Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials
BMJ 2014; 348 doi: https://doi.org/10.1136/bmj.g2035 (Published 01 April 2014) Cite this as: BMJ 2014;348:g2035All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor,
We read with interest the study by Theodoratou and colleagues [1]. Their comprehensive umbrella review of an impressive number of systematic reviews and meta-analyses did not find convincing evidence of a clear role of vitamin D on the vast majority of health outcomes. Interestingly, vitamin D supplementation did not reduce the risk of fractures and falls in older people.
Some comments might be of interest. Before shifting to the other edge of underestimating the value of vitamin D supplementation, it is prudent to recognize that these findings cannot be generalized. First, the anti-fracture and protective role of vitamin D against falls, when it is co-administered with calcium [2], should not be neglected. This has also been shown for fracture risk reduction [3]. Second, baseline 25-hydroxy-vitamin D [25(OH)D] levels may also be an issue, since vitamin D deficient subjects seem to present a greater response in fall risk reduction compared with those with vitamin D sufficient levels [2]. Third, the anti-fracture benefit of vitamin D seems to be dose-dependent, as shown by another recent meta-analysis, in which patients with doses >800 IU/d seem to benefit more than those with lower doses [4]. Fourth, institutionalized patients may also benefit more than community-dwelling patients from vitamin D [3,4].
Regarding the non-skeletal effects of vitamin D, they are difficult to be proven yet with current data, since most studies are of a relatively short duration to show a reduction in outcomes, such as cancer or cardiovascular disease (CVD). It must be also emphasized that most studies had the effect of vitamin D supplementation on bone health as the primary outcome. However, a recent study by Chowdhury and colleagues [5] showed that vitamin D deficiency is associated with a modest but significant increased all-cause and disease-specific mortality, mainly from CVD, compared with 25(OH)D concentrations ≥30 ng/ml. Perhaps greater doses of vitamin D, especially vitamin D3 (i.e. >2,000 IU/d) are needed [5] and the response may depend also on the baseline levels. These questions may be answered by upcoming trials, such as the VITAL study.
We suggest that the need for vitamin D supplementation must be individualized, taking account the patient’s age and status (institution or community), fracture risk, baseline 25(OH)D levels and it should be also co-administered with calcium.
References
1.Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis JP. Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ 2014;348:g2035. doi: 10.1136/bmj.g2035.
2.Murad MH, Elamin KB, Abu Elnour NO, Elamin MB, Alkatib AA, Fatourechi MM, et al. Clinical review: The effect of vitamin D on falls: a systematic review and meta-analysis. J Clin Endocrinol Metab 2011;96:2997-3006.
3.Chung M, Lee J, Terasawa T, Lau J, Trikalinos TA. Vitamin D with or without calcium supplementation for prevention of cancer and fractures: an updated meta-analysis for the U.S. Preventive Services Task Force. Ann Intern Med 2011;155: 827-38.
4.Bischoff-Ferrari HA, Willett WC, Orav EJ, Lips P, Meunier PJ, Lyons RA, et al. A pooled analysis of vitamin D dose requirements for fracture prevention. N Engl J Med 2012; 367: 40-9.
5.Chowdhury R, Kunutsor S, Vitezova A, Oliver-Williams C, Chowdhury S, Kiefte-de-Jong JC, et al. Vitamin D and risk of cause specific death: systematic review and meta-analysis of observational cohort and randomised intervention studies. BMJ. 2014;348:g1903. doi: 10.1136/bmj.g1903.
Competing interests: No competing interests
Boucher [1] presents a number of arguments in favour of the opinion that vitamin D supplementation has substantial benefits, but this is not clearly shown in our review [2]. Given the study selection protocol, the risk of overlap of original studies is minimal for any given outcome, as we only selected one review per outcome for our analysis. The importance of the supplementation dose and of pre-supplementation 25-hydroxyvitamin D (25-OHD) levels is valid. In our discussion we highlighted that inappropriately low dose or short duration of vitamin D supplementation in the randomised controlled trials might be inadequate to raise the body’s vitamin D concentrations enough to show a difference between the arms of a trial. However, there is no certainty that high doses and long duration would necessarily be effective and it is widely used practice to use such arguments to avoid accepting that an intervention is simply ineffective, as documented in vitamin E, hormone treatment, or invasive cardiac interventions in stable patients cases [3,4]. We also mentioned that large differences in baseline plasma concentrations of 25-OHD in different populations could interfere with the effect of the supplementation, but similarly, this could have an impact in either direction.
