EUROPEAN COMMISSION DIRECTORATE-GENERAL JOINT RESEARCH CENTRE Directorate F - Health, Consumer & Reference Materials (Geel) Technical Round Table on Honey Authentication JRC-Geel, Belgium 25 January 2018 Meeting Report March 2018 ARES(2018)1677606
EUROPEAN COMMISSION DIRECTORATE-GENERAL JOINT RESEARCH CENTRE Directorate F - Health, Consumer & Reference Materials (Geel)
Technical Round Table on Honey Authentication
JRC-Geel, Belgium
25 January 2018
Meeting Report
March 2018
ARES(2018)1677606
2
Contents
Executive Summary ................................................................................................................................. 3
Introduction ............................................................................................................................................ 4
Setting the scene ..................................................................................................................................... 5
Outcome of World Café Discussions ....................................................................................................... 8
Conclusions and recommendations ...................................................................................................... 13
Acknowledgements ............................................................................................................................... 15
Annex 1. List of participants. ................................................................................................................. 16
Annex 2. Mind map of areas for improvement of control in the honey sector. .................................. 18
3
Executive Summary
This report presents the outcome of the technical round table on honey authentication, which was
held in Geel (BE) on 25th January 2018. The purpose of the meeting was to collect the opinion of a
broad representation of stakeholders of the honey supply chain on the current challenges to
authenticate honey, to identify the gaps in available tools and knowledge and identify ways of filling
those gaps. In a highly participatory manner the most common forms of honey fraud were discussed
and needs for addressing them in an effective manner identified. Among them were the lack of
analytical methods to detect the addition of certain types of sugar (syrup) and non-authorised
processing of honey such as (ultra)filtration, inappropriate used of bee feeding with sugar (syrup)
and resin treatment and effective infrastructures for the validation of analytical methods and the
provision of quality assurance tools, i.e. reference materials and proficiency testing rounds. Most
important appeared the need for modernised purity criteria of honey that have to go beyond the
basic quality requirements laid down in current EU legislation. Such criteria could take the form of an
electronic collection of chemical fingerprints against which a suspect sample can be compared for
assessing its authenticity and/or correctness of label declaration.
In particular the Round Table suggested that the following actions should be undertaken:
A critical review of the current definition of identity and purity criteria of honey is necessary;
Acceptance / rejection criteria for authenticating honey are needed;
An appropriate analysis of the vulnerability of the honey supply chain should be done and an
improved traceability system implemented;
Screening methods should be developed to economise testing;
Analytical methods to detect emerging fraud cases should be developed and already existing
methods have to be validated;
A mechanism for providing quality assurance tools should be established;
Chemical and biological characteristics of genuine honeys (including blends), bee feeding
products, and products from inappropriate practices should be generated and stored in a
publicly available database.
A coordinated effort involving all stakeholders of the honey supply chain and competent authorities
will be necessary to create an efficient mechanism for delivering the required tools and
infrastructures.
4
Introduction
Honey quality is regulated at international level by the FAO/WHO Codex Alimentarius (CODEX STAN
12-19811) and accordingly at European level by the Council Directive 2001/110/EC2 establishing
methods for their analysis.
European apiculture is a niche sector of agricultural production and is dominated by nonprofessional
beekeepers. Overall, EU honey production has been increasing slowly with annual variations
depending on climatic conditions. However, keeping this level of production is becoming harder for
beekeepers due to the challenges they face in terms of bees' health and environmental constraints.
Despite being the world's second largest honey producer, the EU is a net importer of honey as
domestic production only covers around 60% of consumption. The main supplier of honey imported
into the EU is China, followed by Ukraine and countries in Latin America.
The European Commission has regularly been informed of the presence on the market, in a
potentially significant proportion, of honey that may not meet the composition criteria laid down by
the Council Directive 2001/110/EC and/or that is not the result of the production process required
by the legal definition of honey. Moreover, despite the progress realised so far, scientific knowledge
concerning honey chemistry and technology lags behind the inventiveness of dishonest operators.
Therefore the development of efficient anti-fraud methods is necessary in order to avoid
disturbance of the market and the deterioration of the image of honey.
Thirty-three experts (Annex 1) working in the field of honey authentication were invited by the
Directorate-General Joint Research Centre (DG JRC) to discuss the currently most challenging issues
related to honey quality and authentication and to prioritise actions for solving them.
The main objectives of the technical round table were: i) to evaluate the capabilities and limitations
of currently used methods to monitor honey authenticity, and ii) to collect input for setting up a
roadmap to improve the currently used technology for authenticating honey. For reaching these
objectives it is appropriate to joint efforts at European and international levels, to involve all key
players, and evaluate the advantages / disadvantages of the official methods and of emerging
methods that, ideally, should be faster, cheaper, and more robust and be accepted worldwide.
1 http://www.fao.org/fao-who-codexalimentarius/sh-
proxy/ar/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCODEX%2BSTAN%2B12-1981%252Fcxs_012e.pdf; Revised in 2001. 2 http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:02001L0110-20140623&qid=1517998899798&from=FR;
Consolidated in 2014.
5
Setting the scene
Three presentations summarised the state-of-the-art for detecting adulteration of honeys with
exogenous sugars and for identifying the geographical and botanical origins of honeys.
A. Maquet (DG JRC) summarised the results of honey authenticity testing by liquid
chromatography-isotope ratio mass spectrometry (LC-IRMS).
In 2015 the European Commission organised an EU coordinated control plan (Commission
Recommendation C (2015) 1558) to assess the prevalence on the market of honey adulterated with
sugars and honeys mislabelled with regard to their botanical source or geographical origin3. All 28
Member States plus Norway and Switzerland participated in the plan. They collected over 2000
samples of honey at all stages of the supply chain. The coordinated control plan foresaw a three
tiered approach for the analysis of the collected honey samples:
All samples were analysed by the Member States for sensory characteristics and pollen profiles to check compliance with relevant provisions of the EU Honey Directive (2001/110/EC);
Compliant samples were then submitted to sugar analysis;
The samples which passed all these checks (or seemed suspicious) were then sent to the JRC for testing by LC-IRMS.
