Associations between visceral adipose tissue estimates produced by near-infrared spectroscopy, mobile anthropometrics, and traditional body composition assessments and estimates derived from dual-energy X-ray absorptiometry

Assessments of visceral adipose tissue (VAT) are critical in preventing metabolic disorders; however, there are limited measurement methods that are accurate and accessible for VAT. The purpose of this cross-sectional study was to evaluate the association between VAT estimates from consumer-grade devices and traditional anthropometrics and VAT and subcutaneous adipose tissue (SAT) from dual-energy X-ray absorptiometry (DXA). Data were collected from 182 participants (female = 114; White = 127; Black/African-American (BAA) = 48) which included anthropometrics and indices of VAT produced by near-infrared reactance spectroscopy (NIRS), visual body composition (VBC) and multifrequency BIA (MFBIA). VAT and SAT were collected using DXA. Bivariate and partial correlations were calculated between DXA(VAT) and DXA(SAT) and other VAT estimates. All VAT indices had positive moderate-strong correlations with VAT (all P < 0 center dot 001) and SAT (all P < 0 center dot 001). Only waist:hip (r = 0 center dot 69), VAT(VBC) (r = 0 center dot 84), and VAT(MFBIA) (r = 0 center dot 86) had stronger associations with VAT than SAT (P < 0 center dot 001). Partial associations between VAT(VBC) and VAT(MFBIA) were only stronger for VAT than SAT in White participants (r = 0 center dot 67, P < 0 center dot 001) but not female, male, or BAA participants individually. Partial correlations for waist:hip were stronger for VAT than SAT, but only for male (r = 0 center dot 40, P < 0 center dot 010) or White participants (r = 0 center dot 48, P < 0 center dot 001). NIRS was amongst the weakest predictors of VAT which was highest in male participants (r = 0 center dot 39, P < 0 center dot 010) but non-existent in BAA participants (r = -0 center dot 02, P > 0 center dot 050) after adjusting for SAT. Both anthropometric and consumer-grade VAT indices are consistently better predictors of SAT than VAT. These data highlight the need for a standardised, but convenient, VAT estimation protocol that can account for the relationship between SAT and VAT that differs by sex/race.