Talk:Metformin/Archive 1

Archive 1

Archive 2

older entries

i think the relative importance of using metformin in the treatment of type 2 diabetes mellitus should be underlined given the results of the united kingdom prospective diabetes study (ukpds). it was the only agent that reduced cardiovascular mortality and morbidity - if i had type 2, i'd be the first in line to want metformin, ahead of sulphonylureas etc.

I totally agree. I'll see if I can find the reference. JFW | T@lk 22:08, 27 October 2005 (UTC)

i think these are the refs:

UKPDS 34 Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes. UKPDS Study Group Lancet (1998); 352: 854-865

UKPDS 33 Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. UKPDS Study Group Lancet (1998); 352: 837-853

Well, the first reference was already there and I've updated it further. JFW | T@lk 00:49, 30 October 2005 (UTC)

have added the word "cardiovascular" in front of complications. pernickety, but in UKPDS, sulphonylureas decreased microvascular but not macrovascular complications.

there is also an argument that metformin is no longer being the "only" drug shown to decrease CVS complications - the PROactive trial showed that pioglitazone reduced death, MI, bypass surger, strokes but not peripheral vascular disease interventions. this has led to some controversy as the trial used a composite primary endpoint (death+MI+CVA+amputation etc) which was not significantly reduced. i therefore wonder if the word "only" should be removed.

Codeine

Despite the claim of Jfdwolff (22:38, 27 October 2005 (sorted things a bit - removed the codeine stuff (does not relieve cramping))), codeine is effective for the relief of cramping caused by metformin. Codeine is indicated for diarrhea and pain relief, and as such it is immediately relevant to the common gastrointestinal side effect of metformin. This is not the place for details on the mechanism of action of codeine/morphine, but in short, the resulting morphine has a relaxing, even paralytic effect, that particularly affects the entire gastrointestinal tract (this is the cause of loss of appetite and constipation in long-term morphine users). Whether it is always effective is anther matter, hence the word "should". Caveats, such as codeine being appropriate only for short-term use, might be warranted, but in my opinion belong under [codeine]. --Anthony Duff 22:41, 4 January 2006 (UTC)

So why codeine and not loperamide, a non-centrally acting opioid constipant? JFW | T@lk 23:12, 4 January 2006 (UTC)

That's a good question to which I have found no certain answer. However, it seems "well-known" that loperamide is not effective for relieving diarrhea caused by metformin, while codeine is. Apparently, the (unknown) mechanism of induction of diarrhea caused by metformin is different than for ordinary diarrhea. A distinguishing feature of metformin-induced diarrhea is that it is associated with strong, painful intestinal cramping that occurs irrespective of what is in the colon. The initial bowel motions may be quite solid. Loperamide apparently works by making the colon insensitive to its contents, and metformin-induced diarrhea seems not to be triggered by the physical contents of the colon.

This, at least, is my best guess. With further investigation, I cannot find anything better. I freely admit that I do not know why loperamide doesn't work, if indeed it doesn't.

If metformin is causing pain, cramping and/or diarrhea, then the standard action is for the metformin dose to be reduced. However, I think this brief mention of codeine is important because it gives short-term, immediate relief to the person who is in actual pain. --Anthony Duff 23:22, 5 January 2006 (UTC)

Metformin and aspirin

Why is metformin perscribed with aspirin even though aspirin is a blood thinner and can cause bleeding? Does metformin have to be taken with another drug, such as aspirin?

Not necessarily, but diabetics are at risk for stroke and heart attack, the risk of which is reduced significantly by a low daily dose of aspirin. JFW | T@lk 16:16, 25 January 2006 (UTC)

The last line states that metformin did not receive FDA approval for type 2 DM in 1994 - a typo surely?

Avandamet

Is that Avandamet stuff really necessary on this page? It's entirely tangential to an article about metformin, and would be better put on a page of its own so as not to distract from, er, people reading about metformin! This may just be the slight beef I have with drug company names invariably appearing in the first few lines of every bleedin' drug article - but really this is irrelevant to what metformin is, and pretty uninteresting to boot.

Anyone want to argue to preserve it?

Nmg20 23:14, 3 August 2006 (UTC)

Contraindications

This section appears to give medical advice (use of "should" and recommendation to patients). Per WP:MEDMOS, could it be rewritten to be more encyclopedic language? Or would the language be weakened to be merely passively reporting on guidelines. Perhaps I'm reading it wrong? Wrt the last sentence, is "any radiographic procedure" really the guideline? Surely dental and x-rays of broken bones are OK? Colin°^Talk 22:44, 12 January 2007 (UTC)

I'll see what I can do about that, although AFAIK overuse of passive voice is somewhat frowned upon. As for that last sentence, it's actually "before any radiographic procedure involving iodinated contrast"; the problem is actually the use of contrast, so no problem with plain films, dental stuff etc. Fvasconcellos 23:25, 12 January 2007 (UTC)

Perhaps my second sentence was ambiguous. I wasn't recommending passive voice, I saw that as a possible negative consequence of rephrasing. I'm not sure how best to phrase it and wondered if fixing one "problem" just introduced another. Colin°^Talk 08:50, 13 January 2007 (UTC)

Ah, sorry about that. I've rephrased the sentences anyway, it does indeed look better (to me at least). Thank you for your edits BTW. Fvasconcellos 12:56, 13 January 2007 (UTC)

Suggestions from GA review

• Try to use more patient-friendly language. I changed some, but other parts remain a little technical.
• metformin will not induce hypoglycemia. could use a reference.

Overall, excellent. TimVickers 01:59, 14 January 2007 (UTC)

Pharmacogenetics

doi:10.1172/JCI30558 - organic cation transporter 1 polymorphisms determine metformin action. doi:10.1172/JCI32133 discuss the implications of this study. JFW | T@lk 19:35, 2 May 2007 (UTC)

Very interesting. Shame I don't have full access—this will probably warrant a mention in the future. Fvasconcellos (t·c) 12:38, 4 May 2007 (UTC)

Apprpropriate addition of important miscellaneous info

Metformin is a commonly prescribed and very important generic drug. As such, a well written drug article is necessary---especially in light of Wiki's new emphasis on a well written range of pharmacologically oriented articles. Regardless, metformin is potentially quite noteworthy due to current scrutiny of it as a possible life-extending agent. I had included a blockof text (below) previously, but it was removed from the article---probably correctly so---because it did not fit it. I am requesting that contributors consider the appropriateness of this information concerning metformin, and how it might be represented. The previously removed block of text I'm referring to follows: "Although any healthful effects, per se, other than its approved use in the treatment of diabetes, life extensionists have shown recent interest in Metformin. According to a 2003 Affymetrix gene chip study, the spectrum of human gene activation/suppression is apparently concordant between the actions of metformin and CR (caloric restriction), the latter being a recognized agent in extending life span in laboratory animals. It is unclear whether or not metformin bestows any such related benefit to humans seeking to extend healthy life span; a so-called grail of discovering a drug agent that would mimic CR has long been sought, however." --Thanks, PLK

Certainly sounds interesting. Can we get a reference for the study, please? So long as it fits wikipedia criteria (published in a respected, peer-reviewed journal), there's a place for this in the article. Nmg20 11:42, 27 May 2007 (UTC)

It does sound interesting, but I'm afraid of introducing undue weight (bear with me here) if we devote too much space to this; metformin's use as an antidiabetic must far exceed its use (use, not utility) as a "life-extension agent". Have there been any clinical implications? Perhaps a few well-placed sentences, if we can cite this properly? Fvasconcellos (t·c) 15:13, 27 May 2007 (UTC)

www.lef.org writes Is Metformin an Anti-Aging Therapy? BioMarker may have already found an anti-aging therapy with its unique technology platform. When BioMarker scientists screened several glucoregulatory agents used to treat diabetes, they found that the undisputed star of the group was metformin, which was twice as effective as the other drugs in reproducing the gene-expression effects of caloric restriction.32

Further evidence that metformin may be an authentic anti-aging therapy is that a similar drug—phenformin—was found in the late 1970s to extend life span in mice by 23%. This study by Russian researchers also demonstrated a fourfold reduction in cancer incidence in the phenformin-treated mice.33 In 2002, scientists at the National Institute on Aging presented data showing that metformin extended the life span of mice by 20%.

BioMarker is currently conducting its own study of metformin in 40 mice, seeking to determine the extent to which metformin can extend life span and reduce the incidence of age-related diseases in these animals. This study is being conducted at the University of California at Riverside under the direction of Dr. Stephen Spindler.

peer reviewed studies on metformin http://www.ncbi.nlm.nih.gov/sites/entrez?Db=pubmed&Cmd=ShowDetailView&TermToSearch=16441842&ordinalpos=4&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum Metformin extends the mean and maximum lifespans of female transgenic HER-2/neu mice by 8% and 13.1% in comparison with control mice. Phenformin, a close chemical relative of metformin, extends lifespan and reduces tumor incidence in C3H mice

If you find the lef.org studies online those refs will be valued

"There goes the neighborhood"

Well, "there goes the neighborhood." Since miscellany and trivia are now officially discouraged on Wikipedia, how long will it be before talk/discussion becomes overly 'mentored'? Metformin has some very interesting and potentially breakthrough features that are currently forbidden to be included in the main article. OK, I understand that the Five Pillars are an important guideline for an encyclopedia. Yet, the charm and the power of Wiki originally was similar to that of the Web itself: instantaneous action, response, as well as the immediate observance of cutting edge information. Let's try to hold onto that power, that unique aspect of Wiki that is wholly unknown to other encyclopedias. Comments---even rather impertinent seeming ones---ought to have a shot at being posted in a discussion with an ultimate goal of considering their eventual or potential worthwhile and fruitful inclusion in Wikipedia articles proper. Let’s retain what is special, and I dare say magical, about Wiki: because chasing away spontaneity will close the box whose opening was and is indeed a magic portal! --Paul Kiesow a.k.a. WikiPaull —Preceding unsigned comment added by 71.59.74.194 (talk) 01:35, 6 September 2007 (UTC)

which of the many drug product names are frequent

I found the list of product names a bit long, especially since it is in such a prominent position of the article. Made a quick search to check which ones are really in common use. Here are the hit counts as of 10/07:

2'250'000	Glucophage
53'800	Riomet
35'900	Fortamet
30'500	Glumetza
16'900	Diabex
16'400	Diaformin
154	Cidophage

According to that order I only left the hits above 30 thousand in bold and deleted product name Cidophage is almost not used.

Jakob Suckale 16:50, 31 October 2007 (UTC)

Interesting approach; I've been meaning to clean up the lead but never get around to it. Fvasconcellos (t·c) 16:55, 31 October 2007 (UTC)

LDL decrease

PMID 17638715: among a lot more interesting information, metformin therapy associated with decrease in LDL levels (and modest decrease in triglycerides). Worth mentioning? Fvasconcellos (t·c) 22:33, 6 November 2007 (UTC)

We also need to qualify the fact that heart failure is probably not a contraindication, as per a recent series of articles in the BMJ. JFW | T@lk 10:26, 7 November 2007 (UTC)

Yes, I had a read at this one. How does this sound?

Heart failure has long been considered a contraindication for metformin use, although a 2007 systematic review showed it to be the only anti-diabetic drug not associated with harm in people with heart failure.

—Fvasconcellos (t·c) 15:33, 7 November 2007 (UTC)

Sounds jolly good. JFW | T@lk 21:16, 11 November 2007 (UTC)

Going in the article now :) Fvasconcellos (t·c) 22:19, 11 November 2007 (UTC)

Further evidence of TSH suppression

PMID 17992605. Fvasconcellos (t·c) 20:55, 11 November 2007 (UTC)

Use in PCOS

NEJM entire article on its use in PCOS. JFW | T@lk 08:38, 6 January 2008 (UTC)

Ooh, that's excellent. I don't suppose I could impose and ask you to flesh out the "Indications" section a bit based on this? ;) Fvasconcellos* (t·c) 14:53, 6 January 2008 (UTC)

I have read about the use of metformin in policystic ovarium sindrome and it semms that metformin increases estrogen levels or decreses testoterona levels is there any evidence that could lead to erectil disfunction or decresed sexual desire? DAZL —Preceding unsigned comment added by 201.172.237.37 (talk) 05:18, 20 March 2008 (UTC)

Try the reference desk. This page is for discussions on the article. I would imagine that in women with PCO, estrogen is lowish and testosterone is raised. Hence, treatment will improve this imbalance. Whether it happens in non-PCO diabetics is probably not documented. JFW | T@lk 08:44, 11 April 2008 (UTC)

I don't think 201.172.237.37 that metformin for PCOS, a condition in women, is ever likely to cause them erectile dysfunction (a condition in men) :-) David Ruben ^Talk 22:53, 16 June 2008 (UTC)

No, I think he was wondering if metformin use in men could cause a decrease in testosterone levels. Fvasconcellos (t·c) 14:56, 1 August 2008 (UTC)

PCOS

Metformin is commonly used in the treatment of PCOS (poly Cystic Ovarian Syndrome) also known a PCOD, PCO, or Stein Leventhral Syndrome. It help patients with condtion because it reduces insulin production , which in turn lowers the testosterone produced by the ovaries. Increased testosterone is usally the main cause of PCOS. —Preceding unsigned comment added by 76.197.15.44 (talk) 19:22, 11 May 2008 (UTC)

I understand this was (or is) an off label use. Do you have any references supporting the popularity, or not, of this use? If I just add your text to the article, it may be seen as drug company promotion. --Hroðulf (or Hrothulf) (Talk) 20:02, 11 May 2008 (UTC)

Fair point - although isn't metformin off-licence now? Anyway, the best evidence is probably PMID 12917943, which is a Cochrane database systematic review from 2002. In addition, as ever it's worth checking the PCOS article, which offers PMID 17984248, and the NEJM article mentioned in the previous comment on this page is good, too. Nmg20 (talk) 07:23, 12 May 2008 (UTC)

Hm - it would also have been worth my checking the state of the article before posting! - the Cochrane review, at least, is already in there, and the popularity question is covered, at least in the UK, by the NICE recommendation (which is also referenced). Nmg20 (talk) 07:26, 12 May 2008 (UTC)

Nmg20, Metformin is not off-licence per se, which would imply nolonger a prescription-only medication. It is though off-patent, meaning that generic manufacturers can now make it at a cheeper price than the original branded product... but all of these still remained licensed medications (by medicine regulatory bodies, eg FDA), with specified license indications for treatment of diabetes and available only on prescription. It is the specific use of Metformin for PCOS that is "unlicensed indication". In UK the British National Formulary, de facto official prescribing guide, not only describes use of Metformin for PCOS but also suggests dosages to be used (whilst the BNF then notes "Metformin doses in the BNF may differ from those in the product literature" - UK doctors would cite the BNF as a responsible body of opinion on safe prescribing practice)David Ruben ^Talk 22:49, 16 June 2008 (UTC)

Yup - apologies, slip of the keyboard! I believe the point I was making - that off-patent drugs are less likely (though by no means impossible) to be the target of drug company promotions - stands. Nmg20 (talk) 09:24, 17 June 2008 (UTC)

That's that sourced from multiple directions. Thanks!

I have a general question. What authority does the BNF have in relation to other UK drug assessors: MHRA and NICE? If its data differs from an MHRA approved label, which do we consider to be our reliable source, or do we cite both and note the difference?

--Hroðulf (or Hrothulf) (Talk) 14:15, 17 June 2008 (UTC)

Good question! My feeling would be that the MHRA are more involved in interacting with the drugs companies than with doctors, and that NICE issues guidelines on (cost-)effective treatment of diseases where the BNF issues guidelines on safe and effective use of drugs - so I guess they're telling us slightly different things? As a first stab, perhaps MHRA information where there'd been legail issues as with seroxat, NICE where there are political issues as with memantine or to give recommended treatment info, and the BNF for permissible ones? Nmg20 (talk) 15:13, 17 June 2008 (UTC)

Do we care about the GFR

After papers that metformin is probably safer than other antihyperglycaemics in congestive heart failure, now these renal physicians from London feel that we can be more lax with the GFR as well, as evidence for harm is poor and the other drugs are actually quite rubbish too: PMID 18368503 JFW | T@lk 19:29, 16 June 2008 (UTC)

You UK docs are a bold lot, aren't you? Fvasconcellos (t·c) 19:38, 16 June 2008 (UTC)

Guidence on when to reduce dosage of Metformin in renal impairment got tightened in the most recent (March 2008) British National Formulary. Previous editions quoted BNF's definition of mild renal impairment as GFR<50ml/min, however in the lastest BNF55 there has been a switched to quoting CKD levels for which CKD3 is GFR<60ml/min. Wish those nephrologists were working in my local hospitals :-) With CKD guidelines resulting in us GPs having to send obviously stable patients (unchanged creatinines, just that their increasing age reduces the MDRD eGFR), with well controlled blood pressures off to local Nephrologists (that's Guidelines for you), we're starting to see well controlled diabetics have their dose of Metformin reduced for being on wrong side of GFR=60ml/min ! I can't wait for their next HbA1c's in 6 months time :-( Perhaps there'll be enough real-world support for the views of that paper JFW, that I will then be able to return their Metformin dose back :-) David Ruben ^Talk 22:37, 16 June 2008 (UTC)

Lactic acidosis

doi:10.1093/ndt/gfn152 epidemiology of lactic acidosis - letter to editor so in itself a poor source but likely to provide useful background reading. JFW | T@lk 22:41, 30 June 2008 (UTC)

Meta-analysis of CV risk

Let's remember to add this when the study actually comes out. It's already being reported in the popular press. Fvasconcellos (t·c) 22:12, 28 October 2008 (UTC)

Nice free review

Apparently Merck-sponsored, but can be useful. Fvasconcellos (t·c) 21:57, 4 November 2008 (UTC)

Alzheimer Disease

The study relating Metformin with AD is being removed and there's no justification since that information came from a reliable source. If my text doesn't represent well the source rewrite it instead of deleting. If you don't agree with the importance the source gave to the matter balance it with equally sourced opinion.--Nutriveg (talk) 19:13, 10 March 2009 (UTC)

I tried to explain on your talk page why we don't normally include information from cellular models and animal studies in this kind of articles. Unless there is strong evidence of metformin causing dementia in man, I cannot imagine why the reader would find this relevant. It is certainly not the way WP:PHARM goes along with this kind of studies. I think it should be removed, not because the source is not reliable but because it overemphasises a curious phenomenon in a Petri dish that hasn't even been replicated. I shall await reactions from other contributors because removing your content again, but please be prepared to yield to consensus. JFW | T@lk 23:22, 10 March 2009 (UTC)

I really don't think this should be included yet—this is an early finding and hasn't even made it to animal studies yet. If other contributors feel this should stay, it should at least be reworded to something along the lines of "In vitro, metformin has been found to increase generation of amyloid beta." In retrospect, I'm not even sure the TSH suppression effect (the "Hormonal" section in the article) should be mentioned; it doesn't seem to have generated much interest in the literature, and there haven't been any solid studies. Fvasconcellos (t·c) 23:41, 10 March 2009 (UTC)

The source cited is from a legitimate, peer-reviewed source, and the research is relevant to the article. Though it is very early at this stage, and certainly more experiments are going to be done on this. I don't see any reason why it can't be included, but I've moved it to the very end of the 'adverse effects' section, and re-phrased it to indicate that it ***may*** have an effect, but the jury is still out on this one. Dr. Cash (talk) 15:20, 11 March 2009 (UTC)

I still think we should wait for further studies, or at least until this has been cited a few times! What do you think of the TSH suppression effect? I'd really like more input on this one—right now, I'm not sure it is worth mentioning either. It's been nearly three years since the report, surely it would have generated more research by now. Fvasconcellos (t·c) 15:39, 11 March 2009 (UTC)

You shouldn't assume the readers are all stupid, otherwise little content should be presented at wikipedia because other information should be blinded from those who might interpret it wrongly. The link to the in vitro wikipedia article is present, let the readers decide if it's worthy information or not. Many of them would understand that the precautionary principle applies. If you want to rewrite or balance, feel free to do so, but don't just remove information based on reliable source because you don't like it yourself. Nutriveg (talk) 17:28, 11 March 2009 (UTC)

Three years since the report? The PNAS article was published on 2/20/2009 -- 2-3 weeks ago. It's pretty new. But I do agree with your rephrasing it and combining the sections. That seems appropriate. Dr. Cash (talk) 18:33, 11 March 2009 (UTC)

Fv was referring to the TSH report from 2006, I believe. --ἀνυπόδητος (talk) 18:40, 11 March 2009 (UTC)

The sentence seems to claim (at least, to the unsuspecting reader) that accumulation of amyloid beta causes AD, or at least increases the risk to get AD. Since decrease of amyloid beta has no positive effects on AD (Holmes C; et al. (2008). "Long-term effects of Aβ42 immunisation in Alzheimer's disease: follow-up of a randomised, placebo-controlled phase I trial". The Lancet. 372 (9634): 216–233. {{cite journal}}: Explicit use of et al. in: |author= (help); Unknown parameter |month= ignored (help)), this might not be true at all. The relationship between AD and amyloid beta is less than clear at the moment, and it would need a much more careful formulation to reflect this.

