Variability in feed and total mixed ration neutral detergent fiber and crude protein analyses among commercial laboratories

Abstract

The objective of this study was to assess the variability in amylase-treated neutral detergent fiber (aNDF) and crude protein (CP) analyses of feed and total mixed ration (TMR) samples among feed analysis laboratories. Two TMR were prepared that varied in the dry matter proportion of forage and concentrates: 45% forage (LF-TMR) versus 60% forage (HF-TMR). Replicated TMR and individual feed samples were dried, ground through a 4-mm screen, and sent to 10 commercial and 4 research or development laboratories for aNDF and CP analyses. Laboratories were asked to complete a detailed questionnaire regarding the aNDF procedure used. Variability in aNDF and CP analyses was assessed using univariate statistics and mixed modeling procedures. Significant variability in the aNDF analysis of individual feeds was found among the participating laboratories. The variability was particularly large for low-aNDF feeds such as distillers and barley grains. The variability among laboratories in the aNDF analysis of low-fiber TMR was greater than of high-fiber TMR, with the most likely reason being their greater proportions of grains and protein concentrates and the effect of variation in the aNDF protocols, particularly α-amylase use, on the analysis of these types of feeds. Variability due to the technique used for aNDF analysis was not statistically significant when outlier labs using the filter bag technique were removed; however, laboratories using the filter bag technique tended to produce more variable results than did laboratories using variation of the crucible technique (SE=2.542 vs. 0.930, respectively). Calculated aNDF values for TMR, based on proportions and aNDF analysis of individual feeds, were slightly greater than analyzed aNDF values for TMR. Results from this ringtest emphasize the need for feed analysis laboratories to follow the official aNDF method exactly. Variation within and among laboratories can be reduced by replicating analysis and including reference materials in each analytical run. Results of CP analysis were more consistent among laboratories, and variability in CP analysis of individual feeds or TMR was marginally acceptable.