Articles

More evidence that routine multivitamin use should be avoided

If scientific evidence guides our health decisions, we will look back at the vitamin craze of the last few decades with disbelief. Indiscriminate use is, in most cases, probably useless and potentially harmful. We are collectively throwing away billions of dollars into supplements, chasing the idea of benefits that have never materialized. Multivitamins are marketed with a veneer of science but that image is a mirage – rigorous testing doesn’t support the health claims. But I don’t think the routine use of vitamins will disappear anytime soon. It’s a skillfully-marketed panacea that about half of us buy into.

Not all vitamin and mineral supplementation is useless. They can be used appropriately, when our decisions are informed by scientific evidence: Folic acid prevents neural tube defects in the developing fetus. Vitamin B12 can reverse anemia. Vitamin D is recommended for breastfeeding babies to prevent deficiency. Vitamin K injections in newborns prevent potentially catastrophic bleeding events. But the most common reason for taking vitamins isn’t a clear need, but rather our desire to “improve overall health”. It’s deemed “primary prevention” – the belief that we’re just filling in the gaps in our diet. Others may believe that if vitamins are good, then more vitamins must be better. And there is no debate that we need dietary vitamins to live. The case for indiscriminate supplementation, however, has never been established. We’ve been led to believe, through very effective marketing, that taking vitamins is beneficial to our overall health – even if our health status is reasonably good. So if supplements truly provide real benefits, then we should be able to verify this claim by studying health effects in populations of people that consume vitamins for years at a time. Those studies have been done. Different endpoints, different study populations, and different combinations of vitamins. The evidence is clear. Routine multivitamin supplementation doesn’t offer any meaningful health benefits. The parrot is dead.

This week’s Annals of Internal Medicine published three papers on multivitamins, with a frankly written editorial that should not surprise regular readers of this blog: Enough Is Enough: Stop Wasting Money on Vitamin and Mineral Supplements:

The large body of accumulated evidence has important public health and clinical implications. Evidence is sufficient to advise against routine supplementation, and we should translate null and negative findings into action. The message is simple: Most supplements do not prevent chronic disease or death, their use is not justified, and they should be avoided. This message is especially true for the general population with no clear evidence of micronutrient deficiencies, who represent most supplement users in the United States and in other countries.

While these papers have generated a lot of press, the findings are not surprising. They are consistent with what the accumulated scientific evidence already tells us: there is no compelling reason for most people to take regular vitamin supplements. At best there are an expensive placebo. At worst, they may be harmful.

The U.S. Preventive Services Task Force (USPSTF) Systematic Review

One of the three papers isn’t actually new, as I reviewed it last month. Briefly, it’s the U.S. Preventive Services Task Force Report update to its 2003 guidance on the use of multivitamins, entitled “Vitamin and Mineral Supplements in the Primary Prevention of Cardiovascular Disease and Cancer: An Updated Systematic Evidence Review for the U.S. Preventive Services Task Force“. The intent of this review was to examine the evidence for vitamin and mineral supplementation in community-dwelling, nutrient-sufficient adults. They studied the effect of supplementation on two major killers: cardiovascular disease (CVD) and cancer. Thousands of papers were searched and those of the highest quality were compiled:

  • Multivitamin supplements: 4 trials and 1 cohort study
  • For individual or paired supplements: 18 trials and 5 cohort studies

Overall, multivitamins were found to have no effect on cardiovascular disease or cancer risk. If there is an actual effect, it’s too small to measure in these trials, and too small to be meaningful. From a risk perspective – there was no clear signal of harms from multivitamin supplementation.

Individual supplements also lacked a sound evidence base. There was no evidence of harm or benefit from vitamins A, C, or D, folic acid, calcium (with or without Vitamin D), or selenium. There was more data with beta-carotene, enough to demonstrate that it clearly has no meaningful effects, except in high-risk groups, where it is associated with an increase (yes, an increase) in cancer risk. The evidence for vitamin E was also clear – there were no beneficial effects. The takeaways from this review were stark. Multivitamins have no established role to play in the prevention of cardiovascular disease and cancer. There is currently no persuasive evidence to suggest that routine supplementation offers any meaningful benefits. The USPSTF review joins a long list of studies (1,2,3,4,5,6,7,8,9,10) that have come to similar conclusions.

The Physicians’ Health Study II (PHS II)

The PHS II was a massive study designed to study the effects of vitamins on a number of chronic diseases. This paper reported on the effects of a daily multivitamin on cognition, a secondary outcome in the study. This is another area where the evidence for vitamins has been unimpressive – mainly negative findings albeit with some positive results. The PHS II recruited 5947 male physicians, aged 65 or older. Patients were randomized to beta-carotene or placebo; synthetic vitamin E (400IU) on alternate days, or placebo; vitamin C 500mg daily, or placebo; or a multivitamin (Centrum Silver) or placebo. Patients were assessed through telephone interviews that measured cognition ever 3 years, and followed for a total of 12 years. The reports for beta carotene and vitamins C and E (for cardiovascular disease and for cancer) have been previously reported, with no meaningful benefits shown. This study looked at the multivitamin treatment with Centrum Silver – about 2 900 were randomized to each group.

There was good follow-up with participants and the adherence rate was high, with about 84% of each group taking at least 2/3 of their doses. The findings were also clear – there were no differences between the groups for any outcome at any evaluation, starting with the first cognitive assessment right through to the final assessment about a decade later. The same was observed for verbal memory: no differences between the groups at any assessment. With a very large study of such high quality, we can draw clear conclusions: supplementation with a multivitamin in a generally healthy group of older males appears to have no benefit. If multivitamins have any effect on cognition, then it was too small to be detectable.

The Trial to Assess Chelation Therapy (TACT)

The last trial is the most infamous of all, one that SBM editor Kimball Atwood argued was “unethical, dangerous, pointless, and wasteful.” TACT was a $30 million NCCAM-funded phase 3 trial that assessed chelation therapy versus placebo and high-dose vitamins versus placebo, for the treatment of patients that recently had a heart attack. Both chelation and high-dose vitamins are the darlings of alternative medicine providers who believe these therapies can treat coronary artery disease: not just modestly, but dramatically. The scientific evidence says otherwise: outside of treating heavy metal toxicity, chelation is quackery. There are several posts at Science-Based Medicine on the myriad of problems with the TACT trial, including its ethical issues, its misleading consent process, and the fact that some of the investigators were convicted felons. Few academic research centers participated in this trial. Instead, investigators (who appeared to be advocates of chelation) relied on recruiting from a network of over 130 small centers, some of which already administered chelation treatments and other pseudoscientific treatments. Sites included the “Institute of Integrative Medicine” in New Jersey, an “Integrated Wellness” center in Maine and the “Center for Complementary Medicine” in New York. While designed to be double-blind, there were questions as to whether this blinding could have been maintained effectively. Overall, 1 708 patients aged 50 or over were recruited, all of whom had a heart attack at least six months previously. The vitamin intervention was a massive 28 ingredient supplement – 3 caplets, twice daily, which according to the trial authors, was “designed to reflect the vitamin regimen commonly used by chelation practitioners” giving doses that were in some cases significantly higher than the recommended daily value:

Table 1 TACT Vitamin Content

There were significant problems during the study period. Recruitment was difficult and consequently the target number was reduced. What was expected to take three years to enroll, took seven. In total, 853 patients were randomized to vitamins, 855 to placebo. The endpoint studied was the time to the next major cardiac event (e.g., heart attack) or death. Subjects were followed for a median of about 5 years. The problems didn’t end with recruitment. The lead investigators had to plead with site investigators to spend more time actually tracking and following patients, as they noted as recently as 2007 [PDF]:

Over the last year, the number of patients that have stopped their infusions and vitamins has increased dangerously. If our patients do not receive their treatments, then there is no chance that we could show a difference between groups.

And in what looks like some sort of grade school homework assignment, the investigators even published a list of “Helpful Websites for Finding Patients” to encourage investigators to track patients down that had dropped out of the study. Despite their efforts, the discontinuation rates were indeed dismal: 46% of participants stopped therapy, with no difference between the placebo and the vitamin arms (Perhaps it was the burden of taking six caplets per day.) In the final analysis, no significant difference was found between the high-dose vitamins and placebo for any of the endpoints, including death, or any cardiovascular event. On the positive side, however, the high-dose vitamins were well tolerated and no obvious harms were found. This time, the design and implementation was defined by alternative medicine proponents – and vitamins still failed to deliver any meaningful benefits.

The response from vitamin advocates

As expected, vitamin purveyors don’t take criticism of their product lightly:

The editorial demonstrates a close-minded, one-sided approach that attempts to dismiss even the proven benefits of vitamins and minerals. It’s a shame for consumers that the authors refuse to recognize the real-life need for vitamin and mineral supplementation, living in a fairy-tale world that makes the inaccurate assumption that we’re all eating healthy diets and getting everything we need from food alone. We would not suggest that vitamin supplements are a panacea for preventing chronic disease, but we hope the authors would agree that there is an appropriate place for supplements. Given that government research repeatedly demonstrates that the typical consumer diet is falling short on critical nutrients, vitamin supplements are an appropriate option to meet those needs.

- Steve Mister, President and CEO of the Council for Responsible Nutrition (CRN)

The intention of supplements is to supplement the diet. Don’t expect supplements to cure the common cold or prevent cancer, but they are part of the puzzle of a healthy lifestyle.

- Cara Welch, Sr VP of Scientific & Regulatory Affairs for Natural Products Association (NPA)

It’s telling that advocates won’t directly address the elephant in the room: these products, when studied, demonstrably offer no meaningful health benefits. What the evidence does say is that there is no established role for indiscriminate supplementation, and that they appear to contribute nothing to a healthy lifestyle.

But when it comes to making excuses for supplements, no-one can bring the hysterics like Mike Adams of Natural News:

To make sure these multivitamin studies fail to produce positive results, these studies are universally structured so that they are based on cheap, low-grade, synthetic vitamins and inorganic minerals. Not coincidentally, these brands of low-grade multivitamins are actually manufactured by companies owned by pharmaceutical interests. They really do have a financial incentive to make multivitamins look bad, and so their multivitamin formulations are intentionally designed to fail.

It’s clear Adams didn’t really read the studies at all. If he did, he’d see that Pfizer actually does sell the supplement used in the PHS II study – it’s called Centrum Silver, and a negative trial is probably not what Pfizer is looking for. And guess who designed the supplement used in the TACT study? The TACT authors note:

…complementary and alternative medicine practitioners rather than clinical researchers or supplement companies designed the specific components of the oral treatment regimen, leading to a unique high dose mixture.

TACT was a close as we will probably ever get to seeing a vitamin trial designed by alternative medicine proponents. Furthermore, none of the studies were driven by, or sponsored by, the pharmaceutical industry. The TACT study was sponsored by NCCAM, a regulator with a long history of boosting and supporting supplement use. And why does Adams think these trials repeatedly show that routine supplementation has no meaningful effects? He doesn’t actually refute with actual data. Cue the conspiracy music:

Let’s get down to the real motivation in all this, however. The Annals of Internal Medicine and the scientists behind this extremely deceptive junk science all share the same intention: they want you to trust in drugs, not multivitamins.

Which is an odd statement to make, as I don’t see the Annals or any “scientist” suggesting we should substitute vitamins with drugs for any of the conditions studied. The science is in fact very clear on this point: The healthiest approach, and the science-based approach, is to obtain your vitamins from your food – not from supplements.

Conclusion

Three new papers published in the Annals of Internal Medicine add to an accumulated body of research that has studied the health effects of routine vitamin and mineral supplements in healthy populations. The best available evidence gives us good, reliable information to conclude that multivitamins offer no meaningful health benefits to the generally healthy consumer. It’s time to bring an end to the era of indiscriminate multivitamin use.

References

Guallar E., Stranges S., Mulrow C., Appel L.J. & Miller E.R. (2013). Enough Is Enough: Stop Wasting Money on Vitamin and Mineral Supplements, Annals of Internal Medicine, 159 (12) 850-851. DOI:

Grodstein F., O’Brien J., Kang J.H., Dushkes R., Cook N.R., Okereke O., Manson J.E., Glynn R.J., Buring J.E. & Gaziano J.M. & (2013). Long-Term Multivitamin Supplementation and Cognitive Function in Men, Annals of Internal Medicine, 159 (12) 806-814. DOI:

(2013). High-Dose Multivitamins and Minerals After a Heart Attack, Annals of Internal Medicine, 159 (12) I-20. DOI:

Fortmann S.P., Burda B.U., Senger C.A., Lin J.S. & Whitlock E.P. (2013). Vitamin and Mineral Supplements in the Primary Prevention of Cardiovascular Disease and Cancer: An Updated Systematic Evidence Review for the U.S. Preventive Services Task Force, Annals of Internal Medicine, 159 (12) 824-834. DOI:

Posted in: Herbs & Supplements, Science and Medicine

Leave a Comment (202) ↓

202 thoughts on “More evidence that routine multivitamin use should be avoided

  1. Tom says:

    Let’s drop the pretense and stop saying “…multivitamin supplements may be harmful.” They most definitely are harmful, economically, and that harm should be counted as significant in any thoughtful analysis.

    1. Thor says:

      Hear, hear! It would be a challenge to calculate the immense cost to our family of taking multiple supplements over an extended period of time. We’re talking thousands. And what about the time and effort? What about the skewed, anti-science view that comes with it, as well as any number of faulty beliefs? We’re just thankful that, in the long run, reason prevailed over magical thinking. Big thanks to this SBM site for its immense contribution in facilitating this. Whew…..

      1. irenegoodnight says:

        Good news Thor! Sometimes I think we are only preaching to the already converted!

        1. Thor says:

          Put it this way ign,—I was already converted, but SBM gave power, knowledge, and authority to the trend. A sense of closure, finality. It has been an invaluable tool and trove to have along the path of developing a clearer understanding of reality (the world we live in). And a rock solid support in confronting the barrage of falsity ready at the drop of a dime to rear it’s deluded head.

    2. Jason says:

      Seems like one could use this money instead to buy highly nutritious foods and still have money left over they wouldn’t have had if they kept buying supplements.

      Anyone hear about this Ray Kurzweil guy? He spends thousands a month on supplements! Takes 150/day! Nut job! http://www.dailymail.co.uk/news/article-2467514/Ray-Kurzweil-shares-plans-immortality.html

  2. Drydoc says:

    Thank you. Another fabulous SBM article to share with my medical students and residents.

  3. rork says:

    Was the Physicians’ Health Study II’s conclusions about reduction of cancer incidence found wanting?
    http://www.ncbi.nlm.nih.gov/pubmed/23162860
    The results for lowish dose vs placebo for cancer incidence: HR 0.92; 95% CI, 0.86-0.998; P=.04
    We had discussions about p-values a few days ago and this is a good example. In cancer work I usually don’t pay much attention to novel compounds unless HR exceeds 1.2 (or less than 1/1.2), but that’s in Cox models on smaller data, where the treatment can involve additional suffering, and the decreased hazard might mean you live 3 weeks longer. PHS II had about 15000 subjects and massive (>10 year) followup, and so can detect very small differences. Folks who might argue p=.04 is not enough to go for that treatment, since it is not enough proof, may not be thinking like I do. We have two treatments A and B and we want to know which is better, and here B looks slightly better, so to continue to advocate A, in the absence of data that A is better (here or in other studies), is problematic. That lesser-powered studied failed to find B better could just be proving they lacked power to detect HR of .92 – only rare studies are powered like that. Does it mean I think we should all jump on B – no, it’s just about cancer, not all-cause, and it’s male docs. I might be willing to spend some money on more big studies though, cause even 8% improvement can mean allot if we are talking about vast numbers of people.
    Note I’m not considering costs (money, hassle, quality of life) though, it’s just asking which of two slot machines you want to try next.
    Disclaim: I claim I am not an alt med troll, just skeptic, agent 99 of the math police, cancer researcher, with no nutrition chops. I take no vitamins or supplements. Instead I worry I get too much A cause of the plants I eat (the CARET study). I appreciate further lessons.

    1. WilliamLawrenceUtridge says:

      Meh, even if the detection of harm is spurious, that doesn’t change the fact that the detection of benefit was zero. Even if vitamins aren’t harmful, they’re still a waste of money.

      1. rork says:

        I’m saying the benefit was positive in PHS II and now am saying it was also positive in SU.VI.MAX study. I think editorial and pundits may not have read USPSTF review carefully enough:
        Results:
        “When SU.VI.MAX’s findings in men were pooled with the PHS-II results, the unadjusted relative risk for all cancer incidence was reduced over 10 years of follow-up (unadjusted pooled relative risk, 0.93 [CI, 0.87 to 0.99]).”
        Discussion:
        “We found a statistically significant protective effect from multivitamin supplementation when we pooled data for men in these 2 trials. The borderline significance level in both studies and the lack of an effect in women in SU.VI.MAX suggest we should not try to overgeneralize these results. ”

        SU.VI.MAX found significant sex by treatment interaction – that’s not easy to do as the nerds can attest.

        My posterior says it is likely protective in men. Borderline significance, in two giant and pretty good studies, can combine to give a pretty big Bayes factor. I’d give odds in favor of all-cause mortality too. At least older men, in the U.S. Guys, eat more plants. I don’t mean potato chips.

        1. hellothere says:

          SUVIMAX also had significant differences in baseline status between men and women, with the women having a much greater blood concentration compared to the men. Women in the placebo group tended to have higher blood vitamin status (in most cases) than the men in the active group. Essentially, if the women already were already replete, then how would they get an obsevable effect? The men increased their status with multivitamin treatment, and as such cancer risk was avoided. I would believe (and the study authors seem to agree with me) that the apparent gender effect is likely due to the differences in nutrient status, rather than some physiological difference between men and women.

    2. Geekoid says:

      p .04 doesn’t mean less likely. IT’s means there isn’t enough data to counter the null hypothesis. So, no you should be using it.
      And your A an B is useless without knowing the studies that support both treatments.

      “…may not be thinking like I do.”
      no doubt.

  4. Calli Arcale says:

    But Flintstone’s vitamins are so fun. ;-) “Hey, I got Dino!” “Yeah, well, I got Betty!”

    :-P

    1. David Gorski says:

      Oh, I used to love Flintstones vitamins. Especially purple Dinos.

  5. goodnightirene says:

    If this message can gain widespread traction, there may be some hope for our collective cause! One of my altie “friends” actually referred to this report and she is actually considering its merits. Of course, you have to remember that this crowd have already been convinced that most of the food we eat is devoid of “real” nutrition, having been grown in “depleted” soil, so many will remain convinced that “eating properly” is not an option and will continue using vitamins as “insurance”.

    But those are the well-entrenched. Many will hopefully be influenced by these new studies. In the spirit of Solstice, I am optimistic for a change.

    1. Young CC Prof says:

      If you want to have a good cry, check out the comments section on this particular article about it: http://theweek.com/article/index/254290/how-the-vitamin-industrial-complex-swindled-america

      It makes the comments section of a vaccine article look like a model of rational and purposed debate, and makes a UFO seekers convention look sane. Seriously, how do that many conspiracy theorists even exist?

      1. irenegoodnight says:

        Yes, I find the same thing almost daily in the NY Times. Ignorance is alive and well and thriving on the internet.

        Thanks for bursting my brief bubble of hope! :-(

    2. Halidom says:

      One thing I noticed in the study was the lack of talking about diet. If more of the
      people on the placebo were healthy eaters it skew the results. The only way the study could work is if the eating habits were monitored as well. I agree that no supplement will cure or stop cancer but certain vitamins are required and few people follow really good eating habits.

      1. Chris says:

        Well in the conclusion of this blog article it says: “Three new papers published in the Annals of Internal Medicine add to an accumulated body of research that has studied the health effects of routine vitamin and mineral supplements in healthy populations.”

        So one assumes the study is in healthy people. And if they are healthy they may be eating a decent balanced diet. The study (full paper is free online), Multivitamins in the Prevention of Cancer in Men: The Physicians’ Health Study II Randomized Controlled Trial, did state that part of the screening included analyzing several factors: “Finally, we conducted subgroup analyses stratified by major cancer risk factors, parental history of cancer, selected dietary factors, and other PHS II interventions.”

        Plus it was a study on whether extra nutrients outside of their normal diet affected risk of cancer. It does not.

        You made this claim: “few people follow really good eating habits.”

        I am sorry, but this statement requires an obligatory: Citation needed.

      2. windriven says:

        @Halidom

        “no supplement will cure or stop cancer but certain vitamins are required and few people follow really good eating habits.”

        Breathtaking. I’m trying to parse this out to address each fallacy individually but life is only so long. So I think I will limit myself to the fact that supplements are not a substitute for a proper diet. This isn’t the Jetsons; a good diet does not come in a gel cap. A balanced diet requires far more than a minimum allotment of vitamins and minerals. sCAMsters love to conflate supplements and diet to create the illusion that dietary deficiencies can be corrected with a simple pill. It is horsehockey.

  6. TwistBarbie says:

    This study is fantastic because it’s reaching so many people, and hopefully making them think twice about throwing money away on expensive supplements. Of course, the true believers will probably gripe about the specific forms of vitamins used. It seems like every month there’s a new form of vitamin X (this month it’s liposomal vitamin C) that’s sooooo much more bioavailable than any previously used form (and about 10x the price) and NO WONDER you aren’t experiencing miraculous cures using that proletariat ascorbic acid, and probably a SYNTHETIC too*! Made in a lab or something! How dreadful!

    * I actually had a woman come in to the pharmacy wanting to purchase a bulk mineral of some sort. She asked about the purity and I explained it was USP grade:
    Her: “but is it organic? Or is it made in a lab or something?”
    Me: “….well….. it’s not organic, no, it’s “mineral x”, ….that would be impossible…..”
    Her: “Well, I don’t think I want it then!”

    1. Scott Gavura says:

      “proletariat ascorbic acid” – perfect.

    2. Calli Arcale says:

      Organic mineral supplements . . . and just when I thought the term “organic” couldn’t betray more essential misunderstanding.

      Hey, could Tums be described as organic, since the animals that produced the stuff hundreds of millions of years ago were not fed any antibiotics or artificial pesticides? :-D

    3. Young CC Prof says:

      Ah, misuse of “organic!”

      A couple years back, I went to my mother’s house and helped her cook for some holiday or other. She had a LOT of organic veggies, like half the fridge was stuffed with organic veggies. So I made a vegetable soup, and we served it to company.

      It came out rather well, if I do say so myself. My mother certainly said so, and repeated over and over again about how it was 100% organic. Finally I spoke up.

      “It isn’t.”

      “What?”

      “I put salt in it. Salt is inorganic.”

      1. tariqata says:

        I used to work with someone who informed me that she wouldn’t dare put salt down on her driveway in winter (which is admittedly problematic) – she proudly used “organic sand” instead.

    4. Josh says:

      If you were more of a salesperson, you’d have replied: “Organic? Well, it’s 100% natural!”

      1. Frederick says:

        Organic minerals.. LOL yeah the excavator and caterpillar used to extract it are Bio and free range :-)

  7. Harriet Hall says:

    This is not really news. The Medical Letter and I said so over 5 years ago: http://www.sciencebasedmedicine.org/should-i-take-a-multivitamin/

  8. MTDoc says:

    I found it amusing that the centrum silver commercial on TV makes a point of the fact that it was their product used in the PHSII study. They forgot to mention the negative conclusions, but such is marketing. Afraid I’m guilty of wasting 8 cents a day on a multivitamin, however, based on the rationalization that , being diabetic and needing to keep my weight under control, I consume very small quantities of food. So even an ideal diet in an activity restricted individual raises the question of adequate vitamin intake. When I was more active I never considered supplements either for myself or my patients, and that was long before these studies came out. I was part of PHSI however.

    1. rork says:

      See my PHS II comment at #3. Reduced incidence of cancer.
      That’s a Nov 2012 paper. It used Centrum Silver.

    2. rork says:

      Centrum Silver is what reduced cancer incidence in PHS II. PMID: 23162860.
      I mentioned in at #3, but perhaps commented so much that the effect was minimized. I took it to be depleted uranium rounds against the no good evidence of benefit statements in today’s post.

      1. MTDoc says:

        Sorry, I did sort of miss that point. I’ll probably continue to waste 8 cents a day anyway, since my wife lays it out on the counter each AM. Incidentally it was PHS one I participated in, many years ago, which demonstrated the value of low dose aspirin, at least for male doctors! I suspect there were just not enough female docs in those days, and the end point (CV event) was thought to be more common, and earlier in males.

  9. Reno Hates Me says:

    I wish I could convince my personal doctor to look at the science. She’s a big fan of trendy health issues.

    Got a digestive problem? It’s probably not the three different prescription medications you’re on, instead it’s Gluten. Go on a gluten free diet.

    Feeling down? It can’t be because of your health problems. Instead it’s low vitamin D. Everyone needs to take supplements of vitamin D.

    [Yes, I am shopping for a new doctor! But our area has a major doctor shortage and it's hard to find one who is accepting new patients.]

    1. stanmrak says:

      You would be amazed at the health problems that can clear up if you stop eating wheat. Just try it for a month – it’s the only way to find out. What are you going to do instead? Take another prescription to ‘cure’ the side effects from the other three you’re taking?

      FYI, it’s not possible to get any vitamin D from sunlight for 5-6 months a year unless you live in the tropics. Supplementing in the winter is almost essential. Your doctor is way ahead of most.

      1. windriven says:

        Eff off, stan. This is a thread about multivitamin supplements, not about your gluten obsession and not about the propriety of supplementing vitamin D in higher latitudes during winter. How is it possible for you to cram so much stupid into one 88 word comment?

        1. Nashira says:

          The only appropriate gluten obsession takes place over at The Fresh Loaf, imho.

          1. Dave says:

            A nurse at my hospital claims that his favorite organism is yeast – responsible for bread and for microbrews.

            1. windriven says:

              “responsible for bread and for microbrews.”

              And wine. :-)

  10. D.Simpson says:

    This is a devil’s advocate question. I think Dr. Oz and Dr. Hyman are both goofs, but I’m curious how to answer this:

    http://drhyman.com/blog/2013/11/18/setting-record-straight-supplements/

    (Sorry, new to this, link may not have worked.)

    Dr. Hyman claims that “When Dr. Oz and I tested his studio audience, we weren’t surprised to find widespread nutritional deficiencies.”

    What is likely the truth here?
    - The nutritional deficiencies aren’t nearly as widespread as these two claim?
    - Even if they are as widespread, supplementation would not help? Much of the above discussion focusses on people who have a generally healthy diet, and shows that supplements won’t make you healthier. But can they help people with nutritional deficiencies – notwithstanding of course that the BETTER way to deal with a nutritional deficiency is with a healthier diet?

    1. Young CC Prof says:

      I would ask, “tested how?” Did they actually do blood tests? Or did they do a survey asking about vague symptoms, like appears on that web site? Depression is not necessarily a sign you need vitamins.

      If they did do blood tests, I’d ask what specific deficiencies they found, and in what percentage of the audience. Low vitamin D levels are not uncommon, and iron deficiency sometimes occurs in women, especially during or after pregnancy. This is well known. If they were finding widespread deficiencies of C or A or something, that would be very surprising.

    2. WilliamLawrenceUtridge says:

      How do Hyman & Oz define a “nutritional deficiency”?