Member States submitted to the JRC, 893 samples of honey which they had found to be compliant
during tests in Tier 1 and Tier 2. By using LC-IRMS the JRC found that 14% of the samples tested were
suspicious of containing added sugar. This was further broken down according to the declared
geographical origin, point of collection (i.e. producer, packager or retailer) and type of honey.
Overall, the results indicated that the practice of adding sugars to honey occurs within the EU and in
third countries.
The applied technique (analytical method together with the decision criteria) has not been validated
in multi-laboratory studies conducted at the international level. It relies on empirically determined
benchmark purity criteria, taken from the published literature so that the selection of honeys used
to set the benchmark may influence the compliance decision. Further action is thus necessary to
establish the robustness of the results required for evidence in enforcement action.
A. Charlton (Fera Ltd, UK) presented the potential of Nuclear Magnetic Resonance (NMR) to detect
honey adulteration as a complimentary approach. NMR spectroscopy has been successfully used to
identify biomarkers of botanical and geographical origin for European honey. The accurate and rapid
measurement of methylglyoxal, a biomarker of Manuka honey, using quantitative NMR was also
reported. Thus the potential of NMR to detect adulteration related to botanical and geographical
origin is promising.
3 https://ec.europa.eu/food/safety/official_controls/food_fraud/honey_en
6
Concerning the detection of sugar adulteration, a thorough investigation highlighted that current
databases of honey NMR spectra may not be representative of the international honey market and
should take into account variation due to seasonality and permitted practises such as blending.
Interpretation of NMR spectra of immature honeys and blends seem to be problematic due to
ambiguity about permitted practises. It is therefore recommended to validate the methods of
analysis for honey (particularly NMR, LC-IRMS and DNA based pollen tests), to ensure the quality of
the honey database, to establish more transparent criteria for deciding whether a honey is
adulterated and to improve the understanding of honey production within and particularly outside
of the EU.
G. Kaklamanos (DG JRC) introduced the main aspects of a targeted and untargeted mass
spectrometry based metabolomics workflow to detect honey adulteration. Ultra-High-
Performance Liquid Chromatography coupled to High-Resolution Quadrupole-Orbitrap Mass
Spectrometry (UHPLC-Q-Orbitrap), can be applied to detect the oligosaccharide and polysaccharide
profile of honey samples (targeted metabolomics) and to indicate addition of sugar-based
adulterants or malpractices of bee feeding. Preliminary threshold levels were estimated for a few
oligosaccharides, which could be used as markers for detecting honey adulteration. An untargeted
metabolomics approach combined with advanced multivariate data analysis seemed to allow the
discrimination of honey samples of different geographical and / or botanical origin.
After the overview on the state-of-the-art and challenges in detecting honey adulteration, the
participants were invited to indicate what measures are efficient and what areas would need
improvement or action regarding the fight against honey adulteration (Annex 2).
The existing analytical methods for basic quality control of honey, the efforts to start harmonizing
them, some databases and the current legal framework was considered as appropriate. The
motivation of the entire honey sector to collaborate was also stressed.
The currently unresolved issues highlighted by the participants could be grouped in four main
categories:
Infrastructure needs
Participants mainly stressed that coordinated and robust sampling and testing campaigns
should be organized. Openly accessible and trustworthy databases and expert networks
were also mentioned as being important.
Types of fraud
Participants mainly stressed the need for a better definition of honey authenticity; more
specifically, questions related to acceptable bee feeding practices and industrial processing
of honey were raised.
Analytical tools
The main topics concerned the validation of existing and new methods through ring trials,
the improvement of the methods (accuracy, application domain, data fusion), the availability
of emerging methods (e.g. NMR, High Resolution Mass Spectrometry - HRMS), the
7
harmonization of analytical methods and the validation of the decision criteria for purity and
authenticity.
Legislation
Some participants suggested also providing more detailed provisions for honey authenticity
in the relevant EU legislation.
8
Outcome of World Café Discussions
A participatory approach was used to engage the participants in group work to answer the following
questions:
Types of fraud
o Where in the honey supply chain is the risk of adulteration highest?
o What types of fraud are most frequently observed?
Appropriate analytical tools
o What kind of analytical tools do we have to tackle honey adulterations?
o What are the limitations of these tools?
o What should be improved?
Infrastructure needs
o Which infrastructures are needed to tackle honey adulterations?
o Are sufficiently standardized / harmonised analytical methods available?
o Are the required quality assurance tools (e.g. proficiency tests, reference materials)
available?
o Are database available?
After harvesting the answers / suggestions participants prioritized the issues that should merit
actions from policy makers.
Concerning the types of fraud in the honey sector and where they most frequently occur in the
supply chain, participants felt that honey imported to the EU does not always comply with purity
benchmarks, but a lack of detailed statistics makes a comparison to other food supply chains
difficult. It was therefore generally agreed that there is a knowledge gap about where in the supply
chain fraudulent manipulations occur. However, the importance of identifying the most vulnerable
stage in the supply chain and to better focus honey control activities to those stages was stressed. As
a solution, blockchain technology coupled with sharing of analytical fingerprints was suggested as a
way to trace honey from the beehive to the consumers and to control its quality.
The identified fraud types were prioritized into five major classes. It has been noted that in reality
class boundaries may have varying geometries as honey can be adulterated in several ways.
1. Addition of sugar
Addition of sugar was identified as the most frequently occurring type of fraudulent
manipulation. Exogenous sugar can originate from inappropriate bee feeding and / or from a
direct addition of sugar / syrup to honey, and the difficulty in differentiating the origin was
pointed out. Honey from sugar fed bees was mostly considered a malpractice, whereas the
9
addition of sugar / syrups was considered fraud. A list of possible bee feeding materials and
sugars / syrups used for adulteration could be a useful starting point for analytical control.