Regarding the TSH suppression: I agree with Fv that this is simply not notable. There were no symptoms, no change in thyroid hormone levels, no known mechanism, so the information carried in this sentence is very nearly zero. --ἀνυπόδητος (talk) 18:41, 11 March 2009 (UTC)

I'm going to WP:BOLDly remove both items, as most commentators above seem to agree that it is too early to discuss AD and TSH. JFW | T@lk 19:24, 11 March 2009 (UTC)

Oh, and I insist on discussing all 5207 studies about metformin. In other words, a bit of balance is called for. JFW | T@lk 19:48, 11 March 2009 (UTC)

3 to 2, is just a basic majority, early is to consider there is a majority few minutes after that one opinion. Fell WP:BOLDly to include all the relevant information you find on those articles.--Nutriveg (talk) 20:11, 11 March 2009 (UTC)

Surely the lead section isn't the right place to mention the beta amyloid study? --ἀνυπόδητος (talk) 20:53, 11 March 2009 (UTC)

Probably not, but the other reference for it was removed, if you can find a better place for the information...--Nutriveg (talk) 20:58, 11 March 2009 (UTC)

Moving the information into another section doesn't get us anywhere. I put it back into the Adverse effects section where it belongs, if it does belong anywhere. Please, people, leave it there until there is some consensus. And don't introduce any more typos, if possible. Thank you --ἀνυπόδητος (talk) 21:36, 11 March 2009 (UTC)

This other study, on humans, confirms that Metformin (or glyburide, since the study doesn't differentiate) as monotherapy (without insulin) resulted in higher beta-amyloid plaques than combined therapy. However the study diverges saying that combined therapy is better than control.--Nutriveg (talk) 21:47, 11 March 2009 (UTC)

That's interesting. But it doesn't confirm (I like that word being used in medicine as much as I like the word "cure") that either drug is associated with an increase in plaque formation; it suggests that insulin has a protective effect. Might this be due to metformin increasing biogenesis of Aβ? Certainly—but we don't know yet. I am in no way disputing the merit of the Chen et al. study or its findings, I'm disputing the merit of mentioning these findings when they are so preliminary. For a drug that doesn't even have a fully elucidated mechanism of action yet, some of this research—like much of the research on the diabetes/AD connection—relies far too heavily on presumptions, possibilities, and borderline speculation for my taste. I'd rather see the article focusing on clearly defined facts of definite clinical importance and on "known unknowns" than on "unknown unknowns" :) Fvasconcellos (t·c) 22:08, 11 March 2009 (UTC)

I'm somewhat dismayed that Nutriveg cannot wait until we have achieved some semblance of consensus before readding the same content. This PNAS study simply fails every single notability test we normally employ in pharmacology articles. The fact that it's from a "reliable source" means nothing; there are 5206 other reliable sources about metformin, most of which are clinical studies and not Petri dish studies. I simply cannot understand why we should cover this single unconfirmed preliminary result, and I took several responses above to be in agreement with my view. I'm rapidly going towards having this page protected until the dispute has been resolved. JFW | T@lk 22:59, 11 March 2009 (UTC)

Sorry guys... JFW | T@lk 23:02, 11 March 2009 (UTC)

This study isn't suitable as it's an animal and cell study. How do you propose this relates to human subjects? I mention that because this is a human drug, not an animal one. Placing this on the article prior to a conclusive human study is tantamount to causing panic amongst our diabetic readers, and the danger of placing this here before a proper study has been conducted should not be underestimated. Please wait for a human study, just to enforce what I'm saying, please wait for a human study. There is no rush here. —Cyclonenim (talk · contribs · email) 23:41, 11 March 2009 (UTC)

Besides what JFW claims, the 2009 study was not just about petri dish it included mice as well. The 2008 one was on humans and it brings more evidence that Metformin monotherapy (without insulin) has increased Aβ than combined therapy. The problem with the 2008 study is that it used subjects with same levels of dementia as controls (similar Aβ is expected anyway). There's strong association of Beta amyloid with AD so precautionary principle applies. There's no need to anyone panic, no one is questioning the safety of combined Metformin+Insulin. If you think the information makes the article unbalanced add more of you think it's important. The negative POV is not to blame if it gains evidence only because the article is incomplete. I've no problem in moving it to another section like "Controversies", "Unclear effects", whatever, or to a rewriting, as long as the main information is preserved.--Nutriveg (talk) 02:23, 12 March 2009 (UTC)

You just stated the problem with both studies yourself. You stated that the 2008 one on humans used subjects with the same level of dementia as the control subjects, which immediately disqualifies that study as a good study on humans. The 2009 study is still based on mice and cells, which as we all know, is not the same as producing the same effect in a human. Until a later study which concludes the effect in humans, I'm very wary to add it to the article. It's not so much that it's negative, I don't mind putting negative POV stuff in articles, so long as it's reliable and above all relevant. This, however, is not relevant to humans and should be excluded until that has been fixed. —Cyclonenim (talk · contribs · email) 07:36, 12 March 2009 (UTC)

You shouldn't garbage the whole study because of that control issue (non-diabetics), since when comparing diabetics treatments on that same ground it shows that concomitant therapy performs better than monotherapy on beta amyloid, getting the same, but restated, conclusion of the 2009 one: monotherapy is worse than concomitant therapy. The comparative effect of monotherapy and concomitant on beta amyloid is supported by two studies, on three different approaches (cellular model, mice and humans). The AD association with beta amyloid may be omitted from the text, but the comparative effect of monotherapy on beta amyloid is enough supported, at least to use the Precautionary principle, which I'll cite since no one cares to read "The precautionary principle applies where scientific evidence is insufficient, inconclusive or uncertain and preliminary scientific evaluation indicates that there are reasonable grounds for concern that the potentially dangerous effects on the environment, human, animal or plant health may be inconsistent with the high level of protection chosen by the EU".Nutriveg (talk) 15:11, 12 March 2009 (UTC)

You continue to push for a line of evidence that simply does not exist. Perhaps the time has come for you to explain why you are so passionate about including this single study. There is no evidence that metformin use predisposes to dementia, and I actually find it quite irresponsible of you that you continue to push this. Metformin is the one antidiabetic drug known to reduce cardiovascular disease (UKPDS). Will you, because of your precautionary principle, deny these effects? Now you have come along with a theory that simply lacks a strong epidemiological evidence base and expect us to swallow it whole? You have now been opposed by four different editors, most of whom are very well acquainted with the rigors of evidence and the need to appraise single studies carefully. I urge you in the strongest terms to read WP:MEDRS, a guideline which explains clearly the kind of studies we would normally be citing in pharmacology articles. JFW | T@lk 21:13, 12 March 2009 (UTC)

For now I'm supporting the inclusion of the evidences brought by reliable sources that monotherapy is related to higher beta amyloid than concomitant therapy with insulin. But you're unable to see that as you were when you kept bashing it was just about "petri dish". If you want to investigate dementia alone I'd recommend another study which says that "Patients who were treated with oral antidiabetic medications" had higher risk for dementia then diabetes alone.--Nutriveg (talk) 21:46, 12 March 2009 (UTC)

Unless the full article so states that then your last sentence seems a synthesis, for the abstract makes no comment as to relative risks of diabetes and diabetes-treated with oral agents - yes differing hazzard ratios are stated, but I fail to see (in the abstract) whether one is statistically higher than the other. So unless the full article (to which I have not access) makes that comparison comment, then WP:SYN applies. Also note that WP:PRIMARY states "All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than original analysis of the primary-source material by Wikipedia editors." - and that's policy. So priliminary findings (assuming correctly reliably sourced) are generally not to be used until some secondary source assesses the significance or otherwise of the initial findings, and certainly not in list of adverse effects as there is no chance yet of any real-world consensus on this. Lastly wikipedia is not a news ticker-tape of unfolding events, but is at least somewhat set back waiting for a less myopic view (per Wikipedia:NOT#NEWS). David Ruben ^Talk 22:52, 12 March 2009 (UTC)

I'm just not feeling it, guys. There's too much speculation here: metformin might produce an effect in humans that might represent a greater risk for a disease that a person with well-treated diabetes might live long enough to develop.

Nutriveg, I think we all understand the precautionary principle. I think all the rest of us also recognize that Wikipedia's content policies are not controlled by the Maastricht Treaty. We need to represent the proportional weight of scientific evidence. The worry that "metformin might cause Alzheimer's" is simply not a significant enough issue to justify mentioning it at this time. WhatamIdoing (talk) 23:49, 12 March 2009 (UTC)

The paper does not demonstrate a causal link between metformin use and AD. At present, speculation of this nature is a fringe theory with only a single reference that does not merit entry in Wikipedia's article. Axl ¤ [Talk] 10:38, 13 March 2009 (UTC)

In vitro studies are not of enough significance to qualify for an article on a disease in humans for anything other than basic science. This might be appropriate for an article on ongoing research.--Doc James (talk · contribs · email) 03:40, 15 March 2009 (UTC)

Wiki said "increased risk of Alzheimer Disease" the ref said "increases biogenesis of Alzheimer's amyloid peptides". This ref says nothing about alzheimers disease just about amyloid peptides in petri dishes.--Doc James (talk · contribs · email) 03:45, 15 March 2009 (UTC)

links

Trying to follow the above discussions, and would like to know if the following two links are part of the discussions?

1. Metformin
2. Metformin - Evidence from the UK Prospective Diabetes Study (UKPDS)

—Preceding unsigned comment added by Gioto (talk • contribs) 06:07, 17 April 2009

• The first link (adipocyte.co.uk) is run by a UK Individual and may be an interesting summary but is less useful as a source (see guidance of WP:RS and WP:PSTS). The web pages do point to a number of on-line publications that do look like useful sources however.
• The second link is to the NHS library but the particular summary of UKPDS is itself not a publication (not dated and no authors named) and does not provide links to the publications mentioned or provide full form references for them. For example the first reference given "(National Collaborating Centre for Chronic Conditions, 2008)" implies that there should be a bibliography supporting the summary but no bibliography is provided so the references cannot be verified.

—Teahot (talk) 09:48, 17 April 2009 (UTC)

Okay thanks for the explanation. I also found the following article "Metformin and serious adverse effects - The Medical Journal of Australia" and a more recent one Metformin and lactic acidosis in an Australian community setting: the Fremantle Diabetes Study which may need to be add if relevant. 02:32, 14 May 2009 (UTC)

Interesting; the latter could certainly be worked into the article. That adipocyte.co.uk link is actually very good. Although (as Teahot mentioned) it is not useful as a direct source as it is self-published, it provides and summarizes references that are reliable sources. Fvasconcellos (t·c) 14:04, 14 May 2009 (UTC)

Any direct references to published articles in the Medical Journal of Australia (including the two mentioned), would make excellent sources. If you think they are relevant, I suggest you go ahead and add them.—Teahot (talk) 17:40, 14 May 2009 (UTC)

Pharmacokinetics

A quick question re elimination half-life and reference 39 Robert et al. The stated elimination half-life of 17.6 hours (in the last sentence of this section) is outside of the reported range of 18.5-31.5 hours.

...in red blood cells, with a much longer elimination half-life: 17.6 hours (18.5–31.5 hours in a single-dose study of non-diabetic people).[39]

I read the Robert et al journal article and didn't find the 17.6 hours in the text or tables. Did I overlook it or could this be a typo?

Lkbwiki (talk) 15:51, 9 August 2009 (UTC)

It was actually a source omission—17.6 hours is the nominal half-life reported in most of the literature, including the product information. The Robert et al. study, which reports a different range in non-diabetic subjects, was added as a counterpoint. Fvasconcellos (t·c) 16:01, 9 August 2009 (UTC)

Thanks for the quick response and page edit.Lkbwiki (talk) 16:20, 9 August 2009 (UTC)

Metformin and CD8 function

This study has been making waves in the popular press. It's much too early to add it to the article, but it's definitely something to keep an eye on. Fvasconcellos (t·c) 02:26, 11 June 2009 (UTC)

And: Metformin induces unique biological and molecular responses in triple negative breast cancer cells. Also still a mouse model, very preliminary as the AD research mentioned above, but something to keep an eye on. Fvasconcellos (t·c) 15:48, 12 September 2009 (UTC)

Reduced pancreatic cancer risk?

Will add to the article later. Fvasconcellos (t·c) 21:05, 18 August 2009 (UTC)

Thanks to the people who've worked on this article

I've recently been diagnosed with type2 and this has been really helpful. Your work is appreciated -Thanks! —Preceding unsigned comment added by 222.153.36.128 (talk) 11:11, 12 September 2009 (UTC)

I'm very glad to hear that, and I hope I speak for all contributors in saying "you're welcome". If you have any suggestions for improving the article, they're more than welcome. Fvasconcellos (t·c) 15:41, 12 September 2009 (UTC)

Beta amyloid and Metformin monotherapy

The topic has changed I think you're unable to see my last addition. There's no mention of AD in it, but beta amyloid. It's not about Metformin, but "Metformin monotherapy". And it's not a single paper but two papers supporting the information. Read what you're talking about first.--Nutriveg (talk) 12:19, 13 March 2009 (UTC)

Since my last edit albeit many came here to comment on AD, all failed to comment on "Beta amyloid and Metformin monotherapy" as I exposed, "Axl" and "JFW" said it was a single study, but there are two supporting studies, "WhatamIdoing" commented about Alzheimer, but there's no mention of Alzheimer in my last edit, both "David Ruben" and "JFW" commented on dementia, there's no mention of dementia either. So, having all the comments made to points that were already removed from the text or ignored new information that has been added they are invalid for that last edit. So if no one express a valid argument about it I'll revert to that last edit.--Nutriveg (talk) 21:17, 13 March 2009 (UTC)

No, actually, I think you're going to pay attention to the utter lack of WP:Consensus for its inclusion, not to whether or not you think that the uniform opposition by many independent editors counts as "valid arguments" that persuade you to change your mind about the subject. WhatamIdoing (talk) 21:29, 13 March 2009 (UTC)

The edit link provided does include "diabetic dementia patients", so stop wikilawyering, stop pushing a (currently) minor research observation, and stop attacking other editors (discuss issues not editors), finally don't issue ultimatums of continuing to revert against consensus - that is not how we collaboratively work here and such action risks you being blocked for disruption. David Ruben ^Talk 14:52, 14 March 2009 (UTC)

It's more tricky to find the references that Nutriveg added, because the article has since been changed (reverted). Nutriveg's most recent edit is technically accurate. However it gives undue weight to a finding that requires a lot more research before it becomes relevant to Wikipedia's article. Axl ¤ [Talk] 17:41, 14 March 2009 (UTC)

"WhatamIdoing": just saying you disagree doesn't help reach consensus, I tried to improve the text listening to previous arguments. David, well those were the subject of the study (all of the subjects, including controls, had dementia), it isn't said there was a connection, I may remove that part. Axl, there are two articles supporting the information, while one includes two studies, and I moved the text from "side effects" to "interactions" (with insulin), so there's no undue balance. If three studies aren't enough why not to remove all the remaining text in the same situation but mine?--Nutriveg (talk) 19:18, 14 March 2009 (UTC)

Nutriveg, we're trying to write an encyclopedia, not a breaking-news website. This purported connection is widely considered to be speculative, perhaps worth more research -- but not known. Mentioning it at all is inappropriate at this time. When it hits the textbooks or the reviews, then we can include it. WhatamIdoing (talk) 19:40, 14 March 2009 (UTC)

Agree with WhatamIdoing. This is not significant until it hits a review.--Doc James (talk · contribs · email) 05:20, 15 March 2009 (UTC)

Adding some information:

Activation of AMPK inhibits amyloid beta production [1], Metformin activates AMPK [2].

Untreated diabetic patients with dementia had higher amyloid-beta peptide load in the cerebral cortex.[3] --Nutriveg (talk) 12:23, 17 April 2009 (UTC)

Have there been any further developments in this respect? Fvasconcellos (t·c) 21:06, 18 August 2009 (UTC)

At least one.Liao, F.; Xu, H. (Jun 2009). "Insulin signaling in sporadic Alzheimer's disease". Science signaling. 2 (74): pe36. doi:10.1126/scisignal.274pe36. PMID 19509405.--Nutriveg (talk) 21:29, 18 August 2009 (UTC)

I can't read this article but I think it's related [4]--Nutriveg (talk) 04:01, 19 August 2009 (UTC)

About women: "Metformin reduces the insulin response by decreasing hepatic gluconeogenesis and reducing androgen levels,"[5] "The results showed that (...) following metformin, (...) insulin response was significantly lower"[6] About men: "low early insulin response was associated with increased risk of subsequent Alzheimer’s disease" [7]. "A low insulin response at baseline was associated with a higher cumulative risk of AD"[8]--Nutriveg (talk) 04:19, 19 August 2009 (UTC)

Interesting. I'm sure someone will soon start looking for a direct link between metformin use and incidence of AD. Fvasconcellos (t·c) 12:39, 19 August 2009 (UTC)

"Increased levels of plasma Aβ 42 were found in non-demented subjects who used insulin and biguanides"[9]. That's a review.--Nutriveg (talk) 14:45, 6 October 2009 (UTC)

Reduced insulin response in women

The following phrase was removed but I don't see why:

In women, Metformin reduces insulin response by decreasing hepatic gluconeogenesis and reducing androgen levels.

--Nutriveg (talk) 13:58, 6 October 2009 (UTC)

There are a few problems with your addition. You added a tiny section under "adverse events", while I suspect you wanted to cite these studies as evidence of a beneficial effect. The data on PCOS is already discussed elsewhere in the article (under "off-label use"). Could I also suggest that 11 references to make the same point is also a bit excessive.

Could I recommend that you select 1-2 references that actually add anything to the article and place them in the relevant section (probably "mechanism of action"). JFW | T@lk 17:44, 6 October 2009 (UTC)

Well, I didn't create a new section but a new title under a existing section, to separate from the others. I do think that "reduced insulin response" is an unexpected effect, you may suggest a better section for that phrase.

I'm not talking about PCOS, but "reduced insulin response".

I didn't knew if your problem was with sources, since you said "these studies are not really of sufficient clinical relevance" as a justification to revert my edit, so I added other sources (including other studies) I had available that support that phrase.--Nutriveg (talk) 20:02, 6 October 2009 (UTC)

Since there's no further response so far I'll put it back.--Nutriveg (talk) 21:50, 13 October 2009 (UTC)

"Reduced insulin response" is, actually, neither adverse nor unexpected. It should be mentioned in the article, but not as an adverse effect, and certainly does not merit its own section. By the way, nearly all of those references deal with non-diabetic women with PCOS, in which this effect is beneficial, as hyperinsulinemia plays a major role in PCOS. Fvasconcellos (t·c) 22:37, 13 October 2009 (UTC)

I don't see why it would be expected to delay insulin response, but anyway it deserves citation. It's not a new section but a topic under a existing section, since that section is divided by topics. It's not adverse if one has PCOS but otherwise it increases the risk of Alzheimers.[10].--Nutriveg (talk) 14:42, 14 October 2009 (UTC)

Perhaps it's just your terminology that is nonstandard. Presumably by reduced insulin response you don't mean insulin resistance. Perhaps you should clarify your edits rather than simply reverting back against consensus. JFW | T@lk 21:52, 14 October 2009 (UTC)

I used the same term used by the sources, I just cited them, where they are clear to differentiate both, since insulin response can be affected by other factors. I didn't simple reverted back, I waited for your (lack of) response for more than a week!--Nutriveg (talk) 22:06, 14 October 2009 (UTC)

Evident lack of reliable source

Medical association sources are not reliable, because their primary interest is make profit. Metformin is low effective against diabetes, but it is sold JUST because it is more economically conveniet than testosterone (which is much safer and effectiver, try and see...).

It is clear why it is sold metformin and not testosterone: because it is necessary to take many tablets per day instead than one injection per month (testosterone).--158.194.199.13 (talk) 20:09, 28 October 2009 (UTC)

Please see WP:MEDRS for what actually constitutes a reliable source. Metformin is not "low effective against diabetes". In fact, it is probably the antidiabetic with the most well-established benefits; broadly speaking, all oral antidiabetics are equally effective in lowering blood sugar and Hb_A1C.

There is absolutely no high-quality evidence to support the safety or efficacy of testosterone replacement as a treatment for T2DM. At best, there is the possibility of its supplementation being somewhat beneficial in some men with diabetes.

And, finally, even if testosterone was an established treatment for diabetes, this information would not belong in the Metformin article. Please do not reinsert it. Fvasconcellos (t·c) 00:07, 29 October 2009 (UTC)

Reliable means non profit source. Medical associations are profit associations. There are no serious articles in which is visible an high metformin anti-diabetic effectiveness for a simple reason: it is not effective. If it is a common treatment for merely economic reasons is another fact.

Concerning to the testosterone there are many articles in which it is well explained that testosterone injictions greatly improve sugar tollerance (and of course erection, osteoporosis, anemia, memory and so on...). Testosterone is a good treatment for all T2M patients because they are hypogonadic, it is simply impossible to have type 2 diabetes with good free testosterone levels (>20 picograms/ml).

Testosterone is simply much effectivier and safer than metformin. Try and see!

I would suggest to write in the article:" In order to obtain the maximum profit metforin is sold as anti-diabetic drug. Metformin is unuseful and damaging anti-diabetic drugs. This compound is very effective just on increasing doctor and industries bank account...There are many better alteratives as in example testosterone, unfortunately (for the profit of industries) they must be injected just once per month (and also less)".--Testosterone vs diabetes (talk) 16:10, 29 October 2009 (UTC)

No, that is not what it means. Read WP:MEDRS for what constitutes a Reliable source for Wikipedia's medicine-related articles. There are no high-quality studies in peer-reviewed academic journals to support your suggestions. Metformin is one of the cheapest antidiabetics available (surpassed only by the old sulfonylureas) and is a generic anywhere in the world; trust me, it's no one's idea of a cash cow.

Please stop edit-warring; you're quickly headed for a block. Fvasconcellos (t·c) 17:33, 29 October 2009 (UTC)

It is absolutely false! If you control (even in Wikipedia articles) there are many articles in review with high impact factor related with high efficiency of testosterone in decreasing excessive blood glucose levels. Hypogonadism is very related with diabetes mellitus.

http://www.ncbi.nlm.nih.gov/pubmed/19444934?ordinalpos=2&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum

http://www.ncbi.nlm.nih.gov/pubmed/18832284?ordinalpos=9&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum

http://www.ncbi.nlm.nih.gov/pubmed/18832284?ordinalpos=9&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum

http://tde.sagepub.com/cgi/content/abstract/34/5_suppl/97S

http://www.ncbi.nlm.nih.gov/pubmed/18772488?ordinalpos=24&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum

According to the scientific reviews: testosterone is definitely better than metformin as anti-diabetic drug (and of course as anti erectile disfunction, osteoporosis, Alzhaimer, dementia and so on). Metformin is better than testosterone from a merely economic point of view.

You simply can not accept that diabetes mellitus is easy to recover: just take testosterone in right dosage! You can not accept that metformin is not effective (a part on doctor's bank account of course).

Testosterone is very cheap (more than metformin), and with one injection per month (or also per trimester) diabetes is out. Metformin costs much more because it is also necessary to take many tabs per day.

The first sources in the article have not scientific value and they are not scientific articles, they are just pubblicity--Testosterone vs diabetes (talk) 17:55, 29 October 2009 (UTC)

The nature of a non-profit

Non-profit organizations such as the American Diabetes Association are so named because it is illegal for their members to profit from the organization. Consequently, the claim in the edit summary that "their primary aim is the profit" is absurd. WhatamIdoing (talk) 21:02, 29 October 2009 (UTC)

American Diabetes Association is a non profit association???? Do you know the huge deal of anti-diabetic drugs? Why this "non-profit" association speaks about highly remunerative drugs as metformine, or about highly remunerative foods as in example low colosterol food? Why American Diabetes Association make censorship about the real T2M cause (low testosterone and/or high cortisol levels)? I have never heard a doctor speaking about the tight relationships between these two hormones and diabetes (look upward articles). This is a proof that what is really important for industries is only the profit, not pubblic health! Sorry for my bad english, but I hope you understood. Why they try to prescribe metformin and not testosterone, when it is clear that the steroid is much better?--Testosterone vs diabetes (talk) 21:18, 29 October 2009 (UTC)

Where are the interventional trials showing testosterone to decrease micro- and macrovascular disease? JFW | T@lk 21:42, 29 October 2009 (UTC)

Yes: American Diabetes Association, IRS number 13-1623888, Pub 78 listed is a public charity (the highest classification of non-profit organization under US laws).

