Optimizing dose-dependent effect of phytase during formulation
Exogenous phytases are commonly added to animal diets to breakdown phytic acid from plant-derived ingredients. With the hydrolysis of phytic acid, availability of P is increased resulting to reduced inorganic P supplementation and reduced P excretion to the environment. Anti-nutritional properties of phytic acids are also lost during hydrolysis subsequently improving Ca and trace mineral retention and increasing energy and amino acid digestibility (Cowieson et al., 2011). Studies in broilers and swine showed that use of phytase in low available P diets improved weight gain, efficiency of feed conversion, bone mineralization and nutrient utilization (Harper et al., 1997; Leske and Coon, 1999; Shirley and Edwards, 2003; Dilger et al., 2004; Cowieson et al., 2006; Olukosi et al., 2007; Olukosi and Adeola, 2008; Gurbuz et al., 2009, Selle et al., 2012, Zeng et al., 2014). The magnitude of response to phytase varies according to different factors including phytase biochemistry and microbial source, phytase dose, dietary phytate concentration, protein source and characteristics, cation/anion balance, animal category, and gut physiology (Jongbloed et al., 2004; Cowieson et al., 2011; Bye et al., 2013; Dersjant-li et al., 2018).
Dose response to phytase
Increasing phytase dose has been shown to increase the digestibility of some nutrients in a linear, quadratic and cubic fashion. For instance, apparent total tract digestibility (ATTD) of Na, K, Mg and apparent ileal digestibility (AID) of Lys, Thr and Asp in weanling pigs increased linearly with increasing phytase dose up to 20,000 FTU/kg (Zeng et al., 2014). Meta-analysis conducted by Zouaoui et al. (2018) using data from 34 papers published between 1994 and 2015 also showed that digestibility of Arg, His, Ile, Leu, Met, Thr, Phe and Val in pigs increased linearly with the addition of phytase. In broilers, a linear increase in N retention and AMEn with increasing phytase dose up to 12,000 U/kg was also observed by Shirley and Edwards (2003).
Meanwhile, non-linear responses were observed in other nutrients. For example, in the studies conducted by Bento et al. (2012) and Dersjant-Li et al. (2017) in weanling pigs, P and Ca digestibility showed a quadratic increase with increasing phytase doses up to 2,000 FTU/kg. A quadratic increase in the ATTD of Ca, P and Cu and AID of Tyr in weanling pigs was also observed by Zeng et al. (2014). Moreover, in a meta-analysis conducted by Zouaoui et al. (2018), addition of microbial phytase resulted in a quadratic increase in Lys and Tyr digestibility in pigs. Kies et al. (2016) on the other hand, observed a cubic increase in ATTD of Ca and P and a quadratic increase in ATTD of Mg, Na, K, and Cu with increasing phytase dose up to 15,000 FTU/kg.
In broilers, some nutrients similarly showed curvilinear response. For instance, increasing phytase dose up to 12,000 U/kg resulted to a quadratic increase in Ca and P retention (Shirley and Edwards, 2003). Likewise, the study conducted by Beeson et al. (2017) showed a quadratic increase in nutrient retention in terms of total tract retention of N and P with increasing phytase dose.
Non-linear adjustment during formulation
Considering that phytase has non-linear effect to some nutrients, using a single matrix value for all doses is inaccurate. To take into account the curvilinear response of some nutrients to phytase and keeping in mind that most feed formulation softwares run based on a linear programming system, nutritionists or formulators need to run several simulations to come up with a least cost formula that optimizes the use of phytase considering its effect at different inclusion levels. This results to more time spent on feed formulation and longer listing of raw materials, which can be further lengthened if the nutritionists/formulators are optimizing formulas for two or more species. One solution to this is to use piecewise linear approximation during formulation. Piecewise linear approximation uses the response values at different doses to fit the curvilinear effect of phytase. Because all the values are saved in one profile, the non-linear effect of phytase can be taken into consideration in one formula optimization run. In addition, if the nutritionist/formulator wants to capture the value of phytase down to a single unit, optimum value in between recommended inclusion levels can be quickly generated using piecewise linear approximation.