valid meta-analysis.

A premise in meta-analysis (Wikipedia definition) is that it "combines the result of several studies that address a set of related research hypotheses", allowing one to conclude then that JAMA report was inevitably flawed. Among many variables, the "related hypotheses" are only those addressing antioxidants with mortality, whereas "the wide heterogeneity of patients already ill with varied diseases studied in different trial designs, treatment dosages and durations" describes confounding multiple variables making the JAMA report's conclusions doubtful.

Not acknowledging this interpretation dismisses the report's obvious weaknesses, perhaps misleading visitors to this Wikipedia page in search of an alternate interpretation.

We should not be so strict that a publication in JAMA is irreproachable. The industry voices critiquing the JAMA study have been reasoned and their facts are worthy to state in the Antioxidant Article that its design and conclusions are questionable. --Paul144 03:59, 28 March 2007 (UTC)

No, I don't agree. This JAMA paper is certainly not "obviously flawed". A meta-analysis takes data from multiple trials and re-analyses it, either to test the same hypothesis addressed in the original studies, or a new hypothesis. The method followed in this JAMA paper is a standard in the field and are described here. Describing the Cochrane protocols for meta-analyses as "obviously flawed" is simply incorrect. I am happy to put this finding in context with other equally-valid meta-analyses (following similar protocols). Perhaps if you found some meta-analyses that demonstrated positive results of antioxidants this would be a path out of this impasse? If such positive studies exist, they would balance the negative findings. As the Wikipedia WP:Undue weight policy states.

I think you're in the minority defending the JAMA report as unflawed. The key point is that, while appropriate meta-analysis methods were applied according to the Cochrane protocols, the choice of what disease groups to include were obviously not -- the authors grouped 1) unrelated diseases, 2) patients who were both already morbid with their illnesses and those that were healthier, 3) a wide range of doses of antioxidants, 4) a wide range of study durations, etc. This is poor meta-analysis design and is flawed.

You seem entrenched and are not finding a collaborative, balanced position. I think you should remove yourself from this discussion and allow the wiki process of checks and balances for revision of the article. --Paul144 08:58, 28 March 2007 (UTC)

Disagreements over whether something is approached the Neutral Point Of View (NPOV) way can usually be avoided through the practice of good research. Facts (as defined in the A simple formulation section above) are not Points Of View (POV, here used in the meaning of "opposite of NPOV") in and of themselves. A good way to build a neutral point of view is to find a reputable source for the piece of information you want to add to Wikipedia, and then cite that source.

Industry publications are not reliable sources. Adding non-peer reviewed criticisms from people with an obvious bias to this article is deeply unwise and directly contradicts Wikipedia policy - please read WP:Undue weight. TimVickers 04:41, 28 March 2007 (UTC)

I don't agree with your method of excluding an industry source more because it's not from a typical scientific method published in a peer-reviewed journal, rather than listening to the reasonable objections raised to the JAMA article within the publication like the Nutraceuticals World editorial (which would have been peer-reviewed).

Here is one considered position opposing the JAMA report[1] and another from several individuals simply voicing, as I am, that the JAMA study was not the best-designed meta-analysis, indicating that its conclusions should be tempered with other points of view, including those of Dr. Alex Schauss, a scientist who publishes in peer-reviewed literature[2].

If you can't see this, Tim, then I propose we remove the section about the JAMA study in the Antioxidant Article all together, and wait a few months for further analysis of its validity and additional counterpoints to be presented. --Paul144 08:58, 28 March 2007 (UTC)

I probably appear intransigent here as I am bound (as are you) by Wikipedia's policy. To quote directly from the relevant section "NPOV says that the article should fairly represent all significant viewpoints that have been published by a verifiable source, and should do so in proportion to the prominence of each." If Nutracuticals World or NPIcenter are indeed peer-reviewed journals, what is their peer-review process? are they listed in Medline? what is their impact factor? I will be happy to add material from these sources if they are indeed of high quality. However, the onus is on you to show that these sources pass WP:Attribution. This problem would be very easily resolved if you could find some peer-reviewed meta-analyses or randomised double-blind clinical trials that demonstrated positive health effects for antioxidant supplementation. As the policy says "Disagreements over whether something is approached the Neutral Point Of View (NPOV) way can usually be avoided through the practice of good research." TimVickers 15:41, 28 March 2007 (UTC)

To try to help us come to agreement here, I've asked for some input from the Wikipedia:Mediation Cabal as well as requesting some other opinions at the Wikiproject Medicine and Wikiproject Pharmacology. TimVickers 16:02, 28 March 2007 (UTC)

• I have to agree with TimVickers here—we must insist upon the highest-quality peer reviewed sources in medical articles. If there are significant problems with the JAMA article, that will eventually be published in a peer-reviewed journal; we should err on the side of accuracy over a rush to publish, and we must uphold the highest-quality reliable sources. SandyGeorgia (Talk) 16:05, 28 March 2007 (UTC)

It should be noted that, even discounting the potentially serious methodological flaws, in the words of the aforementioned JAMA analysis "The pooled effect of all supplements vs. placebo or no intervention in all randomized trials was not significant." They did further restriction (throwing out an additional 21 trials for being "high bias" through what they admit is a process with some risk of error) before coming to the final conclusion was reached. --71.218.232.238 20:14, 6 April 2007 (UTC)