      And why is a vitamin pill better than food as a source of vitamins?

      1. stanmrak says:

        No one is claiming that vitamin supplements are better than food as a source of nutrients, except naysayers like you.

        1. windriven says:

          Nice strawman there, stan.

          What has been claimed, or at least implied, is that multivitamins are an appropriate substitute for a healthy diet, especially for the poor. And that, of course, is exactly the kind of exploitive, self-serving nonsense that brings me to so despise your ilk.

          A diet based on prepackaged foods tends to be high in fat, high in carbohydrates, high in salt, high in calories, low in fiber and, yes, deficient in a variety of vitamins and minerals. If you supplement the vitamins and minerals the diet is still high in fat, high in carbohydrates, high in salt, high in calories and low in fiber.

  11. stanmrak says:

    I love all the references to Centrum multivitamins. Any study using Centrum vitamins is a joke from the start… Centrum? Seriously? Centrum is to vitamins as doughnuts are to food. Centrum has numerous ingredients that are toxic. Centrum is humorously known in the healthcare profession as “bedpan bullets” because the pills often come out of the anus virtually intact – enough so that you can still read the logo imprint on the side.

    1. MTDoc says:

      Now I know I’ll continue taking them, After all, I’m a lifetime NRA member.

    2. irenegoodnight says:

      Of course you did, but this report seems to be getting a lot of press–and in a way that is making at least some who have been using vitamins wonder if its worth it.

    3. WilliamLawrenceUtridge says:

      As an expert in marketing Stan, I’m sure you know the best kind of vitamins. Because there’s nothing like marketing to give a science-based, well-rounded medical education.

      You appear to be claiming the results would be different if a different vitamin were used. That’s still an empty assertion based on a complete lack of any proof that a different vitamin would help. Perhaps, but enrobed in your certainty that you as a marketing expert know better, you never take the second step – which is to say “perhaps not”.

      1. stanmrak says:

        Well, we don’t know what the results would be if we used a quality vitamin – they’re almost never used in these studies. Always the cheap junk. It’s like they either want the study to fail, or they’re so clueless that they don’t know the difference between vitamin products. They do see to be clueless about nutrition.

        Let’s say we wanted to find out how much better your car would run by adding a certain performance-enhancing gasoline additive. If we used a cheap knock-off variety that costs $1, I suspect that the results would be different than if we used the $20 top quality additive.

        1. Scott Gavura says:

          TwistBarbie refuted your comment before you even made it. I would be interested to see evidence that there is a relationship between the retail cost of a vitamin and its efficacy (based on some clinically relevant outcome). Failing that, I would be interested to see the evidence that there is relationship between the cost of a vitamin and its bioavailability.

  12. irenegoodnight says:

    Nice to know you’re checking.

  13. Scott is on another anti-vitamin rant I see.

    If you get all your vitamins and minerals from food, you dont need multivitamin he says.
    Well Duh! What else is new Scott, if your gas tank is full you dont need gas?

    How many people get sufficient vitamins from food in the real world? About 25% of people who I see for the first time eat no vegetables at all. Another 30% get insufficient amounts.

    What, you thought avitaminosis happens only on National Geographic channel? 50% of poor neighbourhoods in US have no access to fruit or vegetable products.

    Thats the reality, not everyone eats well.

    1. TwistBarbie says:

      Actually, it says 50% don’t live near a grocery store, that is a very, very far cry from saying they have NO access to fruit or vegetable products, but your leap certainly offers some insight into how your brain works.

      1. stanmrak says:

        If you don’t live near a grocery store and are too poor to afford a car, you might as well say that you have no access to fresh produce. It’s not a huge leap. you obviously don’t know anyone in that predicament. I see these people all the time.
        Add to that the fact that the government subsidizes corn soy wheat instead of broccoli and lettuce, making junk food cheap and real food expensive, and the poor are relegated to poor nutrition. Free market capitalism triumphs again.

        1. nancy brownlee says:

          “If you don’t live near a grocery store and are too poor to afford a car, you might as well say that you have no access to fresh produce…. I see these people all the time.”

          Hey, I was these people. For years and years. In school, and out of it. I rode buses, got rides, bought apples at the 6-12. I saw a whole lot of other people doing the same thing, and I still see them. If you can buy cigarettes and Lotto tickets, you can buy produce. you know what? After I finished school and moved to the country and started working at a little college, I met dozens of young people who had- quite literally- never tasted a salad, and whose idea of vegetables was beans and potatoes – and had never tasted any other. They all lived smack in the middle of agricultural abundance, a few blocks from a huge farmer’s market.

          “the poor are relegated to poor nutrition.” Nope. Bullshit. They may find it easier or more appealing to eat a frozen burrito, but they don’t have to.

          1. Scott Field says:

            You can get apples to snack on, sure. But can you get all the ingredients to cook a healthy, well-balanced meal? And could you afford them if they were available? And would you have the time to cook them after getting home from your 2nd job? Etc. It’s not that poor people *can’t* eat healthy; but there are a number of disincentives that make them less likely to.

            To be fair, I haven’t seen any data correlating food deserts to actual health impacts? It seems to make sense that making healthy food harder to get could make people less likely to eat healthy. But that’s an assumption, not a conclusion. Anyone?

            1. nancy brownlee says:

              It’s only a personal observation, so worth very little- but the working poor, lower-middle-class people I have lived with since my working poor, lower-middle-class birth vary a great deal in their cooking and eating habits. The ones who don’t cook simple, healthy meals don’t do it because it’s much less trouble to buy cheap fast food, or nuke some junk. The give their kids Hot Pockets and donuts for supper because it’s easy. It’s not cheaper than healthy food, but it’s easy. There are lots of low-income people who DO cook for their families. They don’t have more money, they are not less tired at the end of day, they work the same kind of jobs, and they live in the same neighborhoods. It’s not access, it’s effort. Middle class families often make the same choices. People are lazy, aren’t they? People often do the most expedient thing, not the best thing, don’t they? That’s people for ya.

              1. Scott Field says:

                “It’s not access, it’s effort.”
                True. But it stands to reason that the more difficult access is, the more effort is required.

            2. windriven says:

              Many communities have food banks and other programs to help poor people acquire healthy and nutritious foods. My factory is in a fairly rural part of WA and we certainly have abundant sources of fresh, wholesome foods. Further, food stamps are widely distributed and can be used – should be used – for healthy foods. The failure is, in my opinion, more likely to be poor dietary habits than access to fresh foods.

              But most disturbingly your comments seem to suggest vitamin supplements as an alternative to a good diet. They aren’t. A healthy diet includes a range of micronutrients, fiber, protein, fats and carbohydrates as well as more common vitamins and minerals.

              1. Scott Field says:

                “The failure is, in my opinion, more likely to be poor dietary habits than access to fresh foods.”
                Sure. But the more difficult it is to get fresh foods, the harder it is to develop good dietary habits. It’s like saying you can still get a good education at a crappy school with bad teachers if you just have good study habits; of course you can, but it just makes it that much harder.
                .
                “A healthy diet includes a range of micronutrients, fiber, protein, fats and carbohydrates as well as more common vitamins and minerals.”
                Fair point. I didn’t mean to oversimplify.

        2. Sawyer says:

          You and FBA are hereby forbidden to ever whine about Big Pharma “treating the symptoms rather than the cause”, because that’s exactly what you’ve done here.

          Go find me a person with extensive experience running a successful charity that promotes health, education, and frugality among the poor. If they are trying to get rid of a liquor store in a bad neighborhood, I can assure you they aren’t going to bother replacing it with a GNC. It turns out poor people aren’t big fans of out-of-touch, half-measure attempts to barely improve their lives. God forbid we help get them flu vaccines instead of Centrum Silver.

          1. windriven says:

            High five.

          2. Scott Field says:

            I’ve never whined about Big Pharma in my life. Don’t put words in my mouth.

            As for poverty being a “symptom”… there is no sane response to that much stupid.

            1. Sawyer says:

              I’m sorry if the thread system here is tough to navigate, but my reply was to clearly to stan. I hope the symptom comment makes more sense in that context. I don’t like people putting words in my mouth either!

              1. windriven says:

                Yes, the threading drives me band edge too. I’m resorting to the old “@ intended_recipient” for clarity.

            2. windriven says:

              @Scott Field

              Perhaps you wouldn’t be misunderstood if you clearly laid out your position. It seems a perfectly reasonable inference, based on what you’ve written in this thread, that you propose vitamin supplementation as an appropriate response to poor diet in impoverished communities.

              I believe you’ve misinterpreted Sawyer’s use of ‘symptom’. I believe he was counting poor diet a symptom of poverty, not poverty as a symptom of … anything.

        3. WilliamLawrenceUtridge says:

          There may not be access to fresh produce, but canned and frozen produce and vegetables are both excellent, sometimes superior sources of vitamins compared to fresh produce, depending on the length of the supply chain. Not to mention, even fast food usually has some vegetables attached to it – lettuce, onions, tomatoes, pickles, etc.

          But yeah, the systematic incentives of the US agricultural system are definitely worth questioning. I would be very curious to see what the chain of reasoning is to defend the subsidies that are handed out. On the other hand, food security is far more assured by corn, wheat and soy than by broccoli and lettuce – far more calories, fats and protein on a per-acre basis.

          It’s an error to think of broccoli and lettuce as the only “healthy” food. To a starving person, they’re more of a cruel joke than a valuable source of nutrients, you’ll die of starvation far faster than you will die of scurvy or some other micronutrient deficit.

      1. windriven says:

        Once again, only a douche believes the answer to vitamin deficiencies in children (as reported in the Daily Mail!!) is supplements – the kind of douche that is proud of being a fast buck artist.

        A healthy diet is the cornerstone of a healthy body. Oranges, broccoli, nuts, whole grains, fish and a bit of meat. And exercise.

        Suggesting vitamins as a cure for poor diet is like suggesting Cliff’s Notes as a cure for illiteracy.

      2. MadisonMD says:

        So scurvy went from 51 to 94 cases in a country of 62 million according to a 4-year old article in the Daily Mail. FBA’s idea is that poor folk spend their money on multivitamin tabs with NNT of 1,000,000 rather than nutritious food.

    2. windriven says:

      First, “About 25% of people who I see for the first time eat no vegetables at all,” suggests to me that you are either a liar or that 25% of the “people you see” are. Second, the answer to the few who have deficient diets isn’t to sell them crap, it is to encourage them to eat an appropriate diet. There is more to a healthy diet than vitamins.

      “50% of poor neighbourhoods in US have no access to fruit or vegetable products. ”
      Assuming for a moment that this is true, why do you think (a) that stores in poor neighborhoods are more likely to carry multivitamins than they are bunches of broccoli or (b) that poor people will spend their money on Flintstones instead of Fritos? Stores sell what people want to buy. Crappy dietary choices among America’s poor won’t be corrected with supplements of questionable value. It is an important public health issue and it won’t be solved by shoveling money to fast buck artists.

      1. Scott Field says:

        “the kind of douche that is proud of being a fast buck artist”
        You know how the anti-vaccers like to accuse everyone who disagrees with them of being a shill for Big Pharma? You know how it’s bullshit when they do it? It’s bullshit when you do it too. Please stop.

        “suggests to me that you are either a liar or that 25% of the “people you see” are.”
        It suggests to me that you haven’t been paying attention. Countless studies, polls, etc. have shown Americans typically eat at best half the veggies we’re supposed to. It’s a huge issue in public health circles. Seriously, 30 seconds on Google…

        “the answer to the few who have deficient diets isn’t to sell them crap, it is to encourage them to eat an appropriate diet.”
        See above re it’s not only “the few.” But that aside: Doctors have been telling people to eat better and exercise more since forever. And yet every year fewer and fewer of us do so. If you know a way to reverse that, I’m sure we’d all love to hear it. Changing people’s behaviors like diet is *hard* and has a depressingly low success rate. (See obesity rates, cholesterol, lung cancer, et. al.)

        That’s like saying “the answer to obesity is to encourage people to eat less.” Theoretically true, maybe, but demonstrably not working as a practical solution.

        1. windriven says:

          Try to keep up slick: Fast Buck Artist chose his avatar and screen name, not me. FBA preys on vulnerable people with a witches brew of delusional quackery. It offends me. If you don’t like it, don’t read my comments. Problem solved.

          Don’t throw straw men at me. FBA said 25% of the people he sees for the first time “eat no vegetables at all.” Bullcrap. The tomato on a Whopper is still a vegetable. You’d have to live a bizarre life to eat no vegetables at all.

          “Changing peoples behaviors like diet is *hard*.”

          Yeah, much easier to rip people off with worthless supplements. Quitting smoking is hard too. So people should just change to e-cigarettes, right?

          Take your tone trollery and wrap it around a multivitamin suppository, then stuff it. Let me know if you’re ever able to mount a meaningful argument worth more than 12 seconds of my time.

          1. Scott Field says:

            “You’d have to live a bizarre life to eat no vegetables at all.”
            Fair point that “not enough veggies” is not the same thing as “no veggies at all.” Sorry, I misread.

            “Yeah, much easier to rip people off with worthless supplements.”
            Just to clarify: are you arguing that if someone is not getting enough (say) vitamin B in their diet, that taking a vitamin B supplement does *not* actually give them more vitamin B? I mean, I know you believe people can magically change their diets overnight. But assuming someone can’t/won’t, are you actually claiming that taking a vitamin B supplement provides no benefit at all?

            “Quitting smoking is hard too.”
            No shit. Which is why we have things like nicotine patches, gum, etc to help people change their behaviors. (e-cigs seem like a scam to me, but I honestly haven’t researched them. IIRC, the last SBM article on them concluded there wasn’t enough evidence yet to conclude if they’re actually effective as a smoking cessation aid.)

            Based on your “logic” should I assume you’re also opposed to prescribing Lipitor to people with high cholesterol? Cuz they should just eat better amirite? If not, please explain how that’s different?

            “Take your tone trollery and wrap it around a multivitamin suppository, then stuff it. Let me know if you’re ever able to mount a meaningful argument worth more than 12 seconds of my time.”
            Wow, now *that’s* a meaningful argument. Bravo. Let me know when you’re done with insults and want to address the argument I already made, which is that just telling people to eat better has shockingly little effect on changing their behavior. *Of course* it would be better if everyone just ate better; but pretending that they do or will is flat denialism. Or do you not care if your “solutions” don’t actually work in the real world as long as you get to claim moral superiority?

            1. windriven says:

              “Just to clarify: are you arguing that if someone is not getting enough (say) vitamin B in their diet…”

              The topic of this blog was multivitamins, not supplementation to correct one specific deficiency.

              My argument is that poor diet involves far more than a paucity of some combination of vitamins and minerals and cannot easily be replaced with a pill. It is my further argument that better dietary education is a fundamental public health issue and should be addressed as such. Whether this is ‘easy’ or ‘hard’ is immaterial.

              “Which is why we have things like nicotine patches, gum, etc to help people change their behaviors.”

              One would have to be especially delusional to imagine that multivitamins ‘help people to change their behaviors.’

              “Based on your “logic” should I assume you’re also opposed to prescribing Lipitor to people with high cholesterol? Cuz they should just eat better amirite? If not, please explain how that’s different?”

              It is different for two reasons: hypercholesterolemia is not necessarily caused by diet; statin drugs in appropriate populations are linked to fewer CVD risks while multivitamins are not linked to any reduction in risks.

              To answer your final pout, I agree that simply telling people to improve their diets is not particularly effective. But it does not logically follow that multivitamin use leads to either better health outcomes or better diets.

              Finally, this has nothing to do with moral superiority but everything to do with factual superiority.

  14. Bryan says:

    Centrum vitamin pills come out intact, Stan? Maraini et al.* would probably beg to differ:

    “The use of multivitamin-mineral supplements has become increasingly common, but whether the use of such supplements improves micronutrient status remains still unclear. The objective of this report is to investigate how a long-term vitamin-mineral supplementation following the US Recommended Daily Intake (RDI) affected the plasma levels of selected nutrients in a subset (No. = 407) of participants in the Italian-American Clinical Trial of Nutritional Supplements and Age-related Cataract (CTNS). The CTNS was a double-blind, single centre, controlled clinical trial of 1020 participants aged 55-75 years randomized to a daily tablet of Centrum © or placebo.”

    “Participants assigned to Centrum © showed a significant increase (p < 0.005) in mean/median plasma levels of vitamin E, beta-carotene, folate, and vitamin B12, and an improved riboflavin status when compared with participants assigned to placebo."

    * Effects of multivitamin/mineral supplementation on plasma levels of nutrients. Report No. 4 of the Italian-American clinical trial of nutritional supplements and age-related cataract (2009): http://www.ncbi.nlm.nih.gov/m/pubmed/19636163

    1. stanmrak says:

      Your study only mentioned 5 nutrients out of the 30 or so that Centrum claims to have. That would indicate only a 17% rate of absorption at best.

      1. windriven says:

        stan,

        “That would indicate only a 17% rate of absorption at best.”

        Did you fail math and science or did you just not take them? The fact that a study only mentions a particular number of test subjects doesn’t mean that other untested subjects were unaffected.

        stan is a marketing guy. stan doesn’t understand science. Therefore marketing people don’t understand science. Do you grasp the logical failure there?

      2. Scott Gavura says:

        @stamrak:

        Your study only mentioned 5 nutrients out of the 30 or so that Centrum claims to have. That would indicate only a 17% rate of absorption at best.

        [facepalm]

      3. CHotel says:

        Stunning lack of logic as pointed out by windriven aside, 5 out of 30 showing benefit would mean a 17% rate at WORST, not best. You’re implying that all 25 others are null, and that the study could be wrong.

        You’d think for a marketing guy you’d be better with words.

        1. Bryan says:

          Stan, 5 out of 5 micronutrients studied were shown to raise plasma levels significantly. In this universe, that’s as close to 100% as you get.

          Cheap bullets will penetrate no more than 17% of the population if you only examine the holes they made in 5 of 30 people shot?

          Two of the vitamines looked at were B12 and folic acid. Both have been shown in numerous studies to raise blood levels of both, lowering homocysteine at the same time. Nothing ‘high end’ about the cyanocobalamin and synthetic folic acid that were used.

  15. hellothere says:

    This drives me absolutely nuts. In this Annals issue we have a systematic review on multivitamins which concluded that there wasn’t enough evidence to recommend widespread use, a trial examining a combination of IV chelation therapy with high-dose vitamins for the secondary prevention of heart attacks (compared NOT to a placebo, but to a low dose vitamin regimen), and a secondary endpoint from PHS2.

    The US preventive task force report seems a little backwards to me. I understand that given the fact that there’s only 2 trials (really only 1, since SUVIMAX isn’t a “true” multivitamin like Centrum Silver – the levels used were sorta random in some cases and not based on the DRIs), and that the biggest one only had men in the study, the USPTF can’t make a blanket recommendation for multivitamin use and that they recommend further research. Ok, I get that. But when they recognize that 2/3 trials found statistically significant reductions in cancer risk for men (and the one that didn’t was the shortest, smallest, and only one that didn’t have cancer incidence as a primary endpoint), this somehow gets turned into “the lack of an effect for women (albeit in 1 trial), the borderline significance in men in both trials, and the lack of any effect on CVD in either study makes it difficult to conclude that multivitamin supplementation is beneficial.” But the review showed that men can get a significant reduction in cancer risk with multivitamin use – how is that not beneficial??? And borderline significance? Since when did p<0.05 become insufficient to establish significance? I remember my old stats professor telling us re: significance: "you're either pregnant or you're not. it's less than 0.05 or it's not.". Close only counts in horseshoes and hand grenades – with significance, it's either significant or it's not, and in this case, the benefit was significant. Period.

    The chelation study has no business being lumped in with this group. In addition to the fact that it is testing something that the general public has no or limited access to, it is testing a completely different hypothesis (namely secondary MI prevention in MI survivors) than say PHS2 (primary prevention in completely healthy men). Just because the study involves multiple vitamins doesn't make it a "multivitamin" as most people would define it. And the fact that it compares high dose vitamins to low dose vitamins (for some reason people have been calling that a placebo, which is of course inaccurate) would, to my mind, not make it sufficient to judge the efficacy of this intervention – all they showed was that high dose is no better than low dose. Also, the fact that they lost 46% of their patients due to non-compliance, etc, also makes me question the validity of the study, but I digress.

    The PHS2 paper discussion makes my blood boil the most. What amazes me is that, for some reason, when a null finding comes out from PHS2 people jump all over it as the nail in the coffin for multivitamin use, and that Centrum is deceiving the public, and it's all a big money pit blah blah blah. But when they discuss the positive findings from the study, it becomes "well, the effect is so small and we can't really be sure and it's only in men, so….". PHS2 is a randomized, double blind, placebo controlled, long term, massive, well-powered study funded by the NIH and run by Harvard – it's like the Rolls Royce of nutrition studies. It's the shiniest gold standard you could ever ask for. Ok, bummer, no observed effect on cognitive decline, which we should mention was a secondary endpoint for the study (cancer, heart disease, and eye disease were the 3 primary endpoints). But let's not forget that PHS2 showed beneficial outcomes in 2/3 of these primary endpoints, namely a significant reduction in cancer risk, a significant reduction in cataract risk, and while no overall reduction in cardiovascular events, did show a significant reduction in fatal MI incidence (again, secondary endpoint – but if we're counting the cognitive decline data as being solid and it's a secondary endpoint then shouldn't we do the same with the fatal MI data?). I mean COME ON. What else do you want? We're talking about an incredibly cheap (pennies/day) intervention with no side effects that requires virtually no effort that can significantly reduce your risk of cancer, cataracts, and (maybe) fatal MI? And yet it's not a good enough intervention? That's like a public health gift from heaven! Ok sure, the effects aren't massive – but what are you expecting? These aren't statins, they're *multivitamins*. I think we need to manage our expectations a little better here. And, one could argue that since the PHS2 study population was leaner, ate better, exercised more, smoked less, had lower cholesterol, was richer and overall healthier than the general US population, that if they took a true slice of "Americana" the effect would be larger – but at the end of the day, these are still significant, albeit somewhat modest effects from a rock solid study. Let's not poopoo the often insulted 8% reduction in cancer incidence – since about 1.6 million people are diagnosed with cancer annually, that 8% = 128,000 people every year that wouldn't develop a tumor – which is still ALOT of people.

    Yes I agree that we need more research – the lack of data in women seems like an obvious knowledge gap, as a for instance (aside from the already recognized benefits of folic acid for women of childbearing age and the need for calcium/vitamin D to prevent falls/fractures especially for women). But to say that "enough is enough" and that there's no evidence of benefit for vitamins/minerals? I wholeheartedly disagree.

    1. rork says:

      ” I remember my old stats professor telling us re: significance: “you’re either pregnant or you’re not. it’s less than 0.05 or it’s not.”. Close only counts in horseshoes and hand grenades – with significance, it’s either significant or it’s not, and in this case, the benefit was significant.”
      You sound like significant means “true”, but it doesn’t, and there’s nothing magical about .05. P-value merely attempts to summarize the strength of evidence from particular data. It say less about what to do. More below.
      I do agree that the cancer prevention shown by the largest two randomized placebo controlled studies make me think it will help average over 50 male, as I wrote above. And not just cancer incidence – overall mortality too. One might note that the second, PHS II study, was in docs, and we might guess average old fart might eat less well, have more other bad habits, exercise less, and benefit more from the treatment. The effect size was larger in the French data, who were less selected. (Never forget it was bad for the women in that study.)

      I think that the greatest failing with the ‘Enough is Enough” editorial and its insufficiently critical reviews here and elsewhere. Even thinking about it as a hypothesis testing problem is too simplistic: that fails to even admit there is a decision space and loss function. I’d prescribe decision theory, for both individual trying to decide, and groups of docs trying to decide what to advise. I admit that how much of what thing for who will take us forever to get really good at saying, but both patient and doc are faced with the yes or no decision now. It’s not like false positive errors and false negative ones are of completely different natures here – the costs of both are measured in the same units, health and lives of people.
      PS for experts: I don’t even have a point mass in my prior at HR=1. Why would I?

  16. stanmrak says:

    Next week, month, season, there ‘ll be more studies that show exactly the opposite conclusions. There are certainly enough of those studies already. I’ve been following this debate for over 30 years and nothing has changed. On top of that, most of these studies are too poorly designed to mean anything, anyway.
    No need to get your panties in a wad, Just make your own choice, and don’t infringe on someone else’s choice just because they believe someone else and not you.
    If the evidence was so conclusive, they wouldn’t keep doing these studies.

    1. hellothere says:

      PHS2 took nearly 13 years to complete – so no, there won’t be another one of these studies next week/month/season. And that’s the point of the USPSTF report – that there’s not enough of these studes (3, in fact). And what specifically was poorly designed about PHS2?

      1. stanmrak says:

        They used Centrum multivitamins – the processed junk food equivalent of vitamin supplements. The study only proves that Centrum vitamins don’t work very well, but what can you expect for 8 cents a day? A decent multivitamin costs about $1.50 for a daily dose. You get what you pay for. Would you buy a motor oil that was priced at 15 cents a quart?

        1. Bryan says:

          The real question is: why would you buy motor oil at $1.50 a quart if motor oil 10 times cheaper then that is just as good?

        2. CHotel says:

          I mean, I should know better than to bother asking Stan to cite some decent peer-reviewed literature showing that Centrum vitamins are inferior to whatever his favourite flavour is, but if I don’t then people might think he has a point.

  17. Scott Field says:

    So vitamins don’t cure cancer & cardiovascular disease (USPDTF), cognition (PHII), or coronary artery disease (TACT)? …OK, fine. I know there are snake oil salesmen making absurd claims like this, so they need to be refuted, and SBM is great at that. Cool. But I’m not sure you’ve justified your “should be avoided” headline.

    Most people I know who take vitamins routinely have far more modest expectations, like “having more energy,” “getting sick less,” “feeling better,” etc. And yes, I fully recognize these are vague, subjective, and much harder to test. But saying “vitamins don’t cure cancer” isn’t really relevant to someone who takes a multivitamin for minor general health reasons.

    The “minor” case for daily vitamins could be summed up as:
    1. Getting less than the RDA of vitamins is bad for you. (Not generally apocalyptic, but certainly not great.)
    2. Many (most?) Americans do not eat healthy diets, and therefore may not be getting the vitamins they need.
    3. Persuading people to change their diets is notoriously difficult. (See obesity rates, cholesterol, etc.)
    4. Taking a normal* multivitamin dose is dirt-cheap and has basically no side effects.

    If there are other studies refuting any of those four points, please point me to them; I’m entirely open to being persuaded here. I agree the Vitamins-As-Miracle-Cures claims are CAM-scams. But just because someone claims (wrongly) that aspirin cures cancer, doesn’t mean I shouldn’t take aspirin for my headache.

    * I’m discounting the high dose regimens, which I agree are pure quackery, and may be dangerous to boot.

  18. Roman says:

    What about the skewed, anti-science view that comes with it, as well as any number of faulty beliefs? We’re just thankful that, in the long run, reason prevailed over magical thinking. Big thanks to this SBM site for its immense contribution in facilitating this. Whew…..