2. Mislabelling
Mislabelling with respect to botanical- and geographical origin, mono-floral vs. poly-floral
honey, and blossom vs. honeydew honey was highlighted as the second most important
type of fraud. The practice of adding pollen or monofloral honey to ultrafiltered honey and
then labelling it as a monofloral honey was highlighted as a fraudulent practice that is
difficult to identify. Additionally, a lack of EU regulatory limits for the relative amount of
specific pollen types in monofloral honey impedes regulatory follow up. A regulated
minimum content of pollen from a specific plant species in monofloral honey was suggested.
Others suggested guidelines on the interpretation of the EU honey Directive; notably with
respect to pollen content.
3. Resin treatment / ultrafiltration (followed by blending)
Addition of pollen to ultrafiltered honey or the dilution of good quality honey with
ultrafiltered honey was discussed. Some participants reported that natural honey
constituents such as pollen and different enzymes are added back to filtered honey to match
the characteristic of genuine honey, and that these products are sold at different prices
depending on e.g. the enzyme activity of the adulterated honey. Synthetic resins are illegally
used to remove unwanted substances (antibiotics, pesticides, etc.) from honey; a potential
health issue with the use of resins may result.
4. Bee feeding
Bee feeding is widespread and accepted; however, feeding has to stop when nectar flow
starts. Some carry over is practically unavoidable and should be considered in the analytical
control. There was a discussion about the possibility to standardize bee feeding practices,
which is difficult due to climatic differences between EU countries.
To aid the analytical control of exogenous sugar in honey, it was suggested that bee feeding
material could be co-sampled when honey is collected in the national honey control
programmes.
5. Immature honey
It was generally agreed that immature honey is not properly defined in legislation, and a
guidance document is needed. It was argued that it might be difficult in some countries to
reach < 20% moisture before harvest as a result of a humid climate. The discrimination
between industrially dried immature honey and mature honey is an analytical challenge.
Concerning the appropriate analytical tools, their limitations and the needed improvements to
tackle honey adulteration, a wide variety of techniques already exists at different stages of
development / implementation in the concerned laboratories. Participants mentioned the following
techniques:
10
Conventional analysis: physicochemical analysis (e.g. pH, moisture, colour, electrical
conductivity and hydroxymethylfurfural - HMF), foreign enzymes analysis (amylase),
rheological analysis, melissopalynology and sensory analysis (colour, aroma and flavour).
Most of them are official and harmonised methods.
Isotopic measurement techniques: Elemental Analysis - Isotope Ratio Mass Spectrometry
(EA-IRMS) as an official method and LC-IRMS as a benchmark method.
Separation techniques (official and harmonised methods): sugar profiling by High
Performance Liquid Chromatography (HPLC) or Gas-Liquid Chromatography (GLC) (low cost,
readily available screening tool).
Spectrometric techniques: LC-HRMS for targeted and untargeted metabolomics, LC-MS/MS
for marker detection and GLC-MS for aroma profiling.
Spectroscopic techniques: Fourier Transform Infrared (FTIR) or Near-Infrared (NIR)
(screening tool) spectroscopy and NMR for targeted and untargeted metabolomics.
Trace elements: profiling by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS).
Molecular Biology: DNA barcoding and Next Generation Sequencing.
Statistical tools: chemometrics, data fusion of non-targeted methods, multiplexing data
from different techniques and decision tree approaches.
Other: biosensors.
Other topics mentioned during the discussions concerned sample preparation (should be simplified);
conflicting results when using different assays for the determination of diastase activity (Schade vs.
Phadebas assays); establishing ranges of electrical conductivity for botanical origin; limitations on
the dynamic range and signal overlapping in some methods; need for definition of analytical/quality
parameters for non-targeted methods (limit of detection, limit of quantification and specificity); and
specific concerns related to NMR spectroscopy (sensitivity at low concentration levels, lack of
resolution – signal overlapping and need for “expert interpretation"), influence of commercial
treatment (e.g. filtering) on certain characteristics of honey (e.g. NMR profile).
The experts highlighted the following limitations and consequently needs for improvements:
harmonisation, availability of databases, networking and quality assurance tools. As these areas of
activities were also clearly reported by the participants in the next and last topic "Infrastructure
needs", they will be summarised in that paragraph.
The poor cost effectiveness of the available set of analytical methods was criticized as several
methods are currently necessary for comprehensive purity checking of honey, which increases the
overall cost and requires investments that many official laboratories have difficulties to fund.
Moreover, loss of credibility of non-harmonised tests due to different data interpretation and
contradicting testing results provided by several laboratories for the same sample prevent such tests
from being widely accepted.
It was noted that guidance on using complementary analytical methods and contextual awareness in
data interpretation would provide confidence in testing outcomes. In this regard, it should be
recognized that analytical results are not the only criteria for decision making, but other factors such
as, price, traceability, etc., need to be taken into account for triggering legal action/enforcement.
11
The needed infrastructures to tackle honey adulteration were divided in four main areas: regulation,
databases, quality assurance tools and networking (Annex 3).
1. Regulation
International agreements are necessary for a better definition of honey, bee
feeding, bee feeding products, harvesting, Good Manufacturing Practice (GMP);
some could be integrated in the EU Honey Directive (2001/110/EC).
Improved traceability programmes are needed. The potential of Blockchain for
improving traceability and control in the honey sector should be investigated.
Improve and multiply borders control in particular at the main ports where non-EU
honeys enter the EU (e.g. Antwerpen / Rotterdam / Valencia, etc...).
Acceptance / rejection criteria for honey should be put in place as well as a foreign
supplier certification system to prevent fraud and control better the quality of
products.
Need for monitoring plans along the whole supply chain to improve the control
effectiveness (in particular at the beekeeping and honey reselling stage) including
Commission audits.
2. Databases
Databases storing information of compositional characteristics of honey need to be
representative, trustworthy, and accessible (not only to EU official control lab; open
access).
A clear definition is required on what exactly is 'authentic' honey in order to decide
which samples can be used to populate the databases.
Difficulty for getting authentic honey, especially from non-EU countries
(harmonisation of limits, e.g. moisture level when harvested) and to trace them
throughout the food chain (from producer to retailer).