I don't know why you think that they favor "highly remunerative" drugs and foods. Metformin is quite cheap -- about 20 cents per pill for a low dose, which means that a low dose of the typical daily testosterone creams used in the US costs about ten times as much -- and naturally cholesterol-free foods (e.g., vegetables and rice) are cheaper than high-cholesterol foods (e.g., pork). Even if they weren't cheaper, they make no money from those recommendations.

Additionally, I suspect that many women would object to being given testosterone, especially women with polycystic ovarian syndrome, who naturally overproduce testosterone and related androgens, and find that testosterone hurts, not helps, their health. Too much testosterone causes acne, hirsuitism, and infertility in women.

Do please provide JFW with information about the proper clinical trials that you say support your unusual beliefs. WhatamIdoing (talk) 22:06, 29 October 2009 (UTC)

Are you blind? Do you see the articles upward? Is it cheaper testosterone (one injection evert TRE MOUNTHS) or metformin (many uneffective tablets per day)? Is cheaper normal cheese or cheese with a low cholesterol content? Is cheaper normal yogurt or yogurt with low cholesterol content? American diabetes society is a merely profit association which promotes use of expensive and uneffective products as metformin (uneffective but economically convenient), low colesterol food (total colesterol is not related with cardio-vascular diseases or diabetes). American society of diabetes makes censorship about the real causes of diabetes type 2 (low testosterone levels and/or high cortisol levels), the economic reasons are clear.

JFW is a doctor. It means he is payed to prescribe metformin, therefore he get blind (he do not see the articles in this discussion). His opinion is not important because he has interest to defend metformin.

Whatiamdoing, testosterone can not cause acne and hyrsitism, you are making confusion between testosterone and dihydrotestosterone. Bald people for example have low testosterone levels and high dihydrotestosterone levels. The effects oh these hormones are completely different, testosterone acts especially improving sugar absoption and increasing muscle mass (for that reasons diabetic patients have low muscle mass).

There are many way to take testosterone, each of them has vantages and disdventages. For diabetes the intramuscolar injections are better than gels. The injections are very cheap, gels are not cheap.

The role of testosterone in women sugar absorption is not well studied, therefore we can not make easy conclusion, however this hormone is very important for women too, women testosterone replacement therapy is very effective against loss of libido and osteoporosis. If you see the articles the use of testosterone is effective and safe for men, further studies are necessary to be done for woman patients.

Metformin is uneffective and damaging drug, it is sold just because is very remunerative. --Testosterone vs diabetes (talk) 17:30, 30 October 2009 (UTC)

Perhaps things are different in your country, but here in the US, fat-free yogurt cost exactly the same as low-fat yogurt, and full-fat yogurt is very expensive (raises the cost by 30% to 60%) and can be hard to find. Reduced fat cheeses, e.g., part-skim mozzarella, are the cheapest on the market. Skim (fat-free) milk is 5%-10% cheaper than whole milk.

Perhaps testosterone is also cheap in your country. It is very definitely not cheap in the US. Metformin, on the other hand, is widely available in the US for just $10 for a 90-day supply. See here. WhatamIdoing (talk) 22:40, 30 October 2009 (UTC)

In Brazil, a 30-day supply of metformin costs about 50 cents of a dollar through a government program (Farmácia Popular) and about 4 dollars otherwise. Interestingly, low-fat yogurt is no more expensive here either :) Fvasconcellos (t·c) 22:44, 30 October 2009 (UTC)

In Italy Sustanon 250 (testosterone injections, usually one per 3-4 moths) cost 11 euro. It means 11/100 (days)= 0.11 euro per day.

Metformin 1g costs 5 euro (60 tabs. Usually 3 tabs per day)= 0.25 euro per day.

Testosterone is definitely cheaper than metformin! Moreover testosterone is effective against osteoporosis, loss of libido, depression, dementia, blood pressure normalization; metofmin no. Testosterone has also less side effects than metformin. --Testosterone vs diabetes (talk) 15:01, 1 November 2009 (UTC)

Interesting information. Unfortunately there is no proof, and no reliable source claiming that testosterone is a better therapy overall -it has been suggested that could be of some help in hypogonadal men. And such information plays no role in this article. --Cyclopia^talk 15:04, 1 November 2009 (UTC)

I agree: for any alternative treatment to be relevant to this particular article, we would have to have a study that directly compared the two options. WhatamIdoing (talk) 01:44, 2 November 2009 (UTC)

is link to Drugs@FDA clinical/lable info part of template? nice addition

Drugs@ FDA has a bunch of scanned things used to support license and label. Note I have an agenda here, I'm trying to get more people to use this and have FDA stop accepting scanned documents without some attempt at OCR equivalents,http://www.accessdata.fda.gov/drugsatfda_docs/nda/97/020357a_s006.pdf Their site also makes use of posts and generic links making citations hard so I'm encouraging anyone who finds this useful to suggest they make the site better for online usage, including maybe an API. While this is clearly soapboxing, it is quite relevant to the goals of wikipedia. http://www.accessdata.fda.gov/Scripts/cder/DrugsatFDA/ is the main search page and everything from then on seems to use cookies and can't book mark... Nerdseeksblonde (talk) 15:56, 1 November 2009 (UTC)

12 papers on pubmed mention "diabetes cushing testosterone" for debate above,

1. Nakamura, A; Nagai, S; Taniguchi, S; Umetsu, M; Atsumi, T; Wada, N; Yoshioka, N; Ono, Y; Sasano, H; Koike, T (Jan-2007). "Unilateral adrenalectomy improves insulin resistance and polycystic ovaries in a middle-aged woman with virilizing adrenocortical adenoma complicated with Cushing's syndrome". Journal of endocrinological investigation. 30 (1): 65–9. PMID 17318025. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
2. Glintborg, D; Andersen, M; Hagen, C; Hangaard, J; Rasmussen, PE; Schousboe, K; Hermann, AP (Dec-2004). "Prevalence of endocrine diseases and abnormal glucose tolerance tests in 340 Caucasian premenopausal women with hirsutism as the referral diagnosis". Fertility and sterility. 82 (6): 1570–9. doi:10.1016/j.fertnstert.2004.06.040. PMC 10.1016/j.fertnstert.2004.06.040. PMID 15589862. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help)CS1 maint: date and year (link)
3. Afandi, B; Saadi, HF. "Cushing's syndrome caused by unsupervised use of ocular glucocorticoids". Endocrine practice : official journal of the American College of Endocrinology and the American Association of Clinical Endocrinologists. 9 (6): 526–9. PMID 14715481.
4. Fukushima, A; Tanikawa, T; Kawahara, C; Misawa, H; Kanda, K; Morita, E; Sasano, H; Tanaka, Y (Apr-2003). "Virilizing adrenocortical adenoma with Cushing's syndrome, thyroid papillary carcinoma and hypergastrinemia in a middle-aged woman". Endocrine journal. 50 (2): 179–87. PMID 12803238. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
5. Lubin, V; Antoine, JM; Beressi, JP; Vexiau, P (9-Nov-2002). "[Cushing's syndrome during pregnancy]". Presse médicale (Paris, France : 1983). 31 (36): 1706–13. PMID 12467153. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
6. Petruson, K; Petruson, B; Lindstedt, G; Bengtsson, BA (Jul-1997). "Transsphenoidal adenomectomy in Cushing's disease via a lateral rhinotomy approach". Surgical neurology. 48 (1): 37–43, discussion 44-5. PMID 9199682. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
7. Bakiri, F; Aouali, R; Semrouni, M; Derome, P; Chitour, F; Benmiloud, M (Oct-1996). "Treatment of Cushing's disease by transsphenoidal, pituitary microsurgery: prognosis factors and long-term follow-up". Journal of endocrinological investigation. 19 (9): 572–80. PMID 8957739. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
8. Cunningham, SK (Dec-1994). "Dissociation of adrenal androgen and cortisol secretion in Cushing's syndrome". Clinical endocrinology. 41 (6): 795–800. PMID 7889616. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
9. Trainer, PJ; Verhelst, J; Howlett, TA; Lowe, DG; Grossman, AB; Savage, MO; Afshar, F; Besser, GM (Jan-1993). "Transsphenoidal resection in Cushing's disease: undetectable serum cortisol as the definition of successful treatment". Clinical endocrinology. 38 (1): 73–8. PMID 8435888. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
10. Malchoff, CD; DeBold, CR; Kozol, RA; Ramsby, GR; Page, DL; Malchoff, DM; Orth, DN (Apr-1989). "Adrenocorticotropin-independent bilateral macronodular adrenal hyperplasia: an unusual cause of Cushing's syndrome". The Journal of clinical endocrinology and metabolism. 68 (4): 855–60. PMID 2537845. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)
11. Bezverkhaia, TP; Verkhogliadova, LM; Babicheva, ZI; Kovalenko, IV (1986). "[Diagnostic value of determining the levels of testosterone, estradiol and testosterone-estradiol-binding globulin in the plasma]". Laboratornoe delo (6): 330–3. PMID 2427774.{{cite journal}}: CS1 maint: date and year (link)
12. Baba, H (20-Jan-1973). "[Radioimmunoassay of serum testosterone--results of the study of normal subjects and patients with endocrine disorders]". Nippon Naibunpi Gakkai zasshi. 49 (1): 60–79. PMID 4734772. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)

I've appreciated your edit. But the text you added is partially related with diabetes mellitus. Adrenocortical adenomas can cause DM but they are quite rare, and their percentage in subclinical Cushing's syndrome and in diabetes mellitus is quite low. Usually patients with diabetes mellitus type 2 or with subclinical Cushing's syndrome have higher (than normal individuals) cortisol levels, higher ACTH levels (with possibilities of pituitary adenomas) and lower testosterone levels.--Testosterone vs diabetes (talk) 16:55, 1 November 2009 (UTC)

The other articles seem very interesting! --Testosterone vs diabetes (talk) 16:55, 1 November 2009 (UTC)

abs copy vio? link? is this a wikipedia thing or legal? I don't have an easy way to edit but what is fair use? I thought anyone could index these or replay pubmed. Nerdseeksblonde (talk) 17:59, 1 November 2009 (UTC)

I don't think that abstracts are considered fair use. PubMed (ISI, Scirus etc.) is authorized to do so by the journals, but I doubt that non-open access journals are happy to see dozens of them freely distributed. Thus I erred on the side of caution. I personally hate this kind of things, but copyvios are legal stuff, and as such is taken pretty seriously here. It can be a good idea to ask to WP:MCQ (even if this stuff is just text, they probably have sound advice). Sorry if it caused problems. --Cyclopia^talk 18:13, 1 November 2009 (UTC)

I fixed the formatting of the list for you. When you're not trying to make the citations appear elsewhere on the page, just leave off the <ref> part. WhatamIdoing (talk) 01:49, 2 November 2009 (UTC)

TvD, some of these sources are interesting, but what do they have to do with metformin? For this article, we need studies that specifically and directly tell us something about metformin. This is Metformin, not Diabetes treatment. WhatamIdoing (talk) 01:52, 2 November 2009 (UTC)

We won't find them, because there hasn't been enough study in this area. Most of the studies above are not germane to this discussion (or the discussion at Diabetes mellitus for that matter). This whole discussion has meandered way off topic, and I'm inclined to collapse it. Fvasconcellos (t·c) 10:43, 2 November 2009 (UTC)

In this case, esp considering the size of above citation list, I wouldn't object but merit discussions can help add perspective to content. In this case, someone specifically has an issue with sources, withouth a big concern for relevance to this topic. "Metformin cushing" returns a tractable 6 refs,

Heneghan, Marykathleen (22-Sep-2009). "Cushing disease as possible cause of persistent growth failure despite growth hormone therapy in a small for gestational age male". Pituitary: -. doi:10.1007/s11102-009-0201-3. PMC 10.1007/s11102-009-0201-3. PMID 19771523. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help)CS1 maint: date and year (link)

August, P; Brooks, MM; Hardison, RM; Kelsey, SF; MacGregor, JM; Orchard, TJ; Chaitman, BR; Genuth, SM; Goldberg, SH; Hlatky, MA; Jones, TL; Molitch, ME; Nesto, RW; Sako, EY; Sobel, BE (11-Jun-2009). "A randomized trial of therapies for type 2 diabetes and coronary artery disease". The New England journal of medicine. 360 (24): 2503–15. doi:10.1056/NEJMoa0805796. PMC 10.1056/NEJMoa0805796. PMID 19502645. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help); Missing |author1= (help)CS1 maint: date and year (link)

Pop-Busui, R; Lopes, N; Jones, TL (Mar-2009). "Prevalence of diabetic peripheral neuropathy and relation to glycemic control therapies at baseline in the BARI 2D cohort". Journal of the peripheral nervous system : JPNS. 14 (1): 1–13. doi:10.1111/j.1529-8027.2009.00200.x. PMC 10.1111/j.1529-8027.2009.00200.x. PMID 19335534. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help)CS1 maint: date and year (link)

Mapas-Dimaya, AC; Bahtiyar, G; Mejia, JO; Sacerdote, AS (Oct-2008). "Metformin-responsive classic salt-losing congenital adrenal hyperplasia due to 21-hydroxylase deficiency: a case report". Endocrine practice : official journal of the American College of Endocrinology and the American Association of Clinical Endocrinologists. 14 (7): 889–91. PMID 18996819. {{cite journal}}: Check date values in: |date= (help)CS1 maint: date and year (link)

Iwasaki, Y; Nishiyama, M; Tsugita, M; Taguchi, T; Asai, M; Yoshida, M; Kambayashi, M; Hashimoto, K (26-Mar-2008). "Is the metabolic syndrome an intracellular Cushing state? Effects of multiple humoral factors on the transcriptional activity of the hepatic glucocorticoid-activating enzyme (11beta-hydroxysteroid dehydrogenase type 1) gene". Molecular and cellular endocrinology. 285 (1–2): 10–8. doi:10.1016/j.mce.2008.01.012. PMC 10.1016/j.mce.2008.01.012. PMID 18313835. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help)CS1 maint: date and year (link)

Christ-Crain, M; Lolli, F; Fekete, C; Seboek, D; Wittmann, G; Feltrin, D; Igreja, SC; Ajodha, S; Harvey-White, J; Kunos, G; Müller, B; Pralong, F; Aubert, G; Arnaldi, G; Giacchetti, G; Boscaro, M; Grossman, AB; Korbonits, M (Jun-2008). "AMP-activated protein kinase mediates glucocorticoid-induced metabolic changes: a novel mechanism in Cushing's syndrome". The FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 22 (6): 1672–83. doi:10.1096/fj.07-094144. PMC 10.1096/fj.07-094144. PMID 18198220. {{cite journal}}: Check |pmc= value (help); Check date values in: |date= (help)CS1 maint: date and year (link) CS1 maint: unflagged free DOI (link)

Nerdseeksblonde (talk) 15:42, 2 November 2009 (UTC)

I have not inserted the articles. I try to find some study comparing testosterone and metformin.--Testosterone vs diabetes (talk) 16:28, 2 November 2009 (UTC)

Metformin vs testosterone

I did not find articles with strict comparations between metformin and testosterone treatments against diabetes.

http://www.ncbi.nlm.nih.gov/pubmed/12235466?itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=2

http://www.ncbi.nlm.nih.gov/pubmed/11707532?itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_RVDocSum&ordinalpos=14

Anyway metformin decreases testosterone levels, this effect (as earlier stated) can cause or worse diabetes (and many other diseases related to hypogonadism). We may therefore state that metformin is harmful for sugar absorption and many other health issues (bone density, sexuality, memory, cognition, mood, digestion...). Testosterone has no so serious side effects, and greatly improves the upward biological parameters (see diabetes mellitus and testosterone replacement therapy).

Testosterone is a natural product, metformin no (I mean it is not naturally present in the human body). Metformin failed to cure diabetes, testosterone (right now) no.

Metformin decreases TSH respond and may cause serious digestion problems (vomit, nausea, diarrhea, constipation), testosterone no.

We may therefore state that testosterone is definitely a better anti-diabetic drug than metformin.--Testosterone vs diabetes (talk) 19:05, 2 November 2009 (UTC)

This is interesting. But while the effects on the hormonal balance can be added per the sources you brought, the "statements" you talk about definitely cannot be added in any Wikipedia article, because at least they are a mix of WP:OR and WP:SYNTHESIS. The point is: we cannot write deductions that are not explictly stated in reliable sources (no matter how trivial they may seem -and here they do not seem trivial at all, too) --Cyclopia^talk 19:12, 2 November 2009 (UTC)

Of course, right now (in a wikipedia article) we can not state that testosterone is better than metformin aqs anti-diabetic drug. But, i guess, we can compare the anti-diabetic drugs as I did it upward using scientific articles in some specific paragraph.--Testosterone vs diabetes (talk) 19:19, 2 November 2009 (UTC)

I have nothing against the proposal. It doesn't seem to belong to this article, but a general "Comparison of anti-diabetic drugs", separated article, could be interesting, if well done and not limited to metformin and testosterone (I guess there are others). A good idea -that also allows you to learn Wikisyntax etc. and doing useful work- could be for you to work on it on a subpage on your userpage (see WP:SUB for details). After it reaches a "reasonably good" state, you can then move it to the article space. --Cyclopia^talk 19:44, 2 November 2009 (UTC)

There's already an article on that subject: Diabetes management.

The two papers linked above could be used in this article to supporta claim that metformin reduces testosterone levels in men. That's as far as we can go with it, because there's no evidence that this change has any practical consequences.

We also can't support this claim that testosterone is essentially safe, as it definitely has known side effects[11], some of which are considered severe. WhatamIdoing (talk) 04:52, 3 November 2009 (UTC)

Furthermore, testosterone is not an established anti-diabetic, and should not be presented as such (e.g. by including it in the Anti-diabetic drug article). There have been no head-to-head comparisons of testosterone and anti-diabetics, and comparing testosterone to these drugs will invariably be a synthesis of published material. Fvasconcellos (t·c) 10:02, 3 November 2009 (UTC)

I don't want to speak to merit in this discussion but certainly for wikipedia OR is a concern. Just generally there is a problem in the field to extrapolate results and assume causality where none exists. Also, there was something about "natural" above- AFAIK there is no magic here and natural doesn't have any inherent implications on safety and efficay for a specific purpose. Nerdseeksblonde (talk) 14:40, 3 November 2009 (UTC)

Whatamidoing the reduction of testosterone (hypogonadism) is strictly related with T2D, osteoporosis, memory loss, heart attacks and many other severe diseases (at least in men but it seems in women too). Metformin is dangerous for health because it reduces testosterone and tiroid's hormones.

Testosterone is a fully established anti-diabetes compound (up to now for men only), it is used by many doctors in USA and in EU excatly (also) for this purpose, its anti-diabetic use is greatly increasing. In example Dot. Ullis (who follows american olympic athlets and who write a book called "SuperT") uses testosterone for diabetic men (type 1 and 2). WhatamIdoing I invite you to do not attach absourd and nonscientific links with absourd and ridicolous (headache...hahaha...do not take offence) side effects. Yes testosterone can have serious side effects (mainly inhibition of endogenous T production) but absolutely not severe and anyway, metformin has much more side effects.

Metformin is not efficient in diabetes reduction for a simple reason: it does not affect insulin resistance. Metformin reduces gluconeogenesis (which is indeed one of the mechanisms involved in hyperglycemia), but it does not increase insulin respond (glucose absorption). Yes it (slightly...) reduces blood glucose levels, but this has nothing to do with diabetes or with insulin resistance. Diabetes is not excatly hyperglycemia, but it is intracellular glucose deficiency. I've never read (I could do a mistake in this) that metformin increases intracellular glucose levels, instead there are many articles stating the increase of insulin respond after using testosterone.

If a diabetic man takes metformin will not improve his health, if he takes testosterone will greatly improve his health.

Here the adjective "natural" means naturally present in human body (as testosterone in example) and nonnaturally present in human body (metformin is a natural compound but it is not naturally present in the human body). I think if something is similar to our natural compounds, it should be more effective and safer.--Testosterone vs diabetes (talk) 19:27, 3 November 2009 (UTC)

1)This place is not a forum to discuss about medicine, but it is for discussion on what has to be included in the article.

2)All of this would be interesting on a forum, but not here. It has been already repeated several times that what you are reporting above is either your personal opinion, or synthesis between sources. Therefore, it has no place in Wikipedia -and more so, not in this article.

Going into the details:

3) Testosterone is by no means a "fully established" anti-diabetes compound, at least unless sources are found showing academic consensus on that.

4) Your opinion on metformin effectiveness has to be backed up by sources explicitly stating what you say, otherwise, again, it is WP:SYNTHESIS

5) The widespread misconception that if something is "similar to our natural compounds, it should be more effective and safer" is a known falsehood. In our stomachs we have HCl, but you wouldn't want to drink it regularly. Hormones are extremly biologically powerful compounds and taking them without extreme precaution can create a lot of harm. And again, since it is what you think, it is WP:OR

6) Since these points have been repeatedly established, I invite our fellow Testosterone_vs_diabetes to drop the stick on the dead horse and to understand that this is not a soapbox. I invite him/her to collaborate constructively by providing sources and discussing them here, and eventually adding statements that are a)firmly and explicitly backed up by sources b)scientifically consensual. But this article is not going to say that testosterone is better than metformin, because 1)it is not backed up by sources 2)even if it was, it is not the right place to include such information.

I hope we can put an end to this debate. --Cyclopia^talk 19:56, 3 November 2009 (UTC)

Cyclopia there is the full consensus for testosterone use against diabetes. You also inserted articles focused on this topic. Testosterone is used by many doctors for clinical (not research only).

The scientific articles state that metformin reduces testosterone, LH and tiroid hormones. I think it is a very serious side-effect.

I'm perfectly agree with the other points, indeed it is not so serious to make an head to head comparation, but as you told it could be possible to speak about all anti-diabetic drugs.--Testosterone vs diabetes (talk) 20:10, 3 November 2009 (UTC)

there is the full consensus for testosterone use against diabetes. - Find a review paper stating that explicitly and I will agree. --Cyclopia^talk 20:13, 3 November 2009 (UTC)

The literature says that metformin produces lower levels of these hormones in men. The literature then says "The clinical significance of these findings needs further investigation" (PMID 12235466) -- that is, "We don't know if this has practical effects."

For example: There are drugs that dramatically increase the white blood cell count. Taking these drugs does not, however, produce leukemia. It's just a temporary and harmless artifact.

The scientific establishment has determined that metformin reduces these hormones. They have not determined that reducing these hormones this way has any practical health consequences.