Calcium supplementation was discussed in our review mainly because vitamin D and calcium were co-administered in the majority of vitamin D trials. On the other hand, we believe that interactions between multiple vitamins are very difficult to be reliably investigated, unless one specifically designs trials that study the supplementation of multiple vitamins. Finally, to account for the specific problems of observational nutrient intake studies (as the ones described by Ioannidis [5]) we decided to exclude observational meta-analyses of dietary or supplementary vitamin D intake. Overall, we conclude that trends for survival benefits in subgroups of trials should be seen as exploratory and need further validation.
References
1. Letter by Boucher [based on her rapid response http://www.bmj.com/content/348/bmj.g2035/rr/693593 ]
2. Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis JP. Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ. 2014;348:g2035.
3. Siontis GC, Tatsioni A, Katritsis DG, Ioannidis JP. Persistent reservations against contradicted percutaneous coronary intervention indications: citation content analysis. Am Heart J. 2009;157(4):695-701.
4. Tatsioni A, Bonitsis NG, Ioannidis JP. Persistence of contradicted claims in the literature. JAMA. 2007;298(21):2517-26.
5. Ioannidis JP. Implausible results in human nutrition research. BMJ. 2013;347:f6698.
Competing interests: No competing interests
It is asserted that low levels of 25(OH)D accurately reflect vitamin D status; (i.e., vitamin D storage and VDR-mediated control of calcium metabolism and innate immunity). However, measurement of both 1,25(OH)2D and 25(OH)D is essential to accurately assess vitamin D status and to recognize dysregulated vitamin D metabolism. Blaney et al. found that serum 25(OH)D alone is not a sensitive measure of the autoimmune disease state. Measuring both metabolites often demonstrates a positive correlation of elevated 1,25(OH)2D and low 25(OH)D to inflammatory diseases.
1. Thurnham DI. Plasma 25-Hydroxy-Cholecalciferol (Vitamin D) is Depressed by Inflammation: Implications and Parallels with Other Micronutrients. Sight Lif. 2011;25((2)).
2. Blaney GP, Albert PJ, Proal AD. Vitamin D metabolites as clinical markers in autoimmune and chronic disease. Ann N Y Acad Sci. Sep 209;1173:384-90.
Competing interests: No competing interests
To the Editor
The report, “Vitamin D and multiple outcomes; umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials”, and related comment [1a,b], suggests that benefits from higher vitamin D repletion are limited. However, the findings may have been affected by a number of factors that were not adjusted for such as,
1) the likely inclusion of reported data in more than one of the meta-analyses examined,
2) not allowing for the fact that some benefits appear only at higher intakes, e.g., >800 IU/day for fracture reductions in older women[2].
3) that improvements are often seen only in initially deficient subjects, as for muscle strength[3] or
4) that ‘threshold’ 25(OH)D concentrations are reported for the appearance of some benefits, e.g. >80nmol/l for reductions in insulin resistance,[4] and
5) non-exclusion of studies using huge interval doses with initial adverse effects[5], though such problems are mentioned in the Discussion section[1].
Furthermore, failure of supplementation to achieve repletion, or being replete before supplementation, affects nutrient RCT outcomes,[6] but this is not discussed. Though the authors refer to the importance of calcium intakes in studying vitamin D efficacy, they do not discuss the interactions of other nutrients known to interact with vitamin D in humans, e.g., vitamin A[7], vitamin K[8] and magnesium[9] - intakes of vitamin A are woefully inadequate in many developing countries[10] and excessive in most affluent communities, circulating retinoids being in the potentially toxic range in 34% of Europeans.[11] Thus, confounding by lack of data on vitamin A intakes is likely to be considerable.
In addition, the degree of tissue damage when RCTs start matters, since irreversible changes cannot be corrected, (as long known for bony deformity in rickets and already suspected for islet insulin secretion in 1995[12]), and recently highlighted as a potential problem for vitamin D and target soft tissues since health benefits from adequate provision of vitamin D may well require life-long repletion,[13] a concept which has recently been provided with considerable support.[14,15]
It is also surprising that Ioannidis and colleagues do not appear to have allowed for all the problems inherent in nutritional research in humans since a recent BMJ editorial by Ioannidis discusses the many specific problems of nutrient intake studies, including RCTs, and clearly identifies many factors that need to be allowed for in RCTs of nutrients, as compared to RCTs of pharmaceuticals, detailing also why varying intakes of single nutrients can only be expected to produce relatively small (<10%) effects on health.[16] However, Chowdury et al this week report 35% increases in mortality with low vitamin D status on meta-analysis of 73 observational cohorts (n= 849,412 subjects) together with an 11% reduction in overall mortality (from cardiovascular, cancer and non-vascular and non-cancer deaths) after supplementation in 22 randomised intervention studies using solely vitamin D3 (n=30,716 subjects) but not with vitamin D2.
