On 2015 Mar 10, Oliver Gillie commented:

Dr Autier has misunderstood my affairs which are explained in a Corrigendum to the scientific journal, Public Health Nutrition (see reference below). I have always reported conflict of interest accurately. The editor of PHN published the Corrigendum to clarify the situation while agreeing that I had no financial conflict of interest to declare. Dr Autier raises the issue once more but I have already explained my position in the Corrigendum and am surprised he has not seen it.

I worked with the Vitamin D Company Ltd helping them to sell their 5,000 IU tablet in the UK at a time when high dose tablets were difficult to get in the UK. I felt that availability of the 5000 tablet could make an important contribution to our national health. However I never received any sales commission or any money at all from the Vitamin D Company and so have no financial conflict of interest to declare. In fact my 11 year campaign to inform the public about the facts of vitamin D and sunshine has cost me several thousand pounds in publishing and distributing, free of charge, reports on vitamin D (please see below for details). These reports have done much to publicise scientific and health issues surrounding sunshine and vitamin D which have been widely misunderstood by influential bodies such as Cancer Research UK. Over these 11 years I have made no earnings other than from occasional newspaper and magazine articles.

Dr Autier is a staff member of the International Prevention Research Institute (IPRI), a body which is funded by industry and undertakes custom research for industry. This financial security allows Dr Autier the luxury of adopting a fallacious statistical argument, which may be welcomed by the pharmaceutical industry, but does them no credit because it is poorly thought out scientifically and poorly argued. Autier et al have presented lengthy statistical results which incorporate this scientific error, an error well understood by most experimentalists yet denied here by Autier. His response below tells us what we know already but fails to address the scientific issue raised by my article in Public Health Nutrition (ref below).

REFERENCES:

The CORRIGENDUM in PHE:

http://journals.cambridge.org/download.php?file=/PHN/S1368980015000130a.pdf&code=a0e2984cd8793a37d36eebf438633b36

ARTICLE ON AUTIER et al’s ERROR:

“Controlled trials of vitamin D, causality and type 2 statistical error.” Public Health Nutrition 2014

TWO REPORTS on vitamin D written by Gillie (summaries below) are available free from: www.healthresearchforum.org.uk :

“SUNLIGHT ROBBERY - Health benefits of sunlight are denied by current public health policy in the UK.” (2004), drew attention to the health benefits of sunbathing and vitamin D and inadequate public understanding of the issues. Alternative advice for the public on safe sun exposure, given in the book, differs in important ways from advice given by Cancer Research-UK, which has encouraged sun avoidance so risking serious vitamin D insufficiency and increasing overall risks of cancer. Sir Richard Doll commented on the book: “I am most impressed with the way Gillie has collected and presented the evidence”.

“SCOTLAND'S HEALTH DEFICIT - an explanation and a plan (2008)”, has drawn attention to vitamin D insufficiency in Scotland and the link with chronic disease. The central hypothesis of the book is that a substantial and significant portion of the excess mortality of Scotland compared with England is caused by vitamin D insufficiency. This excess mortality, known as “the Scottish effect”, cannot be explained by smoking, alcohol, diet or poverty. A subsidiary hypothesis is that major chronic diseases that are caused at least in part by D insufficiency (e.g. cancer, heart disease, MS etc) have a higher incidence in Scotland than in England. Evidence for these ideas is examined in the book and general support is found for them. Sir Muir Gray commented: “I was very impressed by the strength of evidence and by the conclusion…The work is of importance and a great achievement.”

On 2015 Mar 09, William Grant commented:

Comments on why vitamin D observational and intervention studies disagree

In the paper by Autier et al. [1], it was stated that "The discrepancy between observational and intervention studies suggests that low 25(OH)D is a marker of ill health. Inflammatory processes involved in disease occurrence and clinical course would reduce 25(OH)D, which would explain why low vitamin D status is reported in a wide range of disorders." In my comment I will argue that the primary reason for the discrepancy between observational and intervention studies lies in the design of the intervention studies and that there is evidence that vitamin D3 reduces biomarkers of inflammation.

If intervention trials are intended to evaluate observational studies, i.e., determine whether vitamin D has a beneficial role in health, they should start with an understanding of the 25-hydroxyvitamin D [25(OH)D] concentration-health outcome relation [2]. They should then seek to enroll people in the studies with sufficiently low 25(OH)D concentrations that the vitamin D dose used in the study will significantly change the position along the relation [3]. If this is not done, the intervention study then becomes an exercise in determining whether supplementing the average population with a given amount of vitamin D has any impact on health outcomes, which is a different question.

A typical 25(OH)D concentration-health outcome relation was recently published for breast cancer incidence from case-control studies [4]. The power law fit to the data from 11 studies goes from an odds ratio of 1.77 at 15 nmol/L, to 1.04 at 30 nmol/L, 0.70 at 50 nmol/L, 0.52 at 75 nmol/L, and 0.41 at 100 nmol/L. The 25(OH)D concentration-PTH relation has a similar form, extending to 185 nmol/L [5]. Thus, there is a large change in odds ratio for a small change in 25(OH)D for low 25(OH)D concentrations, with the change decreasing with increasing 25(OH)D concentration. In addition, 25(OH)D concentration changes in a nonlinear fashion with respect to baseline 25(OH)D concentration for a given vitamin D3 oral intake [6], so it takes higher vitamin D3 doses to have an impact at higher baseline concentrations.

