Kevin T Keith, 4th place

Abstract

Dietary mineral supplementation is a perennial topic of intense and widespread concern. A vast literature on the topic exists. This literature suffers from severe limitations in comprehensiveness, relevance of subject populations (i.e., mostly non-human and often very small), and direct applicability to clinical practice. Existing dietary recommendations promulgated by expert bodies are often based on incomplete and weak empirical data, and in many cases merely estimated or extrapolated. However, improving the quality and predictive power of the existing data is not an analytical problem. The problem in identifying clear and reliable dietary mineral supplement recommendations has largely to do with the adequacy of the experimental data base, not with failures to analyze or interpret data which do exist. For a variety of reasons, some having to do with procedural bias in medical research (focus of research efforts, competing agendas between funders, researchers, clinicians, and patients, and the like), some having to with the inherent process of research in clinical medicine (animal studies precede human studies and thus dominate in the literature; pre-clinical research results may point to but do not prove clinical applicability; human subjects research is expensive and time-consuming), existing research on clinical benefits of dietary minerals and their supplemental intake is scattered an unsystematic, lacking in quantitative rigor and often lacking reliability or confidence estimates, and very often not directly applicable to the problem question posed by this contest. The recommendation priors which are available from authoritative scientific bodies (e.g., Institute of Medicine Dietary Reference Intakes) are thus subject to considerable, but unquantified, uncertainties. This is not the same as a lack of a sufficiently “rational” or insightful underlying analysis, however; in fact, the skimpiness of the empirical data ground makes such analysis impossible in the case of many potential mineral supplements, and of limited reliability in others. The developmental process underlying these priors – expert-level literature review and consensus-building – is sensible even though often not quantitatively rigorous. It is unlikely that any of the recommendation priors can be improved upon by further “rational analysis” other than perhaps by way of quantification of confidence limits in specified patient sub-populations. At the same time, the complexity of such a task, especially given a target survey list of over 20 known potentially relevant dietary minerals, makes its successful solution essentially impossible within the constraints of the contest problem. Further, the analytical tools recommended and preferred in the problem proposition – especially Bayesian statistics – are not directly applicable to the problem data. From this perspective, the contest problem better illustrates the limits of certain “rational methods” approaches to clinical decision-making than it does the weaknesses of existing approaches, specifically in the case of large-scale dietary intake recommendations, and possibly others. (Please see “Report Body” for further reflections on these difficulties.)

Recommendation List

No quantified recommendations were developed. See “Report Body” for excruciating confession of abject failure.

Report Body

To Whom It May Concern:

I have attempted and failed to produce a strong and defensible entry to your Quantified Health contest, and as the deadline approaches have finally concluded that my efforts are not sufficient. This was partly circumstantial – I only became aware of the contest very recently, and due to press of obligations had only 3 days to work on my entry; no doubt it is partly due to my inherent limitations as well. But I think it was partly structural – a consequence of the contest rules, the subject matter, and their relationship.

I admire the project your group is undertaking, and think it may produce real benefits. I think there is much to be said for the “rationalized” approach to medical decision-making you promote. But I also have the impression, if I may offer this observation, that this contest does not provide a relevant or useful test of that methodology, or an opportunity for entrants to employ those methods to best effect. Having put considerable effort into my incomplete entry, and gradually come to the above conclusion, I thought it might be of some use at least to pass my thoughts on to you.

Please understand that what I have to say is offered as feedback, or an external perspective, only. This is not an attempt to circumvent the contest rules, or excuse my failure to satisfy them. I have included my partial first-draft report below (under the double line of stars) for illustrative purposes, but in regard to the contest you may consider it a non-submission. I simply did not want that effort to go entirely to waste, and hoped that you might find it useful. You may make of it, and of this feedback, what you will. I hope you will forgive my presumption in offering it.

My reactions to the contest are threefold: (1) it is essentially impossible to satisfy the terms of the contest in a rigorous way; (2) a responsible and accurate answer to the question posed would (mostly) not require or employ the analytical framework the published “Judging and Scoring” criteria anticipate; (3) the “rational” methodology the contest promotes is not the most useful approach to answering the question posed, and, correspondingly, the question posed is not an effective testground of the rationalized methodology.

(1) “Important dietary minerals” is a vague term that requires clarification; any reasonable estimate, however, would identify close to a dozen dietary minerals that have significant physiological effects, many of which could be beneficial as dietary supplements for at least some patient populations. Just narrowing the field to that few target minerals requires consideration of a much larger list: the USDA provides individualized recommendations for dietary values of over 20 minerals on its interactive website; the Institute of Medicine of the US National Academies publishes a standard dietary reference work that has chapter-length analyses of 16 minerals, and shorter analysis of 5 more; the Linus Pauling Institute at Oregon State University has published detailed literature summaries on dietary implications of 14 different minerals.