Overall, the available evidence appears to suggest that the possibility of causality of vitamin D insufficiency for human ill health should not be dismissed.
References
1 (a). Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis JP. Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ. 2014. Apr1;348:g2035. doi: 10.1136/bmj.g2035. (b), Kmietowicz Z. Vitamin D shows no clear evidence of benefits despite hundreds of studies BMJ. 2014 Apr 1;348:g2489. doi: 10.1136/bmj.g2489.
2. Bergman GJ, Fan T, McFetridge JT, Sen SS. Efficacy of vitamin D3 supplementation in preventing fractures in elderly women: a meta-analysis. Curr Med Res Opin. 2010 26:1193-201.
3. Muir SW, Montero-Odasso M. Effect of vitamin D supplementation on muscle strength, gait and balance in older adults: a systematic review and meta-analysis. J Am Geriatr Soc. 2011 59:2291-300. Review
4. von Hurst PR, Stonehouse W, Coad J. Vitamin D supplementation reduces insulin resistance in South Asian women living in New Zealand who are insulin resistant and vitamin D deficient - a randomised, placebo-controlled trial. Br J Nutr. 2010 103:549-55
5. Hansen KE. High-dose vitamin D: helpful or harmful? Curr Rheumatol Rep. 2011 13:257-64.
6. Lappe JM1, Heaney RP. Why randomized controlled trials of calcium and vitamin D sometimes fail. Dermatoendocrinol. 2012 ;4:95-100.
7. Carlberg C. Lipid soluble vitamins in gene regulation. Biofactors. 1999;10:91-7. Review.
8. Torbergsen AC, Watne LO2, Wyller TB3, Frihagen F4, Strømsøe K5, Bøhmer T6, Mowe M7. Vitamin K1 and 25(OH)D are independently and synergistically associated with a risk for hip fracture in an elderly population: A case control study. Clin Nutr. 2014.01.016. [Epub ahead of print]
9. Matsuzaki H, Katsumata S, Kajita Y, Miwa M. Magnesium deficiency regulates vitamin D metabolizing enzymes and type II sodium-phosphate co-transporter mRNA expression in rats. Magnes Res. 2013;26:83-6.
10. Ahmed T, Hossain M, Sanin KI. Global burden of maternal and child undernutrition and micronutrient deficiencies. Ann Nutr Metab. 2012;61 Suppl 1:8-17.
11. Flynn A, Hirvonen T, Mensink GB, Ocké MC, Serra-Majem L, Stos K, Szponar L, Tetens I, Turrini A, Fletcher R, Wildemann T. Intake of selected nutrients from foods, from fortification and from supplements in various European countries. Food Nutr Res. 2009; 12;53.
12. Penniston KL, Tanumihardjo SA. The acute and chronic toxic effects of vitamin A. Am J Clin Nutr. 2006;83:191-201. Review.
13. Afzal S, Brøndum-Jacobsen P, Bojesen SE, Nordestgaard B. Vitamin D concentration, obesity and risk of diabetes: a Mendelian randomisation study. Lancet Diabetes & Endoctinology. 2014;2:2298-306
14. Pilz S, Gaksch M, Tomaschitz A. 28 Jan 2014 - Vitamin D and prevention of diabetes: is lifelong endogenous vitamin D needed? Lancet Diabetes & Endocrinology. 2014;2:267 - 268
15. Ioannidis JPA. Implausible results in human nutrition research. BMJ. 2014;347:7
16. Chowdury R, Kunutsor S, Vitezova A, et al. Vitamin D and risk of cause specific death: systematic review and meta-analysis of observational cohort and randomised intervention studies. BMJ; 2014:11
Competing interests: No competing interests
My initial reaction while reading this (and before reading the responses) was that some fairly opaque statistics were covering for serious methodological flaws. Meta-analysis has insurmountable shortcomings - such as heterogeneity and reliance on and amplification of flaws in the original studies. To repeat this process by meta-analysing the meta-analyses seems to place excessive hope on the virtues of amalgamation of data for apparent mathematical benefits. I would place more trust in a single, well-designed, adequately powered study.
To borrow a painting analogy, all the colours of the rainbow have been mixed and we have come up with brown. I am glad to see that other, better qualified people have similar misgivings with this paper and have eruditely expressed them above.
