Use of bioelectrical impedance spectroscopy to provide a measure of body composition in sows.

The ability to accurately estimate fat mass and fat-free mass (FFM) has the potential to improve the way in which sow body condition can be managed in a breeding herd. Bioelectrical impedance spectroscopy (BIS) has been evaluated as a practical technique for assessment of body composition in several livestock species, but similar work is lacking in sows. Bioelectrical impedance uses population-specific algorithms that require values for the apparent resistivities of body fluids and body proportion factors. This study comprised three major aims: (i) to derive apparent resistivity coefficients for extracellular water (ECW) and intracellular water (ICW) required for validation of BIS predictions of total body water (TBW) in live sows against standard reference tracer dilution methods; (ii) to develop predictions of TBW to body composition prediction algorithms, namely FFM, by developing a body geometry correction factor (Kb) and (iii) to compare the BIS predictions of FFM against existing impedance predictors and published prediction equations for use in sows, based on physical measurements of back-fat depth and BW (P2-based predictors). Whole body impedance measurements and the determination of TBW by deuterium dilution and ECW by bromide dilution were performed on 40 Large White x Landrace sows. Mean apparent resistivity coefficients of body fluids were 431.1 Ω.cm for ECW and 1827.8 Ω.cm for ICW. Using these coefficients, TBW and ECW were over-estimated by 6.5 and 3.3%, respectively, compared to measured reference values, although these differences were not statistically different (P > 0.05). Mean Kb was 1.09 ± 0.14. Fat-free mass predictions were 194.9 kg, which equates to 60.9% of total sow weight, and 183.0 kg for BIS and the deuterium dilution method, respectively. Mean differences between the predicted and measured FFM values ranged from -8.2 to 32.7%, but were not statistically different (P > 0.05). Method validation (leave-one-out procedure) revealed that mean differences between predicted and measured values were not statistically significant (P > 0.05). Of the impedance-based predictors, equivalence testing revealed that BIS displayed the lowest test bias of 11.9 kg (8.2%), although the P2-based prediction equations exhibited the lowest bias and percentage equivalence, with narrow limits of agreement. Results indicate although differences between mean predicted and measured values were not significantly different, relatively wide limits of agreement suggest BIS as an impractical option for assessing body composition in individual sows compared to the use of existing prediction equations based on BW and back fat.