    I heartily second Thor’s sentiments. Being daily swamped in a sludgy swell of woo in both the real and the online world, SBM articles are Drano for the mind. An evening’s read of SBM has become a virtual addiction and what an education it is! As well, over and above the published articles, I find the appended comments to be remarkably thoughtful and equally instructive. Sincere thanks to all contributors!

  19. Roman100 says:

    What about the skewed, anti-science view that comes with it, as well as any number of faulty beliefs? We’re just thankful that, in the long run, reason prevailed over magical thinking. Big thanks to this SBM site for its immense contribution in facilitating this. Whew…..

    I heartily second Thor’s sentiment. Being daily swamped in a sludge of woo both in the real and the online worlds, SBM is a little dose of sanity, of Drano for the mind, as the saying goes. An evening’s perusal of SBM has become, for me, an education and a virtual obsession. I find the articles to be of the highest value, as well as the thoughtful and often penetrating comments.

  20. Roman100 says:

    Golly. I didn’t think my first post would go through and so tried recomposing it from memory. Sorry about that!

  21. skeptic says:

    1) Did not mention that in TACT, there was a statistically significant benefit of high dose oral vitamin therapy in patients who were not taking statins, with a statistically positive interaction test.

    2) Did not mention that PHS II documented a reduction in cancers, as well as fatal MI, in physicians randomized to the Centrum multivitamin.

    3) Did not mention that large segments of the population – the elderly, pregnant women, vegetarians, people in institutions – do typically need some form of vitamin supplementation (often multiple) to prevent disease.

    This site loves bashing vitamins as a necessary corrective to the pro-vitamin hype. But do not throw out the science with the bathwater. Let’s not forget, there have been exceptions – vitamin E in patients with specific haptoglobin genotypes; vitamin D in institutionalized elderly females; folic acid to prevent NTDs – just to cite three examples. In our trials we are lumping a bunch of genotype-phenotype combinations together without any consideration of the subsets who might benefit the most from vitamin therapy – the overall signal is therefore often a wash. That’s where the money now is – in teasing out the people who for whatever reason can still benefit from vitamins. Personalized medicine is here to stay.

    1. Vicki says:

      They do in fact mention things that you claim they don’t: the post and the study it’s describing are explicitly about the general community-dwelling adult population. So not elderly women, or men, living in institutions. Pregnant women, again, are a specific sub-population, and Dr. Gavura explicitly mentions folic acid supplements to prevent possible birth defects.

      He may not have written a post you agree with, or the post you wish he had written. But you’re disagreeing with things he didn’t write.

  22. Scottynuke says:

    And yet another data point on supplements of all stripes (not to mention another DSHEA slam):

    http://www.nytimes.com/2013/12/22/us/spike-in-harm-to-liver-is-tied-to-dietary-aids.html

    1. squirrelelite says:

      I saw that same article, scottynuke.

      Evidently, green tea extracts are the main culprit.
      But, the general lack of regulation and quality controls in the supplement industry is the real problem.

  23. MF34 says:

    Multivitamins are toxic, period. Chromium, which is used for blood sugar problems is very toxic and dangerous. Copper, zinc, iron, and all metals and minerals can be very very dangerous. Please get minerals from food. MSM sulfer is safe and can be beneficial, but totally healthy people probably don’t need it. Please use diet and organic food for blood sugar problems. Minerals build up in the body. That’s why adding toxic fluoride to my water makes me very angry. Multivitamins are evil, yet excess fluoride is forced on me. Good vitamins can be low dose natural b complex, low-moderate vitamin D, Mk-7, low dose occasional Methylcobalamin B-12 and if you need vitamin C , E, or A, get it from food. Even natural minerals might change in the body to toxic forms – Chromium could change to the toxic type – who knows. Chromium supplements have shown to cause DNA damage. No form is safe in my view. Folic acid is not real folate. Excess vitamin A is bad. Selenium is highly toxic if you take just a little too much.

    1. Chris says:

      “Minerals build up in the body.”

      Citation needed.

      “Chromium supplements have shown to cause DNA damage.”

      Citation needed.

      “No form is safe in my view.”

      Truthfully we are not interested in your view. Just provide us the science to back up your claims.

    2. Chris says:

      “Minerals build up in the body.”

      Citation needed.

      “Chromium supplements have shown to cause DNA damage.”

      Citation needed.

      “No form is safe in my view.”

      We are not interested in you “view.” If you make a claim, you need to back it up with verifiable scientific documentation. In the future, instead of unsupported assertions try links to actual scientific studies.

    3. Chris says:

      Apologies for the double post. WordPress hates me by showing my comment, and then not showing it.

  24. MF34 says:

    I don’t have any citations. I thought I could share my opinions. I will make sure to not post here again. Please disregard my post. I can’t delete it.

    1. windriven says:

      The clue should have been the ‘Science’ in the title. It’s a tough crowd and personal opinions won’t generally get you very far. But if you have good science to support your positions …

    2. Chris says:

      You are welcome to stick around. That is if you have a mind open to the culture here that “if you make a claim, you must support that claim with scientific evidence.” Because, while you are welcome to have your own opinions, but not to your own facts.

    3. MadisonMD says:

      @MF34
      I don’t think folks meant to scare you away. I think most agree with your idea that taking multivitamins is not warranted and you raise some legitimate concerns about potential toxicities. However, this group will ask for evidence if you make assertions.

      So, you are welcome here. Read. Comment. Perhaps I would suggest that when you post, pose your concern as a question rather than an assertion. If you do so, others may actually even provide evidence for/against the concern you raise and we all learn.

      1. Chris says:

        Now, I kind of feel bad. I did not know I was so harsh.

        But we have so many people who come here making incredible claims, and yet they refuse to back them up. It can be frustrating that they think we can take them at their word.

        By the way, the term “Citation needed” is from Wikipedia, where it is absolutely required for an editor to source their inputs.

  25. Lebenleber says:

    Thank you for the article. I’d be interested in your response to this response:
    http://lpi.oregonstate.edu/news/enoughisenough-response.html

    Furthermore, what is your opinion on the RDA for potassium of 4.7 g/d, which is very hard to reach even with a carefully planned diet?
    Not too long ago, I kept a food diary for 51 days and then analyzed the approximate averaged daily nutrient intake by using various food databases:
    http://www.mediafire.com/download/ia0ttkagm9ncndv/fooddiary_51d_07102012_nowater_sanitized.pdf
    Notes: I followed a high-fat [mostly canola & olive oil] pescetarian [mostly oily fish] diet. Carbohydrates came mostly from whole-grain food. “Nowater” means that I did not document intake of non-caloric beverages [mostly tap water & tea]. “Sanitized” means that I excluded some nutrients (e.g. biotin, choline) for which the food databases or the scientific knowledge were too incomplete to make any conclusions. The sodium and sugar intakes may be somewhat exaggerated due to database constraints.

  26. Lebenleber says:

    Thank you for the article. I’d be interested in your response to this response:
    http://lpi.oregonstate.edu/news/enoughisenough-response.html

    Furthermore, what is your opinion on the RDA for potassium of 4.7 g/d, which is very hard to reach even with a carefully planned diet?
    Not too long ago, I kept a food diary for 51 days and then analyzed the approximate averaged daily nutrient intake by using various food databases:
    http://www.mediafire.com/download/ia0ttkagm9ncndv/fooddiary_51d_07102012_nowater_sanitized.pdf
    Notes: I followed a high-fat [mostly canola & olive oil] pescetarian [mostly oily fish] diet. Carbohydrates came mostly from whole-grain food. “Nowater” means that I did not document intake of non-caloric beverages [mostly tap water & tea]. “Sanitized” means that I excluded some nutrients (e.g. biotin, choline) for which the food databases or the scientific knowledge were too incomplete to make any conclusions. The sodium and sugar intakes may be somewhat exaggerated due to database constraints. The dates are obiously bogus.

    1. windriven says:

      I’ve read the Linus Pauling Institute piece you linked. I was struck by several things. First, the author is more interested in pushing multivitamins than in reshaping dietary habits. Unfortunately poor dietary habits have consequences beyond deficiencies in a few vitamins so the multivitamin prescription is neither sufficient not necessary to solve the actual problem.

      Second, the paper dismisses as “non-significant” the 0.1% of US adults who exceed the Tolerable Upper Intake Level for vitamins E. If we use 200 million as the number of US adults that would be 200,000 people! Moreover, that paper links to a Pauling Institute vitamin E puff piece that has nothing to do with the 0.1% claim. Bad form, life lover.

      So ultimately I would dismiss this as marketing masquerading as science. But if you have some actual science to offer on this subject I’m sure it would be well received here.

      1. MadisonMD says:

        Second, the paper dismisses as “non-significant” the 0.1% of US adults who exceed the Tolerable Upper Intake Level for vitamins E…. that would be 200,000 people!

        Wow. If they could only offer some evidence that such multivitamins benefit >0.1% of those who take them… beyond of course the straw man– “most people don’t get enough vitamins!”

        1. Lebenleber says:

          “[...] beyond of course the straw man– ‘most people don’t get enough vitamins!’”

          Did you just gloss over average Joes and Janes being provably not “well-nourished” in key micronutrients, according to Estimated Average Requirements (reference 2)?

          1. MadisonMD says:

            Did you just gloss over average Joes and Janes being provably not “well-nourished” in key micronutrients, according to Estimated Average Requirements (reference 2)?

            Actually, no, I didn’t just gloss over it. I actually pulled up the referenced article, determined how intake of many nutrients was estimated and averaged over a period of 2 days. Then I looked at the results and was not very surprised to see that for many of them, one or more was found to be short of the “Estimated Average Requirements” over this small sampling time. Then I considered this, looked up appropriate references on vitamin stores and wrote out my reasoned response below– see straw man #2.

            In short my point is that there is a very large difference between:

            (a) Falling short of “Estimated Average Requirements” in a period of time that is much shorter than the period of time for vitamin stores. Not a health issue (see below for discussion of vitamin stores).
            and
            (b) Getting a disease of vitamin deficiency. This is a health issue.

            (a) is far more likely than (b). So if your institute’s major purpose is to show the value of vitamins, you do research that demonstrates (a) rather than (b).

            I suppose from your question, you also read reference #2? How did you determine the results to mean that “Joes and Janes being provably not well nourished?”

            1. Lebenleber says:

              “I actually pulled up the referenced article, determined how intake of many nutrients was estimated and averaged over a period of 2 days.”

              I think you’re misrepresenting this study’s method. The researchers did not estimate nutrient intake over a period of 2 days. Instead, they collected 2 days worth of more-or-less accurate dietary intake data from ~ 16 k participants and then used the National Cancer Institute’s method for estimating _usual_ dietary intakes:
              http://appliedresearch.cancer.gov/diet/usualintakes/method.html
              Therefore, your argumentation of “[f]alling short of [EARs] in a period of time that is much shorter than the period of time for vitamin stores” does not apply.
              Granted, I am not familiar with that method, but it seems to be a well-established one for this purpose. Do you discount this study because you disagree with this method? Would you only accept a method based on sufficiently long food diaries, even though this would be a much more complex undertaking and longer food diaries generally suffer from poor compliance?

              As for your point of (a) not necessarily leading to (b), well, for the purpose of public health, you either agree with the Estimated Average Requirements as they are currently defined or you don’t. And I already gave vitamin D as an example of how the definitions of vitamin deficiencies may be extended in the future.

              1. MadisonMD says:

                I think you’re misrepresenting this study’s method. The researchers did not estimate nutrient intake over a period of 2 days. Instead, they collected 2 days worth of more-or-less accurate dietary intake data from ~ 16 k participants and then used the National Cancer Institute’s method for estimating _usual_ dietary intakes

                In either case it is an estimate based on two days of collected data. I fail to see why you think there is a substantial difference between the two. If someone has very poor intake of vitamin A over these two days but takes in a lot at another time, do you think this would be captured? Also, thank you for linking to the NCI methodology page. It reveals another reason why intake might be underestimated: “Many studies have found misreporting of energy intake on both 24-hour recalls and food frequency instruments, almost always in the direction of underreporting; this suggests that some foods are underreported.” So it is likely vitamin intake from food is underreported and underestimated in this study.

                As for your point of (a) not necessarily leading to (b), well, for the purpose of public health, you either agree with the Estimated Average Requirements as they are currently defined or you don’t.

                I seem to have not made myself clear. I was trying to find the reason why, for example, this study would report that 25% of individuals fall short of recommended Vitamin C intake, whereas the incidence of disease caused by vitamin C deficiency, i.e. scurvy, is far far lower. I proposed that this was because the individuals with estimated low intake may actually take in sufficient average Vitamin C over longer periods of time and this overall intake might have been underrepresented when collecting only two days of data.

                But thank you for pointing this out. I also found another reason. I looked up the definition of Estimated Average Requirements and found it defined as such:
                “An Estimated Average Requirement (EAR) is the average daily nutrient intake level estimated to meet the requirements of half of the healthy individuals in a group.”
                So if the Fulgoni study is finding that more than half of the population are meeting the EAR, then shouldn’t we conclude we are right on target?

                Well, anyway, I would be happy to hear your thoughts on why diseases of vitamin deficiency are much lower than what this article would predict.

                And I already gave vitamin D as an example of how the definitions of vitamin deficiencies may be extended in the future.

                Hell, yeah. Anything may or may not be extended in the future. We may be commuting to Mars in the future… or not. I’m not sure what your point is here.

              2. MadisonMD says:

                It’s so cute how my accidentally nested blockquote got so small.

              3. Lebenleber says:

                “In either case it is an estimate based on two days of collected data. I fail to see why you think there is a substantial difference between the two.”

                It just shows that you didn’t, in fact, properly read the study. And the NCI method does claim that, statistically speaking, two days of collected data are sufficient to estimate usual intake. You seem to be switching subjects. At first, you misrepresented the study’s method, now that you know better, you dismiss the NCI method.
                Of course, ideally, such a study should be based on accurate sufficiently long food diaries, but considering that it can take years to deplete vitamin B12 stores, this is practically infeasible.

                “I was trying to find the reason why, for example, this study would report that 25% of individuals fall short of recommended Vitamin C intake, whereas the incidence of disease caused by vitamin C deficiency, i.e. scurvy, is far far lower.”

                The reason for this is probably that the EARs for vitamin C (60 – 75 mg/d) are already much higher than what is required to prevent overt scurvy (7 – 8 mg/d)*. Would you argue that 8 mg/d of vitamin C ought to be enough?

                * Hastingly googled: http://www.nrv.gov.au/nutrients/vitamin%20c.htm

                “So if the Fulgoni study is finding that more than half of the population are meeting the EAR, then shouldn’t we conclude we are right on target?”

                No, because it doesn’t tell you much about distribution. Only if 100 % of the people would meet or exceed all the EARs could we proclaim that at least 50 % of the population were well-nourished in all micronutrients.

              4. MadisonMD says:

                It just shows that you didn’t, in fact, properly read the study.

                OK. You seem to have a point here. I’ve looked more closely at the model and see that it uses a correlation matrix to estimate consumption of occasional foods over long periods of time from 2 days of recalled dietary intake. So this is attempting to account for some of the variation in diets over longer periods of time.

                And the NCI method does claim that, statistically speaking, two days of collected data are sufficient to estimate usual intake.

                It can claim anything it desires, but claims alone are not particularly convincing. Is it validated to get an accurate estimate of over 17 micronutrients? If so, what is the accuracy of the method for this purpose? I’ve looked through the references on the NCI methodology page and am not able to get numbers on this.

                You seem to be switching subjects. At first, you misrepresented the study’s method, now that you know better, you dismiss the NCI method

                Misrepresented? Dismiss? No not really. Um, you seem to think I care whether science will show that multivitamins are needed or not. Frankly, I don’t give a damn how the science comes out. However, I do feel that in order to interpret the results of a single study, one should consider why whether the results are consonant or dissonant with other factual evidence. You think I’m trying to discount the value of this study. I’m actually trying to determine why the findings are at odds with known facts. Under-reporting on diet recall is one such reason.

                The reason for this is probably that the EARs for vitamin C (60 – 75 mg/d) are already much higher than what is required to prevent overt scurvy (7 – 8 mg/d)*. Would you argue that 8 mg/d of vitamin C ought to be enough?

                This seems to be the crux of the discussion here. If 8mg/d prevents scurvy in the entire population, then– please do enlighten me– why isn’t it enough?

                [I realize that you will need a margin of safety and account for individual variation so I doubt that 8mg/d would prevent scurvy in the entire population.]

                No, because it doesn’t tell you much about distribution.

                Well, life-liver you found a weakness in my argument here–you are right. The distribution of who falls short of EAR may not match the distribution of who can go with less.

                Only if 100 % of the people would meet or exceed all the EARs could we proclaim that at least 50 % of the population were well-nourished in all micronutrients.

                This does not compute. But closer than I was. Let’s let that go for now.

                The Crux of the matter: You’ve made some good points. If you want to argue the above points more, fine. But why don’t we just cut to the quick:

                If the majority of the US population is not taking in sufficient vitamins, then where in the dickens are all the folks with scurvy, night blindness, beriberi, and pellagra?

      2. Lebenleber says:

        “First, the author is more interested in pushing multivitamins than in reshaping
        dietary habits.”

        The author states that “a well-balanced diet is the best way to get all of one’s essential nutrients”, so I disagree with your impression. From a public health perspective, however, and as most physicians know, convincing people to pop a pill is usually easier than convincing them to change their lifestyle.

        “Moreover, that paper links to a Pauling Institute vitamin E puff piece that has nothing to do with the 0.1% claim.”

        While I agree about the linked article being a puff piece, their “non-significant fraction (0.1%)” claim is properly backed by reference 2, table 1:
        http://jn.nutrition.org/content/141/10/1847/T1.expansion.html
        I personally consider this a minor matter in the whole context of the article, though.

        “So ultimately I would dismiss this as marketing masquerading as science. But if you have some actual science to offer on this subject I’m sure it would be well received here.”

        Did you just gloss over the mention of the RCTs actually finding small, but statistically significant benefits from multivitamin supplementation, a cheap, low-risk measure? I find your overly superficial reading of this article and your inappropriately snide response disappointing and more telling of your personal confirmation bias. The LPI’s advice is to buy the cheap, no-name 100 % RDA supermarket multivitamins (an advice shared by the Harvard School of Public Health), so they cannot rightfully be lumped with the quacks selling overpriced supplements as a panacea to all diseases.
        Disclaimer: I am not affiliated with the LPI or the supplement industry. And “Lebenleber” means “life liver”, intended to be ironic (“Live your life, goddamnit! #YOLO”), not “life lover”.

        1. windriven says:

          “I disagree with your impression.”

          Disagree, then. Yes, the author made sort of a quack Miranda disclaimer but the impression left by the article, to say nothing of the agenda of the Pauling Institute, is very pro-vitamin-supplements. My objection to this remains: there is more to a balanced diet than a selection of vitamins and minerals. That changing dietary habits is difficult is beside the point.

          “their “non-significant fraction (0.1%)” claim is properly backed by reference 2, table 1:”

          Fine. That should have been your reference, not the puff piece. Moreover, the fact remains that we’re talking about 200,000 people, a significant number.

          “Did you just gloss over the mention of the RCTs actually finding small, but statistically significant benefits from multivitamin supplementation, a cheap, low-risk measure?”

          No, I just don’t find it compelling.

          ” And “Lebenleber” means “life liver”, intended to be ironic (“Live your life, goddamnit! #YOLO”), not “life lover”.”

          Quite so. I misread your screen name as Lebenlieber rather than Lebenleber. In my piss-poor German they are homophones, or nearly so. My apologies if you are offended.

          1. Lebenleber says:

            “Yes, the author made sort of a quack Miranda disclaimer but the impression left by the article, to say nothing of the agenda of the Pauling Institute, is very pro-vitamin-supplements. My objection to this remains: there is more to a balanced diet than a selection of vitamins and minerals.”

            Well, I, for one, do consider the superlative “best” in “a well-balanced diet is the best way to get all of one’s essential nutrients” unambiguous. The LPI considers a daily 100 % RDA multivitamin (+ additional vitamin D) a cheap, low-risk “nutritional insurance” for the average American, as does the equally quacky Harvard School of Public Health. And aside from that, both emphatically recommend healthy habits such as a good diet (their definitions are similar) and regular exercise*.
            * LPI’s advice: http://lpi.oregonstate.edu/lpirx2.html

            “That changing dietary habits is difficult is beside the point.”

            I will disagree, then. Which measures are more likely to work for an entire population is exactly what public health is mainly about. Why fortify flour with folic acid, salt with iodine, milk with vitamin D etc. when the people could just be more mindful about their diets?

            “That should have been your reference, not the puff piece.”

            Your reading is superficial, again. What you rightfully called a puff piece was an non-essential article[1] linked to by the article I referenced[2]. See, two different articles. The article I linked to has its claims referenced in scientific journals, including the “non-significant fraction (0.1%)” claim.

            [1] http://blogs.oregonstate.edu/linuspaulinginstitute/2013/09/18/all-about-e2/
            [2] http://lpi.oregonstate.edu/news/enoughisenough-response.html

            “Moreover, the fact remains that we’re talking about 200,000 people, a significant number.”
            “No, I just don’t find it compelling.”

            Interesting. You are worried about ~ 200k people overdosing on vitamin E through supplementation, yet find the small, but relatively larger benefits found in the RCTs not compelling. I mean properly run RCTs are just the gold standard, why should we care about their findings?
            As a side note, if you blame multivitamin supplements for people overdosing on vitamin E most likely through quacky megadoses (RDA: 15 mg; UL: 1000 mg), do you also blame paracetamol because many more dead people didn’t read the PIL?
            Finally, I must contest that you fing the RCTs’ positive results not “compelling” is different from your original statement that the LPI’s response were “marketing masquerading as science”, lacking “actual science”.

            1. windriven says:

              Yes, RCTs are the gold standard and should be carefully considered. But the RCTs covering multivitamin use are ambiguous, they do not uniformly support the Pauling position. It is disingenuous to pretend that they do.

              I’m disturbed that you can easily dismiss a potential for 200k overdoses of vitamin E and just as easily embrace multivitamin use that has weak evidence supporting it. Please review Madison’s thoughtful response below. We are not, to the best of my knowledge, suffering outbreaks of scurvy and beriberi in this country.

              As an aside, a few cherry picked studies do not constitute science.

              Let’s agree that the universal adoption of a good diet rich in protein, fiber, vitamins and minerals and low in saturated fat is a worthy objective. Because we continue to disagree about the desirability of broad use of multivitamins.

              1. Lebenleber says:

                “But the RCTs covering multivitamin use are ambiguous, they do not uniformly support the Pauling position. It is disingenuous to pretend that they do.”

                No, they do not, and neither I or the LPI claimed that. The LPI mainly disagreed with the editorial (“Enough Is Enough”), pointing out that this absolute stance is not supported by the RCT’s positive findings.

                “I’m disturbed that you can easily dismiss a potential for 200k overdoses of vitamin E and just as easily embrace multivitamin use that has weak evidence supporting it.”

                I’m disturbed that you can easily dismiss a potential for “8% reduction in total and epithelial cell cancer incidence in male physicians”, “a 12% reduction in total cancer incidence excluding prostate cancer [in male physicians]” and “a significant 9% reduction in the incidence of total cataract” from one of the best-designed studies ever undertaken concering this topic.
                And I have just shown that a common 100 % RDA multivitamin supplement containing 15 mg of vitamin E combined with all other food sources is very, very unlikely to cause overdose (UL: 1000 mg), not to mention that the ULs are already set very cautiously to begin with. If I had to fear overdosing a vitamin, it’d be vitamin A.

                “As an aside, a few cherry picked studies do not constitute science.”

                From your superficial reading of the article I referenced, to the wasting of time that ensued trying to resolve the resultant misunderstandings, to you ignoring the RCTs’ positive findings, yet accepting their negative findings and the editorial, I am inclined to believe that you are the one suffering from an overly selective perception.

              2. Lebenleber says:

                To avoid a misunderstanding, regarding the PHSII, “negative findings” should be “null findings”.

              3. windriven says:

                ” I am inclined to believe that you are the one suffering from an overly selective perception.”

                And you would be correct. The Pauling Institute is more an advocacy group than a research one. You’ve come here peddling a claim that falls outside the current consensus. Extraordinary claims require extraordinary proofs. But your first citations were a Pauling news release and your own food diary. And then you kvetch that I misunderstand otr don’t take you seriously?

                If you wish to be taken seriously here, mount a compelling case and don’t expect me or anyone else here to do it for you. Waving news releases, puff pieces and a table from a class project ain’t getting it done.

                The current consensus is that daily multivitamins are of little value. I will say again, a few cherry picked studies aren’t going to change that.

              4. Lebenleber says:

                “If you wish to be taken seriously here, mount a compelling case and don’t expect me or anyone else here to do it for you.”

                I wasn’t originally mounting a case, but merely asking for a response to the LPI response because I like hearing all well-thought-out perspectives. Instead, you came, didn’t properly read the article, didn’t properly check the references, spouted some useless noise (e.g. “Bad form, life lover.”) and wasted both of your time. You now don’t even respond to my arguments anymore. Thank you for confirming my general impression that comment sections are a cesspit.

              5. windriven says:

                “I wasn’t originally mounting a case, but merely asking for a response to the LPI response”

                It was not at all clear what the point of your original comment was. It asked for a response to a news release, raised an interesting point about potassium, and went into a long chat about a 51 day collection of anecdotal dietary information. What is to be made of that?

                As I have said, the news release from Pauling is what it is – Pauling is an advocacy group and the release argued that the benefits of daily multivitamins outweigh the risks. OK. That is one group’s opinion but it is not consistent with current mainstream thought on the subject.

                The one interesting question you posed was the one about potassium. But it was buried in a flurry of not so interesting stuff. And lost. I would have been interested to hear Scott Gavura or Angora Rabbit weigh in on this but apparently it was buried deep enough to escape their attention.

              6. Lebenleber says:

                “It was not at all clear what the point of your original comment was.”

                You truly don’t get it, do you? Not everyone commenting on an article is looking to sublimate their repressed anger or reaffirm their opinions by preaching to the choir. I was genuinely looking for interesting viewpoints and perhaps an insightful rebuttal of the LPI piece, but then you came and destroyed all hope.

                “It asked for a response to a news release, raised an interesting point about potassium, and went into a long chat about a 51 day collection of anecdotal dietary information. What is to be made of that?”

                I used two paragraphs to separate the subjects. I initially wanted to introduce my question about potassium by telling the story of the food diary and how potassium came on my radar, but then decided for conciseness.

                I’m done wasting my time with you.

              7. windriven says:

                “I’m done wasting my time with you.”

                Oh good, something we can agree on!

    2. MadisonMD says:

      Staw men.

      From Linus Pauling Institute Website:

      More than 40 years ago, Dr. Pauling concluded that vitamins and other essential micronutrients play a significant role in enhancing human health and preventing chronic diseases, not just deficiency diseases.

      Now, more than 40 years later, we have yet to discover that vitamins do anything other than prevent diseases of vitamin deficiency. It’s role in deficiencies is well known, well documented, and is the subject of the second paragraph of Scott Gauvra’s post.