Access to authentic samples is a critical point for populating databases: sufficient
samples (not only honey reflecting worldwide production but also bee feeding
products, sugar syrups, and products from inappropriate practices) and in sufficient
amounts should be collected directly from beekeepers or from suppliers by a person
qualified for sampling. This is an expensive process (several years of production from
the same sites should be foreseen) and will require putting in place local networks.
Metadata is important and should include information regarding
botanical/geographical origin, bee species, season / year of production, storage
practices, bee feeding practices, processing characteristics, blending practices,
composition, filtering, etc.
Already existing initiatives for sharing data: Food Industry Intelligence Network
(FIIN) in UK, a network of technical leaders to share knowledge on food authenticity
and traceability, or the European Reference Centre for Control in the Wine Sector
(ERC-CWS) and its EU wine database.
12
3. Quality assurance tools (harmonisation / standardisation, reference materials, proficiency
testing, etc.)
Reference materials are needed for sugar syrups, adulterated / non adulterated
honey (e.g. TUBITAK certified reference material for carbon isotope ratios in honey),
treated honey ((ultra)filtration, resin treatment).
The main issue with regard to honey reference materials will be the shelf life of the
reference material (only one year for certain parameters – the stability of some
sugars could also be an issue (i.e. trisaccharides / oligosaccharides).
Standardised methods are needed. Some methods are mature enough for already
organising proficiency tests, in particular for EA/LC-IRMS. This work should lead to
the harmonisation of acceptance limits for authentic honey accepted by all control
laboratories and possibly to the standardisation of the method.
Harmonised approaches by authorities within EU-28: definition of monofloral honey;
melissopalynology including interpretation of results; set-up limits / threshold levels
for bee feeding; statistics and compliance criteria (agreement on confidence levels);
guidance on recovery correction and application of measurement uncertainty.
4. Networking
Coordination by an independent body (EU Commission / other EU body) is needed
to better inform / train official control laboratories in terms of developing guidelines
for sampling and analytical methods (e.g. selection of methods to support
legislation); and identifying Reference Laboratories.
Network of competent laboratories equipped with state-of-the-art analytical
techniques. Several good initiatives already exist such as the German National
Reference Centre for Authenticity and Integrity of the Food Chain and the UK virtual
network hosted by LGC (composed of 14 participating laboratories).
Enhance and facilitate networking and communication (e.g. webpages, electronic
working groups, physical meetings, feed into Virtual Authenticity Network, involving
Food Authenticity Centres of Expertise).
Network needs also to be extended at the international level: exchanges of
experience and stakeholder collaboration.
Gain knowledge from already existing networks (e.g. Sure-Global-Fair - SGF in fruit
juices).
Need of a referee (independent authority) in case of dispute (related to methods
and results).
Funding is very important. National initiatives could contribute (e.g. UK may be able
to obtain some national funding) but there would be a need to join forces.
13
Conclusions and recommendations
All participants agreed that there is an urgent need to initiate actions to better control the honey
sector.
A critical review of the current definition of identity and purity criteria of honey is necessary. It
was acknowledged that adapting the EU Honey Directive would also require action at the FAO/WHO
Codex Alimentarius level to revise the respective Codex honey standard (STAN 12-1981).
Alternatively, complementary purity standards as well as corresponding analytical methods could be
created via Standard Developing Organisations such as the International Organization for
Standardization (ISO) or the International Honey Commission (IHC). Clearer product definitions are
needed in particular for monofloral honeys, bee feeding and industrial practices such as resin
treatment, drying of immature honey and (ultra)filtration in order to enable regulatory compliance
testing by control authorities.
Acceptance / rejection criteria for authenticating honey are needed. A survey among Member
States' competent authorities should be conducted to canvass national provisions and practices
beyond the EU Honey Directive to authenticate honey. The outcome of this task should be a
guidance document on good practices as well as currently used decision criteria concerning honey
authentication. Lack of such agreed criteria impedes regulatory follow up and in some Member
States investments into infrastructure to fight honey adulteration.
An appropriate analysis of the vulnerability of the honey supply chain should be done and an
improved traceability system implemented. Close monitoring of honey production statistics and
trade flow data combined with a vulnerability model of the honey supply chain is a pre-requisite for
informing auto-control actions by private stakeholders as well as targeted controls by the competent
authorities. Blockchain technology seems to be a promising way to improve product traceability and
transparency along the supply chain but unfortunately its application to fight food fraud is still in its
infancy.
Screening methods should be developed to economise testing. Generally, the laboratories are well
equipped for detecting conventional honey frauds and particularly for basic quality controls
(moisture, electrical conductivity, etc); however, there is a need for screening tools to cope with the
huge amount of samples to be tested in an economically feasible manner. Moreover, as several
analytical methods are needed for confirming the genuineness of a honey sample, a general screen
for singling out suspicious samples for further testing will improve the cost effectiveness of the
control system.
Analytical methods to detect emerging fraud cases should be developed and already existing
methods should be validated. Participants identified a need for methods to detect the addition of
industrially dried immature honey and/or (ultra)filtered honey to extend genuine honey; addition of
pollen and or enzymes to (ultra)filtered honey; intentional overfeeding bees with sugar (syrups).
Several modern spectroscopic methods which are already in use by service providers or research
institutions have to be validated and standardised so that they can be used for official control
purposes. The scope of a method shall be clearly described as well as the application of
complementary methods for confirming a suspicious result.
14
A mechanism for providing quality assurance tools should be established. Reference materials and
proficiency testing schemes are needed to provide evidence that methods are correctly applied by
laboratories and for establishing trust in testing results by all involved parties.
Chemical and biological characteristics of genuine honeys (including blends), bee feeding products,
and products from inappropriate practices should be generated and stored in a publicly available
database. This process would require obtaining samples by authorised personnel from carefully
selected honey producers. The (bio)chemical and biological composition of those samples and their
characteristic fingerprints obtained by modern analytical techniques have to be generated by
competent laboratories and the resulting data stored in an appropriate database. Private sector
mentioned that the authenticity of a sample will have to be defined beforehand. As databases
already exist, there was a discussion on the need to rebuild a new one; alternatively, the possibility
to effectively share the information under which conditions should be investigated. The database
should be (openly) accessible and for reasons of trustworthiness and neutrality it should be created
and curated by an independent institution. It was noted that the validation of the information
contained in the database will be a prerequisite for its use in official control activities.
