Research and analysis

Review of methods applied in the ‘From the Hives’ survey on honey authenticity

Published 13 February 2025

Background

The European Commission (EC) ‘From the Hives’ report, published in 2023, presented the results of a survey investigating possible adulteration of honey with added sugars.

The report focused on honey imported into the European Union (EU) via border control posts. A total of 320 samples were taken, originating from 20 countries, including 13 samples from the UK. Samples were sent to the EC’s Joint Research Centre (JRC) for analysis aimed at detecting the presence of exogenous sugar syrups using several different methods and markers.

Defra and the Food Standards Agency (FSA) subsequently sought an independent scientific expert opinion on the analytical approach taken in the survey including interpretation of testing results for UK samples. A small Authenticity Methodology Working Group (AMWG) technical subgroup, with honey testing, referee analysis and Isotope Ratio Mass Spectrometry (IRMS) expertise was established. The group was asked to provide their expert opinion on the overall approach taken and suitability of the methods used as part of the EC honey survey, also advising on how honey testing approaches could be strengthened in the future.

The group framed discussions around the level of analytical confidence in the survey approach from a testing perspective – reviewing evidence on each of the methods, markers and thresholds applied in the survey and considering how different honey production practices and different honey types might impact on interpretation of results.

This paper summarises the conclusions reached and recommendations made by the group. This opinion was used to inform follow-up for the UK honey samples and UK government’s formal response to the EC.

Summary of the AMWG technical subgroup views

The group raised concerns over the use of the term ‘suspicious’ in the report and the different ways this could be interpreted, noting that the term suspicious is generally not considered as proof. They recommended applying the phrase ‘flag for further investigation’ in relation to results where a potential marker of adulteration was detected and where follow-up would be worthwhile.

They agreed that having relevant information on the honey type and origin of the samples was invaluable when interpreting the testing data. This information facilitated a clearer understanding of why specific markers might be present and whether this was a result of likely adulteration or potentially as a result of some other production practice (such as carryover of bee feed or mechanical moisture reduction).

The group noted that performance characteristics such as limit of detection (LoD) and limit of quantification (LoQ) were not supplied for the tests used to detect difructose anhydride (DFA), 2-acetylfuran-3-glucopyranoside (AFGP) and mannose. This made it challenging to provide meaningful interpretation of the analytical data for these tests as in some cases there is evidence that these markers can be naturally present in some honeys.

Without quantitative testing data for these methods or information on the method limits of detection, it was difficult to assess whether the markers could be present legitimately in these samples or whether they are at levels that suggest addition of exogenous sugars.

The group reviewed the published evidence around the methods used in the survey and the markers or thresholds used as the basis for adulteration testing in the survey. In some areas they noted that published data were scant to support or refute the use of specific markers (particularly around oligosaccharide degree of polymerisation and DFAs) and identified a need for more research and data on the occurrence of these and other putative markers for adulteration. On reviewing the results for the 13 individual UK samples, the group agreed that the weight of analytical evidence varied for each sample, as more weight could be given to some tests than others. 

The group made several recommendations around future honey sampling and testing approaches that could form the basis of any future UK surveillance and provided views on which tests could be given more weight. The group considered EA/LC-IRMS analysis, LC-HRMS analysis for AFGP, and the usual tests for honey quality the most important to apply. Less weight should be given to tests looking at oligosaccharide chain length and mannose because of the potential for these to be found naturally in honey. DFA analysis should be included as there is no firm conclusion either way on the weight of this test at present.

AMWG technical subgroup views, conclusions and recommendations

Response to question 1: What is the level of analytical confidence in the approach from a testing perspective, are the conclusions arrived at valid based on the analysis undertaken?

The group discussed each method applied in the survey in turn and the evidence base around the markers and thresholds used as the basis of these methods. The group agreed to frame their discussion around the testing methods in the context of European and European-type honey versus mechanically moisture-reduced honeys, as this has an impact on how results are interpreted.

