Vitamin D, in high doses, prevents falls

Clinical question Can supplementation with vitamin D prevent falls in older people?

Bottom line Dosages of 700 IU to 1,000 IU daily of vitamin D will prevent one additional fall for every 11 patients who take it regularly. Lower doses are not effective, and active forms are no more effective. Given its price, availability, and lack of side effects, vitamin D (with or without calcium) should be a regular supplement in older patients because approximately 1 in 3 older people will fall each year. (Level of evidence = 1a)

Synopsis Vitamin D, in addition to its effect on bone strength, also affects muscle strength and, in patients at risk for vitamin D deficiency, increases strength, function, and balance. The authors of this meta-analysis combined the results of eight randomized controlled trials that enrolled a total of 2,426 older persons. They identified these studies by searching four databases (including the Cochrane central register), by consulting experts, and by reviewing the bibliographies of identified studies. They did not find evidence of publication bias. They excluded studies that were not randomized, were short term, studied intramuscular vitamin D, or did not assess falls during the entire study period. Two researchers independently abstracted the data. Five of the eight studies also gave supplemental calcium to patients in both treatment groups. Heterogeneity was found but was removed when high-dose and low-dose studies were considered separately. High dosages (700-1,000 IU/d) of vitamin D decreased fall risk by 19% over 2 to 36 months of treatment; one additional fall was prevented for every 11 patients given high-dose vitamin D instead of placebo (number needed to treat = 11; 95% CI, 7-20). Lower doses of vitamin D did not reduce falls. Active forms of vitamin D produced similar results as high-dose supplemental vitamin D.

Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, et al. Fall prevention with supplemental and active forms of vitamin D: meta-analysis of randomised controlled trials. BMJ. 2009;339:b3692. doi:10.1136/bmj.b3692.