Of course, as the contest website notes, it is useful to focus on just those minerals that are “important”. But it requires as much research to rule a mineral out as it does to include one, and, under the “show your work” requirement of the contest rules, almost as much space to provide the references and data to justify those distinctions. This raises two logistical problems: the research burden of the contest problem is unrealistic (PubMed commonly returns several thousand hits for a typical search on a given dietary mineral, using likely keywords), and there is insufficient space in the suggested 5,000-word report length to treat each target mineral in sufficient detail ( the Linus Pauling Institute’s writeup on dietary calcium alone runs over 6,500 words, and does not include quantified supplement recommendations; not all of that may be relevant to the contest problem, but the difficulty should be apparent).

In addition, the scope of the project is . . . daunting. With a target field of over 20 minerals, and an immense literature base for most of them, meeting the terms of the judging criteria becomes laughably impossible. Just the first two expectations – “How much of the important, high quality and relevant scientific literature was found and cited?” and “Was scientific literature that was potentially relevant but not useful identified and the reasons for it being not useful justified?” – imply an exhaustive literature review that would be unmanageable for even a single mineral. From there, providing dietary recommendations individualized to specific patient populations, with quantified benefit estimates and cost/benefit analysis, imposes an extensive further burden of literature meta-analysis and sensitivity testing, reiterated 10 – 20 times. Of course, each entrant can decide for themselves how much effort it is reasonable to them to put into the contest, but I would suggest that a focus on just one specified mineral would have been appropriate to the terms of the contest and the report length; as offered, the contest terms virtually guarantee an inadequate analysis of the entire list.

(2) The discussion on the contest website, and the published judging criteria, imply an application of the types of analytical methods promoted by the “Less Wrong” framework. In particular, the emphasis on “ill-framed or potentially misleading statistical framework(s)”, biases in the literature, and “benefits and drawbacks under uncertainty”, echo the emphasis on the Less Wrong website on statistical meta-analysis and Bayesian statistics, and the fundamental assumption there that scientific literature is often biased. However, the problem in answering the question posed in the contest is not a lack of rational analysis of information on the subject.

It is true that there are systematic biases in the literature – in particular, most of the literature for any dietary mineral involves animal models; experimental arms in human research often have absurdly small n and are sometimes uncontrolled; dietary implications are frequently indirect or speculative; and there is little research on beneficial supra-RDA supplement levels, as opposed to pathological excess or deficient levels – but there is certainly a great deal of literature available and it has been extensively analyzed. The several summary analyses and reference standards referenced above are the product of expert-level academic panels taking into consideration a range of literature, from a perspective of knowledge and experience, that dwarfs what would be possible for any single amateur working alone over a matter of days or weeks; these reference works also conform closely to similar publications from expert panels of national scientific bodies in other scientifically-advanced countries.

It is valuable and appropriate to remind oneself that experts can be biased and scientific consensus can be wrong, but to assume that experts are presumptively not just wrong but irrational about one of the most heavily-studied fields in medicine is, frankly, itself a form of bias. In short, the question about dietary supplements has not been definitively answered yet because it is an inherently difficult problem involving complex and synergistic mechanisms, and the problem is still in the information-gathering stage in most respects, not because the scientific community hasn’t been “rational” or clever enough. No application of “rational methods” by outsiders is going to resolve that question on the basis of existing information, nor are the outsiders going to discover answers that have not been already considered by the relevant experts. (I speak here of the dietary literature, but the point may be relevant in other cases as well. Before any analysis of the kind required by this contest is undertaken, it may be well to conduct a sanity check as to whether the particular problem stands to benefit from that approach; I suspect that the existing medical “received wisdom” is in many cases more robust than the “less wrong” perspective assumes.)

Most important, the contest website states that the best solutions will be posted “so everyone can benefit from the knowledge therein”. No doubt the usual disclaimers (“this is not medical advice”, “consult your physician”) will be deployed, but it is clearly anticipated that the recommendations made by the winning entrants will be made available to the public, implicitly endorsed as having greater analytical rigor than the existing expert recommendations. On the basis of the concerns expressed above, this raises grave questions of responsibility and safety.