Competing interests: No competing interests
Vitamin D can be synthesized by human skin from 7-dehydrocholesterol with the help of ultraviolet light obtained from sunlight. Calcitriol is the active form of vitamin D which is formed after two hydroxylation reactions; one at 25th position occurs in liver by 25-α-hydroxylase and the other in kidney at 1st position by 1-α-hydroxylase enzymes. The presence of extra renal 1-α-hydroxylase enzyme in many tissues including macrophages, B cells and T cells has led to the proposal that vitamin D also play important role in these tissues. 25-hydroxy vitamin D is the major circulating form of vitamin D and the extra renal tissues convert it into calcitriol where it acts in a paracrine manner.
As a result of this, many disease conditions including autoimmune diseases, infectious diseases including HIV, type 2 diabetes and metabolic syndrome and some types of cancers are associated with low serum levels of 25-hyroxy vitamin D. In these conditions, whether decreased 25-hydroxy vitamin D in serum is the cause or effect is debatable. The role of vitamin D in modulating the immune responses involves production of calcitriol by extra renal 1-α-hydroxylase enzyme acting on 25-hydroxy vitamin D obtained possibly from plasma. This could possibly explain the decreased level of 25-hydroxy vitamin D observed in the plasma of patients with various diseases. Supplementing vitamin D in such cases would then depend on the effect of vitamin D in the overall disease process. The overall effect of vitamin D may be even more complicated in inflammatory conditions in metabolically active tissues such as adipose tissue. Since inflammation is one of the common events observed in many of the clinical outcomes studied, it is very important that future research on vitamin D should be conducted to decipher the effects of vitamin D in mediating both immune and metabolic functions.
Lastly, a deeper understanding of the vitamin D metabolism and better biomarker to assess its deficiency are needed.
Competing interests: No competing interests
Umbrella review of vitamin D and multiple health outcomes: Which way up was the umbrella…did this review collect knowledge or shield us from it?
Author John Ioannidis described umbrella reviews [1] as “subjective and suboptimal”, “limited by the data in the primary studies” and the process of “patching together pre-existing reviews”, thus open to subjective interpretations based on the authors opinions.
This umbrella review of Vitamin D by four authors including John Ioannidis [2] found a null result without exploring the reasons for their findings. The reason Vitamin D meta-analyses and reviews fail to produce useful results is thought to be biological flaws in primary studies [3, 4]. These flaws lead to null results as the intervention does not change the Vitamin D status; however these flaws may be overlooked when evaluating the research for Vitamin D and other nutrients [5, 6]. Furthermore, the methodology for differentiating flawed and non-flawed studies has been described by Heaney [7].
The review authors [2] describe flaws as “difficulties in relation to RCTs” such as “low dose vitamin D supplementation”, “large differences in baseline plasma concentrations of 25-hydroxyvitamin D” and “contamination with private use of vitamin D” in the randomised controlled trials that might be “inadequate to raise the body’s vitamin D concentrations enough to show a difference between the arms of a trial”. The review authors ascribe the concepts to Autier (2014) [8] although previously described by Heaney and others as biological flaws [5, 6]. Whilst this umbrella review [2] recognised these “difficulties” or biological flaws in their primary studies, the authors choose not to compare meta-analyses of flawed and unflawed primary studies.
The review authors [2] identify 107 systematic reviews and 74 meta-analyses with null and conflicting results, yet they have simply repeated the same process. Repeating the same thing over and over again expecting a different result has been described as insanity (a quote ascribed to Albert Einstein).
A previous review of Vitamin D [9] focused on studies without flaws concluded there was evidence to support the use of vitamin D supplementation in many non-skeletal diseases. Furthermore, there is a systematic review and meta-analysis on vitamin D supplementation and depression [10] that shows a positive result for primary studies without flaws, a negative result for studies with flaws and a null effect when studies are combined. These findings need validating by further research for depression and in other disorders.
Thus, there is a theoretical basis, a published methodology, and a meta-analysis to support the contention that Vitamin D reviews find and will continue to find null results unless primary studies with these ‘difficulties’ or biological flaws are excluded from the process.
The question remains why did these authors not present a review of flawed and unflawed studies in a review claimed to be all-encompassing?
References:
1. Ioannidis JP. Integration of evidence from multiple meta-analyses: a primer on umbrella reviews, treatment networks and multiple treatments meta-analyses. CMAJ2009;181:488-93.
2. Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis JP. Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ2014 Apr 1;348:g2035. doi: 10.1136/bmj.g2035.
