Vitamin D and Health

6.142 Another meta-analysis (Murad et al., 2011) of 26 trials (n=45,782; mean age, 76y) reported vitamin D
supplementation significantly reduced fall risk (OR of at least 1 fall=0.86; 95% CI, 0.77-0.96) but noted
substantial heterogeneity across studies (I 2 =66%). Subgroup analysis showed risk reduction was
greater in participants who were considered vitamin D deficient 58 at baseline but the mean serum
25(OH)D concentration in this subgroup was not specified. Risk reduction was also greater when
calcium was co-administered (vitamin D + calcium vs placebo, 10 trials, OR=0.83; 95% CI=0.72-0.93;
vitamin D + calcium vs calcium, 10 trials, OR=0.63; 95% CI, 0.50-0.81). Vitamin D alone vs placebo did
not reduce risk of fall reduction (OR=0.97; 95% CI, 0.84-1.11). The authors concluded that vitamin D
with calcium reduces the risk of falls however publication bias had exaggerated the estimates of risk
reduction.
6.143 A meta-analysis of 20 trials (n=29,535; mean age range, 71-89 y) (Bolland et al., 2014) reported no
effect of vitamin D supplementation, with or without calcium, on risk of falls (RR=0.96; 95% CI, 0.91-
1.01). Subgroup analyses did not find significant interactions between baseline or achieved serum
25(OH)D concentration and fall risk. Sixteen trials reported serum 25(OH)D concentration (mean
range, 22-75 nmol) with baseline serum concentration < 50 nmol/L in 12 trials. All trials reported
increases in serum 25(OH)D concentration and 14 trials reported that serum 25(OH)D concentration
increased to > 50 nmol/L in the vitamin D intervention group. The need for further randomised trials
on effects of vitamin D supplements on falls was assessed using trial sequential analysis 59 with a risk
reduction threshold of 15%. In the 20 RCTs included in the analysis, the effect estimate for vitamin D
supplementation (+/- calcium) on falls lay within the futility boundary, indicating that it does not alter
the relative risk of falls by 15% or more.
Intervention studies
6.144 An RCT of community-dwelling women (n=2256; median age 76y) randomised to receive a vitamin D
supplement (12,500 µg/500,000 IU) or placebo on an annual basis for 3-5 years (Sanders et al., 2010)
reported a significant increase in rate of falls (RaR=1.15; 95% CI, 1.02-1.30) and fractures (RR=1.26,
95% CI 1.00-1.59) in the vitamin D supplemented group compared to placebo. A post hoc analysis
indicated that the excess falls and fractures occurred in the 3 months after dosing, when median
serum 25(OH)D concentration was approximately 120 nmol/L after 1 month and 90 nmol/L after 3
months. This RCT used very high supplementation doses (provided as annual bolus) which may
explain why effects differed from those observed with daily supplementation.
6.145 In another RCT (Bischoff-Ferrari et al., 2016), community-dwelling men and women (n=200; mean age,
78y with a prior fall) were randomised to receive a monthly dose of either 600 µg (24,000 IU) vitamin
D 3 , 1500 µg (60,000 IU) vitamin D 3, or 600 µg (24,000 IU) vitamin D 3 + 300 µg 25(OH)D 3 for 12 months.
The primary outcome was improvement in lower extremity function and achieving serum 25(OH)D
concentrations ≥ 75 nmol/L; a secondary outcome was monthly reported falls. The groups receiving
1500 µg (60,000 IU) vitamin D 3 and 600 µg (24,000 IU) vitamin D 3 + 300 µg 25(OH)D 3 were more likely
to achieve serum 25(OH)D concentrations ≥ 75 nmol/L than the 600 µg (24,000 IU) vitamin D 3 alone
group. There was no difference between groups in lower extremity function but the incidence of falls
was significantly higher in the 1500 µg (60,000 IU) vitamin D 3 group (66.9%; 95% CI, 54.4-77.5%) and
58
Studies categorised as having a vitamin D-deficient vs not deficient population based on: author description; baseline serum 25(OH)D
concentration or inclusion of participants with at least 2 vitamin D deficiency risk factors (old age, dark skin, living in a nursing home, living far from
the equator, winter season, sunscreen use, wearing a veil, smoking, obesity, malabsorption disease, renal or liver disease, and use of medication.
59 Trial sequential analysis permits estimation of the point at which the evidence base is large and consistent enough to make further trials futile
because of the low probability that they will affect results of existing meta-analyses (Wetterslev et al, 2008).
65