PFAS Residues in Animal-Based Food

Per- and Polyfluoroalkyl Compounds (PFAS) are synthetic chemicals that are widely used due to their extraordinary properties. The added value they offer in daily life is significant, and they are long-lasting and highly resilient. Nevertheless, PFAS are harmful to our health, making them objects of scrutiny by food safety and health authorities. We previously reported on the determination of PFAS in drinking water. In this article, we take aim at PFAS in animal-based food for human consumption. To reach the required low levels of determination in a sustainable manner, the matrix requires highly efficient sample preparation, as described in GERSTEL AppNote 247[3]. Using the described sample prep solution system and associated method, PFAS is determined in Food in accordance with the EU Guidance Document on Analytical Parameters for the Determination of Per- and Polyfluoroalkyl Substances (PFAS) in Food and Feed, Version 1.2, 11 May 2022.). The described system meets the latest requirements with lower maximum levels for perfluoroalkyl substances in food of animal origin of 0.2 μg/kg set forth in Commission regulation (EU) 2020/2388.   
Mandatory labeling of food enables consumers to distinguish nutritious and healthy products from other ones that many may prefer. Among the things the labels that decorate our food packaging don’t provide information about, however, is residues of chemical contaminants. The risk of contamination is always tangible when food is produced using chemical ingredients or is processed industrially. Further, contaminants can find their way into our food delivery chain through contamination of animal feed, pastures, or water. Even the packaging must be rounded up with the usual suspects. The legislative takes all these facts into account by issuing guidelines and setting maximum acceptable values for contaminants. However, as is regularly demonstrated by headline grabbing events, consumer protection requires monitoring. And monitoring for PFAS requires both expert knowledge and high-performance analytical technology. 

In the crosshairs of inspectors

Food safety monitoring in its early days focused mainly on compliance with hygiene requirements. Over the past decades, significant and increasing attention has been paid to potential contamination with harmful residues of chemical substances used in food production, such as pesticides in crops or active pharmaceutical ingredients used to treat farm animals. Substances that are hazardous to human health can even be formed during storage or transport of food, including molds and the toxins they form. Further, process contaminants can be formed or added during processing of food and beverages, and hazardous compounds can even migrate from packaging materials into the food product. Per- and Polyfluoroalkyl Compounds (PFAS) are synthetic chemicals that are widely used due to their extraordinary properties. The added value they offer in daily life is significant, and they are long-lasting and highly resilient. Nevertheless, PFAS are harmful to our health, making them objects of scrutiny by food safety and health authorities. In addition, their solubility in water means that they are transported distributed into the environment at large and their persistence lets them accumulate in pockets of the environment including in the food chain. We previously reported on the determination of 20 especially toxic PFAS compounds in drinking water in accordance with the EU Water Framework Directive 2020/2184. In this article, we take aim at PFAS in animal-based food for human consumption. To reach the required low levels of determination in an efficient and sustainable manner, the matrix requires highly efficient sample preparation.

The Chemistry of PFAS

PFAS are man-made chemical compounds tailored for use in numerous domestic and industrial applications. PFAS are used as additives in food packaging, cookware and carpets, clothing, cleaning agents and fire-fighting foams. A look at the molecular structure explains the wide range of applications: PFAS are highly fluorinated organic chemicals based, among other things, on carboxylic and sulfonic acids with a chain length of C4 to C18. A distinction is made between perfluorinated alkyl sulfonates (PFAS) with perfluoro octane sulfonate (PFOS) as the most well-known representative, and perfluorinated carboxylic acids (PFCA), among which perfluorooctanoic acid (PFOA) is the most notorious. PFAS are synthesized by substituting hydrogen atoms with fluorine atoms, resulting in major changes in chemical properties: the PFAS carbon chain is hydrophobic, the head of the molecule is hydrophilic. The amphiphilic character explains the use of some PFAS as surfactants. Compared to classic surfactants, however, the carbon chain of the PFAS is also lipophobic: it not only repels water, but also oil, grease, and dirt. In addition, PFAS are extremely heat-resistant and water-soluble, which has enabled them to spread in ground and surface water – our most important drinking water reservoirs. PFAS can accumulate in pockets of the environment, including the food chain. 