      So what is the response from Linus Pauling Institute (LPI), Mike Adams and the Vitamin industry? That millions of Americans are not getting the US RDA allowance of vitamins each and every day. Two massive straw men:

      (1) The claim was vitamins do more than prevent deficiency. We all know vitamin supplements can prevent deficiencies. Like, duh. That’s how they were discovered. The claim was that beyond preventing deficiencies vitamins enhance human health and prevent chronic disease. Perhaps Adams and LPI now concede that vitamins do no such thing? Without a reason to exist, LPI may wish to close its doors.

      (2) It is irrelevant that many people do not meet RDA each and every day. Most vitamin deficiencies in the U.S. remain extremely rare (exceptions noted by Scott Gauvra in paragraph #2.) The vitamin pill pushers want us to believe that if we ingest no vitamins for a day, we will wake up the next morning bleeding from every orifice and with pellagra, beriberi, scurvy, megaloblastic anemia. Wrong. Why? Vitamins are stored. Vitamin C stores are smaller than others, yet it takes 6-8 MONTHS of absolutely 0 intake to get scurvy. Moreover, you can prevent scurvy with as little as 10mg Vitamin C per day–11% of US RDA (ibid).

      So when folks tell us we need to take vitamins to prevent disease, please look for these straw men. Can vitamin pills prevent deficiency? Yes, of course. Need multivitamins be taken by the average worried-well American? No.

      1. Lebenleber says:

        Vitamin D is a good example of the semantics conundrum. Chronic vitamin D deficiency causes rickets, so the age- & sex-relative RDAs were set to prevent rickets. Yet some epidemiological studies found that vitamin D intake above the RDA was negatively correlated with some cancers, (AFAIR) osteoporosis and all-cause mortality, relative to merely meeting the RDA. So, depending on whether these associations are causal, the definition of chronic vitamin D deficiency may in the future be extended to increased risk of some cancers, osteoporosis and all-cause mortality; and the RDAs be increased accordingly.

        1. WilliamLawrenceUtridge says:

          Vitamin D is a good example of an exception – yes, low levels of vitamin D causes rickets and the guidelines are aimed at preventing rickets. And yes it is the one vitamin where chronically low levels has been associated with more adverse outcomes, and yes, the RDA will probably be increased on the basis of the gradual accumulation of evidence. But that doesn’t mean that all other vitamins are similarly low (nor that our evolutionary history was filled with much higher doses of vitamin intake). Of course, vitamin D itself is problematic since blood levels are not merely associated with dietary intake, and it’s an interesting question whether higher levels of vitamin D intake through sun exposure has a positive trade-off compared to the cancer risks.

          Medicine. It’s complicated.

          1. Lebenleber says:

            Well, it’s ultimately all a semantics game about how to define “deficiency”. I mentioned vitamin D because it is the best current example to counter his argument:
            “Now, more than 40 years later, we have yet to discover that vitamins do anything other than prevent diseases of vitamin deficiency.”

            I do, however, disagree with your claim that vitamin D were the only such exception. Take vitamin B12, deficiency of which may cause pernicious anemia and irreversible neurological damage somewhat akin to Parkinson’s. Yet subclinical vitamin B12 deficiency, possibly masked by adequate folate intake, can cause hyperhomocysteinemia, which is positively correlated with some CVD incidences, according to epidemiological studies*. Of course, this is well known, and the definition of vitamin B12 deficiency has already been extended, but his other handwavy argument
            “Anything may or may not be extended in the future. We may be commuting to Mars in the future… or not. I’m not sure what your point is here.”
            can be used to categorically dismiss all such hypotheses, including the current cause of vitamin D.

            * I don’t know to what extent cause and effect are already clear.

            1. MadisonMD says:

              Says I: After 40 years we have yet to discover that vitamins do anything other than prevent diseases of vitamin deficiency.

              Says you: We might discover it with Vitamin D in the future!

              Now how could I *possibly* counter such a a brilliant, piercing argument without resorting to handwaving? :)

            2. Lebenleber says:

              Now that I think about it, a better established example than vitamin D would be therapeutic megadoses of niacin as a statin-like drug.

              1. MadisonMD says:

                Brilliant. Genius. Verily, you thrive on technicality. You are, of course, right– prescription dose slow-release niacin can improve lipid profiles albeit with common and severe side effects. I stand corrected.

                Although I could add an asterisk with this small exception to the premise of my argument, it has nothing to do with the conclusion: There is no need for the average person to take a daily multivitamin.

                Thanks for playing.

    3. WilliamLawrenceUtridge says:

      Let’s not forget that the recommended daily intakes are specifically designed to ensure that a high-90% of the population is at no risk for deficiency, and thus do not represent the amount needed to stave off deficiency for most people. They are essentially set arbitrarily high in order to ensure most people will never be deficient.

      1. MadisonMD says:

        Yes, WLU, this is a good point that I overlooked (except to mention as an aside that only 10mg Vit C is sufficient to stave off scurvy.) It became more clear when lebenleber claimed:

        Did you just gloss over average Joes and Janes being provably not “well-nourished” in key micronutrients, according to Estimated Average Requirements (reference 2)?

        where his favorite study that ‘proved’ this merely suggested some people do not actually achieve EAR. Well it looks that he actually was technically correct– he hedged with the scare quotes around “well-nourished.” I should have noticed he was joker.

        1. Lebenleber says:

          I specifically wrote “[...] according to Estimated Average Requirements”, but this isn’t the first time you had trouble with reading comprehension. After handwaving, not understanding and more handwaving, you final argument now seems to be that the EARs as defined by the IOM “don’t real”.

          “I should have noticed he was joker.”
          You resort to personal attacks, you lose.

          1. MadisonMD says:

            you final argument now seems to be that the EARs as defined by the IOM “don’t real”.

            Maybe my reading comprehension would be better if you could write something comprehensible. I guess windriven was right about you after all.

            1. windriven says:

              Madison, This is a classic example of JAQing off. Lebenlubber appears, stammering and shuffling his feet, saying “I’m only asking a question.” In fact, he is arguing the efficacy of multivitamins by picking around the edges of Scott’s blog. By JAQing off he is freed from articulating a complete argument and defending it. He can pick and poke and whine that we didn’t read the citations to the citations of some effing press release. It is a strategy as common as it is chickensh*t.

              The topic of the blog was the paucity of evidence that daily multivitamin use is efficacious. There are rare but real risks. The JAQer doesn’t address this though, arguing obliquely about the efficacies of one or another specific vitamin without addressing the underlying issue that specific deficiencies deserve to be evaluated and treated on their own – that multivitamins are a shotgun that may correct a specific deficiency while seriously overdosing others. The point of the technique is to obscure Scott’s point rather than to refute it.

              1. MadisonMD says:

                Yes, windriven. I see this clearly now. He was toying with me.

                Whereas I was taking about the likelihood of people being malnourished, he was talking about the likelihood of people being not “well-nourished.” Whereas I was trying to uncover the truth by looking at his favorite reference, evaluating it and placing it in the context of general knowledge, and acknowledging what what I had said was, in any way factually wrong, well, his purpose was to goad me by saying I was “misrepresenting” and “dismissing” evidence until such point as I would call him a name. Then, he could claim victory in this odd childish little game of his.” Well, bravo– it was well done.

                I just wanted to make clear I did not mean to call Lebenleber names by referring to him as a ‘joker,’ but merely wished to acknowledge that I now realize that the joke was on me.

                To show my sincerity in this regard, I wish to point out that, if in fact, I had meant to call him a name, I probably would have said something something more like nasty bastard jackass or, perhaps, rotten little insolent troll. Those would have aligned more closely with my thoughts and also my general preference to avoid obscenities.

                In any case, I’ll let him blather on to his trollish heart’s desire and then allow him to JAQ off at his next stop. I do wish, however, that he would place some tape over his laptop camera so we don’t have to see anymore of this.

              2. windriven says:

                Your jpeg link is hysterical!

                There are so many interesting topics to explore. When trolls, JAQs, or just plain delusionals hijack threads I find it hard to respond with equanimity anymore.

      2. Lebenleber says:

        Yes, the RDAs are generously defined to meet the needs of 97-98 % of the population and each person should ideally have easy access to blood tests, but this isn’t the case and you therefore have to deal with uncertainties, from a public health perspective. I’ll try to summarize my personal opinon on this:
        Even ignoring special cases like vitamin D intake, the vast majority of people do not eat well, including intake of various micronutrients. These people would benefit from a 100 % RDA multivitamin supplement, but would vastly more benefit from an overall healthier lifestyle (diet, exercise, sleep, etc.). However, convincing people to change their ingrained habits is vastly more difficult than convincing them to simply pop a pill. Of course, there are also downsides to this pragmaticism, e.g. the “health halo” or “health alibi”, i.e., people believing that this one magic pill will fix all their bad habits, or people believing it could treat and prevent all diseases. Since vitamins are in general of very low toxicity (except preformed vitamin A), non-megadosed supplements combined with regular food intake do not pose a realistic overdose threat.

        1. Lebenleber says:

          I forgot to mention that educating and trying to convince people to improve their lifestyle is, of course, nevertheless very important, but unfortunately not very successful, from an epidemiological standpoint.

  27. Bryan says:

    @MF34: Although methylcobalamin (MeCbl, an alkylcobalamin) is one of the coenzyme forms of vitamin B12, research strongly suggests that once inside your cells the methyl ligand from methyl-B12 gets removed and fresh methyl-B12 is assembled de novo: 

    “Intracellular processing leads to methylcobalamin (MeCbl)3 and 5′-deoxyadenosylcobalamin (AdoCbl), which support the activities of methionine synthase and methylmalonyl-CoA mutase, respectively. In fact, even provision of an active cofactor form, e.g. MeCbl, the dominant form of the cofactor found in human plasma, demands its conversion to an intermediate that can be subsequently partitioned for AdoCbl and MeCbl synthesis.” *

    “The current work shows that newly internalized alkylcobalamins undergo dealkylation processing, a likely prerequisite for generating the biologically active cobalamin forms AdoCbl and MeCbl, and that the dealkylase activity requires the cblC protein.” **

    Not too surprising, considering that the enzym methionine synthase binds B12 in an intermediate, bivalent state, thus allowing folate in the form of methyltetrahydrofolate (MTHF) to donate its methyl group to cobalamin, creating (trivalent) methylcobalamin and THF. The methyl group is then used to methylate homocysteine, creating methionine. 

    In other words, your body transiently creates its own methylcobalamin, demethylating folate in the form of MTHF, regardless of the type of B12 that’s in your food or supplement. There is no reason to suspect that studies performed with cheaper cyanocobalamin or hydroxocobalamin fail to show more positive outcomes due to a lack of ‘high end’, ‘biologically superior’ methylcobalamin.

    * Kim et al, A human vitamin B12 trafficking protein uses glutathione transferase activity for processing alkylcobalamins, J Biol Chem 2009: http://www.jbc.org/content/284/48/33418.long

    ** Hannibal et al, Processing of alkylcobalamins in mammalian cells: a role for the MMACHC (cblC) gene product, Mol Genet Metab 2009: 
    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2709701

  28. jason says:

    Is vitamin D one of those things you should take over the winter months if you live in a northern latitude with little sun exposure?

    1. windriven says:

      @jason says

      You should consult with your physician. I live in the PNW and most days take a vitamin D supplement from about November through April. In my case one could clearly see changes in my serum level over the course of the year.

      But how effective this is at much of anything is fairly speculative. I take it because it is easy, inexpensive, and hasn’t caused any unwanted side effects that I am aware of. But I’m not betting on it having any profound upside either.

      You might also spend a little time on PubMed as there exists a blizzard of literature on the subject.

      1. jason says:

        I will do that for sure, PubMed and talk to my Doc. I understand the PNW gets a lot of rain and very little sunshine during the fall and winter so we are dealing with similar situations.

        1. Chris says:

          The rain depends on what side of the Cascades you are on, most of eastern Oregon, Washington and British Columbia are fairly arid. The problem is the latitude, I just recently learned that Lake Washington, which forms Seattle’s eastern border, is more north than all but one of the Great Lakes:
          http://blogs.seattletimes.com/northwesttraveler/2013/12/28/geography-quiz-test-your-travel-smarts/

          It gets very dim just after 4pm. It is not unusual to commute both to and from work in the dark. Today I need to remember to harvest garden herbs by 3pm for a recipe.

          1. windriven says:

            “It gets very dim just after 4pm.”

            That must be when Oz comes on.

            1. Chris says:

              ;-)

              I have no idea. But the light outside has dimmed, and it is 4:20 in the afternoon.

  29. Jason says:

    Science Daily press release from the LPI
    http://www.sciencedaily.com/releases/2013/12/131230135106.htm
    They make some very dubious claims to create an argument that wasn’t even being made in the first place, like this
    “Needed are new methodologies that accurately measure baseline nutrient levels, provide supplements or dietary changes only to subjects who clearly are inadequate or deficient, and then study the resulting changes in their health.”
    I don’t recall anywhere in the study or write up of this that it said people with deficiencies should not take vitamin supplements. The argument was that adequately fed populations would have no need for supplements, not that deficient people should not use them. Surely they would recognize such logical inconsistencies.

    But wait, there’s more. They also question whether researchers actually know the difference between in vitro and in vivo and that humans are not rats.
    “Beyond that, many scientists studying these topics are unaware of ways in which nutrients may behave differently in something like a cell culture or lab animal, compared to the human body. This raises special challenges with vitamin C research in particular.”

    The LPI is showing itself to be right in line with quack sCAMmers.

    1. MadisonMD says:

      Jason:
      You nailed it. That LPI post from sciencedaily is silly. They want new methodologies to measure nutrients so that they can claim people are deficient. They want to set the highest possible benchmark for the definiition of “well-nourished,” in order to the highest possible number of people deficient.
      In reality, the endpoint is health, not some number on a blood test. If nobody has disease from deficiency of a given nutrient, then there is no need to test for it nor to ask them to take a vitamin pill supplement.
      Pauling’s and LPI’s favorite vitamin is C, but there is nary a case of scurvy in the US. I’ve only met one doctor who has diagnosed a case of scurvy– in a gentleman that had a diet limited to whisky and vodka. We’ve already read above that incidence in the UK of this disease is approximately 1 in a million.
      This is just one of these times when the best scientific evidence favors an approach that results in less money for big pharma. Too bad– facts are facts.

      1. Andrey Pavlov says:

        @madison:

        Which is precisely why scurvy is tested on our medical boards. LOL. The joke goes that likelihood of it being on the boards is inversely proportional to how likely the condition is in the real world. I can’t tell you how many “old ladies” on a “tea and toast” diet had scurvy in my practice questions for the boards.

        “An 80 year old lady on a ‘tea and toast’ diet comes into your office with the complaint of bleeding from the gums when brushing her teeth”

        Dx: scurvy.

        I have yet to see a case and have not heard of anyone I know seeing a case.

        1. Chris says:

          Fortunately this old lady likes lemon with her tea. ;-)

          (though this morning it is coffee and toast)

      2. Andrés says:

        Ascorbic acid is needed in the generation of collagen. Not enough ascorbic acid to this function brings up scurby.

        As I have already said ascorbic acid is needed for a good function of the immune system in all of the vertebrates including us. As I have already said we are still learning new effects such attenuating lung injury bring up by bacterial toxins. As I have already said ascorbic acid level in ICU patients seems to be quite low even under total parenteral nutrition with something like 200mg/day of vitamin C. As I have already said replenishment to normal physiological level in those patients with measured low ones doesn’t need any clinical trial of any kind in my view. It is plausible that it will get a positive effect. Clinical trials are needed to check supraphysiological levels though.

        Is it possible to be low in vitamin C (let’s say more susceptible to viral infections) without having scurby? Yes. No, I don’t think that vitamin C supplementation is going to have a preventive effect in the general population but from Vitamin C for preventing and treating the common cold:

        Five trials involving a total of 598 marathon runners, skiers and soldiers on subarctic exercises yielded a pooled RR of 0.48 (95% CI 0.35 to 0.64).

        Hence my point: low vitamin C without scurby. Moreover, it seems that you may die of sepsis without scurby.

        Is there anyone out there checking ascorbic acid levels in severely ill patients?

        1. windriven says:

          “Not enough ascorbic acid to this function brings up scurby.”

          Scurvy.

          “Five trials involving a total of 598 marathon runners, skiers and soldiers on subarctic exercises yielded a pooled RR of 0.48 (95% CI 0.35 to 0.64).”

          Do you understand what a 0.48 RR means? I’m thinking not.

          Do you understand that comparing a population of high performance athletes with a general population* is a veritable minefield of confounding factors?

          *apologies to those just grazing the comments if this lacks context. Andres links to a meta that is inane even by Cochrane’s sloppy standards. I’m not sufficiently motivated to deconstruct the whole mess. You can find the meta here if you wish to analyze it yourself.

          1. MadisonMD says:

            Yet, as inane as it is, the major conclusion of the Cochrane review (before the weasel words begin) is:

            The failure of vitamin C supplementation to reduce the incidence of colds in the general population indicates that routine vitamin C supplementation is not justified

            1. Andrey Pavlov says:

              @Andres:

              Pubmed came back up and I read the article. It also does not support anything you are trying to claim, let alone the fact that the trial consisted of only nine! individuals. I mean it is very much as nearly worthless a reference as you could muster. I won’t bother deconstructing exactly why it doesn’t say what you think it says because even if it did, a study of 9 people back in 1987 is simply irrelevant anyways.

              Also, you are misunderstanding and misusing Popper in the same way as you are significant figures.

          2. Andrés says:

            windriven said:

            “Not enough ascorbic acid to this function brings up scurby.”

            Scurvy.

            Oops! I was so sure (the logical inference from “ascorbic” and the Spanish “escorbuto”) when writing it that I didn’t mind the spell-checker warning. I stand corrected.

            Do you understand what a 0.48 RR means? I’m thinking not.

            You are aware that my misspellings doesn’t do me stupid, aren’t you? Yes, I understand Relative Risk. As a further disclosure. I understand Odds Ratio. I understand the idea behind Hazard Ratio although I haven’t bother with the nuances of its computation.

            I remind you that Hemilä and Chalker said in their Selection criteria Section:

            We restricted our review to placebo-controlled trials.

            I don’t have access to the 2013 review (I already gave myself a link with the title on my previous message) but the 2005 one is freely accessible with a very enlightening Heterogeneity of results subsection on page 10:

            A subgroup analysis is shown in Figure 1 in which the six studies which involved marathon runners, skiers, and Canadian soldiers in a subarctic exercise were moved to a separate subgroup in the meta-analysis. This resulted in two distinct groups of trials which were significantly different from each other in their pooled estimates of effect. Furthermore, the two subgroups were not heterogeneous within the two pools, as indicated by the high p-values in chi-square test, and the zero values for the I square value.

            Just in case I have verified the “soldiers on subarctic exercises” study and certainly it was placebo controlled. If you are going to keep the incompetence blame (“inane even by Cochrane’s sloppy standards”) on Hemilä and Chalker I think you have to substantiate it yourself.

            I am having doubts myself about your understanding of the 0.48 RR and its 95% confidence interval of (0.35,0.64). That CI so much far away from 1.0 means a quite low probability of the result having been brought up by chance. The heterogeneity between studies on the general population and those on athletes and soldiers under stress conditions and the lack of it within each one of those two groups grants the analysis of each of the two by itself. 48% lower cold incidence of those under active prophylactic treatment with vitamin C than those under placebo on the same subpopulation is interesting enough. It certainly is not a RR “comparing a population of high performance athletes with a general population”.

            Certainly it points toward the need of a large scale trial on a similar subpopulation to refute it. I certainly think it is already enough to substantiate my claim: vitamin C deficiency without scurvy happens.

            1. MadisonMD says:

              So Andres,
              Your conclusion is that if you are a soldier on subarctic exercises, a marathon runner, or a skier, then taking you should take extra vitamin C to prevent colds?
              I’m concerned that this conclusion differs from the initial hypothesis.

              1. Jason says:

                It sounds to me like Andres is trying to make a case for labeling the common cold as a vitamin C deficiency.

              2. Andrey Pavlov says:

                I certainly think it is already enough to substantiate my claim: vitamin C deficiency without scurvy happens.

                OK…. so what? What does that vitamin C deficiency actually mean? What is happening from it that is deleterious and what are we preventing or ameliorating by supplementing it? I can make up numbers and make all sorts of “deficiencies” but if it doesn’t have identifiable harms prevented or benefits attained from supplementation I’ve accomplished nothing except treating a number with no meaning.

                Now, if you want to discuss how elite athletes or those under extreme environmental stresses may be more prone to clinical deficiency and this is reasonable cause for prophylactic supplementation that’s one thing and, depending on the specifics may be reasonable. But that has nothing to do with supplementing the general population nor with defining deficiency for the general population.

              3. Andrey Pavlov says:

                @MadisonMD:

                Thanks for clarifying the threading issue. It is a bit frustrating to me. I’ve tacked this on somewhere in reply to you in our conversation so hopefully you manage to pick up on it. Also, thank you for the compliment. I very much enjoy reading your comments as well. In fact, these days I skim most but yours are ones I always pause to read in full (and often re-read).

                You say:

                This would only seem to matter if you control with a cohort outside of your study. Doesn’t the critical care community use control arms in the trials

                The use of control arms in critical care is rather difficult. Clinical equipoise makes it hard for us to randomize groups to treatment, but so does the actual patient state. They are necessarily complex and physiologic demands may make experimental treatments difficult to implement.

                For example, using VitC. Suppose we do an RCT with VitC and randomize patients to receive either it or placebo. Well, that VitC must come with some volume of fluid. If a patient is volume overloaded and we know that we have a 50% chance of giving them just saline that would unquestionably worsen their prognosis we are hard pressed to justify giving it both ethically and because it would skew the results – giving critically ill patients fluid beyond initial resuscitation is known to increase mortality. If we just randomize a treatment arm, with no placebo control, then we run into the issue of providers knowing who is getting what and that alone can change outcomes, particularly in critically ill patients.

                So for these reasons (and more) in general, most research in critical care has used a “before and after” model where they continue doing whatever it was standard at the institution for [x] amount of time or patients and then introducing the new intervention for the following [x] amount of time or patients. You can, I’m sure, imagine the drawbacks of these sorts of studies.

                There are exceptions to this of course, but they are extremely involved and expensive and lead to big things like ARDSnet protocols. Basically, the field is difficult to do research in because the sand is constantly shifting under our feet and, by definition, our patient population has a massive amount of confounders plus a high expected mortality.

                I presented a poster at the ATS conference in Philly last year and listened to some great talks, including one by an amazing guy in the field named Derek Angus over at UPitt. He and his panel outlined a lot of these difficulties including a rather interesting one. ICU patients should, by definition, have a high mortality. Otherwise we are doing too many soft admits to the unit. But at the same time we are always striving to decrease mortality. As our knowledge, tech, and process of care improves people who would have been in the ICU a decade or two ago can now safely be on the floor instead. So we have this strange sort of push-pull of where our mortality should be, what we want it to be, and how that actually fits within reality and our models (we do a lot of Hosmer-Lemeshow goodness of fit analyses). This is further compounded by the fact that our most solidly validated mortality metric – the APACHEII score – is incredibly out of date. It is, in all reality, useless in predicting actual mortality. But since it is so well validated and older studies used it, we still use it as nothing more than a baseline comparator to prove our populations are equally sick (benefits in mortality of an intervention could, after all, be purely due to luck of the patient population being more or less sick). The APACHEIV is likely to supplant this over time and there are other measures which are less used like the SOFA score. We essentially have an invisible “floor” of mortality that is constantly shifting under our feet.

                So at any given time trying to compare outcomes proves to be tricky since it is hard to say what mortality should be and to really tease out confounders to demonstrate an effect. In fact, at that same talk Derek Young showed by example that if a hypothetical drug decreased ventilator associated pneumonia (VAP) by 50%, that would decrease mortality from VAP by 50% and since VAP accounts for up to 20% of deaths in the ICU, should decrease mortality by 10%. But to have a study powered well enough to detect that difference would require n=10,000 at 90% CI (these may not be exact numbers, I’m trying to recall off the top of my head from almost a year ago, but the ballpark is about right). And that is for a drug that, by any standards, would be a genuine miracle drug. This is largely due to that invisible floor of mortality, which is another way of saying “lots and lots of stuff leads to the death of critically ill patients.”

                One thing we know leads to deaths is delays in treatment and less aggressive treatment. Manny Rivers basically started this entire concept with the idea of Early Goal Directed Therapy (EGDT) which has become standard of care for septic patients. The problem is that overall it vastly improves mortality to follow EGDT and meets goals across the board, but we know it doesn’t fit for everyone and that certain parts of the EGDT protocol don’t actually have a good rationale. Yet, it is exceedingly hard to tease this out. For example, my own research is pretty uniquely awesome in that we are – to the best of my knowledge – the only institution to have a complete prospective, intent to treat, cohort. And with over 1,100 patients in my database we are still underpowered to detect a mortality difference for specific aspects of our protocol. And this is all compounded by the fact that in-hospital mortality is by far the easiest metric but 30-day mortality is really the gold standard (but not often used because it is so hard to obtain). So confounders upon confounders upon confounders.

                To finally tie this back to my comment to Andres and your specific question, one thing we do know – and Manny Rivers started and basically everyone since then has confirmed – is that doing nothing different except protocolizing the treatments in an order bundle and having hard limits for time to treatment and time to goals improves mortality. And depending on the data series you look at, this can account for anywhere from about a 5% to 45% relative reduction in mortality. You’ve done nothing differently except ensure rapid, aggressive, and timely care.

                So in the case of these VitC repletion studies – specifically the on Andres cites – a 30ish% relative reduction in mortality could be entirely due to merely being in a trial. Because at that point all treatments to the patient population will protocolized in order to maintain the integrity of the actually studied aspect (the VitC) and that alone would improve mortality. Determining how much of the mortality reduction is actually due to that vs the VitC is tricky and would require large cohorts with fancy statistical analysis (like H-L goodness of fit). I wasn’t exactly trying to argue that the VitC actually did nothing, but merely that the study Andres cited is not as convincing as he thinks because it could be due to everything else besides to VitC which means that at least some of the effect size is not VitC mediated. Possibly all. My own data showed a 10% absolute risk reduction in mortality (which is about 35% relative risk reduction) just by having early bundled order sets initiated. Even more interestingly is showed that missing parts of the set didn’t matter much at all – it was basically just doing it in the first place. And more interestingly it showed that once you missed the boat and didn’t initiate promptly mortality was higher and stayed flat. Meaning that if you figured it out 6 or 48 hours later (and everywhere in between) and then began the protocol it helped, but didn’t make much difference. (this last part is the paper I am currently writing).

                So when something I find somewhat implausible like VitC repletion shows effect sizes like that, I’m skeptical because of all the other things in critical care that can produce such an effect size that are a byproduct of being studied rather than the actual intervention itself. I also do not find compelling the fact that critically ill patients who die have lower VitC levels than those who survive. Interesting, yes. Causative? Not established and easily not. In a critical patient who dies everything is out of whack and it is more likely that the VitC is an incidental effect of all the other stuff rather than a contributing cause of mortality.

                I could go on caveating and explaining how it could be interesting and yadda yadda but I reckon I’ve rambled on long enough and hopefully it made some sense to you. I certainly do not have all the answers and could very well be mistaken on things. I am, after all, just graduated from medical school and still have 6 more years of GME to actually become a critical care attending. But I absolutely love the field and have done research in it for the last two years so I do know a couple things.