(3) Bayesian analysis adds little to this question. Bayesian statistics are a method for estimating probabilities of events in light of known information regarding related events. The question of recommendations for dietary supplements is not probabilistic. Possibly it could be modeled along the lines of the likelihood of truth in a test result – the classic case of Bayesian analysis – but only if the dietary recommendations in question were based on tests with known probabilities of true/false positive/negative results. Instead, most dietary recommendations are based on simple correlations between observed pathology and dietary intake in subject populations, with values (e.g., “Recommended Dietary Intake” or “Tolerable Upper Limit”) set with reference to a cutoff population percentile or, worse, often simply extrapolated from lower values (“Adequate Intake” estimates); in most cases there is no reliability estimate for these percentile figures. Applying Bayesian techniques to such undefined estimates would be either impossible or, at best, an exercise in shoehorning rough or unquantified data into an inappropriate technical framework. (It might be concluded then that “more research is required”, but of course that’s always true.)

Non-Bayesian “rational” methods are similarly limited. As noted above, the biggest problem related to the contest subject is the adequacy of the existing data base, not a lack of sophistication in analyzing it. In most cases, a common-sense appreciation of the reliability or limitations of existing research is the most relevant tool for identifying best recommendations; sophisticated analysis is neither possible nor necessary given the data at hand.

It appears that the contest problem may have been chosen as being simply an interesting and commonly-asked medical question that appears to be quantitative (“what quantities of supplements should be taken?”), but is not answered by rigorous quantitative analysis. The contest rules appear to anticipate a more rigorous analysis than the data allow, leading either to an inability to meet the terms of the contest or a temptation to over-analyze very rough data. (The fact that the rules anticipate a lower report length of 1,000 words – again, on a target list of over 20 substances recognized in the literature as common dietary supplements – only underscores the possibility that there was a considerable mismatch between the anticipated contest results and the actual scope of the problem assigned. Also, the monetary cost of most mineral supplements works out to pennies per dose, making financial cost, at least, an almost negligible consideration in any case of significant benefit. The contest rules do not seem to recognize this, but do require the analysis anyway. Again, the contest rules seem in some sense generic, not tailored to the actual data ground of the problem in question.)

In my own case, as I worked through the material I found I was simply producing very limited and inadequate summaries of the existing literature on each mineral; there was little gist for rigorous analysis, and little resulting content beyond what already existed in that literature. The systematic “methodology” I laid out for approaching the problem even as a literature review turned out to be absurdly too ambitious for the scope of the task, and still did not admit of analytical rigor. Perhaps someone more skillful and insightful, and with more time to commit, could have produced a more sophisticated analysis. But, especially in light of the concerns expressed above, it seems likely that the best recommendations for dietary supplements in most cases are going to be the recommendations that have already been identified in the relevant research and expert meta-analysis. If that is the case, then there is little to be added through the methodology of this contest; at best, the entrants can hope to accurately identify what already exists.

In closing, let me note that my own efforts on this problem (below) are ample demonstration of the difficulties it poses. After days of effort, I found I had succeeded in profiling only about a third of the target list of minerals, and those only the ones that could be most easily eliminated. My profiles, as noted, were essentially just references to existing literature with little real analysis and only superficial reliability estimates at most. Yet even such brief summaries took up more space than was allowable under the contest terms, and were entirely inadequate as reliable and responsible recommendations for actual human health intervention.

As I said, this is not a letter of criticism; I think the perspective implicit in this contest is a valuable one, and its continued promotion stands to bring great benefits. It is also not a letter of complaint: you posed your terms, and I accepted but failed to meet them. However, as I have also said, and tried to explain, that fact may have been in some part a result of the clash between the implicit expectations and, if I may, biases motivating the contest, and the practical reality of making clinical recommendations on the basis of scientific research. I endorse and applaud the perception that much existing clinical wisdom is sorely flawed, but perhaps it is also the case that the scientists and experts who created it aren’t as dumb as you think. Both they and you, no doubt, want to do the best they can with the data available; it may be, however, and this contest may serve to demonstrate, that whatever limitations that effort may reveal cannot be resolved simply by being “less wrong”. My intention in writing was to offer that suggestion; I know I can rely on you to evaluate it at its worth.

Thanks very much for the chance to attempt this problem. It was fascinating and challenging, at the very least. Best of luck in the outcome, and with your ongoing project.

Sincerely,

Kevin T Keith

* * * * * * * * * * * * * * * * * *

This report attempts to answer the question: “What are the best recommendations for what quantities adults (ages 20-60) should take the important dietary minerals in, and what are the costs and benefits of various amounts?” It is assumed that “quantities adults . . . should take” refers to the supplementation of inherent dietary quantities by the use of mineral tablets, specialized diets, or the like, and not to the ordinary intake of minerals through a typical balanced diet.

