Lavender honey authenticity analysis
Publié le 04/03/2025
Product Quality
Anomalies uncovered
As part of a study on the quality of French honey, inconsistent results have been observed in the analysis of lavender-lavandin honey. This honey, known for its specificities and subject to a derogation for its sucrose content, has already required the adaptation of several existing analysis methods. At issue today is the “maximum difference” criterion, calculated in the LC-IRMS analysis.
Following warnings from several players in the beekeeping sector, the interprofessional organization InterApi has commissioned and funded a three-year project to study the quality of French honey and the reliability of analyses. This project began in June 2023, led by ITSAP-Institut de l’abeille, the French technical institute for beekeeping and pollination, in collaboration with the ADA network, local beekeeping development associations.
Lavender-lavandin honey is recognized for its qualities and specificities. It is covered by a derogation in the directive 2001/110/EC on honey concerning its sucrose content, which generally exceeds 5% but must remain below 15%. Its diastase level can be low, which would classify it amongst the honeys with a low natural enzyme content. Finally, this honey has low pollen content, with lavender pollen generally under-represented[1].
Following warnings from several players in the beekeeping sector, the interprofessional organization InterApi has commissioned and funded a three-year project to study the quality of French honey and the reliability of analyses. This project began in June 2023, led by ITSAP-Institut de l’abeille, the French technical institute for beekeeping and pollination, in collaboration with the ADA network, local beekeeping development associations.
Lavender-lavandin honey is recognized for its qualities and specificities. It is covered by a derogation in the directive 2001/110/EC on honey concerning its sucrose content, which generally exceeds 5% but must remain below 15%. Its diastase level can be low, which would classify it amongst the honeys with a low natural enzyme content. Finally, this honey has low pollen content, with lavender pollen generally under-represented[1].
Analysing the authenticity of lavender honey is therefore complicated due to these specific characteristics, and inconsistencies have been notified in recent years involving several analysis methods. The 1H-NMR profiling[2] was specifically studied in 2021-2022, through a collaboration between ADAPI[3] and several laboratories, in order to enhance the databases and improve the analysis of this honey. This work has paid off, as shown by the coherency of the results obtained since then. However, in 2023 and 2024, repeated inconsistencies in LC-IRMS[4] analysis were identified for lavender-lavandin honeys produced in Provence.
Analysing the authenticity of lavender honey is therefore complicated due to these specific characteristics, and inconsistencies have been notified in recent years involving several analysis methods. The 1H-NMR profiling[2] was specifically studied in 2021-2022, through a collaboration between ADAPI[3] and several laboratories, in order to enhance the databases and improve the analysis of this honey. This work has paid off, as shown by the coherency of the results obtained since then. However, in 2023 and 2024, repeated inconsistencies in LC-IRMS[4] analysis were identified for lavender-lavandin honeys produced in Provence.
Data: source and method
As part of the quality project led by ITSAP-Institut de l’abeille, 60 samples, including 20 lavender honeys produced in Provence in 2024 were screened. The traceability of each hive was rigorously monitored by ADAPI: beekeeping practices, locations, transhumance routes, feeding periods, quantities consumed, and the types of products used. For sampling, the first super of each hive was extracted manually. The samples were sent to 6 reference laboratories in France and Europe to analyse several quality criteria, with an emphasis on adulteration analysis: validation of the botanical designation using melissopalynological, sensory and physicochemical analyses, hydroxymethylfurfural content, diastase, invertase, EA-IRMS[5], LC-IRMS, 1H-NMR profiling and oligosaccharides.
The analyses were distributed amongst the laboratories and carried out blindly, i.e. without disclosing the traceability data beforehand.
As part of the quality project led by ITSAP-Institut de l’abeille, 60 samples, including 20 lavender honeys produced in Provence in 2024 were screened. The traceability of each hive was rigorously monitored by ADAPI: beekeeping practices, locations, transhumance routes, feeding periods, quantities consumed, and the types of products used. For sampling, the first super of each hive was extracted manually. The samples were sent to 6 reference laboratories in France and Europe to analyse several quality criteria, with an emphasis on adulteration analysis: validation of the botanical designation using melissopalynological, sensory and physicochemical analyses, hydroxymethylfurfural content, diastase, invertase, EA-IRMS[5], LC-IRMS, 1H-NMR profiling and oligosaccharides.
