Quantifying long chain polyunsaturated fatty acids (LC-PUFA) in fish oil concentrates and algal oils − choosing the correct method

Lipid Technology August 2017, Vol. 29, No. 7-8 71 DOI 10.1002/lite.201700024 Feature Quantifying long chain polyunsaturated fatty acids (LC-PUFA) in fish oil concentrates and algal oils choosing the correct method Charlie Scrimgeour and Claire Traynor Charlie Scrimgeour is Consultant and Claire Traynor is Head of Mylnefield Lipid Analysis , James Hutton Limited. Invergowrie, DD2 5DA, UK E-mail: Claire.Traynor@huttonltd.com  st 1 Place Award ‘GOED Nutraceutical Oils Series’ in the AOCS Laboratory Proficiency Programme 2016/17 Traditionally, fish oils used for nutritional purposes as a result of their contents of long chain omega-3 fatty acids have been in their natural triglyceride form. Contents of eicosapentaenoic (EPA) and docosahexaenoic acid (DHA) have been relatively low, typically up to 30% maximum of EPA plus DHA combined. However, fish oil concentrates, in either ethyl ester or reconstituted triglyceride form, and algal oils rich in LC-PUFA are increasingly available in the speciality oil market. These may have EPA plus DHA contents of 40% to 75% in the nutritional supplements sector and even higher when used for pharmaceutical purposes. In recent years there have been a number of reports and counter reports on the quality of supplements produced with one of the complaints being products failing to meet label claims in terms of content of active ingredients [1, 2]. One reason for this has been the use of inappropriate methods for measuring long chain omega-3 in concentrated products. Accurate analysis of the LC-PUFA is required for production control, labeling claims and trading standard tests. It is important the most appropriate analytical method is used with such materials to ensure compliance with specifications, label claims and customer requirements. In the majority of cases the fatty acid compositions of oils and fats are measured using gas chromatography (GC), with the samples of interest first being methylated to create fatty acid methyl esters (FAMES). The primary result from GC analysis of a fatty acid methyl ester mixture is a trace of detector response against time (Figure 1), from which the fatty acids can be identified and quantified. The normalised detector response is often referred to as the area % of the individual fatty acids. Since the flame ionisation detector responds to the amount of carbon eluting from the GC column over a wide dynamic range, the area% is close to the weight% of fatty acids in the mixture. The weight % can be accurately calculated using theoretical response factors based on the empirical formula of the individual fatty acids. For the range of chain-length and unsaturation found in most plant oils this small correction (Table 1) is all that is required to obtain reliable weight% results (note the weight%, as described above, is the amount of each fatty acid in the fatty acid profile, NOT its absolute amount in the sample). The absolute amount may be less than the weight% because:  Sample rarely comprise only fatty acids most oils are triacylglycerols which contain 4% glycerol.  Minor components such as tocopherols and sterols are present. © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim  Oxidation or other degradation has occurred.  The oil may be only part of the formulation. To measure absolute amounts, the detector response has to be calibrated against a known amount of standard. Odd chain-length fatty acids are rare in nature and can be used as internal standards, added at the start of the analysis and subject to the same extraction and derivatization procedures as the sample. 