Some examples from the journal literature are instructive. A vitamin D3 trial regarding respiratory infections in adults conducted in New Zealand had baseline 25(OH)D concentration of 73 nmol/L and achieved 25(OH)D concentration above 120 nmol/L but found no reduction [7]. In contrast, a related study in Mongolia involving school children with baseline 25(OH)D concentrations of 18 nmol/L given 300 IU/d vitamin D3 found a 50% reduction in acute respiratory infections [8].

It is noted that the history of the understanding of the health benefits of vitamin A went through some of the same concerns as the understanding of those for vitamin D is today. About half of the intervention studies in the 1930s found beneficial effects in reducing infections [9]. In that paper "As suggested by Hess et al. (1933), no therapeutic benefit of vitamin A was noted for infants who were already sufficient in vitamin A." [10]. Now vitamin A supplementation prevents blindness and saves many lives from death due to infections in developing countries [11].

Regarding inflammation, a paper was recently published in response to Ref. 1 in which all vitamin D randomized controlled trials found were examined for beneficial effects on biomarkers of inflammation along with baseline and achieved 25(OH)D concentration [12]. For those trials using vitamin D3 with baseline 25(OH)D concentration between 17 and 47 nmol/L, 49% found a beneficial effect. For those trials with baseline 25(OH)D concentration between 49 and 77 nmol/L, only 27% did. In addition, 55% those with achieved 25(OH)D concentrations between 35 and 82 nmol/L had a beneficial finding contrasted with 29% for those with achieved concentrations between 83 and 187 nmol/L. Thus, baseline 25(OH)D concentration was much more important than achieved 25(OH)D concentration.

Well-designed vitamin D3 intervention studies should confirm many of the findings from observational studies. However, for now, observational studies remain the best guide to optimal 25(OH)D concentrations [13].

References 1. Autier P, Boniol M, Pizot C, Mullie, P. Vitamin D status and ill health: a systematic review. Lancet Diabetes & Endocrinology. 2014;2(1):76-89. 2. Grant WB. Using findings from observational studies to guide vitamin D randomized controlled trials. J Intern Med. 2015;277(1):83-6. 3. Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. Nutr Rev. 2014;72(1):48-54. 4. Grant WB. 25-Hydroxyvitamin D and breast cancer, colorectal cancer, and colorectal adenomas: case–control versus nested case–control studies, Anticancer Res. 2015;35(2):1153-60. 5. Valcour A, Blocki F, Hawkins DM, Rao SD. Effects of age and serum 25-OH-vitamin D on serum parathyroid hormone levels. J Clin Endocrinol Metab. 2012;97(11):3989-95. 6. Garland CF, French CB, Baggerly LL, Heaney RP. Vitamin D supplement doses and serum 25-hydroxyvitamin D in the range associated with cancer prevention. Anticancer Res 2011;31:617-22. 7. Murdoch DR, Slow S, Chambers ST, Jennings LC, Stewart AW, Priest PC, Florkowski CM, Livesey JH, Camargo CA, Scragg R. Effect of vitamin D3 supplementation on upper respiratory tract infections in healthy adults: the VIDARIS randomized controlled trial. JAMA. 2012;308(13):1333-9. 8. Camargo CA Jr, Ganmaa D, Frazier AL, Kirchberg FF, Stuart JJ, Kleinman K, Sumberzul N, Rich-Edwards JW. Randomized trial of vitamin D supplementation and risk of acute respiratory infection in Mongolia. Pediatrics. 2012;130(3):e561-7. 9. Semba RD. Vitamin A as "anti-infective" therapy, 1920-1940. J Nutr. 1999;129(4):783-91 10. Hess AF, Lewis, JM, Barenberg LH. Does our dietary require vitamin A supplement? J. Am. Med. Assoc. 1933;101:657–63. 11. Sommer A. Preventing blindness and saving lives: the centenary of vitamin A. JAMA Ophthalmol. 2014;132(1):115-7. 12. Cannell JJ, Grant WB, Holick MF. Vitamin D and inflammation. Dermatoendocrinol. 2014;6(1): e983401-1-10. Published online Jan. 29, 2015 http://www.tandfonline.com/doi/pdf/10.4161/19381980.2014.983401 13. Grant WB, Wimalawansa SJ, Holick MF, Cannell JJ, Pludowski P, Lappe JM, Pittaway M, May P. Emphasizing the health benefits of vitamin D for those with neurodevelopmental disorders and intellectual disabilities. Nutrients. 2015;7:1538-64.

Disclosure My organization, Sunlight, Nutrition and Health Research Center, receives funding from Bio-Tech Pharmacal, Inc. (Fayetteville, AR) and the Vitamin D Council (San Luis Obispo, CA). In the past it also received funding from), the UV Foundation (McLean, VA), the Vitamin D Society (Woodstock, Ontario), the Sunlight Research Forum (Veldhoven), and MediSun Technology (Highland Park, IL.

