Comparison of different body composition models in acromegaly.

Comparison of different body composition models in acromegaly. Brummer RJ, Lonn L, Bengtsson BA, Kvist H, Bosaeus I, Sjostrom L. Department of Clinical Nutrition, Sahlgrenska Hospital, Goteborg, Sweden. The aberrant body composition of 10 patients with active acromegaly was used to evaluate the validity and limitations of several models and methods to assess body composition. Body composition was determined using either a two-compartment model, dividing the body in a body fat (BF) compartment and a fat-free mass (FFM) compartment, or a four-compartment model in which the FFM compartment comprises the three following components: body cell mass, extracellular water and the fat-free extracellular solids. The measurement techniques consisted of anthropometry, bioelectrical impedance analysis (BIA)-applying various established regression equations-tritiated water dilution, whole body 40K-counting, and whole body computed tomography (CT). This latter method was used as the reference technique. Assessment of total body water using BIA - applying the RJL or Kushner equation-correlated significantly with the assessment using tritiated water dilution (P < 0.01). Body fat assessment using the two-compartment model based on either tritiated water dilution or BIA-applying the RJL or Lukaski equation-as well as body fat assessment using the four-compartment model based on tritiated water dilution and whole body 40K-counting were significantly correlated with body fat assessment using CT (P < 0.01) and resulted in good agreement with each other with respect to the absolute values of the body fat determination. BIA using other regression equations overestimated body fat by 7.2-13.7 kg. Whole body 40K-counting was significantly correlated with CT-determined muscle plus skin volume (P < 0.001). CT-calibrated anthropometric predictions significantly overestimated body fat. It is concluded that in patients with active acromegaly, the determination of body composition using either certain two-compartment models based on measurement of total body water or bioelectrical impedance, or a four-compartment model based on total body water and total body potassium measurements show good agreement with CT-determined body composition. Publication Types: Clinical Trial