Prediction of body composition in anorexia nervosa: Results from a retrospective study

Background & aims The assessment of body composition is crucial in evaluating nutritional status in female subjects with anorexia nervosa (AN) and improving their clinical management. The aim of this retrospective study was to assess the accuracy of selected BIA (bioimpedance analysis) equations for fat-free mass (FFM) in female AN subjects and to formulate a specific equation for these subjects. Methods Eighty-two restrictive female AN subjects (age 20.5 ± 3.7 yrs, BMI 15.7 ± 1.7 kg/m 2 ) were studied. Body composition was determined with dual-energy X-ray absorptiometry (DXA) and estimated by BIA using five different equations. Linear correlation analysis was carried out to evaluate the association of FFM with selected variables. Multiple regression analysis was used to formulate specific equations to predict FFM in AN. Results All predictive equations underestimated FFM at the population level with a bias from −5.6 to −11.7%, while the percentage of accurate predictions varied from 12.2% to 35.4%. More interestingly, multiple regression analysis clearly indicates that, in addition to weight, ZI 100 or RI also emerged as independent predictors of DXA-derived FFM, increasing the prediction power of the equation well above that observed with anthropometric characteristics only. Conclusions This study shows that the selected predictive BIA equations considered exhibit an insufficient accuracy at the population and the individual level. Predictive formulas based on body weight plus BIA parameters such as RI and ZI 100 offer a rather accurate prediction of FFM (with high R squared).