The way forward. A concerted action including all stakeholders of the honey supply chain is needed
to fight adulteration and malpractices in the honey sector in order to protect the reputation of
European beekeepers and honey packers as well as basic consumer rights. As several tasks have to
be initiated in parallel; therefore, effective coordination by a body independent of national and
commercial interest will be needed. Participants suggested that at the technical level the DG JRC
should be entrusted with this task while for political decisions existing expert groups already
coordinated by DG AGRI shall take the lead. Collaboration among European stakeholders and
between them and international players is a key element in a future action plan. In Europe, DG
SANTE and DG AGRI, the competent authorities of the Member States and their national apiculture
programmes, industry, beekeeper associations and consumer association are players that shall
contribute to the implementation and execution of an action plan. In addition, reaching out to the
international level by involving Standard Developing Organisations such as ISO and IHC for setting
specifications and standardising test methods as well as to several honey exporting countries for
getting access to authentic honey samples will complement activities at the European scale and
enlarge the understanding on constraints and needs related to the control of honey authenticity.
The success of such an action plan will only be possible if appropriate resources are provided and the
work is carried out in a well-coordinated manner by a network of relevant stakeholders. Initiatives
started by competent authorities in the Member States or other stakeholder are welcome but need
to be embedded in a wider network in order to avoid non-harmonised standard setting that could in
the end create barriers to the free movement of goods in the Internal Market.
15
Acknowledgements
JRC would like to thank all participants for their contributions during the Round Table discussion.
In order to stay updated on the development of further actions on honey adulteration as well as on
the documents provided during the technical round table please consult the following link:
https://ec.europa.eu/jrc/en/science-update/how-ensure-genuine-honey-market
16
Annex 1. List of participants.
First Name Last Name Nationality Organisation
Filippo ABRUZZO ITA European Commission (DG SANTE)
Vadims BARTKEVICS LVA Institute of Food Safety- Animal Health and Environment "BIOR"
Patricia BEAUNE FRA Famille Michaud Apiculteurs / FEEDM
Gudrun BECKH DEU IHC (respectively QSI)
Klaus BECKMANN DEU Intertek Food Services GmbH
Viktorija BELSAK SVN Administration of the Republic of Slovenia for Food Safety- Veterinary Sector and Plant Protection
Cynthia BENITES FRA Copa-Cogeca
Etienne BRUNEAU BEL Copa-Cogeca + CARI
Ana CABANERO ESP Laboratorio Arbitral Agroalimentario. Ministry of Agriculture and Fishery- Food and Environmental Affairs
Adrian CHARLTON GBR Fera Science Ltd
Arne DUEBECKE DEU Quality Services International GmbH (QSI)
Selvarani ELAHI GBR Laboratory of the Government Chemist- LGC
Lutz ELFLEIN DEU Eurofins Food Integrity Control Services (also member of IHC)
Cathal HENIGAN IRL Valeo Foods (via Fera; UK Honey Assoc.)
Santiago HERRERO ESP Matrunita (via Mr Quaglia; FEEDM)
Tatjana KARAČIĆ HRV Ministry of Agriculture
Muriel LANDURÉ FRA SCL - Laboratoire de Marseille
Dario LASIĆ HRV Teaching Institute of Public Health Andrija Stampar
Sara LOPEZ-VARELA ESP AECOSAN-CICC
Pernille Lundquist MADSEN DNK Danish Veterinary and food Administration
Michelle MCQUILLAN GBR Defra (UK Ministry)
Anna OBEL POL Główny Inspektorat Jakosći Handlowej Artykułów Rolno-Spożywczych
Helena PASTELL FIN Finnish Food Safety Authority Evira
Giancarlo QUAGLIA ITA Floramo (also Head of FEEDM)
Gerhard RIMKUS DEU Intertek Food Services
17
Maryline SALLE FRA SCL - Laboratoire de Marseille
M.Teresa SANCHO ESP University of Burgos (also IHC)
Georg SCHREIBER DEU European Commission (DG SANTE)
Stephan SCHWARZINGER AUT Research Center for Bio-Macromolecules - University of Bayreuth
David SENCHERMÉS ESP ASEMIEL-ANIMPA
Dalibor TITERA CZE Vyzkumny ustav vcelarsky (Bee Research Institute)
Michael WALKER GBR Laboratory of the Government Chemist- LGC
John WARREN GBR Laboratory of the Government Chemist- LGC
Jorge SORRIBES ESP ASEMIEL-ANIMPA
18
Annex 2. Mind map of areas for improvement of control in the honey sector.
Potential of NMR to detect honey adulteration
Honey Round Table, JRC-Geel, Belgium, 25 January 2018
Fera Science Ltd.
• A leading supplier of scientific solutions, evidence
and advice across the Agri-Food supply chain in
Europe – “from farm to fork”.
• Joint venture between UK government and Capita
www.fera.co.uk
Honey authenticity
NMR to detect sugar addition?
• C3 plants predominate and include sugar beet
• C4 plants approx 3% of all vascular plants and
include sugar cane and corn
• Most honey is therefore made from the nectar of
C3 plants
• C4 sugar addition is easily detected using EA-
IRMS
• C3 sugar addition is difficult to detect and current
LC-IRMS method is limited mainly due to honey
being predominantly a C3 sugar mixture
• NMR has been proposed as a complimentary
approach to detect sugar adulteration
NMR spectroscopy
O
OHOH
OH
OH
OH
OH
OOH
N
N
NH2
O
OH OH
O O
NH2
OH
O
NH2
OH
O
NH2
O
NH
N
OH
NH2
OH
O
OH OH
OH
NH2
OH
O O
NH2
N
N
NH2
O
OHN
O
OHOH
OH
OH
OH
OH
OOH
N
N
NH2
O
OH OH
O O
NH2
OH
O
NH2
OH
O
NH2
O
NH
N
OH
NH2
OH
O
OH OH
OH
NH2
OH
O O
NH2
N
N
NH2
O
OHN
O
OHOH
OH
OH
OH
OH
OOH
N
N
NH2
O
OH OH
O O
NH2
OH
O
NH2
OH
O
NH2
O
NH
N
OH
NH2
OH
O
OH OH
OH
NH2
OH
O O
NH2
N
N
NH2
O
OHN
O
OHOH
OH
OH
OH
OH
OOH
N
N
NH2
O
OH OH
O O
NH2
OH
O
NH2
OH
O
NH2
O
NH
N
OH
NH2
OH
O
OH OH
OH
NH2
OH
O O
NH2
N
N
NH2
O
OHN
Excitation Pulse
(Radiowave)
9 8 7 6 5 4 3 2 1 ppm
NMR Spectrum
Free Induction
Decay
Fourier
Transform
NMR data
1D spectra: High sensitivity, accurate chemical shifts.