On 2015 Mar 09, Philippe Autier commented:

In our systematic review, we concluded that the association between low vitamin D status (as assessed by serum 25(OH)D concentration) and ill health found in most observational studies is not causal and could just reflect the influence of inflammatory processes present in many diseases (1). Our conclusion was essentially based on the finding that in the vast majority of randomized trials, vitamin D supplementation had no impact on health outcomes. O Gillie refutes our conclusions on the basis that type 2 errors could underlie the absence of trial result compatible with health benefit of vitamin D supplementation. This refutation is wrong because it equates to the common mistake of post-hoc power calculation (2). A type 2 error consists in rejecting the alternative hypothesis (i.e., vitamin D supplementation can prevent or cure diseases) and accepting the null hypothesis (i.e., vitamin D supplementation cannot prevent or cure diseases) when ultimately, the alternative hypothesis would turn out to be true. The avoidance of type 2 errors in randomized trials largely governs the size of trials, because the more subjects are included in a trial, the lower the probability of a type 2 error. This concept is termed the “power” of a trial that is the capacity of a trial to demonstrate the likelihood of the alternative hypothesis. The power is computed prior to randomization as the probability that a trial of a given size could confirm the possibility of a truly existing association between the intervention and the outcome. But once a trial has been conducted, the interpretation of results should be solely driven by the probability of type 1 error to avoid fallacious interpretation based on post-hoc power calculation.

For example, observational studies on cardiovascular diseases have documented risk reductions of about 40% for high versus low quantile of serum 25(OH)D concentration. At inception of their meta-analysis of randomized trials on vitamin D supplements and the risk of cardiovascular death, given the numbers of subjects and of cardiovascular events reported by trials, Bjelakovic et al. (3) had a 96.1% power to detect a 20% reduction in the risk of cardiovascular death (bilateral test with an alpha risk of 5%). By usual standards, a 96.1% power is huge. Once the meta-analysis was done, the observed relative risk was of 1.02 (95%CI 0.91-1.13). The probability that a significant real risk reduction of 20% could have been missed by the meta-analysis is a type 1 error whose probability can be estimated as 1 out of 58548.

In our review, many randomized trials on vitamin D supplementation and their meta-analyses were sufficiently powered for identifying risk reductions of 20% or more. But the flat results of most trials indicate that the probability of type 1 error for missing risk reductions of 20% is negligible. In lay terms, this can be referred as an extremely small chance to have missed a true relationship between vitamin D supplementation and an outcome.

O Gillie insists much on vitamin D status in children. Probably that because growing evidence reveals the unsoundness of the hypothesis of an “epidemic of vitamin D deficiency” in adults, “vitamin D deficiency” in children is taken as the substitute culprit for ill health during adult life. However, in many countries where bright sunshine is not habitual, vitamin D supplementation for small children and pregnant women is common practice since 50 years or more. If occasional cases of rickets occur in latitudes remote from the equator, these cases concentrate in children with naturally dark skin. It is therefore highly unlikely that a large proportion of ill health in adults would be due to rampant vitamin D deficiency in children.

O Gillie alludes to physiological mechanisms by which a lack of vitamin D would harm individuals, including cardiac remodeling and auto-immune diseases. But the scientific literature is replete with speculations on biological mechanism evoked by believers for promoting the putative healthy virtues of antioxidants, vitamins, and many other micro or macro nutrients. Unfortunately, ugly facts, often under the form of results of randomized trials (e.g., ref. 4 and 5), practically always demonstrate the inanity of these speculations.<br> We were surprised that O Gillie declared to have no conflict of interest (http://www.vitdco.com/pages/about-us).

Philippe Autier MD, Mathieu Boniol PhD, Patrick Mullie PhD

1. Autier P, Boniol M, Pizot C, Mullie P. Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol 2014;2(1):76-89.
2. Hoenig, John M. and Heisey, Dennis M. (2001), The Abuse of Power: The Pervasive Fallacy of Power Calculations for Data Analysis, The American Statistician, 55, 19-24.
3. Bjelakovic G, Gluud LL, Nikolova D, Whitfield K, Wetterslev J, Simonetti RG, Bjelakovic M, Gluud C. Vitamin D supplementation for prevention of mortality in adults. Cochrane Database Syst Rev 2011; (7):CD007470.
4. Miller ER 3rd, Pastor-Barriuso R, Dalal D, Riemersma RA, Appel LJ, Guallar E. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med. 2005 Jan 4;142(1):37-46.
5. Lawlor DA, Davey Smith G, Kundu D, Bruckdorfer KR, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet. 2004 May 22;363(9422):1724-7.

On 2015 Feb 25, Oliver Gillie commented:

There is a serious flaw in the scientific reasoning of Autier et al 1,2. A negative result in a clinical trial of a vitamin in adult disease does not prove that the vitamin cannot have caused the disease. A vitamin deficiency may have occurred earlier in life causing irreversible metabolic damage.

Title: Clinical trials, causation and type 2 error. By Oliver Gillie PhD Health Research Forum, 68 Whitehall Park, London N19 3TN Tel: ++44 20 7561 9677

I have pointed out Autier’s mistake which may be classified as a “type 2 statistical error” 2. Regrettably The Lancet (diabetes and endocrinology) has refused to publish a correction even though the error may be seen as one of editing as much as of original research.

Autier et al point out that raising the blood level of 25(OH)D (the standard reference measure of vitamin D) with a vitamin D supplement in clinical trials has not generally been found to modify the occurrence or clinical course of diseases associated with low D in observational studies. And they conclude: “Hence, associations between 25(OH)D and health disorders reported by investigators of observational studies are not causal.” 1

Autier is even more forthright in a Lancet press release where he is quoted as saying: "If the health benefits of high vitamin D concentrations shown by data from observational studies are not reproduced in randomised trials (the gold standard method for assessing a causal relation between an exposure and an outcome) then the relation between vitamin D status and disorders are probably the result of confounding or physiological events involved in these disorders… What this discrepancy suggests is that decreases in vitamin D levels are a marker of deteriorating health." 3

However, Autier et al have misunderstood the literature on trials and causation. They reference Byar et al.4 as referring to RCTs as a “gold standard” in establishing causation. However Byar et al. do not use the term “gold standard” and only consider causation very briefly. Causation may be proved in a clinical trial when supplementation succeeds in correcting a defect, but not when it fails to do so. A null result may be obtained simply because the trial took place too long after an irreversible insult occurring at a much earlier time.