3. Scragg RDo we need to take calcium with vitamin D supplements to prevent falls, fractures, and death? Curr Opin Clin Nutr Metab Care. 2012 Nov;15(6):614-24. doi: 10.1097/MCO.0b013e328359ef21.
4. Lappe, J.M.; Heaney, R.P. Why randomized controlled trials of calcium and Vitamin D sometimes fail. Dermatoendocrinol2012, 4, 95-100.
5. Heaney RP. Vitamin D: basal status and effective dose. N Engl J Med2012; 367:77–78.
6. Biesalski, H.K.; Aggett, P.J.; Anton, R.; Bernstein, P.S.; Blumberg, J.; Heaney, R.P. Scientific substantiation of health claims: Evidence-based nutrition. 26th Hohenheim Consensus Conference; Nutr2011, 27, S1–S20.
7. Heaney RP. Guidelines for optimizing design and analysis of clinical studies of nutrient effects. Nutr Rev2014; 72: 48054.
8. Autier P, Boniol M, Pizot C, Mullie P. Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol2014; 2: 76-89.
9. Spedding S, Vanlint S, Morris H, Scragg R. Does Vitamin D Sufficiency Equate to a Single Serum 25-Hydroxyvitamin D Level or Are Different Levels Required for Non-Skeletal Diseases? Nutrients2013; 5(12):5127-5139.
109. Spedding S. Vitamin D and depression: A systematic review and meta-analysis comparing studies with and without biological flaws. Nutrients2014 in press.
Competing interests: No competing interests
Vitamin D is one of the essential fat soluble vitamins required for normal mineralization and regulation of calcium and phosphorus metabolism. Its active form Calcitriol requires normal functioning liver and kidney for 25 hydroxylation and 1 hydroxylation respectively. It is now labeled as a steroid hormone which has specific cytosolic receptors in specific target organs like instestine, bone and kidney to exert its physiological functions. Other organs which have such receptors are subjected to calcitriol action possibly involving calcium and phosphorus linked metabolic functions. Metabolic studies involving such dietary components and also many such studies are carried out in isolated conditions or in testing specific research questions that are needed for fully evaluating the potential benefits derived from such a vitamin. Such an approach to find out what other outcomes can be derived from Vitamin D on human health apart from its known established effects on Calcium and phosphorus in bone metabolism may sometimes take away its real function to other generalized functions. Therefore let us be convinced that vitamin D is important for bone mineralization and demineralization.
Competing interests: No competing interests
The big problem with randomised trials of vitamin D supplementation is that woefully inadequate doses of vitamin D are used (e.g. 800 IU per day). Whilst the optimal blood level of (25-hydroxy) vitamin D is still a matter of debate, 100-200 nmol/L might be the ideal; the daily dose of vitamin D3 required to achieve this in adults is typically 5,000-10,000 IU per day (such doses are quite safe, as potential toxicity appears to require a sustained dose of 40,000 IU per day, and blood levels in excess of 375 nmol/L). It is thus hardly surprising when significant results are not seen – how many pharmaceutical drug trials would show a positive result if the drug being tested was given in an amount of only one-tenth of the therapeutic dose?!
Competing interests: No competing interests
Let us treat those who are deficient, not everybody!
The authors of the extensive review on observational and interventional vitamin D studies have to be congratulated for doing the important and challenging work to sum up the available vitamin D research in humans. They can however, by definition not overcome the surprising problem that is clearly evident from many of the primary intervention studies: often vitamin D deficiency was not a necessary inclusion criterion for entry in these studies.
We all know how important antibiotics, insulin, thyroxin, aspirin etc. are - but what if we performed studies on these very powerful medications in patients who do not actually need them, say aspirin for everybody, antibiotics without bacterial infection, insulin without diabetes, thyroxin with normal thyroid function. What would we conclude? That these are inefficient treatments.
This is what is currently happening to vitamin D: without deficiency, no treatment is necessary, and obviously no advantage is possible.
Besides, many of the available RCTs are fraught with other problems as well: (very) small sample size (let the statisticians see this!), usually without prior power calculation, (too) small doses, (too) short follow-up, no baseline (and post-intervention) vitamin D level measured, etc.
Unfortunately, some of the ongoing vitamin D trials continue to make these mistakes, and I will not surprised if they too fail to find a benefit.
So let us stay rational and perform properly powered and designed RCTs to finally show that vitamin D is powerful as well, but only in those who are vitamin D deficient!
Competing interests: I have received speaker honoraria from Fresenius Kabi and I am a lead investigator in the VITdAL@ICU study - ClinicalTrials.gov identifier: NCT01130181