The Sample determines the Analysis Strategy

Drinking water is not only used for human consumption, in many cases it is used to water animals whose meat and eggs we eat and whose milk we drink. And even when animals are watered using surface or well water, based on the distribution of PFAS in the environment, it is only logical to assume that PFAS will also be present in milk, eggs, meat, and fish. This has made PFAS a target of food inspectors. The analysis of food of animal origin to determine whether it contains PFAS is not just an option, it is active consumer protection and is therefore a requirement.
There are parallels, of course, when determining PFAS in different sample types. But analyzing the complex matrix of food requires a different approach than the one needed when analyzing water. To determine PFAS in food, interfering sample matrix must be removed. In other words, the compounds are extracted, and the extract is prepared for LC-MS/MS analysis. The QuEChERS method [2] has established itself as the method of choice for this process. Originally, QuEChERS was developed for extraction of low-fat fruit and vegetable samples. Thanks to regular modifications and new implementations, it has now also become the method of choice for extracting complex matrices, such as cereals, and fatty foods, such as oils, milk, and meat products. Following extraction and QuEChERS clean-up, PFAS are determined in the extract by LC-MS/MS. 

Unlocking the Potential

The State Research Institute for Health and Veterinary Affairs (LUA) in Saxony, in cooperation with the Technical University of Dresden, and the GERSTEL Innovation and Technology Department in Mülheim an der Ruhr, all in Germany, have shown that there is significant optimization potential in the methods used for the determination of PFAS in food, in terms of both efficiency and sensitivity[3]. By modifying the analysis method and using updated technology, Sauer et al. succeeded in developing an elegantly automated method for the determination of PFAS, reaching lower limits of quantification while reducing solvent consumption and cost per analysis. The sample preparation is based on online SPE with replaceable cartridges coupled directly to HPLC-MS/MS. A dedicated online SPE module (SPExos II, Gerstel) based on cartridges with an ID of 1 mm was used. The cartridges are inserted into the HPLC mobile phase and can be eluted directly and thus quantitatively onto the HPLC column for optimum analyte recovery. 

Online, Connected - and Clean when the sample is loaded

Online SPE in combination with a suitable autosampler sets the stage for efficient, accurate, and reliable determination of PFAS. The SPExos II system, which was used in the work presented here, performs all the classical steps connected with SPE sample preparation: Conditioning, loading, rinsing, eluting, as well as replacing the cartridge as specified. “Just for the sake of completeness”, says Thomas Brandsch, Ph.D., "in the case of water samples, the autosampler we use (MultiPurpose Sampler, GERSTEL) flushes residues of PFAS compounds that are adsorbed on vial and tubing surfaces onto the replaceable SPE cartridge. This step improves analyte recovery significantly while reducing sample to sample carry-over to an absolute minimum". Dr. Brandsch is an Application Scientist in the GERSTEL Innovation and Technology Laboratories and co-author of the study [3]. After the analytes have been eluted, the SPExos II system removes the cartridge from the mobile phase flow path and prepares the system for the next analysis. The PrepAhead mode enables simultaneous HPLC-MS/MS analysis and preparation of the next following sample for best possible throughput and system utilization

PFAS Analysis, starting with Water

The method was developed using standards in aqueous solution and then validated by analyzing real samples. Sauer et al. relied on an automated method previously developed by Dr. Brandsch and his colleague Oliver Lerch, Ph.D. for the determination of 20 PFAS [1] listed in the EU Water Framework Directive 2020/2184. An autosampler (MPS robotic, GERSTEL) accurately loads sample to the online SPE system (SPExos II, GERSTEL), which is directly coupled to the HPLC-MS/MS system consisting of a 1290 Infinity II pump and a 6495C Triple Quadrupole Mass Spectrometer (both Agilent Technologies). The differences between the food and water analysis methods are essentially minor, but foodstuffs generally have complex matrices and require additional extract clean-up in a QuEChERS extraction process before online-SPE-LC-MS/MS analysis of the extract is performed. 

