Family Medicine Physician, Lebanon, PA
Dr. Orner reports no financial relationships relevant to this field of study.
- Vitamin D supplementation does not have clinically meaningful effects on bone mineral density nor prevention of fractures or falls.
- Using a threshold of 15%, 10%, or 7.5% relative risk reduction in falls or fractures, further similar trials are unlikely to alter this study’s conclusion.
- The effect of vitamin D doses ≥ 800 IU appears similar to doses < 800 IU/day.
SYNOPSIS: The authors of this systemic review, random-effect meta-analysis, and trial sequence analysis suggested that vitamin D supplementation does not have clinically significant effects on bone mineral density or the prevention of fractures or falls.
SOURCE: Bolland MJ, Grey A, Avenell A. Effects of vitamin D supplementation on musculoskeletal health: A systematic review, meta-analysis, and trial sequential analysis. Lancet Diabetes Endocrinol 2018;6:847-858.
Evidence from the early 2000s seemed to suggest benefit to musculoskeletal health from vitamin D supplementation.1 However, more recent studies refute this claim. Authors of a Cochrane Review published in 2014 found that taking only vitamin D was unlikely to prevent osteoporotic fractures in older adults.2 However, vitamin D supplementation combined with calcium seemed to slightly reduce the likelihood of fractures.2
Of late, data suggest that vitamin D supplementation may not prevent fractures or falls. Furthermore, supplementation may not have any clinically significant effect on bone mineral density (BMD). Bolland et al conducted a systematic review, random-effects meta-analysis, and trial sequential analysis to ascertain the effects of vitamin D supplementation on fractures, falls, and bone density.3 Funded by the Health Research Council of New Zealand, the study included 81 randomized, controlled trials that reported fractures, falls, or BMD.
This authors did not focus on coadministration of calcium and vitamin D supplements, since no large trials analyzing falls or fractures have become available since the previous systematic reviews of 2014. Nor did they evaluate dietary intake of calcium.
The researchers began by using their literature searches from previous meta-analyses published in 2014-2015. Subsequently, they searched PubMed in December 2015 for randomized, controlled trials and systematic reviews of vitamin D in adults. From these searches, they identified studies with outcomes of fractures, falls, or BMD. Finally, in September 2017 and February 2018, they conducted a search of “vitamin D” and other key words using PubMed, Embase, Cochrane Central, and two clinical trial databases.
Inclusion criteria were randomized, controlled trials in adults > 18 years of age; trials comparing vitamin D supplements with untreated controls, placebo, or lower-dose vitamin D supplements (i.e., doses < 800 IU per day); multiple intervention trials if the study groups differed only by the use of vitamin D; and quasi-randomized and open-label trials. Exclusion criteria were trials of hydroxylated vitamin D analogues, such as paricalcitol or doxercalciferol, and trials reporting BMD using techniques other than dual-energy X-ray absorptiometry.
They included randomized, controlled trials for specific disease or conditions likely to affect bone turnover but initially analyzed the data separately. If analysis results were similar, the trials were pooled in subsequent analyses. For trials comparing different dosages of vitamin D supplementation, high-dose vitamin D supplementation was considered ≥ 800 IU/day and low-dose was < 800 IU/day.
Primary endpoints were at least one fracture or fall. Secondary endpoints were the percentage change in BMD from baseline of the total body, total hip, femoral neck, forearm, and lumbar spine.
Trial sequence analysis (TSA) was completed for each outcome. Although TSA has been in use for more than 40 years, software that can perform TSA more easily is available free of charge through the Copenhagen Trial Unit, allowing more researchers to use this type of analysis.4 This cumulative meta-analysis reduces the risk of false-positive results by estimating the optimum sample size for statistical inference, thresholds for statistical significance, and thresholds for futility. This type of evaluation is useful when the number of participants in a meta-analysis may be insufficient because using a 95% confidence interval can lead to false-positive conclusions. Bolland et al noted this is the first TSA of vitamin D supplementation and BMD.
Researchers also assessed the futility boundary for the primary and secondary outcomes. Futility boundaries identify settings in which it is fruitless to continue a clinical trial or study. One example is if no useful information will be obtained by adding similar studies to a meta-analysis.5
For falls and fractures, the initial relative risk reduction threshold was 15%. Increasingly smaller thresholds were used until the optimal sample size exceeded the actual sample size. For BMD, the threshold for BMD increase initially was set at 3%, and then researchers evaluated progressively smaller thresholds.
When available, researchers also assessed the 25-hydroxyvitamin D (25(OH)D) levels reported in the studies. Of the 81 trials, 72 were completed in populations with a mean baseline 25(OH)D concentration < 50 nmol/L (20 ng/mL), while concentrations were < 25 nmol/L (10 mg/mL) in four trials. Moreover, 76 trials reported achieving a 25(OH)D concentration of ≥ 50 nmol/L. Populations in 44 studies achieved a concentration of 75 nmol/L (30 ng/mL).
In the majority of randomized, controlled trials, researchers studied vitamin D supplementation as monotherapy in populations of community-dwelling women. The preponderance of participants were 65 years of age or older. The average dose of vitamin D was > 800 IU/daily for a duration of less than one year.
