Bioelectric impedance analysis versus prediction equations for percent body fat in healthy nigerian adults: correlation or conflict

Background and objective: There is dearth of studies on the validation of equations for predicting Percent Body Fat (PBF) in sub-Saharan Africa. This study compared Bio-electrical Impedance Analysis (BIA) estimated PBF and PBF predicted from equations. Materials and Methods: 1,350 volunteers whose ages ranged between 18 and 65 years participated in this study. Anthropometric parameters and indices were assessed following standardized procedures. BIA was used to estimate PBF based on standard formulae. Deurenberg -#1, Deurenberg-#2, Gallagher and Jackson-Pollock equations were used to compute PBF. Data was analyzed using descriptive and inferential statistics. Alpha was set at 0.05. Results: The mean values for PBF obtained from BIA, Deurenberg #1, Deurenberg #2, Gallagher, Jackson-Pollock and the present study’s equations were 21.5±9.3, 11.5±6.76, 9.3±7.09, 9.4±7.38, and 7.9±7.87. BIA estimated PBF was significantly higher than predictive PBF (p= 0.001). BIA estimated PBF was significantly correlated with each of Deurenberg #1 (r = 0.122; p = 0.001); Deurenberg-#2 (r=0.102; p=0.001); Gallagher (r=0.080; p=0.001); Jackson-Pollock (r=0.060; p=0.016) and the present study’s (r=0.997; p=0.001) equations. Socio-demographic and anthropometric variable except WHR were significantly correlated with BIA estimated PBF (p<0.05). Age, sex and BMI could significantly predict PBF at 68.3% level of variability. Conclusion: There were significant inverse but weak relationships between BIA estimated PBF and each of Deurenberg-#1, Deurenberg-#2, Jackson-Pollock and Gallagher prediction equation. The BIA estimated PBF was significantly higher than predicted PBF. Socio-demographic and anthropometrical variables were significantly determinants of BIA estimated PBF. The prediction equation derived from this study provides an easy method of estimating PBF