Meta-Analysis on the relative effectiveness of the liquid hydroxy analogue of methionine compared with DL-Met in broilers using multi-exponential regression
- Publication Type
- Contribution to conference
- Authors
- Sauer N., K. Emrich, H.P. Piepho, A. Lemme, M. Redshaw, R. Mosenthin
- Year of publication
- 2007
- Published in
- 16th European Symposium on Poultry Nutrition
- Editor
- World's Poultry Science Association
- Page (from - to)
- 227-230
- Conference name
- European Symposium on Poultry Nutrition
- Conference location
- Strasbourg (France)
- Conference date
- 26.08.2007
- Keywords
- Broiler, DL-methionine, Liquid Methionine Hydroxy analogue
Plenty of experiments comparing the nutritional value of DL-methionine (dlMet) or the liquid methionine hydroxy analogue free acid (dlMHAFA) have been published in the past. Therefore, a meta-analysis was performed to estimate the relative efficiency of dlMHAFA compared with dlMet by means of exponential regression analysis taking into account daily weight gain (DWG) and gain to feed ratio (GF) as performance criteria. Each referenced study hereunder was individually analysed by exponential regression (Y=a+ß*exp(-yX)) and only those which converged and which parameter estimates fulfilled the constraints a>0, ß>0, and y>0 were eligible for further analysis. Data from 38 (GF) to 40 (DWG) dose-response experiments, extracted from a total of 27 peer-reviewed papers were used. Mixed-effects nonlinear regression was extended to allow for testing heterogeneity among plateaus. The analysis revealed that plateaus of both methionine sources can be assumed identical. Inclusion of the co-variable "age at start of experiment (ASE)" had a significant effect on the models (p<0.05) which were: DWG = (32.83 + 1.03 * ASE) + 10.46 * (1 -exp(-13.58 * dlMet - 10.96 * dlMHAFA); GF = (0.516 - 0.0026 * ASE) + 0.119 * (1 - exp(-8.41 dlMet - 6.66 * dlMHAFA). Regression analysis on DWG and GF revealed relative efficiencies of 80.7 % and 79.2 %, respectively, for dlMHAFA, based on equimolar comparison with dlMet, and the differences in efficiency proved to be significant (p<0.05).