If you've been following along, you'll understand why. Towards the end of May, the US Preventive Services Task Force announced that vitamin D supplementation was a good thing in that it reduced falls by 17% in community-dwelling elderly at risk for falls. Then last week, this same USPSTF announced that vitamin D wasn't such a good thing when they didn't recommend low dose vitamin D to prevent fracture in non-institutionalized post-menopausal women.
retrospective, observational cohort study that found a reverse J-curve association between vitamin D & all-cause mortality. The authors followed 247,574 Danish participants for 3 years and noted that lowest mortality was found at 25OH vitamin D of 50-60nmol/L, which is equivalent to 20-24ng/mL. Those with 25OH vitamin D of just 10nmol/L (or 4ng/mL) had over twice the relative risk of all-cause mortality, while those w/25OH vitamin D of 140nmol/L (or 56ng/mL) had 42% greater risk, both compared to those w/50nmol/L.
Of course, this study really muddles things up b/c the level at which the Danes found lowest mortality is typically considered deficient (20-29ng/mL) here in the States, w/normal reference range being 30-100ng/mL. Could it be that normal isn't necessarily optimal? So what are we to make of this study's conclusions? Even the authors readily admitted that "the study did not allow inference of causality", their fancy way of saying that proof of correlation is not proof of causation.
Before we jump to conclusions and hit the ball back over the net, let's take a look at study #2.