              4. MadisonMD says:

                …doing nothing different except protocolizing the treatments in an order bundle and having hard limits for time to treatment and time to goals improves mortality. And depending on the data series you look at, this can account for anywhere from about a 5% to 45% relative reduction in mortality.

                If true, then the standard of care should be that everything done in ICU follows a protocol, no?

                I see that it would be difficult to use placebo (saline infusion) or blind an intervention study in the ICU. But you could still randomize, which would be better than using sequential cohorts. (Or, at least you could require that the same protocol is followed +/- intervention for both of the sequential cohorts.)

                Your criticism of the Vit C experiment is valid. But why are we doing such experiments? I think we need to use the best possible control in clinical studies to avoid having useless results.

                From Feynman’s Cargo Cult Science:

                I was shocked to hear of an experiment being done at the big accelerator at the National Accelerator Laboratory, where a person used deuterium. In order to compare his heavy hydrogen results to what might happen with light hydrogen, he had to use data from someone else’s experiment on light hydrogen, which was done on different apparatus. When asked why, he said it was because he couldn’t get time on the program (because there’s so little time and it’s such expensive apparatus) to do the experiment with light hydrogen on this apparatus because there wouldn’t be any new result. And so the men in charge of programs at NAL are so anxious for new results, in order to get more money to keep the thing going for public relations purposes, they are destroying–possibly–the value of the experiments themselves, which is the whole purpose of the thing. It is often hard for the experimenters there to complete their work as their scientific integrity demands.

              5. Andrey Pavlov says:

                @MadisonMD:

                If true, then the standard of care should be that everything done in ICU follows a protocol, no?

                Yes, that is precisely what we have been arguing. And in fact it is becoming a reality. You’ll be hard pressed to find an ICU in an academic or major referral center that doesn’t follow a protocolized order set.

                The problem is that we physicians have a long history of being cowboys. And in the ICU that is even more true. It is, to borrow a phrase, the Wild West of medicine. Relatively little known, drama all the time, and stupendous wins and losses abound. Physicians in general dislike using bundled order sets – to those that don’t like it the common complaint is that it takes away the clinical judgement of medicine. And if it isn’t for clinical judgement, what are we there for? There are a million ways to justify why THAT specific patient didn’t meet the requirements for the protocol. And in fact I personally have argued with residents about whether a patient meets inclusion criteria for the protocol. It was rather unfair since the resident was not a research oriented guy and he had no idea that I have basically been running and designing the database for the protocol so I knew exactly what the criteria were.

                But you could still randomize, which would be better than using sequential cohorts.

                Yes, and this has been the predominant mode of testing in critical care. We call it a “before and after” study. It is acknowledged as being significantly flawed (particularly in answering certain questions) but for practical purposes usually the best we’ve got.

                Your criticism of the Vit C experiment is valid. But why are we doing such experiments

                An interesting – and loaded – question indeed. I could be flippant and say it is because people have an obsession with vitamins, particularly C, and are still hung up on the antioxidant theory of aging and cell damage. And that is probably at least somewhat true.

                A valid reason is because it does have some a priori likelihood of being reasonable and actually having an effect. The reality is that in sepsis specifically (my specific area of interest and research) we essentially have no treatments. None. In a nutshell our entire strategy of treating sepsis is to get antibiotics on board to eliminate the bug(s) causing it and let the body fix itself. In the meantime we do a lot of temporizing measures to try and prevent end organ damage from becoming so significant that the patient can’t recover. Basically everything we do in the treatment of sepsis (and many other critically ill patients who essentially have the exact same biochemical cascades triggered by something other than infection) is to temporize things until we can help get rid of the initial insult. VitC could possibly be a temporizing measure. It almost certainly will have no significant effect on the actual physiology going on or help the body fight off infection or whatever. But it could, hypothetically, help prevent some end organ damage from rampant over activation of the immune system and the subsequent free radical production. If it does that, it could buy a little extra time for the organs to let the antibiotics (or whatever else we are doing) actually fix the problem.

                Of course, in this case other antioxidants have at least similar prior plausibility… and they are being investigated as well. And so far nothing particularly promising. But that could be because the effect is small and hard to tease out… but would still be useful. With 800,000 cases per year of sepsis alone in the US alone, every little bit can save lives.

                Which ties in with the last reason (off the top of my head): we’re desperate. These are people in grave condition with an average of 20-50% mortality. We’ve got nothing except antibiotics and temporizing measures. Anything else useful to add to our armamentarium is desperately needed. Which is, in part, why Xigris was so widely adopted so rapidly and then found to do a whole lot of nothing.

                Study design in CCM is very difficult and honestly I just don’t know enough to really propose robust designs. I know enough to spot some of the more obvious flaws, difficulties with interpreting data, and background to have a gestalt about things but I am still a ways off from designing protocols myself.

                I think we need to use the best possible control in clinical studies to avoid having useless results.

                Absolutely agreed. And Feynman is TheMan. It is just really difficult to do in CCM is all. Derek Young had some excellent ideas and I think there may be a paradigm shift in how we do trials in CCM in the next decade or so. Not a complete replacement, but an augmentation in order to power our studies better. But it is tricky and I don’t fully understand it yet myself.

              6. MadisonMD says:

                Me:

                our criticism of the Vit C experiment is valid. But why are we doing such experiments

                Andrey:

                An interesting – and loaded – question indeed. I could be flippant and say it is because people have an obsession with vitamins, particularly C, and are still hung up on the antioxidant theory of aging and cell damage. And that is probably at least somewhat true.

                Ooops I was unclear. I meant why are we doing trials without proper controls.

                Anyway, thanks and good luck battling the cowboy attitude when you run the show.

              7. Andrey Pavlov says:

                Oi! The threading, the threading. One thing I have noticed though is that if you screw up and navigate away from the page your comment is saved so you don’t lose everything you’ve written. I can’t tell you how many times I’ve been crestfallen at losing a long and/or well thought out comment!

                @Andres:

                We’re not off to a good start here, Andres. Your very first link that you put up actually rather soundly argues against your position. Did you actually read the paper? Did you understand it?

                Firstly, the paper discusses all the problems with and inaccuracies of objectively testing for changes in free radical and AO interactions. In other words, it clearly states that there is no gold standard for these sorts of measurements and that one needs multiple surrogate markers to hope at achieving some sort of reliable result. It then goes on to discuss the data so far and concludes, rightly so, that there is absolutely no compelling evidence to think that athletes have an additional need for AO supplementation and that AO supplementation does not actually produce any sort of positive outcomes. In all cases they call the data equivocal and without good rationale or known mechanisms of action. I think you only made it to the last sentence of the abstract where they basically try and spin the data to conclude that in the absence of evidence and without evidence of harm, why not supplement athletes? But the rest of the paper really doesn’t support this conclusion and absolutely in no way whatsoever supports your contention that there is an actual deficiency possible in regards to VitC that isn’t scurvy. In other words, a resounding slap against your thesis. Your thesis may be right, but the fact that you choose this paper to represent it shows that you either did not read it or do not understand it.

                Let’s toss up a few good quotes from the paper (all emphasis mine):

                The reason for this interest in antioxidants is the finding that highly reactive chemical species, called free radicals, may increase during exercise…. In the latter study, some of the change at 5 min after exercise is probably due to the 6% decrease in plasma volume. [NB: Exactly the confounder I mentioned at some point in our discussion)... However, they did not account for exercise-induced changes in plasma volume.... It is not clear why studies examining concentrations of vitamins C and E during and after exercise show various responses [extreme heterogeneity in responses and measurements, with no clear and consistent change in any metric measured, with various methods, and failure to account for things like plasma volume changes... not exactly a bunch of good data to make your claims with Andres]…They found that trained men who were exercised to exhaustion on a treadmill had increased blood amounts of GSSG immediately after exercise, but values returned to rest within 1 h. Blood amounts of GSH did not change significantly… It has been suggested that decreased plasma GSH after exercise reflects its consumption by skeletal muscle, which results in a reduced export rate from muscle into plasma [in other words, not an actual reflection of a diminished AO capacity but of shunting to appropriately manage free radical production which then returns to baseline quickly; in other words not a sign of deficiency]… In contrast to the previous studies, Camus et al (43) observed no change in blood GSH or GSSG after uphill walking or downhill running for 35 min, and Marin et al (53) reported no change in blood GSH and GSSG during 30 min of treadmill running… Currently, it is difficult to explain these equivocal results… the discrepant results may be due to the high intersubject variability in MDA and the nonspecificity of the assay… The vitamin C supplement appeared to reduce the intensity of soreness. For approximately one-half of the subjects there was a >33% reduction in soreness compared with the placebo. [NB: highly subjective measure in a non-blinded study]… Several early studies examined the effects of vitamin C supplementation on exercise performance other than its effect on soreness. The results regarding vitamin C supplementation are equivocal, but most well-controlled studies report no beneficial effect on either endurance or strength performance… [how much more clear can you get than that?]… Likewise, studies of vitamin C restriction showed that a marginal vitamin C deficiency did not affect performance [so no AO supplementation seems to actually affect performance]… The vitamin C restriction had no harmful effect on health and did not affect V̇O2max. Thus, vitamin C supplementation in those with adequate or even inadequate status does not appear to improve exercise performance…

                I could go on and on with the rest of this paper and all of the negative findings they have, but you may as well just actually read the paper at that point. I’ll just chuck up a couple from the conclusion:

                However, the theoretical basis for why antioxidants should enhance performance is not clear. Studies have generally found that antioxidant supplements do not improve performance…It should be noted that most studies that have assessed antioxidant changes after exercise or training have used as endpoints various antioxidants or oxidative byproducts (eg, GSSG) in the blood. Increased or decreased blood concentrations do not necessarily reflect changes at the tissue level and may have minimal physiologic implications. [Hmmm... where have we heard that before? Perhaps in my own previous comments?]

                And the money quote:

                At present, data are insufficient to recommend antioxidant supplements for athletes or other persons who exercise regularly. Some researchers have suggested that megadoses and long-term use of antioxidants can be harmful

                I’m almost tempted to not even bother with the rest of your post if you can manage to cite this paper as supporting your contention that there is utility to supplementation of VitC in athletes considering how resoundingly opposite to your claim this paper is.

                But I’ve got a little bit of time and I can’t wait to see what your next reference says….

                Next one is a rather paltry and outdated support of VitC for colds. Yes, we know that PMN’s use VitC as part of their oxidative burst for killing bacteria and viruses. But some of the speculation in the article has since been demonstrated false and the conclusion of the paragraph is all you really need anyways:

                …administration of more than 1g/day [of Vitc] had no consistent effect on the incidence of common colds, but supported a moderate benefit on duration and severity of symptoms…

                We’ve discussed here plenty of times why that “moderate benefit” is a crock. Basically you are taking a highly variable disease and examining a subjective component of it and then trying to claim you can quantify duration of a cold down to the number of hours. The majority of the data on this that shows a reduction in duration is statistically significant… at something like 3-6 hours less duration. I call bullshit. You can’t tell me your cold ended at exactly 4:35pm and then factor that in. The significant figures used to calculate the statistical significance is incongruent with the precision of the measurements available and is not scientifically valid. And that’s all you’ve got with this paper as well (and in general).

                The paper concludes by saying that vitamins and minerals are important for immune function. No duh. What it doesn’t demonstrate in any way at all is that there is some sort of “in between” deficiency like you are trying to claim. All it does is speculate that maybe there is something going on because they play a role. Other data and better data show this is not supported.

                Interesting that you pick up on the part that is an in vitro study of chemotaxis but not all the rest of it that is basically just pure conjecture and a lot of maybe’s undergirded by a general comment that vitamins and minerals are important for cellular function. The chemotaxis means very little, if anything. You have no data that this effect is preserved in vivo nor that if it were it actually does anything to improve outcomes. All you have is that VitC in a petri dish makes lymphocytes move faster and that it is necessary for some differentiation. Yet nothing that indicates this is beneficial in vivo.

                Another big swing and a miss.

                You next study you pull the mini-abstract from the VitC section header and leave it at that. Once again, it is uncontroversial that VitC is necessary for proper immune function. What is in contention is that levels higher than what is necessary to induce scurvy but lower than some “normal” level (as yet undefined) lead to immune dysfunction. This paper once again does not demonstrate that and is once again full of nothing but speculation without any good data to support it. In fact, from just a couple paragraphs down from your quoted mini-abstract:

                However, most studies have complicated the interpretation of the results by administering a number of antioxidant nutrients in addition to vitamin C.

                And then more discussion of the heterogeneous, but largely negative studies on VitC supplementation and immune function. Then more animal and in vitro studies. I’ll let the conclusion speak for itself:

                Unfortunately, we are far from being able to define the optimal levels of intake required to maintain an optimal immune response to prevent or treat viral or other infectious diseases.

                Pretty thin gruel, Andres.

                I will leave it at that for now because, as much fun as it is to dissect studies that actually argue against your thesis, the Niners game is about to start and my lovely and I will be going to a sports bar for lunch and to watch the game.

                I’ll get back to the rest of it later on today, particularly the critical care stuff which also doesn’t quite support your thesis. Interesting indeed – I’ll absolutely grant that – and much more prior plausibility than the other stuff you’ve been chucking up here, but far from conclusive. From the paper itself:

                The historical cohort study design has inherent limitations. Because patients were not randomized to therapy, the mortality benefit associated with the AO group cannot be specifically attributed to this treatment.

              8. Andrey Pavlov says:

                @Andres:

                I still think you are missing the point. You state that you cherry picked some data as if that is a valid means by which to demonstrate your point. It doesn’t work that way. As MadisonMD pointed out, you are grasping at little threads that seem to support your hypothesis without realizing that those threads are from completely different tapestries that cannot be woven together. In other words, you can’t take a whole bunch of studies, with different methodologies, end points, surrogate markers, populations, and take the tiny bit from each one that gives some support of your hypothesis and think that weaves together to actually demonstrate your hypothesis. What that does is perhaps give us reason to do further specific investigation but it does not “prove” your hypothesis about VitC and the immune system such that it is now my job to falsify it. The burden of proof still rests with you. Particularly considering how the corpus of data is, at absolute best, largely equivocal which means that it is for the most part indistinguishable from noise in the data. The idea that the immune system requires ascorbate to function is banal. We accept it and that is not at all an issue of contention. What the implications of that are is, and you are going far beyond the evidence (and indeed at times contrary to the evidence) in that regard.

                Hence my comment about tooth fairy science. You haven’t established that the effect exists yet are trying to tie together strings of other evidence as if it does. Obviously you disagree otherwise you wouldn’t be making the claims you are, but nonetheless that is the case.

                I cited it because it is a more complete reference and I don’t have access to the paper by Gleeson et alter.

                So you are using a reference which provides a lot of lines of evidence against your hypothesis (that diminished levels of ascorbate have some impact on physiology without frank scurvy) in order to pull a citation from it that you think supports your hypothesis in reference to a different paper which you don’t have access to?

                Do you understand how incredibly poor scientific argumentation that is? Essentially you are saying “my hypothesis is legitimate based on a quote I cherry picked from two papers I didn’t read.” I mean, I don’t even know that I actually have to argue against that. If there was ever a more clear example of having a conclusion and then digging around to find confirming evidence I haven’t seen it (outside of young earth creationists). This is quite literally the exact opposite of good scientific inquiry.

                I don’t think it is enough to falsify the hypothesis.

                You haven’t established the hypothesis in the first place!. That is the point, Andres. And that is why you are engaging in tooth fairy science. These are the data you should be using to establish your hypothesis. Instead you’ve merely assumed it and then taken that data to be insufficient to falsify your assumption. You’ve put the cart before the horse.

                You even admit to it in the very next sentence yet fail to see the implications.

                I concede that we don’t have much human in vivo data about ascorbic effect on the immune system.

                We don’t have data about how it works, but you assume the hypothesis as valid and ask us to falsify it. The data that does exist is where you must begin. And when you yourself admit that there really isn’t much, you cannot then assume the hypothesis in the absence of the data and then demand we falsify your assumption. And what data there is is simply not convincing to support your hypothesis beyond, perhaps, further investigation.

                You have to be aware that a reduction of 3-6 hours of the mean value is most likely driven by some people under treatment getting no effect while others getting a higher time reduction than 6 hours

                No, you’ve missed the point Andres. It doesn’t matter either way. Your claim is contingent upon significant heterogeneity in the results which in and of itself makes the conclusion suspect. But even then, you seem to not understand the idea of significant figures. That was not a term I was bandying around because it sounded good. It is a scientific term with a precise meaning that has real world consequences in terms of the validity of measured results.

                Even if your assertion is true (which it is not, when you look at the actual data) the means are simply invalid as a metric by which to judge statistical significance. When you are measuring something, the numbers you compute must comport with the least accurate measurement that you have. So for example if you are measuring the average volume of oranges and you use a tape measure that has markings to 1mm and another tape measure the diameter. Let’s say you have 5 oranges with diameter of 55.1, 45.3, 50.3, 60.0, 48.9mm each (when you measure, you estimate the final digit at one beyond the maximum accuracy of your measuring device, in this case tenths of a mm). Your average diameter would be

                259.6mm/5 = 51.92mm.

                The problem is that your maximum accuracy is to the 0.1mm not 0.01mm so if you reported that value it would be scientifically inaccurate. You must report 51.9mm as the average. Now you want to calculate the volume so you find that:

                V = 4/3*pi*R^3 = 4/3*pi*(51.9/2)^3 = 73,198.35087322mm^3

                But once again, our least accurate measurement was 0.1mm, so the correct answer is 73,198.4mm

                My argument is that the inherent accuracy of measuring when a viral illness “ends” is not measurable in hours. So when you report that the mean duration of a cold was 3 hours less, that is invalid for the same reason as the diameter and volume calculations above are invalid. There is no rigorous way to define exactly how precise our measurement can be on how long your cold lasts, but anything less than 12 hours is simply ridiculous to try and assert. So a 3 hour difference is, by the use of significant figures, the same as zero since our ability to measure the difference cannot tell us whether there is a difference between 3 hours and 0 hours. If there were data showing that the colds lasted 18 or 24 hours less, then I would be much more convinced. The fact that these studies calculated data to a mean of 3-6 hours is an indication of bad use of significant figures. And that is absolutely regardless of whether it was because 10 patients had no response and 10 patients had a 12 hour response. That averages to 6 hours but, because of the limitations of the accuracy of defining when a cold “ends” the computed mean must be rounded to zero to be scientifically valid. Otherwise you must throw out the 10 with no response, leave 10 with a 12 hour response and that will be something more reasonable, but then you would have to explain to me how you justified throwing out the ones with zero response which would be very hard to do and ultimately also be very likely to invalidate the data.

                So yes, I understand how these means are computed. What you don’t understand (but hopefully do now) is why that is invalid.

                my opinion on the reduction in duration of the common cold under treatment: I think it grants further study of higher doses

                So in addition to a decrease in duration you will need to demonstrate a dose-response relationship that is distinct from any sort of study artifact or placebo response in order to think that higher doses warrant further study. The first premise (above) has yet to be established as does the second. The fact that you think it can’t be blinded further alludes to the noise in the signal inherent in these studies and gives us even less confidence that there is actually any signal to be found.

                Granted. It is inherently not conclusive. It is inherently asking for replication though.

                Now here we can agree. But once again, this sort of data does not at all support your contention that this is some sort of in-between deficiency of ascorbate that affects the immune function of otherwise healthy individuals but does not give them scurvy. In fact, even if the critical care trial did show causality clearly that still wouldn’t support your thesis – critical care patients are a very, very special population and their results cannot be generalized to other populations.

                So no matter how you try and slice it you are still left holding the bag. You cannot assume your hypothesis, there is no data to support the idea of a non-scurvy yet clinically significant deficiency of ascorbate, there is not even a good and validated level at which we might consider ascorbate to be adequate or not, regardless of clinical state. You cannot try and tie all these threads together between critically ill patients, athletes, people with colds, and you especially cannot do it using two papers you haven’t even read because you wanted to cherry pick a quote from one of the other!

                You are lacking the broad knowledge about the limitations of these sorts of studies, the confounders in how ascorbate might be handled in the body and how that may affect measured levels, the fact that there is no established means by which to define an ascorbate deficiency, nor the actual role that ascorbate plays in the immune system.

                You do realize that ascorbate – to the best we know – acts to quench the oxidative burst of lymphocytes? Which is why there is some plausibility for the use in critically ill patients – they are in a state of physiological extremis which means that all resources are mobilized and used. Generating new lymphocytes is an incredibly energy intensive process. If you lose lymphs to their own oxidative burst you sap resources for production of new ones. If you give them extra ascorbate so that each lymph may act longer before self immolation with free radicals, you allow a differential shunting resources to maintain physiology and recover organ function. At least, that is a plausible mechanism in critically ill patients. And even then it is far from established. The same rationale does not exist for non-critically ill individuals. Which is why you cannot use that research to support your other claim.

              9. Andrey Pavlov says:

                @Andres:

                You do realize that what you have just done is exactly what you are doing wrong, don’t you? You have had your thesis refuted soundly – individuals with scurvy don’t have immune deficiency issues – and then went and dug hard to cherry pick a few studies that don’t actually support your claim of immune deficiency but seem tangentially related. You didn’t even care that the articles you picked were of the poorest quality evidence – case studies and extremely old data. And then you handily conclude that it demonstrates moderate immune deficiency in pre-scurvy VitC deficient people. Once again, bullshit. The data you have cited does not in any way, shape, or form support that thesis. Period.

                Of course it is going to be difficult to distinguish infection caused either by skin breach overwhelming or properly immune system dysfunction.

                Actually it is not. Because we can do immune function assays directly on the immune cells themselves. There is no such direct data, because patients with scurvy are not considered an immunodeficient population. Your citations do nothing to demonstrate that they are, as you yourself noted. You are not doing science when you’ve have already determined your conclusion and then, in the absence of evidence, twist other studies to try and comport to your conclusion. Skin breakdown and collagen failure are the reasons for infections in late stage scurvy and nothing you have provided demonstrates anything about immune modulation in earlier stages.

                You must also realize that there is a very serious confounding factor here – who gets scurvy, Andres? Is it people who are otherwise well nourished, healthy, and by some incredible stroke of bad luck have an isolated, severe, and protracted VitC deficiency? No! It is indigent, alcoholic, drug abusers, and other people who are chronically malnourished. And it is absolutely uncontroversial that severe malnourishment leads to immunodeficiency. In the ICU we deal with this regularly – mounting a serious immune response is extremely energy intensive. The cell turnover is downright astounding. So you need a lot of caloric intake along with micro and macronutrients for a sustained immune response.

                But that has nothing to do with VitC in isolation. And you have brought absolutely zero data to bear to demonstrate that a decline in immune function in relation to VitC exists. And certainly none to demonstrate that supplementation in an otherwise healthy individual will do anything. In fact, as far back as 1978 it was recognized that VitC did nothing for the average person in terms of cold/flu duration.

                So when you say:

                If I am mistaken perhaps I am not alone.

                Please do not twist the writing of these physicians to suit your conclusion – they do not agree with you, at least not by the articles you have referenced.

                Nevertheless scurvy afflicted individuals’ immune system doesn’t approach anergy until the last stage of scurvy. From the third stage forward as late Dr. St. Medard described

                Funny how you decided to latch on to the section describing an infection and neglected this part:

                …a white blood cell count of 4,700 per cubic millimeter with a normal differential, and a platelet count of 202,000 per cubic millimeter…A hematologist recommended treatment with vitamin K for the mild elevation of the prothrombin time and in the context of the patient’s malnourished presentation…he patient was treated with multivitamins: 500 mg vitamin C, three times per day, 10 mg vitamin K, subcutaneously for 3 days, and 200 mg vitamin B12 intramuscularly for 1 dose.

                Indicating that there was a normal distribution and number of white blood cells, that the patient was globally malnourished, and that the treatment addressed all the nutritional deficiencies. There was no direct assay of immune cell function, but the point is that everything mentioned indicates all likelihood of normal immune cell functioning and you would need to speculate that they were not functioning properly – in other words, not any actual evidence to support your claim. Again.

                plasma ascorbic acid concentration has epidemiologically being correlated with immune exposure/response to some gingival pathogens. From Pussinen et alter

                Once again, extremely weak evidence of correlation, not causation, with multiple confounding issues. A hypothesis generator at best, but not in the context of the rest of the literature and the rest of our discussion. It asks questions, not answers them. So you cannot use this study to support your claim. You are grabbing for straws to try and cherry pick data to fit your conclusion. Again.

                On the first paper by Leggot et alter:

                Once again, very old data with an n=11. And you really think that this bolsters your claim? This is almost the flimsiest data you can possibly throw at me. You are digging so hard to try and prove your conclusion you resorted to this study as if it demonstrated anything? Once again, back in 1985 would have been a good hypothesis generator, but the paper itself does not add evidence to your claim. It really doesn’t matter what it says at all. But of course, you didn’t understand what that limited data tried to tell you anyways. It was noted that there was some difference in bleeding of the gums based on VitC levels. That has nothing to do with immune function but exactly to do with collagen synthesis. There is a reason why gums bleed first in the earliest stages of scurvy – they have a very high level of turnover and high demand for collagen synthesis. IT stands to reason that this would bleed in a deficient person. Now, the data really tells us nothing because it is so small, but even if we actually accepted at face value what the study tried to show us, that is all it would show. As the authors themselves say in the discussion:

                Ascorbic acid deficiency has been related to a loss of integrity of the periodontal vasculature, increase permeability of the crevicular epithelium to bacterial products, and enhanced tissue histamine sensitivity

                In other words, exactly what we would expect and nothing at all to do with immune function. Good job on cherry picking though – cherry picking an old study, with a very small n, and then cherry picking a few passages from it that superficially seem to support your thesis. But, once again, a complete fail to do so.

                Vogel et alter (1986) in The Effects of Megadoses of Ascorbic Acid on PMN Chemotaxis and Experimental Gingivitis centered on patients with “mean daily ascorbate intake level of approximately twice the recommended daily allowances”. It is not clear to me if there was any defficient patient (my curiosity didn’t get over the 20$ yet) but it seems unlikely.

                I’m now confused – you are putting up articles that argue against your position… and yet you still hold it. Of course all of it is extremely poor quality data. This one with an n=24. Don’t pay the $20 – besides the fact that I can send it to you if you really want, it doesn’t support your claim. And both groups had the same baseline ascorbate levels (0.85mg%). And, once again, supplementing them did nothing; and that includes a direct assay of the one thing we may reasonably expect VitC to do – alter PMN function.

                Sulaiman and Shehadeh (2010) in Assessment of Total Antioxidant Capacity and the Use of Vitamin C in the Treatment of Non-Smokers With Chronic Periodontitis didn’t found an statistical significant impact of adjuvant vitamin C this time over a general patient population it seems. I am considering paying 20$ for this one.

                Once again, don’t waste your money. The abstract alone should show you that it doesn’t demonstrate what you want it to anyways.

                However, the adjunctive dose of vitamin C did
                not offer additional effect

                Besides the fact that it is still a smallish study (n=60), they subdivided the trial arms into two arms (n=15 each) which further decreased the power of the study.