Members defined moisture reduction for the purposes of this part of the discussions as either:

• an immature honey that has not matured in a hive and moisture is then reduced mechanically (as with much of the Chinese honey produced)
• honey that has been produced in a tropical, humid environment where the moisture will not naturally reduce and mechanical moisture reduction outside the hive is undertaken

Moisture reduction can occur to different extents depending on the methods used which vary in different countries. Moisture reduction can be an important part of honey production to maintain product shelf life, quality and safety.

C4 and C3 sugar syrup detection by Elemental Analyser/Liquid Chromatography – Isotope Ratio Mass Spectrometry (EA/LC- IRMS) Analysis

EA/LC-IRMS is a technique for the detection of added sugars by the determination of the ratios of stable carbon isotopes in the bulk honey, in its various carbohydrates and in an extracted protein fraction used as an internal standard.

Plants synthesise sugars by 2 main chemical pathways during photosynthesis. Most plants use a ‘C3’ route, fewer plants use a ‘C4’ route. The nectar collected by bees is largely from C3 plants whereas added sugars may originate from C4 plants such as maize (corn) or sugar cane or C3 plants such as beet sugar and rice syrups.

Differences between the stable carbon isotope ratios of the separated carbohydrate and proteinaceous fractions of pure honey are distributed about zero with an assumed normal distribution the spread of which depends on natural biological variation and the measurement uncertainty of the measurement. Differences above these combined benchmarks are usually good indications of the addition of exogenous sugars. The stable carbon isotope ratios (‘delta values’) of interest in honey analysis occur on a negative scale. Given the above, detection of C3 sugars in honey has traditionally been more challenging for analysts.

The group agreed on the following conclusions:

• The benchmark figures applied in the analysis were derived from a database of 451 samples sourced from traders and beekeepers and the study authors took appropriate steps to verify the authenticity of and characterise those samples. However, the differences in C isotope values for protein and sugar fractions could be further influenced by natural biological variation, and this should be explored further.
• There can be a process with high methylglyoxal honey (such as Mãnuka honey) which will result in depletion of the carbon 13 in the proteinaceous fraction. Therefore, Mãnuka samples flagged as suspicious using EA/LC-IRMS need careful consideration, particularly as the original 451 sample database did not appear to include any genuine Mãnuka honey.
• The measurement uncertainty aspects of the method are accepted as they are based on an interlaboratory trial. However, care is needed to ensure calibration to the same standard when comparing carbon isotope delta values obtained by LC-IRMS to those obtained by EA-IRMS as this is prone to hidden biases that might well vary in extent between laboratories. This would clearly affect the magnitude of the thresholds for differences between honey protein and the honey sugars. Comparability of any measurement results requires measurands that are traceable to the same reference.
• Nevertheless, with the above caveats, a finding indicating potential sugar addition to honey, by EA/LC-IRMS should not be dismissed, it requires further investigation of the honey.

Oligosaccharides detection using High Performance Anion Exchange Chromatography – Pulsed Amperometric Detector (HPAEC-PAD) and Liquid Chromatography – High Resolution Mass Spectrometry (LC-HRMS)

Oligosaccharides have been used as a marker for sugar syrups as they arise in syrups through enzymatic hydrolysis as well as thermal hydrolysis. However, the extent to which oligosaccharides occur naturally in honey is poorly understood as there is little published research in this area. Honey that has been in the hive for a period of time (matured) has been exposed to enzymatic activity and the oligosaccharides that naturally came from nectar have been broken down into mono-saccharides, di-saccharides, tri-saccharides and short chain oligosaccharides. However, if the honey has not been in the hive for any great length of time it will not have been in contact with the enzymes for long and there will not have been time for the breakdown of oligosaccharides to occur.