QuEChERS Extract Cleanup required

Sauer et al. validated their SPE-HPLC-MS/MS method using samples of animal-based foods: Eggs, meat, and fish were treated to a QuEChERS-based sample preparation process. An internal standard was first added to each five-gram food sample to monitor the extraction process. Each sample was then extracted twice with acetonitrile under alkaline conditions to ensure good PFAS extraction recovery. The phases were separated by adding sodium chloride, and the combined organic phases for each sample collected, acidified with formic acid, and frozen. The following day, the organic phase, in which the analytes are concentrated, was subjected to a cleanup process based on dispersive solid phase extraction using magnesium sulfate (MgSO4) and Envi-Carb cartridges. The resulting extract was concentrated by evaporation to 0.3 mL under inert atmosphere. Internal standards for the HPLC-MS/MS analysis were added, and the mixture topped up to a volume of one mL. The method was calibrated in the concentration range from 0.025 to 5 ng/mL (internal standards 1 ng/mL), corresponding to 0.005 to 1 µg/kg food in a 5 g sample. 

Sample Preparation and Sample Introduction 

A 25 µL aliquot of the extract was introduced using online SPE. The HPLC separation following the SPE was performed on a Poroshell 120 EC-C18 column, 3x100 mm, 2.7 µm, Agilent® Technologies. SPE cartridge conditioning and washing was performed in multiple steps: First, using 300 µL of a mixture of acetonitrile, acetone and formic acid (50/50/1, vvv) and secondly with 300 µL of 0.1- percent formic acid in methanol. For the final organic wash step, 300 µL of acetonitrile was used. For comparison, to document the cleanup and concentration effects of the SPE step, a 2 µL aliquot was introduced directly to the HPLC (Zorbax Eclipse Plus C18 2.1x50 mm, 1.8 µm, Agilent Technologies). As Sauer et al. report, the organic wash has a considerable effect on the analysis result. 

Analysis Details

Final cleanup of the QuEChERS extracts was performed automatically using online SPE cartridges (Polymer WAX, GERSTEL), which are significantly smaller than conventional SPE cartridges. Sauer et al.: “Typically, in online SPE, elution is performed with a solvent gradient that is provided by the analytical pump. However, the WAX cartridges are eluted with ammonia (NH3) in methanol (MeOH). That mixture cannot be used for transfer directly to the HPLC system since the analytes wouldn’t be retained on the column. In the method described here, an additional (isocratic) HPLC pump is used to elute cartridges and the resulting eluate combined with the starting buffer of the binary analytical mobile phase inside the SPExos.” When the SPE cartridge elution is complete, the chromatography starts. The binary pump delivers the necessary gradient: 0.1 percent formic acid in water (eluent A) and 0.25 percent ammonia and 0.05 percent formic acid in methanol (eluent B). While the HPLC separation is in progress, SPExos II starts preparing the next sample (PrepAhead).

A closer look at Matrix Effects

To assess possible matrix effects, Sauer et al. extracted real samples: Egg, fish and meat without adding internal standards. A larger amount of extract was prepared, which was then spiked with standard solution and internal standard in small portions to obtain a calibration series. Separate method calibrations, based on the individual sample types, were carried out identically and compared with method calibrations, obtained from solvent standards ranging from 0.1 to 2 ng/mL (equivalent to 0.01 to 0.2 μg/kg in food based on a five-gram sample). All solutions were injected both directly (2 µL) and using online SPE, and the analysis results were compared. To assess the influence of the organic cartridge wash, Sauer et al. also performed online SPE with and without it and compared the analysis results. According to the authors, switching between the two modes was easily performed using the SPExos system. Different sample volumes were injected for comparison, such as, for example, 25 μL by online SPE and 2 μL by direct injection to HPLC-MS/MS. The cleanup effect of the online SPE process makes it possible to inject larger amounts of QuEChERS extracts without experiencing matrix effects, as Sauer et al. reported. The comparison rested on the assumption that the peak area would increase by a factor of 12.5, which, as the authors report, was very close to the actual result: "Using 12.5 times more sample and online SPE resulted in ten to thirteen times larger peak areas compared with those obtained by direct injection." At the same time, the signal-to-noise ratio was improved. In addition, some analytes generated significantly larger peak areas when the samples loaded to the online SPE cartridges had been subjected to the organic wash. This is a clear indication that online SPE cleanup helps reduce matrix-induced ion suppression.

