Background: The purpose of this pilot study was to compare estimates of percentage body fat mass (FM) by bioelectric impedance analysis (BIA) and skinfold measurements (SF) with estimates obtained from dual-energy x-ray absorptiometry (DEXA) in 47 HIV-infected male subjects receiving highly active antiretroviral therapy (HAART). As different patterns of abdominal fat accumulation might affect the body FM estimation, correlation and agreement of these methods were also compared in patients with waist to hip ratio (WHR) < or =0.9 and >0.9.
Methods: Body FM was estimated by BIA and by measuring skinfold thickness at biceps, triceps, and subscapular area, and was compared with DEXA as the reference method using paired t-test.
Results: Estimates by SF were significantly higher and by BIA were significantly lower compared with DEXA for all subjects. This relationship persisted only in those with WHR >0.9. Both BIA and SF correlated significantly with DEXA, but they did not agree. However, both techniques showed a small intermethod bias, and the precision was within the acceptable range. This relationship persisted in those with WHR >0.9. In comparison with measurement by BIA, SF showed poorer agreement (larger bias and error).
Conclusion: For population studies and perhaps to monitor changes over time for intervention studies, the bias for both BIA and SF methods is relatively small and errors and precisions are within the acceptable range when compared with DEXA, and thus all 3 techniques can be used for routine monitoring of total body FM in male subjects with HIV infection.