Physical limitations of the TOBEC method: accuracy and long-term stability.

While measurement of electrical conductivity provides one way of estimating body composition in live animals, the accuracy of methods based on this principle requires further study. In this work, we evaluate the effect of the ambient conditions and sample geometry on the response given by an SA-2 model (EM-SCAN) apparatus as well as the reproducibility of the measurements over time. A 2 degrees C variation of the sample temperature (a normal range for living animals) can result in a 6-10% variation in the response. When the conductive mass of a sample is increased in length, the response of the apparatus does not increase once the sample length reaches half the length of the measurement chamber. Samples having the same conductive mass but different shapes can show up to a 17-fold variation in signal. The geometry of the sample itself appears to be of prime importance for determining the strength of the response. We find that the reference phantom provided by the manufacturer is inadequate for calibration and is unable to detect the 10% variation over time of the signal of the apparatus. Until these problems are resolved, the usefulness of the EM-SCAN SA-2 to investigate body composition accurately is questionable.