2D spectra: High information content and resolution.
ppm
7.07.58.08.5 ppm
6.8
7.0
7.2
7.4
7.6
7.8
8.0
8.2
8.41H – 1H TOCSY
Kynurenic acid in honey (2008)
Variable Chemical Shift Assignment
6328 9.129
7012 8.837
8461 8.218
8465 8.216
11298 7.006
11339 6.988
25221 1.056
Kynurenic acid7.07.58.0 ppmppm
7.07.58.08.5 ppm
105
110
115
120
125
130
135
1H – 13C HSQC
Profiling of Manuka honey (2008)
• Upper trace: Kanuka honey • Centre trace: EU honey • Lower trace: Manuka honey (5+)
Active ingredient precursor (Dihydroxyacetone)
Active ingredient (methylglyoxal)
qNMR of honey (2010)
Methylglyoxal
• Alkyl proton unresolved from water signal so methyl signal used for quantification
• MGO has a mono and a dihydrate in equilibrium so need to use both signals
5+
10+
15+
2,5-dihydroxyphenylacetic acid (2008)
Known biomarker of Strawberry Tree honey
High levels of Strawberry Tree biomarkers in Autumn Maquis
Honey is natural, definitions are man made!
Spectral fingerprinting
Computationally intensive
Data handling and
bioinformatics tools required Multivariate
Statistics
Observation
GM
Control
PC1
.004.02
-.01
0.00
.002.01
.01
.02
PC6PC2
0.0000.00-.002-.01
Artificial
intelligence
Univariate
Statistics
Spectral fingerprint
Classification rates (2016)
Leptosperin 4 markers combined
Ideal
space
Low
false
-ve
Low false +ve False +ve: non-Manuka classified as Manuka
False -ve: Manuka classified as non-Manuka
Best balance is probably: ca 15% false positive & ca 5% false negative
2-methoxyacetophenone
Lepteredine
POBA
We are in a good, if not an ideal space due to low false –ve rate
Calculated threshold 99.5 mg/kg
- Leptosperin
• For a threshold at 99.5 mg/kg Leptosperin
somewhere between 84 and 94% of New
Zealand Commercial “Other Honeys” will be
identified as non-Manuka and between 85 and
95% of Manuka honeys will produce a result
above this threshold.
Leptosperin Threshold?
1
10
100
1000
10000
Lep
tosp
eri
n (
mg
kg-1
)
Sample
Manuka
Non-Manuka
Warning
Action
Percentage of Samples
Leptosperin Threshold (mg kg-1) Manuka (n=220) Non-Manuka
(n=135)
>100 (Wholly or Mainly Manuka) 94.1% 2.2%
>50 and <100 (Manuka Blend) 5.9% 11.9%
< 50 (Non-Manuka) 0.0% 85.9%
Statistical models
• Generalised from real data using statistical
distribution models and measurement
uncertainties (errors)
• Ranges and therefore errors include variability
introduced into the data by factors such as:
collection location, vintage, analytical error
• Errors and ranges are best reduced and
threshold refined by analysing more samples of
known provenance
PCA plots shows the difference between years Leptosperin is very stable between the 2 years > 5% difference
Vintage (2016)
Detection of sugar adulteration (2017)
• Recent work at Fera has focussed on determining
the robustness of NMR spectroscopy for the
detection of sugar syrups in honey
• This was required following the use of the
technology by commercial laboratories which
resulted in claims of sugar adulteration being made
against some honey producers
• A thorough investigation into the marketing and
application of the technology has highlighted several
findings
Key findings
1. Current databases of honey NMR spectra may not be
representative of international market sources
2. Databases should take into account variation due to
seasonality and permitted practises such as blending
3. Potential for unexpected overlapping resonances at
lower field strengths and impact on quantification not
fully explored
4. Some key markers used to imply adulteration have
not been identified/ disclosed so cannot be validated
5. NMR analysis results for immature honeys and
blends seem to be most problematic due to confusion
about permitted practises
General recommendations
• Validate methods of analysis for honey, particularly
NMR, LC-IRMS and DNA based pollen tests
• Database QA is critical for implementing a successful
honey monitoring programme
• Criteria for stating that a honey is adulterated need to
be more transparent
• Improve understanding of honey production within
and particularly outside of the EU
• Unify approach internationally as similar work is
being undertaken particularly in US but also China
and NZ
Fera Colleagues EU
Rowse
UMFHA
Analytica
Comvita
University of York
Acknowledgements
The European Commission’s science and knowledge service
Joint Research Centre
Technical round table on honey authenticity
Targeted and untargeted mass spectrometry
George Kaklamanos
Technical round table on honey authenticity, Geel, BE – 25/01/2018
Composition of authentic honey
Fructose ~ 31 - 49%
Glucose ~ 23 - 41%
Sucrose ~ 0.2 - 10% & Water ~ 17%
Other ingredients up to 6%
Oligosaccharides 3-5%
Volatile compounds, phenolic acids, flavonoids, amino acids,
organic acids, proteins, vitamins, trace elements etc.