You and I might believe that it's probable that there will be practical health consequences, but we have no reliable sources that say that reducing these hormones by taking metformin (and not due to any other cause) is a problem. WhatamIdoing (talk) 20:20, 3 November 2009 (UTC)

About the "headache" joke: You might choose to read the source again. The SEVERE side effects are listed thus: "Severe allergic reactions (rash; hives; itching; difficulty breathing; tightness in the chest; swelling of the mouth, face, lips, or tongue); breast growth or pain; change in the size or shape of the testicles; dark urine or light-colored bowel movements; depression or mood changes; dizziness; gingivitis; interrupted breathing while sleeping; loss of appetite; nausea; painful or prolonged erection; stomach pain; swelling of the ankles or legs; urination problems; weight gain; yellowing of the skin or eyes."

IMO, anaphylaxis, kidney failure, liver failure, etc. are not jokes. People die from these side effects. WhatamIdoing (talk) 20:22, 3 November 2009 (UTC)

Ok, whatamidoing, I understand you perfectly...The side affects of your source are completely wrong! Testosterone does not cause that side effects (in case find articles, but I'm very sceptic). Talking seriously testosterone can have other side effects (inhibition of T endogenous synthesis, reduction of testicular volume. In some individuals is possible to have gynecomastia (effect due to estrogens which are testosterone derivatives, problem easily solved by anastrozol) and baldness (effect due to dihydrotestosterone, another testosterone derivative, problem easily solved by finasteride)

Conserning to the value of this testosterone (and other hormones too) reduction, I'm agree we have not clear results on it, on the other hand we have no reasons to think that this testosterone and TSH minor respond are different than others.

The anti-sexual effect of metformin was not studied because the people who take it have already erectile disfunction [PMID: 18727737] (it is almost impossible to have T2M with good free testosterone levels), therefore they do not recognize this effect.

I've found on internet a 30 years old user that took metformin and he started to have sexual problems http://www.progettodiabete.org/indice_ie1000.html?expert/2008/e2_02962.html

"Tre mesi fa ho ricevuto la diagnosi di IGT, e mi è stato prescritto del Metforal 500 (ora 3 compresse al giorno). Sto rapidamente riacquistando il pesoforma, a fronte di qualche banale disordine intestinale di cui ero stato avvisato. Purtroppo però, da quando ho incominciato ad assumere il Metforal ho notato una notevole diminuzione della libido: nessun problema erettile, direi piuttosto un disinteresse per l’attività sessuale."

(tre moths ago I've got IGT diagnosis, Metforal 500 (metformin) was prescribed to me (now 3 tabs per day). I'm rapidly getting the right weight, but with some intestinal disorder which I knew before. Unfortunatley, I had a strong libido decrease after Metforal use: no erectile problems, but disinterest for sexual activity).

Cyclopia, are the articles that we inserted good enough for that purpose? Is the book of Dot. Ullis a reliable source? Are the articles in a gym review reliable sources (in italian...)?--Testosterone vs diabetes (talk) 21:02, 3 November 2009 (UTC)

About WhatamIdoing concerns: If you think his source is wrong, find another source of equal reliability contradicting it.

About sources: For sure a forum about diabetes, like the one you translated, is not a RS. The "sources above" are also not good enough because -to my understanding- there is not a sentence saying explicitly that it is universally accepted as a cure. If I am wrong. quote it and link it. The book of Dr.Ullis, I don't know it, so please provide a link or more information -if it is an academic book (like a collection of reviews or an academic manual), may be. Articles in a "gym review" (??) don't look like a RS, but provide links to be sure. In short: I would consider RS recent, peer-reviewed academic reviews published in a decent-impact journal; possibly more than one. --Cyclopia^talk 21:11, 3 November 2009 (UTC)

Cyclopia, I accept the statements made by PMID 12235466. The statements made by this source are significantly smaller than the statements made by TvD. The source says that metformin (1) reduces several hormone levels (2) in men (3) in a short-term study involving a (4) very small sample size. The source also plainly and directly states that they don't know whether the lowered hormone levels produces the negative effects of hypogonadism. Thus I would support inclusion of an adverse effect: "reduced hormone levels in some men", but not anything further (e.g., "testosterone is definitely a better anti-diabetic drug than metformin").

TvD, the list of testosterone side effects comes from the US FDA's mandatory labeling for these products. The FDA is notorious for prohibiting manufacturers from mentioning any side effect that isn't proven to the FDA's satisfaction. It does not say that these are common side effects -- only that these severe side effects happen often enough for the FDA to require the manufacturer to disclose their existence to patients. WhatamIdoing (talk) 22:48, 3 November 2009 (UTC)

WhatamIdoing, sorry but it was me being not clear -I was answering to TvD concerns about you, not to you, sorry! --Cyclopia^talk 11:43, 4 November 2009 (UTC)

The list of your side effects is completely invented, do you have real scientific articles? FDA is a nonscientific source, and moreover its interests (to sell or to do not sell medicines) are too high to assume its good faith. FDA invent absoud T side effect in order to sell more remunerative drugs as in example metformin. Anyway here only scientific articles can be taken in consideration, it means your citation can not be considered. I've readen many testosterone articles and I know personally many testosterone users, I can ensure you the effects of your nonscientific source are really invented. Testosterone use can have serious side effects, but not severe. See testosterone mandatory sheets. Many people can not take metformin (because they vomit or have diahrrea), everyone can take testosterone.

Metformin is not a good anti/diabetic drug, testosterone yes. Metformin failed to cope diabetes, testosterone no.

I guess you are woman, testosterone is used in women too in order to cope osteoporosis and loss of libido (testosterone is essential in women too). Testosterone is not exactly a man hormone. --Testosterone vs diabetes (talk) 10:01, 4 November 2009 (UTC)

TvD, even if you were completely right (and while the side effects source above is indeed not the best, you have showed no proof of your position either), testosterone is not going to be mentioned here, because it is, at the very best, highly WP:UNDUE. --Cyclopia^talk 11:57, 4 November 2009 (UTC)

I guess if you can find a reliable source that uses both the words "metfomin" and "testosterone" you could argue for some passing mention but it is unlikely to be favorable or comprehensive (" some others have examined blah[]"). Personally I'm big on inclusion but then you need to argue over wording. Nerdseeksblonde (talk) 13:17, 4 November 2009 (UTC)

I can find good data only for women, comparing with oral contraceptive pill [12].--Nutriveg (talk) 13:30, 4 November 2009 (UTC)

And this which I can't call good enough.--Nutriveg (talk) 13:46, 4 November 2009 (UTC)

Although insulin sensitivity is associated with testosterone deficiency, there is no evidence that insulin sensitizers, including metformin and thiazolidinediones, are able to elevate testosterone levels in men with diabetes.--Nutriveg (talk) 13:52, 4 November 2009 (UTC)

"at least after 3-month treatment, metformin and dexfenfluramine do not modify the effects of diet on anthropometric, metabolic and hormonal parameters"--Nutriveg (talk) 14:05, 4 November 2009 (UTC)

Nutriveg, 3 of 4 our citations state that metformin reduces total and free testosterone (another one state that do not increase testosterone levels. What means? reduction?). Moreover metformin reduces Thyrotropin levels http://jcem.endojournals.org/cgi/content/full/91/1/225. Reduction of testosterone and TSH levels are very serious concerns.

Moreover I would remeber that metformin does not increase sugar absorption, it means metformin does not reduce insulin resistance.

Metformin is a not universally stated cure against T2M--Testosterone vs diabetes (talk) 17:00, 4 November 2009 (UTC)

Let's see if this makes more sense:

• YES: A substantial reduction of testosterone is usually a health problem.
• YES: Metformin reduces testosterone in some men.
• NO: Reduction of testosterone caused by metformin are very serious concerns.

We cannot make the last claim -- the claim that adds up the two previous ones as if they were algebraic equations -- because we have no reliable source that says exactly that. We have sources that say an unnaturally low testosterone level is a health problem. We have a source that says metformin reduces testosterone in some men. We do not have a source that this specific method of hormone reduction is a very serious concern. In fact, we have a source that says they're aren't sure if this particular reduction is actually associated with health effects. Therefore we cannot include a claim that they "definitely are", because we don't know that. We suspect it might be a problem, and you apparently believe it is a problem -- but we have no source that says that this effect creates practical health problems. We cannot go beyond our sources. WhatamIdoing (talk) 18:19, 4 November 2009 (UTC)

For me a 30 years old boy who have had sexual problems by metformin is a reliable source, and I'm sure there are many others.

What means "unnaturally low testosterone level"? Hypogonadism is not natural? Hypogonadism is very common and, unfortunately it is natural too. We have many user that have serious problems of hormones given by metformin (testosterone, LH and TSH levels), their reduction is always a serious problem. We have no reasons to think that "metformin hypogonadism or hypotyroidism" are different than others, therefore we have to say that assumption of metformin is harmful.

We have no serious sources stating a decrease of insulin resistance (which means better intracellular sugar absorption). We can not say seriously that metformin cures diabetes--Testosterone vs diabetes (talk) 18:49, 4 November 2009 (UTC)

TvD, we understand, but there is no point in insisting. A forum is not a RS for WP, sorry. We can add the fact that metformin sometimes reduces testorone levels, but nothing more. --Cyclopia^talk 19:39, 4 November 2009 (UTC)

Ok, and TSH too.--158.194.199.13 (talk) 20:02, 4 November 2009 (UTC)

Mechanism of lactic acidosis

I'm trying to figure out how Metformin causes lactic acidosis (if the information is out there in the medical literature. I don't want to do Original Research)

I'm going to post abstracts here to see if there's any relevant information. I only have an undergrad in biotech, so some of these abstracts are above my level.

these data indicate that biguanide-induced lactic acidosis can be attributed to acceleration of glycolysis in response to mitochondrial impairment. Indeed, the desired clinical outcome, viz., decreased blood glucose, could be due to increased glucose uptake and glycolytic flux in response to drug-induced mitochondrial dysfunction.

[13]

The development of AMPK(AMP-activated Kinase) activators which do not share metformin's modest risk of inducing lactic acidosis--apparently reflecting an inhibition of mitochondrial complex 1 that is not intrinsic to AMPK activity [14]

Metformin is, however, associated with a very low incidence of lactic acidosis because, differently from phenformin, it does not undergo liver metabolism and, as a consequence, there are no high-risk groups, displaying an impaired metabolic handling.

[http://www.ncbi.nlm.nih.gov/pubmed/7862618?ordinalpos=5&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum ] phenformin is similar to metformin.

A mechanism of action for phenformin-associated lactic-acidosis, attributable to impaired mitochondrial function arising from inactivation of Ca2+-sensitive, NAD+-dependent mitochondrial dehydrogenases (e.g. 2-oxoglutarate dehydrogenase) due to alteration in mitochondrial calcium content, is proposed.

[http://www.ncbi.nlm.nih.gov/pubmed/2449214?ordinalpos=7&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum ] When the concentration required to reduce the oxygen consumption to 75% of the control value (from 0.40 to 0.29 micromol/min/mg protein) was compared with the EC(50) value obtained in vivo, a clear correlation was observed among the three biguanides, suggesting that oxygen consumption in isolated rat hepatocytes can be used as an index of the incidence of lactic acidosis.

[15]

(from 2008) Biguanide-induced mitochondrial dysfunction yields increased lactate production and cytotoxicity of aerobically-poised HepG2 cells and human hepatocytes in vitro.

As a class, the biguanides induce lactic acidosis, a hallmark of mitochondrial impairment...In toto, these data indicate that biguanide-induced lactic acidosis can be attributed to acceleration of glycolysis in response to mitochondrial impairment. Indeed, the desired clinical outcome, viz., decreased blood glucose, could be due to increased glucose uptake and glycolytic flux in response to drug-induced mitochondrial dysfunction.

[16]

There also seems to be some attempt to argue that Metformin doesn't cause lactic acidosis. It doesn't seem valid to me, but the view is being pushed (with the assertion that those cases of Metformin related lactic acidosis which are presented are incidental to metformin therapy)

CONCLUSIONS: While the term 'metformin-associated lactic acidosis' is commonly used to depict all situations of lactic acidosis in metformin therapy, true metformin-associated lactic acidosis, i.e. one which refers to metformin and concurrent pathologies as co-precipitating factors, was never observed in the studied reports. As there was no mortality due to metformin alone, it is important that physicians are familiar with the range of other risk factors that contribute to lactic acidosis in patients treated with metformin. [17]

Here, we report a 43-year-old woman with type 2 diabetes mellitus and chronic renal insufficiency who developed hypoglycemia, hypothermia, tachycardia and lactic acidosis after a suicide attempt with a metformin overdose. [http://www.ncbi.nlm.nih.gov/pubmed/19321413?ordinalpos=11&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum ] --Ryan Wise (talk) 16:35, 24 June 2009 (UTC)

Here's some good information from various textbooks.

I propose that we remove the increased intestinal anaerobic glycolysis from the lactic acidosis mechanism section. The actual mechanism is believed to be that lactate is a substrate for gluconeogenesis, and metformin inhibits hepatic (liver) uptake of lactate to prevent gluconeogenesis. In normal individuals, there will not be a significant increase because the kidneys will simply clear it, but in those with impaired renal function there will be no renal clearance and it will build up. The previously-cited research article that the lactic acidosis mechanism is based on even said it was an unproven hypothesis, and it does not appear to play an important role in humans. Ryan Pedigo (talk) 23:36, 28 November 2009 (UTC)

From: Davis Stephen N, "Chapter 60. Insulin, Oral Hypoglycemic Agents, and the Pharmacology of the Endocrine Pancreas" (Chapter). Brunton LL, Lazo JS, Parker KL: Goodman & Gilman's The Pharmacological Basis of Therapeutics, 11e: http://www.accessmedicine.com/content.aspx?aID=958974.

"Metformin also may decrease plasma glucose by reducing the absorption of glucose from the intestine, but this action has not been shown to have clinical relevance."

From: Masharani Umesh, "Chapter 27. Diabetes Mellitus & Hypoglycemia" (Chapter). McPhee SJ, Papadakis MA, Tierney LM, Jr.: CURRENT Medical Diagnosis & Treatment 2010: http://www.accessmedicine.com/content.aspx?aID=15524.

"Lactic acidosis has been reported as a side effect but is uncommon with metformin in contrast to phenformin. While therapeutic doses of metformin reduce lactate uptake by the liver, serum lactate levels rise only minimally if at all, since other organs such as the kidney can remove the slight excess. However, if tissue hypoxia occurs, the metformin-treated patient is at higher risk for lactic acidosis due to compromised lactate removal. Similarly, when kidney function deteriorates, affecting not only lactate removal by the kidney but also metformin excretion, plasma levels of metformin rise far above the therapeutic range and block hepatic uptake enough to provoke lactic acidosis without associated increases in lactic acid production. Almost all reported cases have involved subjects with associated risk factors that should have contraindicated its use (kidney, liver, or cardiorespiratory insufficiency, alcoholism, advanced age). Acute kidney failure can occur rarely in certain patients receiving radiocontrast agents. Metformin therapy should therefore be temporarily halted on the day of the test and for 2 days following injection of radiocontrast agents to avoid potential lactic acidosis if end-stage chronic kidney failure occurs." Ryan Pedigo (talk) 23:36, 28 November 2009 (UTC)

From: Golan, E., David et.al. "Chapter 29. Pharmacology of the Endocrine Pancreas". Principles of Pharmacology, 2008.

A potentially more serious adverse effect is lactic acidosis. Because biguanides decrease the flux of metabolic acids through gluconeogenic pathways, lactic acid can accumulate to dangerous levels in biguanide-treated patients. With metformin, introduced in the United States in 1995 and the only biguanide currently available in the US, the incidence of lactic acidosis is low and predictable. Lactic acidosis is more common when metformin is taken by patients who have other conditions predisposing to metabolic acidosis. Thus, metformin should not be administered to patients with hepatic disease, heart failure, respiratory disease, hypoxemia, severe infection, alcohol abuse, a tendency to ketoacidosis, or renal disease (the latter because biguanides are excreted by the kidneys). Ryan Pedigo (talk) 00:15, 29 November 2009 (UTC)

I have edited the lactic acidosis section to reflect what is currently accepted in the medical field, please edit it accordingly. Thanks! Ryan Pedigo (talk) 23:53, 28 November 2009 (UTC)

Nice, this is actually pretty non-controversial. I am not completely happy with the statement supported by Current, though. What I gather from the actual source is that MALA in patients with renal impairment is mostly due to reduced metformin excretion, as well as lactate removal. This makes more sense than the current text. Lactate is indeed removed by the kidneys, but I was unaware that renal excretion played any real role, especially since it is all but negligible at baseline (see e.g. PMID 12225607). Fvasconcellos (t·c) 01:42, 29 November 2009 (UTC)

Thanks for fixing up and clarifying what I wrote! Also thanks for consolidating the references, I could not figure out how to do so (new to editing). I think it looks much better now. Ryan Pedigo (talk) 02:20, 29 November 2009 (UTC) Made a small edit and removed "heart failure" as one of the lactic acidosis contraindications because earlier in the article they cited a large trial that showed that heart failure did not cause issues with this. Ryan Pedigo (talk) 05:20, 29 November 2009 (UTC)

Well, it's still an established contraindication. What recent research is showing is that it doesn't have to be :) You're welcome, and thank you for your additions; welcome to Wikipedia. If you need any help with further edits, please leave me a message. Fvasconcellos (t·c) 11:41, 29 November 2009 (UTC)

general comments on drug sources to try to resolve above debate, getting off topic and soapboxing a bit but recurring problem

Usually the clinical trial results are the best source of information for documented effects in humans. All drugs approved for sale in the US in the past few years have their clinical trial details more or less available from Drugs@FDA ( with issues I mentioned above). These would make a good template item if the FDA would provide bookmarkable links ( I think you need to post or use cookies or something since links are generic, anyone know a soapbox on which to post this concern would help as I've tried to contact FDA already). Older drug trials results probably are available by FOIA as this seems the be rationale for even providing the facility at all. The AERS, Adverse Events Reporting System, is excellent for some exploratory original research but probably not useful for wikipedia and unlikely to be conclusive for any hypothesis. Case reports and open label tests have their uses and probably deserve mention but you could end up writing your own review article and there is the problem with introducing bias by cherry picking observations.

Causal inference has been a recurring problem in medicine and everywhere else- you would hope this would be better than economics but probably you would be wrong LOL. Confirmation bias, interpretation of data in a self-affirming light, is often used to turn associations or correlations into a causal link. I don't want to bash anyone but I happen to be aware of some recent trial failures at PFE and CEGE where it seems more people were killed that cured. The 2 from PFE related to a cholesterol transport inhibitor and an insulin related target for cancer- both of these relate to systems which contain a lot of literature with questionable interpretations. Certainly early results and expert trial design are no guarantee of success. This is probably due to time lag also as many trials start to look questionable prior to their results being announced due to new research literature coming out after the trial designs have been cast in stone ( or at least a monolith of paper).

On natural products, see some recent tirades of mine on vitamin C versus citric acid and anecdotes about fruit juice and more glib handwaiving analysis that can be summarized as "[oxidataive] damage must be bad." It turns out, ( I think the recently killed yale grad student was working on one class of molecules that do this) many molecules signal to each other with damaging oxygen species (ROS) and simply mega dosing on vitamin C can cause problems. In the meantime, while hyping "too obvious to be wrong" anti-oxidants, everyone ignores the "inert" critic acid in fruit juice but now see recent literature in this area.

My point here is that the best service to the reader and self is to try to examine the data carefully without inflicting our own desires and "too obvious to be wrong" attributions upon it. I started this witch hunt thinking it may be required to know al the lab procedures and look for obscure artifacts in methods but often you don't even need to look that deeply to find problems in various conclusions that other latch onto. fwiw. Nerdseeksblonde (talk) 11:23, 4 November 2009 (UTC)

Sorry, WP:TLDR. JFW | T@lk 10:41, 29 November 2009 (UTC)

Hormonal

Hypogonadism is related to diabetes mellitus, therefore if metformin's hypogonadism is related with diabetes mellitus it is unknown (not so much...anyway...)--Testosterone vs diabetes (talk) 21:36, 28 November 2009 (UTC)

Sorry, I'm not following. Could you clarify, ideally with the help of reliable sources? MastCell ^Talk 21:44, 28 November 2009 (UTC)

I thought the diabetes came before the metformin. I can't follow either? JFW | T@lk 23:49, 28 November 2009 (UTC)

The articles say that hypogonadism can cause diabetes and also that metformin cause hypogonadism. The articles are definitely reliable sources...if you see upward there is the wikipedia consensus on it--Testosterone vs diabetes (talk) 12:14, 29 November 2009 (UTC)

Metformin for diabetes + heart failure

Our lead currently says "studies suggest it may be the best choice for people with heart failure" and later in the body says "Heart failure has long been considered a contraindication for metformin use, although a 2007 systematic review showed metformin to be the only anti-diabetic drug not associated with harm in people with heart failure." Both these statements are sourced to PMID 17761999, which is a systematic review. That journal article concludes "Metformin was the only antidiabetic agent not associated with harm in patients with heart failure and diabetes. It was associated with reduced all cause mortality in two of the three studies.", which is not the same thing as saying it may be the "best choice", which is a clinical decision that involves more than just an "association" or not. Also the phrasing of the second comment places a clinical guideline (a contraindication) in opposition to research results for which the reader is left to make a judgement as to whether the guideline is still appropriate.

I've updated the ADA's "Standards of medical care in diabetes" source to the 2009 edition. This says: "Metformin is often contraindicated because of renal insufficiency or significant heart failure." It does not (as far as I can see) recommend Metformin for this patient group. The PubMed page for the 2007 systematic review lists several sources of "comments on". I haven't access to those but not all seem to confirm the result and one, titled "Review: evidence for major benefits and harms of antidiabetic agents for diabetes with heart failure is limited." indicates our article text may not be on solid ground.

I'm no expert here so perhaps someone can enlighten me as to the medical current thinking. IMO we should be noting clinical recommendations by citing clinical guidelines, not by citing studies or systematic reviews that make no such recommendations.

