Correlations among differing quantitative definitions of lipid oxidative stability in commodity fats and oils

The objective of this study was to explore the compatibility of various oxidative stability quantifications in vegetable oils. A diverse sample set (n=50) of edible fats and oils was examined. Oil stability index (OSI) data were collected with the initial lipids, and four classical methods (peroxide value [PV], conjugated dienes/trienes [CDT], p-anisidine value [p-AnV], and 2-thiobarbituric acid-reactive substances [TBARS]) were utilized to assess lipid decomposition over a period of 2 months of accelerated storage (60°C). Resultant oxidation curves were summarized by seven numerical interpretation techniques, and submitted to correlation analyses. The correlations among the different quantitations of stability ranged from negligible (r²<0.10) to strong (r²>0.90). The area-under-the-curve (AUC) and maximum value data demonstrated the greatest consistencies among the different assays. The strongest predictive relationship between 1° and 2° outputs occurred between maximum values for CDT and p-AnV (r²=0.71). OSI values showed moderate power regression relationships (r²=0.59-0.70) with the AUC data. The data demonstrate that the quantitation methods used to assess oxidative stability can dramatically affect results and conclusions. Caution, and consideration of these correlations, is recommended when designing oxidative stability studies, and also when comparing conclusions resulting from differing stability assessments. Practical applications: The information reported can assist investigators to make reasoned decisions in the design of studies regarding the assessment and summarization of oxidative stability. It can also help researchers when making comparisons to prior conclusions regarding oxidative stability, with more detailed consideration of the effect upon results of the different means by which stability may be assessed and defined.