Adulteration of honey by sugar syrups
LC-HRMS metabolomics – Workflow
Planning Ahead / Experimental design
Conditioning
QC-format
Samples
≥7 QC injections
DDA
MSn
QC
Dilution
Multiple
extraction-blanks
Last QC 1st QC
Dilution
test
Test
mix QC
Run a test mix
• Check RT
• Accurate mass
• Frequent adducts
Unbiased tandem MS acquisition
• Run Data Dependent Acquisition (DDA)
• MSn
(Use pooled/QC samples)
Run Neg mode
Run sample
preparation blanks
Concept of sugar profiling
When honey is adulterated with syrups you obtain degradation
products from the enzymatic hydrolysis of starch
Oligosaccharides
Degree of Polymerization 3−11
Polysaccharides
Degree of Polymerization 12−19 or higher
Degradation products (AFGPRS, DFA IS and HFCS)
Sugar profiling
Analysis of different type of sugar syrups (glucose, maltose, rice, HFCS,
sugar cane, sugar beet, agave, maple, spelt, palm etc.).
Analysis of bee feeding products from bee feeding stores.
Analysis of a large number of honey samples.
Analysis of possible authentic honey samples from producers.
Quantification of main dominant sugars & oligo- with DP 4-10.
Establishment of threshold levels for DP4, DP6 and DP10.
Sugar profiling
C:\Xcalibur\...\Authentic samples\s400 06/28/16 13:19:24
RT: 0.00 - 16.01 SM: 5B
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Time (min)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
Rel
ativ
e Ab
unda
nce
7.34
7.24
8.04
6.21 8.89
7.998.54
9.17
6.05
0.63
9.42
9.63
9.83
9.994.656.93
10.150.54 5.76 10.28
5.623.553.480.75 3.84 5.26 10.522.821.901.0810.80 11.86
12.07 16.0012.60 13.74 14.07 15.65
NL:4.05E8
TIC F: FTMS - p ESI Full ms [224.00-3143.00] MS s400
Total Ion Chromatogram (TIC) of a standard solution of maltodextrins DP 4-19
Sugar profiling
Type syrup DP4 DP5 DP6 DP7 DP8 DP9 DP10 DP11 DP12 DP13 DP14 DP15 DP16 DP17 DP18 DP19
Feed (pollen) + + + + - - - - - - - - - - - -
Feed (Nektapoll) + + + + + - - - - - - - - - - -
Feed - SB + + + + + + + + + + + + + + + +
Sugarcane + + + + + - - - - - - - - - - -
Sugarcane + + + + + - - - - - - - - - - -
Sugarcane + + + + + - - - - - - - - - - -
HFCS + + + + + + + + + + + + - - - -
Agave + + + + + - - - - - - - - - - -
Agave + + + + + + + + + + + + + + + +
Agave + + + + + + + + + + + + + + + +
Agave + + + + + + + + + + + + + + + +
Maple + + + - - - - - - - - - - - - -
Maple + + + + - - - - - - - - - - - -
Wheat (Maltose) + + + + + + + + + + + + + + + +
Wheat (Glucose) + + + + + + + + + + + + + + + +
Palm + + + + + + + + + + + + + + + +
Spelt + + + + + + + + + + + + + + + +
Rice + + + + + + + + + + + + + + + +
Syrup+honey + + + + + + + + + + + + + + + +
Syrup+honey + + + + + + + + + + + + + + + +
Sugar profiling
Results obtained by UPLC-HRMS raised interesting issues:
Is it realistic to observe oligosaccharides of DP 4-8 in authentic honey
samples?
Could honeydew be a natural source of polysaccharides in honey?
Can the good practice of bee feeding affect the presence of oligo-&
polysaccharides in honey and at which extend?
Can thresholds levels (solid quantified values) wisely be selected?
Untargeted Pre-processing workflow
Unsupervised/
supervised methods
UPLC-MS raw data
Filter & identify
peaks Grouping, MinFrac
Ret. time correction
Re-grouping
Fill missing data
Normalisation Data pre-processing
Multivariate analysis
Converted in mzXML files
using MSconvert
Using XCMS package
CV filtering
Pre-processing workflow (XCMS)
Quality Control and Data Filtering
QC – robust LOESS Signal Correction
LOESS: Locally weighted scatter-plot smoother which is a strongly related non-parametric regression method that combines multiple regression models in a nearest-neighbor-based meta-model
Bank QC True samples (5~10) QC True samples (5~10) QC QC
Repeat
(n)
Experimental design
QC-RLSC algorithm: The
measured data of QC samples is smoothed by the
Loess method. The coefficient values between
QC samples are interpolated by the cubic-spline. The
entire datasets is aligned to the spline result.
Untargeted metabolomics
Potential to identify adulterated honey, geo- and/or the botanical origins.
Analysis of possible authentic honey samples.
Analysis of honey samples (217) with a detectable poly- profile.
Evaluation of classification models for threshold levels for DP6 & DP10.
Analysis of a broad range of representative honey samples (286) and
confirmation of the robustness of the threshold level.
Classification of geographical origin and floral composition.
Adulterated and non-adulterated
(threshold level DP10=55ppm)
217 honey samples with a detectable polysaccharide profile R2=0.9, Q2=0.758
Application of threshold level
(DP10=55ppm)
286 honey samples representative batch of samples R2=0.902, Q2=0.873
Metabolite Identification
p-value
VIP>2.0
-0.002<CoeffCS<0.002
Mass accuracy <5 ppm
Isotopic patterns
Variable Magnitude
Rel
iab
ilit
y (
mo
d.