I am a bit concerned that our first three sources (top rate clinical guidelines which include discussions of the literature) are only used for the lead sentence. Consensus documents such as these are gold-dust and their discussions of the literature carry more weight than a literature review by a single academic. We should make more use of them. Colin°^Talk 20:48, 19 December 2009 (UTC)

Good points all around. As far as I can say after skimming through it, the systematic review compares metformin with other treatments (insulin, sulfonylureas, thiazolidinediones) and concludes that metformin does not increase overall mortality or hospital admissions. The review did not compare metformin with with no diabetes treatment, which is what we would like to know. Therefore, it is still possible that metformin may increase mortality in diabetic patients with heart failure, and saying that "metformin is best choice for people with heart failure" is unjustified. The Sceptical Chymist (talk) 13:22, 20 December 2009 (UTC)

Yes, I agree. Please reword as you feel necessary and I'll have a look later. Unfortunately, I am swamped with work and will not have time to do any editing until after the holidays. Fvasconcellos (t·c) 13:57, 20 December 2009 (UTC)

History: Metformin vs others

The statement

"Being a generic drug, it was at a commercial disadvantage; phenformin and buformin were marketed as being more potent and were supported by large pharmaceutical companies with a large sales force. As a result, metformin was mostly used in Scotland and Northern Ireland, while phenformin and buformin achieved worldwide clinical acceptance."

is sourced to this article. This source makes no pharmaceutical comparison to phenformin so we can't use it to say it was "marketed as being more potent". It mentions only one "large pharmaceutical company", the one behind phenformin. It doesn't mention buformin. It doesn't say the reason metformin has low sales or a limited market was because it was a generic drug. Rather it is merely the size of the drug company and its sales force. The article doesn't say it was "mostly used in Scotland and Northen Ireland". The article indicate the reason "metformin found an understanding reception in Belfast and in Edinburgh" was the "friendly approach" of the sales person and perhaps a hint of support for the underdog (but I'm reading between the lines here). Wrt the rest of the world, it says "In the south of England it was not widely used, and it was never marketed in the USA". This cannot be used to indicate the worldwide geographic spread of these three drugs. Perhaps it was big in Liverpool or in all of Europe? Or everywhere but the USA and the SE of England. Even if the text here was a closer match to its source material, the article is not a serious analysis of marketing or drug history, but just a personal recollection: "I well remember as a junior doctor....". It could be used to explain that metformin was marketed by a tiny drug company compared to the much larger company behind phenformin. That's about it.

If we are to take this article to FA, we need to be scrupulous about saying only what our sources let us say. Colin°^Talk 20:44, 21 December 2009 (UTC)

Furthermore, "marketed" is a term of the art in the pharma industry, and it means something much closer to "legally available for sale" than to "possesses a slick advertising campaign and a dedicated sales force". Metformin is clearly marketed in the US, as a brief visit to nearly any pharmacy's website will prove. WhatamIdoing (talk) 21:24, 21 December 2009 (UTC)

Sorry, I missed off the " at that time." wrt "marketed in the USA". The source-article isn't incorrect AFAIK on that point. Like with so many useful drugs, the US is often the last to know. But you are right we need to be careful with the word "marketed" that we don't confused our readers. Colin°^Talk 22:05, 21 December 2009 (UTC)

Please go ahead and fix the issue. Use "a serious analysis of marketing" of metformin if you can find one. The Sceptical Chymist (talk) 00:26, 22 December 2009 (UTC)

Metformin book

• Metformin: The Gold Standard: A Scientific Handbook; Ed. By Clifford J. Bailey. New York: John Wiley. 2007. ISBN 0-470-72541-9.

The first chapter is available here and covers only the herbal lead-up to metformin's discovery. The first section of the book (36 pages) covers the full history.

Should be worth trying to get second-hand, from a library or buying on Amazon market place and re-selling later. It appears to be the only book dedicated to the subject. We might fail the "well-researched" criteria at FAC without it, or at least a good reason why our sources are better. Colin°^Talk 20:44, 21 December 2009 (UTC)

I am sorry but "Metformin: The Gold Standard" is not well-researched, so I doubt that using it would help us with the "well-researched" criteria. For example, "G. officinalis is a rich source of guanidine" is incorrect. In another sentence, "biguanide derivatives [23,24]. Two of these, decamethylene diaguanide (known as Synthalin A)", the authors are confusing biguanides (metformin, buformin, phenformin) with diguanidines (synthalins). See, for example, this warning from "Archives of Internal Medicine" [18]. The Sceptical Chymist (talk) 00:51, 22 December 2009 (UTC)

Those two examples are not good examples of where this book is inferior to our sources:

1. You say the statement "G. officinalis is a rich source of guanidine" is incorrect. Have you got a source for that? The full text in the book chapter is "G. officinalis is a rich source of guanidine and related molecules, which account for its biological effects." One may parse this as "rich source of" and "guanidine and related molecules": in other words, the related molecules could form a good proportion of what is being described. Regardless, the statement is backed up by other sources, including those we currently use in the article.

• "G. officinalis was found to be rich in guanidine" from CJ Bailey(2004) Metformin: its botanical background (same author as one of the editors of the book).
• "This blooming G. officinalis (Goat’s rue; French lilac; Italian fitch) is rich in guanidine." from LA Witters(2001) The blooming of the French lilac.
• "Galega officinalis is rich in guanidine, the hypoglycemic component." from L Dey, et al (2002) Alternative Therapies for Type 2 Diabetes, which cites three sources for that fact:

• British Herbal Medicine Association: British Herbal Pharmaceopoeia. Keighley. UK; 1979.
• Hermann M. Herbs and medicinal Flowers. New York, NY: Galahad; 1973.
• Petricic J, Kalogjera Z: Bestimmung des galegins und die antidiabetische wirkung der drog herba galegae. Planta Med 1982;45:140.

• And others...

2. I think you have misread the "biguanide derivatives" section, though the text is ambiguous and could have been better written. The full chapter text is:

In addition, numerous other guanidine derivatives were studied around this time, including a series of biguanide derivatives [23,24]. Two of these, decamethylene diaguanide (known as Synthalin A) and dodecamethylene diaguanide (Synthalin B) did achieve some clinical use [25].

The phrase "two of these" refers to "numerous other guanidine derivatives were studied around this time" and not "a series of biguanide derivatives". The synthalins are "guanidine derivatives" so the text is correct.

The book is published by John Wiley & Sons Ltd, which is a respected publisher. According to the blub, "this book is edited by a world class team of expert diabetologists ... It also includes a bibliography of all papers published on metformin and a complete list of all authors on those papers." The latter fact alone makes it an important resource for anyone attempting "comprehensive research". The lead author/editor is Clifford J Bailey (see here for bio). Another author, Ian Campbell is mentioned here in a press release that discusses the book. Campbell claims "This scientific handbook reviews in detail everything there is to know about metformin and looks to its future potential uses such as diabetes prevention and other conditions associated with insulin resistance such as polycystic ovary disease, non-alcohol fatty liver and cancer prevention." Both Campbell and Bailey are senior university researchers on diabetes. Colin°^Talk 10:09, 23 December 2009 (UTC)

As I am the one who originally pointed out those inaccuracies/ambiguities, I feel compelled to say that I find them to be minor and should not be taken to reflect on the whole book until we've had a chance to read it. The claim that Galega contains guanidine is perpetuated in a lot of respectable scholarly literature (and has been since the 1920s), so I can't fault the book for reproducing it as well.

TSC, if you read half of the textbook chapters I translate and proofread, I guarantee you'd stop trusting many respected publishers :) I still dare you to send me a copy of one of those inaccuracy-free articles ;) Fvasconcellos (t·c) 12:17, 23 December 2009 (UTC)

Ah, I've only just read Fvasconcellos' workpage where you guys have discussed this book and other issues. It would still be handy to have a source to hand that says Galega contains "guanidines" but not "guanidine" itself. If this is a common mistake (as highlighted by the list above, all using suspiciously similar phrasing) then it would be good to be armed with a defence against it. From my own limited experience, I don't think historical articles (or historical sections) in medical journals are reviewed particularly thoroughly. Perhaps they are regarded as a bit of entertainment in amongst the serious research. A book has much more room to expand on a topic. On the ketogenic diet, my textbook's chapters were perhaps 10 times longer than the equivalent journal article on the same topic (and written by the same folk). When drawing on source material, it is much easier to condense than to merely rephrase an already highly condensed work. It is hard to write an engaging story when the source only contains the bones and little flesh. Colin°^Talk 13:48, 23 December 2009 (UTC)

I don't think we'll be able to find any source that says "Galega contains a guanidine derivative, but not guanidine" :) (I mean, can actual guanidine even be isolated from a plant source?) Good sources on the constituents of G. officinalis, but probably not germane to this article, are doi:10.1007/s10298-008-0284-4 and [19]. The most useful handy source is probably Drugs.com (yes). And before someone asks what makes Drugs.com reliable, this is just an online version of Facts & Comparisons, one of the most trusted and widely used sources of pharmaceutical information out there. Fvasconcellos (t·c) 14:22, 23 December 2009 (UTC)

"TSC, if you read half of the textbook chapters I translate and proofread, I guarantee you'd stop trusting many respected publishers :)" -?? -- I think this is addressed to Colin, since I am the one who does not trust that book. The Sceptical Chymist (talk) 02:27, 24 December 2009 (UTC)

Re guanidine in Galega. This issue clearly shows the pitfalls of writing based only on secondary sources. Poor quality secondary sources tend to quote other secondary sources and so on until you get to the original secondary source where the mistake was made (somebody probably mixed up guanidine derivative galegin with guanidine). I challenge you to find a primary article where somebody actually found guanidine in Galega. I was not able to do that. The Sceptical Chymist (talk) 02:35, 24 December 2009 (UTC)

So far, we've found one minor issue with a chapter on herbs. And that minor issue is repeated throughout the medical literature (including papers we cite). And we found one ambiguous sentence pair. I'm really struggling to see how the case has been made to throw-out a multi-author book, written and edited by professors who have published dozens of papers on the subject. I think it is fair to say we need to read this book, even if we don't end up citing it. We can't claim to have thoroughly researched the topic if we ignore the only academic book on the subject, a book that claims to "reviews in detail everything there is to know about metformin".

The issue doesn't show the pitfalls of writing based on secondary sources at all. Achieving distance-from-the-source will fix errors too, as well as adding the odd few. Primary sources aren't more accurate than secondary sources. History shows many to be just plain wrong, or contain errors. Anyway, the PSTS argument is not one for us to debate here.Wikipedia has made its decision and it is primarily based on secondary sources. Wikipedia will have to live with the pluses and minuses that result from that just as it lives with the "not written by named experts" decision that I'm sure many would consider loopy. Colin°^Talk 09:30, 24 December 2009 (UTC)

Metformin vs. glyburide in gestational diabetes

PMID 20027034. Might be an interesting read, despite the small sample size and statistically insignificant findings. Fvasconcellos (t·c) 12:24, 23 December 2009 (UTC)

Citation crud in the wiki body text

Suggestion: we take advantage of the new WP:FOOTNOTES system to move all the citation template crud to the References section. See discussion at Wikipedia talk:WikiProject Medicine/Archive 15#New footnotes formatting system. Have a look at the wikitext in Ketogenic diet. You can actually read the body text! In contrast, Metformin's 100 references make reading the wikitext hard work. Having all the refs in one place means you can format them readably and order them neat and tidy. I found that naming and ordering them Harvard-style (AuthorYEAR) made it easy for me to find if I'd used a ref already and reuse it. I didn't use citation templates on ketogenic diet but that's not a change I'm suggesting here. The new style is 100% compatible with the old, so if you forget or whatever and add a ref into the body text, it all still works. Colin°^Talk 20:46, 23 December 2009 (UTC)

Yes, excellent. I've been meaning to do this for a while. Fvasconcellos (t·c) 21:13, 23 December 2009 (UTC)

What do The Sceptical Chymist and WhatamIdoing think? Any other major contributors? Do you have a preference for the naming scheme for refs? Colin°^Talk 22:04, 23 December 2009 (UTC)

Sounds reasonable. JFW | T@lk 22:31, 23 December 2009 (UTC)

I've no objection, provided that someone else is willing to do the grunt work. (I keep hoping for a script that will convert whole pages at once.) WhatamIdoing (talk) 23:02, 23 December 2009 (UTC)

Meh, it's not that hard; I'll do it later if no one objects. A script would be nice, though.

Converting from citation templates to Vancouver style, or vice versa—now that would be a little bit more work :) Fvasconcellos (t·c) 23:13, 23 December 2009 (UTC)

I did the KD convert by hand but the above linked discussion mentions using a ref tool. Actually, you can convert from WP-style citations to Vancouver style by dropping the "e" off of "cite journal", etc. Eubulides has been working on some alternative templates, mostly with the aim of getting rid of the baggage the existing templates add to the HTML page. Colin°^Talk 23:53, 23 December 2009 (UTC)

It's pretty easy to switch {{cite journal}} and {{cite book}} to {{cit journal}} and {{cit book}}. I just now did that; please revert if you see problems or Just Don't Like It. Apart from the change to Vancouver format, this change shrunk the size of the article's generated HTML from 243 kB to 160 kB, a savings of 34%. I haven't yet implemented {{cit press release}} or {{cit pmid}} so I left those alone for now. Eubulides (talk) 01:06, 24 December 2009 (UTC)

Oh, I'm not sure Fvasconcellos was suggesting a move to Vancouver style as I don't think he cares for it much. I think it is fine. Expect Fvasconcellos to find the flaws in your template's formatting! Colin°^Talk 09:34, 24 December 2009 (UTC)

Heh, I am indeed not fond at all of Vancouver style. It could just be the result of staring at it nearly every day, though :) Seriously, I haven't checked the formatting, but I'm sure it's fine. Fvasconcellos (t·c) 12:32, 24 December 2009 (UTC)

Deletion of Trade Names

The alternate trade names should be kept. Without the trade names, if one does a search on, for example Obimet, a reader arrives at the Metformin page without any indication of why the redirect occurred, leaving the reader puzzled. The identification of Metformin as a INN name is useful to people who do not know about INN, and who then can be directed to the article on INN. Deletion of useful material that one person thinks is superfluous is inappropriate. There are at least two people who think the information is reasonably included, the person(s) who included it in the first place, and myself. Darrell_Greenwood (talk) 19:30, 2 December 2009 (UTC)

Yes, I agree. Please see discussion at User:Fvasconcellos/Fvasconcellos#Metformin history; you're more than welcome to join in and help further improvement of the article! Fvasconcellos (t·c) 23:49, 2 December 2009 (UTC)

Thank you. Darrell_Greenwood (talk) 04:43, 3 December 2009 (UTC)

I also support the inclusion of the trade names, and I am sorry to see MEDMOS inaccurately invoked as an excuse for removing trade names, because MEDMOS permits them. WhatamIdoing (talk) 00:12, 4 December 2009 (UTC)

MEDMOS recommends but not, of course, mandates only the main trade name (The lead should highlight the name of the drug as per normal guidelines. The BAN or USAN variant may also be mentioned, with the word in bold. The initial brand name and manufacturer follows, in parentheses.) However,WP:FACR mandates the article to be well-written -- and the long names list makes the lead less readable. I also challenge DG and WAID to actually write some content here, instead of reverting or engaging in non-productive discussions. The Sceptical Chymist (talk) 11:46, 4 December 2009 (UTC)

TSC, there is absolutely no need to be combative. The reader's point of view is of paramount importance—no one needs to add content to state an opinion or have it taken into account. You yourself are reverting to what is still clearly a minority opinion. And why remove the pronunciation? This is an encyclopedia article; many things we believe to be self-evident are, in fact, not at all so. Fvasconcellos (t·c) 12:58, 4 December 2009 (UTC)

Ages ago, we had a proposed outline for the lead sentence of drug articles, which would go something like this:

Drug name (INN) or alternate name (USAN) is a class of drug used...

It never seems to have explicitly made its way into MEDMOS (although, if I had my way, it'd be added now). Using a format as succinct as Drug name (Trade name) is not very informative. Consider, for instance, that is reverses the usual package insert/box format of "trade name/generic name". We must note that the first name is a non-proprietary name and that the second (and third, and fourth if need be) is a trade name. This explicit format is used in Bupropion and Linezolid and, more succinctly, in Sertraline. In fact, in Linezolid it is expanded further into a more prose-like solution, with trade names noted and explained at the end of the first paragraph. Fvasconcellos (t·c) 13:06, 4 December 2009 (UTC)

I respectfully disagree. The letter of WP:MEDMOS does not recommend multiple trade names. The spirit of Wikipedia is to make the article readable. The invocation of hypothetical idiot-readers who cannot read the label does not hold water. Such readers are virtually non-existent.

For example, according to stats.grok.se [[20]] in November there were

• 64078 searches on Metformin
• 1170 - on Glucophage
• 17 - on Glucophage_XR
• 46 - on Riomet
• 41 - on Fortamet
• 97 - on Glumetza
• 6 - on Obimet
• 0 - on Dianben
• 58 - on Diabex
• 44 - on Diaformin

• • Thus, all other names account for 0.5% of all searches. To insist that they are useful is to defy common sense. The Sceptical Chymist (talk) 04:40, 5 December 2009 (UTC)
    • I'm sure those statistics will be useful. Can we agree to not touch this particular area until a consensus is reached? Fvasconcellos (t·c) 11:19, 5 December 2009 (UTC)
    • And this is no "invocation of hypothetica idiot-readers" (!). See this excellent argument for inclusions that most of us would find blindingly obvious. Fvasconcellos (t·c) 14:24, 5 December 2009 (UTC)
      • Agreed with the first part. Thank you for listening, and working constructively and creatively towards consensus. It still would be good, as we agreed, to find 2-4 most important trade names. The above statistics indicates that only Glucophage is worth mentioning. I'll mark the section with "needs consensus" to serve as a reminder. Is that OK with you? If not, please feel free to remove it. The Sceptical Chymist (talk) 14:47, 5 December 2009 (UTC)
      • The second part asks, in effect, Should we exclude Bob from Wikipedia in favor of Robert? The practical answer is that both should be included (Bob was queried 9221 times and Robert - 10537). Unlike the case with Metformin/Glucophage, there is no sharp cut-off for the less used names. Let's be practical people and leave improbable hypotheticals and comparisons of apples and oranges to Judge Antonin Scalia of the US Supreme Court. The Sceptical Chymist (talk) 14:47, 5 December 2009 (UTC)```

Well, I do hate these tags (does the reader need them? No) but they can be useful. I'll ping the projects and see if anyone is interested in adding their opinions... Fvasconcellos (t·::::c) 15:29, 5 December 2009 (UTC)

Support inclusion of trade names. According to the statistic above, there are 10+ searches for one of the lesser known trade names a day. Some WP articles have less hits than that, and still are notable. I am quite sure many readers would be confused by redirects from trade names that are not mentioned in the first few sentences – at least to judge from all the people I try to convince every day that "Glucophage (active ingredient: metformin)" is the same as "Metformin Stada (active ingredient: metformin)" [I am a pharmacist] :-)

I must admit that lots of bold-faced trade names in the lead look ugly, but I am sure accessability is more important than looking nice. Also, I seem to recall an article that solved the problem along the lines of "The drug is sold under the trade names Trade name 1, Trade name 2 and others; for a full list see [[#Trade names|below]]." --ἀνυπόδητος (talk) 16:17, 5 December 2009 (UTC)

No, you are mistaken. None of the lesser names gets more than 4 hits per day. And how can you be "sure many users would be confused by redirects"? That is a hypothetical with no evidence whatsoever. The Sceptical Chymist (talk) 02:46, 6 December 2009 (UTC)

I suggest moving this discussion to WP:MEDMOS as this debate applies to all pharmacology articles and is about wording to MEDMOS guideline. Discussion should surround this sentence "The initial brand name and manufacturer follows, in parentheses." If there are lots of brand names, I would usually restrict them to english only speaking countries. If the list is still too long (eg more than 4 or 5 trade names), then restrict further to most common/notable ones with the option of creating a trade names section further down the page. Further reasoning for my views are here, Wikipedia_talk:WikiProject_Medicine#Trade_names.--Literaturegeek | T@1k? 16:47, 5 December 2009 (UTC)

Please make your case at Wikipedia_talk:Manual of Style (medicine-related articles)#Trade names. Colin°^Talk 17:59, 5 December 2009 (UTC)

Yeah, guys, you go ahead. I am sure you will work out something reasonable. I'll do more meaningful things meanwhile. The Sceptical Chymist (talk) 02:46, 6 December 2009 (UTC)

Trade names are useful as sometimes people only know the brand name and use Wikipedia in an effort to find the generic name. I see nothing wrong with listing many of this just not too many in the lead. Could go in there own section.Doc James (talk · contribs · email) 19:48, 6 December 2009 (UTC)

"Useful for those who only know brand name..."? -- That is nothing but theorizing. The statistics show that such people are less than 0.5% of readers. The rest has to suffer through an unreadable chunk of trivia. And at what point "many" becomes "too many"? Again, query statistics (see above) suggests that only Glucophage is of any significance. The Sceptical Chymist (talk) 12:01, 8 December 2009 (UTC)

You are mistaken. stats.grok.se does not measure http referrer data but just which article is accessed, so cannot be used for what you are using it for. Stats.grok.se only measures how many times an article is accessed, not the http referrer data from search engines. So it will only count how many people either clicked on a redirect in another wiki article or else used the internal wiki search engine. For example, lets say someone typed into google "xanax", the "alprazolam" wikipedia article comes up, not the xanax redirect thus a person would be logged as visiting the alprazolam article. The about stat page,[21] says this "It counts the title the page was accessed under". For what you are using it for, it should say "it analyses http referrer data for keywords" or words to that effect. The high stats for Glucophage compared to other keywords is likely because there are wiki internal links to it on the following pages,Pill_splitting#Uniformity_of_split, List_of_drugs:_G#glu, Merck_Serono which people and search engine bots are following or clicking which is contributing to statistics.--Literaturegeek | T@1k? 12:53, 8 December 2009 (UTC)

So, what is the problem? According to you, grok.se would count how many people either clicked on a redirect in another wiki article or else used the internal wiki search engine. That is exactly what I am counting -- how many times the word was used in the internal search engine or internal redirect.

What Google does or what keyword is used in Google is of no concern to us. If a user puts Blahblahfag into Google and it gives him metformin page, that is exactly what is needed. Google essentially tells him that Blahblahfag is metformin. I see no problem here. The Sceptical Chymist (talk) 00:20, 9 December 2009 (UTC)

The tool does not differentiate between how many clicked on a redirect in an article and how many people used the internal search engine. The tool is only relevant for comparing how many people read say an article on drug abuse versus an article on alcoholism. I guess my point is that its statistics are not relevant to this discussion.--Literaturegeek | T@1k? 02:27, 9 December 2009 (UTC)

• OK, as I expected, a reasonable compromise was worked out at Wikipedia_talk:Manual of Style (medicine-related articles)#Trade names. It also seems to me that most of the participants in that discussion were against using the "listcrud" of obscure proprietary names anywhere in the article. Is my reading correct? The Sceptical Chymist (talk) 14:05, 20 December 2009 (UTC)

Yes as you probably gathered there was consensus for initial brand names by the patent holder to be mentioned in the lead, generally there only are between 1 - 3 initial brand names. There was a consensus to avoid generic trade names, so we can omit generic trade names from the article. Do you know which are generic trade names and which are brand names? It is still unclear what to do with other brand names, I will post on MEDMOS about this.--Literaturegeek | T@1k? 15:25, 26 December 2009 (UTC)

Contraindications - radiographic procedures

We currently state that "metformin be temporarily discontinued before any radiographic study involving iodinated contrast". The two cited sources (British and European) do not recommend it be discontinued "before" but rather "at the time of the investigation" or "from the time of the study". Can this text be rephrased to match the sources.

