Statistics from Altmetric.com
Editor,—In their study of cystic fibrosis patients who were nutritionally assessed at the start and end of a 14 day period of home intravenous antibiotic treatment for chest disease, Vicet al concluded that increased weight was a result of increased fat storage.1 While this may in part be true, their methodology dictates caution in drawing conclusions.
They used bioelectrical impedance analysis (BIA) at 50 Hz to estimate body water, and four site skinfold thickness to estimate the percentage body fat in their patients. From these data and weights at day 1 and day 14, they estimated changes in fat mass by two indirect methods. The validity of this approach rests on changes in body composition being reflected in changes in BIA and skinfolds respectively. The authors detail neither the changes over their study period in BIA and skinfold thickness, nor the variability of the skinfold measurements. These data are necessary to support their assertion that body fat stores have changed.
In the course of nutritional rehabilitation the initial weight gain is often a result of increased total body water. The apparent increases in body fat in this study may be an artefact resulting from the reliance of both indirect methods on body weight at some point in the calculations. Azcue et al who performed a validation of BIA in patients with cystic fibrosis found a standard error of the estimate of total body water of 3.19 litres.2 Such a large standard error makes the interpretation of short term changes in BIA estimated body water difficult. Changes in total body water content may be better estimated from swept frequency BIA, but whether single or swept frequency techniques are used there is a need to test the assumption that short term changes in BIA reflect changes in labelled water (H2 18O) distribution. We would not therefore support the routine use of BIA in monitoring nutrition in individual cystic fibrosis patients. In experimental studies the current paper would justify the interchangeable use of BIA and anthropometry to estimate body fat when an underestimate of 0.7 kg and an overestimate of 1.1 kg is not of biological significance.