Associations among 25-hydroxyvitamin D, diet quality, and metabolic disturbance differ by adiposity in adults in the United States

Abstract

Context: Recent evidence indicates that a higher plasma level of 25-hydroxyvitamin D [25(OH)D] is associated with lower adiposity and a reduced number of metabolic disturbances (MetD).

Objectives: We examined associations among dietary quality, 25(OH)D, percent body fat (%BF), and MetD, and a pathway linking them, across central obesity.

Design: This cross-sectional nationally representative study used extensive data from the National Health and Nutrition Examination Surveys of 2001-2004.

Participants: U.S. adults aged at least 20 yr were stratified by central obesity (CO) status. Sample sizes ranged from 1943 (all MetD combined) to 7796 (each component).

Main outcome measures: %BF was measured using dual-energy x-ray absorptiometry, and MetD was measured with individual continuous nonadiposity outcomes (e.g. fasting plasma glucose) and with a composite count index of binary MetD with prespecified cutoff points (Index I).

Results: A higher 25(OH)D was associated with better dietary quality, lower %BF, and lower number of MetD. These inverse 25(OH)D-%BF and 25(OH)D-MetD associations (i.e. fasting blood glucose, homeostatic model assessment of insulin resistance, C-reactive protein, and Index I) were significantly stronger among the CO+ group. Finally, the pathway linking the dairy component of the Healthy Eating Index (HEIdairy) to Index I through 25(OH)D and %BF indicated complete mediation among the CO- group, but HEIdairy and 25(OH)D had direct inverse associations with Index I among the CO+ group.

Conclusions: Due to potential genetic differences between CO- and CO+ groups, empowering U.S. adults with central obesity to make related behavioral changes may be especially effective in improving their vitamin D status and metabolic profile.

Figure 1

Flow diagram for study sample selection, NHANES 2001–2004.

Figure 2

SM stratified by central obesity (CO) status: testing mediation of HEI-Index I MetD through 25(OH)D and %BF. A theoretical model was set up in which sociodemographic and lifestyle factors were exogenous to the main variables of interest. The model consisted of simultaneous equations with outcomes being HEI, 25(OH)D, %BF, and number of MetD (Index I range between 0 and 7), stratified by central obesity status. HEI_total and HEI_dairy were considered alternatively. Effect modification by central obesity status was determined for the main associations of interest by comparing 95% CI of path coefficients. The following equations correspond to the general setup of the structural equations model: 1 2 3 4 where HEI (total or dairy component; z-score) was allowed to predict all subsequent outcomes, plasma 25(OH)D level (vitamin D; z-score) was allowed to predict only %BF and number of MetD (Index I; ordinal), and BF was allowed to predict only MetD. Z_j is the vector of other exogenous variables (e.g. sociodemographic and lifestyle factors; continuous terms such as age were entered as continuous), whereas e₁, e₂, e₃, and e₄ are the error terms. Correlations between all exogenous variables except error terms (assumed to be zero) were estimated.

Figure 3

Mean difference in 25(OH)D by central obesity status (CO⁺ or CO⁻): OLS multiple linear regression findings (β with 95% CI), adjusting for sociodemographic, lifestyle, dietary factors, and % BF (NHANES 2001–2004). *, P < 0.05 for null hypothesis that adjusted β = 0 from OLS multiple linear regression, Wald test. For model 1, 25(OH)D is regressed on central obesity, adjusted for age and sex (n = 8184); model 2 adjusted for all sociodemographic, lifestyle, and dietary variables (see Covariates in Materials and Methods) (n = 7300); and model 3 further adjusted for %BF (n = 7278).