The analyses were distributed amongst the laboratories and carried out blindly, i.e. without disclosing the traceability data beforehand.
Furthermore, the Provence Miel cooperative has shared data from 51 drums of lavender honey produced in 2023. The production areas, feeding periods, quantities and types of products are also known. Monitoring was conducted at the apiary level. These samples were analysed using 1H-NMR profiling, EA-IRMS and LC-IRMS in a laboratory accredited by COFRAC to ISO 17025. The analytical plan follows two stages: an “average sample” is constituted by taking honey from several drums. If this sample does not comply, each drum is then studied individually. Drums for which the first “average sample” was deemed authentic were not examined individually.
Furthermore, the Provence Miel cooperative has shared data from 51 drums of lavender honey produced in 2023. The production areas, feeding periods, quantities and types of products are also known. Monitoring was conducted at the apiary level. These samples were analysed using 1H-NMR profiling, EA-IRMS and LC-IRMS in a laboratory accredited by COFRAC to ISO 17025. The analytical plan follows two stages: an “average sample” is constituted by taking honey from several drums. If this sample does not comply, each drum is then studied individually. Drums for which the first “average sample” was deemed authentic were not examined individually.
What are the results?
Out of the dataset collected by ITSAP-Institut de l’abeille, all 20 samples comply with the oligosaccharides and EA-IRMS analyses. The NMR detects exogenous sugars in only one sample. For these methods, the results are globally consistent between the laboratories and with regard to the traceability.
Adulteration detections with LC-IRMS were observed in 4 samples by laboratory D, with results falling within the measurement uncertainty zone, and in 7 samples by laboratory C. These results are not consistent between laboratories.
Out of the dataset collected by ITSAP-Institut de l’abeille, all 20 samples comply with the oligosaccharides and EA-IRMS analyses. The NMR detects exogenous sugars in only one sample. For these methods, the results are globally consistent between the laboratories and with regard to the traceability.
Adulteration detections with LC-IRMS were observed in 4 samples by laboratory D, with results falling within the measurement uncertainty zone, and in 7 samples by laboratory C. These results are not consistent between laboratories.
Amongst the non-compliant samples, samples 1 to 4 come from the same apiary and were fed similar quantities of sugar. However, only the first one is deemed authentic, while the others have variable results depending on the laboratory. Samples 6 to 20 all come from hives that had a chestnut harvest between the last feeding and the start of the lavender honey flow. For some, like samples 11 to 14, the last feeding occurred very early in the season, at the beginning of February. Yet, they do not comply with the LC-IRMS analysis.
Amongst the non-compliant samples, samples 1 to 4 come from the same apiary and were fed similar quantities of sugar. However, only the first one is deemed authentic, while the others have variable results depending on the laboratory. Samples 6 to 20 all come from hives that had a chestnut harvest between the last feeding and the start of the lavender honey flow. For some, like samples 11 to 14, the last feeding occurred very early in the season, at the beginning of February. Yet, they do not comply with the LC-IRMS analysis.
From the data collected by Provence Miel, results for each drum show that all complied with EA-IRMS and NMR analyses, while 27 drums, or 52%, failed the LC-IRMS analysis.
From the data collected by Provence Miel, results for each drum show that all complied with EA-IRMS and NMR analyses, while 27 drums, or 52%, failed the LC-IRMS analysis.
For 73% of these samples deemed adulterated, the result fell within the measurement uncertainty zone. Of these 27 drums, 6 were not fed at all in 2023, and 14 come from apiaries that had an intermediate harvest between the last feeding and the start of the lavender honey flow.
For 73% of these samples deemed adulterated, the result fell within the measurement uncertainty zone. Of these 27 drums, 6 were not fed at all in 2023, and 14 come from apiaries that had an intermediate harvest between the last feeding and the start of the lavender honey flow.
The limits of IRMS analysis
EA-IRMS analysis is based on measuring the isotopic carbon ratio of sugars and protein in honey, and the difference between the two. In the LC-IRMS method, chromatography is first used to separate the different sugar fractions in honey (fructose, glucose, disaccharides, trisaccharides, oligosaccharides), before measuring the isotopic ratio of each. Several criteria are then calculated: the fructose-glucose difference (the “F-G” criterion) and the maximum difference between the lowest and highest isotope ratio measurement amongst the sugar fractions (“maximum difference” or “delta max” criterion).
