Spectral profiles of commercial omega-3 supplements: an exploratory analysis by ATR-FTIR and ¹H NMR

Abstract

Most of the population is dependent on supplemental products to reach the recommended level of omega-3 polyunsaturated fatty acid (ω-3 PUFA) intake. Thus, knowledge about the quality of ω-3 supplements is important for their safe consumption. In this work, attenuated total reflectance–Fourier transform infrared (ATR-FTIR) and nuclear magnetic resonance (NMR) spectroscopy were applied to assess the quality of fourteen commercial ω-3 supplements. Using ATR-FTIR data, we could identify whether ω-3 PUFA was esterified as either triacylglyceride (71%) or ethyl (29%) esters in ω-3 supplements. The type of esterification is rarely included in the product labels, although the consumer should have the right to choose which form of the supplement to consume. On the other hand, ¹H NMR spectra were useful to determine the relative concentration of eicosapentaenoic (EPA) and docosahexaenoic (DHA) acids, and ω-3 PUFA in these commercial samples. Ethyl esters have higher concentrations of unsaturated fatty acids. The NMR results showed a good agreement between the obtained and declared DHA and EPA amounts on the product labels, except for one sample whose high level of ω-3 PUFA indicated it to be a vegetable oil-enriched supplement. Moreover, ω-3 supplements from Schizochytrium sp. microalgae oil revealed higher levels of DHA and ω-3 PUFA, but lower levels of EPA than fish oil. These findings indicate the need for a constant assessment of the quality of commercial products whose ATR-FTIR spectra could be routinely used for the evaluation of PUFA esterification, and NMR analysis could be used to provide advanced quantitative information on commercial ω-3 supplements.

Electronic supplementary material

The online version of this article (10.1007/s13197-019-04157-y) contains supplementary material, which is available to authorized users.

Introduction

Omega-3 polyunsaturated fatty acids (ω-3 PUFA) have been shown to reduce the risk of coronary artery disease, hypertension, diabetes, arthritis, other inflammatory and autoimmune disorders, and cancer (Dias et al. 2017; Huerta-Yépez et al. 2016; Punia et al. 2019; Simopoulos 2002). Some PUFA are essential because they are not synthesized by humans and must be obtained from dietary sources. A diet rich in seafood provides an adequate daily intake of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), but most of the population depends on supplemental products to reach the recommended daily doses of these fatty acids (Vestland et al. 2016).

The primary sources of ω-3 PUFA are fish oils, but alga-, moss- and fungus-based sources of PUFA have also been employed in food supplements (Shapaval et al. 2014; Punia et al. 2019). One limitation of unrefined fish oil is the relatively low concentration of ω-3 PUFA, naturally found as a triacylglyceride ester (TG). Higher concentrations of ω-3 PUFA can be achieved through a transethylation process. During this process, the glycerol backbone of TG is removed from all fatty acids. Then, the PUFA, as free fatty acids, is esterified again as an ethyl ester (EE), while some of the shorter chain fatty acids are taken out (Leite et al. 2006). Several clinical comparative studies have compared ω-3 PUFA in TG and EE forms. Some results suggested that the bioavailability, safety, efficacy, and absorption of ω-3 PUFA are similar in the TG and EE forms when a steady-state has been achieved (Hansen et al. 1993; Krokan et al. 1993; Nordøy et al. 1991), while others suggested problems in their bioequivalence demonstration (Maki et al. 2017) and absorption (Lopez-Toledano et al. 2017a, b).

Traditionally, ω-3 PUFA have been analyzed by gas chromatography coupled to flame ionization detection or mass spectrometry (Li and Srigley 2017). In these techniques, lipids are first extracted from biological samples and then converted to methyl esters, which are separated and analyzed by gas chromatography. The disadvantages of these methods include possible oxidation of PUFA, artifact formation during the extraction and transmethylation processes, and laborious and time-consuming analysis (Sacchi et al. 1993; Igarashi et al. 2002). Currently, nuclear magnetic resonance (NMR) has been employed in oil analysis, which allows reduced analysis time, minimum sample preparation, and good agreement with gas chromatography results (Aursand 1993). Igarashi et al. (2002) developed a method to quantify DHA and ω-3 PUFA in fish oils using a protocol validated by an IUPAC interlaboratory study. Several studies have employed NMR in edible oils analysis (Nieva-Echevarría et al. 2016; Popescu et al. 2015; Ruiz-Aracama et al. 2017). Alternatively, attenuated total reflectance– Fourier-transform infrared spectroscopy (ATR-FTIR) has been employed to monitor PUFA levels. Yoshida and Yoshida (2004) developed a non-invasive and reagent-free method to observe PUFA behavior in the human oral mucosa. In turn, Shapaval et al. (2014) developed an ATR-FTIR-based approach for the screening of PUFA produced by Mucor fungi.