The methodology used in this report includes: definition of the target substances (“important dietary minerals”); winnowing of the list to remove minerals well-characterized as not requiring dietary supplementation in the healthy adult human (either because of ready availability without supplementation, or because there is no evidence that supplementation is beneficial); review of the remaining potentially needed minerals with regard to reliable empirical evidence of their necessity and benefits in the target population; characterization of the benefits, risks, and costs (financial and other) of those minerals identified as potential supplements, in quantitative terms to the extent possible, with regard to the target population, both male and female; review of non-typical population cases (pregnancy, lactation, or extremes of age or health status within the group of adult humans aged 20 – 60); review of “alternative” (including commercial) literature for potentially reliable recommendations outside the scientific mainstream; and the generation of final recommendations in light of the above. Sources cited will be characterized as to reliability with regard to potential bias, methodology used and empirical confirmation of data, limits of experimental samples and data collection, and other relevant factors. Recommendations will be characterized as to reliability, applicability, and quantitative uncertainty.

With reference to methodological adequacy and comprehensiveness, the state of the literature on this subject must be considered. The published literature in mainstream scientific journals alone, on dietary minerals and supplementation, is vast [10]. Non-peer-reviewed material from commercial and “alternative” sources is essentially unlimited. Therefore it is impossible within the scope of this project to undertake a comprehensive literature review.

Instead, a “best available” standard is adopted to establish a baseline scientific consensus and determine the evidentiary basis for that consensus. This approach is made possible by the fact that nutrition guidelines are published regularly by respected national scientific bodies in the US, which have also undertaken extensive meta-analysis projects, based on the published scientific literature and conducted by panels of academic experts, aimed at precisely the question posed above [3], [6]. It is acknowledged that the dietary recommendations resulting from these reviews are themselves the subject of controversy, and that this project is intended to evaluate and extend the consensus which largely rests on that work. However, this report accepts the “Recommended Dietary Allowances” (RDA), “Adequate Intake” (AI) estimates, and “Tolerable Upper Limit” (UL) estimates for individual nutrients, published by the USDA and Institute of Medicine, as relevant analytical priors for comparative purposes.

As a framing consideration, it should be noted that the first approximated answer to the question regarding appropriate level of dietary supplement for the average healthy human, regardless of population sub-type, is: “None”. Although it is known that deficiencies of most of the recognized dietary nutrients may be pathological, it is only very recently in human history that reliable non-dietary (or diet-related, such as from natural salt formations) sources of mineral nutrients have become available. Human evolution has determined that these nutrients are necessary, but the human species has flourished over its evolutionary history with only naturally-obtainable sources of them. Even granting the considerable changes in human diet that have taken place since the rise of agriculture, it is evident that an ordinary diet with no mineral supplements is sufficient for life for the average person, and in most cases for good health. Therefore the question for this report can be assumed to be not whether mineral supplements are necessary for most healthy patients, but whether particular added benefits can be obtained by supplementing mineral intake above natural levels resulting from a typical balanced diet. Supplements for the purpose of maintaining ordinary good health will be assumed to be unnecessary except where need has been demonstrated on a reliable empirical basis (most typically in cases of illness, pregnancy, etc.).

For clarification, “dietary minerals” may be defined as dietary substances that are “elements that originate in the Earth and cannot be made by living organisms” [7]. This definition excludes vitamins and complex nutrients such as carbohydrates or proteins (which are complex organic molecules), as well as medicinal substances used to treat specific diagnosed diseases. The definition of “dietary minerals” is also commonly taken to exclude the ubiquitous chemical components of biological organic molecules, Carbon (C), Hydrogen (H), Oxygen (O), and Nitrogen (N).

Dietary minerals commonly identified as such in standard professional references include: Arsenic (As), Boron (B), Calcium (Ca), Chlorine (Cl), Chromium (Cr), Copper (Cu), Fluorine (F), Iodine (I), Iron (Fe), Magnesium (Mg), Manganese (Mn), Molybdenum (Mo), Nickel (Ni), Phosphorus (P), Potassium (K), Silicon (Si), Selenium (Se), Sulfur (S), Vanadium (V), and Zinc (Z) [3], [5].

Not all of these are proven to be biologically necessary in humans, and some are found in sufficient quantities in many ordinary foods and thus do not require dietary supplementation. The “important dietary minerals” may be taken to be those from among the list above which do not meet any of the following disqualifying criteria: (a) no benefits are known from elevated doses in any patient sub-population; (b) elevated doses are known to be dangerous in all patient sub-populations; (c) is commonly present due to inclusion in an ordinary diet but plays no role in human biology in any patient sub-population; (d) is commonly present due to inclusion in an ordinary diet but is negligible because observable benefits of supplementation are small compared to other safe means of obtaining such benefits, in all patient sub-populations. Minerals which fall under any of (a) – (d) above are not candidates for dietary supplementation and need not be considered in this report. Minerals which fall under any of (a) – (d) above for most patient sub-populations, with the exception of particular sub-groups or marginal cases (such as illness or metabolic disorder) will be considered only in light of the relevant sub-groups.

