Yesterday’s Alleged Vitamin D Dethroning Is a Good Example of Why It Is Wise to Withhold Judgment on Widely Hyped Dietary Medical Benefit Findings — Ask Instead, Where Is the Evidence and What Is Its Quality? — and Do Not Confuse Markers and Associations with Causation

06 December 2013

Citation — to the meta-analysis that appears to debunk the widely alleged non-bone benefits of vitamin D supplementation

Philippe Autier, Prof Mathieu Boniol, Cécile Pizot, and Patrick Mullie, Vitamin D status and ill health: a systematic review, Lancet Diabetes & Endocrinology, DOI:10.1016/S2213-8587(13)70165-7 (early online publication, 06 December 2013)

Citation — to The Lancet’s editorial comment regarding the study

Editorial, Vitamin D: chasing a myth?, Lancet Diabetes & Endocrinology, DOI:10.1016/S2213-8587(13)70164-5 (early online publication, 06 December 2013)

I seize on this meta-analysis as a way of reacquainting people with the two common analytical problems

These are:

(1) The media’s dissemination of purported, but poorly proven dietary supplementation benefits

and

(2) the tendency for everyone, including scientists, to confuse associations and medical condition markers with causation.

Language

I begin, for the scientifically literate, with the study’s findings as delivered in its own language.

Readers who doubt their ability to penetrate medical jargon should skip down to the lay language section, located just below the immediately following one.

For the scientifically literate — the findings in scientific language

From the paper’s abstract:

Low serum concentrations of 25-hydroxyvitamin D (25[OH]D) have been associated with many non-skeletal disorders. However, whether low 25(OH)D is the cause or result of ill health is not known.

We did a systematic search of prospective and intervention studies that assessed the effect of 25(OH)D concentrations on non-skeletal health outcomes in individuals aged 18 years or older.

We identified 290 prospective cohort studies (279 on disease occurrence or mortality, and 11 on cancer characteristics or survival), and 172 randomised trials of major health outcomes and of physiological parameters related to disease risk or inflammatory status.

Investigators of most prospective studies reported moderate to strong inverse associations between 25(OH)D concentrations and cardiovascular diseases, serum lipid concentrations, inflammation, glucose metabolism disorders, weight gain, infectious diseases, multiple sclerosis, mood disorders, declining cognitive function, impaired physical functioning, and all-cause mortality.

High 25(OH)D concentrations were not associated with a lower risk of cancer, except colorectal cancer.

Results from intervention studies did not show an effect of vitamin D supplementation on disease occurrence, including colorectal cancer. In 34 intervention studies including 2805 individuals with mean 25(OH)D concentration lower than 50 nmol/L at baseline supplementation with 50 μg per day or more did not show better results.

Supplementation in elderly people (mainly women) with 20 μg vitamin D per day seemed to slightly reduce all-cause mortality.

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 elderly people, restoration of vitamin D deficits due to ageing and lifestyle changes induced by ill health could explain why low-dose supplementation leads to slight gains in survival.

The last couple of years saw a noticeable rise in exhortations to boost one’s vitamin D intake, so as to reduce the risk of some developing some common medical conditions — it turns out that this probably a mistaken effort

© 2013 Philippe Autier, Prof Mathieu Boniol, Cécile Pizot, and Patrick Mullie, Vitamin D status and ill health: a systematic review, Lancet Diabetes & Endocrinology, DOI:10.1016/S2213-8587(13)70165-7 (early online publication, 06 December 2013) (at Abstract) (paragraph split)

Study’s vitamin D findings in ordinary language

From Reuters’ Kate Kelland — in her admirably transparent writing style:

Researchers cast doubt on the prevailing wisdom that vitamin D supplements can prevent conditions like cancer, diabetes and heart disease, saying on Friday low vitamin D may be a consequence, not a cause, of ill health.

[S]ome [previous] observational studies [had] suggested a link between low levels of vitamin D and greater risks of many acute and chronic diseases.

But it is not clear whether this is a cause-and-effect relationship, so various large trials have been conducted to try to test whether vitamin D supplementation can reduce the risk of developing disease.

Researchers led by Philippe Autier . . . analysed data from several hundred observational studies and clinical trials examining the effects of vitamin D levels on so-called non-bone health - including links to illness such as cancer, diabetes and cardiovascular disease.

They found that the benefits of high vitamin D levels seen in observational studies - including reduced risk of cardiovascular events, diabetes and colorectal cancer - were not replicated in randomised trials where participants were given vitamin D to see if it would protect against illness.

"What this discrepancy suggests is that decreases in vitamin D levels are a marker of deteriorating health," said Autier.

Is this a persuasive debunking? — Maybe not, which is how science works

From editors for Lancet Diabetes & Endocrinology:

For those who ‘believe’, the lack of benefit found in most trials completed thus far can be attributed to issues including inadequate supplementation, testing of a population not sufficiently vitamin D deficient at baseline, incorrect formulation, underpowering [in the statistical sense — meaning too few people in the pertinent study], or insufficient follow-up.

In other words, the editors’ argument is that some of the studies that the debunking analysis looked at may have been incorrectly done. Which, of course, would cast doubt on the meta-analysis’s tentative conclusions.

My two points — reiterated

Whether this meta-analysis presents a definitive debunking of vitamin D’s supplementation benefit, outside skeletal health, is unimportant — except as:

(a) a reminder of how difficult it is to separate causation from associations with and markers of ill health

and

(b) an indicator of how easy it is for scientifically illiterate media to sensationalize unproven study results.

The public needs to be aware of the latter point and scientists and medical researchers, the former.

The moral? — Be skeptical of “one bat” solutions to medical problems

Biology, physiology, and pathology are almost always far more complex than simple-minded solutions would have us believe.

The multi-factorial nature of most ill health confounds our ability to be definitively certain of what to do, especially in public health settings that mix a bunch of different genomes and patient environments into one pot.

And keep in mind that there is money to be made by hyping solutions that aren’t. Not everyone cares about what is true.

PeteFree.com

Yesterday’s Alleged Vitamin D Dethroning Is a Good Example of Why It Is Wise to Withhold Judgment on Widely Hyped Dietary Medical Benefit Findings — Ask Instead, Where Is the Evidence and What Is Its Quality? — and Do Not Confuse Markers and Associations with Causation