Due to the health benefits related to ω-3 PUFA consumption, the market for supplement capsules has substantially and rapidly expanded. Consequently, there is a corresponding demand for analytical methods that can routinely monitor these products and assess the quality of production. In this work, ATR-FTIR was used to identify whether ω-3 PUFA was esterified as either TG or EE esters in ω-3 supplements. ¹H NMR was employed to determine the relative concentration of DHA, EPA, and ω-3 PUFA in fourteen commercial ω-3 supplements.

Materials and methods

Commercial ω-3 supplements

Fourteen samples of commercial ω-3 supplements from different Brazilian manufacturers were purchased from a local market (Campo Grande-MS, Brazil).

ATR-FTIR measurements

For ATR-FTIR analysis, the ω-3 supplement capsules were punctured with a glass pipette, and three drops of the oil were placed directly in the ATR accessory. Following background correction, the samples were analyzed as films. Measurements were performed on a PerkinElmer Spectrum 100 FTIR spectrometer equipped with a GeSe (germanium selenium) ATR accessory using the PerkinElmer Spectrum software version 10.03.07. The spectra were obtained in the region between 4000 and 600 cm⁻¹, with a spectral resolution of 4 cm⁻¹. All spectra were the average of 16 scans. All samples were analyzed in triplicate. The achieved spectra are available in Figure S1.

¹H NMR analysis and data processing

For ¹H NMR analysis, the ω-3 supplement capsules were punctured with a glass pipette to remove the oil, and 50 mg of the oil was dissolved in deuterated chloroform (500 µL, CDCl₃) with tetramethylsilane (1%, TMS) as a chemical shift reference. Mixtures were homogenized and transferred into 5-mm NMR tubes. All samples were analyzed in triplicate.

Experiments were carried out without spinning at 25 °C in a Bruker DPX300 7.05 T spectrometer equipped with a 5 mm dual-probe at a frequency of 300 MHz for hydrogen. An IUPAC interlaboratory study demonstrated that less expensive 300 MHz instruments were comparable with 400–500 MHz instruments for quantification of the main PUFA (Igarashi et al. 2002). The experimental parameters for ¹H NMR experiments were as follows: data acquisition, 2.73 s; relaxation delay, 1.0 s (repetition time, 3.73 s); data points, 32 k; spectral width, 20 ppm; dummy scans, 2; scans, 64; pulse angle, 30°; temperature, 25 °C; and frequency offset, 5.56 ppm.

Data processing included zero filling to 64 k, line-broadening multiplication by 0.3 Hz, and Fourier transformation. All spectra were automatically phase- and baseline-corrected and referenced to TMS (set to 0.000 ppm), followed by manual correction when necessary. The signals were integrated following bias and slope correction using TopSpin software (v3.5 pl7, Bruker Biospin). The achieved spectra are available in Figure S3.

Relative concentrations of DHA, EPA and ω-3 PUFA

Relative concentrations (mol/mol) of DHA, EPA, and ω-3 PUFA were obtained from ¹H NMR spectra (Bratu et al. 2013; Igarashi et al. 2002; Tyl et al. 2008). The relative proportion of DHA was calculated according to Eq. 1:

$$DHA\%=\frac{I_I/2}{I_H+I_I/2}\times 100$$ 1

where I_I represents the signal areas from the DHA α-methylene group, and I_H represents the signal areas from the other FA α-methylene groups. The relative proportion of EPA was calculated according to Eq. 2:

$$EPA\%=\frac{I_F}{\left(I_F+I_E+I_I/2\right)}\times 100$$ 2

where I_F represents the signal areas from the EPA β-methylene group, I_E represents the signal from other β-methylene groups, and I_I represents the signal areas from the DHA α-methylene group. The relative proportion of ω-3 PUFA was calculated according to Eq. 3:

$$\omega -3PUFA\%=\frac{I_B}{(I_A+I_B)}\times 100$$ 3

where I_B represents the signal areas from the terminal methyl groups of ω-3 PUFA and I_A represents the signal areas from the terminal methyl groups of other FA.

Assignment of ¹H NMR and ATR-FTIR spectra

The assignments of ATR-FTIR spectral features were completed by comparing them with the literature data (Plans et al. 2015; Rohman and Man 2010) and are presented in Fig. 1 and Table S1. The assignments of ¹H NMR spectral features were completed by comparing them with the literature data (Aursand et al. 2007; Aursand 1993; Bratu et al. 2013) and are presented in Fig. 2 and Table S2. The assignments were confirmed by heteronuclear single quantum coherence spectroscopy (2D ¹H-¹³C HSQC), homonuclear total correlation spectroscopy (2D ¹H-¹H TOCSY), correlation spectroscopy (2D ¹H-¹H COSY), and two-dimensional J-resolved NMR spectroscopy (JRES) using Bruker’s available pulse sequences.

