Pitfalls in Digestibility Techniques for Evaluation of Phytases.


Phosphorus (P) and calcium sparing values for phytases have been estimated from digestibility, growth and bone ash studies where comparisons with an inorganic source yield an equivalency value. Such studies lack precision, particularly in establishing the true extent of deficiency of the reference. Low P diets depress intake, and often more phytase is required to equilibrate intake than AvP with the positive control, such that calculated P equivalency exceed that achieved in commercial practice. Dietary phytate enhances and phytase reduces endogenous losses in what appears to be a dose dependant manner, thus phytase will favour partitioning of AME to NE of gain rather than NE of maintenance, an effect which may be missed in digestibility studies. Concomitant growth and bone ash studies using appropriate controls are needed to establish any defined nutrient matrices.


The use of phytase globally has increased dramatically since its first use in the early 1990’s. Initially its use was effectively mandated in areas such as the Netherlands where environmental concerns with regards to phosphorus pollution made animal agriculture almost impossible in the absence of this enzyme (Simons et al., 1990). Phytase use reduced phosphorus pollution through its ability to hydrolyse phosphorus from plant phytates, making it available to the animal and thus enabling feed compounders to reduce inclusion rates of inorganic phosphates (Simons et al., 1990; Simons & Versteegh, 1990; Cromwell, 1991). The use of phytase originally did not spread widely since the cost of the enzyme often exceeded the savings in inorganic phosphates, but nevertheless its use did establish the application of a phosphorus “equivalency” value of the enzyme. Much of the early work established this equivalency value based on digestibility eg (Eeckhout & de Paepe, 1991; Beers & Jongbloed, 1992; Schoner et al., 1993) and performance studies eg (Schoner et al., 1993), some of which compared phytase dosage against an inorganic source of P in P deficient diets. Such work was effective in providing an estimate of the amount of AvP that could be replaced by a given amount of phytase, typically this being approximately 0.1% AvP for the broiler when 500 units of phytase are used per kg feed. – (Vogt, 1992; Beers & Jongbloed, 1992). Nevertheless, despite the apparently simplistic relationship between phytase dose and P release from dietary phytate, in practice there is a great deal of variation in the responses observed in vivo, particularly between classes of animals or as the animal ages. For example Kornegay estimated that 500 units of phytase was equivalent to as little 0.037% digestible phosphorus from a review of 23 poultry experiments from 13 references (Kornegay, 2001) whereas Van der Klis (van der Klis et al., 1997) estimated only 192 units of phytase was needed to replace 0.1% AvP in laying hens. The notion that a given dose of phytase can consistently replace a fixed amount of digestible phosphorus is therefore open to question from an academic viewpoint, but it is clear from commercial use that in practice such an assumption has been applied with success for the most part


For both growth and digestibility studies the key to success is use of an appropriate positive and negative control. Ideally the positive control should supply the nutrient of interest at precisely the requirement of the animal for the response parameter tested. In such a case the deficiency of the negative control would be equivalent to the differences in content of that nutrient between the two controls. Any increment in the availability of the nutrient through the activity of the phytase will result in improved performance up to that of the positive control. In practice, however, such precision is rarely if ever achieved due to variation in animal performance and ingredient composition. As a result, if an exogenous nutrient source is not simultaneously added to the negative control to construct a standard dose response curve, (eg an inorganic phosphorus source) then the true extent of the deficiency of the negative control is unknown. A diet, which is nominally deficient in AvP by 0.2%, may actually be far less deficient than predicted. Such an error would be observed if there were a dose titration of an inorganic source into the negative control, but if this were not the case then the resulting growth studies could be overly optimistic in describing the value of the phytase. This issue is also relevant for some digestibility studies. If the positive control employed provides P in excess of requirement, and the negative control when supplemented with phytase just approaches P requirement, then the apparent benefit of addition of a phytase will appear greater in metabolisability compared with ileal or faecal digestibility studies. An example with growing pigs is given below in Table 1. Gain did not differ significantly between the positive and negative controls (6.5 vs 3.7g/kg phosphorus between 35 and 88kg suggesting the NC was at best marginal and the PC probably over-fortified in phosphorus.

The digestibility of the phosphorus at the faecal level was 54% for the PC and fell to 45% on the NC due to removal of the highly digestible inorganic source. Adding phytase to the NC raised this to 66% with the result that the calculated digestible P content of the phytase containing diet was 2.44g/kg, or 0.76 g/kg more then the NC. Thus in this trial 1200U of phytase has an “equivalence” of 0.76g of P, bringing the digestible P content of the ration to 69% of that of the positive control.

Table 1. Comparison of phytase effects on digestible and retainable P in pigs. Adapted from (Nasi & Helander, 1994)

   Pos Control  Neg controlNC + 1200 U /kg% of pos control
Total P g/kg6.503.703.7057%
Intake P g/d14.808.408.50 
Absorbed g/g8.003.805.60 
Diet digestible P content g/kg  3.51  1.67  2.44  69%
Digestible P value g/kg  0.76 
Urine P g/d2.180.090.42 
P retained g/d5.803.705.18 
Met P diet content g/kg2.551.632.2589%
Apparent Met P value g/kg  0.63 

