[Effect of refeeding on the body composition of females with restrictive anorexia nervosa; anthropometry versus bioelectrical impedance].

OBJECTIVE

To assess the body composition in a group of malnourished patients with anorexia nervosa as compared to healthy controls, before and after nutritional support, by means of anthropometry and bioimpedance.

METHODS

Prospective observational study. Complete anthropometry was performed as well as bioimpedance analysis in 12 women (24.5 years) with restrictive anorexia nervosa at hospital admission and weekly thereafter during re-feeding. The control group was formed by 24 healthy women (21 years). The Student's t test, the Mann-Whitney U test, the Student's t test for repeated measurements, and the Wilcoxon's test were applied. The level of agreement between anthropometry and BIA was calculated by the interclass correlation coefficient and the Bland-Altman s test.

RESULTS

The patients had significant improvements in all indexes of body composition throughout their hospital staying although their values at discharge still were lower than those of the control subjects. The average weight gain was 5.22 kg (SD: 1.42), of which 51.4% was fat mass, preferentially centrally distributed. In the control subjects, the BIA equation that correlated the best with anthropometrics was Sun s equation (CCI = 0.896); in the patients, the level of agreement was weaker, both at hospital admission and at hospital discharge.

CONCLUSIONS

Re-feeding produces weight gain, essentially at the expense of fat mass, which is centrally distributed; the nutritional status is not reestablished. The level of agreement between anthropometry and bioimpedance for studying body composition is acceptable, especially in healthy subjects. In those cases with severe changes in body composition and/or water balance, anthropometry is recommended when vectorial BIA or some other gold standard method are not available for the analysis of body composition.