Eliminating Matrix-Induced Deviations

When a matrix matched calibration cannot be performed, a solution-based calibration is the quantification of choice, including corrections based on internal standards, as Sauer et al. explain. If an isotopically labeled analyte is available, the relative response (the ratio of the peak areas of the analyte and internal standard) can be determined independently of possible matrix effects. However, if the internal standard only resembles the analyte, the relative responses may vary, and accurate quantification may not be possible. The researchers cite perfluorohexadecanoic acid (PFHxDA) as an example, it is quantified with isotope-labeled perfluorotetradecanoic acid (13C2-PFTeDA): "Using direct injection of 2 μL leads to slight differences between those calibration curves that are based on solvent-based standards and those based on matrix-matched standards,” Sauer et al. report. When injecting 25 μL through online SPE without organic wash of the cartridge, the calibration curves obtained for the different matrices also differed from the calibration curve based on standards in solution. The researchers explain this phenomenon with matrix effects that arise when injecting significantly larger amounts of sample. The organic cartridge wash before elution significantly reduces those matrix effects, resulting in calibration curve slopes that are comparable with those obtained with calibration standards.

Limits of Quantification

In summary, the goal pursued by the authors of the study was to lower the limits of determination. Spiked samples at different concentration levels were analyzed in replicate to determine accuracy and repeatability of the analysis method. The lowest concentration at which certain analytical quality criteria can be met is regarded as the limit of quantification [4]. The researchers report that online SPE enrichment and purification enabled the injection of larger sample amounts along with a reduction in matrix effects. Without the organic cartridge wash step, the recovery of some compounds suffered, resulting in a lower signal increase than expected, for example in the case of 13C 2-PFTeDA, leading to increased deviations for PFHxDA and PFODA. When the cartridge wash was included, recovery of the internal standards increased significantly, reducing potential deviations. 

Conclusion

According to Sauer et al. the automated online SPE cleanup of food extracts enables the determination of PFAS compounds in the ng/kg range in foods of animal origin. The cleanup effect results in limits of detection of 0.01 µg/kg for most compounds, which means significantly lower than those achieved by direct injection. Some sulfonic acids exhibited more pronounced matrix effects resulting in higher limits of quantification (0.05 μg/kg). Perfluorooctanesulfonamide (PFOSA) and N-Ethyl-perfluorooctanesulfonamide (N-EtFOSA) could not be determined when an organic cartridge wash was performed. Without organic wash, limits of quantification of 0.05 μg/kg and 0.5 μg/kg respectively were reached.
The organic solvent cartridge wash before elution effectively removes matrix interference, improving the accuracy of the results, as the researchers report. The results they obtained make it clear that online SPE cleanup prior to HPLC-MS/MS determination greatly improves the results obtained for selected PFAS in animal foods.
The application expert Thomas Brandsch, Ph.D. will present his work in a webinar to be held on February 15, 2023, at 11 a.m.  CET. Participation in the online lecture, which is held in English, is free of charge and only requires registration at https://www.gerstel.com/en/onlineseminar-PFAS-Animal-Food-2023

Guido Deußing, Redaktionsbüro, Neuss, E-Mail: guido.deussing@pressetextkom.de

References

[1] Thomas Brandsch, Oliver Lerch, Determination of PFAS in Water according to EU 2020/2184 and DIN 38407-42 using online-SPE-LC-MS/MS. GERSTEL AppNote 237.
[2] Michelangelo Anastassiades, Steven J. Lehotay, Darinka Stajnbaher, Frank J. Schenck, Fast and easy multiresidue method employing acetonitrile extraction/partitioning and "dispersive solid-phase extraction" for the determination of pesticide residues in produce, J AOAC Int. 2003 Mar-Apr;86(2):412-31. 
[3] Claudia Sauer, Christin Pleger, Thomas Frenzel, Thomas Simat, Thomas Brandsch, Oliver Lerch, Determination of PFAS in Food of Animal Origin using online SPE Cleanup and LC-MS/MS, GERSTEL AppNote 247, 2022 https://www.gerstel.com/en/Determination_of_PFAS_in_Food_of_Animal_Origin_using_online_SPE%20Cleanup_and_LC-MS/MS 
[4] Guidance Document on Analytical Parameters for the Determination of Per- and Polyfluoroalkyl Substances (PFAS) in Food and Feed, Version 1.2, 11 May 2022.

Weight
1100