Limits of Food Authentication: A Separation Science Approach to Maximum Tolerable Levels

Nicolai Zederkopff Ballin from the Danish Veterinary and Food Authority examines chromatography’s role in helping to implement the maximum tolerable level (MTL) framework for food adulteration and translating analytical results into compliance decisions. Rather than measuring adulterants directly, chromatographic methods typically quantify secondary markers—such as 16-O-methylcafestol for Robusta-in-Arabica coffee fraud or species-specific peptides for meat substitution—which are then converted into an estimated content of the non-authentic component and compared against the MTL.

Ballin emphasizes that chromatography can detect and quantify but cannot establish intent; distinguishing deliberate adulteration from natural variation or accidental contamination falls to regulators and courts, not laboratories. Robust method validation (accuracy, precision, selectivity, recovery), appropriate limits of quantification (LOQ), and rigorous accounting for measurement uncertainty are essential, particularly when results are close to the MTL.

Targeted approaches such as high-performance liquid chromatography–mass spectrometry (HPLC–MS) and gas chromatography–mass spectrometry (GC–MS) suit routine enforcement with known markers and are often International Organization for Standardization (ISO)-standardized, whereas non-targeted fingerprinting methods better address complex questions like geographic origin, although they typically classify rather than quantify. Both approaches, when validated, feed into the same harmonized MTL, which combines an industrial baseline with a severity-grading system reflecting cultural, contextual, and sustainability expectations—making chromatography indispensable to transparent and consistent enforcement.

You recently published a paper called “Towards Harmonisation in Food Authentication: How to Establish Maximum Tolerable Levels of Food Adulterants”.¹ How can chromatographic techniques be used to detect and quantify non-authentic components in food samples in relation to the proposed MTLs?
Chromatographic techniques are commonly used to quantify primary or secondary markers of non-authentic components. For example, chromatographic analysis of the secondary marker 16-O-methylcafestol (16-OMC) can be used to estimate the proportion of Robusta coffee in Arabica coffee, while species-specific peptides can be used to estimate the proportion of a non-authentic meat species. The chromatographic result must therefore be translated, using an established relationship, into an estimated content of the non-authentic component. This provides a common and understandable entity that can be compared to the MTL, i.e., the highest level of a non-authentic component considered tolerable. The MTL is derived as a balance between industrial realities and consumer expectations.¹

How can chromatographers distinguish between genuine natural variation in food composition and the presence of intentional adulteration using chromatographic analysis?
The short answer is that they cannot. Chromatographic analysis alone cannot determine intent. Laboratories measure the concentration of a marker, but the analytical result does not reveal whether its presence results from deliberate substitution, accidental mixing, or cross-contamination. Natural variation is particularly relevant when a secondary marker is used to estimate the content of a non-authentic component, because the marker concentration may vary substantially between authentic reference materials.²

The relevant question is therefore whether the estimated content of the non-authentic component exceeds the MTL. In the proposed framework, non-authentic components include both intentional adulteration and unintentional process-related admixtures. The assessment of intent, economic motivation, and potential fraud requires additional evidence and is a matter for competent authorities and the judicial system rather than the analytical laboratory.

How do limits of detection (LOD) affect the ability of chromatographic methods to determine whether an adulterant exceeds its maximum tolerable level?
A result above the LOQ is normally required to quantify a secondary marker and translate it into an estimated content of a non-authentic component. LOD may be sufficient for qualitative detection or screening, whereas LOQ is required when a quantitative compliance decision is needed.

The calculation would be straightforward if Robusta coffee always contained a fixed concentration of 16-OMC. However, 16-OMC concentrations vary between Robusta coffees. This natural variation must be included when translating the marker concentration into a proportion of Robusta coffee, and the translation should be conservative in order to minimize the risk of falsely declaring a compliant sample non-compliant. Consequently, the method LOQ must be low enough to allow reliable quantification of the marker at concentrations relevant to the MTL. The choice of marker, chromatographic technique, and appropriate method validation are therefore critical.

