Structural equation modeling of antinutrients in rainbow trout diets and their impact on feed intake and growth

Abstract The inclusion of plant proteins in rainbow trout diets significantly impacts growth and feed intake. This effect has been ascribed to the presence of antinutritional factors (ANF) present in plant ingredients, although the magnitude of the effects of ANF has not been determined. A series of six 56-day growth trials were performed to determine the effect of feeding 0, 75, 150, 225 or 300 g/kg of pea meal (PM), pea protein concentrate (PPC), soybean meal (SBM), soy protein concentrate (SPC), canola meal (CM) and canola protein concentrate (CPC) on the growth of rainbow trout. Diets were formulated to be equal in digestible nutrient content, with 17.6 MJ/kg digestible energy and 386.2 g/kg digestible crude protein. For each ingredient trial, specific growth rates (SGR) and average daily feed intakes (ADFI) of fish fed experimental diets were transformed to a percentage of the response of fish fed the 0 g/kg diet for each ingredient trial, (resulting in tSGR and tADFI, respectively). SGR and ADFI values were transformed to enable comparisons between experiments (tSGR and tADFI, respectively). Structural equation modeling was used to determine the highest likelihood model for the effect of ANF (starch, phytic acid, glucosinolates, tannins, isoflavones, total NSP, soluble NSP, insoluble NSP and saponins) on tADFI and tSGR, as well as the effect of tADFI on tSGR. All possible models were examined. The Akaike Information Criteria 0 was used to determine the model with the highest likelihood. This model contained 29 parameters and six degrees of freedom. Insoluble NSP, glucosinolates, saponins, tannins and phytic acid were statistically significant within the model. There were significant correlations between all ANFs in the structural equation model, except between phytic acid and saponins. Glucosinolates, saponins and phytic acid in the diet had a significantly negative impact on tADFI, whereas the tannins had a significantly positive impact on tADFI. SGR was negatively influenced by saponins and positively influenced by tADFI. This model provides a basis for the design of experiments to determine the effect of dietary ANF on the growth performance of rainbow trout. It can be expanded in the future using additional ingredients and ANF and can be tested by feeding diets containing varying levels of these ANF to determine if the resulting growth and feed intakes are the same as would be predicted by the model.