Translate Register Log In Login with facebookLogin and Register

Falls are prevented by vitamin D - provided you get enough - Dec 2014

Study, comment, reply to the comment, then VitaminDWiki references

Differences in Overlapping Meta-Analyses of Vitamin D Supplements and Falls

Journal of Clinical Endocrinology and Metabolism, DOI: http://dx.doi.org/10.1210/jc.2014-2562
Received: June 04, 2014, Published Online: August 05, 2014
Mark J. Bolland, Andrew Grey, and Ian R. Reid
Department of Medicine, University of Auckland, Auckland 1142, New Zealand

Overlapping meta-analyses on the same topic are common and often report discordant results. An Endocrine Society (ES) meta-analysis reported that vitamin D with calcium reduced the risk of falls, whereas vitamin D monotherapy had no effect. Despite meta-analyzing an overlapping set of trials, we concluded that vitamin D with or without calcium had no effect on falls.
The objective of the study was to determine the reasons for the different conclusions from the two meta-analyses.
We extracted raw data from the 25 randomized controlled trials included in the ES meta-analysis, calculated the treatment effect for each trial, compared them with the published ES meta-analysis, and determined the reason for any differences.
Of the 25 trials, there were differences in 14 results (56%) between the two meta-analyses. In the ES meta-analysis, data were used from a subset of falls or participants (four trials), from trial completers (three trials), or were imputed from fracture data (one trial). Other differences resulted from use of adjusted results from original papers (two trials), differences in incorporating data from multiarm and factorial studies (three trials), and inconsistent group numbers reported in original papers (two trials). In a reanalysis of the ES meta-analysis using unadjusted analyses and all randomized participants, there was a marginal effect of vitamin D with or without calcium on falls [relative risk (RR) 0.95, 95% confidence interval (CI) 0.90–1.00], with the subgroup analysis showing no effect of vitamin D monotherapy (RR 1.00, 95% CI 0.92–1.08) or vitamin D with calcium vs the placebo/controls (RR 0.95, 95% CI 0.88–1.02), but a modest effect of vitamin D with calcium vs calcium (RR 0.84, 95% CI 0.76–0.92) was shown.
Methodological differences in utilizing data from the same trials directly led to substantially different conclusions between meta-analyses about the efficacy of vitamin D supplements on falls. More detailed reporting of meta-analyses is necessary to allow readers to understand the discordant results from overlapping meta-analyses.

Vitamin D and Falls: Treat the Deficiency (and the Ones at Risk)

K. Amrein 1, S. Pilz1, and H.A. Bischoff-Ferrari 2,3
Published online on December 3, 2014
1 Department of Internal Medicine, Division of Endocrinology and Metabolism, Medical University of Graz, Austria
2 Geriatric Clinic, University Hospital of Zurich, Zurich, Switzerland
3 Centre on Aging and Mobility, University Hospital of Zurich and City Hospital Waid, Zurich Switzerland

We read with interest the recent analysis by Bolland and coworkers on vitamin D and falls (1). The authors concluded that methodological differences were responsible for substantially different conclusions between their own and the Endocrine Society’s 2011 meta-analysis (1,2). We agree with the authors that for vitamin D supplementation in general, and falling as an outcome particularly, methods and the selection of trials included in a meta-analysis matter. Earlier meta-analyses have demonstrated a protective effect of vitamin D in double-blind RCTs where falls as the endpoint were pre-defined and assessed throughout the trial (3-5). In a selection of 8 high quality trials (5), a significant 27% reduction in the odds of falling was demonstrated overall (odds ratio 0.73 (95% confidence interval 0.62 to 0.87)), while trials that provided a higher dose of vitamin D (700 to 1000 IU) reduced the odds of falling even more pronounced by 34% (odds ratio 0.66 (0.53 to 0.82)), while no reduction was seen with low dose vitamin D (odds ratio 1.14 (0.69 to 1.87)).