I note that

• Goergen SK, Rumbold G, Compton G, Harris C (2010). "Systematic review of current guidelines, and their evidence base, on risk of lactic acidosis after administration of contrast medium for patients receiving metformin". Radiology. 254 (1): 261–9. doi:10.1148/radiol.09090690. PMID 20032157. {{cite journal}}: Unknown parameter |month= ignored (help)CS1 maint: multiple names: authors list (link)

indicates there are five guidelines (I'm not sure if that includes the manufacturer's prescribing information leaflet) and that those guidelines don't completely agree and are based on weak evidence. I don't have access to the full paper. I think the disagreement is probably at a level of detail finer than we currently provide, but it would be good to check.

Since that systematic review is up-to-date and comprehensive, it would be best to cite it instead of individual guidelines (some of which are getting very old). Provided all the current guidelines agree on the aspects we mention in our article, this would be the best source. And if they disagree significantly, then perhaps we need to reword what we say. Colin°^Talk 22:34, 26 December 2009 (UTC)

Historical and foreign-language sources

The purpose of the References section on Wikipedia is to provide evidence of where an editor drew his material and provide a resource for readers to confirm the article text. For the former, the editor should have read the full text of the source. For the latter, it helps if the resource is easily available to readers and in English. The following "sources" are suspicious. If we only read the secondary sources that cite these primary sources, then we should say so.

• Werner E, Bell J (1921). "The preparation of methylguanidine, and of ββ-dimethylguanidine by the interaction of dicyanodiamide, and methylammonium and dimethylammonium chlorides respectively". J. Chem. Soc., Transactions 121: 1790–5. doi:10.1039/CT9222101790
• Short WF, Hobday GI, Oxley P (1948). "Improvements in the manufacture of diguanide compounds". GB Patent 610,379.
• Quoted from Chemical Abstracts, v.45, 24828 (1951)Garcia EY (1950). "Fluamine, a new synthetic analgesic and antiflu drug". J. Philippine Med. Assoc 26: 287–93.
• SUPNIEWSKI J, CHRUSCIEL T (1954). "[N-dimethyl-di-guanide and its biological properties.]" (in Polish). Arch. Immunol. Ther. Exp. (Warsz.) 2: 1–15. PMID 13269290
• Quoted from Chemical Abstracts, v.49, 74699 (1955)Supniewski J, Krupinska, J. (1954). "[Effect of biguanide derivatives on experimental cowpox in rabbits.]" (in French). Bulletin de l'Academie Polonaise des Sciences, Classe 3: Mathematique, Astronomie, Physique, Chimie, Geologie et Geographie 2(Classe II): 161–5.

Colin°^Talk 21:09, 19 December 2009 (UTC)

Sorry for missing some some of those. Check the chapter now, I fixed most of the issues except for:

• "Quoted from Chemical Abstracts" means that a secondary source (Chemical Abstracts) was used, does not it?
• Anyone can get the full text of any European patent from the European Patent Office site.
• I did read Werner and Bell.

The Sceptical Chymist (talk) 13:44, 20 December 2009 (UTC)

I don't think that the Chemical Abstracts Service is a secondary source. My impression is that it's more like PubMed. Reprinting the abstract from a primary source doesn't make the abstract become a secondary source. WhatamIdoing (talk) 22:44, 20 December 2009 (UTC)

Agree with WhatamIdoing that if "Chemical Abstracts" merely lists an abstract, then it is no better than a PubMed abstract. The "Quoted from" phrase threw me as I assumed the article was making a direct quotation, not an indirect citation. The patent above is a GB patent, not a European one. The GB patent office indicates it is not available online. The issue isn't so much whether readers could track down these sources (though it wouldn't do much good in the case of the Polish or French sources for most readers), but whether editors here actually consulted them directly as sources. Also, without secondary sources, we are at a loss to gauge whether some facts (such as the Boots Pure Drug Company patent) is relevant/notable in our history story. Colin°^Talk 19:43, 21 December 2009 (UTC)

Since I did not read the original references, I referred to abstracts from CAS as is customary in scientific literature. PubMed does not have abstracts for these articles, otherwise I would have used PubMed.

Let me repeat. Anyone can get the full text of any European patent from the European Patent Office site, and that is exactly what I did with GB 610379. You can do that too. GB is in Europe, by the way. The Sceptical Chymist (talk) 00:19, 22 December 2009 (UTC)

Wrt the abstracts: are these "abstracts" a summary of the original paper or are they in fact the entire original work? If the former, this this is no better than seeing the abstract on PubMed and citing the entire paper as a "source". I don't know how the chemistry scientific literature works, but I was taught at university not to cite papers I hadn't actually read. And Wikipedia is not a scientific literature review: our references section isn't a list of seminal works but is a list of sources we've consulted.

I have eventually managed to get the European Patent Office site to divulge the patent: searching by number didn't work. The use of this source fails WP:NOR, which states "Without a secondary source, a primary source may be used only to make descriptive claims, the accuracy of which is verifiable by a reasonable, educated person without specialist knowledge." I cannot tell that the patent describes "an improved method of preparing metformin and its analogs". The patent claims "This inventoin relates to the production of organic sulphonic acid salts of diguanides and to the diguanides which may be prepared from them and in particular to the sulphonic acid salts of either diguanide itself or such diguanides as are substituted in one NH₂ group by one or more alkyl, aralkyl or aryl groups.". Metformin, being a biguanide, doesn't seem to be directly within the scope of the patent. Perhaps my ignorance of chemistry is showing here. Even if metformin is/was mentioned as the desired product of the improved process, we need a secondary source in order to gauge whether this fact is at all relevent or notable in the history of the drug. Colin°^Talk 09:05, 22 December 2009 (UTC)

"I was taught at university not to cite papers I hadn't actually read" - Ask somebody in the know. You will find out that citing an abstract is allowed for difficult to obtain references, if you disclose that.

"Perhaps my ignorance of chemistry is showing here." - True. The patent is called "Improvements in the preparation of diguanid compounds". Diguanides or biguanides are metformin and its analogues (do not confuse with synthalins which are diguanidINes). Example 3 is the preparation of metformin. The patent application was approved by the Patent Office, thus confirming the claims on a better method for the preparation of metformin.

"we need a secondary source in order to gauge whether this fact is at all relevent or notable in the history of the drug." - No we do not. Determining whether the fact is notable is up to us. You are welcome to present your arguments as to why the Boots patent is superfluous. The Sceptical Chymist (talk) 03:04, 23 December 2009 (UTC)

If the citation is for the entire paper (or book, or whatever), then, no, no academic institution will accept reading only the abstract (or introduction, or first page, or appendix, or whatever). Some citation systems have specific methods for citing partial texts (|title="Abstract of 'Title of Paper Here'" is a simple approach). Since abstracts don't always quite match up with the promised content, relying on their completeness and accuracy is a dangerous practice.

You seriously misrepresent the patent examiner's job. Patents require only that the invention claimed be novel and non-obvious, not that it is a "better" production method. It happens that patented compositions and methods are frequently improvements, but the patent office has no way of verifying any information presented in a specification, and it frankly doesn't care. People patent processes that are known to be worse than existing systems every day of the year. This process might be better, and it might not be better, but the patent office's actual finding -- which is that it is novel and non-obvious -- does not give you license to claim that any source says that it's better. WhatamIdoing (talk) 07:55, 23 December 2009 (UTC)

I accept that if our citation explicitly cites an abstract then we are saying "the abstract (not the paper) is our source". But a condensed snippet of a primary source isn't an ideal choice for Wikipedia.

Thank-you for clarifying the "Diguanides or biguanides" issue: so these are synonyms? I agree with WhatamIdoing that this doesn't indicate it was "better". Indeed the "best" system for preparing those compounds might have already been patented and the Boots patent is a 2nd-best-but-avoids-Bayer's-patent. We do need a secondary source. Please read WP:WEIGHT. If no secondary sources mention the Boots patent wrt the story of metformin then mentioning this fact is very likely to "be disproportionate to [its] overall significance to the article topic". Same goes for EY Garcia and his flu: I've failed to find any online reference to his paper, or him and "Fluamine". The factor that determines the WP:WEIGHT to give to a viewpoint or fact is "prevalence in reliable sources, not its prevalence among Wikipedia editors".

Our history section contains too much original research. The core story and all the key facts must come from secondary sources. By all means read and cite the historical papers/patents to get a little extra info to add some flesh to the story, if that is required. But it is not our job to be historians. WP:PSTS: "Wikipedia articles should rely mainly on published reliable secondary sources". Colin°^Talk 09:10, 23 December 2009 (UTC)

"Same goes for EY Garcia and his flu: I've failed to find any online reference to his paper, or him and "Fluamine"" I categorically disagree with the implication that Garcias role is insignificant. Garcia ran a first study of metformin in humans ans was first to note its blood sugar lowering effect in humans. Indeed Sterne in one of his interviews says that a paper prom Philippines gave him the idea to study metformin. I submit that Sterne's role in discovering metformin is insignificant. He did not do anything new. He was just lucky not to be working in a 3rd world country.

That is a statement of opinion, which has no place in a Wikipedia article. Sterne's role in discovering metformin may indeed have been insignificant, but his role in developing metformin as an anti-diabetic was pivotal. Surely you do not take issue with that. Fvasconcellos (t·c) 14:04, 24 December 2009 (UTC)

My point exactly. Excluding Sterne is against common sense, so is excluding Garcia, even though Baily could not be bothered to look up Garcia's paper. The Sceptical Chymist (talk) 16:59, 24 December 2009 (UTC)

• There is no OR in the History section. Using primary sources is allowed when describing facts.

• There is no such sentence in WP:WEIGHT -- "Wikipedia articles should rely mainly on published reliable secondary sources".

• The following rule from WP:WEIGHT--Wikipedia aims to present competing views in proportion to their representation in reliable sources on the subject.-- does not apply because we are not talking about including a viewpoint but about including a fact.

• However, this rule from WP:WEIGHT applies: "Undue weight applies to more than just viewpoints. An article should not give undue weight to any aspects of the subject but should strive to treat each aspect with a weight appropriate to its significance to the subject." I submit that Garcia's research was a significant development in the history of metformin (see reasons above). Similarly, Boots patent is significant because it represents the first example of preparative synthesis of metformin in the literature. The Sceptical Chymist (talk) 03:05, 24 December 2009 (UTC)

• "If the citation is for the entire paper (or book, or whatever), then, no, no academic institution will accept reading only the abstract"--Wrong. For example in the ACS Style Guide, p 185, the following citation is an example: "Ferch, H. Highly Dispersive Filler for Emulsion Paints. Plastics, Paint Rubber 1966, 10, 85-86; Chem. Abstr., 1966, 17198g." Another example, from JACS, one of the most prestigious chemistry journals [[22]], citation 23 reads: "23.(a) Sakamoto, N., Tan, H., Hata, E., and Kihara, N. JP 04164087, 1992; Chem. Abstr. 1992, 117, 192139.(b) Tan, H., Sakamoto, N., Hata, E., Ishitoku, T., and Kihara, N. US 5037977, 1990; Chem. Abstr. 1990, 113, 6663."

• "You seriously misrepresent the patent examiner's job. Patents require only that the invention claimed be novel and non-obvious, not that it is a "better" production method." -- Wrong again. There are actually five patentability requirements. One of them is usefulness. For example in the US Law: "Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor". Clearly, a process that is worse than the existing one, would not be useful. The USPTO’s official interpretation of the utility requirement "provides that the utility of an invention has to be (i) specific, (ii) substantial and (iii) credible... when the examiner has reason to believe that the claim is not for a practical application that produces a useful result, the claim should be rejected." Thus the patent examiner has to understand the claims of "improved process" and judge them credible in order for the patent application to be approved. The Sceptical Chymist (talk) 04:29, 24 December 2009 (UTC)

• The Boots patent is titled "Improvements in the manufacture of diguanide compounds". Any issue with what it describes can be immediately resolved by following the source. Simple as this. Fvasconcellos (t·c) 14:04, 24 December 2009 (UTC)

• The sentence "Wikipedia articles should rely mainly on published reliable secondary sources" is from WP:PSTS as stated. It is a solid aspect of policy that has been around for years and stood many attempts to question it from those who would rather cherry pick primary sources to make their POV (I'm not suggesting that here, but it is one of the reasons the policy is so important).
• The viewpoint vs fact issue is a non-issue as shown by the next part of WP:WEIGHT you quote, which goes on to say "discussion of isolated events, criticisms, or news reports about a subject may be verifiable and neutral, but still be disproportionate to their overall significance to the article topic". In addition, the presentation of a fact from history as a "key" fact, is a viewpoint. And I repeat from WP:WEIGHT: "Keep in mind that, in determining proper weight, we consider a viewpoint's prevalence in reliable sources, not its prevalence among Wikipedia editors".

The issue here is that you surreptitiously change the subject of WP:WEIGHT. In dealing with viewpoints we do need secondary sources to stay neutral, in dealing with facts it is sufficient to use reliable primary source. The Sceptical Chymist (talk) 12:59, 24 December 2009 (UTC)

• I've now found Garcia mentioned in Bailey's "Metformin: its botanical background" paper, which we cite. I'm still getting up to speed here. It would really help if you would make your case source-based rather than "I submit that...". All that matters is what our sources say.
• The Boots patent does not establish that it was "the first example of preparative synthesis of metformin in the literature". To claim first, you need a source that has done the literature search and where the source says the Boots patent was first. Without that, this is WP:OR. Even without the claim of "first" or the issue of whether it is at all notable, that Boots patent concerns the preparation of metformin is not "verifiable by a reasonable, educated person without specialist knowledge" -- ie. me. We need a source that claims Boots' patent is notable in this story.

The Boots patent does concern the preparation of metformin. I am sorry that your knowledge is insufficient to verify that. Most of chemistry, physics and mathematics articles would have not existed if they needed your understanding. The Sceptical Chymist (talk) 12:59, 24 December 2009 (UTC)

• We've already accepted that explicitly citing "Chem. Abstr." makes it clear one hasn't read the paper, so the examples you cite are doing the same.

• It looked to me like WAID said it is unacceptable, so I had to present additional evidence. The Sceptical Chymist (talk) 12:59, 24 December 2009 (UTC)

• "Clearly, a process that is worse than the existing one, would not be useful." Not true and being better than all previous methods is not a requirement of patent law (nor could it possibly be verifiable by a patent clerk). How does one measure "better"? Is it a process with fewer steps, involve cheaper ingredients, produce fewer by-products, produce more (useful) by-products, make use of existing machinery, use less energy, take up less room, be less likely to go bang... This sort of argument over what a primary source does or does not say or establish is exactly why Wikipedia strongly prefers secondary souces.

Colin°^Talk 09:11, 24 December 2009 (UTC)

• I see here that you have your own interpretation of patent laws. The Sceptical Chymist (talk) 12:59, 24 December 2009 (UTC)

Why are you arguing about some interpretation of US patent law? This is a GB patent, which only requires it

"be new ; have an inventive step that is not obvious to someone with knowledge and experience in the subject; be capable of being made or used in some kind of industry and not be ...( a few things ) "

The authors of that patent may claim it is an improved process. Big deal. One can find some aspect of any novel or tweaked process that someone could subjectively call an improvement. That's not notable and the sentence in this article currently stands alone and without context. Companies file patents all the time. By including this patent in our short history section, we are making a bold claim that it is significant in the story of metformin. So far, the person with that viewpoint is The Sceptical Chymist. If you can find published reliable sources that share this view on the Boots patent, we can make progress. Colin°^Talk 14:09, 24 December 2009 (UTC)

Where are reliable secondary sources that state that Boots patent is of no note? By excluding this patent from our short history section, we would be making a bold claim that it is insignificant in the story of metformin. So far, the person with that viewpoint is Colin. The Sceptical Chymist (talk) 15:10, 24 December 2009 (UTC)

It doesn't work that way round. Notability has to be proven. Lack of notability can't be proven in the same way that I can't cite any sources saying your are not Santa Claus.

What is to stop me listing any one of the 800+ patents that the European Patent Office lists with "metformin" in the title or abstract? Why shouldn't we mention, for example, that in 2008, Fertin Pharma patented Metformin chewing gum?[23]. Or what about listing any of the 5799 papers concerning metformin that PubMed lists, or one of the 2644 papers that have "metformin" in their title? Perhaps our readers need to be told of A Parrow's research into the treatment of DB with metformin, published in 1968, in Swedish?(PMID 5742151) The thing that stops me is WP:WEIGHT. Colin°^Talk 15:57, 24 December 2009 (UTC)

Common sense and the fact that you never get consensus is to stop you from that. The Sceptical Chymist (talk) 16:53, 24 December 2009 (UTC)

Do I understand your position correctly: no information from a (non-review) scientific paper/patent (primary sources) can be included in Wikipedia unless this scientific paper/patent was cited somewhere in a review (secondary source)? The Sceptical Chymist (talk) 16:50, 24 December 2009 (UTC)

No, but it not ideal. If the WEIGHT is obvious or undisputed then it may be fine. We're trying for FA here, not just expanding a stub with what scraps we can find. As a reader, I can't make sense of why the Boots patent should be mentioned in the article. There's no context. When I look at the source there's no context. It is just some random fact like the chewing gum patent. I'm sure Merck has dozens of patents on improved processes for making metformin tablets. What makes the Boots one special? BTW: the previous sentence "Interest in metformin, however, picked up at the end of the 1940s" is unsourced. That patent certainly doesn't let us make such a claim. Colin°^Talk 18:01, 24 December 2009 (UTC)

So your personal view is that no information from a (non-review) scientific paper/patent (primary sources) should be included in Wikipedia unless this scientific paper/patent was cited somewhere in a review (secondary source). Would it be correct to sum up your reasoning thus: The significance of a primary source is unknown without a secondary source referring to it. Including such a primary source would require a presumption of its importance, which is an OR.

What makes the Boots patent special? It is a first preparative synthesis of metformin described in the literature. The sentence: "Interest in metformin, however, picked up at the end of the 1940s" is a simple summation of a fact that there were 2 publications on metformin until then, and 7 between 1948 and 1956. Summations are allowed. If you insist though, we can say "according to Chemical Abstracts, there were 2 publications on metformin until 1948, and 7 between 1948 and 1956." or "number of publications on metformin increased after 1948." Any reasonably intelligent reader can go to the library and count the CAS hits for the corresponding years. That does not require any special education. The Sceptical Chymist (talk) 18:35, 24 December 2009 (UTC)

TSC, your claims about the patent process are simply wrong, but this isn't the time and place to run you through an agent's course. The fact is that if you want to write "This patented process is better than the previous one," then Wikipedia requires you to provide an independent/third-party source that actually makes this claim, in words that are plain enough that no one could possibly believe mean anything other than "This is better than that, and that is worse than this." The Boots patent might be special in the real world, but Wikipedia requires a source that says it is special. WhatamIdoing (talk) 19:48, 24 December 2009 (UTC)

As I said, this is not ideal. Just as WP:V doesn't require every uncontestable fact be sourced, neither do we require a second-opinion for every issue that may concern the notability of a fact. The notability of the Boots patent is not established by our source and isn't explained in the body text. It is not the significance of the primary source that is the issue (the actual patent record, and whether someone cites it, is irrelevant). It is the significance of the event or fact that the primary source documents (Boots doing some research, inventing an improved process, holding a patent).

I'm afraid the analysis and interpretation of the Chem Abs. search results is original research. And database search results aren't considered reliable sources on WP (they aren't, strictly speaking, published material). If you want, I'm sure there's some noticeboard you could ask. The important point is that if these key-event issues were important to the story of metformin, somebody, somewhere would have covered it. If they haven't, and we feel they should have, Wikipedia is not the place to right that wrong. Nobody will complain we weren't comprehensive if we fail to mention something nobody else has mentioned. Colin°^Talk 22:24, 24 December 2009 (UTC)

The important point is that if these key-event issues were important to the story of metformin, somebody, somewhere would have covered it. This could be true if metformin were a big name drug or had multibillion sales. Then the pharma money and publicity machine would have taken care of that. However, most of the historical material about metformin has been written by a couple of diabetologists or promoted by a mid-ranking pharma company. It is far from comprehensive. The said diabetologists are not better historians than you and I. This follows from the inaccuracies in their writings and their failure to find the groundbreaking Garcia article, which would only require a 5-minute database search. The Sceptical Chymist (talk) 12:02, 26 December 2009 (UTC)

I realize this can be excruciating for a research scientist (as I presume you are), but that is of little relevance to Wikipedia. For our purposes, these diabetologists are better than you and I because their work on the issue at hand has been published and partially constitutes the scholarly literature on the matter. To mention an under-recognized contribution (such as Garcia's) while abiding by WP:WEIGHT, as you have done, is good editorial judgment. To sidestep the existing secondary sources because of their flaws (which are, I maintain, minor) and favor the primary literature creates the problem of veering into WP:OR territory. Bearing in mind that (a) we are talking about historical information and (b) the patent was being used to make a single descriptive claim, I don't think we have deviated from source-based research yet, on this point, but we are dangerously close to it. Fvasconcellos (t·c) 14:09, 26 December 2009 (UTC)

All I can say is that WP stresses good quality sources. Most often they are secondary. But exceptions do exist, and I am abiding by WP:IGNORE, that is by common sense. I also did not sidestep those Baily sources completely; please look at what I cited. By the way, I am trying to get Colin agree not to cite any guidelines at each other (Jove be my witness: we know them well enough :(( ) and rather argue on merits and common sense. Care to join the pact? The Sceptical Chymist (talk) 14:27, 26 December 2009 (UTC)

I personally feel this whole patent debacle is a moot point in light of recent changes to the section, but if you'd rather continue, be my guest(s). I'll go make some popcorn. Fvasconcellos (t·c) 14:36, 26 December 2009 (UTC)

WAID, your claims about the patent process are simply wrong, (as are your claims about referring to abstracts in scientific publications) but this isn't the time and place to run you through an agent's course. Rather, I would like to point out that your comments here are not constructive. They distract us from our main goal - writing a good article, instead of helping us to do it. In the last month you wrote 434 words on this Talk page and not a single one adding content to the article. How about reversing the ratio -- one word in Talk per 500 words in the article? The Sceptical Chymist (talk) 12:17, 26 December 2009 (UTC)

As I have mentioned before, discussion can be as constructive as content work, and often more so. I'm sure WAID has had excellent reasons for confining her contributions to discussion on improvement of the article rather than editing the article directly; this builds consensus and avoids unnecessary back-and-forth in mainspace edits. Fvasconcellos (t·c) 14:09, 26 December 2009 (UTC)

It "can" be, but in the case of WAID it is not. OK, I agree 1:500 is an impossible task for anyone. How about writing at least 50 content words per 500 counterproductive words? Considering how prodigious WAID is with her opinions, even 10:1 ratio will be a great help to us. The Sceptical Chymist (talk) 14:27, 26 December 2009 (UTC)

Well, you're not so shy with those opinions yourself, good sir :) Why don't we all agree that sarcasm is really helpful and move on before good contributors are driven away by a discussion that, to use a woefully stale metaphor, is clearly generating more heat than light at the moment? Fvasconcellos (t·c) 14:36, 26 December 2009 (UTC)

My point exactly. I am trying to get WAID to be a meaningful contributor. And I love your metaphor! To continue with it, interactions with you generate mostly light, with Colin - both light and heat, but WAID only generates heat through friction, and if there ever any light, it comes from flames. The Sceptical Chymist (talk) 14:59, 26 December 2009 (UTC)

Actually, metformin is a billion dollar drug and there are some aspects of that I'll bring up later. Whether or not you or I think we could do better than the authors of our sources is a fact of life we will have to live with and move on from. They got their material edited and published and they have a reputation and they have qualifications longer than your arm and they have more experience with the drug than I probably have years of life. Remember, verifiability, not truth is the standard we must uphold. We have to try hard to be neutral reporters of what our sources say, not investigative journalists trying to tell the story nobody has heard.