Co
rrel
ati
on
)
Highlights markers: DP4-6 & DP8-10
Application of threshold level
(DP6=200ppm)
217 honey samples with a detectable polysaccharide profile R2=0.9, Q2=0.897
Adulterated and non-adulterated
(threshold level DP6=200ppm)
286 honey samples representative batch of samples R2=0.992, Q2=0.643 Similar evaluation for DP4 at threshold level > 700ppm
Floral composition of 9 varieties
From 286 honey samples including only monofloral varieties with (n>3), removing polyfloral and unknown – Total: 96 samples R2=0.9, Q2=0.535
Floral composition of 6 varieties
From 286 honey samples including only monofloral varieties with (n>3), removing polyfloral and unknown – Total: 80 samples R2=0.917, Q2=0.588
Floral composition of 3 varieties
From 286 honey samples including only monofloral varieties with (n>3), removing polyfloral and unknown – Total: 14 samples R2=0.987, Q2=0.73
Geographical origin
Authentic honey samples – Total: 76 samples R2=0.9, Q2=0.538
Geographical origin (clustered area)
Authentic honey samples – Total: 58 samples R2=0.9, Q2=0.553
Conclusions
Potential of applying an LC-HRMS sugar profiling approach for the
detection of honey adulteration.
Establishment of quantified threshold levels for honey adulteration.
Broader applicability to other analytical techniques (HPLC-PAD, MS
etc.).
Need of authentic honeys, materials from bee feeding
experiments, larger set of sugar syrups and bee feeding products.
Conclusions
Preliminary results showed the potential of applying an HRMS
untargeted metabolomics approach for honey authentication studies.
Need of including a larger sample-size, taking into account the
possible influence of geographical, botanical origin and honey
production/processing.
Identifying possible discriminating markers and set-up of targeted
profiling approaches.
Conclusions
Assess a complementary untargeted metabolomics approach by two
analytical strategies based on NMR and LC-HRMS.
Data fusion between NMR and HRMS to increase the discriminating
potential.
Thank you Any questions?
The European Commission’s science and knowledge service
Joint Research Centre
Results of honey authenticity testing by liquid chromatography-isotope ratio mass spectrometry
JRC-Directorate F:
Health, Consumers and Reference Materials
Eric Aries, Julien Burton, Luis Carrasco,
Olivier De Rudder, Alain Maquet
Technical round table on honey authentication, JRC-Geel, BE – 25/01/2018
Fraud cases
Coordinated control plan on authenticity of honey to detect fraudulent practices
Honey mislabelled with regard to its
geographical and/or botanical origin;
Products declared or presented as honey
although containing exogenous sugars or
sugar products.
Proposed coordinated control plan
JRC's contribution
Analyse the samples received from the Member States by the in-house
validated Liquid Chromatography – Isotope Ratio Mass Spectrometry
(LC-IRMS) method and interpret the results on the basis of published
authenticity criteria.
Sampling by the Member States
Source type Samples collected
Border inspection 35 1.5%
Distributor 157 6.9%
Importer 63 2.8%
Packaging companies 134 5.9%
Processor 81 3.6%
Producer 152 6.7%
Retailer 1010 44.6%
Storage companies 60 2.7%
Wholesaler 81 3.6%
Unknown 491 21.7%
Total 2264 100.0%
Sampling by the Member States
Samples Number
Collected by Member States (plus Norway and Switzerland) 2264
Sent to JRC and analysed by LC-IRMS
1069
-) of which without meta-data 38
-) non-compliant by applying the tests of Tiers 1 and 2 and EA-
IRMS in the Member States
138
-) compliant by applying the tests of Tiers 1 and 2 and EA-IRMS
in the Member States
893
Honey composition
Water ~ 18%
Fructose ~ 31 - 49%
Glucose ~ 23 - 41%
Disaccharides (i.e. sucrose) ~ 0.2 - 10%
Oligosaccharides 3-5% including:
trisaccharides (melezitose, raffinose, erlose, etc.)
traces of tetra-saccharides and penta-saccharides
Other ingredients up to 6% including anti-oxidants
(flavonoids), organic acids, minerals, proteins and amino-
acids
Source: Intertek, 2015
Benchmark purity criteria – EA/LC-IRMS Raezke (Intertek, 2015) & Elflein method (eFood Lab, 3/2015)
N= > 20,000 honeys (Intertek's database)
Measurement uncertainty in dubio pro reo
Parameter Std dev.
13Cprotein 0.16‰
13Cfructose 0.11‰
13Cglucose 0.12‰
13Cdisaccharides 0.18‰
13Ctrisaccharides 0.23‰
Percent peak area
oligosaccharides
0.66%
𝑈 = 2 ∗ 𝑆𝐷2 𝑎 + 𝑆𝐷2(𝑏)
Compliant polyfloral honey
Trisaccharides
(-26.3 ‰)
Disaccharides
(-25.1 ‰)
Glucose
(-25.2 ‰)
Fructose
(-24.9 ‰)
13C protein = -25.3‰
Δ 13C max = | ts – f | = 1.4‰ (compliant)
HS-0315
Suspicious honey
13C protein = -29.3‰
Δ 13C max = | f – p | = 3.9‰
Oligosaccharides peak area = 12%
(suspicious)
Preliminary conclusions on adulteration experiments using the LC-IRMS
Source adulterant Potential detection limit Pre-condition
C4 sugar > 1%
C3 sugar > 10% Oligosaccharides present
Issues with trisaccharide: peaks with low intensities
Results
Origin Samples
(n)
Suspicion of non-compliance
(n) (%)
Blend of EU honeys 96 19 19.8
Blend of EU and non-EU honeys 426 40 9.4
Blend of non-EU honeys 30 3 10.0
Single EU Member State 275 53 19.3
Single non-EU country 55 11 20.0
Unknown 11 1 9.1
TOTAL 893 127 14.2
Results
Category Samples
(n)
Suspicion of non-compliance
(n) (%)
Border 4 0 0
Distributor 106 8 7.6
Importer 21 2 9.5
Packager 29 4 13.8
Processor 36 3 8.3
Producer 51 5 9.8
Retailer 563 92 16.3
Storage 22 3 13.6
Wholesaler 56 10 17.9
Unknown 5 0 0
TOTAL 893 127 14.2
Recommendations
Harmonization of analytical methods
Biobank of honeys, sugar syrups and bee feeding products
European honey reference database
Validation of emerging analytical methods
Any questions?
Stay in touch
EU Science Hub:
ec.europa.eu/jrc
Twitter:
@EU_ScienceHub
YouTube:
EU Science Hub
Facebook:
EU Science Hub – Joint Research Centre
LinkedIn:
Joint Research Centre