The group agreed on the following conclusions:

• For European and European-style honey matured in the hive, the presence of oligosaccharides can be a flag for further investigation depending on the number and concentrations of oligosaccharides present as they do occur naturally in honey at unknown levels.
• Oligosaccharides may persist in mechanically moisture-reduced honeys, particularly immature honey, owing to reduced contact time with bee enzymes.
• There may be a homologous series sequence appearance to the chromatograms of syrup-derived oligosaccharides, which does not tend to be present in natural honey-derived oligosaccharides.
• A honey blend with a low percentage of mechanically moisture-reduced or immature honey in the blend may have oligosaccharide profiles within expected natural variation. However, if the blend is a more even mix of European-style and moisture-reduced or immature honeys presence of oligosaccharides does not necessarily indicate adulteration, but would be a flag for further investigation. Without data relating to the levels of oligosaccharides in honey and analytical parameters such as LoD and LoQ it is not possible to comment further.

2-acetylfuran-3-glucopyranoside (AFGP) detection by LC-HRMS

AFGP, referenced in the literature as a marker for rice syrup, is more commonly known as SMR (specific marker for rice). It was one of the first tests introduced for detecting rice syrup adulteration of honeys and is a good marker for rice syrup addition. The amount of the marker substance varies in different rice syrups which makes developing a quantitative test difficult. There are rice syrups with very little of the marker and some with a lot, depending on how the syrup was manufactured and what raw material was used.

The group agreed on the following conclusions:

• Members agreed that if the AFGP marker is detected in honey it should be a flag for further investigation, as it is not reported to occur naturally in honey. It may however be present due to carryover of bee feed and therefore becomes a matter of interpretation.
• If the LoD of the method was known, it would facilitate making a reasonable judgement on whether the analysis undertaken was suitable for making a conclusion about adulteration.
• AFGP does appear to be a good marker. However, there can be cross-contamination at trace amounts, sub 2 milligrams per kilogram (mg/kg), arising from many routes. This should not necessarily trigger the assumption that exogenous sugars are present as a result of addition of syrup. Nevertheless, finding of AFGP at low mg/kg amounts certainly flags honey for further investigation.
• For any given level of AFGP found in a honey sample, this could arise from a small amount of a high marker-containing syrup or a large amount of a low marker-containing syrup. If any kind of indication of the quantity of syrup is given it should be heavily caveated and a quantitative range should be given, making it clear that there is equal likelihood of the results being at the low end or high end of the range because the amount of marker in the original syrup is unknown.

Difructose anhydride (DFA) detection by LC-HRMS

DFAs are formed as part of the caramelisation process (heating of fructose). There appears to be limited research in the literature on the occurrence of DFAs in honey including in Asian or mechanically moisture-reduced honeys, so the range that could be expected in honeys for legitimate reasons is not known.

Any amount of excessive heating of honey could potentially lead to formation of DFAs and there is the potential for mechanical moisture reduction to expose honey to heat, particularly the honey film close to the evaporators or tubing when vacuum evaporation is used, which could lead to an amount of caramelised honey being present. No LoD was published for this method in the JRC survey report and therefore it’s unclear what levels of DFAs are leading to samples being classed as suspicious. Because of the lack of published information, DFA is a marker that flags further investigation, but it is difficult to assign a weight to it.

A database of syrups and honeys containing the ranges of these markers would be valuable and would allow assessment of any correlations between markers to assist in interpreting results. For example, there may be a correlation between increased hydroxymethylfurfural (HMF) and DFA levels if the honey had been heated. However, if the heating has occurred as part of the manufacturing process for the sugar syrup, then you would not expect to see an increase in the HMF level.

The group agreed that the presence of DFAs is a flag for further investigation as it’s an indicator of a different practice or process having taken place, but is not necessarily indicative of sugar syrup adulteration.