                None, and I do mean none, of the clinical indicators showed any difference in any group. The only difference was at baseline “total antioxidant concentration” or TOAC between the study and case-matched control group. That difference disappeared after the first month regardless of VitC supplementation!.

                In other words, this paper, even if it were absolutely flawless and awesome and well powered (which it isn’t), says absolutely nothing useful about VitC!. All it says is that treating chronic periodontitis increases TOAC. Actually, it does say one thing about VitC – adding it to standard treatment doesn’t further increase TOAC. In other words, the VitC did nothing.

                Which is why I find it absolutely stunning that you can take all of that and say:

                So, I haven’t found evidence that falsifies the usefulness of a vitamin C repletion intervention in those periodontal patients deficient in vitamin C.

                Actually, the very studies you linked me to tend to do exactly that! In fact the newest and best study demonstrates very, very clearly that VitC played absolutely no role whatsoever in anything they did! But more to the point, you are assuming the conclusion and then trying to find data to falsify it. That is not how science works, and that is not what Popper was talking about. You have a profound misunderstanding of the process of science and how to read the body of literature on a topic.

                You are mistaken. I have already explained to you (as clearly as I am able to) why your argument is invalid when we are estimating the mean value of the difference between symptoms duration (continuous random variable) both for the control and the intervention groups and the problem with the noise (precision) in both the control and intervention group measurements is identically distributed (same random variable E) since we are blinding the experiments.

                I’m sorry Andres, but you are mistaken. I hate arguing from authority, but in this case I must. I’ve tried explaining it to you multiple times, using multiple different examples. At this point, it is not an arguable point anymore. I am vastly more educated, trained, experienced, and work in the relevant fields than you are (on this particular topic) and so you can either accept that you are wrong and try and find out why or not. It has nothing to do with blinding, and comparing means does not invalidate significant figures. It is a basic concept and you are trying to spin it by going many levels above it to try and demonstrate that you can plug numbers into an equation and get numbers out. The point is that the numbers you are plugging in are invalid. Until you understand that you will be doomed to failure in these endeavors.

                So once again you have done nothing but demonstrate how a highly motivated person with a conclusion already in mind can dig around and find a whole bunch of dreck to cherry pick and support that conclusion. And the sad part is that you haven’t even really done that. Because much of what you have put up doesn’t even do that much for you!

                You are better served learning how to do science and read studies properly than continuing to chase your false ideas here. But don’t waste your money on articles you will misunderstand anyways. Just shoot me a message or a comment and I’ll try and get it for you. At least I can try and help you from wasting your money if not your time.

            2. Andrés says:

              So, let’s get context.

              MadisonMD said:

              I was trying to find the reason why, for example, this study would report that 25% of individuals fall short of recommended Vitamin C intake, whereas the incidence of disease caused by vitamin C deficiency, i.e. scurvy, is far far lower.

              This seems to be the crux of the discussion here. If 8mg/d prevents scurvy in the entire population, then– please do enlighten me– why isn’t it enough?

              If the majority of the US population is not taking in sufficient vitamins, then where in the dickens are all the folks with scurvy, night blindness, beriberi, and pellagra?

              From Linus Pauling Institute Website:

              More than 40 years ago, Dr. Pauling concluded that vitamins and other essential micronutrients play a significant role in enhancing human health and preventing chronic diseases, not just deficiency diseases.

              Now, more than 40 years later, we have yet to discover that vitamins do anything other than prevent diseases of vitamin deficiency.

              They want to set the highest possible benchmark for the definiition of “well-nourished,” in order to the highest possible number of people deficient.

              Now, to the question by MadisonMD:

              So Andres,
              Your conclusion is that if you are a soldier on subarctic exercises, a marathon runner, or a skier, then taking you should take extra vitamin C to prevent colds?

              No. I don’t bother about the common cold prevention although it certainly was Pauling’s fault as I have already concede. It is just a piece of data pointing toward vitamin C deficiency without scurvy.

              And to the question by Andrey Pavlov:

              OK…. so what? What does that
              vitamin C deficiency actually mean? What is happening from it that is
              deleterious and what are we preventing or ameliorating by supplementing
              it?

              I am concerned about so many seriously ill patients going on —or dying— with low plasma levels of vitamin C (median value of 11μM in 62 ICU patients —even under total parenteral nutrition with 200mg/day of vitamin C— versus 61.8μM in 34 healthy controls without interquartile overlap) when we already know it is used by our immune system and nobody seems to bother checking and correcting it to a healthy physiological level when it may have a positive effect. From the Australian Ministry of Health page linked by Lebenleber:

              Newton et al (1983) showed that for intakes up to 30 mg/day, plasma concentrations are about 11 µmol/L (or 0.2 mg/dL). Above this intake, plasma concentrations increase steeply to 60 µmol/L and plateau at 80 µmol/L, the renal threshold.

              60μM seems quite a good arbitrary goal already achieved by a sample of a healthy population. Later we may gather more data.

              Of course I would like to see clinical trials testing supraphysiological levels on infectious diseases in my lifetime too.

              Jason said:

              It sounds to me like Andres is trying to make a case for labeling the common cold as a vitamin C deficiency.

              I have already conceded that it has positively been refuted that vitamin C prevents colds. Nevertheless incidence will be higher in very specific subpopulations presumably due to lower plasma levels.

              1. Andrey Pavlov says:

                So I’ve read the articles you linked to. Interesting stuff, though very, very preliminary. Also not particularly impressive, though certainly worth investigating. However, patients in ICU’s are a very, very special population with completely aberrant physiologies that we as physicians are manipulating very heavily. More to the point, this data has literally absolutely nothing to do with this post. Multivitamin use, vitamin use in the general population, and supplementation to supraphysiological levels is lightyears away from looking at the role of free radical/oxidant/anti-oxidant derangement in critically ill individuals. BTW, I do research in critical care and have presented a poster of some of my work at the American Thoracic Society International Conference last year so I know a thing or two about the critical care literature and management of critically ill patients.

                They demonstrate roughly a 28% relative decrease in mortality. I won’t dissect out all the details of each study and why that number is in and of itself suspect since, as I said above, none of this has anything to do with the topic at hand. However, to put it in perspective my own research data shows a 36% RRR in patients simply by diagnosing patients earlier and following a protocol. And lots of data series have shown that in critically ill patients merely following a protocol with bundled order sets decreases mortality anywhere from 5-45% RRR. And that is without doing anything new or different – just doing it in a timely and complete manner.

                So when you take critically ill patients and do something, anything in a protocolized manner you can expect some decrease in mortality just from that. The study designs used – while valid – are also very well known to be fraught with confounders that cannot be accounted for. So at best what this is telling us is that there is a derangement of free radical management and that maybe, maybe some sort of repletion of an antioxidant to help an overwhelmed system could be helpful. The first part is a completely “no duh” statement if you have any knowledge of the physiology of critically ill patients. The second is highly speculative and, sadly, not particularly well supported by the literature. There is probably some effect buried out there but it isn’t hugely robust. Not that I wouldn’t use every possible edge I could in the ICU but once again, that is a very special population and does not have any bearing on any other population of human beings.

                This would be like us talking about how Product X does not do what it is advertised for fuel efficiency in cars and you saying that Product X has some possible use in Formula 1 racing to increase top speed. Tomatoes and oranges.

                As for the finding that critically ill patients have lower serum levels of VitC… so what? Critically ill patients have scores of electrolyte and micro/macro nutrient derangements. They also tend to always be fluid overloaded and anemic which in and of itself could explain lower serum concentrations (and certainly explains part of it). The real question is “will supplementing it do anything useful?” And the answer so far is… maybe. Probably not, but maybe something small in certain cases. And that’s about all you’ve got.

                So when you say:

                60μM seems quite a good arbitrary goal already achieved by a sample of a healthy population.

                There is simply no basis for this assertion. Now if you want to do an actual STUDY and pick some arbitrary points as cutoffs… sure. But all you’ve managed to do is generate a hypothesis for a topic that has nothing to do with the discussion on this post.

                I am concerned about so many seriously ill patients going on —or dying— with low plasma levels of vitamin C

                I mean honestly the only response is “big whoop.” Why aren’t you concerned about how many of them are anemic, hypomagnesemic, hypokalemic, hypo and hyper natremic, hypoalbuminemic, etc? So many derangements to pick from why get so fixated on VitC? And we’ve studied supplementation and control of various physiological and lab value parameters. And we’ve learned a few things, but one really important underlying principle. In all cases, but especially critical care, doing more tends to be more harmful. Strict glucose control is harmful. Colloids for volume repletion are neutral at best, but likely harmful. Certain ones more than others. Blood transfusions – even in astoundingly anemic patients! – is harmful. Boy, you’d think if there was ONE thing that would be a slam dunk would be to transfuse a critically ill patient with a crit of 6. Except the data shows us that with the exception of patients with KNOWN coronary artery disease, it is more harmful than just letting them be anemic. And even then, going higher than 10 is harmful.

                So really, Andres, what is your ultimate point here? That there is some interesting stuff to suss out in critical care regarding free radical derangements and potential uses of antioxidants? I agree. And that research is done, but so far not particularly conclusive or significantly promising. Beyond that, you’ve got nothing except some particular fetish for ascorbic acid.

              2. MadisonMD says:

                Andres,
                First of all I want to thank you for challenging my ideas, making coherent argument and providing citations. It caused me to re-evaluate the statements I made above and that you nicely summarized in your post.
                I also want to apologize for not having reading all your previous posts that you are citing. Honestly I haven’t read them all and don’t plan to. I hope this is acceptable to you as I don’t really expect you to read things I’ve posted elsewhere either. I have reviewed the articles you cited, however– and I do thank you for linking to them.
                If I understand you correctly, your argument it is that:
                (a) Vitamin C has favorable immunologic effects in preclinical studies.
                AND
                (b) Vitamin C can prevent colds by half in skiers, soldiers on subarctic exercises, and marathon runners, validating the idea that Vitamin C has clinically significant positive immunologic effects in humans.
                THEREFORE
                (c) Severely ill patients are dying because of weak immunity which arises from suboptimal levels of Vitamin C.
                ————–
                There are two major weaknesses with your argument:
                1. The link between the premise and conclusion is tenuous. If Vitamin C ‘boosts immunity’ then the positive immunologic effect should be present in more than skiers, soldiers, and marathon runners. If it boosts immunity of only for elite athletes, then I find it difficult to extrapolate to the most severely ill patients. These two populations seem to be on diametrically opposite and non-overlapping areas in the spectrum of health or (if you prefer a more precise definition used in oncology) performance status.
                2. Premise (b) is highly questionable. The major problem is that this is a post-hoc unplanned subset analysis. If you include 29 trials and 11,077 patients you see no effect of vitamin C. If you select 6 of these 29 studies post-hoc, it is no surprise you can find a subset with any RR you would like to find (and the relationship among them is tenouous– are marathon runners, skiiers, and soldiers on subartic exercises really similar? Are they really distinct form the other populations?).
                [For an example of how post-hoc study selection can influence results, see Gorski's post about how a post-hoc subset analysis by a homeopath 'proved' homeopathy.]
                An additional weakness in premise (b) is that one of these studies (skiing school) was the basis for Pauling formulating his hypothesis in the first place. It should be excluded from any subsequent studies seeking to validate the hypothesis.
                The weakness in premise (b) is not just my analysis. It is actually noted by the PLOS Medicine authors who state:

                However, great caution should be exercised in generalizing from this finding, which is based mainly on marathon runners.

                So Andres, thank you again for making the argument and challenging my views. However, I do not find the argument compelling.

              3. Andrés says:

                @ Andrey Pavlov: You aren’t implying that I have been the first going somewhat off topic in this post comments, are you? Perhaps I wouldn’t have posted any comment on this Scott Gavura’s post if the topic of deficiency only when severely deficiency illness shows up (scurvy for vitamin C case) hadn’t been brought up. Nevertheless, thanks for your comment. It certainly has enlightened me as to why physicians aren’t going directly to replenish vitC in ICU patients without waiting for any clinical trial results. If it has been tried and plasma levels have been measured at the same time please provide the reference. I have just searched  and located one paper. From Ascorbic acid dynamics in the seriously ill and injured (60μM=1.05672mg/dl):

                Materials and Methods. Ascorbic acid levels were determined in 12 critically injured patients and 2 patients with severe surgical infections. Each patient received TPN supplemented with increasing doses of ascorbic acid over a 6-day period. Therapeutic responses were determined by plasma and urine measurements using high-pressure liquid chromatography.

                Results. The initial mean ± SEM baseline plasma ascorbic acid concentration was depressed (0.11 ± 0.03 mg/dl) and unresponsive following 2 days on 300 mg/day supplementation (0.14 ± 0.03; P = 1.0) and only approached low normal plasma levels following 2 days on 1000 mg/day (0.32 ± 0.08; P = 0.36). A significant increase was noted following 2 days on 3000 mg/day (1.2 ± 0.03; P = 0.005).

                Conclusion. We confirmed extremely low plasma levels of ascorbic acid following trauma and infection. Maximal early repletion of this vitamin requires rapid pool filling early in the post-injury period using supraphysiologic doses for 3 or more days.

                I don’t know why we haven’t had a double blind randomized clinical trial about vitamin C replenishment in seriously ill patients since 2003.

                Dr. Pavlov said:

                Beyond that, you’ve got nothing except some particular fetish for ascorbic acid.

                I have profusely pointed out why just like its natural history and immune system effects (some quite recently discovered perhaps involved in Allan Smith’s anecdote).

                As a personal disclosure I have confessed having even more fetishes and as a matter of fact vitamin D sufficiency has also been found to be associated with ICU mortality (so vitamin D deficiency without rickets happens). It seems that blood irradiation in the middle of the 20th century may had been somewhat effective against serious infections because of the generation of pharmacological doses of vitamin D as I have already commented.

                MadisonMD said:

                I also want to apologize for not having reading all your previous posts that you are citing.

                Fair enough. I have to admit a hyperlink fetish too.

                MadisonMD said:

                If I understand you correctly, your argument it is that:
                (a) Vitamin C has favorable immunologic effects in preclinical studies.
                AND
                (b) Vitamin C can prevent colds by half in skiers, soldiers on subarctic exercises, and marathon runners, validating the idea that Vitamin C has clinically significant positive immunologic effects in humans.
                THEREFORE
                (c) Severely ill patients are dying because of weak immunity which arises from suboptimal levels of Vitamin C.

                Actually my argument is (a), (b) and (c) point toward existence of vitamin C deficiency without scurvy. This being said I consider the hypothesis “severely ill patients will have a lower mortality if vitamin C replete” plausible enough to grant a refutation trial on severely ill patients. From the paper I already have quoted above:

                Increased free radical levels appear to compromise the antioxidant levels in the critically ill and adversely affect outcome. In this regard, Miyagatani et al. reported that high levels of ascorbic acid (133 mg/kg/h) resulted in an 80% survival (P = 0.05) of septic rats compared to 50% without ascorbic acid [25]. They also noted increased levels of hepatic glutathione, the principal intracellular free radical scavenger, in the high-dose ascorbic acid-treated group.

                I know, rats. But ascorbic acid supplementation having the same effect on us is a very plausible hypothesis. It certainly should be refuted.

                I consider all the common cold research brought up by Pauling a distraction from quite more important questions such as refutation of the antiviral properties of intravenously administered sodium ascorbate in high enough doses (beginning with successful clinical experience of late Dr. Klenner) as I have already said. The only recent intervention study I am aware of is a retrospective cohort one for the control group with an intervention of 3g/day and several other substances for seven days pointing toward a positive effect that I have been linking.

                MadisonMD said:

                If it boosts immunity of only for elite athletes, then I find it difficult to extrapolate to the most severely ill patients.

                If elite athletes showing a supplementation benefit don’t have lower plasma levels than a healthy population sample (like the one with median 61.8μM) to begin with then my point really weakens. It is a pity plasma levels didn’t get measured in those trials. I have cherry picked (it seems that not so many studies have measured a lowering of ascorbic plasma level in athletes) some results from Antioxidants: what role do they play in physical activity and health?:

                Gleeson et al (41) reported that the plasma concentration of ascorbic acid increased from 52.7 mmol/L to 67.0 mmol/L immediately after a 21-km running race. However, at 24 h after the race, the ascorbic acid concentrations decreased to 20% below pre-exercise values and remained low for the next 48 h.

                I don’t think that preventive oral vitamin C in doses lower than 10g/day is going to boost immunity in order to prevent the common cold in a repleted population (let’s say around 60μM).

                MadisonMD said:

                2. Premise (b) is highly questionable. The major problem is that this is a post-hoc unplanned subset analysis. If you include 29 trials and 11,077 patients you see no effect of vitamin C. If you select 6 of these 29 studies post-hoc, it is no surprise you can find a subset with any RR you would like to find (and the relationship among them is tenouous– are marathon runners, skiiers, and soldiers on subartic exercises really similar? Are they really distinct form the other populations?).

                That’s the reason I included the Heterogeneity of results subsection. I think that it deserves further intervention trials to refute it. Dr. Groski doesn’t mention any heterogeneity analysis on the meta-analysis of the homeopathy trials. Not that it matters since vitamin C difference lies in its prior probability brought up by its natural history.

                MadisonMD said:

                So Andres, thank you again for making the argument and challenging my views. However, I do not find the argument compelling.

                You’re welcome. If the proper double blind randomized clinical trial of intravenous vitamin C on seriously ill ICU patients were brought to existence the argument would be more compelling or completely refuted. Granted.

              4. MadisonMD says:

                That’s the reason I included the Heterogeneity of results subsection. I think that it deserves further intervention trials to refute it.

                Ah, somehow I missed this part of your argument. Even though the homeopathy analysis was of a different type (the effect of adding trials stepwise to the final results) the fact is that there was heterogeneity in the trial results for homeopathy. That is the very reason that meta-analysis is powerful to eliminate this heterogeneity by including additional studies and data.

                My point is it is very possible–nay likely– that you can get a different results with a 642-patient subgroup of 6 studies in a meta-analysis of 11077-patient and 29-trials. The heterogeneity analysis you point to simply shows that these six trials were pulled out post-hoc. This analysis hinges on the idea that there is some similarity between soldiers, skiers, and marathon runners that they make them somehow similar to each other yet distinct from the populations in other studies.

                Anyway, what you hypothesize about Vitamin C doing more than preventing scurvy is plausible. But I expect we would have had more positive Vitamin C findings by know after decades of research. I’d really like to see more exciting lab findings before spending more money on expensive and minimally powered clinical research.

                Looking at your website link, you certainly do have a fetish for vitamin C (I can read just enough spanish to see it). As a thought experiment, imagine that we, like most mammals, had retained the ability to synthesize ascorbate and it was not a Vitamin. Would you then be so fixated on this metabolite to the exclusion of nutrients? Also– do you think most non-primate animal models of disease invalid because these animals do synthesize ascorbate? (It’s interesting that you cite a rat study– they make the Vitamin C they need!)

                I have nothing against Vitamin C, but I’m very tired of seeing arguments about it– we still argue about whether it prevents cold when we have a–holy blazes– 29 trials testing the hypothesis! There is so much unknown about human biology and disease. Couldn’t we invest in something a little less studied and perhaps with fewer negative results?

              5. Andrey Pavlov says:

                @andres:

                For whatever reason, threading just refuses to work with my browser, so sorry for the misplaced comment.

                A few points.

                You aren’t implying that I have been the first going somewhat off topic in this post comments, are you? Perhaps I wouldn’t have posted any comment on this Scott Gavura’s post if the topic of deficiency only when severely deficiency illness shows up (scurvy for vitamin C case) hadn’t been brought up.

                Fair enough. I apologize for not following closely enough all of the comment threads to see that you had made an intentional change of topic. I was operating under the impression that this was still a bid to justify routine supplementation of VitC and/or other vitamins in otherwise normal individual (i.e. those without scurvy).

                Conclusion. We confirmed extremely low plasma levels of ascorbic acid following trauma and infection. Maximal early repletion of this vitamin requires rapid pool filling early in the post-injury period using supraphysiologic doses for 3 or more days.

                This is interesting about the pool, availability, and metabolism of VitC. But it doesn’t tell us if these changes in serum level actually effect any physiological change. It does, in fact, give yet one more confounder in the study of VitC.

                perhaps involved in Allan Smith’s anecdote

                A highly, highly dubious anecdote from what appears to be a highly dubious source. I would not even begin to consider it as part of the evidence nor have any bearing on anything related to the topic.

                d as a matter of fact vitamin D sufficiency has also been found to be associated with ICU mortality (so vitamin D deficiency without rickets happens)

                An interesting point to make. One that, I think, does not support your thesis. We have done heaps of studies on VitD and it seems that there are other manifestations of deficiencies and that a rather large percentage of the population is at least somewhat deplete. The point being that we have also done heaps of studies on VitC and yet do not come to similar conclusions. In reading the literature I am more convinced about VitD than VitC, yet we have an abundance of studies with C. That should tell you something and that something does not support the idea of sub-clinical VitC deficiency. At least not one that is amenable to supplementation.

                And the key here is that we have begun to identify something other than rickets that VitD deficiency can cause. It is still rudimentary and we still find that people technically “low” but otherwise completely asymptomatic do not benefit from supplementation. But we have nothing analogous in VitC.

                So you can study all sorts of things you want and link them to VitC but until you have first proven that something actually exists as a result of this non-scurvy inducing hypovitaminosis C you are practicing tooth fairy science.

                The animal studies are interesting and I agree worth pursuing, especially in the context of critically ill patients. But, as MadisonMD pointed out, here the models are even worse than usual since VitC production and utilization is an evolutionarily quirky one.

                Actually my argument is (a), (b) and (c) point toward existence of vitamin C deficiency without scurvy.

                Then what you have presented so far fails to demonstrate this. By that same token me giving acyclovir to a patient with shingles and him getting better points to an acyclovir deficiency. These data demonstrate – at best – a possible immunomodulatory effect of VitC which could play a role in changing outcomes and durations of viral illness to some degree. It does not, directly, demonstrate that a mild deficiency of VitC leads to immune dysfunction.

                This being said I consider the hypothesis “severely ill patients will have a lower mortality if vitamin C replete” plausible enough to grant a refutation trial on severely ill patients

                Sure, I think it may be reasonable to give it a try if someone is so inclined, but I don’t think it is quite so plausible. Because, as I said before, the results could all be chalked up to many things other than VitC repletion and we have no data to support that some arbitrary level of VitC has any bearing on mortality at all. It could – and so far seems likely to be – that the low VitC is a byproduct of whatever else is going on that actually leads to the mortality. Repletion doesn’t actually help improve mortality, but being in any trial does.

                In other words, critically ill patients die and they happen to die with low VitC levels. Critically ill patients with better process of care die less. Critically ill patients in a trial to give them VitC get better process of care. The effect size of the VitC trial falls within the effect size regularly seen and well established by process of care trials. So you have a confounder that can completely explain away the effect with a somewhat low level of prior plausibility in the first place.

                Once again, I still admit there may be something here, but it either hasn’t been studied well or is small enough to be hidden in other confounders. Either way, based on the general prior probability of animal studies (particularly in sepsis) translating into effective therapy for humans being rather low coupled with what is at best a small-to-modest effect size leaves me playing the odds and saying it won’t pan out. But I would love to see the data on it for someone interested in pursuing it. Your conclusions, however, are currently not supported by the state of evidence.

                If elite athletes showing a supplementation benefit don’t have lower plasma levels than a healthy population sample (like the one with median 61.8μM) to begin with then my point really weakens

                Agreed, but this idea brings in the confounder I mentioned above – how do we know that serum levels accurately reflect the levels relevant to the physiology and biochemistry involved? It could very well be that elite athletes are actually deficient but have a normal serum value because they shift pools of VitC. Or vice versa. This is something that Angora Rabbit could certainly shed a lot of light on (I wish I could invoke her like a magic fairy in my everyday life).

                I don’t think that preventive oral vitamin C in doses lower than 10g/day is going to boost immunity in order to prevent the common cold in a repleted population

                Another evidence free assertion and one that is essentially a moving of the goalposts. If VitC supplementation doesn’t work to help prevents colds in otherwise replete persons, then we just keep upping the dose, right? Maybe 15g/day is the magic number? Yet this tells us, in context with the fact that all other data on all other vitamins indicate no benefit from supranormal supplementation, that there is likely no effect or at least none worth pursuing.

                If the proper double blind randomized clinical trial of intravenous vitamin C on seriously ill ICU patients were brought to existence the argument would be more compelling or completely refuted. Granted.

                Granted indeed, but vastly more complicated than I think you realize. I could write an entire paper on just the confounders, difficulties, and practical limitations of the study you are proposing.

              6. MadisonMD says:

                Andrey,
                I enjoy reading your analysis. I’m wondering if you explain this (which puzzled me in two of your posts):

                Repletion doesn’t actually help improve mortality, but being in any trial does.

                This would only seem to matter if you control with a cohort outside of your study. Doesn’t the critical care community use control arms in the trials?

                [PS: It's not your browser-- SBM only seems to allow 4 levels of replies-- we are in the 4th.]

              7. Andrés says:

                MadisonMD said:

                The heterogeneity analysis you point to simply shows that these six trials were pulled out post-hoc. This analysis hinges on the idea that there is some similarity between soldiers, skiers, and marathon runners that they make them somehow similar to each other yet distinct from the populations in other studies.

                I would bet on intensive physical exertion lowering the body ascorbate pool just as pointed out by my previous quote from Antioxidants: what role do they play in physical activity and health? (though I miswrote the link).

                MadisonMD said:

                Anyway, what you hypothesize about Vitamin C doing more than preventing scurvy is plausible.

                Scurvy implies low enough vitamin C to impair properly formed collagen. It has already been demonstrated more functions of vitamin C than this. From Selected vitamins and trace elements support immune function by strengthening epithelial barriers and cellular and humoral immune responses that seems to restrict to the human species (I have already pointed out mixed species research about vitamin C and the immune system):

                The immune-enhancing role of vitamin C has recently been reviewed50. Vitamin C is highly concentrated in leukocytes and is used rapidly during infection. In fact, it has been defined as a stimulant of leukocyte functions, especially of neutrophil and monocyte movement. Vitamin C supplements have been shown to enhance neutrophil chemotaxis in healthy adults (1 –3 g/day) and children (20 mg/kg/day)51. In addition, supplementation with vitamin C has been demonstrated to stimulate the immune system by enhancing T-lymphocyte proliferation in response to infection increasing cytokine production and synthesis of immunoglobulins52. Vitamin C may also play a significant role in the regulation of the inflammatory response53.

                The question is if we would have impaired immune system function with low vitamin C level but it high enough for avoiding scurvy. From Nutrients and their role in host resistance to infection:

                Reduced concentrations of this vitamin in leukocytes is associated with reduced immune function [78]. In humans, the essentiality of vitamin C to the immune system is most clearly illustrated during the clinical deficiency disease, scurvy, where infections occur, and there is anergy (poor or immeasurable immune response) in almost every component of the immune system [73]. Indeed, a common method to assess vitamin C status is to measure the concentration of the vitamin in leukocytes [73].

                Yes, reference [78] is on ascorbate deficient pigs and [73] is a book I have no access to. Nevertheless if the cite is correct there will be a range of unhealthyness between scurvy (anergy of the immune system) and health.

