Technical Round Table on Honey Authentication · the honey database, to establish more transparent criteria for deciding whether a honey is adulterated and to improve the understanding

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

[email protected]

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