Please don't turn this personal regarding who contributes what. We aren't just "writing a good article". We are writing an article on Wikipedia, which is aiming for Featured status. Few of Wikipedia's policies and guidelines are concerned with writing a good article, like one might read in a book or magazine. But they fundamentally affect how we build an article, so they are important. Colin°^Talk 14:52, 26 December 2009 (UTC)

Actually, metformin is not a billion dollar drug, at least not in the US (see here,[http://drugtopics.modernmedicine.com/drugtopics/Clinical+News/2008-Top-200-generic-drugs-by-retail-dollars/ArticleStandard/Article/detail/597086?contextCategoryId=7604 it may be useful later). Even if you talk worldwide, there is no way it could be a multibillion dollar drug. Most likely, somewhere barely above $1 billion. The Sceptical Chymist (talk) 15:17, 26 December 2009 (UTC)

Glucophage was worth $2 billion, and fell to $200 million when it lost exclusivity in the US.[24] The article you list mentioned generics sales. The brands Glucophage and Glucophage XR combined exceed $500 million in the US. Then add the combination drugs (Glucovance is over $400 million). And Europe is much the same size as the US. Plus the rest of the world.

The story of metformin, in the US at least, should cover how one company achieved exclusivity for a drug long out of patent and how it campaigned to prolong that exclusivity through methods that achieved bad press for big pharma and allegations of corporate influence on US politics. Colin°^Talk 17:24, 26 December 2009 (UTC)

1) Sales numbers are not quite correct. Glucophage was $ 2 billion. The brands Glucophage and Glucophage XR combined exceeded $500 million in the US in 2003 (you refer to a 2003 chart). In 2008 they were below the radar. Glucovance is a separate medication under a separate patent, and also was below the radar in the US in 2008. The US is about 40% of the world pharmaceutical market and 20% (relative) larger than Europe. Overall, worldwide metformin sales barely above $1 billion is a charitable estimate. The Sceptical Chymist (talk) 05:43, 27 December 2009 (UTC)

2) That would be interesting. Is it possible to do that without breaching into serious OR? The Sceptical Chymist (talk) 05:43, 27 December 2009 (UTC)

(outdent) OK. It is no longer such a big drug. Is the reason for this fall just that it has become dirt cheap or has another drug or class of drugs taken market share. If the latter, we should perhaps indicate what has superseded it.

It was at one point a multi-billion dollar drug (just). The fact that we were talking billions (and a major part of a drug company's turnover) meant that millions were spent influencing legislation, conducting trials in an effort to find new licence indications, and developing minor variations that would take over when the generics came. From the brief searches I've done, I think telling this story should be possible by consulting respectable newspapers and business journals. We may need to find someone with access to the historical archives of some US newspapers. Let's start a new section on this. Colin°^Talk 08:08, 27 December 2009 (UTC)

Break

In removing the sentence that the Boots patent represents an "improvement", we have made some progress. However, we still have related problems: a source from 1959 is hardly a good source for determining what the "usual" (=modern) synthesis method is, or that the process named in one patent is described in "multiple later patents and publications".

Overall, this looks like an effort to name the Boots patent as a reference regardless of whether it is actually necessary or appropriate (a point on which I'm keeping an open mind). WhatamIdoing (talk) 06:37, 27 December 2009 (UTC)

Book citations

• DiPiro, Joseph T.; Talbert, Robert L.; Yee, Gary C.; Matzke, Gary R.; Wells, Barbara G.; Posey, L. Michael (2005). Pharmacotherapy: a pathophysiologic approach. New York: McGraw-Hill. ISBN 0071416137.{{cite book}}: CS1 maint: multiple names: authors list (link)
• Drug Facts and Comparisons 2005. St. Louis, Mo: Facts and Comparisons. 2004. ISBN 1574391933. {{cite book}}: Unknown parameter |month= ignored (help)
• Maharani U (2009). "Chapter 27: Diabetes Mellitus & Hypoglycemia". In Papadakis MA, McPhee SJ (ed.). CURRENT Medical Diagnosis and Treatment 2010 (49 ed.). McGraw-Hill Medical. ISBN 0-07-162444-9.
• Golan ED; et al. (2005). "Chapter 29: Pharmacology of the Endocrine Pancreas". Principles of pharmacology: the pathophysiologic basis of drug therapy. Philadelphia: Lippincott, Williams & Wilkins. ISBN 0-7817-4678-7. {{cite book}}: Cite has empty unknown parameter: |author.= (help); Explicit use of et al. in: |author= (help)
• Davis SN (2006). "Chapter 60: Insulin, Oral Hypoglycemic Agents, and the Pharmacology of the Endocrine Pancreas". In Brunton L, Lazo J, Parker K (ed.). Goodman & Gilman's The Pharmacological Basis of Therapeutics (11^th ed.). New York: McGraw-Hill. ISBN 978-0071422802.{{cite book}}: CS1 maint: multiple names: editors list (link)

The first two need pages (and optionally chapters). The latter three have chapters and may benefit from pages if the chapters are long and/or listing one or two pages would help the reader identify the source.

The "Drug Facts and Comparisons" book could be updated to a more recent edition (2010 was published in October 2009).

Colin°^Talk 20:44, 26 December 2009 (UTC)

Right now, texbooks are mis-utilized IMHO. The facts cited to refs 24–26 should preferably be cited either to a good general review or to a single textbook (Goodman & Gilman's is my personal favorite). The Facts & Comparisons statement can probably be cited to the PI instead (freely available and a reliable source). I can get my hands on CURRENT quite easily, and Principles of Pharmacology is available on Google Book Search. Fvasconcellos (t·c) 21:03, 26 December 2009 (UTC)

When reading those, pay attention to the possibly controversial claims of rarity of hypoglycemia (4% for mild one, higher than placebo according to UKPDS) and about weight loss (none on the long term according to UKPDS). The Sceptical Chymist (talk) 05:15, 27 December 2009 (UTC)

The rarity of hypoglycemia is not at all controversial in the current literature: all the major recent textbooks (including the 2010 edition of CURRENT and G&G) categorically state that metformin does not induce it at therapeutic dosages (in fact, this article does not mention the classic classification of metformin as an "antihyperglycemic" rather than a hypoglycemic agent), and the pooled data in Bolen et al. supports this. Was it significantly higher than placebo in UKPDS, or just higher?

As for weight loss, yes, the current knowledge based on high-quality evidence is that metformin has no effect on body weight. Statements that claim mild weight loss should be removed or updated. The Ann Intern Med systematic reviews should be used better in this respect. Fvasconcellos (t·c) 19:47, 27 December 2009 (UTC)

Because of the size of UKPDS it has large weight in systematic reviews and is bound to determine their conclusions to a significant degree. Here is what it says in Lancet (352, 860): "Over 10 years of follow-up among patients taking therapy as allocated, the proportions of patients per year who had one or more major hypoglycaemic attacks in the conventional, chlorpropamide, glybenclamide, insulin, and metformin groups were 0.7%, 0.6%, 2.5%, 0.3% and 0% respectively; for any hypoglycaemic episode the corresponding proportions were 0.9%, 12.1%, 17.5%, 34.0% and 4.2%." The best wording would be "metformin has much lower rates of hypoglycemia than ..." and "no very rare major hypoglycemic attacks". The Sceptical Chymist (talk) 03:03, 28 December 2009 (UTC)

If you want a systematic review, see a report sponsored by AHRQ here [[25]] p. 92-98. It pretty much says it all. And there were cases of severe hypoglycemic reactions, so I scratched "no" in the above comment. The Sceptical Chymist (talk) 03:25, 28 December 2009 (UTC)

I'm familiar with that report, and I'm not as fond of it as I'd like to be. I find it rather excessively inclusive, if you know what I mean—too many small studies (and I mean small), too much heterogeneity... But never mind; it's still a good source. As for severe events, do you mean the one event reported in ten-year follow-up of UKPDS, or the 0.5% mean annual percentage reported at three years? By the way, if you have a copy of UKPDS 34, I'd love to read it. Fvasconcellos (t·c) 11:58, 28 December 2009 (UTC)

Check your e-mail. The Sceptical Chymist (talk) 14:44, 28 December 2009 (UTC)

Contraindications

Our Contraditions section currently states that "creatinine levels over 150 μmol/l" are a contraindication. The source for that is an editorial, which is an opinion piece, not a clinical guideline. The editorial contains "Suggested revised contraindications and guidelines for withdrawing metformin", which includes the arbitrary "150 μmol/l" figure. Was this suggestion adopted?

We then go on to explicitly cite the prescribing information leaflet in saying "unstable or acute congestive heart failure increases risk of lactic acidosis with metformin". I don't see any text in the PI leaflet that states there is an association of "increased risk" with the combination of metformin and "unstable or acute congestive heart failure". It isn't listed as a contraindication. The PI leaflet recommends care to be taken and in some cases that metformin be stopped. But it doesn't say there is evidence of "increased risk".

The BNF currently lists as contraindications:

• ketoacidosis, see also Lactic Acidosis below
• use of iodine-containing X-ray contrast media (do not restart metformin until renal function returns to normal) and use of general anaesthesia (suspend metformin on the morning of surgery and restart when renal function returns to normal)
• pregnancy
• breast-feeding

and goes on to specifically warn about lactic acidosis:

Metformin should be used cautiously in renal impairment because of the increased risk of lactic acidosis: it is contra-indicated in patients with significant renal impairment. NICE recommends that the dose of metformin should be reviewed if estimated glomerular filtration rate (eGFR) falls below 45 mL/minute/1.73 m2 and to avoid if eGFR less than 30 mL/minute/1.73 m2.To reduce the risk of lactic acidosis, metformin should be stopped or temporarily withdrawn in those at risk of tissue hypoxia or sudden deterioration in renal function, such as those with dehydration, severe infection, shock, sepsis, acute heart failure, respiratory failure or hepatic impairment, or those who have recently had a myocardial infarction

I've already mentioned in a section further above that the systematic review comment doesn't belong here.

I suggest our contraindications section isn't comprehensive or detailed enough and should be based on clinical guidelines (NICE, BNF, PI leaflet, FDA, professional diabetes associations, etc). Colin°^Talk 23:33, 26 December 2009 (UTC)

I doubt that liver disease is a contraindication for metformin since it is excreted through kidneys. PI does not seem to mention it. Pregnancy is not an outright contraindication, there seems to be no confirmed cases of harm. I doubt that metformin gets through the placenta or into the breastmilk. Beware of BNF, it is too brief and just lumps all side effects together without any analysis. The Sceptical Chymist (talk) 05:07, 27 December 2009 (UTC)

I believe that the guideline above is not best understood as 'metformin causes lactic acidosis in liver patients', but that decompensated liver disease (and the other conditions named) is a known risk factor for lactic acidosis, and it would be safer not to increase the risk in these temporary high-risk situations.

Colin, you might like to look at PMID 19321413. WhatamIdoing (talk) 06:55, 27 December 2009 (UTC)

The classical contraindications are just that, classical, and will be found nearly identical in the BNF, the prescribing information, and most general med/pharm textbooks. What's being discussed in the literature is the extent to which these traditional contraindications are warranted (use in pregnancy, for instance, is increasingly being found safe, although efficacy is another matter). Two "landmark" articles in BMJ and CMAJ (2003 and 2005 respectively, if memory serves) are probably worth reading, and perhaps citing, in this respect; I'll try to track them down. Fvasconcellos (t·c) 20:06, 27 December 2009 (UTC)

• Contraindications to the use of metformin. Evidence suggests that it is time to amend the list (BMJ Editorial): "If adherence to the published contraindications, all of which relate to the feared risk of lactic acidosis, were to be strict, metformin would, or rather should, be seldom prescribed at all. The British National Formulary says that any predisposition to lactic acidosis is a contraindication (http://bnf.org/). As diabetes itself isa predisposition to accumulation of lactate,3 should we stop using the drug altogether in the treatment of diabetes? The BNF and other publications also use the terms “renal or hepatic impairment.” These terms are vague and therefore unhelpful." [26]]

• Response 1: "The editorial by Jones et al (1) is extremely timely and and illustrates just how easily pharmacotherapeutic dogma can still become entrenched in the BNF and similar publications, despite a paucity of evidence." Richard Quinton, Consultant & Senior Lecturer in Endocrinology Royal Victoria Infirmary, Newcastle-upon-Tyne [27]

• Response 2: "Your recent editorial on contraindications to the use of metformin1 did not mention pregnancy, according to the British National Formulary (http://bnf.org/), another contraindication to its use. In women with gestational diabetes (GDM), or pre-gestational type 2 diabetes, whose blood glucose values cannot be controlled by lifestyle measures, insulin therapy is usually recommended. Given that pregnancy is a state of insulin resistance, metformin might be a logical alternative: it reduces hepatic glucose output, improves peripheral glucose uptake and suppresses fatty acid oxidation, cardinal features of both GDM and type 2 diabetes" William M Hague, Consultant Physician in Obstetric Medicine Queen Elizabeth Hospital, Woodville SA 5011 Australia [28]

The Sceptical Chymist (talk) 22:30, 30 December 2009 (UTC)

Adverse effects: lactic acidosis

We cite this source on "The most serious potential adverse effect of biguanide use is lactic acidosis". However, that source makes a strong case that metformin plays no role in elevating the risk of lactic acidosis and that the incidence of this symptom is no higher when metformin is used then when it isn't (for the population that is typically prescribed metformin) and that cases of lactic acidosis are therefore due to the underlying medical conditions. Also we say the risk is not increased "so long as it is not prescribed to known high-risk groups". That isn't what that review says. For example it says "In addition to the trials analyzed in this review, several other trials have confirmed that metformin treatment does not significantly elevate blood lactate levels, even in the presence of renal impairment or advanced age." Perhaps we have interpreted "prescribed under the study conditions" as "prescribed according to guidelines on contraindications". But the article shows that most metformin prescriptions fail to follow the contraindication guidelines to the letter of the law.

A recent letter in the BMJ (which I can't read the full text of) indicates blaming metformin for cases of lactic acidosis may be unjustified. Clearly this is a topic of controversy. Colin°^Talk 23:33, 26 December 2009 (UTC)

Yes, it is controversial indeed. Some authors have even suggested that metformin-associated lactic acidosis is a non-entity, while the authors of recent case reports have rather furiously disagreed. Fvasconcellos (t·c) 19:56, 27 December 2009 (UTC)

NICE guidelines

The 2008 NICE guidelines for Type 2 diabetes contain some info that would be useful:

Page 51 says:

Glucose control deteriorates continually with time in most people with Type 2 diabetes -- it is not a chronic stable condition. This is known to be due to progressive failure of insulin secretion. Accordingly therapy has to be stepped up with time, one drug added to another until such time as only exogenous insulin replacement will suffice.

We should briefly explain T2 diabetes and where metformin fits into the therapy.

Page 85 says:

An evidence call on the use of extended-release metformin preparations did not find that their use in unselected patients reduced GI side effects. differences in cost, and lack of other documented benefit, led to the conclusion that these therapies should be used only where intolerance to the immediate-release preparation had been documented.

A fairly firm "don't waste your money on these" statement.

Pages 86-87 give the recommendations, which include indications, contraindications and take-care:

R26: Start metformin treatment in a person who is overweight or obese (tailoring the assessment of body weight associated risk according to ethnic group*) and whose blood glucose is inadequately controlled (see recommendation 16) by lifestyle interventions (nutrition and exercise) alone.

R27: Consider metformin as an option for first-line glucose-lowering therapy for a person who is not overweight.

R28: Continue with metformin if blood glucose control remains or becomes inadequate (see recommendation 16) and another oral glucose-lowering medication (usually a sulfonylurea) is added.

R29: Step up metformin therapy gradually over weeks to minimise risk of gastrointestinal side effects. Consider a trial of extended absorption metformin tablets where gastrointestinal tolerability prevents continuation of metformin therapy.

R30: Review the dose of metformin if the serum creatinine exceeds 130 micromol/l or the eGFR is below 45 ml/minute/1.73 m2.

• Stop the metformin if the serum creatinine exceeds 150 micromol/l or the eGFR is below 30 ml/minute/1.73 m2.
• Prescribe metformin with caution for those at risk of a sudden deterioration in kidney function and those at risk of eGFR falling below 45 ml/minute/1.73 m2.

R31 The benefits of metformin therapy should be discussed with a person with mild to moderate liver dysfunction or cardiac impairment so that:

• due consideration can be given to the cardiovascular-protective effects of the drug
• an informed decision can be made on whether to continue or stop the metformin.

Colin°^Talk 09:09, 27 December 2009 (UTC)

I'll add some other guidelines to this pot and we can see what common ground emerges. The article currently takes a "studies have shown" approach to the drug. For example, indications are discussed as the end-product of research, which is often detailed. The important message is lost among all the percentages and evidence. Who should take the drug? Who should avoid the drug or be cautious. How is the drug introduced. What is monitored when they are taking the drug. When do people come off the drug, or have another added.

It might help to move the discussion of the effectiveness of the drug in various populations for various conditions to a section of its own. That can be free to discuss and present the best trial data, leaving the other sections to present the state-of-the-art of our knowledge. Colin°^Talk 14:25, 27 December 2009 (UTC)

Ideally, there would be an "effectiveness" section, summarizing UKPDS/DPP data, mentioning reductions in A1C levels and fasting glucose, comparing it to the sulfonylureas and insulin-containing regimens, etc. That's what I'm going for. There is certainly enough literature to support it. Fvasconcellos (t·c) 19:59, 27 December 2009 (UTC)

"Effectiveness" expressed via proxy assays (fasting glucose, A1C levels) is meaningless without real-life outcomes such as mortality and cardiovascular events. As the experience with thiazolidinediones shows, reliance on surrogate outcomes may even be harmful. There is no need to invoke A1C levels when we have good data for real outcomes. Perhaps the A1C levels could be mentioned in pharmacodynamics. The Sceptical Chymist (talk) 02:40, 28 December 2009 (UTC)

Do you really feel there's enough good data on hard endpoints (hmm, that somehow didn't come out right...) regarding metformin? The difference between agents in effect on A1C is practically negligible; I'm suggesting mention, not reliance. Fvasconcellos (t·c) 03:01, 28 December 2009 (UTC)

To get the hard data was the whole point of running UKPDS, a huge and incredibly long study. Believe it or not but they got to a 50% mortality point at about 25 years so the results are very solid. The comparative numbers are quite convincing, since other agents had minuscule effect of were actually harmful. The Sceptical Chymist (talk) 03:31, 28 December 2009 (UTC)

Metformin and cancer

Does the possibility that metformin decrease cancer in diabetic patients deserve its own section Metformin#Investigational_findings ? All the studies cited there are for diabetic patients, and metformin is mostly compared with other treatments -- not with placebo or dietary advice treatment. I suggest merging this paragraph into Treatment of Diabetes. It would make a nice additional explanation for the lower all-causes mortality in the metformin group in UKPDS 34 study. The Sceptical Chymist (talk) 00:00, 30 December 2009 (UTC)

That section isn't meant for these studies alone, although they may predominate at the moment. Their very preliminary nature is the reason they are qualified and mentioned rather offhandedly. I don't think they should be moved to the diabetes section; we mustn't allow them to "make a nice additional explanation for the lower all-causes mortality" in UKPDS because even implying that would veer out of the editorial license lane and into the original research overpass :) Fvasconcellos (t·c) 01:22, 30 December 2009 (UTC)

Very well. But please consider that all treatment of diabetics should logically be in the diabetes section. There is no difference between preventing cardiovascular disease and preventing cancer in diabetics. The purely statistical character of the cancer-prevention research is captured well enough in the text. This paragraph will be separate from UKPDS 34, and so no connection will be implied. There is other stuff I am thinking about putting into the Investigiational section, for example "metformin is one of a handful of medications that reliably increase lifespan in mammals" :). The Sceptical Chymist (talk) 15:54, 30 December 2009 (UTC)

Can metformin help with cancer in non-diabetics? I would guess that to be unlikely. In the only study addressing this issue PMID 19487376, diabetics on metformin after chemotherapy had similar pCR (quick and dirty surrogate for response) to non-diabetics, although higher pCR than diabetics on other drugs. IMHO, metformin only restored the response of diabetics to normal level. The same as in observational studies metformin only decreased the cancer mortality of diabetics to the level of general population. The Sceptical Chymist (talk) 00:00, 30 December 2009 (UTC)

Pancreatic cancer research study

Currently, the article says the following about antidiabetic drugs and pancreatic cancer:

A large case-control study conducted at M.D. Anderson Cancer Center has suggested that metformin may protect against pancreatic cancer. The risk of pancreatic cancer in study participants who took metformin was found to be 62% lower than in participants who had never taken it, whereas participants who had used insulin or secretagogues (such as the sulfonylureas) were found to have a 5-fold and 2.5-fold higher risk of pancreatic cancer, respectively, compared to participants that had been treated with neither. The study had several limitations, however, and the reason for this risk reduction is still unclear.

