Dear Readers,
As you know, vitamin D is a subject of great interest; we are daily bombarded with news on the topic in different fields of specialization, some swaying us in one direction and some in another… You may have noticed the growing scientific interest on this subject, as attested by the increasing number of publications on PubMed since 2000, which even today stands at a high level, even if the number has begun to plateau.
It comes as no surprise that in addition to its noted impact on human skeleton and phosphocalcic metabolism, vitamin D can also have extraskeletal effects. There are at least five good reasons in support of this: vitamin D receptors are present in numerous cells – I would say they are nearly ubiquitous; vitamin D controls the transcription of numerous genes; it has endocrine effects, and not only calciotropic ones; the activation and catabolism of vitamin D take place in several organs and tissues; and it has intracrine and paracrine effects in numerous cells of various natures.
Most available studies on the effects of vitamin D are preclinical or observational. The latter often describe associations between a lack of vitamin D and the incidence, activity or outcomes of many illnesses, but they have an intrinsic limit of not being able to document a sure causal relationship. On the other hand, randomized, double-blind, placebo controlled clinical trials relative to supplementation – the only ones able to scientifically verify the effects of vitamin D – are few in number, sometimes for ethical reasons, and often suffer from bias.
The most frequent bias is the treatment of subjects without deficiency, forgetting that as a nutrient vitamin D can only have effects when it is lacking. Recently we have seen, for example, a publication of a study [1] in which subjects, who were for the most part not deficient, were given supplements: the conclusion of the ineffectiveness, in terms of preventing fractures and falls, of vitamin D supplements for adults living in senior communities created confusion among both doctors and patients. Instead, the researchers should have first verified – by means of an epidemiologic study – the prevalence of vitamin D deficiency to understand whether the administration of a supplement would be at best useless, if not harmful, in that kind of community and in that population group. When studies are carried out which make little sense – like the one we have just described – which are of poor quality, which are conducted with extremely variable doses and administration modes of vitamin D, and which adopt different or unknown protocols in completely different clinical conditions – in these cases, meta-analyses may produce misleading results and conclusions.
For example, the recent meta-analysis conducted by Zhao et al., published in JAMA [2], mixed together studies using D2 or D3, with doses that ranged from 400 IU/day of vitamin D to 500000 IU/year ± variable doses of calcium, in subjects with completely different or unknown vitamin D profiles and calcium intake, and with extremely variable – or worse, unknown – fracture risk conditions. It should not surprise us that the results are not statistically significant. Likewise, the attempt in this meta-analysis to rationalize the analysis by having recourse to the evaluation of a subgroup with baseline serum levels of 25(OH)D < 20 ng/mL is tainted by the fact that this fundamental datum is only available in very few studies; it was therefore mostly estimated on the basis of the dosage in a small subgroup of subjects, who were not necessarily representative of the entire population under examination.
Furthermore, available studies almost always lack verification of the 25(OH)D serum level at the end of the study in the untreatedplacebo group; as has been observed in some studies [3], the control group does not turn out to be mostly deficient, probably as a result of the widespread and common tendency today of self-managed vitamin D supplementation.