Vitamin D prevents colds? News stories turn blind eye to limitations of sunshine vitamin study

Kathlyn Stone is an associate editor with She tweets at @KatKStone.

Vitamin D is a hot research topic, but the public gets a lot of mixed messages about its usefulness and safety.

Case in point: One of the latest studies, published Wednesday, Feb. 15, analyzed results from 25 previous trials and found that vitamin D supplementation might help prevent respiratory infections in some people who have low vitamin D levels.

We were optimistic that the news release from the BMJ – the journal that published the study — would help steer news coverage in an informed direction. That’s because it referenced some of the limitations of the study in a linked editorial that accompanies the National Health Service (NHS)-sponsored study by Martineau and colleagues. The editorial is written by researchers who have previously published their own work on vitamin D but had no involvement in the Martineau study. They concluded that the study is a “hypothesis…requiring confirmation” in larger controlled trials.

Most news reports skipped the limitations

But optimism gave way to a feeling of “foiled again” when we looked at February 16’s crop of vitamin D articles. Most of the stories we read contained none of the numerous cautions and limitations described in the editorial. The few that did add some caveats had them buried far down in the story. And none of the concerns with the research were reflected in the headlines. 

The Guardian’s article on the study offered perhaps the strongest caution of any story we saw, but its original headline (“Adding vitamin D to food reduces deaths, say scientists”) was inaccurate at best. The study doesn’t claim that vitamin D reduces death. (The headline has since been updated to “Vitamin D ‘proved to cut risk of colds and flu'”–but this is still misleading; that hypothesis still requires confirmation, as the linked editorial points out.)

But the Guardian article improves from there. Lower in the article it includes snippets from the editorial:

“Although the authors consider the case proven, scientists are still divided. Mark Bolland from the University of Auckland and Alison Avenell from the University of Aberdeen say in an editorial in the BMJ that large randomised controlled trials – comparing people taking vitamin D with others who do not – are still needed.

“Current evidence does not support the use of vitamin D supplementation to prevent disease, except for those at high risk of osteomalacia (weak bones and muscles due to low blood vitamin D levels, currently defined as less than 25 nmol/L),” they write.”

NPR’s approach on the limitations (A Bit More Vitamin D Might Help Prevent Colds And Flu) was to provide some context about how hard it is to determine the right amount of vitamin D.

“Over the last 10 years, a number of studies have suggested that the sunshine vitamin can help prevent disease. That has led people to think that higher doses of supplements are better. But Abrams says he’s not convinced there’s a benefit of taking a supplement for people who are not deficient. “It needs further studies to confirm.”

“Abrams [an expert who had contributed to Institute of Medicine guidelines on vitamin D intake] says the importance of the new study is that it’s a summary of 25 controlled trials. “And it shows that people with very low vitamin D [levels] do better when they’re given supplements.” He says this is not too surprising. “If you’re deficient, getting an adequate amount will make a difference.”

“In other words, if you’re getting the recommended 600 IUs of vitamin D from your diet, a supplement may not lead to any further benefit. But the growing interest in vitamin D has lots of people curious about their levels.”

TIME’s single-source story contained no caveats, nor did the Harvard Gazette, which provided an even more glowing summary of the research. It begins, “A new global collaborative study has confirmed that vitamin D supplementation can help protect against acute respiratory infections.” There’s no mention of the very modest risk reduction or of the common problems associated with meta-analyses.

Allison Dostal, PhD, RD, a nutrition researcher at the University of Minnesota Medical Center and contributor, pointed out that the results of the study were “arrived at through sub-group analyses and may not have had the statistical power to fully support their assertions. Further, racial and ethnic subgroups were not studied in detail, which limits the translation of the results to specific populations at risk for vitamin D deficiency.”

The issue with subgroup analyses is that results found in one group (in this case based on age) don’t necessarily reflect the results found across all groups. For example, an intervention found beneficial in one subgroup may be inconsequential or even harmful in another.

Dostal said information from the meta-analyses could potentially help in development of vitamin D supplementation guidelines for prevention of acute respiratory infection — if the results are supported by more randomized trials designed to look specifically at the outcomes of interest.

While this study showed that daily or weekly dosing had a protective effect against acute respiratory infection, “randomized trials designed to study the difference in dosing frequencies, with appropriate control groups, are necessary to fully support this conclusion,” Dostal said.

Kathleen Fairfield, MD, MPH, DrPH, an internist and researcher at Maine Medical Center Research Institute and frequent contributor, had a similar view. “Meta-analyses are only as good as their source studies.” Many smaller studies with differences in their size and intent can’t be summed together to make useful conclusions, she explained. “There is absolutely no reason to be talking about population supplementation of vitamin D on the basis of this,” said Fairfield.

Key limitations that should have been noted

With the exception of the Guardian, all of the stories we saw glossed over or ignored the limitations outlined in the BMJ editorial. Here are some of the key ones:

  • The definition of “acute respiratory tract infection” varied among the 25 studies – ranging from acute ear inflammation to lab-confirmed influenza to self-reported colds and confirmed pneumonia. Is a reduction in this mixture of conditions applicable to the general population?
  • The meta-analysis of 25 different studies showed only a 2% overall absolute risk reduction. (Absolute risk numbers are important since they show the true size of the benefit.)
  • The benefits from vitamin D augmentation appear from the study data to apply mainly to children ages 13 months to 15 years, 9 months. Here the absolute risk reduction was 13%. But in all other age groups, “the absolute reductions were small and statistically insignificant, ranging from 0-3%,” according to the editorial authors.

That finding in a small age group shouldn’t translate into the generalization that vitamin D augmentation reduces the chances for colds and flu in the general population. And health writers may have missed an easy opportunity to help educate rather than obfuscate findings from another vitamin D study.

You might also like

Comments (1)

Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

February 17, 2017 at 12:03 pm

This is an excellent blog on the issue of Vitamin D. I would quibble about only one word and that is the word “only”, as in, “The meta-analysis of 25 different studies showed only a 2% overall absolute risk reduction.” For many widely prescribed drugs for prevention, such as cholesterol-lowering drugs, antihypertensives, anti-osteoporosis or anti-diabetic treatments on the absolute benefits are often under 5%, often in the range of “only” 1-3% but when they are reported, very few commentators or critics qualify those benefits with an “only”. The main problem, of course, is that these slim benefits, where maybe one in 50 patients might benefit, are reported in misleading absolute terms, and so everyone gets the sense they are much more effective than they actually are. So to end this quibble, I’d say that if we are going to qualify the apparent low benefits of vitamins or alternatives with an “only” then we should also be doing the same with many drugs whose benefits are often quite miniscule. The most widely prescribed drug for osteoporosis, for example, which has had sales in the billions over the last 20 years was sold on a chance of benefiting 1% of highly osteoporotic women with reduced hip fractures. Strange thing was, you almost never heard anyone write “only 1%.”