EA-IRMS analysis is based on measuring the isotopic carbon ratio of sugars and protein in honey, and the difference between the two. In the LC-IRMS method, chromatography is first used to separate the different sugar fractions in honey (fructose, glucose, disaccharides, trisaccharides, oligosaccharides), before measuring the isotopic ratio of each. Several criteria are then calculated: the fructose-glucose difference (the “F-G” criterion) and the maximum difference between the lowest and highest isotope ratio measurement amongst the sugar fractions (“maximum difference” or “delta max” criterion).
However, EA-IRMS and LC-IRMS analyses have limitations documented by several authors and can yield false positive, meaning non-compliant analysis results that are incorrect, for certain types of honey with atypical profiles. Many false positives have been observed, for example for manuka honey in New Zealand[6], citrus honey in the United States[7] or mesquite honey in Mexico. Honeys with a naturally low protein content or a high yeast level could also be prone to false positive in EA-IRMS analysis.
However, EA-IRMS and LC-IRMS analyses have limitations documented by several authors and can yield false positive, meaning non-compliant analysis results that are incorrect, for certain types of honey with atypical profiles. Many false positives have been observed, for example for manuka honey in New Zealand[6], citrus honey in the United States[7] or mesquite honey in Mexico. Honeys with a naturally low protein content or a high yeast level could also be prone to false positive in EA-IRMS analysis.
Regarding LC-IRMS, the maximum value set at ± 2.1‰[8] for the "maximum difference" criterion is debated and is not uniformly applied by all laboratories. Some extend this limit by adding measurement uncertainty, which can lead to different conclusions from one laboratory to another. A method was recently proposed to take this uncertainty into account when assessing honey purity[9].
Regarding LC-IRMS, the maximum value set at ± 2.1‰[8] for the "maximum difference" criterion is debated and is not uniformly applied by all laboratories. Some extend this limit by adding measurement uncertainty, which can lead to different conclusions from one laboratory to another. A method was recently proposed to take this uncertainty into account when assessing honey purity[9].
In the case of lavender-lavandin honey, work had already been carried out in 2017 by several laboratories to adapt the LC-IRMS to honeys naturally rich in sucrose. Large differences were observed between the isotopic ratio of disaccharides and monosaccharides, leading to numerous false positives. As a result, laboratories now exclude disaccharides from the "maximum difference" criterion calculation for this type of honey.
In the case of lavender-lavandin honey, work had already been carried out in 2017 by several laboratories to adapt the LC-IRMS to honeys naturally rich in sucrose. Large differences were observed between the isotopic ratio of disaccharides and monosaccharides, leading to numerous false positives. As a result, laboratories now exclude disaccharides from the "maximum difference" criterion calculation for this type of honey.
In the present study, for both datasets, we note that 100% of the non-conformities come from LC-IRMS analysis. They are all linked to the "maximum difference" criterion, and in 100% of cases, the high delta is due to trisaccharides, as disaccharides had already been excluded from the calculation.
In the present study, for both datasets, we note that 100% of the non-conformities come from LC-IRMS analysis. They are all linked to the "maximum difference" criterion, and in 100% of cases, the high delta is due to trisaccharides, as disaccharides had already been excluded from the calculation.
In conclusion
Based on the results of this study analysing lavender-lavandin honeys produced in Provence and the known traceability of each sample, we can estimate that the non-compliances obtained are cases of false positives. LC-IRMS analysis does not appear to be suitable for examining this type of honey. Given the current state of knowledge and to avoid an excessive number of biased results, we recommend setting aside LC-IRMS for the time being and relying on other existing methods to assess its authenticity. New analyses are planned for the 2025 production.
Based on the results of this study analysing lavender-lavandin honeys produced in Provence and the known traceability of each sample, we can estimate that the non-compliances obtained are cases of false positives. LC-IRMS analysis does not appear to be suitable for examining this type of honey. Given the current state of knowledge and to avoid an excessive number of biased results, we recommend setting aside LC-IRMS for the time being and relying on other existing methods to assess its authenticity. New analyses are planned for the 2025 production.