The authors chose an alternative comprehensive approach aiming to include any trial irrespective of its quality and, not surprisingly, could not show an overall benefit. Most of the analyzed 25 RCTs (1) did not define how falls were assessed or did not assess falls throughout the trial period, and 20% of trials did not have a double-blind design. Further, 40% of RCTs either had no measure of 25-hydroxyvitamin D status at baseline, or 25-hydroxyvitamin D levels were only measured in a small subgroup of the study population (typically less than 5%). Thus, we do not know whether these individuals started in a replete vitamin D state or received sufficient vitamin D to expect a benefit. In addition, when reported, compliance with supplementation was often poor. Finally, the authors included five RCTs with oral or injected large vitamin D doses that were given once-only or yearly, which may need to be assessed separately as those regimens are not supported by recent guidelines on fall prevention with vitamin D.
Future high-quality research is needed to clarify individual needs for vitamin D beyond the repletion of vitamin D deficiency. This should be addressed in multi-dose trials that ideally target vitamin D deficient individuals and measure baseline and follow-up 25(OH)D status in all participants.
1.Bolland MJ, Grey A, Reid IR. Differences in overlapping meta-analyses of vitamin D supplements and falls. J Clin Endocrinol Metab 2014:jc20142562.
2.Murad MH, Elamin KB, Abu Elnour NO, et al. Clinical review: The effect of vitamin D on falls: a systematic review and meta-analysis. J Clin Endocrinol Metab 2011;96:2997-3006.
3.Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, et al. Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials. BMJ 2009;339:b3692.
4.Ross AC, Taylor CL, Yaktine AL, Del Valle HB. Dietary Reference Intakes for Calcium and Vitamin D: Institute of Medicine; 2011.
5.Bischoff-Ferrari HA, Willett WC, Orav EJ, Kiel DP, Dawson-Hughes B. Re: Fall prevention with Vitamin D. Clarifications needed. http://www.bmj.com/content/339/bmjb3692?tab=responses (access: Feb132012) 2011.
Conflict of Interest:
KA: speaker honoraria and investigator initiated funding by Fresenius Kabi
PZ: none
HAB-F: speaker honoraria and independent and investigator initiated funding by MSD, DSM, Nestle, Roche Diagnostics, WILD, and Pfizer

Response to Amrein et al.

Mark Bolland, Andrew Grey, and Ian Reid
University of Auckland
Published online on December 3, 2014

Amrein and colleagues contrast the discordant results from an earlier meta-analysis of vitamin D on falls by Bischoff-Ferrari (1) with our recent meta-analysis (2). A recommended approach to assess discordant overlapping meta-analyses is to consider 6 domains- the clinical question asked, study selection and inclusion, data extraction, study quality, ability to combine studies, and statistical methods (3). The two meta-analyses differed in important ways in 3 of those domains.

Study selection: 14/15 trials included in the Bischoff-Ferrari meta-analysis were included in our meta-analysis, with 1 cluster randomized controlled trial excluded (4). 14/20 trials included in our meta-analysis were included in the Bischoff-Ferrari meta-analysis: 2 trials were excluded because of intramuscular administration (5, 6), 1 study was excluded because of its short duration (7), 1 study was not included although it appears to meet the inclusion criteria (8), and 2 trials were published after the Bischoff-Ferrari analysis (9, 10). Another cluster randomized controlled trial we excluded also seems to meet the inclusion criteria for the Bischoff-Ferrari analysis (11). The rationale for the exclusion by Bischoff-Ferrari of trials of intramuscular administration is weak- studies were excluded because the “technique is invasive and has resulted in small but variable increases in 25(OH)D concentrations” (1).

Data extraction: Of the 14 trials included in both meta-analyses, the extracted data differed for 6/14 (43%) of trials. We think the differences are due to use by Bischoff-Ferrari of data only from participants that completed the trial, rather than from all randomized participants, for 2 studies (12, 13), rounding for 1 study (14), inconsistent data in the primary papers for 2 studies (15, 16), and for reasons we could not determine for 1 study (17).

Ability to combine studies: In our pooled analysis of all 20 trials the relative risk (RR) was 0.96 (95% CI 0.91-1.01). If we restrict the analysis to the 15 studies included in the Bischoff-Ferrari analysis and the missed study, the pooled RR is 0.95 (0.89-1.00), compared to the pooled result of all trials reported by Bischoff-Ferrari, 0.92 (0.85-0.99).

Bischoff-Ferrari then carried out a primary analysis of a subset of 8 trials they termed “high quality”. These trials were double blind, had falls pre-specified as an endpoint, included a definition of falls and how they were assessed, and assessed falls for the entire trial period. However, 2 of the 8 trials did not meet these criteria- in one, falls were not pre-specified (18), and in another falls were only assessed in a subset of participants for a short part of the trial (14). A further trial that appears eligible for inclusion in the primary analysis was excluded because “patients were in unstable health” (13), seemingly because the trial recruited participants admitted to geriatric rehabilitation units, either inpatient or day wards, although 90% were living in the community. However, health stability was not stated in the list of criteria for inclusion in the primary analysis. Furthermore, other trials that were included in the primary analysis recruited patients from institutions (14, 18-20) who likely had similar or poorer levels of health to those in the excluded trial. When the two ineligible trials are excluded from the primary analysis by Bischoff-Ferrari, and the excluded trial included, the pooled result from 7 trials is RR 0.86 (0.76-0.97), compared with the result reported in the paper RR 0.81 (0.71-0.92). After adding 2 later eligible trials (9, 10), the pooled result is RR 0.91 (0.81-1.03).