Dietary mineral levels are reported by the US Institute of Medicine in the form of “Dietary Reference Intakes” (DRI); the two most salient data points for each mineral are the “Recommended Dietary Allowance” (RDA), which is defined as the level determined on the basis of empirical research to meet the dietary requirements of 97-98% of a given healthy population sub-group, and the “Adequate Intake” level (AI), which is an estimated level “believed to cover the needs of all healthy individuals in the group” where “sufficient scientific evidence is not available to . . . calculate an RDA”. There is also the “Tolerable Upper Level” (UL), defined as “the highest level of daily nutrient intake that is likely to pose no risk”. [3]

The use of standardized reference values (RDA/AI) for estimating dietary nutrient sufficiency is subject to a variety of methodological weaknesses. Nutrient values of foodstuffs included in standard (e.g., USDA or Institute of Medicine) reference databases are often estimated indirectly, either calculated from similar or component forms of a given food, or on the basis of reported or standardized ingredient content of prepared foods. Standardized “nutrient retention” values are used to estimate the nutrient content of cooked foods, based on data obtained from raw foods. In some cases, values are simply assumed equal to those obtained from analysis of different plants of the same genus or general type (“leafy green”, “tropical”) [8]. Nutrient analysis of sample diets on the basis of ingredient nutrient values is purely additive (i.e., it is assumed no synergistic effects are present), and standardized reference values do not directly reflect the nutrient content of the actual foods a given person eats. Estimated nutrient values are generally reported without limits of precision. Recommended daily average or maximum intake levels for the standardized individual are uncorrected for age, sex, health status, or other factors, and are usually given as total absolute intake levels uncorrected for bodyweight (implicitly assuming every person has the same biochemistry, and is the same size, regardless of age, sex, or individual health or body morphology).

In addition, standard dietary mineral recommendations often exhibit a weak empirical or analytical basis. Recognized authorities admit weaknesses in the experimental record, and the use of non-empirical estimations (“available data [on certain nutrients] were often sparse or drawn from studies with significant limitations . . . . Thus, although governed by scientific rationales, informed judgments were often required in setting these reference values” [1]). Interpretation can be suspect (the USDA asserts that “potassium deficiency is characterized by hypokalemia” – a tautological statement [1]; also, the Institute of Medicine Dietary Reference Intake reports often claim “no clear biological function in humans” for minerals where literature searches reveal reported physiological roles at dietary levels, and observed effects from supplementation – for example, in the cases of Si and V). Thus, known priors are of questionable reliability, but quantitative estimates of confidence intervals are usually unreported.

A review of the historically recommended intake (RDA and DRI) of major minerals over the period 1968 – 2001 shows that, while some changed by only 0 – 10% of the original values (Ca, P, I, Mg, Zn) over that time period, some recommendations rose, fell, or fluctuated by larger percentages, and some historical recommendations were given in the form of ranges of several hundred percent, or in one case greater than an order of magnitude (the 1974 RDA for Molybdenum was “45 – 500mg”) [9]. This illustrates the imprecision inherent in these estimates, even derived from best-available empirical evidence.

Bearing in mind the limitations of the existing literature, and the (unquantified, but obvious) uncertainties in existing dietary need estimates (priors), review of the literature identifying relevant reported physiological effects (or lack thereof) of the previously-listed dietary minerals may be summarized as below.

Minerals to be excluded from consideration as “important” (meaning that there is no known benefit to be gained by supplementing the dietary intake of such minerals), by way of the criteria (a) – (d) above, include the following:

Arsenic is not an important dietary mineral in healthy human populations. It is famously poisonous, and, if there is a minimal necessary level, it appears commonly obtainable by way of trace As ingested through an ordinary diet. The US Institute of Medicine states “there is no justification for adding arsenic to food or supplements” [4]. However, artificially induced arsenic deficiency has been observed in animals, resulting in death apparently related to disruption of metabolism of the essential amino acid methionine [15]. Some research indicates that arsenic may be depleted in humans by medical treatments such as hemodialysis [14], or in conditions that place a high demand on methionine metabolism, such as pregnancy, lactation, or others [16]. This suggests that there may be some danger of extreme As deprivation in specific human populations, and thus a need for minimal As intake; however, this level has not been reliably characterized. Extrapolations from animal data indicate that it is at or below the level normally obtained through diet alone [16]. Conclusion: As supplementation cannot be recommended; at-risk patient populations should take care to maintain a normal balanced diet.