A “type 2 error” mistakenly accepts the null hypothesis, when an alternative hypothesis is or could be the true state of nature. That is: Autier et al find no significant difference between treated and untreated subjects in adult trials of vitamin D supplements and use this finding to support a null hypothesis. In doing so they wrongly rule out an alternative hypothesis that treatment at an earlier stage, e.g. during growth and development, might be effective. Much more extensive trials in all age groups and in pregnancy are required before a null conclusion could be safely reached.

Rickets is the classical example of a disease which may be cured in early life but not in adulthood.5 It causes alteration of normal bone formation and deformation of limbs which may be corrected in childhood by supplementation with vitamin D. If however the deformations, whether gross or minor, are not corrected by vitamin D while the bones are growing, the bones become set in a pathological form that cannot be corrected by later supplementation.

Cardiac structure and some cardiac diseases such as hypertension may be associated causally with low 25(OH)D levels. The Baltimore Longitudinal Study of Aging has found that 25(OH)D levels are positively correlated with left ventricle wall thickness and there is a relationship between 25(OH)D and left-ventricle concentric remodelling.6 Hypertension in this study was also linked to left ventricle hypertrophy and low 25(OH)D. Experiments with young rats show that deprivation of vitamin D causes specific cardiac abnormalities similar to those found in the Baltimore study: cardiac hypertrophy, left-chamber alterations and systolic dysfunction, which follow on from cardiac inflammation, fibrosis and apoptosis.6 This strongly suggests that the association of low 25(OH)D with cardiac pathology is the result of lifelong low vitamin D levels which caused heart abnormalities during early growth. These observations can reasonably be regarded as proof that heart anatomy, and diseases arising from pathological changes in heart anatomy, can be caused by early vitamin D deficiency.

However, if we follow the reasoning of Autier et al. failure of vitamin D to induce beneficial heart remodeling in adults with abnormal heart anatomy would lead us to misleading conclusions i.e. that low vitamin D associated with abnormal heart anatomy in the Baltimore study is the result of reverse causation.

Autoimmune and other diseases are associated in significantly elevated rates with hospital admission for vitamin D deficiency, osteomalacia and rickets.8 Failure in determination of the immune system early in life by escape of T cells from thymic deletion could explain this common association of low vitamin D with autoimmune disease. Vitamin D is well known to be involved in immune processes.9

Many of the molecular mechanisms involving vitamin D are well known and provide a substantial basis for further exploration with clinical trials.9 However clinical trials of vitamin D must be conducted bearing in mind the additional difficulties presented when testing a nutrient rather than a drug.10 It is very difficult and enormously expensive to undertake clinical trials which continue for 20 or 30 years. Indeed it may not be possible. Other approaches are therefore required such as the animal experiments which complement the Baltimore study.

Further examples showing how vitamin D deficiency may cause disease at various stages of life are given in an article2 by the author together with examples of the serious harms which may ensue from literal or dogmatic adherence to policy advice provided by Autier and colleagues. Conflicts of interest: I have no conflicts of interest.

References 1. Autier PBM, Boniol M, Pizot C & Mullie P (2013) Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol 2, 76–89. 2. Gillie O. Controlled trials of vitamin D, causality and type 2 statistical error. Public Health Nutrition 2014. 3. Anon. (2013) Further doubt cast on benefit of vitamin D supplementation for disease prevention. Press release promoting Autier et al. (2013); available at : http://www.eurekalert.org/pub_releases/2013-12/l-fdc120313.php 4. Byar DP, Simon RM, Friedewald WT et al. (1976) Randomized clinical trials. Perspectives on some recent ideas. N Engl J Med 95, 74–80. 5. Holick MF (2006) Resurrection of vitamin D deficiency and rickets. J Clin Invest 116, 2062–2072. 6. Ameri P, Canepa M, Milaneschi Y et al. (2013) Relationship between vitamin D status and left ventricular geometry in a healthy population: results from the Baltimore Longitudinal Study of Aging. J Intern Med 273, 253–262. 7. Assalin HB, Rafacho BP, dos Santos PP et al. (2013) Impact of the length of vitamin D deficiency on cardiac remodeling. Circ Heart Fail 6, 809–816. 8. Ramagopalan SV, Maugeri NJ, Handunnetthi L et al. (2013) Hospital admissions for vitamin D related conditions and subsequent immune-mediated disease: record-linkage studies. BMC Med 11, 171. 9. Pludowski P, Holick MF, Pilz S et al. Vitamin D effects on mitochondrial health, immunity, autoimmunity, cardiovascular disease, cancer, fertility, pregnancy, dementia and mortality – a review of recent evidence. Autoimmune Rev 2013; 12: 976-89 10. Grant WB. Using findings from observational studies to guide vitamin D randomized controlled trials. J Int Med. Published online: 5 May 2014 DOI: 10.1111/joim.12260

On 2015 Feb 25, Oliver Gillie commented:

There is a serious flaw in the scientific reasoning of Autier et al 1,2. A negative result in a clinical trial of a vitamin in adult disease does not prove that the vitamin cannot have caused the disease. A vitamin deficiency may have occurred earlier in life causing irreversible metabolic damage.