An analysis of 36 trials showed that vitamin D supplementation had no effect on total fracture (n = 44,790; relative risk [RR], 1.00; 95% confidence interval [CI], 0.93-1.07). Also, supplementation had no effect on hip fracture (20 trials; n = 36,655; RR, 1.11; 95% CI, 0.97-1.26). Regarding falls, an analysis of 37 trials showed no significant effect from supplementation (n = 34,144; RR, 0.97; 95% CI, 0.93-1.02).
Bolland et al reported there was no consistent increase in BMD over the trial duration. Also, the difference in BMD between study groups (vitamin D vs. control, high-dose vs. low-dose vitamin D) at the trial endpoints was low. Specifically, the differences were: 0.13% for total body (-0.16 to 0.42), 0.34% (0.13-0.55) for total hip, 1.12% (0.58-1.65) for femoral neck, 0.25% (0.00-0.49) for lumbar spine, and -0.16% (-0.46 to 0.13) for forearm. Notably, one study that had a between-group difference of 10.6% later was excluded from analyses because it was determined to be an outlier. This led to the femoral neck difference changing to 1.12% (0.58-1.65). For the primary endpoints in this study, the estimated effect size lay within the futility boundary for RRs of 15%, 10%, and 7.5%, offering support to the determination that vitamin D supplementation does not reduce fractures or falls. Further similar studies are unlikely to change this conclusion. For BMD, the estimates for total body, total hip, and forearm were within the futility boundary for a between-group difference of 0.5%. For the lumbar spine and femoral neck, a between-group difference of 1.0% was within the futility boundary.
The most interesting part of this study was the use of TSA to determine if the addition of future studies would change the outcome. It appears that for the primary and secondary endpoints, future studies would not make a significant difference. However, there is room for further evaluation of supplementation at lower 25(OH)D concentrations, as well as combination therapy with calcium or magnesium.
Only four trials included reported baseline 25(OH)D concentrations < 25 nmol/L (10 mg/mL). In a recent study published in the Journal of Bone and Mineral Research, researchers reported significant treatment effects at the spine and hip with vitamin D 1,000 IU/day over a year in participants with a baseline 25(OH)D ≤ 30 nmol/L (12 ng/mL).6 In a study of 344 middle-aged women, researchers suggested 25(OH)D levels of 29-33 nmol/L are required for optimal musculoskeletal health.7 It would be beneficial to conduct a large meta-analysis and TSA that included more studies with baseline 25(OH)D of ≤ 30 nmol/L to determine if this population may benefit from supplementation.
It also is important to note that although the researchers did not find increased BMD differences over the trial duration between trial groups, the pooled analysis using the final trial timepoints showed differences between the trial groups. Whether these differences were present at the onset of the trials was not discussed. This appears to be a flaw in their study and report. More commentary from the researchers would be helpful.
The meta-analysis included quasi-randomized trials, which can lead to greater risk of selection bias because participant allocation to different trial arms may not truly be random and may be based on factors such as order of recruitment or medical record number.
There was no comment on the quality of the vitamin D supplements. However, it is not clear that using high-quality supplements would change the outcomes. Correspondingly, it would be thought-provoking to see research on the comparison of vitamin D supplementation vs. dietary sources of vitamin D with and/or without calcium.
Based on this trial, vitamin D supplementation would not be useful for reducing falls or fractures in those with 25(OH)D levels > 30 nmol/L. More commentary from the researchers is needed to clarify the usefulness of vitamin D supplementation on BMD. Providers should check vitamin D levels prior to starting patients on vitamin D supplementation if considering it as a preventive therapy for falls or fractures.
- Boonen S, Bischoff-Ferrari HA, Cooper C, et al. Addressing the musculoskeletal components of fracture risk with calcium and vitamin D: A review of the evidence. Calcif Tissue Int 2006;78:257-270.
- Avenell A, Mak JC, O’Connell D. Vitamin D and vitamin D analogues for preventing fractures in post-menopausal women and older men. Cochrane Database Syst Rev 2014;(4):CD000227.
- Bolland MJ, Grey A, Avenell A. Effects of vitamin D supplementation on musculoskeletal health: A systematic review, meta-analysis, and trial sequential analysis. Lancet Diabetes Endocrinol 2018;6:847-858.
- Thorlund K, Engstrøm J, Wetterslev J, et al. User Manual for Trial Sequential Analysis (TSA). Available at: http://www.ctu.dk/tsa/files/TSA_manual.pdf. Accessed Jan. 12, 2019.
- Emerson SS, Kittelson JM, Gillen DL. Stochastic Curtailment in Group Sequential Designs. Available at: https://biostats.bepress.com/uwbiostat/paper243/. Accessed Jan. 1, 2019.
- Macdonald HM, Reid IR, Gamble GD, Fet al. 25-hydroxyvitamin D threshold for the effects of vitamin D supplements on bone density: Secondary analysis of a randomized controlled trial. J Bone Miner Res 2018;33:1464-1469.
- Wu F, Wills K, Laslett LL, et al. Cut-points for associations between vitamin D status and multiple musculoskeletal outcomes in middle-aged women. Osteoporos Int 2017;28:505-515.