Collaboration with laboratories has already proven effective in adapting the 1H-NMR profiling analysis. It is necessary to continue improving methods and monitor the impact of lavender honey flow evolution on isotopic analysis results, in order to better understand the source of these inconsistencies and harmonize conformity criteria.
This is partly the aim of the work carried out by the JRC[10] as part of the HarmHoney project to harmonise analysis methods, to which ITSAP-Institut de l’abeille and the ADA network have contributed by sending authentic and traced honey samples.
Collaboration with laboratories has already proven effective in adapting the 1H-NMR profiling analysis. It is necessary to continue improving methods and monitor the impact of lavender honey flow evolution on isotopic analysis results, in order to better understand the source of these inconsistencies and harmonize conformity criteria.
This is partly the aim of the work carried out by the JRC[10] as part of the HarmHoney project to harmonise analysis methods, to which ITSAP-Institut de l’abeille and the ADA network have contributed by sending authentic and traced honey samples.
Notes
•[1] Loublier, Y., Piana, M.-L., Pham-Delègue, M.-H., Bomeck R. (1994). “Caractérisation pollinique des miels français de lavande: premiers résultats . » Grana 33: 231-238.
•[2] Nuclear magnetic resonance. The « NMR profiling » analysis consists of comparing the sample’s profile with a database.
•[3] Association for the development of beekeeping in Provence
•[4] Liquid chromatography coupled with isotope ratio mass spectroscopy.
•[5] Elemental analysis for isotope ratio mass spectroscopy.
•[6] Grainger, M. N. C. (2022). "C4 sugar adulteration methodology: Understanding false-positive results for mānuka honey." Food Chemistry Advances 1 : 100128.
•[1] Loublier, Y., Piana, M.-L., Pham-Delègue, M.-H., Bomeck R. (1994). “Caractérisation pollinique des miels français de lavande: premiers résultats . » Grana 33: 231-238.
•[2] Nuclear magnetic resonance. The « NMR profiling » analysis consists of comparing the sample’s profile with a database.
•[3] Association for the development of beekeeping in Provence
•[4] Liquid chromatography coupled with isotope ratio mass spectroscopy.
•[5] Elemental analysis for isotope ratio mass spectroscopy.
•[6] Grainger, M. N. C. (2022). "C4 sugar adulteration methodology: Understanding false-positive results for mānuka honey." Food Chemistry Advances 1 : 100128.
•[7] White, J. W. and F. A. Robinson (1983). "13C/12C Ratios of Citrus Honeys and Nectars and Their Regulatory Implications " Journal of AOAC International 66(1): 1-4.
•[8] Elflein, L. and Raezke, K.-P. (2008). “Improved detection of honey adulteration by measuring differences between 13 C/12 C 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 (δ13 C-EA/LC-IRMS)” Apidologie 3: 574-587.
•[9] Ulberth, F., Aries, E., De Rudder, O., Kaklamanos, G., Maquet, A. (2024). « Purity assessment of honey based on compound specific stable carbon isotope ratios obtained by LC-IRMS” Journal of AOAC International 107(5) : 884-887.
•[10] Joint Research Center.
•[7] White, J. W. and F. A. Robinson (1983). "13C/12C Ratios of Citrus Honeys and Nectars and Their Regulatory Implications " Journal of AOAC International 66(1): 1-4.
•[8] Elflein, L. and Raezke, K.-P. (2008). “Improved detection of honey adulteration by measuring differences between 13 C/12 C 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 (δ13 C-EA/LC-IRMS)” Apidologie 3: 574-587.
•[9] Ulberth, F., Aries, E., De Rudder, O., Kaklamanos, G., Maquet, A. (2024). « Purity assessment of honey based on compound specific stable carbon isotope ratios obtained by LC-IRMS” Journal of AOAC International 107(5) : 884-887.
•[10] Joint Research Center.
Author :
Héloïse Descotes-Genon (ITSAP)
Proofreaders :
Cécile Ferrus (ITSAP), Emilie Tourlet (ADAPI), Patrick Molle (InterApi), Alban Maisonnasse (Provence Miel)
s'inscrire à la newsletter
Vous y retrouverez les actualités de nos dernières recherches, événements, publications, infos clés à savoir en tant qu'apiculteur.