Mannose detection by LC-HRMS and Proton Nuclear Magnetic Resonance Spectroscopy (1H-NMR)

Mannose is known to be naturally present in honey, in the literature its presence has been demonstrated in Linden and Chestnut honeys. It is fairly ubiquitous in insects and may therefore be in animal cells found in honey. It is also a breakdown product of galactomannans in yeast (honeys with high yeast contamination are more likely to contain mannose as yeasts are composed of galactomannans, which could theoretically degrade to mannose as the moisture is reduced, especially at elevated temperatures).

Without quantitative results or information on the LoD used in the JRC analysis, interpreting the testing results in the survey is difficult. Interpretation of a positive mannose result would depend on the level of mannose present, the honey origin, and the origin in terms of other processing it has undergone.

The group agreed that the presence of mannose is a flag for further investigation, but the weight accorded to the presence of mannose is low owing to the possibility of natural occurrence from various routes. Therefore, the finding of mannose alone, without supporting analytical or other evidence, would not make it a priority for further investigation.

General observations

The JRC should think carefully about the language used to describe the outcomes of the tests that were applied. The use of the word suspicious may be interpreted in different ways.

Bearing in mind that data from a number of the tests applied would be interpreted in relation to established thresholds, we suggest that the JRC report method performance characteristics (including limit of detection, LoD and limit of quantification, LoQ). This would enable an outside reader to make a more informed assessment of the analysis undertaken and how any conclusions have been reached.

Bee feeding

Initial enquiries did not uncover any specific regulations covering bee feeding practices in the UK or other countries. It was noted the UK honey regulations require that honey should not have any food ingredients or food additive added to it and this would include exogenous sugars.

The group agreed on the following conclusions:

• Bee feeding is acknowledged to be a practice that is widely carried out and is often necessary to maintain bee health and welfare.
• It is known that carryover of bee feed can persist in the hive. It was suggested by experts in the group that this can be for up to 8 weeks, however there is no published evidence to support this time frame.
• Bee feeding can be problematic for different honeys depending on the transition time between the last bee feeding activities, nectar flow onset and honey gathering.
• Because of the variety of sugar syrups and their diverse composition, it will be difficult, for the foreseeable future, to establish a threshold for sugar syrup markers that takes into account potential carryover from legitimate bee feeding.
• Nevertheless, a level below 5-10% sugar syrup in honey is unlikely to be due to economically motivated adulteration and may in fact reflect carryover of sugar and associated markers from bee feed.
• Caution should therefore be taken to avoid over interpretation of results where sugar syrup markers are found with the caveat that bee feeding could be responsible.
• Further work to establish a database of sugar syrups and assess the implications of their composition on honey testing methods and interpretation of results would be welcomed.

Response to question 2: How could honey testing approaches be strengthened for future surveillance?

The group agreed on the following recommendations:

• For future honey surveillance in the UK, we need to carefully consider the analytical tests that could be applied bearing in mind what has been discussed with regards to the weight that can be attributed to each method. The most important tests to apply are the EA/LC-IRMS analysis, LC-HRMS analysis for AFGP, and the usual tests for honey quality such as moisture, HMF and diastase activity. Less weight should be given to tests looking at oligosaccharide chain length and mannose because of the potential for these to be found naturally in honey. DFA analysis should be included as there is no firm conclusion either way on the weight of this test at present.
• Interpretation of results should be undertaken in light of all the evidence captured. Nothing should be published in relation to findings until further investigation has been completed, prioritised by the strength of evidence in relation to sample results and the LoD’s and LoQ’s of tests applied and recorded.

References reviewed as part of this work

2-acetylfuran-3-glucopyranoside/ Specific Marker for Rice (AFGP/SMR)

Xue X, Wang Q, Li Y, Wu L, Chen L, Zhao J and Liu F, 2013. 2-Acetylfuran-3 glucopyranoside as a novel marker for the detection of honey adulterated with rice syrup. Journal of agricultural and food chemistry, 61(31), pages 7488-7493. 