Fluorine is not an important dietary mineral in the meaning of this report. The literature on dietary fluorine is almost exclusively confined to its role in preventing tooth decay. This benefit is well-established, but is obtained in most cases passively through fluoridation of water supplies and toothpaste. Oral supplementation with fluoride compounds is recommended only in individual cases, usually for those who do not live in areas with fluoridated water [20]. Conclusion: fluoride supplementation is relevant only to the issue of tooth decay, and beneficial only in individual cases of special need, in which cases it functions as a dental treatment rather than a dietary supplement.

Nickel is not an important dietary mineral. Ni metabolism in humans is not well-characterized [3], and the literature focuses largely on Ni allergies or toxicity, which can be severe [31]. One study (n = 7) indicates that Vitamin A supplements (100 mg/d) may reduce bodily Ni levels [26]. However, there is no established RDA for Ni, and the established UL from all sources for all relevant population groups is 1mg/d, suggesting that the difference between normal dietary intake and safe supplemental levels, even in cases of partial depletion, is small [4]. Conclusion: Ni supplementation has no observed benefits and may lead to toxicity; Ni depletion is possible in certain cases but this does not warrant dietary supplementation.

Vanadium is not shown to be an important dietary mineral. Its role in a variety of biochemical pathways is well-documented, but professional opinion on its importance in humans “ferments a great deal of contradiction – from toxicity to essentiality”; there is inconclusive clinical data on the use of vanadium in the treatment of certain diseases [21]. Recent early-stage research (Phase II, n = 7) suggests some benefits in diabetes [22]. There is no identified RDA or UL in most population sub-groups [2]. Conclusion: V supplementation may bring clinical benefits in the treatment of certain diseases, but this is not yet proven, and its use in healthy patients has not been shown to provide benefits.

Minerals that cannot be ruled out as “important dietary minerals” under criteria (a) – (d) above are characterized below:

Boron may be an important dietary mineral, particularly with respect to specific population sub-groups. Some sources indicate that B is necessary for human nutrition and B deficiencies may arise from artificial diets (such as total parenteral nutrition – “intravenous feedings”), but even in those cases the necessary physiological levels may be obtained from ordinary foods [3], [27]. Recent research (n = 8) indicates that B supplements of 10mg/d uncorrected for body weight may increase production of steroidal hormones, including sex hormones, in men [29], and there is speculation that B supplements may therefore be performance-enhancing in male athletes [28]. Other research (n = 18) returns similar results, including an increase in estrogen production which is known to be protective in atherosclerosis [30]. One study (n = 7) indicates that Vitamin A supplements (100 mg/d) may reduce bodily B levels [26]. The estimated UL is 20mg/d from all sources, indicating that the above supplement levels are near the recommended maximum safe dose. Conclusion: B supplements are not shown to be beneficial in the typical adult human, but may offset B depletion resulting from other supplements, and at a level of 10mg/d (bioavailable B) do produce physiological changes that may have health benefits for sports performance or atherosclerosis in males; the latter effect has not been quantified as to magnitude or dose-responsiveness, nor demonstrated in women.

Calcium

Chlorine

Chromium

Copper is well-known to play important metabolic roles, but few health effects are observed from changes in Cu intake level [18]. The RDA for Cu (0.9mg/d in males, rising to 1.3mg/d in lactating females) is easily obtained from dietary sources, including less than 1 ounce of beef or 2 ounces of common nuts [6]. The UL for Cu is well above the RDA, but there is evidence that Cu may accumulate to toxic levels even at doses below UL (≤ 7.8mg/d) [17]. An extensive review of the peer-reviewed literature on dietary Copper revealed limited benefit from supplementation in the healthy population, although Cu supplementation may be used as a medical treatment for known disease conditions [19]. Conclusion: dietary supplementation with Cu is not known to provide benefits in the healthy population, although it may be safe as a precaution for those whose diets do not include Cu-rich foods; at levels of 6 – 8x RDA, Cu supplementation creates a risk of toxicity.

Iodine

Iron

Magnesium

Manganese

Molybdenum

Phosphorus

Potassium

Silicon has a limited role in dietary supplementation, specifically in regard to osteoporosis. The role of Si in bone formation has only recently been established [23]. Beneficial effects of Si supplementation in cases of osteoporosis [23], [24] and the bioabsorption of Aluminum (implicated in but not proven to cause Alzheimer Disease) [25] have been hypothesized but clinical results have not yet been demonstrated. There is no established RDA or UL for Silicon due to lack of data, and no AI has been estimated because the average response of humans to Si intake has not been calibrated [3]. Conclusion: Si supplements may plausibly provide protection from osteoporosis and possibly other conditions in at-risk populations, but clinical efficacy and dose-response in these roles is not proven; there is no evidence favoring routine Si supplementation in the healthy population.

Selenium

Sulfur

Sodium

Zinc

In addition to the above conclusions, based in most cases on small samples of healthy (often athletic) subjects, specific impacts on metabolism of certain dietary minerals can be identified in specific patient sub-groups, as discussed below.