Fig. 1.

Representative ATR-FTIR spectra from ω-3 supplements in a triacylglyceride ester (TG) and b ethyl ester (EE)

Fig. 2.

Representative ¹H NMR spectra from ω-3 supplements in a triacylglyceride ester (TG) and b ethyl ester (EE)

Chemometric analyses

Principal component analysis (PCA) was employed in an exploratory analysis of the ATR-FTIR data. PCA was performed on mean-centered data following standard normal variate (SNV) transformation (Dhanoa et al. 1989) in the 3150–2700 cm⁻¹ and 1900–600 cm⁻¹ regions. An optimal number of the principal components (PC) was selected by scree plot analysis. The presence of outliers was evaluated from the graph of Q residual versus Hotelling’s T². All chemometric analyses and calculations were performed in MATLAB^®.

Results and discussion

Assignments of ATR-FTIR and ¹H NMR spectra

Representative ATR-FTIR spectra are displayed in Fig. 1. A complete assignment from ATR-FTIR spectra is presented in Table S1. The peaks at 1743 and 1734 cm⁻¹ correspond to the carbonyl stretch from TG and EE, respectively (Figure S2). These signals allow ATR-FTIR to distinguish ω-3 PUFA as either TG or EE.

Representative ¹H NMR spectra of commercial ω-3 supplement are presented in Fig. 2. A complete assignment from ¹H NMR spectra is presented in Table S2. The signals I_I (2.37 ppm), I_F (1.75 ppm), and I_B (0.95 ppm) allow for the determination of DHA, EPA, and ω-3 PUFA, respectively.

Differentiation between TG and EE

To differentiate TG and EE, PCA was performed from ATR-FTIR data. The PCA model with two principal components (PC) explained 96.83% of data variance for mean-centered data, after standard normal variate (SNV) transformation. The PCA model showed a complete separation between TG and EE along the first PC. TG samples showed negative scores in PC1. In contrast, EE samples showed positive scores (Fig. 3).

Fig. 3.

Score plot for the PCA model from the ATR-FTIR data (2 factors; 96.83% of the variance, mean-centered data after SNV transformation). TG (○) and EE (*)

The first loading for the PCA model showed that the signals related to the symmetric C–H stretches (CH₃ and CH₂) (2871 and 2855 cm⁻¹) and the TG carbonyl stretch (1743 cm⁻¹) contributed to the positive values in PC1, while vibrational stretches of = C–H (3009 cm⁻¹) and the EE carbonyl stretch (1734 cm⁻¹) contributed to the negative values in PC1 (Fig. 4).

Fig. 4.

Representative ATR-FTIR spectrum of a TG, b EE, and c first loading of the PCA model

The score plot showed that ten samples (71%) have ω-3 PUFA as TG and four (29%) as EE. The loading plot showed that EE carbonyl is correlated to the sp² carbons signal, indicating that EE has a higher concentration of unsaturated fatty acids than TG. Comparative studies about bioavailability and effectiveness of ω-3 supplementation have demonstrated similar absorption of EPA and DHA from ω-3 supplements by organisms in their TG and EE forms. However, information on absorption from other ω-3 PUFA has been conflicting to some extent. Some studies have shown no difference in absorption (Hansen et al. 1993; Krokan et al. 1993; Nordøy et al. 1991), while others have suggested a lower absorption of ω-3 PUFA in the EE form (Boustani et al. 1987; Lawson and Hughes 1988; Lopez-Toledano et al. 2017a, b; Maki et al. 2017). Regardless of these debates, consumers have the right to know detailed information about ω-3 supplements in order to make an informed choice, but neither information about the type of esterification is present in the product label.

The main vantage of ATR-FTIR, when compared to traditional gas chromatography, is that the analysis can be carried out directly on the oils without the need for sample preparation, providing faster and lower cost analyzes. Therefore, ATR-FTIR may be used to provide a spectral profile of the ω-3 PUFA in supplement capsules for quality control or authenticity analyses.

Relative concentrations of DHA, EPA, and ω-3 PUFA

The relative concentrations (mol/mol) of DHA, EPA, and ω-3 PUFA were determinate by ¹H NMR spectra (Table S2). DHA was quantified by the α-methylene protons (I_I, ca. 2.37 ppm), which are upfield from the α-methylene protons from other FA (I_H, ca. 2.31 ppm) due to the inductively withdrawing nature and anisotropic effects from the carbonyl group and the C4 double bond. Only ω-3 hexadecatetraenoic acid has a methylene chemical shift similar to the methylene protons of DHA. However, its concentration is generally found to be < 0.5% in fish oils, which allows the concentration of DHA to be determined by high-resolution ¹H NMR (Ando et al. 1989; Igarashi et al. 2002). EPA was quantified by the β-methylene protons (I_F, ca. 1.75 ppm), that differed from other FA (I_E, ca. 1.65 ppm) (Tyl et al. 2008). The ω-3 PUFA differs from other FA because of the anisotropic effect from the pi-electrons in the double bond. The inductively withdrawing nature of the sp²-carbons moves the methyl proton of ω-3 PUFA (I_B, 0.95 ppm) downfield by 0.10 ppm relative to those of non-ω-3 PUFA (I_A, 0.85 ppm) (Bratu et al. 2013).