Difructose anhydride DFA

Montilla A, Ruiz-Matute AI, Sanz ML, Martínez-Castro I and Del Castillo MD, 2006. Difructose anhydrides as quality markers of honey and coffee. Food research international, 39(7), pages 801-806. 

Ruiz-Matute AI, Rodríguez-Sánchez S, Sanz ML and Martínez-Castro I, 2010. Detection of adulterations of honey with high fructose syrups from inulin by GC analysis. Journal of Food Composition and Analysis, 23(3), pages 273-276. 

Mannose

Jia Liu, Xiaoping Zhang, Gang Duan, 2013. Study on Mannose Formation During High Fructose Syrup Production. Journal of Food Science and Biotechnology,32(10).

Missler J, Wiezorek T and Beckh G, 2016. Mannose: A marker for adulteration with syrup or resin treatment of blossom honey. In Proceedings of the XIII International Conference on the Applications of Magnetic Resonance in Food Science (pages 17-20).

Schievano E, Sbrizza M, Zuccato V, Piana L and Tessari M, 2020. NMR carbohydrate profile in tracing acacia honey authenticity. Food chemistry, 309, pages 1257-88.

Elemental Analyser/Liquid Chromatography – Isotope Ratio Mass Spectrometry (EA/LC-IRMS)

Aries E, De Rudder O, Kaklamanos G, Maquet A and Ulberth F, 2021. Results of an Interlaboratory Comparison of a Liquid Chromatography–Isotope Ratio Mass Spectrometry Method for the Determination of 13C/12C Ratios of Saccharides in Honey. Journal of AOAC International, 104(6), pages 1698-1702.

Beckmann K, Beckh G and Lüllmann C, Detection of honey adulteration with 13C isotope ratio mass spectrometry of single sugar fractions.

Cabanero AI, Recio JL and Ruperez M, 2006. Liquid chromatography coupled to isotope ratio mass spectrometry: a new perspective on honey adulteration detection. Journal of agricultural and food chemistry, 54(26), pages 9719-9727.

Frew R, McComb K, Croudis L, Clark D and Van Hale R, 2013. Modified sugar adulteration test applied to New Zealand honey. Food chemistry, 141(4), pages 4127-4131.

Lutz Elflein, Kurt-Peter Raezke. Improved detection of honey adulteration by measuring differences between 13C/12C stable carbon isotope ratios of protein and sugar compounds with a combination of elemental analyzer - isotope ratio mass spectrometry and liquid chromatography - isotope ratio mass spectrometry (_13C-EA/LC-IRMS). Apidologie, Springer Verlag, 2008, 39 (5), pages 574-587.

Rogers KM, Somerton K, Rogers P and Cox J, 2010. Eliminating false positive C4 sugar tests on New Zealand Manuka honey. Rapid Communications in Mass Spectrometry, 24(16), pages 2370-2374.

Rogers KM, Cook JM, Krueger D and Beckmann K, 2013. Modification of AOAC official method SM 998.12 to add filtration and/or centrifugation: interlaboratory comparison exercise. Journal of AOAC International, 96(3), pages 607-614.

Rogers KM, Grainger M and Manley-Harris M, 2014. The unique manuka effect: why New Zealand manuka honey fails the AOAC 998.12 C-4 sugar method. Journal of agricultural and food chemistry, 62(12), pages 2615-2622.

Rogers KM, Sim M, Stewart S, Phillips A, Cooper J, Douance C, Pyne R and Rogers P, 2014. Investigating C-4 sugar contamination of manuka honey and other New Zealand honey varieties using carbon isotopes. Journal of agricultural and food chemistry, 62(12), pages 2605-2614.

Group membership

• Michael Walker (Co-opted Independent Consultant) (Chair)
• Selvarani Elahi (Deputy Government Chemist, AMWG Chair)
• Adrian Charlton (Fera, AMWG member)
• David Hoyland (Co-opted Independent Consultant)
• Philip Dunn (LGC, Co-opted IRMS expert)