                MadisonMD said:

                As a thought experiment, imagine that we, like most mammals, had retained the ability to synthesize ascorbate and it was not a Vitamin. Would you then be so fixated on this metabolite to the exclusion of nutrients? Also– do you think most non-primate animal models of disease invalid because these animals do synthesize ascorbate? (It’s interesting that you cite a rat study– they make the Vitamin C they need!)

                I have already pointed out benefits of vitamin C supplementation on species generating their own ascorbic acid when writing about its natural history. Rats make all the vitamin C they need when healthy. It is just more important in us since our livers don’t make it.

                Dr. Pavlov said:

                So in the case of these VitC repletion studies – specifically the on Andres cites – a 30ish% relative reduction in mortality could be entirely due to merely being in a trial. Because at that point all treatments to the patient population will protocolized in order to maintain the integrity of the actually studied aspect (the VitC) and that alone would improve mortality.

                Unlikely. Since I am under the impression that you haven’t read it I will cite from Impact of High-Dose Antioxidants on Outcomes in Acutely Injured Patients (my bolds):

                Vanderbilt University is nationally and internationally recognized for excellence in bioinformatics, including the physician order entry system WizOrder. This system includes admission protocols for all trauma patients, of which the AO protocol is a part. The compliance with this protocol is well above 90% according to hospital and pharmacy administrative databases.

                The unadjusted relative risk indicates that patients receiving AO have a 30% less risk of dying during their hospitalization (OR 0.70, 95% confidence interval (CI), 0.56-0.88). After adjusting for age, gender, and probability of survival, AO exposure was associated with an even stronger protective effect (OR 0.32, 95% CI 0.22-0.46).

                Expected Survivors had a probability of survival >50% (TRISS >0.50) and Expected Deaths had a probability of survival of <50% (TRISS <0.50). For the Expected Survivors group, there appeared to be no benefit to AO treatment (OR 1.0, 95% CI 0.70-1.4, P = .98), whereas the treatment effect was even greater in the Expected Deaths group (OR 0.24, 95% CI 0.15-0.37, P < .001), adjusted and unadjusted.

                As mentioned in the methods section, major improvements to critical care such as glucose control, ventilator-associated pneumonia-prevention bundles, goal-oriented resuscitation, the use of steroids for adrenal insufficiency, and deep venous thrombosis and stress gastritis prophylaxis preventions were all in place prior to October 1, 2005. No other significant critical care change could be observed during the AO+ group time period of October 1, 2005 moving forward.

                Dr. Pavlov said:

                In reading the literature I am more convinced about VitD than VitC, yet we have an abundance of studies with C. That should tell you something and that something does not support the idea of sub-clinical VitC deficiency. At least not one that is amenable to supplementation.

                It tells me it is much more difficult to be vitamin C than vitamin D deficient. It seems our kidneys are very good not letting it drop for the majority of the population. That it is rarer doesn’t mean it doesn’t happen.

                Dr. Pavlov said:

                So you can study all sorts of things you want and link them to VitC but until you have first proven that something actually exists as a result of this non-scurvy inducing hypovitaminosis C you are practicing tooth fairy science.

                I have already linked to the study showing a lowering of plasma vitamin C after physical exertion which likely is the common factor of the subpopulation showing a benefit (lower cold incidence) of vitamin C supplementation. I think that is an hypothesis that explains the data and that is consistent with what we know of the effects of ascorbic acid on the immune system.

                I must have misunderstood Dr. Hall whenever she talks about Fairy Tale Science. It seems to me she’s talking about the futility of applying the scientific method to unplausible treatments like homeopathy or acupuncture.

                Dr. Pavlov said:

                I don’t think that preventive oral vitamin C in doses
                lower than 10g/day is going to boost immunity in order to prevent the
                common cold in a repleted population

                Another evidence free assertion and one that is essentially a moving of the goalposts.

                Sorry, I was trying to be completely inclusive with those doses that have been shown not to prevent the common cold in the general population.

                Dr. Pavlov said:

                If VitC supplementation doesn’t work to help prevents colds in otherwise replete persons, then we just keep upping the dose, right? Maybe 15g/day is the magic number?

                I know that higher preventative doses have been postulated by late Pauling, late Dr. Klenner and late Dr. Cathcart at least. I am not convinced at all that we are going to get any measurable improve over lower doses though.

                Dr. Pavlov said:

                If the proper double blind randomized clinical trial of intravenous vitamin C on seriously ill ICU patients were brought to existence the argument would be more compelling or completely refuted. Granted.

                Granted indeed, but vastly more complicated than I think you realize. I could write an entire paper on just the confounders, difficulties, and practical limitations of the study you are proposing.

                I am starting to see that. Of course confounders would be dealt with if it could be randomized and double blind. At least a small randomized one already took place on critically ill surgical patients.

              8. Harriet Hall says:

                “I must have misunderstood Dr. Hall whenever she talks about Fairy Tale Science”

                I coined the phrase “tooth fairy science” to refer to doing studies on something that has not been proven to exist, like the tooth fairy, acupuncture points, the memory of water, and the “human energy field.” The key point is not lack of plausibility, but lack of evidence for existence. It’s an extension of Ray Hyman’s categorical imperative: the idea that a phenomenon should be confirmed to exist before attempting an explanation. A phenomenon should be confirmed to exist before clinical studies are done on it.

              9. Andrés says:

                Dr. Hall said:

                The key point is not lack of plausibility, but lack of evidence for existence. It’s an extension of Ray Hyman’s categorical imperative: the idea that a phenomenon should be confirmed to exist before attempting an explanation.

                Ok. Thanks for the clarification. I understand Dr. Pavlov’s point now, not that I shared it. I think that there is evidence enough to support that the immune system doesn’t function at full gear with low ascorbate levels. Not that the evidence is so clear cut to say that it has been proven (I more interested on falsifying than proving though).

                Dr. Pavlov said:

                We’re not off to a good start here, Andres. Your very first link that you put up actually rather soundly argues against your position. Did you actually read the paper? Did you understand it?

                I didn’t read it apart the Subsection Changes in ascorbic acid and tocopherol since I am not interested on athletic performance in the least. I cited it because it is a more complete reference and I don’t have access to the paper by Gleeson et alter. When citing that paper I explicitly said (bolds added now and link corrected):

                I have cherry picked (it seems that not so many studies have measured a lowering of ascorbic plasma level in athletes) some results from Antioxidants: what role do they play in physical activity and health?

                So finding lower levels of plasma ascorbate seems to depend on how intense the exercise is and when you measure. In Clarkson’s and Thompson’s words:

                It is not clear why studies examining concentrations of vitamins C and E during and after exercise show various responses. This variability may be due to the differences in the mode of exercise used, the time points examined, the level of training of the subjects, environmental factors (eg, altitude), or lack of control for changes in plasma volume.

                I don’t think it is enough to falsify the hypothesis.

                I concede that we don’t have much human in vivo data about ascorbic effect on the immune system. I know that although ascorbic acid natural history is ancient its effect in different mammal species’ immune systems is not going to be exactly the same but I don’t think it is going to be completely different either.

                Dr. Pavlov said:

                The majority of the data on this that shows a reduction in duration is statistically significant… at something like 3-6 hours less duration. I call bullshit.

                You have to be aware that a reduction of 3-6 hours of the mean value is most likely driven by some people under treatment getting no effect while others getting a higher time reduction than 6 hours. I have already stated my opinion on the reduction in duration of the common cold under treatment: I think it grants further study of higher doses but it is not going to be possible to keep it blinded.

                Dr. Pavlov quoted (your bolds) from Impact of High-Dose Antioxidants on Outcomes in Acutely Injured Patients

                The historical cohort study design has inherent limitations. Because patients were not randomized to therapy, the mortality benefit associated with the AO group cannot be specifically attributed to this treatment.

                Granted. It is inherently not conclusive. It is inherently asking for replication though.

              10. MadisonMD says:

                @Andres:

                I think that there is evidence enough to support that the immune system doesn’t function at full gear with low ascorbate levels. Not that the evidence is so clear cut to say that it has been proven

                (I more interested on falsifying than proving though).

                Granted. It is inherently not conclusive. It is inherently asking for replication though.

                The fundamental problem with everything you are saying here is they are hypotheses based on threads of evidence. You admit that yourself. You say it is not “clear cut,” not proven, and “not conclusive.” I agree. But it is not our job to falsify your hypothesis.

                There are an infinite number of potential hypotheses. The job of scientists is not to disprove yours. IMHO, your hypothesis is unlikely to be true. I’m investing my efforts in others, thank you. But that does not preclude you or others for investigating further.

                Come back– I’d like to hear from you– when you have more evidence.

              11. Andrés says:

                Dr Pavlov said:

                The burden of proof still rests with you.

                It is not mathematics. I’m just interested on any information that clearly falsifies it à la Popper. I am trying to expose the supporting evidence as thin as it is (not my fault) since I haven’t found the falsifying one.

                We have more effects on the immune system than chemotaxis. I have already commented upon the new discovered role of ascorbic acid on the regulation of neutrophil extracelular trap (NET) formation. From Vitamin C: A Novel Regulator of Neutrophil Extracellular Trap Formation:

                VitC also significantly attenuated PMA [phorbolmyristate acetate] induced NETosis in PMNs [polymorphonuclear neutrophils] from healthy human volunteers.

                Yes, in vitro. There are more in vitro studies pointing toward more effects of ascorbic acid on the human immune system. From The Effect of Ascorbic Acid on Production of Human Interferon and the Antiviral Activity in Vitro:

                Ascorbic acid enhanced the interferon levels produced by human embryo skin and human embryo lung fibroblasts, induced by Newcastle disease virus and by polyinosinic-polycytidylic acid.

                Although other virus didn’t.

                We have some in vivo/in vitro mixed data too. From Vitamin C for the treatment of recurrent furunculosis in patients with imparied neutrophil functions:

                The effect of vitamin C treatment on 23 patients with a history of recurrent furunculosis with negative nasal cultures was studied. Neutrophil functions (chemotaxis, phagocytosis, or superoxide generation) of 12 patients were significantly lower than those of the matched controls. In this group, treatment with vitamin C (1 g/day) caused a dramatic clinical response as well as a significant improvement of neutrophil functions, reaching values similar to those of the controls. Two patients remained vitamin C-dependent. In the patients with normal neutrophil functions, vitamin C treatment neither affected neutrophil activity nor caused a clinical response.

                Dr. Pavlov said:

                So you are using a reference which provides a lot of lines of evidence against your hypothesis (that diminished levels of ascorbate have some impact on physiology without frank scurvy) in order to pull a citation from it that you think supports your hypothesis in reference to a different paper which you don’t have access to?

                I repeat. I am not concerned about athletic performance. It is interesting the epidemiological paper showing a correlation of elderly people’s physical performance with their plasma ascorbate levels though. I was interested in looking for an effect of physical exertion on vitamin C levels since the positive common cold trials didn’t measured it. The impact on health of not so much diminished ascorbate levels is not going to be so strong though. Well, the measured data shown in the abstract of Gleeson et alter is clear enough. They measured a lowering of ascorbic plasma levels between one and three days after a strong physical exertion. I am not going to apologize for not having access to the full paper.

                Dr. Pavlov said:

                When you are measuring something, the numbers you compute must comport with the least accurate measurement that you have.

                I don’t agree. If you measure mortality of a treatment you have only 0 for survival and 1 for death. That doesn’t preclude you from computing the average in order to estimate its mean value: the mortality probability. With respect to colds I don’t know it they measured anything smaller than days. I’ll try to check it. Just to discuss the concept let’s assume they didn’t. That certainly doesn’t preclude from computing averages and their difference in order to study any effect of the treatment. If the difference we find is day/8 = 3 hours or day/4 = 6 hours it is just misleading thinking of this mean value estimation of the difference as a measured difference in some cases of 3 or 6 hours. It is not. This difference being statistically significant just points to the treatment having a statistically significant impact on duration of the common cold.

              12. Andrey Pavlov says:

                @Andres:

                You are digging a deeper hole for yourself.

                I am not going to apologize for not having access to the full paper.

                No, of course you shouldn’t apologize. But you also shouldn’t make claims based on papers you haven’t actually read. Do you not understand why that is a problem? That would be like me claiming cold fusion works and referencing one paragraph from a paper and when it is pointed out that the paper doesn’t actually say what I am claiming it says, my response is “Well, I won’t apologize for not having access to the full paper… but my hypothesis still stands and is valid.”

                As if that weren’t bad enough you are now further demonstrating that you don’t actually understand the papers you are referencing. You are literally just skimming abstracts (and whatever full articles you can) to pull out whatever sounds like it might support your hypothesis and using that as evidence that it does, regardless of whether it actually does or not. To someone not skilled in reading literature and without background knowledge that can seem convincing (which is why you are convinced). But you continue to reference sepsis and critical care to a person who is both skilled at reading the literature, has background knowledge you lack, and a specific interest in sepsis and critical care.

                Case in point. You say:

                VitC also significantly attenuated PMA [phorbolmyristate acetate] induced NETosis in PMNs [polymorphonuclear neutrophils] from healthy human volunteers.

                I believe English is not your first language (it is not mine either) so this is not a dig at your language skills, but you must actually understand the words of the article in order to use them as references. Do you know what “attenuate” means? It means to decrease the activity of. Yes, I know about PNETs and the role they play in immune activity. But this article shows how VitC decreases the activity of PNETs. Which is actually nice in sepsis, because sepsis is an over activation of immune response which causes collateral damage. But it would be a bad thing in otherwise heathy people who get sick. Assuming that VitC had the same effect in healthy people as it does in the article you referenced, it would actually worsen their response to an infection!

                But of course the data shows that VitC doesn’t do anything for colds in healthy people so it can’t be having the same effect (or it is but doesn’t matter) as in the septic patients. Which is exactly why I have been saying this entire time that your referencing to septic and critically ill populations is simply not relevant to the overarching discussion. Now if you want to have a separate discussion on just septic and critically ill patients fine, but even there you are failing to demonstrate anything beyond the fact that VitC is a possibly reasonable thing to study further. Which I would agree with. But you cannot draw any conclusions based on the data so far and you absolutely cannot use that data to support any other use of VitC.

                Next:

                Ascorbic acid enhanced the interferon levels produced by human embryo skin and human embryo lung fibroblasts, induced by Newcastle disease virus and by polyinosinic-polycytidylic acid.

                And? What do you think this demonstrates for you? If you think critically ill patients are a special population where data does not translate to other populations, fetal cells are vastly more so. This literally demonstrates nothing except that VitC has a couple of interesting in vitro effects in fetal fibroblasts. You absolutely and unequivocally cannot translate that data into anything relevant to adult (or even child) immune function. Period.

                Next:

                The effect of vitamin C treatment on 23 patients with a history of recurrent furunculosis with negative nasal cultures was studied. Neutrophil functions (chemotaxis, phagocytosis, or superoxide generation) of 12 patients were significantly lower than those of the matched controls. In this group, treatment with vitamin C (1 g/day) caused a dramatic clinical response as well as a significant improvement of neutrophil functions, reaching values similar to those of the controls. Two patients remained vitamin C-dependent. In the patients with normal neutrophil functions, vitamin C treatment neither affected neutrophil activity nor caused a clinical response.

                Once again, data that is irrelevant to the discussion. It is a completely different beast to take people with documented PMN dysfunction and supplement them than to supplement people with otherwise normal PMN function. It is actually well known that in many cases like this dysfunction on a biochemical level can be a result of protein malfunction which makes it bind less effectively to cofactors necessary for its function. By giving larger amounts of the cofactor you can induce better function by shifting the Michaelis-Menten curve such that more cofactor is bound.

                So what this data tells us is that perhaps (and only perhaps) that the particular PMN dysfunction noted in chronic furunculosis could be due to changes in the VitC binding pocket of some relevant enzyme. It could also be that VitC shifts the equilibrium state of some other cofactor that is relevant to said binding pocket. Or it could be completely spurious. That is all we can tell from this data. What it does not tell us is that VitC improves PMN function.

                Next you reference an epidemiological paper as if that demonstrates anything to support your hypothesis. It does not. Firstly, and most importantly, epidemiological studies cannot demonstrate an arrow of causality. So even if the study actually looked at the question “Do people with higher VitC levels also have better immune function” (however you wish to define that, it would not be able to establish that it was the VitC that lead to the better immune function. It could be that those who had better immune function merely happened to have higher VitC levels, or that better immune function was correlated with something else (either biochemically, socioeconomically, or nutritionally) that lead to higher VitC levels.

                But even that isn’t the case here. This is once again a highly specific population – elderly Japanese females – and it had nothing to do with immune function! It had to do with grip strength and balance.

                This is even remotely relevant.

                You are very much just grabbing whatever tiny bits of information you can that seem to say anything positive about VitC and trying to make them into some horribly deformed Frankenstein of a claim that VitC does anything particularly related to health or immune function. You haven’t even come within lightyears of demonstrating that and your incredibly poor choice of references shows why.

                I was going to read through the full reference for Gleeson et al, but – and I kid you not – PubMed has gone down just now. I’ve tried accessing it multiple ways using my various institutional VPNs and PubMed is simply crashed. It is off the DNS as of right now (09:10 CST). Anyways, no matter what it actually said it simply cannot salvage your hypothesis and the rest of what you have put up is bad enough.

                Now here is the kicker. I said:

                When you are measuring something, the numbers you compute must comport with the least accurate measurement that you have

                And your response? “I don’t agree.”

                I’m sorry, Andres, but you don’t just get to “disagree” with a fundamental principle of how science is done. The concept of significant figures is just not up for agreement or disagreement. It is a fact of how measured analysis is done.

                But of course it is clear that you don’t agree because you don’t even understand what it means or how it applies. Your rebut involves mortality counts. You don’t measure mortality. You count it. At that point, significant figures don’t apply because you aren’t measuring something. When you are talking about duration of a cold, you are measuring time. If you just wanted the binary of “Did you get sick or not?” then significant figures would not apply since that is a count.

                You then say:

                With respect to colds… That certainly doesn’t preclude from computing averages and their difference in order to study any effect of the treatment.

                Of course it doesn’t preclude the fact that you can measure it. It makes the resulting measurements invalid. As per my example with the oranges I can measure anything to any arbitrary accuracy I claim. But that doesn’t mean my results are valid. Just because someone can write numbers down doesn’t mean they are valid. And in the case of duration of colds, significant figures does apply, and when you compute mean values from that and do statistics on that mean value you are doing stats on an invalid value!. If you are trying to claim that colds last for 6h less, and the accuracy of your ability to measure colds is >12h (and I am being generous here) then by definition your value is already indistinguishable from zero. So doing statistics on that value means nothing, even though yes, you can write down whatever numbers you want and put them into equations and get more numbers out.

                I’ll try one last time to explain briefly. The idea is that you are trying to see if there is any statistical difference between two means of cold duration. Let’s say one group had colds that lasted 5d18h (on average) and the other group 5d12h (on average). You could do stats on those numbers and come up with an answer. But, by definition, you already cannot tell the difference between the two groups because your accuracy in measuring is less than the difference between the groups. So the stats don’t matter at all.

                This difference being statistically significant just points to the treatment having a statistically significant impact on duration of the common cold.

                No, it doesn’t. The statistical significance points to the fact that if you plug numbers into an equation you get numbers out. But the numbers you plug in need to be valid in order for the output to be valid. And in this case, because of the inherent inaccuracy in our ability to measure the input numbers, the output numbers are equally as meaningless.

                So at this point I won’t be responding much to you. You’ve demonstrated that you don’t have the necessary understanding of how studies work, what their limitations are, how to apply study data, that you will cherry pick literally anything that sounds good to your hypothesis without even understanding what the study actually says, and will do so all without even having read the actual full paper in question. You don’t understand the difference between measurements and counts, nor the implications of significant figures and how they do and don’t apply. You are conflating evidence and very, very obviously just trying to weave together whatever evidence you can in order to prove a conclusion you already hold; in other words the exact opposite of science.

                You’ve had the benefit of MadisonMD and myself taking the time to explain the mistakes you are making and why they are mistakes. If you are genuinely interested in learning something then you can go back and learn from these lessons. You can always come back and ask questions and people here (myself included) will be happy to answer. As in “does this study support the idea of….” rather than asserting that “this study supports…” will get you a lot farther and you will have a genuine opportunity to learn a lot from the people here. And the added bonus is you will stop making a (scientific) fool of yourself by citing things which you obviously misunderstood and actually argue against your claim.

                But the fact that you already decided to “just disagree” with a simple statement of fact regarding significant figures and your obvious infatuation with VitC I don’t hold out much hope. I’d be happy if you proved me wrong though.

              13. Andrés says:

                Dr. Pavlov said:

                You are digging a deeper hole for yourself.

                I’ll keep digging. Stubborn me!

                Dr. Pavlov said:

                Yes, I know about PNETs and the role they play in immune activity. But this article shows how VitC decreases the activity of PNETs. Which is actually nice in sepsis, because sepsis is an over activation of immune response which causes collateral damage. But it would be a bad thing in otherwise heathy people who get sick. Assuming that VitC had the same effect in healthy people as it does in the article you referenced, it would actually worsen their response to an infection!

                I never said the effects of ascorbic acid to be trivial. I said it has been anciently involved in the evolution of vertebrates’ (at least) immune systems. My main point keeps being that it is unlikely that a substance with so many mechanisms of action is going to show effects of depletion just coinciding with collagen problems. With respect to PNETs I don’t think it is likely a global outcome on human subjects much worse than in the animal models tested. I have already pointed out to the rest of the papers by the research group from the Virginia Commonwealth University. Again from their last one:

                Although NETosis plays a crucial role in host defense during local infection by trapping and killing pathogens, excessive NET formation during systemic infections becomes self-defeating by promoting tissue injury and organ damage [5].

                I don’t see how attenuating excessive NET formation (last desperate all cannon salvo with double powder load) has anything to do with immune system dysfunction under a low serious infection.

                Dr. Pavlov said:

                It is a completely different beast to take people with documented PMN dysfunction and supplement them than to supplement people with otherwise normal PMN function.

                Not that I disagree. From the paper by Levy et alter under discussion (my bolds):

                The levels of vitamin C in plasma were determined before and after treatment and during follow-up. The pretreatment vitamin C concentration in patients with neutrophil dysfunction was 44.4 ± 21 /µmol/L (range, 20-60), lower (P < .05) than the levels in healthy controls of 58.1 ± 25 /µmol/L (range, 30-95). Vitamin C therapy significantly (P < .001) increased vitamin C levels in these patients (75 ± 15 /µmol/L). One year after treatment, the vitamin C plasma level in these patients was 51 ± 18 /µmol/L.
                The patients with negative nasal cultures and normal neutrophil functions (superoxide generation, chemotaxis, and phagocytosis) did not differ from the matched healthy controls.

                The initial blood concentration of vitamin C in patients with normal neutrophil function was
                higher (P < .05) than in patients with neutrophil dysfunction.

                So, I think it is quite relevant to the “vitamin C deficiency without scurvy” hypothesis.

                Dr. Pavlov said:

                By giving larger amounts of the cofactor you can induce better function by shifting the Michaelis-Menten curve such that more cofactor is bound.

                Agree. All that this argument does is pointing out toward individual variability of vitamin level goals depending on the individual genome makeup. This point has already been tested with SNPs implicated on methylation pathways involving B group vitamins.

                Dr. Pavlov said:

                Next you reference an epidemiological paper as if that demonstrates anything to support your hypothesis.

                I didn’t claim demonstration of anything (I add bolds now):

                It is interesting the epidemiological paper showing a correlation of elderly people’s physical performance with their plasma ascorbate levels though.

                It was an aside comment due to yours about athletic performance.

                Dr. Pavlov said:

                Your rebut involves mortality counts. You don’t measure mortality. You count it. At that point, significant figures don’t apply because you aren’t measuring something. When you are talking about duration of a cold, you are measuring time.

                You can count days with symptoms. What you can’t do is directly extrapolate the difference of their averages toward the difference of the mean values of continuous in time duration of a cold both under treatment and not.

                Sorry everyone (just in case is anyone left reading), but I will have to formalize.

                Lets say we have X  (resp. Y) real evaluated random variable measuring the duration of symptoms of a cold under active treatment (resp. placebo). If we are doing a sampling with only complete or partial days with symptoms we will have two integer evaluated random variables W=ceil(X) and Z=ceil(Y) with “ceil” giving us the next equal or higher integer number.

                Of course we may find distributions for X and Y to get discordant (different sign) results about the mean value of the differences. For example with Pr[X=0.9]=1/3, Pr[X=2.1]=2/3 (E[X] = (2.1∙2+0.9)/3 = 1.7, E[W] = (3∙2+1)/3 = 7/3) and Y=1.9 (deterministic) we get E[X-Y] = E[X]-E[Y] = 1.7-1.9 = -0.2 versus E[W-Z] = E[W]-E[Z] = 7/3-2 = +1/3. I don’t think it is likely to happen if we assume mild realistic restrictions such as supposing Gaussian* distributions for X and Y giving rise to  X having the same distribution than a∙Y+b (a and b constants). Under this supposition the problem of estimating E[X-Y] from a sample (a set of measures for each one) of  W and Z seems amenable to the maximum likelihood technique too. Not that I am certain though.

                Nevertheless I said (my bolds added now):

                This difference being statistically significant just points to the treatment having a statistically significant impact on duration of the common cold.

                I didn’t say “demonstrate”. I didn’t say an estimation of E[W-Z] was a good estimation of E[X-Y]. I think that it will be much more informative studying the effect of the treatment through the estimation of Pr[W=k]-Pr[Z=k] for every k, where significant figures will depend on sample size.

                *It may be even impossible under this supposition to have discordant mean value differences. I haven’t tried to demonstrate it though. Statistics are always tricky.

            3. windriven says:

              “Yes, I understand Relative Risk.”

              An RR of .48 is essentially a coin toss. That looks meaningful only when compared to an unrelated group with a different RR. But that comparison would have to be very carefully controlled for confounding factors. In the case you mentioned that was simply not the case.

              1. weing says:

                “An RR of .48 is essentially a coin toss.”
                I thought that a relative risk of 0.48 meant that your risk is about half of that of the control group. A relative risk of 1 would be a coin toss.

              2. MadisonMD says:

                Yes, Windriven.

                Andres and Weing are correct. Relative risk of 0.5 would mean half the risk of the control group. The weakness in the argument seems to be that this is a post-hoc unplanned subgroup analysis. Let me read Andreas’ PLOS Medicine article and then I can explain/comment a bit more.

              3. windriven says:

                @ weing and Madison

                Indeed, I misspoke. Thanks for the dope slap and apologies to Andres.

                My point that the comparison between the elite athletes and the quite different control group is deeply flawed remains.

            4. Andrey Pavlov says:

              @andres:

              And now you are just saying nothing. Literally absolutely nothing. I demonstrate why the study doesn’t support your hypothesis and you claim it isn’t your hypothesis, something else is. You put up more studies and I show why that doesn’t work and now it is just an “incidental interesting” thing to note that doesn’t have anything to do with the discussion.

              You even take exactly what I said and just say it again as if you said it in the first place and it demonstrates some point that we have in contention, which we don’t, as if that demonstrates you’ve been saying anything useful at all (e.g. the PNETs discussion).