Title: Clinical trials, causation and type 2 error. By Oliver Gillie PhD Health Research Forum, 68 Whitehall Park, London N19 3TN Tel: ++44 20 7561 9677

I have pointed out Autier’s mistake which may be classified as a “type 2 statistical error” 2. Regrettably The Lancet (diabetes and endocrinology) has refused to publish a correction even though the error may be seen as one of editing as much as of original research.

Autier et al point out that raising the blood level of 25(OH)D (the standard reference measure of vitamin D) with a vitamin D supplement in clinical trials has not generally been found to modify the occurrence or clinical course of diseases associated with low D in observational studies. And they conclude: “Hence, associations between 25(OH)D and health disorders reported by investigators of observational studies are not causal.” 1

Autier is even more forthright in a Lancet press release where he is quoted as saying: "If the health benefits of high vitamin D concentrations shown by data from observational studies are not reproduced in randomised trials (the gold standard method for assessing a causal relation between an exposure and an outcome) then the relation between vitamin D status and disorders are probably the result of confounding or physiological events involved in these disorders… What this discrepancy suggests is that decreases in vitamin D levels are a marker of deteriorating health." 3

However, Autier et al have misunderstood the literature on trials and causation. They reference Byar et al.4 as referring to RCTs as a “gold standard” in establishing causation. However Byar et al. do not use the term “gold standard” and only consider causation very briefly. Causation may be proved in a clinical trial when supplementation succeeds in correcting a defect, but not when it fails to do so. A null result may be obtained simply because the trial took place too long after an irreversible insult occurring at a much earlier time.

A “type 2 error” mistakenly accepts the null hypothesis, when an alternative hypothesis is or could be the true state of nature. That is: Autier et al find no significant difference between treated and untreated subjects in adult trials of vitamin D supplements and use this finding to support a null hypothesis. In doing so they wrongly rule out an alternative hypothesis that treatment at an earlier stage, e.g. during growth and development, might be effective. Much more extensive trials in all age groups and in pregnancy are required before a null conclusion could be safely reached.

Rickets is the classical example of a disease which may be cured in early life but not in adulthood.5 It causes alteration of normal bone formation and deformation of limbs which may be corrected in childhood by supplementation with vitamin D. If however the deformations, whether gross or minor, are not corrected by vitamin D while the bones are growing, the bones become set in a pathological form that cannot be corrected by later supplementation.

Cardiac structure and some cardiac diseases such as hypertension may be associated causally with low 25(OH)D levels. The Baltimore Longitudinal Study of Aging has found that 25(OH)D levels are positively correlated with left ventricle wall thickness and there is a relationship between 25(OH)D and left-ventricle concentric remodelling.6 Hypertension in this study was also linked to left ventricle hypertrophy and low 25(OH)D. Experiments with young rats show that deprivation of vitamin D causes specific cardiac abnormalities similar to those found in the Baltimore study: cardiac hypertrophy, left-chamber alterations and systolic dysfunction, which follow on from cardiac inflammation, fibrosis and apoptosis.6 This strongly suggests that the association of low 25(OH)D with cardiac pathology is the result of lifelong low vitamin D levels which caused heart abnormalities during early growth. These observations can reasonably be regarded as proof that heart anatomy, and diseases arising from pathological changes in heart anatomy, can be caused by early vitamin D deficiency.

However, if we follow the reasoning of Autier et al. failure of vitamin D to induce beneficial heart remodeling in adults with abnormal heart anatomy would lead us to misleading conclusions i.e. that low vitamin D associated with abnormal heart anatomy in the Baltimore study is the result of reverse causation.

Autoimmune and other diseases are associated in significantly elevated rates with hospital admission for vitamin D deficiency, osteomalacia and rickets.8 Failure in determination of the immune system early in life by escape of T cells from thymic deletion could explain this common association of low vitamin D with autoimmune disease. Vitamin D is well known to be involved in immune processes.9

Many of the molecular mechanisms involving vitamin D are well known and provide a substantial basis for further exploration with clinical trials.9 However clinical trials of vitamin D must be conducted bearing in mind the additional difficulties presented when testing a nutrient rather than a drug.10 It is very difficult and enormously expensive to undertake clinical trials which continue for 20 or 30 years. Indeed it may not be possible. Other approaches are therefore required such as the animal experiments which complement the Baltimore study.

Further examples showing how vitamin D deficiency may cause disease at various stages of life are given in an article2 by the author together with examples of the serious harms which may ensue from literal or dogmatic adherence to policy advice provided by Autier and colleagues. Conflicts of interest: I have no conflicts of interest.

