Twenty-six studies were included (45,782 participants) for the analysis of falls. A further 12 studies that reported on other functional outcomes were noted but not analysed due to heterogeneity. Two studies used cluster randomisation. Allocation was concealed in 18 out of 26 studies. Patients and caregivers were blinded in 18 trials. Loss to follow-up was not reported in nine studies and where reported ranged from zero to 52% (mean 10%). Funding included for-profit resources in one third of the studies.
Vitamin D was associated with a significant reduction in risk of falls (OR 0.86, 95% CI 0.77 to 0.96; Ι²=66%; 26 studies).
Subgroup analyses to explore sources of heterogeneity found no significant subgroup-effect interactions related to patients' dwelling, Vitamin D administration route, type of Vitamin D intervention, study quality and adherence in the intervention group.
Significant interactions were reported based on patient's vitamin D status (deficient versus not deficient). These suggested a greater reduction in falls in deficient patients.
A statistically significant interaction was found between the risk of falling and calcium coadministration. This suggested that there was a greater reduction in risk of fall when both study arms received calcium. The significant benefit of vitamin D on reducing risk of falling was observed only in trials where vitamin D was given with calcium and compared with either calcium plus placebo (10 trials) or calcium alone (six trials). The benefit was not observed in trials that compared vitamin D plus calcium versus placebo alone (10 trials).
The effect of study duration on fall risk as assessed by meta-regression found no significant association.
No significant sex-fall risk interactions were found across all the studies but two studies reported within-study analyses that showed greater reduction in risk of falls in women.
Further subgroup and sensitivity analyses were reported. Publication bias may have been present.