              And now you are left with nothing but the most vague assertion you can muster:

              My main point keeps being that it is unlikely that a substance with so many mechanisms of action is going to show effects of depletion just coinciding with collagen problems.

              So you are arguing that VitC is important in a lot of biochemical functions beyond just collagen and that this biochemical pathway is evolutionarily ancient. Yes, that is correct. Yes, I and everyone else would agree. And yes, that is as useful a statement as “water is the essence of wetness”

              But the assumption you draw from that is, at best, evidence free and, most likely, wrong. You seem to think that because it is so ubiquitous in our biochemistry and used in a function as important as immunology that there must be some sort of deficiency that occurs before frank scurvy. Well, there is absolutely no evidence to support this. None. And you haven’t provided any. If that has been your goal, you have failed. None of the data you presented supports that thesis.

              But even moreso, it doesn’t logically follow. If something is so ubiquitous and integral to so many biochemical pathways – especially if it is evolutionarily ancient – we find that it is more difficult to perturb the system! And, of course, your entire thesis is completely sunk when you realize that people with frank scurvy don’t have an immune deficiency.

              Explain that Andres. You are arguing that there will be some sort of deficiency before levels get low enough to induce scurvy and you argue that this will likely be an immune deficiency. So why aren’t patients with scurvy immune deficient? Why are they not more susceptible to infections? Why do we not treat them like our chemo or AIDS patients in the hospital?

              I’m sorry Andres, but your thesis is either so vague as to be meaningless or completely dead in the water.

              The answer, by the way, is because as an ancient and conserved biochemical pathway, the body shunts ascorbate to where it is needed for vital functions. This is manageable by most until levels get so low that collagen synthesis is the first to suffer the consequences. Therefore it doesn’t follow that there is an in-between deficiency, supplementation in non-special populations will only have the effect of making your urine high in ascorbate, and thus the immunomodulatory effects will only possibly be present in patients with already extremely deranged immune systems if at all.

              You also still don’t understand the use of significant figures and how that applies here, despite your attempt to obfuscate that using ceiling and floor functions which have nothing at all to do with the discussion. The studies don’t use counts of days with symptoms, they use duration of symptoms. So right out the gate it doesn’t apply. But even if it did, that would still necessitate the knowledge of sig figs since doing a count of days with symptoms means that if you calculate a mean as anything less than a day it is still indistinguishable from zero and doing stats on it is still just as meaningless. Your link about sig figs and sample size is also completely irrelevant.

              So all of your pointless “this just points to” and “i’m not saying it is demonstrating” and “i’m just pointing out it is interesting” is simply pointless. What are you wasting so much time trying to actually say here, Andres? That VitC research as a broad category is interesting? Sure, I’ll agree with that. But you try and assert more than that and specifics and when shown how wrong you are you just revert to the old tactic of JAQing off.

              Honestly, this will really be my last response to you. It has become entirely fruitless and painfully obvious you know just enough to not realize how much you don’t know. Take some time to learn – it will be much more worth your while. But I assure you your ““vitamin C deficiency without scurvy” hypothesis” is completely dead in the water.

            5. Andrés says:

              Dr. Pavlov said:

              And, of course, your entire thesis is completely sunk when you realize that people with frank scurvy don’t have an immune deficiency.

              I don’t know if they are more vulnerable to infection diseases like it seems to be the case with athletes and the common cold (I will suppose this likely while waiting for refutation of the experiments). Has it been measured?

              Dr. Pavlov said (my bolds):

              The answer, by the way, is because as an ancient and conserved biochemical pathway, the body shunts ascorbate to where it is needed for vital functions. This is manageable by most until levels get so low that collagen synthesis is the first to suffer the consequences.

              Of course that’s the other possibility.

              Dr. Pavlov said:

              The studies don’t use counts of days with symptoms, they use duration of symptoms.

              Ok. I have reread everything you said about the difference of averages and significant figures. We all have our biases and I have been focusing on the hard problem.

              Dealing with your argument (I hope I understand it this time around) we have to talk about accuracy and precision. First, accuracy is not important at all when comparing two systems and we are interested only on checking if a difference exists between them without focusing on which units we use. As an example we may use an ad hoc rule not using neither inches nor centimeters but some others of our own choosing. It is important to have consistent measurement procedures both in the treatment and control groups though. Second, precision when double blinding the study should be the same between the two groups meaning we may model it with the same distribution (let’s call E the random error in the measurement), whichever it is. So we are left with two random variables measuring symptoms duration in the two groups: V=X+E and A=Y+E. So it becomes obvious that this error while consistently measurements are performed in both groups is not going to have much impact in our comparison (its noise will drive a need for larger sample size though): E[V-A] = E[V]-E[A] = E[X]-E[E] – E[Y]+E[E] = E[X]-E[Y] = E[X-Y].

              Your point is valid for the estimation of either E[X] or E[Y]. It is not for the estimation of the difference E[X-Y].

              If you are going to persist in that we can’t conclude anything about the difference you should provide a reference to the statistical literature.

              Dr. Pavlov said:

              But you try and assert more than that and specifics and when shown how wrong you are you just revert to the old tactic of JAQing off.

              Thanks for the link! I didn’t know what it meant. Phonetics don’t help either. No, I don’t think that I am keeping just asking questions. I keep looking for evidence clearly falsifying the hypothesis I am interested in.

              Nevertheless it is your time and of course you may use it as you see fit.

              1. Andrey Pavlov says:

                @andres:

                I keep looking for evidence clearly falsifying the hypothesis I am interested in.

                You’ve got the whole things backwards. This is what I meant when I said tooth fairy science and putting the cart before the horse. You don’t get to say “my hypothesis is that VitC is important in immunomodulation and there exists a deficient state before scurvy that is physiologically important” and say “now falsify that.” That is a form of begging the question. You must prove your hypothesis based on the null assumption that VitC does not have these properties you are claiming. And there is no positive case to support your hypothesis. Until there is, you cannot assume it to be the case.

                I don’t know if they are more vulnerable to infection diseases like it seems to be the case with athletes and the common cold (I will suppose this likely while waiting for refutation of the experiments). Has it been measured?

                You’ve missed my point. If the VitC deficiency is severe enough to lead to scurvy and your hypothesis were correct you would expect patients with scurvy to be vastly more immune compromised than your athletes with the common cold. They are not. Your postulate is that athletes without VitC supplementation are more susceptible to the common cold because of immune system dysfunction, correct? And this is, by definition, a lesser VitC deficiency than would be experienced in scurvy. So, it logically follows that if you are so depleted that you get scurvy you should also be much more susceptible to infection. To the point where you should be considered a special case like chemo and AIDS patients. But they are not. We do not admit scurvy patients to special rooms like we do chemo and AIDS patients, we do not take special precautions with them, and the do not get the same kinds of infections as we would expect from immunodeficient patients. You have to realize Andres, that being immune deficient means you are in a completely different diagnostic algorithm and you must start thinking of different diseases they may have. This is standard knowledge in medical training and is often even used as a “curve ball” when we get examined. You’ll be asked what is the differential for a patient, come up with answers, and then the examiner will say “What if it is the exact same patient but has mild immunodeficiency? Or severe?” and you are expected to come up with different answers. You will not get asked “what if the patient has scurvy” because that does not change the answers.

                And once again, I do not need to prove that scurvy patients don’t have immunodeficiency. You need to prove to me that they do.

                Ok. I have reread everything you said about the difference of averages and significant figures. We all have our biases and I have been focusing on the hard problem.

                No, you are still misunderstanding. I very much understand the difference between precision and accuracy and I agree that accuracy does not matter in comparing two sets of data. But precision does. And what I am saying is that in the measurement of length of colds the precision is so low that a difference of 3-6 hours is smaller than the precision of our measurement and is therefore invalid. I do not need to show you statistics texts to demonstrate this because it is not a statistics problem. It is a basic measurement problem. So all the rest of your maths is completely irrelevant to the point I have been trying to make. I’ve tried to explain it multiple times in multiple ways, so this is, I promise my last time.

                If you come to me and say “I measured this person’s cold as having last 5 days 7 hours and 23 minutes” I will say “Bullshit. You cannot have measured that. You do not have the ability to measure to that level of precision. Period”

                And that’s it. That’s the point. No stats needed. Now when you do the math and calculate averages, that same limitation applies. So if instead you said “I measured this person’s duration of a cold to 5 and one half days” I would accept that as plausible. But if you then said “I took the mean length of colds of all the people in my group and it averaged to 5 days 7 hours and 23 minutes” I will once again say “Bullshit. Your precision in your initial measurements cannot be the good, so your mean cannot be that good, you must round it to 5 days 12 hours in order to be valid.”

                So if you say to me “The average for group 1 was 5 days 7 hours and 23 minutes and for group 2 was 5 days 11 hours and 45 minutes, therefore a difference of 4 hours and 22 minutes exists” I will once again say “Bullshit. You must round group 1 to 5 days 12 hours and group 2 to 5 days and 12 hours, so there is no difference between the groups that we can measure“.

                And if you then say to me “That difference of 4 hours 22 minutes was statistically significant” I will, once again, say “Bullshit. You can’t actually resolve the difference between the means in the first place so the numbers you plugged in to calculate the p-value were invalid, so your p-value is invalid as well.”

                That is what significant figures means. And it has nothing to do with stats, floor and ceiling functions, or anything beyond the inherent lack of precision in measuring the variable you are interested in.

                Try and understand that, please.

              2. Andrés says:

                Dr. Pavlov said:

                And, of course, your entire thesis is completely sunk when you realize that people with frank scurvy don’t have an immune deficiency.

                If I am mistaken perhaps I am not alone.
                From Drs. Stephen and Utetch’s case report (my bolds):

                It is fatal if left untreated, the proximate cause being infection or sudden death.

                Of course it is going to be difficult to distinguish infection caused either by skin breach overwhelming or properly immune system dysfunction.

                Nevertheless scurvy afflicted individuals’ immune system doesn’t approach anergy until the last stage of scurvy. From the third stage forward as late Dr. St. Medard described (via Dr. Pimentel, my bolds):

                In the third state, the gums at length grow putrid, with a cadaverous smell; when they are inflamed, blood distils from them, and a gangrene ensues . . . there are often fatal hemorrhages, which break out from the external skin . . .. Obstinate ulcers of the skin arise, of the very worst kind, which no applications will cure, and which are apt to turn to a gangrene; they break out in all parts, but especially the legs, and are attended with a stench . . .. There are gnawing, rending pains, quickly shifting from place to place, which grow more violent in the night, affecting all the joints, bones, and viscera.

                In the fourth state, there are fevers of various kinds, which bring on an atrophy; sometimes diarrhea, dysenteries, or violent stranguries (painful urination);

                So, in the first stages of scurvy we are going to find very mild immune impairment at most when compared to a vitamin C replete individual. Since Dr. Pimentel includes ulcerative gingivitis on the differential diagnosis (it seems standard) I think it is interesting to take a look at the correlation between periodontal health and plasma ascorbic acid concentration. There are at least two recent epidemiological (no causation proof of course) studies pointing toward a dose-response.

                Chapple et alter (erratum: concentration of plasma ascorbic acid given as mmol/l instead of μmol/l) found the following ORs at the second forward quintiles of plasma vitamin C for severe periodontitis when “Adjusted for age, gender, race/ethnicity, cigarette smoking, OC/HRT use, diabetes, poverty-income ratio, and education and accounting for NHANES III sampling weights, stratification, and clustering (full model).” (Table 3): 0.88, 0.65, 0.58, 0.53 (only the second quintile 95%CI includes 1). ORs get lower when focusing on never-smokers: 0.67, 0.63, 0.47, 0.38 (again all 95%CIs but the second quintile one exclude 1).

                Iwasaki et alter measuring ‘periodontal disease events’ found in their multivariate adjusted model the following RRs for the second and first tertiles of plasma ascorbic acid concentration: 1.12, 1.30 (both 95%CIs exclude 1).

                Moreover, plasma ascorbic acid concentration has epidemiologically being correlated with immune exposure/response to some gingival pathogens. From Pussinen et alter:

                In the combined Finnish and Russian population, the antibody levels to P. gingivalis were negatively correlated with vitamin C concentrations (r = -0.22; P < 0.001); this association remained statistically significant (P = 0.010) in a linear regression model after adjustment for confounding factors.

                In conclusion, P. gingivalis infection is associated with low concentrations of vitamin C in plasma, which may increase colonization of P. gingivalis or disturb the healing of the infected periodontium.

                I have found one research group doing intervention trials about vitamin C in periodontal health on healthy individuals for a short period of time (depletion diet for 4 weeks). On the first paper by Leggot et alter:

                Ascorbate concentrations in body fluids and leukocytes responded rapidly to changes in ascorbic acid intake. No mucosal pathoses or changes in plaque accumulation or probing depths were noted during any of the periods of depletion or supplementation. However, measures of gingival inflammation were directly related to the ascorbic acid status. The results suggest that ascorbic acid may influence early stages of gingivitis, particularly crevicular bleeding.

                I have also located three intervention trials on patients with periodontal problems.

                Woolfe et alter (1984) in Relationship of ascorbic acid levels of blood and gingival tissue with response to periodontal therapy (I have paid 5$ in order to read it) treated with 250mg q.i.d. not deficient patients (average blood ascorbate concentration of 1.560mg/dl on the placebo group and 1.956mg/dl on the treatment group, almost all of them over the kidney threshold around 1.4mg/dl) founding no difference in gingival outcomes.

                Vogel et alter (1986) in The Effects of Megadoses of Ascorbic Acid on PMN Chemotaxis and Experimental Gingivitis centered on patients with “mean daily ascorbate intake level of approximately twice the recommended daily allowances”. It is not clear to me if there was any defficient patient (my curiosity didn’t get over the 20$ yet) but it seems unlikely. No difference on outcomes either.

                Sulaiman and Shehadeh (2010) in Assessment of Total Antioxidant Capacity and the Use of Vitamin C in the Treatment of Non-Smokers With Chronic Periodontitis didn’t found an statistical significant impact of adjuvant vitamin C this time over a general patient population it seems. I am considering paying 20$ for this one.

                So, I haven’t found evidence that falsifies the usefulness of a vitamin C repletion intervention in those periodontal patients deficient in vitamin C.

                Dr. Pavlov said:

                I will, once again, say “Bullshit. You can’t actually resolve the
                difference between the means in the first place so the numbers you
                plugged in to calculate the p-value were invalid, so your p-value is
                invalid as well.”

                You are mistaken. I have already explained to you (as clearly as I am able to) why your argument is invalid when we are estimating the mean value of the difference between symptoms duration (continuous random variable) both for the control and the intervention groups and the problem with the noise (precision) in both the control and intervention group measurements is identically distributed (same random variable E) since we are blinding the experiments.

                Dr. Pavlov said:

                Try and understand that, please.

                I do understand your argument. If the experiments weren’t blinded then your complaint about precision would be correct.

                This being said, of course I much prefer hard outcomes and that is the reason to find the findings by the Virginia Commonwealth University group the most interesting of them all.

              3. Andrés says:

                Dr. Pavlov said:

                You didn’t even care that the articles you picked were of the poorest quality evidence – case studies and extremely old data.

                Not my fault. That is what I have been able to find. With respect to clinical trials of vitamin C in periodontal health I have cited every single one I have found. Not my fault that older ones focused on replete subpopulations.

                Dr. Pavlov said:

                Skin breakdown and collagen failure are the reasons for infections in late stage scurvy and nothing you have provided demonstrates anything about immune modulation in earlier stages.

                Nevertheless we have several lines of evidence that points toward a mild lower immune system effectiveness when not vitamin C replete that I have already commented upon: lower cold incidence on soldiers/athletes under vitamin C supplementation (weak but not refuted yet), lower cold duration under vitamin C supplementation (of course it can be measured, of course it may be due to a mild antihistamine effect too), effectiveness of vitamin C treatment on those furunculosis patients with imparied neutrophil functions (their blood ascorbate concentration before treatment clearly higher than scurvy levels), effectiveness of vitamin C repletion on critically ill patients. You prefer to look closely at the trees. I prefer to look at the forest from a distance.

                Dr. Pavlov said:

                In fact, as far back as 1978 it was recognized that VitC did nothing for the average person in terms of cold/flu duration.

                Thomas and Holt said in 1978 (my bolds):

                There is general agreement that ascorbate supplementation is ineffective in reducing the incidence of cold and winter illness (Anderson, Reid & Beaton, 1972; Anderson et al., 1975; Beaton & Whalen, 1971; Wilson & Loh, 1973a; Karlowski et al., 1975; Chalmers, 1975; Tyrrell et al., 1977). Similarly, Walker, Bynoe & Tyrrell (1967) failed to demonstrate an effect of ascorbate on the infection of volunteers challenged with common cold preparations or on the susceptibility of tissue cultures to virus infection. There is also general agreement that ascorbate supplementation does produce a modest reduction in the severity of symptoms following infection (Charleston & Clegg, 1972; Anderson et al., 1972, 1975; Anderson, Suranyi & Beaton, 1974; Wilson & Loh, 1973a; Coulehan et al., 1974).

                I read the full paper and I couldn’t find anything about duration.

                Dr. Pavlov said:

                I hate arguing from authority, but in this case I must.

                You don’t hate it enough. In this case you shouldn’t have done it.

                I am vastly more educated, trained, experienced, and work in the relevant fields than you are (on this particular topic) and so you can either accept that you are wrong and try and find out why or not.

                Not in basic statistics. I have already explained to you why your extrapolation from the estimation of the mean value of a single random variable to the estimation of the difference of two random variables is not correct. Since I really don’t think you are stupid there are only two possibilities left. Either your statistics teacher just skimp through basics and focused on more applied techniques or simply you haven’t spent even two minutes trying to grasp my argument.

                I have already disclosed that I am an engineer not an MD. Some pieces of my research (three clicks apart through my blog) were actually quite focused on statistics although centered on queueing theory, simulation, traffic modeling and statistics estimation applied to computer networks.

                You may either take a closer look at my argument or bring anyone else with a good grasp of basic statistics to find the potential error on my argument. I know I am not infallible. Being focused on hard outcomes I don’t think it is going to have any impact on your research though and it is of course your choice to pass and don’t spend any more time on the subject.

                Outside of this very definite topic of course you are vastly more educated than me on the subjects dealt on this blog. The problem resides in that there are people with enough background on human biology and medicine that reviewing the literature has come to a different prior probability about vitamin C therapy than yours. At the time being I found their interpretation more compatible with the results obtained by the Virginia Commonwealth University group using (not always) intravenous vitamin C for example.

                I don’t consider sheer appeal to authority convincing anymore (as I have already said).

                Of course I would like to read myself those two papers I don’t have access to. So if you still feel generous my e-mail address is asuarez_at_det_point_uvigo_anotherpoint_es. I will restrict them to my reading.

              4. Andrés says:

                Dr. Pavlov said:

                And what I am saying is that in the measurement of length of colds the precision is so low that a difference of 3-6 hours is smaller than the precision of our measurement and is therefore invalid.

                Since the topic keeps coming up, I will try to explain why your statement does not apply to the estimation of the difference one last time.

                Let’s assume the terminology of the first figure of the wikipedia entry:

                Accuracy is the proximity of measurement results to the true value; precision, the repeatability, or reproducibility of the measurement

                In a measurement of V=X+E (respectively A=Y+E) we will have: a sample of V (resp. A) will be the measured value, with the associated sample of X (resp. Y) being the “Reference value” and the associated sample of the error E giving rise to both the accuracy through its mean value and the precision through its standard deviation. Let’s say that the error E has as mean value μ (accuracy) and as standard deviation σ (precision). In the case at hand of such a subjective measure as end of symptoms of a cold it is clearly not a good idea supposing a null value for the mean μ so we face a serious limitation in the estimation of E[X] and E[Y]. Now to the influence of the precision (σ) on the estimation of the mean value of the difference (I should use two different errors E and E’ both identically distributed but I don’t think it will improve the clarity of the exposition): E[V-A] = E[V]-E[A] = E[X]-E[E] – E[Y]+E[E] = E[X]-E[Y] = E[X-Y].

                Simply stated E[X-Y] is not constrained in the least by the precision (σ) because the error E is not having any influence but increasing the variance of the estimator of the difference (Var[V-A] = Var[X] + Var[Y] + 2 x Var[E], with Var[E]=σ²) if we do have blinded correctly the subjects in both groups.

                Dr. Pavlov said:

                If you come to me and say “I measured this person’s cold as having last 5 days 7 hours and 23 minutes” I will say “Bullshit. You cannot have measured that. You do not have the ability to measure to that level of precision. Period”

                Yes, we really don’t know how much precision (σ) we will have but perhaps it could be something like 1 hour. Yes, perhaps the accuracy (μ) shouldn’t be expected to be much lower than 1 hour or more. Yes, we could refer to the duration of this particular cold in this patient to have last 5 days and 7 hours or even 5 days plus 6 to 8 hours. No, we shouldn’t throw away the minutes away since neither the accuracy (μ) nor the precision (σ) bear any effect on the mean value of the difference of the duration.

              5. Andrey Pavlov says:

                Sorry Andres. You are still missing the point and still wrong.

                A 1 hour measurement of when a cold “ends” is absolutely ludicrous. The fundamental crux of my argument is that it is completely nonsensical to even think you could accurately measure the duration of a cold to that level of precision.

                But then you follow it up by saying we shouldn’t throw away the minutes, blah, blah. Yes, we should. That is how significant figures works. Yes, when you are doing active calculations for metrics specific to the measurement at hand you can retain more than significant figures but then at the end you are required to ditch all the decimals that don’t comport to the worst precision you have.

                So no, you still don’t understand it.

              6. Andrés says:

                Dr. Pavlov said:

                But then you follow it up by saying we shouldn’t throw away the minutes, blah, blah. Yes, we should. That is how significant figures works. Yes, when you are doing active calculations for metrics specific to the measurement at hand you can retain more than significant figures but then at the end you are required to ditch all the decimals that don’t comport to the worst precision you have.

                Your argument is just like if significant figures would be a foundational basis of statistics and not the other way around. I am not the one that doesn’t grasp it. I understand your argument perfectly and it is still wrong. Significant figures is only a convenient way to not having to specify a confidence interval. Nothing more.

                I will repeat: You may either take a closer look at my argument and point out where exactly is in error or bring anyone else with a good grasp of basic statistics to try to find it.

  30. Andrés says:

    I don’t have access to those three papers but I have zero interest in any supplementation with xenobiotic substances like dl-α-tocopherol anyway, as I have already said.

    Apart from vitamin C deficiency without scurby that I pointed out yesterday, we can find vitamin B deficiencies without either beriberi, pellagra or pernicious anemia if we broaden our view a little and include cognitive impairment avoidance in those with high homocysteine levels (via Dr. Briffa).

    Of course if we narrow our definition of healthy enough daily vitamin supplementation is going to have no effect at all.

  31. Jason says:

    One other thing I thought of regarding this whole vitamin debate is the pro vitamin folks who love to say “well adequate doesn’t mean optimal.” To me this seems like a fluff argument because no one who makes this statement ever seems to know what the optimal levels are. If you have enough to avoid deficiency, but not so much to suffer an overdose, how could that not be optimal?
    I am no expert by any means so is this adequate vs optimal levels even a constructive argument? To me it seems to lack substance.

    1. Andrey Pavlov says:

      @jason:

      I am no expert by any means so is this adequate vs optimal levels even a constructive argument? To me it seems to lack substance.

      Nope. You are pretty much spot on. It IS difficult to actually determine what is a “normal” level of just about anything and that is why there is some debate regarding VitD. However, after some time and lots of decent data we are confident enough to say that VitD deficiency is actually probably pretty prevalent and that supplementation to levels higher the 30ng/dL makes sense. I think it is still reasonable to say that in the absence of any complaints or specific risk factors for osteoporosis there is no impetus for doing so, but it does seem that low level of D can indeed induce states of fatigue, malaise, and generalized dysphoria and that supplementation seems to help. Now, I am fully open to further evidence demonstrating that early preventative supplementation in the absence of complaints decreases the incidence of osteoporosis but so far the data is not exactly convincing on that front.

      In pretty much all other cases, however, we have yet to identify anything new that a sub-clinical “hypovitaminosis” may lead to or what we could be preventing by supplementing to “optimal” levels. Physiologically that idea doesn’t even quite make sense since “optimal” would be whatever level is necessary to ensure that your cellular and metabolic processes proceed without problems. All the evidence so far tells us that MORE than that level does nothing.

      1. Jason says:

        “Physiologically that idea doesn’t even quite make sense since “optimal” would be whatever level is necessary to ensure that your cellular and metabolic processes proceed without problems. All the evidence so far tells us that MORE than that level does nothing.”
        That’s kind of what I figured but I have no training in anything medical

        1. Andrey Pavlov says:

          @jason:

          You don’t need medical training to apply critical thinking skills. Good on ya!

  32. Michael J Gonzalez says:

    Looks like a lot of people missed the lecture on vitamins in their biochemistry class.
    Please read this!

    http://ajcn.nutrition.org/content/75/4/616.full.pdf+html

    1. Scottynuke says:

      Somehow I doubt that paper will ever be included in a biochemistry lecture.

      Could it appear in an advanced genetics class? Perhaps, although the more learned SBMers are in a better position to note how preliminary that work is.

    2. Andrey Pavlov says:

      Dangit… meant that as a reply to this (apologies for the double post):

      How is this even remotely relevant to the conversation? This is a paper about very specific genetic diseases that occur because of mutations in the genes that lead to less functional enzymes/proteins that utilize vitamins as co-factors. The idea is that if the mutation is such that the vitamin co-factor cannot bind well to the mutated protein that the function of the protein will diminish and lead to disease. Thus, if you increase the concentration of the vitamin you will, by simple Michaelis-Menten kinetics, increase the stochastic binding of the vitamin to the binding domain of the protein and thus boost the functionality of the protein to ameliorate (but not cure) the disease in question. In other words, taking a busted lock and forcing a key into it to make it work a little bit better. They go a bit further to say that high dose vitamins may stimulate variant protein production which has a better binding affinity which is indeed interesting.

      But…. this has nothing to do with vitamin use in any population besides the extremely specific and extremely small populations listed in the paper. And has absolutely no bearing on people with otherwise normally functioning proteins.

  33. Andrey Pavlov says:

    How is this even remotely relevant to the conversation? This is a paper about very specific genetic diseases that occur because of mutations in the genes that lead to less functional enzymes/proteins that utilize vitamins as co-factors. The idea is that if the mutation is such that the vitamin co-factor cannot bind well to the mutated protein that the function of the protein will diminish and lead to disease. Thus, if you increase the concentration of the vitamin you will, by simple Michaelis-Menten kinetics, increase the stochastic binding of the vitamin to the binding domain of the protein and thus boost the functionality of the protein to ameliorate (but not cure) the disease in question. In other words, taking a busted lock and forcing a key into it to make it work a little bit better. They go a bit further to say that high dose vitamins may stimulate variant protein production which has a better binding affinity which is indeed interesting.

    But…. this has nothing to do with vitamin use in any population besides the extremely specific and extremely small populations listed in the paper. And has absolutely no bearing on people with otherwise normally functioning proteins.

Comments are closed.