Thus, we think the conclusions of our analysis (21) of the Endocrine Society meta-analysis (22) also apply to the analyses carried out by Bischoff-Ferrari and colleagues. Namely, that methodological differences between meta-analyses favor positive effects of vitamin D on falls.
Our analyses show that existing trials of vitamin D supplements, whether restricted to “high quality” trials or including larger pragmatic trials do not provide evidence for a meaningful clinical effect of vitamin D on the risk of falling.

Amrein and colleagues then argue that vitamin D measurements in current trials are inadequate. This oft-repeated claim is incorrect: in our analysis 16 of 20 trials reported baseline 25OHD levels, and 17 of 20 trials reported 25OHD on treatment (2). Measurements in a sample of participants is all that is required if the purpose is to define the vitamin D status of the cohort either prior to or during treatment. Individual measurements for the entire study population are only needed if there are inclusion or exclusion criteria based on 25OHD levels, if supplementing to a target 25OHD is part of the study design, or if sub-group analyses based on baseline or achieved 25OHD levels are planned.

Even if it produced positive results, a trial that mandates 25OHD baseline and follow-up measurements for all participants would have limited generalizability. A strategy of measuring 25OHD in all participants at risk of vitamin D deficiency, followed by supplementation with vitamin D of those with low 25OHD levels and confirming that 25OHD levels reach a target level with supplementation is unlikely to be cost-effective or feasible at a population level. The contrasting strategy of supplementation with vitamin D of individuals who are at risk of vitamin D deficiency without measuring 25OHD has been tested in many trials. Current evidence excludes a clinically important benefit on falls, fracture, vascular disease and cancer in trials that tested this strategy, and further trials or meta-analyses are unlikely to alter that conclusion (2, 23).
1.Bischoff-Ferrari HA, Dawson-Hughes B, Staehelin HB, Orav JE, Stuck AE, Theiler R, Wong JB, Egli A, Kiel DP, Henschkowski J 2009 Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials. BMJ 339:b3692.
2.Bolland MJ, Grey A, Gamble GD, Reid IR 2014 Vitamin D supplementation and falls: a trial sequential meta-analysis. Lancet Diabetes Endocrinol 2:573-580.
3.Jadad AR, Cook DJ, Browman GP 1997 A guide to interpreting discordant systematic reviews. CMAJ 156:1411-1416.
4.Law M, Withers H, Morris J, Anderson F 2006 Vitamin D supplementation and the prevention of fractures and falls: results of a randomised trial in elderly people in residential accommodation. Age Ageing 35:482-486.
5.Dhesi JK, Jackson SH, Bearne LM, Moniz C, Hurley MV, Swift CG, Allain TJ 2004 Vitamin D supplementation improves neuromuscular function in older people who fall. Age Ageing 33:589-595.
6.Smith H, Anderson F, Raphael H, Maslin P, Crozier S, Cooper C 2007 Effect of annual intramuscular vitamin D on fracture risk in elderly men and women- - a population-based, randomized, double-blind, placebo-controlled trial. Rheumatology (Oxford) 46:1852-1857.
7.Burleigh E, McColl J, Potter J 2007 Does vitamin D stop inpatients falling? A randomised controlled trial. Age Ageing 36:507-513.
8.Porthouse J, Cockayne S, King C, Saxon L, Steele E, Aspray T, Baverstock M, Birks Y, Dumville J, Francis R, Iglesias C, Puffer S, Sutcliffe A, Watt I, Torgerson DJ 2005 Randomised controlled trial of calcium and supplementation with cholecalciferol (vitamin D3) for prevention of fractures in primary care. BMJ 330:1003.

See also VitaminDWiki

The TOP articles in Fractures and Vitamin D are listed here:

Meta-analysis of Vitamin D category listing has these comments of problems with meta-analyses with Vitamin D
1) Dose sizes are ignored: Studies assume benefit from 400 IU is the same as for 4,000 IU
2) Durations are ignored: Studies assume benefit of 1 month dosing is the same as 1 year dosing
3) Existing Vitamin D levels are often ignored.
Studies ignore that a 20 ng addition to a 10ng person has more benefit than a 20 ng addtion to a 40 ng person

See any problem with this page? Report it (FINALLY WORKS)