ORACLE vs. Conventional Fat Methods: Challenges and Limitations of Gravimetric and Gas Chromatography Methodology
Though ubiquitous for foodstuff production and process control applications, gravimetric methods (regarded as “crude” fat) are often not optimized for a given matrix and the results are heavily dependent on a host of experimental conditions such as extraction time, temperature, solvent composition, and hydrolysis conditions (if applicable). Moreover, gravimetric methods can tend to overestimate fat due to extraction of nonfat components such as low molecular weight carbohydrates, amino acids, organic acids, and glycerol.1 Low moisture samples with high protein-to-fat ratios (e.g. Whey protein concentrates) have been shown to be particularly susceptible to overestimation, with nonfat components comprising nearly 50 % of the extraction value.2
The gas chromatography method, often used for food labeling and research applications, is able to provide very detailed information about the fat composition including the amounts of individual fatty acids. While it is viewed as being more accurate than gravimetric methods (since it is less prone to overestimation), results are heavily dependent on experimental factors such as derivatization conditions and efficiency, and column and hydrolysis conditions. Accurate quantitation can also be difficult due both to imprecision in chromatographic peak integration as well as volatility of short chain fatty acids.3
Illustrating the challenges in the reproducibility of conventional methods, the USDA recently published an extensive study whereby commonly available foodstuff CRM samples (sourced from NIST and LGC) were submitted to several contract laboratories for analysis. As summarized in Table 1, it was found that nearly 30 % of the gravimetric method results were outside the 95 % confidence interval (CI) of the certified values. In addition, approximately 17 – 56 % of the gas chromatography results (depending on fatty acid composition) were outside the 95 % confidence intervals.4 Significantly more variation was observed for mono and polyunsaturated fatty acids as compared to saturated fatty acids, which suggests results for samples with significant amounts of vegetable oils (e.g. food dressings) may be particularly unreliable.
|Read the Full Article With Data|
(1) Aued-Pimentel, et al. Quim. Nova 2010, 33, 76-84.
(2) Vaghela, M.N. and Kilara, A. JAOCS 1995, 72, 1117–1121.
(3) Kostic, Vesna. IOSR Journal of Pharmacy 2015, 5, 11–19.
(4) Phillips, Katherine et al. Anal Bioanal. Chem. 2007, 389, 219–229.