Exercise without Weight Loss Prevents Seasonal Decline in Vitamin D Metabolites: The VitaDEx Randomized Controlled Trial

Abstract

Many adults become vitamin D deficient or insufficient during winter at northerly latitudes when cutaneous vitamin D synthesis does not occur. Vitamin D accumulates in adipose tissue and people with overweight or obesity are more likely to have low systemic vitamin D. This randomized controlled trial demonstrates that regular exercise completely maintains serum concentrations of the active vitamin D metabolite 1,25(OH)₂D₃ over winter and may ameliorate the decline in 25(OH)D status in overweight men and women, even without weight loss. The binding of 1,25(OH)₂D₃ to the vitamin D receptor mediates the crucial role for vitamin D in the healthy function of multiple organ systems and vitamin D supplementation does not impact circulating 1,25(OH)₂D₃. Thus, the VitaDEx study provides causal evidence that exercise plays an important role in vitamin D metabolism that is distinct from the effects of oral supplementation.

Keywords: 1,25(OH)D; Exercise; VitaDEx; Vitamin D; Weight Stable; Winter.

Figure 1

CONSORT flow diagram. FMI; fat mass index, PA; physical activity, PAL; physical activity level as a ratio of total to resting energy expenditure, PAR‐Q; Physical Activity Readiness Questionnaire.

Figure 2

Responses of a,c,e) serum total 25(OH)D and b,d,f) 1,25(OH)₂D₃ with the control group in blue and exercise group in orange throughout. The mean ± SD change over 10 weeks of winter for serum concentrations of total 25(OH)D and 1,25(OH)₂D₃ are displayed in (a) and (b), respectively. Individual participant changes from baseline‐to‐post intervention, with mean [95% CI], are displayed in (c) and (d), with cut‐offs for vitamin D insufficiency and deficiency based on serum 25(OH)D are demarcated in (c). Cohen's d effect size [bootstrapped 95% CI]^{[ 27 ]} is presented in (e) and (f) for the difference in baseline‐to‐post‐intervention change in total 25(OH)D concentration between groups with the control group in blue and exercise group in orange and means represented as p _x = p value for interaction effects, and p _t = p value for the main effect of time, comparing the change over time between groups with LMEM.

Figure 3

Responses of serum vitamin D metabolites, ratios, and half‐life. Mean with individual data points are presented for serum a) 25(OH)D₃, b) 25(OH)D₂, c) 24,25(OH)₂D₃, d) vitamin D₃, e) vitamin D metabolite ratios 25(OH)D₃:24,25(OH)₂D₃, and f) 1,25(OH)₂D₃:24,25(OH)₂D₃, g) serum Epi‐25(OH)D₃, and calculated bioavailable h) 25(OH)D and i) free 25(OH)D. Individual (Control n = 11, Exercise n = 14) time course responses over the four weeks to ingestion of 60 nmol of labeled d₃‐25(OH)D₃ j) immediately prior to the intervention or k) during the last four weeks of the intervention from which the d₃‐25(OH)D₃ half‐life was estimate by linear regression. l,m) Mean with individual data points, and individual participant responses from baseline‐to‐post‐intervention are displayed for d₃‐25(OH)D₃ tracer half‐life control group in blue and exercise group in orange. Free and bioavailable 25(OH)D data were only available for n = 20 for the Exercise group due to missing albumin data for one participant. p _x = p value for interaction effects, and p _t = p value for the main effect of time, comparing the change over time between groups with LMEM.

Figure 4

Adipose tissue vitamin D concentration and gene expression. Mean ± SD is presented with individual participant responses from baseline‐to‐post‐intervention for a) adipose tissue 25(OH)D and b) vitamin D₃ concentrations with control group in blue and exercise group in orange. c) The expression of adipose tissue genes involved in vitamin D metabolism is presented as mean ± SD at baseline and post‐intervention from n = 16 control and n = 16 exercise participants, d) as identified based on the GO Terms listed. Genes in red text were not expressed in adipose. p _x = p value for interaction effects, and p _t = p value for the main effect of time, comparing the change over time between groups with LMEM.

Figure 5

Exploratory analysis of relationships between concentration and change in vitamin D metabolites. Heat maps displaying the strength of Pearson's correlations as r values for concentrations and change scores (Δ) of vitamin D metabolite concentrations measured in serum and adipose tissue for the a) control group and b) exercise groups calculated using the Oldham method.^{[ 28 ]} Group level correlations (control group in blue and exercise group in orange) are displayed for correlations that differed between groups according to Fisher's z and Zou's 95% confidence intervals^{[ 30 ]} in (c) for change in serum 25(OH)D, with total serum 25(OH)D, serum vitamin D₃, 24,25(OH)₂D_3, 3‐Epi‐25(OH)D₃, and adipose 25(OH)D₃, and also change in 24,25(OH)₂D₃ with 24,25(OH)₂D₃, i.e., instance where there was a stronger negative correlation in the control group, and in (d) for change in serum vitamin D₃ with total serum 25(OH)D, serum vitamin D₃, 24,25(OH)₂D_3, 3‐Epi‐25(OH)D₃, and adipose D₃, i.e., where there were stronger negative correlations in the exercise group compared to the control group.

Figure 6

A schematic overview of the study design for participants completing the tracer assessment of whole‐body vitamin D expenditure rate. For participants not undertaking this measurement, the trial was completed in six or seven laboratory visits, matching the same timeline for measurement and intervention periods. DXA, dual‐energy X‐ray absorptiometry; PA, physical activity monitoring with Actiheart 5; pQCT, peripheral quantitative computed tomography; FAT_MAX, treadmill‐based incremental exercise test.