Evaluation of body fat changes during weight loss by using improved anthropometric predictive equations

Background/Aim: Skinfold-based equations are widely used to evaluate body fat (BF), but over-/underestimation is often reported. We evaluate the capacity of improved skinfold-based equations to estimate BF changes during weight reduction and compare them against well-established equations. Methods: Overweight adults (n = 44) participated in a 4-month weight reduction intervention. Dual-energy X-ray absorptiometry (DXA) and anthropometric measurements were taken at baseline and after intervention. The BF% was calculated using Garcia, Peterson, and Durnin and Womersley (DW) equations. Results: Baseline and postintervention BF% measured by DXA correlated highest with BF% predicted according to Garcia (r = 0.934 and r = 0.948, respectively), followed by Peterson (r = 0.941 and r = 0.932, respectively) and DW (r = 0.557 and r = 0.402, respectively); only a slight systematic error in overestimating the BF% was observed in estimates according to Garcia (r = 0.147 and r = 0.104, respectively; p < 0.001), while increasing errors occurred using the Peterson (r = 0.624 and r = 0.712, respectively; p < 0.001) and DW (r = 0.767 and r = 0.769, respectively; p < 0.001) equations. Moderate correlations between BF changes (kg) measured by DXA and predicted by DW (r = 0.7211), Peterson (r = 0.697), and Garcia (r = 0.645) were observed. Conclusion: Improved skin-fold equations cannot accurately measure changes in BF after weight reduction.