Body density estimates from upper-body skinfold thicknesses compared to air-displacement plethysmography

Abstract

Determine the accuracy of body density (Db) estimated with upper-body skinfold thickness (SFT) measurements compared to air-displacement plethysmography (ADP) and ascertain whether body mass index (BMI) impacts the accuracy of SFT to assess Db. We estimated Db with SFT and ADP in 131 healthy men and women with normal (N; 18.5-24.9kg/m(2)), overweight (OW; 25-29.9kg/m(2)), and obese (OB; 30-39.9kg/m(2)) BMI. Compared with ADP, SFT overestimated (p<0.05) Db in OW and OB females and in OB males (-0.0047, -0.0164 and -0.0119g/cc, respectively), and underestimated (p<0.05) Db in N females and males (0.0050 and 0.0068g/cc, respectively) but did not differently estimate Db in OW males. The gender by BMI group interaction was not significant. SFT underestimated (p<0.05; 0.0058g/cc) Db in the N and overestimated (p<0.05; 0.0113g/cc) Db in the OB BMI groups. The error in predicting Db did not change significantly over the range of Db within the N (r=0.239, p=0.06) and OB (r=0.160, p=0.934) BMI groups. Limits of agreement were -0.0165 to 0.0284g/cc and -0.0365 to 0.0085g/cc for the N and OB BMI groups, respectively. The error of estimating Db with SFT was correlated with mean Db in the aggregate sample (r=0.495, p<0.0001) and the OW group (r=0.394, p<0.009). The regression-based limits of agreement were +/-0.0226g/cc in the total group and +/-0.0168g/cc in the OW group. Although SFT offer practical advantages, the validity of SFT to estimate Db among individuals with N and OB BMI is adversely affected.