Note, however, the significant urinary losses of P from the positive control, suggesting there is an excess of phosphorus in this diet. The urinary losses result in the metabolisability of P falling to 39% from a digestibility of 54%, resulting in the apparent metabolisable P

content of the diet falling to 2.55g/kg from a digestible P content of 3.51g/kg. Urinary P losses on the negative control are minimal, suggesting that the diet was actually deficient. Adding phytase clearly enhances P digestibility, and as a result it will increase plasma phosphorus (Perney et al., 1993; Carlos & Edwards, 1998; Shirley & Edwards, Jr., 2003). This may result in marginal increments in urinary P compared with the negative control, which is the case in this study, but in comparison with a positive control the urinary losses are small. Large urinary P losses in poultry have bee reported elsewhere on positive controls (Rodehutscord et al., 2002). The benefit of the phytase compared with its negative control is calculated as 0.63 g/kg metabolisable P, this reduction compared with the digestible P value being a result of the relative urinary losses discussed. However, as a result of significant urinary losses on the positive control, the phytase containing diet appears to be much closer to equivalence with the positive control at the level of metabolisable compared with digestible P. Some may interpret this as the enzyme being able to almost account for the difference in P content between the positive and negative control. Thus digestibility studies are difficult to interpret at best, with retainable P clearly being preferred to digested, but even so a standard curve using inorganic sources should be employed to confirm the degree of deficiency of the negative control and against which the phytase can be compared.


A more serious issue is that if the use of low P diets not only reduces dietary AvP but also reduces intake (Rosen, 2001; Rodehutscord & Dieckmann, 2005; Cowieson et al., 2006). Thus for performance to be restored, a restoration of both dietary AvP and intake is required. In some but not all cases the calculated AvP content of the ration is restored to that of the positive control at relatively low levels of phytase inclusion, but significantly more phytase is required to equilibrate intake An example from the literature is given below in table 2.

Table 2. influence of graded levels of phytase on phosphorus metabolisability, intake and performance of broiler chickens to 14d of age (Cowieson et al., 2006)

DietDiet P content %P metabolisability %Diet Metabolisable P content %Intake g (1-14d)Intake of Metabolisable P gGain /g 1-14dToe ash %DM
Positive control0.8548.30.414171.7122514.78
Negative control0.6559.00.383481.3315812.55

These data clearly show that if diet metabolisable P content was the benchmark of choice, then it could be argued that the use of 150 U/kg feed of phytase resulted in an equivalent diet to that of the positive control. However it is clear even from this measure that there are problems with reliance on measurement of metabolisability of P alone. The positive control was formulated to contain 0.5% metabolisable P, and clearly this was not the case. It is possible that the bird excreted the “excess” in the urine resulting in measurement of only

0.41% metabolisable P. The negative control was formulated to contain 0.3% metabolisable P

but clearly the bird retained P more efficiently than predicted, presumably by upregulating absorption and down-regulating urinary excretion. As a result there was a difference of only 0.03% in metabolisable P content in the negative compared with the positive control in stark contrast to the formulated difference of 0.20 %. As a result, addition of even a small dose of phytase was all that was required to bridge this very small difference. However, it is clear from the intake data that the effect of reducing diet P content was far more dramatic on intake than diet metabolisable P content. As a result the intake of metabolisable P over the 14d period was not equilibrated with that of the positive control until more than 1200 U of phytase had been added. It is clear from both gain and toe ash that performance equilibration with the positive control required more than 1200 units of phytase also. It is clear, therefore, that measurement of metabolisability of P in the absence of simultaneous determination of intake or performance could result in significant errors being made. This principal is just as valid, probably moreso, for energy and amino acids.since the effects of phytase on the metabolisability of these nutrients is markedly less than that on phosphorus, hence intake effects are proportionately more important.


Recent observations that phytate can increase and phytase condomitantly decrease endogenous secretions of mucin (Cowieson et al., 2004; Onyango et al., 2004; Cowieson & Ravindran, 2007) have lead to the realisation that not all of the effects of phytase can be determined through direct digestibility studies. Clearly endogenous secretions of mucin result in expenditure of energy and amino acids, nutrients which have been metabolised, but which would otherwise be directed towards growth. Since the utilisation of phytase depresses the losses of mucin it is clear that a greater proportion of metabolisable energy is able to deposit as net energy of gain rather than be “wasted” as net energy of maintenance. Thus there may be situations where the application of a phytase results in no discernible improvement in energy digestibility or metabolisability, but it does result in improved FCR or Net energy of gain. An example is given below in table 3.

Table 3. Influence of phytase on energy metabolisability and performance in broilers from 7- 17d. (Pirgozliev et al, 2007, submitted)

   Intake g/d  Gain  FCEAME MJ/kgNE MJ/kg

This effect of phytase is in many respects confounded with its effect on intake. If use of incremental doses of phytase simultaneously reduces mucin losses and increases intake, then not only is the enzyme permitting a more efficient conversion of metabolisable energy into net energy of gain, but by virtue of the fact that there is a greater intake of metabolisable energy per day, then there is a greater proportion of the daily intake which can be dedicated to growth rather than maintenance. As a result there is a clear problem with reliance on digestibility or metabolisability studies alone in evaluation of this particular enzyme.

In conclusion, much can be learned about the mechanism of action and to some degree the extent of action of phytases through metabolic studies. However, the nutrient matrices for these enzymes derived from metabolic work needs to be reliant on concomitant validated performance studies. Over reliance on metabolic studies alone will likely lead to significant errors in estimation of the achievable nutrient matrix of a given enzyme, resulting either in an undervaluation of the enzyme and thus loss of revenues or worse, an overvaluation of the enzyme and potential loss of performance and possibly health of the animal.


1AB-Vista Feed Ingredients, Marlborough UK. SN8 4AN.