This site uses cookies to store information on your computer. Learn more x


Providing perspectives on recent research into vitamins and nutritionals


Meta-Analyses of Vitamin D and the Nuances of Nutrition Research

By Julia Bird

In the British Medical Journal today, two articles attempt to synthesize research into vitamin D and a variety of different health outcomes (lead authors are Chowdhury, Theodoratou)  and there is a comment by Welsh and Sattar in the same issue.

To briefly summarize the results, the Chowdhury article looked into 73 studies that included 850,000 participants in observational studies of vitamin D serum levels, and 22 RCTs using vitamin D supplements versus placebo or no treatment, and looked at the effects on death from various causes. There was an increased risk of death in participants with low vitamin D levels (lowest third of vitamin D levels), and each decrease of 10 ng/ml vitamin D increased risk of death by 16%. The supplementation studies showed a decrease in risk of death when vitamin D3 was used as a supplement by 11%.

The Theodoratou article conducted an umbrella review, which is a systematic review of systematic reviews and meta-analyses. Three types of systematic reviews and meta-analyses were included: systematic reviews of observational studies (76 reviews), meta-analyses of observational studies (48 meta-analyses), and meta-analyses of supplementation studies (57 meta-analyses. The authors here found that many observational studies found that higher levels of vitamin D were protective for a great variety of conditions. The meta-analyses of intervention studies were only positive for prevention of low birth weight, dental caries and non-vertebral fractures.

The commentary was rather critical of vitamin D supplementation and suggested caution in measuring vitamin D levels in asymptomatic patients.

While the collation of research seems like a good idea at the outset, allowing us to see the forest rather than looking at the individual trees, the approach is perhaps not ideal for nutrients, and also fails to capture differences in the study design or population that may affect the outcome.

Vitamin D is an essential nutrient and a certain minimum level is required in everyone’s diet. Very low levels of vitamin D are a cause of death (see reports of a recent case of rickets in an infant, and two cases reports of serious complications of vitamin D deficiency in older adults). Despite calls that observational research that low levels of vitamin D may be a marker of poor health in general, there should be no doubt in anyone’s mind that vitamin D deficiency should be avoided. Frank deficiency is found in many populations. For example, we have previously reported that 20% of adults in the United Arab Emirates, and 15% of Australian adults are deficient with levels under established cutoffs for osteomalacia. The current controversy over “optimal” serum concentrations and doses should not overshadow concern about frank deficiency, which is prevalent and should be a concern for health professionals.

In this vein, adequate levels of vitamin D are required over the entire lifespan. Many of the endpoints that showed a positive association in observational research are unable to be adequately tested in randomized controlled trials. For example, diseases such as cancer develop over decades and are relatively uncommon in younger people, therefore a long trial conducted over decades with a large number of subjects would probably be needed to answer the question of whether vitamin D prevents cancer. Even large intervention trials such as VITAL that is currently ongoing is unlikely to be able to answer this question. The Theodoratou article found a lack of concordance between observational and intervention studies for many conditions; the practical issues with supplementing an essential nutrient over a long period of time means that observational research may be the best that we have to go by for many conditions.

The other issue with collating research is that baseline vitamin D levels varied in the research populations used, as did the dose and length of time supplemented. The Theodoratou paper stated outright that investigating the effect of dose was beyond the scope of their literature review. The actual serum concentrations were also not used in their analysis. The Chowdhury paper used blunt categories to look at the effect of dose and intervention time. Intervention studies are only likely to be effective when levels are inadequate, the dose is sufficient to improve vitamin D levels and has been given for a long enough time to have an effect on the outcome.

Vitamin D supplements do not give people “false reassurance”. Frank deficiency is found in a substantial proportion of the population. Individual lifestyle and diet factors should be taken into account when deciding whether or people are at risk. A supplement is an inexpensive, safe and convenient way for people to ensure vitamin D adequacy while we are waiting for more definitive results to come in about whether there is are added benefits of a higher dose beyond preventing deficiency.

Main citations:

Chowdhury 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;348:g1903 doi: 10.1136/bmj.g1903.

Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis J. 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.

Welsh P, Sattar N. Vitamin D and chronic disease prevention. BMJ. 2014;348:g2280 doi: 10.1136/bmj.g2280.