*Li D, Yeung SC, Hassan MM, Konopleva M, Abbruzzese JL (2009). "Antidiabetic therapies affect risk of pancreatic cancer". Gastroenterology. 137 (2): 482–8. doi:10.1053/j.gastro.2009.04.013. PMID 19375425. {{cite journal}}: Unknown parameter |laydate= ignored (help); Unknown parameter |laysource= ignored (help); Unknown parameter |laysummary= ignored (help); Unknown parameter |month= ignored (help)CS1 maint: multiple names: authors list (link)</ref>

• That the case-controlled study was "large" implies importance. I don't think this is an aspect Wikipedians should judge as it probably varies depending on condition/therapy/size-of-effect/study-type/etc. This opinion doesn't come from the study text or the accompanying editorial but from the Medscape news article (which, btw, isn't any more lay-friendly than the original article -- it is aimed at doctors after all). However, the accompanying editorial says "This single-center, case-control study included a relatively small number of patients with DM, with the conclusions generated from a few users of these medications". Doesn't look nearly so authoritative now?
• That the study was "conducted at M.D. Anderson Cancer Center" seems to be a distracting detail for this section. Perhaps in the History section, the whos, wheres and whens help make the story but here it makes me think why is this centre being mentioned.
• "suggested that metformin may protect against pancreatic cancer" this is good but comes from a primary research paper rather than a review and it needs to be said "in diabetic patients". There isn't any research on the protective effect in the general population.
• "metformin .. risk .. 62% lower .." has too much accuracy and detail. The odds ratio 95% CI was 0.22–0.69 so unless we want to indicate this wide spread to the reader (and I suggest we don't) then we should just stick with "a significantly lower risk".
• "insulin .. 5-fold .. higher risk" this could easily be interpreted by the general reader that taking insulin increases my chance of developing a deadly cancer by 5×. The full story is more complex and depends on the person's glycemic control and how long the person has been taking insulin. At <2 years, the apparent increased risk is enormous (33×). The authors admit that in the <2 years group, the cause-effect is reversed and these people are most likely taking insulin because they have actually developed pancreatic cancer (which causes/worsens diabetes). For the longer term group, they admit they have a "very small number of study subjects" and that the association with long-term insulin use requires a larger study.
• "secretagogues (such as the sulfonylureas) ... 2.5-fold higher... risk" The study produced mixed results on insulin secretagogues. Long and short-term use produced different results and they admit to having few appropriate controls. Again they request a larger study to investigate this association, largely because other studies suggest a link. The accompanying editorial also admits the results on these drugs "were less conclusive".
• "5-fold / 2.5-fold higher risk". These come from the study body-text details. However such indications of relative risk are not the best means of communicating risk to the general reader. See this article. It is probably not possible to re-phrase the details of the risk increase without engaging in OR, though it is possible to leave it less detailed (the editorial talks of a "moderately higher risk" for the long-term insulin group). There is a significant danger the general reader may grossly misinterpret the risk and solidity of the research behind the figures.
• "The study had several limitations" Many (most?) primary research papers document their study limitations. So on one level, this is meaningless when coming from the primary source. On another it may read as editorialising that the reader should beware the information. But they are given no further reasons as to why the information may be of limited value.
• Is this study a good one? One article (which reads more like an opinion piece than a review: PMID 19890202, full text) points out a number of issues with the study, the main one concerning obesity, which the authors suggest may be a more important factor for pancreatic cancer than the diabetes.
• Is this study important enough to pick out and discuss here?

We need a recent review. And here is one: PMID 20007645, full text. This tells me there are lots of interesting studies on metformin and cancer. While it does point out that epidemiological research indicates metformin lowers the risk of some cancers, it also says: "It remains unclear whether metformin exerts an actual protective effect, blocks a potential mitogenic property of insulin, or is primarily used in patients with less severe diabetes." The review also highlights the danger of citing primary research: research studies on "insulin glargine" proved highly controversial. This review summarises the current research wrt metformin and cancer and I suggest we use it instead to provide a balanced "cancer" sub-section.

I'm a bit uncomfortable with "investigational findings" sitting under "off-label use", which implies clinical use even if not sanctioned by some authority. Perhaps we need a "Research" heading, into which we could have a "cancer" sub-heading? Colin°^Talk 22:11, 30 December 2009 (UTC)

There are several reasonably well-written epidemiological studies pointing out to the decrease of cancer in diabetics on metformin. This fininding is related to the treatment of diabetes and it should be there. As I noted earlier, and in accordance with Colin's points the wording has to be corrected. I'll give it a try. The Sceptical Chymist (talk) 22:37, 30 December 2009 (UTC)

I'm a bit confused about your "related to the treatment of diabetes and it should be there". Cancer treatment or prevention is not currently a therapeutic use of metformin. We should not give the impression that it is. This is research yet to find clinical application. I agree it relates to the diabetic population only but that's no reason to include it in a section that implies current clinical use. Colin°^Talk 22:41, 30 December 2009 (UTC) Oh, and that review mentions a "tumor suppressor pathway" that seems independent of diabetes. So although the human research has necessarily been restricted to diabetics, there is some evidence to suggest a role outside of diabetes. Colin°^Talk 22:44, 30 December 2009 (UTC)

Let's be logical. Is prevention of cardiovascular effects a current indication for metformin? No. The indication is diabetes. However, cardiovascular prevention IS described there. Similarly, lower cancer incidence finding should be there. Speculations about possible applicability of this to non-diabetics can go into the Investigational section. The Sceptical Chymist (talk) 22:56, 30 December 2009 (UTC)

Possibly briefly, perhaps a short sentence or less. The fuller text should be elsewhere. Currently, the "Therapeutic uses" says very little about metformin therapy and mostly presents the results from long-term trials effects on secondary conditions. How is metformin used in the treatment of diabetes? The article just doesn't say! Colin°^Talk 23:09, 30 December 2009 (UTC)

Done. Cancer effects being unconfirmed do not deserve much space. As for the "secondary" conditions, they are what kills people with diabetes. If these "secondary conditions" were not increased, diabetes would have not been a disease. Of course, you are welcome to add info on metformin sugar-lowering ability. Not including it was an oversight on my part. FV was going to do it, I believe. The Sceptical Chymist (talk) 23:21, 30 December 2009 (UTC)

Someone needs to wield the hammer here

The article claims that hypoglycemia does not occur when using Metformin alone in standard doses. Since Metformin inhibits gluconeogenesis (just like ethanol does), this would seem a pretty broad claim to make. Reference given is 25 and 26; 25 is dead trees, 26 is a boring PDF that does deal with a somewhat related topic. I cannot find that claim in it though (may need to read more toroughly); various drugs are meta-analytically compared and Metformin seems to have a comparable low risk of 0-10% hypoglycemia episodes in several years. Futhermore, the article claims that diabetics' gluconeogenesis procudes three times as much glucose as healthy subjects'. Reference is No. 99. If you actually read the PDF which is, thankfully, actually available, you will find out that n=7. This may be irrelevant to the actual purpose of the study, which was to find out the mechanism of gng inhibition by Metformin, with isotopes and woah! However, it is certainly not irrelevant when making such a broad claim about the rate (and it was more 200% not 300%, see table 4, GNG - hepatic cycling). Even worse, the diabetic group was, on average, older than the control group. In my opinion, this is a worthless "reference". It is my impression that there are a lot of references which do not actually support anything, and that the article fell prey to people who read too much into things. --88.74.161.114 (talk) 19:46, 16 July 2010 (UTC)

I agree that a statement saying something does not occur at all is not a good one to make. The systemic review of oral medications for DMII lists hypoglycemia as "more frequent" in second-generation sulfonylureas than metformin or thiazolidinediones. The mechanism reference (the one with n=7) does seem both small and out of date (year 2000); there is certainly better and broader research on metformin's mechanism today. Perigrini (talk) 01:09, 4 October 2010 (UTC)

Just wanted to share this new article on Metformin, it has 100 notes/references. http://www.lef.org/magazine/mag2010/nov2010_The-Drug-Virtually-Everyone-Should-Ask-their-Doctor-About_01.htm Jules105 (talk) 02:52, 12 November 2010 (UTC)

Metformin use in Type 1 diabetes not mentioned

Hi all, I was recently diagnosed with type 1 diabetes, and I am being prescribed Metformin to improve my Insulin sensitivity - This study[29] states: "Adjuvant metformin can improve QOL, insulin sensitivity and glycaemic control in overweight adults with T1DM."

There is no mention of type 1 diabetes at all on the article, perhaps it would be worth adding something about it's use in type 1 patients? - Revelation13 (talk) 01:25, 16 November 2010 (UTC)

I've just noticed there's a to-do list mentioning this, oops! - Revelation13 (talk) 01:27, 16 November 2010 (UTC)

I ALSO have type 1, and have been prescribed this medication, and advised by my doctor to stop using insulin at all. since this drug reduces the glucose produced by the liver, the need to counter it with insulin is reduced. just want to add "I'm very very happy about not having to inject myself anymore, but the stomach issues are a bit of a pain, still ill take them over the shots anyday! — Preceding unsigned comment added by 108.230.190.35 (talk) 01:46, 12 September 2012 (UTC)

Metformin and glycation of LDL

Jfdwolff has twice reverted my attempt to link to a recent study, which suggests that the mechanism behind reduced incidence of ischemic events in persons taking metformin is inhibition of the glycation of LDL. The first revert objected to my use of a link to a news site rather directly to the Diabetes article. I corrected that, adding the Diabetes link to the news site link. Jfdwolff then reverted the link again. I do not fully understand this. Diabetes is a reputable peer-reviewed journal; I found the study quite interesting, but it neither make extraordinary claims nor is it particularly at odds with the previous understanding of Type II diabetes; and the study does not argue for a different standard of care for Type II diabetes, since metformin is already the drug of choice for treating Type II diabetes. In other words, the study seems neither likely to prove controversial nor likely to be misunderstood by doctors and patients in any way that could result in less effective treatment. The cite simply offers one possible explanation for the sentence that precedes it in the Wikipedia article, which says that "it [metformin] appears to decrease mortality primarily through decreasing heart attacks, strokes and other cardiovascular complications." I was careful to word the link such that it was clear this was a recent finding by a single research team. I believe Jfdwolff is being overzealous in invoking WP:MDRES. --Yaush (talk) 14:11, 2 June 2011 (UTC)

SI Units

Minor edit to include SI units ... The old kcal is now in brackets — Preceding unsigned comment added by 86.162.50.222 (talk) 14:27, 25 July 2012 (UTC)

Dispersible tables

This article does not discuss the dispersible tablets. My mother started using them recently and she is having less stomach issues with it. It would be great if someone could contribute information about these dispersible tables.Halukakin (talk) 23:07, 27 April 2013 (UTC)

Metformin - AMPK - Alzheimer's

User:Jfdwolff writes rm highly preliminary result - nobody has showed that metformin increases dementia risk (which would require at least a case control study and ideally a prospective clinical study) - let's not alarm our readers with sheer speculation

There are other studies linking Metformin with an increased risk of Alzheimer's. See here: Metformin, Other Antidiabetic Drugs, and Risk of Alzheimer's Disease. Money quote:

Results As compared with nonusers, long-term users of 60 or more metformin prescriptions were at greater risk of developing AD (adjusted OR (AOR) = 1.71, 95% CI = 1.12–2.60), but there was no consistent trend with increasing number of prescriptions. Long-term use of other antidiabetic drugs such as sulfonylureas (AOR = 1.01, 95% CI = 0.72–1.42), thiazolidinediones (AOR = 0.87, 95% CI = 0.31–2.40), or insulin (AOR = 1.01, 95% CI = 0.58–1.73) was not related to an altered risk of developing AD.

Conclusion Long-term use of sulfonylureas, thiazolidinediones, or insulin was not associated with an altered risk of developing AD. There was a suggestion of a slightly higher risk of AD in long-term users of metformin.

"let's not alarm our readers with sheer speculation" This sounds like censorship, weasle words and/or a non NPV. These aren't "speculations". They are research results. Cloudswrest (talk) 22:26, 29 April 2013 (UTC)

Biguanides for cancer - JCI review

doi:10.1172/JCI67232 would be useful to support changes in the section on cancer. JFW | T@lk 20:27, 3 September 2013 (UTC)

Primary sources

This article uses a lot of primary sources and uses primary sources to refute secondary ones. Thus not really a GA. Needs clean up with lots of good secondary sources available. Doc James (talk · contribs · email) (if I write on your page reply on mine) 16:10, 2 February 2014 (UTC)

Strange wording and personal experience

At the start of the article it stated "Metformin works by suppressing glucose production by the liver." The liver doesn't produce any glucose, it simply stores it until required. It doesn't even do that really, it stores glycogen.

I didn't add this next bit because it is personal experience but someone might like to look into it.

I have recently developed type 2 diabetes mellitus. I took Metformin, both IR (instant release) and SR/MR (slow or modified release) for 3 months. I wrote a computer program recommending doses based on BGLs(Blood Glucose Levels), from the day(s) before. I measured BGL between 10 and 30 times per day and I have the finger tips to prove it. I also counted every gram of sugar (glucose, fructose, galactose, sucrose etc. etc.) I took in from all sources as best as I could although some figures were based on averages for various foods like fresh fruit.

I managed to get my HbA1c from 11.6 to 7.5 over that 3 month period but had to increase doses to do it. Luckily I had only 2 slight hypoglycaemic episodes (BGL<4mmol/L) during that time.

My health was suffering so I discontinued the metformin and started injecting insulin. For the first 5 days my BGL was over 20mmol/L regardless of moderate/normal amounts of insulin.

The point is, the glycogen retained by the liver has to go somewhere. It seems all it did was build up in my liver for 3 months. After the metformin wore off over the 5 day period, BGL dropped significantly. My pancreas is still trying to do it's job because I now only need 25IU/day of external insulin.

I feel heaps better. I must admit I hated the idea of injecting myself every day but, with the pens and ultrafine needles, it actually hurts a lot less than measuring my BGL.Euc (talk) 17:51, 16 April 2014 (UTC)

The liver does, in fact, produce glucose, through the process called gluconeogenesis. Also, the liver does not store up glycogen indefinitely; the liver can convert carbohydrates to fatty acids.

Your personal experience sounds like rebound effect when you came off the metformin. Regardless, it really can't go in the article, since it's anecdotal. --Yaush (talk) 18:09, 16 April 2014 (UTC)

Effect on Cardiovascular Risk

On May 21 the sentence "A preponderance of evidence suggests that metformin may prevent the cardiovascular and possibly the cancer complications of diabetes" was changed, replacing the phrase "a preponderance of evidence" to "a vast body of evidence". The supporting citations included a range of pages in the British National Formulary, which said nothing about this subject, a review of studies in mice, which cannot be used to support this statement per MEDRS, and a third paper, which stated "However, randomized controlled clinical trials do not provide enough evidence for a strong protective effect of metformin on cancer incidence or mortality". Much better references are needed for this type of statement, especially if one is going to contradict a position paper of the American Diabetes Association. Formerly 98 (talk) 12:58, 4 June 2014 (UTC)

weight loss

I find it a little disturbing to see the entire page on metformin doesnt treat its use to combat obesity (Weight loss). It is commonly used throughout latin america for weight loss, and even in the USA doctors use it specifically for weight loss in non-diabetics an example Some details covering this should be added. --— robbie ^{page talk} 22:11, 25 July 2012 (UTC)

I would find it disturbing for this article to describe such a questionable off-label use of the drug. --Yaush (talk) 20:36, 3 September 2013 (UTC)

Nonsense! If it's true that people use it off-label, then that is a fact. There is no reason not to report on off-label uses as long as any claims of its efficacy are supported with evidence or are qualified as unsubstantiated. At the very least, it could serve as a warning of what not to do. — Preceding unsigned comment added by KugelaP (talk • contribs) 16:47, 4 September 2013 (UTC)

There is evidence that Metformin can cause significant weight gain. http://www.rxlist.com/glucophage-side-effects-drug-center.htm — Preceding unsigned comment added by 92.13.61.181 (talk) 23:52, 5 September 2014 (UTC)

Insight

doi:10.1056/NEJMcibr1409796 is a secondary source that elaborates on doi:10.1038/nature13270 (a primary source) that looks very much like a landmark study. It demonstrates that the mitochondrial isoform of glycerophosphate dehydrogenase is the target of metformin. This definitely needs to be worked into the article. (Thanks to the colleague who drew my attention to this!) JFW | T@lk 11:55, 7 November 2014 (UTC)

Safety in chronic kidney disease

doi:10.1001/jama.2014.15298 JAMA. JFW | T@lk 09:17, 24 December 2014 (UTC)

Combinations

Perhaps this section is not meant (or does not need) to be a complete list, but there is a recent addition, Invokamet, which is a combination of metformin and canagliflozin. 148.177.1.211 (talk) 20:11, 12 January 2015 (UTC)

Anti Aging Effect

Shouldn't this be mentioned? There are numerous studies proving it.

These need to be studies in reputable peer-reviewed journals. --Yaush (talk) 19:50, 19 November 2014 (UTC)

Your sources are not reliable sources for a medical claim. They are either popular press - bordering on tabloid - or a proposed study to test claims that has not even begun yet. The idea that metformin releases oxygen into cells is unsupported by its chemistry, and it's not clear why oxygen release would even be beneficial. --Yaush (talk) 22:34, 30 November 2015 (UTC)

Is "The Proceedings of The National Academy of Sciences" considered credible? See this paper "Metformin promotes lifespan through mitohormesis via the peroxiredoxin PRDX-2" http://www.pnas.org/content/111/24/E2501 68.7.246.26 (talk) 06:04, 1 December 2015 (UTC)

this study is done on roundworms, a species very different from homo sapiens

https://clinicaltrials.gov/ct2/show/NCT02432287

this is an ongoing trial that has yet to be concluded

Just checked - PNAS is peer reviewed, according to the FAQ page: "Are PNAS Plus articles peer-reviewed? Yes, these articles—the full article and the Significance Statement—follow the same review process as either Contributed or Direct Submissions." 68.7.246.26 (talk) 06:12, 1 December 2015 (UTC)

There is also a review article, which makes this MEDRS-compliant. See PMID 26475449. SarahSV ^(talk) 06:10, 1 December 2015 (UTC)

Which part of the paper states "antiaging effects"? Doc James (talk · contribs · email) 11:05, 1 December 2015 (UTC)

Ah found it and added a summary to the research section. Doc James (talk · contribs · email) 11:25, 1 December 2015 (UTC)

Interesting commentary on yesterday's news coverage in today's news coverage here. It's an object lesson in why we should not use news sources for health claims. LeadSongDog come howl! 16:22, 1 December 2015 (UTC)

Yup we should generally use no popular press. We should definitely never use the Sun, the Mirror, or the Telegraph. They are not news sources. They are tabloids. Doc James (talk · contribs · email) 06:05, 2 December 2015 (UTC)

• this is pop press [30] Doc James (talk · contribs · email) 12:32, 10 December 2015 (UTC)

While it is popular press, it is not a tabloid. Further, the text in question only reported news, and made no medical/health claims. To wit:

"The FDA approved in 2015 a double-blind, multi-centered trial with 3,000 subjects known as Targeting Aging with Metformin (TAME). One goal is to take use RNA deep sequencing to assess aging-related gene expression using muscle and fat biopsies of volunteers before and after taking the drug." Lfstevens (talk) 17:32, 11 December 2015 (UTC)

Lots of studies occur. We do not list them all based. Singularity Hub does not look like a very good source. Doc James (talk · contribs · email) 09:24, 12 December 2015 (UTC)

True. That's why I only write about those that have picked up general notice, this one in part because of its unusual sponsorship. Do you have any specific criticisms or examples of Singularity's bias or frivolity? Otherwise, it's just another press outlet. Lfstevens (talk) 23:39, 12 December 2015 (UTC)

If it has been picketed up, than one should be able to find a review about it or a major new source. Doc James (talk · contribs · email) 08:35, 13 December 2015 (UTC)

Useful data should already exist

The article says metformin "was introduced to the United Kingdom in 1958, Canada in 1972, and the United States in 1995." If it has a significant effect on longevity, couldn't that be detected relatively easily by looking at data from the large population of long-term users? Novel compound (talk) 14:25, 12 December 2015 (UTC)

Not really given to otherwise healthy people. Doc James (talk · contribs · email) 16:43, 12 December 2015 (UTC)

Mechanism of action

This section is wrong and out of date.

This study

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4074244/

Looks to spot on - meformin works on the mitochondria at the glycerophosphate shuttle - not due to earlier hand-waving explaination still in the article.

Good layman's explanation at http://high-fat-nutrition.blogspot.com/2015/11/just-brief-note-on-metformin.html

Summary: Metformin suppresses hepatic glucose output through decreased gluconeogenesis by inhibiting mtG3Pdh. Lots of evidence.

This also explains why they are looking at it as a possible epilepsy drug.

http://high-fat-nutrition.blogspot.com/2015/12/protons-42-metformin-as-next-epilepsy.html

OK - you can see the article is wrong - I will leave the writing up to someone else. — Preceding unsigned comment added by 108.243.106.82 (talk) 01:22, 27 December 2015 (UTC)

"Metformin has little or no effect on body weight compared with placebo in type 2 diabetes,[23] although it causes weight loss compared with sulfonylureas, since sulfonylureas are associated with weight gain.[23]"

'Weight loss compared with sulfonylureas' is non-sensical. Either metformin causes weight loss or it doesn't. Since the article claims that sulfonylureas cause weight _gain_, it would be appropriate to note that metformin causes less (or no) weight gain campared with sulfonylureas. Also, the claim that metformin has little or no effect on body weight is controversial, and should be considered for removal. There are numerous peer reviewed studies that show the opposite. If the science is unclear (and it seems to be unclear at this point), the claim should simply be removed or noted as unresolved. A peer-reviewed article claiming metformin does cause weight loss is here: http://www.ncbi.nlm.nih.gov/pubmed/9526970/. — Preceding unsigned comment added by 24.136.172.22 (talk) 10:39, 10 March 2016 (UTC)

Metformin HydroChloride is the only commercial available form?

Is there any theoretical or shown benefit from other salts, instead of HydroChloride? I would guess HydroChloride is not that good for gastric mucosa, just guessing, no hard facts.

At least in Finland one can only buy HydroChloride, I just checked 2016.05.

91.214.167.2 (talk) 07:52, 16 May 2016 (UTC)

Benefits and Adverse effects, Metformin vs Berberine?

At least for HbA1c Berberine seems to work similarly. About Cancer prevention in prediabetic and diabetic humans or rodents I don't know that much.

https://examine.com/supplements/berberine/ "The reduction of HbA1c associated with berberine, according to a meta-analysis of diabetics using 1,000-1,500mg berberine daily, was −0.72% (95% CI −0.97 to −0.47) more than placebo".

91.214.167.2 (talk) 08:02, 16 May 2016 (UTC)