References 1. Autier PBM, Boniol M, Pizot C & Mullie P (2013) Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol 2, 76–89. 2. Gillie O. Controlled trials of vitamin D, causality and type 2 statistical error. Public Health Nutrition 2014. 3. Anon. (2013) Further doubt cast on benefit of vitamin D supplementation for disease prevention. Press release promoting Autier et al. (2013); available at : http://www.eurekalert.org/pub_releases/2013-12/l-fdc120313.php 4. Byar DP, Simon RM, Friedewald WT et al. (1976) Randomized clinical trials. Perspectives on some recent ideas. N Engl J Med 95, 74–80. 5. Holick MF (2006) Resurrection of vitamin D deficiency and rickets. J Clin Invest 116, 2062–2072. 6. Ameri P, Canepa M, Milaneschi Y et al. (2013) Relationship between vitamin D status and left ventricular geometry in a healthy population: results from the Baltimore Longitudinal Study of Aging. J Intern Med 273, 253–262. 7. Assalin HB, Rafacho BP, dos Santos PP et al. (2013) Impact of the length of vitamin D deficiency on cardiac remodeling. Circ Heart Fail 6, 809–816. 8. Ramagopalan SV, Maugeri NJ, Handunnetthi L et al. (2013) Hospital admissions for vitamin D related conditions and subsequent immune-mediated disease: record-linkage studies. BMC Med 11, 171. 9. Pludowski P, Holick MF, Pilz S et al. Vitamin D effects on mitochondrial health, immunity, autoimmunity, cardiovascular disease, cancer, fertility, pregnancy, dementia and mortality – a review of recent evidence. Autoimmune Rev 2013; 12: 976-89 10. Grant WB. Using findings from observational studies to guide vitamin D randomized controlled trials. J Int Med. Published online: 5 May 2014 DOI: 10.1111/joim.12260

On 2015 Mar 09, Philippe Autier commented:

In our systematic review, we concluded that the association between low vitamin D status (as assessed by serum 25(OH)D concentration) and ill health found in most observational studies is not causal and could just reflect the influence of inflammatory processes present in many diseases (1). Our conclusion was essentially based on the finding that in the vast majority of randomized trials, vitamin D supplementation had no impact on health outcomes. O Gillie refutes our conclusions on the basis that type 2 errors could underlie the absence of trial result compatible with health benefit of vitamin D supplementation. This refutation is wrong because it equates to the common mistake of post-hoc power calculation (2). A type 2 error consists in rejecting the alternative hypothesis (i.e., vitamin D supplementation can prevent or cure diseases) and accepting the null hypothesis (i.e., vitamin D supplementation cannot prevent or cure diseases) when ultimately, the alternative hypothesis would turn out to be true. The avoidance of type 2 errors in randomized trials largely governs the size of trials, because the more subjects are included in a trial, the lower the probability of a type 2 error. This concept is termed the “power” of a trial that is the capacity of a trial to demonstrate the likelihood of the alternative hypothesis. The power is computed prior to randomization as the probability that a trial of a given size could confirm the possibility of a truly existing association between the intervention and the outcome. But once a trial has been conducted, the interpretation of results should be solely driven by the probability of type 1 error to avoid fallacious interpretation based on post-hoc power calculation.

For example, observational studies on cardiovascular diseases have documented risk reductions of about 40% for high versus low quantile of serum 25(OH)D concentration. At inception of their meta-analysis of randomized trials on vitamin D supplements and the risk of cardiovascular death, given the numbers of subjects and of cardiovascular events reported by trials, Bjelakovic et al. (3) had a 96.1% power to detect a 20% reduction in the risk of cardiovascular death (bilateral test with an alpha risk of 5%). By usual standards, a 96.1% power is huge. Once the meta-analysis was done, the observed relative risk was of 1.02 (95%CI 0.91-1.13). The probability that a significant real risk reduction of 20% could have been missed by the meta-analysis is a type 1 error whose probability can be estimated as 1 out of 58548.

In our review, many randomized trials on vitamin D supplementation and their meta-analyses were sufficiently powered for identifying risk reductions of 20% or more. But the flat results of most trials indicate that the probability of type 1 error for missing risk reductions of 20% is negligible. In lay terms, this can be referred as an extremely small chance to have missed a true relationship between vitamin D supplementation and an outcome.

O Gillie insists much on vitamin D status in children. Probably that because growing evidence reveals the unsoundness of the hypothesis of an “epidemic of vitamin D deficiency” in adults, “vitamin D deficiency” in children is taken as the substitute culprit for ill health during adult life. However, in many countries where bright sunshine is not habitual, vitamin D supplementation for small children and pregnant women is common practice since 50 years or more. If occasional cases of rickets occur in latitudes remote from the equator, these cases concentrate in children with naturally dark skin. It is therefore highly unlikely that a large proportion of ill health in adults would be due to rampant vitamin D deficiency in children.

O Gillie alludes to physiological mechanisms by which a lack of vitamin D would harm individuals, including cardiac remodeling and auto-immune diseases. But the scientific literature is replete with speculations on biological mechanism evoked by believers for promoting the putative healthy virtues of antioxidants, vitamins, and many other micro or macro nutrients. Unfortunately, ugly facts, often under the form of results of randomized trials (e.g., ref. 4 and 5), practically always demonstrate the inanity of these speculations.<br> We were surprised that O Gillie declared to have no conflict of interest (http://www.vitdco.com/pages/about-us).

Philippe Autier MD, Mathieu Boniol PhD, Patrick Mullie PhD

1. Autier P, Boniol M, Pizot C, Mullie P. Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol 2014;2(1):76-89.
2. Hoenig, John M. and Heisey, Dennis M. (2001), The Abuse of Power: The Pervasive Fallacy of Power Calculations for Data Analysis, The American Statistician, 55, 19-24.
3. Bjelakovic G, Gluud LL, Nikolova D, Whitfield K, Wetterslev J, Simonetti RG, Bjelakovic M, Gluud C. Vitamin D supplementation for prevention of mortality in adults. Cochrane Database Syst Rev 2011; (7):CD007470.
4. Miller ER 3rd, Pastor-Barriuso R, Dalal D, Riemersma RA, Appel LJ, Guallar E. Meta-analysis: high-dosage vitamin E supplementation may increase all-cause mortality. Ann Intern Med. 2005 Jan 4;142(1):37-46.
5. Lawlor DA, Davey Smith G, Kundu D, Bruckdorfer KR, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet. 2004 May 22;363(9422):1724-7.