The values achieved by ¹H NMR were compared with those declared on the product label. The analysis revealed a good agreement between the values obtained and declared on the product label, with the exception of one sample. In sample 3, DHA was not detected, while the product label mentioned 13.0% of this compound; the proportion of EPA found in this sample was 0.87 ± 0.01%, significantly less than declared (20.0%). However, this sample showed a high level of ω-3 PUFA, which can indicate that this sample is rich in vegetable oils, which has elevated levels of linoleic acid, but low levels of EPA and DHA. Due to the possibility that sample three has been adulterated, it was excluded from the following discussion.

The results showed great variability in the relative concentrations of DHA, EPA, and ω-3 PUFA from different brands. Relative concentrations of DHA ranged from 9.01 ± 0.14% (sample 5) to 48.03 ± 0.34% (sample 12). The relative concentrations of EPA ranged from 8.85 ± 0.02% (sample 6) to 61.71 ± 2.43% (sample 10). The relative concentrations of ω-3 PUFA ranged from 28.29 ± 0.11% (sample 6) to 92.62 ± 0.10% (sample 10). The results also showed that samples commercialized as EE have higher levels of DHA (24.33 ± 16.95%), EPA (35.37 ± 18.19%), and ω-3 PUFA (72.87 ± 21.55%), compared to TG samples (DHA 12.20 ± 2.56%, EPA 14.99 ± 6.28%, and ω-3 PUFA 33.27 ± 6.21%). This fact is associated with processes for sample concentration, as previously discussed. It is well established that TG samples have more concordant values of DHA, EPA, and ω-3 PUFA than EE samples. Supplements submitted to the sample concentration process by transethylation are more heterogeneous than natural TG samples (have higher standard deviations of average). The heterogeneity of DHA, EPA, and ω-3 PUFA in EE samples may be associated with differences in the concentration process, such as the use of fractional distillation or urea complexation (Kralovec et al. 2012).

Sample 14 corresponded to a ω-3 supplement produced from Schizochytrium sp. microalgae oil. It had higher levels of DHA (34.97 ± 1.73%) and ω-3 PUFA (42.16 ± 1.03%), but a lower level of EPA (1.42 ± 0.90%) than naturally found in fish oil. The most common fishes used for PUFA extraction are cod, anchovy, herring, and sardine (Derner et al. 2006). However, the extraction process of fish PUFA has disadvantages such as unpleasant odor, low stability, the presence of cholesterol, heavy metals contamination, and variable production (Medina et al. 1998). Furthermore, the use of microalgae oils seems to be a more sustainable alternative because the cultivation occurs in a controlled environment (Jenkins et al. 2009). The chemical composition of microalgae oil is simpler, with PUFA concentrations between 25 and 60% of the total lipids (Zittelli et al. 1999). In this way, the microalgae PUFA purification process is convenient and does not require transesterification of TG to EE.

Conclusion

The present work examined the chemical profiles of fourteen samples of ω-3 supplement capsules by ATR-FTIR and ¹H NMR spectroscopy. Analysis utilizing ATR-FTIR revealed that ten samples corresponded to the ω-3 PUFA as TG and four as EE. Since the literature has demonstrated advantages of ω-3 supplements in both forms (TG or EE), the consumer should have the right to choose which form of the supplement to consume, but this information is rarely declared on the product label. Therefore, ATR-FTIR can be routinely employed to classify the type of ω-3 PUFA during quality control and/or authenticity analyses. ¹H NMR revealed that the EE samples normally have higher levels of DHA, EPA, and ω-3 PUFA than TG, due to the concentration process performed by the manufacturer. The analysis revealed a good agreement between the obtained and declared values on the product label, except for one sample. The analysis of ω-3 supplements from Schizochytrium sp. microalgae oil revealed that microalgae oil has higher levels of DHA and ω-3 PUFA, but lower level of EPA than fish oil.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ATR-FTIR

Attenuated total reflectance-Fourier transform infrared spectroscopy

DHA

docosahexaenoic acid

EPA

eicosapentaenoic acid

EE

ethyl esters

FA

fatty acid

GC

gas chromatography

NMR

nuclear magnetic resonance

ω-3

omega-3

PUFA

polyunsaturated fatty acid

PCA

principal component analysis

PC

principal components

SNV

standard normal variate

TG

triacylglyceride