How should measurement uncertainty in chromatographic analysis be considered when a measured concentration of an adulterant is close to the proposed maximum tolerable level?

As mentioned above, in many cases the laboratory does not measure the adulterant directly, but rather a secondary marker that is translated into the content of a non-authentic component.

Measurement uncertainty (MU) must be determined and included when translating the secondary marker into the content of the non-authentic component. If a conservative concentration of the secondary marker is applied in the translation to the non-authentic component, the MTL can be regarded as a tolerance level that does not require an additional confidence interval for the final compliance assessment. This approach reduces the risk of falsely declaring a compliant product non-compliant.

How can chromatographic method validation, including accuracy, precision, selectivity, and recovery, support the reliability of food-authentication results used for regulatory decisions?
Chromatographic methods are essential in food authentication because they can provide quantitative results for both primary and secondary markers. Primary markers include, for example, prohibited Sudan dyes in spices, whereas secondary markers can include compounds or peptides used to estimate the proportion of a non-authentic component.

Appropriate method validation is required to demonstrate that the method is fit for purpose. For secondary markers, the conversion from marker concentration to the content of the non-authentic component must also be scientifically justified and validated. Where results are intended for official control, the validated analytical method should preferably be operated under an appropriate accredited scope.

To what extent can targeted and non-targeted chromatographic approaches, such as HPLC–MS and GC–MS, contribute to the harmonization of food-adulteration testing across different laboratories and jurisdictions?
Targeted and non-targeted chromatographic approaches are complementary. Targeted chromatographic methods are particularly useful when the relevant marker is known. They can provide high selectivity, low quantification limits, and quantitative results that can be converted into an estimated content of a non-authentic component and compared with a maximum level or an MTL. Such methods are well suited for routine enforcement, and several are standardized through ISO or other standardization bodies. Different analytical approaches may be used to estimate the percentage of a non-authentic component, but the resulting estimate should be assessed against the same harmonized MTL.

We have previously proposed that fingerprinting should be regarded as non-targeted, whereas profiling should be regarded as targeted.³ Both approaches can measure numerous compounds or data points and may therefore support more complex authentication questions, including geographical origin or production system, such as conventional versus organic production.

At present, many non-targeted approaches rely on classification models that assign a sample to a reference group or identify it as atypical. Such models do not necessarily provide a percentage of a non-authentic component. However, where robust quantitative models are developed and validated, chromatographic data may also be used to estimate the content of the non-authentic component for comparison with an MTL.

How could chromatographic data be integrated with the proposed low-, moderate-, and high-severity grading system to establish scientifically defensible MTLs for food adulterants?
Chromatographic methods form an integral part of the workflow from analysis to the final estimate of the non-authentic component that is compared with the MTL. The chromatographic result provides the quantitative analytical evidence, whereas the MTL is established from an industrial baseline level (BL) combined with a severity grading system that balances industrial realities with consumer expectations.

The severity grading reflects how strongly the non-authentic component conflicts with cultural, contextual, and sustainability expectations. Under the proposed framework, increasing severity reduces the MTL relative to the BL. Thus, chromatography is indispensable for operationalizing MTLs and assessing compliance in food authentication. The MTL framework supports more transparent and consistent regulatory enforcement.

References

1. Ballin, N. Z.; Laursen, K. H.; Mihailova, A.; Riedl, J. Towards Harmonisation in Food Authentication: How to Establish Maximum Tolerable Levels of Food Adulterants? Trends Food Sci. Technol. 2026, 170, 105609. DOI: 10.1016/j.tifs.2026.105609
2. Bayen, S.; Elliott, C.; Arlorio, M.; et al. Towards a Harmonized Approach for Food Authenticity Marker Validation and Accreditation. Trends Food Sci. Technol. 2024, 149, 104550. DOI: 10.1016/j.tifs.2024.104550
3. Ballin, N. Z.; Laursen, K. H. To Target or Not to Target? Definitions and Nomenclature for Targeted versus Non-Targeted Analytical Food Authentication. Trends Food Sci. Technol. 2019, 86, 537–543. DOI: 10.1016/j.tifs.2018.09.025