Reference Chemistry Limitations and Challenges with Mineral Additives

The testing of animal feeds has great importance in part due to the raising of various animals for consumption. In order to obtain the highest quality meats, the diets of the animals are strictly controlled to maximize their profitability. Obtaining accurate and repeatable reference results on these animal feeds is of paramount importance. In the US, animal feed is tested under AOAC 920.29, an ether extraction method, while in Europe, it is tested under ISO 11085, an acid hydrolysis method. This difference in technique often leads to differences in fat values. The AOAC extraction method utilizes ether as the solvent and cannot extract any bound fat. Conversely, the ISO acid hydrolysis method can potentially extract non-lipid components, thereby inflating the fat values. In addition to the risk of inaccurate results from under- or over-extraction, both methods suffer from poor repeatability, caused by the wide and often changing variety of grains, protein sources, roughage, and other added minerals that comprise animal feeds.

Since the ORACLE™ was introduced in 2016, it has proven to be incredibly successful in accurately determining fat values in countless applications, without the need for calibration or method development. In several independent studies, the ORACLE has proven to be just as accurate and more precise than the chemical analysis methods for a broad range of applications. However, animal feed samples showed significantly lower results on the ORACLE, compared to reference chemistry methods. To determine the cause of this discrepancy between the two results, several different animal feed certified reference materials (CRM) were sourced and analyzed with the ORACLE, as well as both extraction methods listed above.

When analyzing animal feeds on the ORACLE, some samples had lower than expected fat values. Several hypotheses were considered, most of which centered on a comparison of the reference methods and rapid techniques. The initial hypothesis was that the extraction techniques were over extracting nonlipid components, similar to what has been observed with Whey Protein Concentrates (WPC).1 Upon testing the CRM animal feed samples, it was observed that most of the samples produced results comparable to the reference method with excellent precision. Certain samples, however, had low results that were outside of the acceptable statistical range. Upon analysis of the NMR data, it was discovered that there was a rapidly relaxing fat signal that corresponded to the difference between the ORACLE result and the reference technique. To understand why this portion of the fat signal behaved differently, a detailed study of the composition of these feeds was performed.