Hemodialysis is believed to cause depletion of certain trace minerals (“microminerals”), with potential health consequences in kidney dialysis patients. The reason is that the chemical makeup of dialysis fluid is tightly controlled for major ionic macrominerals (Na, K, Cl), but not for trace elements. Serum levels of these minerals are also not commonly monitored in dialysis patients, thus the actual effect of dialysis in a given patient cannot be known, and current research has only pointed to the potential for a problem without characterizing it quantitatively [14]. Specifically, levels of Zn, Mn, and Se may be diminished in dialysis, as established by meta-analysis of at least 7 (Mn), to in excess of 30 (Zn, Se), published clinical studies (sample sizes unknown) [12]. Conclusion: supplementation of these minerals in hemodialysis patients may be beneficial but optimal supplement quantities cannot be established; careful testing of serum levels for trace elements may establish optimal supplement levels in individual patients. Alternatively, since the Recommended Dietary Allowance for Zn, Mg, and Se is far less than half the estimated Tolerable Upper Limit for those minerals in all relevant patient sub-populations [4], a daily supplement of Zn, Mn, and Se at the RDA level could not exceed the estimated safe upper limit and would suffice to replace more than the maximum possible depletion of those minerals through dialysis; therefore, this can be accepted as a safe, though possibly unnecessary, level of supplement for dialysis patients. (Note also that levels of some trace elements, such as Cr and V, may actually be increased by dialysis [12], but this problem cannot be corrected by dietary supplementation. Given these variations and uncertainties, serum mineral testing may be a useful precaution for dialysis patients in general [12], but that issue is outside the scope of this report.)

There is evidence to suggest that established recommendations for dietary mineral allowances in the elderly are not accurate (specifically, that the recommendations for Fe, Zn, Se, and Cu are adequate, but those for Mg and Cr may be too high, that for Ca too low, and that those for P, I, Mn, F, and Mo are based on inadequate evidence) [11]

In conclusion [. . . I got nothin’ . . .]

References

REFERENCE DATABASES

[1] Dietary Reference Intakes for Water, Potassium, Sodium, Chloride, and Sulfate, Panel on Dietary Reference Intakes for Electrolytes and Water, Standing Committee on Scientific Evaluation of Dietary Reference Intakes, Food and Nutrition Board, Institute of Medicine of the National Academies, The National Academies Press, Washington, DC, 2004

[2] Dietary Reference Intakes: Recommended Intakes for Individuals, National Academy of Sciences, Institute of Medicine, Food and Nutrition Board, http://www.iom.edu/Activities/Nutrition/SummaryDRIs/~/media/Files/Activity%20Files/Nutrition/DRIs/5_Summary%20Table%20Tables%201-4.pdf [1/13/2012]

[3] Dietary Reference Intakes: The Essential Guide to Nutrient Requirements, Jennifer J. Otten, Jennifer Pitzi Hellwig, and Linda D. Meyers (eds.), Institute of Medicine of the National Academies, The National Academies Press, Washington, DC, 2006

[4] Dietary Reference Intakes (DRIs): Elements, Food and Nutrition Board, Institute of Medicine of the National Academies, http://www.iom.edu/Home/Global/News%20Announcements/~/media/Files/Activity%20Files/Nutrition/DRIs/DRI_Elements.ashx [1/14/2012]

[5] USDA Interactive DRI Tool for Healthcare Professionals, Institute of Medicine of the National Academy of Science, http://fnic.nal.usda.gov/interactiveDRI/ [1/13/2012]

[6] U.S. Department of Agriculture, Agricultural Research Service 2011. USDA National Nutrient Database for Standard Reference, Release 24, http://www.ars.usda.gov/Services/docs.htm?docid=8964 [1/13/2012]

[7] Minerals, Linus Pauling Institute, Micronutrient Research for Optimum Health, http://lpi.oregonstate.edu/infocenter/minerals.html [1/15/2012]

METHODOLOGY

[8] Sally F. Schakel, I. Marilyn Buzzard, and Susan E. Gebhardt, Procedures for Estimating Nutrient Values for Food Composition Databases, J. Food Composition and Analysis, 10, 102-114 (1997)

[9] Historical Comparison of RDIs, RDAs, and DRIs, 1968 to Present for Minerals, Council for Responsible Nutrition, http://www.crnusa.org/about_recs4.html [1/13/2012]

[10] For example, a PubMed search on “calcium dietary supplement” returns over 2,800 primary clinical studies and meta-analytical reviews; similar searches on other minerals in many cases return 400 – 600 citations, and these searches are likely not exhaustive of the relevant literature.

