Analysis of fatty acids in 12 Mediterranean fish species: advantages and limitations of a new GC-FID/GC-MS based technique.

When fatty acids in fish are analyzed, results in percentage form (profile analysis) are mostly reported. However, the much more useful results expressed as mg/100 g (absolute analysis) is the main information required. Absolute methods based on calibration curves are of good accuracy but with a high degree of complexity if applied to a great number of analytes. Procedures based on the sequence profile analysis-total FA determination-absolute analysis may be suitable for routine use, but suffer from a number of uncertainties that have never been really resolved. These uncertainties are mainly related to the profile analysis. In fact, most profile analyses reported in the literature disagree about the number and type of fatty acids monitored as well as about the total percentage to assign to their sum so leading to possible inaccuracies; in addition the instrumental response factor for all FAME (fatty acid methyl esters) is often considered as a constant, but this is not exactly true. In this work, a set of 24 fatty acids was selected and studied on 12 fish species in the Mediterranean area (variable in lipid content and month of sampling): in our results, and in these species, this set constitutes, on average, 90 ± 3 % of the total fatty acid content. Moreover the error derived from the assumption of a unique response factor was investigated. Two different detection techniques (GC-FID and GC-MS) together with two capillary columns (different in length and polarity) were used in order to acquire complementary data on the same sample. With the protocol here proposed absolute analyses on the 12 cited species are easily achievable by the total FA determination procedure. The accuracy of this approach is good in general, but in some cases (DHA for example) is lower than the accuracy of calibration-based methods. The differences were evaluated on a case by case basis.