On 2015 Mar 09, William Grant commented:

Comments on why vitamin D observational and intervention studies disagree

In the paper by Autier et al. [1], it was stated that "The discrepancy between observational and intervention studies suggests that low 25(OH)D is a marker of ill health. Inflammatory processes involved in disease occurrence and clinical course would reduce 25(OH)D, which would explain why low vitamin D status is reported in a wide range of disorders." In my comment I will argue that the primary reason for the discrepancy between observational and intervention studies lies in the design of the intervention studies and that there is evidence that vitamin D3 reduces biomarkers of inflammation.

If intervention trials are intended to evaluate observational studies, i.e., determine whether vitamin D has a beneficial role in health, they should start with an understanding of the 25-hydroxyvitamin D [25(OH)D] concentration-health outcome relation [2]. They should then seek to enroll people in the studies with sufficiently low 25(OH)D concentrations that the vitamin D dose used in the study will significantly change the position along the relation [3]. If this is not done, the intervention study then becomes an exercise in determining whether supplementing the average population with a given amount of vitamin D has any impact on health outcomes, which is a different question.

A typical 25(OH)D concentration-health outcome relation was recently published for breast cancer incidence from case-control studies [4]. The power law fit to the data from 11 studies goes from an odds ratio of 1.77 at 15 nmol/L, to 1.04 at 30 nmol/L, 0.70 at 50 nmol/L, 0.52 at 75 nmol/L, and 0.41 at 100 nmol/L. The 25(OH)D concentration-PTH relation has a similar form, extending to 185 nmol/L [5]. Thus, there is a large change in odds ratio for a small change in 25(OH)D for low 25(OH)D concentrations, with the change decreasing with increasing 25(OH)D concentration. In addition, 25(OH)D concentration changes in a nonlinear fashion with respect to baseline 25(OH)D concentration for a given vitamin D3 oral intake [6], so it takes higher vitamin D3 doses to have an impact at higher baseline concentrations.

Some examples from the journal literature are instructive. A vitamin D3 trial regarding respiratory infections in adults conducted in New Zealand had baseline 25(OH)D concentration of 73 nmol/L and achieved 25(OH)D concentration above 120 nmol/L but found no reduction [7]. In contrast, a related study in Mongolia involving school children with baseline 25(OH)D concentrations of 18 nmol/L given 300 IU/d vitamin D3 found a 50% reduction in acute respiratory infections [8].

It is noted that the history of the understanding of the health benefits of vitamin A went through some of the same concerns as the understanding of those for vitamin D is today. About half of the intervention studies in the 1930s found beneficial effects in reducing infections [9]. In that paper "As suggested by Hess et al. (1933), no therapeutic benefit of vitamin A was noted for infants who were already sufficient in vitamin A." [10]. Now vitamin A supplementation prevents blindness and saves many lives from death due to infections in developing countries [11].

Regarding inflammation, a paper was recently published in response to Ref. 1 in which all vitamin D randomized controlled trials found were examined for beneficial effects on biomarkers of inflammation along with baseline and achieved 25(OH)D concentration [12]. For those trials using vitamin D3 with baseline 25(OH)D concentration between 17 and 47 nmol/L, 49% found a beneficial effect. For those trials with baseline 25(OH)D concentration between 49 and 77 nmol/L, only 27% did. In addition, 55% those with achieved 25(OH)D concentrations between 35 and 82 nmol/L had a beneficial finding contrasted with 29% for those with achieved concentrations between 83 and 187 nmol/L. Thus, baseline 25(OH)D concentration was much more important than achieved 25(OH)D concentration.

Well-designed vitamin D3 intervention studies should confirm many of the findings from observational studies. However, for now, observational studies remain the best guide to optimal 25(OH)D concentrations [13].