17:0 is used for plant oils and 23:0 for samples with a wider range of chain lengths such as fish oils (Figure 2a). For example, the commonly used method of the American Oil Chemist’s Society (AOCS) Ce 1b-89, uses 1 mg 23:0 as internal standard in 25 mg sample [3]. The theoretical response factors are then used to obtain the weight% for the sample plus standard and hence the absolute amounts of each fatty acid, usually expressed as mg/g. AOCS Ce 1b-89 was developed for the analysis of fish oils which seldom have >20% of either EPA or DHA. Fish oil concentrates and algal oils with >30% EPA and/ or DHA are now widely available and the accuracy of this method has been questioned for such samples. The weakness of AOCS Ce 1b-89 is its reliance on theoretical response factors for the highly unsaturated long chain polyenes. Long-chain polyenes such as EPA and DHA are vulnerable to degradation during GC analysis. Their high molecular weight means that they are late eluting from the GC column and are exposed to the maximum oven temperature for several minutes, DHA being almost the last component to elute. The long exposure to high temperature and their high unsaturation makes them liable to reaction with any traces of oxygen or water in the carrier gas and to reactions catalyzed by any active sites on the injector or the column. Consequently, there may be significant losses of LC-PUFA relative to more saturated components during analysis. The alternative method (European Pharmacopeia, EP) 2.4.29 [4], developed particularly for concentrates, uses measured response factors for EPA and DHA relative to 23:0 to correct for losses during GC analysis, and also requires a number of system suitability tests to ensure that GC and column performance is adequate. 60 to 90 mg amounts of high purity methyl 23:0, ethyl EPA and ethyl DHA are required for each calibration and analysis. In consequence, the EP method is more expensive and time consuming to perform. According to their 2015 ver. 5 Voluntary Monograph and associated Technical Guidance Document, the Global Organisation for www.lipid-technology.com 72 August 2017, Vol. 29, No. 7-8 Lipid Technology Figure 1. Typical GC trace. EPA and DHA (GOED) recommended method for analysis of fish oils and fish oil concentrates is based on the EP 2.4.29 method. System suitability Detector response theoretical and measured response factors for a gravimetric mixture of 16:0 to 22:0 saturated standards must agree within set limits. Column resolution there must be adequate resolution between 23:0 and any 21:5(n-3) present (Figure 2b) and between DHA and 24:1 (Figure 2c). Figure 2. a) Location of internal standard between EPA and DHA, b) separation of internal standard from 21:5(n-3), c) separation of DHA and 24:1. Response factors Table 2. Comparison of measured and theoretical response factors. Measured response factors for a typical test mixture (Table 2) are of the order of 10% lower than theoretical values. There is significant day to day variation in the response factors, requiring them to be checked with each batch of samples. EPA measured response theoretical response % factor factor difference 1.07 0.98 8.7 DHA 1.09 0.97 12.0 Comparison of methods Results for a sample analysed by both methods (Table 3) show that the absolute amounts (mg/g) are around 10% higher for the EP method. Area% results are similar for both methods. The difference in mg/g results closely reflects the difference between theoretical and measured response factors. Table 1. Theoretical response factors. Theoretical response factors (relative to 18:0) Fatty acid Response factor 16:0 1.02 18:0 1.00 18:1 0.99 18:2 0.99 20:0 0.98 20:5 0.95 22:0 0.97 22:6 0.94 www.lipid-technology.com Table 3. Comparison of results. AOCS EP mg/g mg/g % difference EPA 8.2 8.9 7.8 DHA 343.5 394.5 12.9 A% A% % EPA 1.00 0.98 2.0 DHA 42.5 43.39 2.0 Conclusion Method EP 2.4.29 is strongly recommended for samples expected to contain 30% or more of EPA and DHA combined. Method AOCS Ce 1b-89 gives consistently low results for these samples due to its reliance on theoretical response factors. While differences between the two methods may not be significant at lower concentrations, for a 1000mg soft gel capsule containing © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Lipid Technology highly concentrated fish oil, the difference may be as much as 50mg or 75mg of active ingredient per dose. References [1] Albert, B.B., et al., Nat. Sci. Rep. 2015, 5: 7928. © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim August 2017, Vol. 29, No. 7-8 73 [2] Nicholls, P.D., et al., Nutrients 2016, 8, 703. [3] AOCS, Fatty Acid Composition of Marine Oils by GLC, AOCS Official Method Ce 1b-89, Official Methods and Recommended Practices of the AOCS, (6th Ed.). [4] Composition of Fatty Acids in Oils Rich in Omega-3 Acids (01/2016:20429), European Pharmacopoeia (9th Ed.) www.lipid-technology.com