Mark J Bolland1, Andrew Grey2, Greg D Gamble2, Ian R Reid2. 1. Department of Medicine, University of Auckland, Auckland, New Zealand. Electronic address: m.bolland@auckland.ac.nz. 2. Department of Medicine, University of Auckland, Auckland, New Zealand.
Abstract
BACKGROUND: Vitamin D supplementation is often recommended to prevent falls, although vitamin D trials and meta-analyses of these trials have reported conflicting results for this outcome. We aimed to assess if there was a need for further research. METHODS: We explored the value of doing further randomised controlled trials assessing the effects of vitamin D supplements on falls with trial sequential analysis with a risk reduction threshold of 15%. All analyses were done using the numbers of participants who had a fall in intention-to-treat analyses. Trial sequential analysis performs a cumulative meta-analysis, but reduces the risk of false-positive results from repetitive statistical testing by maintaining the overall risk of type 1 error at 5%. FINDINGS: In 20 existing randomised controlled trials (n=29,535), the effect estimate for vitamin D with or without calcium on falls lay within the futility boundary, providing evidence that vitamin D supplementation does not alter the relative risk by 15% or more. In a sensitivity analysis using a risk reduction threshold of 10%, the effect estimate also lay within the futility boundary. In subgroup analyses using a risk reduction threshold of 15%, the effect estimate also lay within the futility boundary for trials of vitamin D supplementation (16 trials, n=22,291) and trials of vitamin D with calcium (six trials, n=9919). INTERPRETATION: In pooled analyses, supplementation with vitamin D, with or without calcium, does not reduce falls by 15% or more. Future trials with similar designs are unlikely to alter these conclusions. At present, there is little justification for prescribing vitamin D supplements to prevent falls. FUNDING: Health Research Council of New Zealand.
BACKGROUND:Vitamin D supplementation is often recommended to prevent falls, although vitamin D trials and meta-analyses of these trials have reported conflicting results for this outcome. We aimed to assess if there was a need for further research. METHODS: We explored the value of doing further randomised controlled trials assessing the effects of vitamin D supplements on falls with trial sequential analysis with a risk reduction threshold of 15%. All analyses were done using the numbers of participants who had a fall in intention-to-treat analyses. Trial sequential analysis performs a cumulative meta-analysis, but reduces the risk of false-positive results from repetitive statistical testing by maintaining the overall risk of type 1 error at 5%. FINDINGS: In 20 existing randomised controlled trials (n=29,535), the effect estimate for vitamin D with or without calcium on falls lay within the futility boundary, providing evidence that vitamin D supplementation does not alter the relative risk by 15% or more. In a sensitivity analysis using a risk reduction threshold of 10%, the effect estimate also lay within the futility boundary. In subgroup analyses using a risk reduction threshold of 15%, the effect estimate also lay within the futility boundary for trials of vitamin D supplementation (16 trials, n=22,291) and trials of vitamin D with calcium (six trials, n=9919). INTERPRETATION: In pooled analyses, supplementation with vitamin D, with or without calcium, does not reduce falls by 15% or more. Future trials with similar designs are unlikely to alter these conclusions. At present, there is little justification for prescribing vitamin D supplements to prevent falls. FUNDING: Health Research Council of New Zealand.
Authors: Erin D Michos; Christine M Mitchell; Edgar R Miller; Alice L Sternberg; Stephen P Juraschek; Jennifer A Schrack; Sarah L Szanton; Jeremy D Walston; Rita R Kalyani; Timothy B Plante; Robert H Christenson; Dave Shade; James Tonascia; David L Roth; Lawrence J Appel Journal: Contemp Clin Trials Date: 2018-08-20 Impact factor: 2.226
Authors: G Michael Allan; Lynda Cranston; Adrienne Lindblad; James McCormack; Michael R Kolber; Scott Garrison; Christina Korownyk Journal: J Gen Intern Med Date: 2016-03-07 Impact factor: 5.128