Dear Ms. Peris, Thank you for your comments. The first objective of this trial was to determine the relative bioavailability of Met sources in broilers. As stated by Jansman et al., (2003) proper studies are those that include a basal diet clearly deficient in Met, and at least 3 levels of supplementation of the test products. As in any dose-response trial, it is crucial that enough data points are included to properly describe the different sections of the response curve. Thus, it is important that the basal diet (starting point) is clearly deficient in the nutrient of interest. In this study, the basal diet was clearly deficient in Met confirmed by the significant reduction in growth performance. The 5 increasing levels of each Met source improved growth performance following - as expected - the law of diminishing returns and with data points well distributed along the complete response curve (deficiency and above plateau). Therefore, all criteria for a proper bioavailability study were met. Accurate “bioavailability values” cannot be determined or extrapolated by only comparing nutrient sources above the requirement. Using this approach, data from this trial will suggest that at the highest Met inclusion level a diluted DL-Met to 65% purity (DLM65) is equally effective as DL-Met with 99% purity! Certainly, this is not correct. As these points are in the plateau section of the response curve (requirement already met), no further responses are expected above this level (law of diminishing returns) which does not allow for any proper comparison of nutrient sources. Therefore, to properly determine the bioavailability of nutrient sources, trials need to include a basal diet clearly deficient in the nutrient of study with enough data points representing the different sections of the response curve.
The second objective of this trial was to validate the mathematical suitability of the multi-exponential regression analysis for estimating the relative bioavailability. For this, we used diluted DL-Met to a purity of 65% (DLM65) from which its bioavailability is known a priori (close to 65%). As DLM65 was included, equimolar comparison would have not been possible in this experiment (same inclusion as DLM). However, proper experiments designed either on equimolar or product basis led to the same conclusion (Jansman et al., 2003; Hoehler et al., 2005, EFSA, 2018). For example, based on an extensive literature study Jansman et al. 2003 reported an average bioefficacy of liquid HMTBA of 77% on equimolar basis compared to DL-Met in broilers. This translates to a bioefficacy of 68% on product basis compared with DL-Met (0.77 * 88). Similarly, the European Food Safety Authority (EFSA, 2018, Appendix A) released a scientific opinion on liquid HMTBA and its calcium salt, confirming a bioavailability of 75% on equimolar basis which is equivalent to 66% (0.75 * 88) on product-to-product basis for liquid MHA-FA. These previous results are in agreement with those obtained in the current study which resulted in an average bioavailability of HMTBA (MHA-FA) and DLM65 of 65 and 61%, respectively. Therefore, proper studies conducted either on equimolar or product basis will lead to the same conclusion.