PATHOLOGY

[11] Wood RJ, Suter PM, Russell RM, Mineral requirements of elderly people, Am J Clin Nutr. 1995 Sep;62(3):493-505.

[12] Tonelli M, Wiebe N, Hemmelgarn B, Klarenbach S, Field C, Manns B, Thadhani R, Gill J; Trace elements in hemodialysis patients: a systematic review and meta-analysis, Alberta Kidney Disease Network, BMC Med. 2009 May 19;7:25.

[13] Rucker D, Thadhani R, Tonelli M., Trace element status in hemodialysis patients, Semin Dial, 2010 Jul-Aug;23(4):389-95. Epub 2010 Jun 14.

ARSENIC

[14] Mayer DR, Kosmus W, Pogglitsch H, Mayer D, Beyer W, Essential trace elements in humans: serum arsenic concentrations in hemodialysis patients in comparison to healthy controls, Biol Trace Elem Res., 1993 Apr; 37(1):27-38

[15] Eric O. Uthus, Arsenic essentiality and factors affecting its importance, in Arsenic Exposure and Health, Willard R. Chappell, Charles O. Abernathy, and C. Richard Cothern (eds.), Science and Technology Letters Monograph Series, 1994

[16] Uthus, E. O., Evidence for arsenic essentiality, Environmental Geochemistry and Health, Volume 14, Number 2, 55-58, DOI: 10.1007/BF01783629

COPPER

[17] Turnlund JR, Keyes WR, Kim SK, Domek JM. Long-term high copper intake: effects on copper absorption, retention, and homeostasis in men. Am J Clin Nutr. 2005;81(4):822-828.

[18] Araya M, Olivares M, Pizarro F, González M, Speisky H, Uauy R., Copper exposure and potential biomarkers of copper metabolism, Biometals. 2003 Mar;16(1):199-204

[19] Copper, Jane Higdon and Victoria J. Drake, Linus Pauling Institute, Micronutrient Research for Optimum Health, 2007, http://lpi.oregonstate.edu/infocenter/minerals/copper/ [1/14/2012]

FLUORINE

[20] Sampaio FC, Levy SM, Systemic fluoride, Monogr Oral Sci. 2011;22:133-45. Epub 2011 Jun 23.

VANADIUM

[21] Mukherjee B, Patra B, Mahapatra S, Banerjee P, Tiwari A, Chatterjee M, Vanadium--an element of atypical biological significance, Toxicol Lett. 2004 Apr 21;150(2):135-43.

[22] Thompson KH, Lichter J, LeBel C, Scaife MC, McNeill JH, Orvig C, Vanadium treatment of type 2 diabetes: a view to the future, J Inorg Biochem. 2009 Apr;103(4):554-8. Epub 2008 Dec 24.

SILICON

[23] Schröder HC, Wiens M, Wang X, Schloßmacher U, Müller WE, Biosilica-Based Strategies for Treatment of Osteoporosis and Other Bone Diseases, Prog Mol Subcell Biol. 2011;52:283-312.

[24] Li Z, Karp H, Zerlin A, Lee TY, Carpenter C, Heber D, Absorption of silicon from artesian aquifer water and its impact on bone health in postmenopausal women: a 12 week pilot study, Nutr J. 2010 Oct 14;9:44.

[25] Domingo JL, Gómez M, Colomina MT, Oral silicon supplementation: an effective therapy for preventing oral aluminum absorption and retention in mammals, Nutr Rev. 2011 Jan;69(1):41-51. doi: 10.1111/j.1753-4887.2010.00360.x. Review.

BORON

[26] Patlar S, Boyali E, Baltaci AK, Mogulkoc R, The effect of vitamin A supplementation on various elements in elite taekwondo players, Biol Trace Elem Res. 2011 Mar;139(3):296-300. Epub 2010 Mar 18.

[27] Nielsen FH, Is boron nutritionally relevant?, Nutr Rev. 2008 Apr;66(4):183-91.

[28] Naghii MR., The significance of dietary boron, with particular reference to athletes, Nutr Health. 1999;13(1):31-7.

[29] Naghii MR, Mofid M, Asgari AR, Hedayati M, Daneshpour MS, Comparative effects of daily and weekly boron supplementation on plasma steroid hormones and proinflammatory cytokines, J Trace Elem Med Biol. 2011 Jan;25(1):54-8. Epub 2010 Dec 3

[30] Naghii MR, Samman S., The effect of boron supplementation on its urinary excretion and selected cardiovascular risk factors in healthy male subjects, Biol Trace Elem Res. 1997 Mar;56(3):273-86.

NICKEL

[31] Aleksandra Duda-Chodak, Urszula B³aszczyk, The impact of nickel on human health, J. Elementol. 2008, 13(4): 685-696