References 1. Autier P, Boniol M, Pizot C, Mullie, P. Vitamin D status and ill health: a systematic review. Lancet Diabetes & Endocrinology. 2014;2(1):76-89. 2. Grant WB. Using findings from observational studies to guide vitamin D randomized controlled trials. J Intern Med. 2015;277(1):83-6. 3. Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. Nutr Rev. 2014;72(1):48-54. 4. Grant WB. 25-Hydroxyvitamin D and breast cancer, colorectal cancer, and colorectal adenomas: case–control versus nested case–control studies, Anticancer Res. 2015;35(2):1153-60. 5. Valcour A, Blocki F, Hawkins DM, Rao SD. Effects of age and serum 25-OH-vitamin D on serum parathyroid hormone levels. J Clin Endocrinol Metab. 2012;97(11):3989-95. 6. Garland CF, French CB, Baggerly LL, Heaney RP. Vitamin D supplement doses and serum 25-hydroxyvitamin D in the range associated with cancer prevention. Anticancer Res 2011;31:617-22. 7. Murdoch DR, Slow S, Chambers ST, Jennings LC, Stewart AW, Priest PC, Florkowski CM, Livesey JH, Camargo CA, Scragg R. Effect of vitamin D3 supplementation on upper respiratory tract infections in healthy adults: the VIDARIS randomized controlled trial. JAMA. 2012;308(13):1333-9. 8. Camargo CA Jr, Ganmaa D, Frazier AL, Kirchberg FF, Stuart JJ, Kleinman K, Sumberzul N, Rich-Edwards JW. Randomized trial of vitamin D supplementation and risk of acute respiratory infection in Mongolia. Pediatrics. 2012;130(3):e561-7. 9. Semba RD. Vitamin A as "anti-infective" therapy, 1920-1940. J Nutr. 1999;129(4):783-91 10. Hess AF, Lewis, JM, Barenberg LH. Does our dietary require vitamin A supplement? J. Am. Med. Assoc. 1933;101:657–63. 11. Sommer A. Preventing blindness and saving lives: the centenary of vitamin A. JAMA Ophthalmol. 2014;132(1):115-7. 12. Cannell JJ, Grant WB, Holick MF. Vitamin D and inflammation. Dermatoendocrinol. 2014;6(1): e983401-1-10. Published online Jan. 29, 2015 http://www.tandfonline.com/doi/pdf/10.4161/19381980.2014.983401 13. Grant WB, Wimalawansa SJ, Holick MF, Cannell JJ, Pludowski P, Lappe JM, Pittaway M, May P. Emphasizing the health benefits of vitamin D for those with neurodevelopmental disorders and intellectual disabilities. Nutrients. 2015;7:1538-64.

Disclosure My organization, Sunlight, Nutrition and Health Research Center, receives funding from Bio-Tech Pharmacal, Inc. (Fayetteville, AR) and the Vitamin D Council (San Luis Obispo, CA). In the past it also received funding from), the UV Foundation (McLean, VA), the Vitamin D Society (Woodstock, Ontario), the Sunlight Research Forum (Veldhoven), and MediSun Technology (Highland Park, IL.

On 2015 Mar 10, Oliver Gillie commented:

Dr Autier has misunderstood my affairs which are explained in a Corrigendum to the scientific journal, Public Health Nutrition (see reference below). I have always reported conflict of interest accurately. The editor of PHN published the Corrigendum to clarify the situation while agreeing that I had no financial conflict of interest to declare. Dr Autier raises the issue once more but I have already explained my position in the Corrigendum and am surprised he has not seen it.

I worked with the Vitamin D Company Ltd helping them to sell their 5,000 IU tablet in the UK at a time when high dose tablets were difficult to get in the UK. I felt that availability of the 5000 tablet could make an important contribution to our national health. However I never received any sales commission or any money at all from the Vitamin D Company and so have no financial conflict of interest to declare. In fact my 11 year campaign to inform the public about the facts of vitamin D and sunshine has cost me several thousand pounds in publishing and distributing, free of charge, reports on vitamin D (please see below for details). These reports have done much to publicise scientific and health issues surrounding sunshine and vitamin D which have been widely misunderstood by influential bodies such as Cancer Research UK. Over these 11 years I have made no earnings other than from occasional newspaper and magazine articles.

Dr Autier is a staff member of the International Prevention Research Institute (IPRI), a body which is funded by industry and undertakes custom research for industry. This financial security allows Dr Autier the luxury of adopting a fallacious statistical argument, which may be welcomed by the pharmaceutical industry, but does them no credit because it is poorly thought out scientifically and poorly argued. Autier et al have presented lengthy statistical results which incorporate this scientific error, an error well understood by most experimentalists yet denied here by Autier. His response below tells us what we know already but fails to address the scientific issue raised by my article in Public Health Nutrition (ref below).

REFERENCES:

The CORRIGENDUM in PHE:

http://journals.cambridge.org/download.php?file=/PHN/S1368980015000130a.pdf&code=a0e2984cd8793a37d36eebf438633b36

ARTICLE ON AUTIER et al’s ERROR:

“Controlled trials of vitamin D, causality and type 2 statistical error.” Public Health Nutrition 2014

TWO REPORTS on vitamin D written by Gillie (summaries below) are available free from: www.healthresearchforum.org.uk :

“SUNLIGHT ROBBERY - Health benefits of sunlight are denied by current public health policy in the UK.” (2004), drew attention to the health benefits of sunbathing and vitamin D and inadequate public understanding of the issues. Alternative advice for the public on safe sun exposure, given in the book, differs in important ways from advice given by Cancer Research-UK, which has encouraged sun avoidance so risking serious vitamin D insufficiency and increasing overall risks of cancer. Sir Richard Doll commented on the book: “I am most impressed with the way Gillie has collected and presented the evidence”.

“SCOTLAND'S HEALTH DEFICIT - an explanation and a plan (2008)”, has drawn attention to vitamin D insufficiency in Scotland and the link with chronic disease. The central hypothesis of the book is that a substantial and significant portion of the excess mortality of Scotland compared with England is caused by vitamin D insufficiency. This excess mortality, known as “the Scottish effect”, cannot be explained by smoking, alcohol, diet or poverty. A subsidiary hypothesis is that major chronic diseases that are caused at least in part by D insufficiency (e.g. cancer, heart disease, MS etc) have a higher incidence in Scotland than in England. Evidence for these ideas is examined in the book and general support is found for them. Sir Muir Gray commented: “I was very impressed by the strength of evidence and by the conclusion…The work is of importance and a great achievement.”