The sticky problem of detecting fake honey

A A Milne’s first Winnie-the-Pooh story book was published on 14 October 1926. For the past 100 years, generations of children have been enchanted by the tales of this bear and his friends. Pooh’s love of honey is central to these books and sees him embark on adventures such as floating up to a beehive with a balloon to get more of his favourite sweet treat.

But the story of honey as a food dates back many thousands of years before A A Milne created the Hundred Acre Woods. Archaeologists have found evidence for the widespread exploitation of honeybees by early Neolithic farmers nearly 9000 years ago. Rock paintings in Spanish caves depicting people gathering honey from wild bees date back to a similar time.

Honey has remained a staple in the human diet ever since. But the distance that most our honey travels today from the hive to our kitchen tables has extended far beyond what those first farmers, and A A Milne, could possibly have imagined. This is especially true in the UK where only about 10% of the honey consumed is made by UK-based bees. The rest is imported, usually in bulk containers. This then undergoes further processing and blending before being packaged and sold as branded and supermarket own-brand honey. China is the source of more than three quarters of the honey imported into the UK.

But there have been growing concerns in recent years that the honey we buy may not be what we expect. ‘The thing that focused people’s attention on the authenticity of food was the horse meat scandal in 2013 and [since then] there’s been a much closer look taken at the integrity and the vulnerability of the food supply chain,’ says Michael Walker, a researcher in honey authenticity at the Institute for Global Food Security at Queen’s University of Belfast in Northern Ireland. Fraudsters target high-value foods and honey – depending on its botanical and geographical origins – can command from around £1 to over £15 a jar.

How widespread is honey fraud?

It is suspected that fraudsters are using low-cost sugar syrups to either bulk out or entirely replace honey in honey jars. A number of recent high-profile analytical studies seem to support such concerns. The European Commission’s 2023 From the Hives report, for example, stated that 46% of the over 300 samples of imported honey tested by its scientists at the Joint Research Centre (JRC) in Belgium showed apparent signs of adulteration.

Significant media attention followed these, and other recent similar, scientific findings. However, a closer look behind the headlines reveals a great deal of scientific uncertainty about the accuracy of honey authenticity tests. There is also significant polarisation within the honey industry, with each stakeholder briefing the media on their side of the story, which is adding to the confusion.

‘It suits people to have strong opinions one way or the other,’ explains Adrian Charlton, principal scientist at Fera Science, a company part-owned by the UK Department for Food Environment and Rural Affairs (Defra), based in York. ‘The [UK and EU] beekeepers say that there’s a lot of fraud, because that helps them to drive up the price of what they can assure is genuine honey, and then on the other side, the honey packers say that there’s no fraud at all, because they want to sell their [imported] products to the supermarket.’

The reality is that current adulteration rates are unknown. ‘There isn’t any clear evidence of fraud at the moment, but that’s partially down to the methods not being fit for purpose for detecting the fraud,’ Charlton says, adding that he believes some adulteration is likely, but at a much lower rate than the media hype suggests.

Why is honey so hard to analyse?

The analytical problems with honey authentication had been known for a while, but the 2023 From the Hives report served as a wake-up call for the industry. Since then, there has been a flurry of activity and international collaborations aimed at better arming those working to combat honey fraud.

‘Honey is a unique matrix analytically,’ says David Hoyland, a technical adviser to the British Honey Importers and Packers Association. Due to its sticky nature, high acidity and a huge amount of natural variation, honey is extremely difficult to analyse. It typically contains around 60–85% fructose and glucose, and around 15–20% water. A cocktail of minor ingredients makes up the remainder. ‘It’s got hundreds, if not thousands, of different trace compounds in there,’ says Hoyland.

The additional components of honey include trace sugars such as sucrose and maltose, enzymes and other proteins, vitamins and minerals, and flavour and aromatic organic compounds. ‘The complex composition of honey varies widely depending on what plants the bees forage on’ and the geographical location of those plants, explains Walker. ‘How the beekeeper manages the hive and how the honey which appears on the shelf is processed’ also has a large impact.

The hive management technique that causes the most analytical confusion is bee feeding. Colonies make honey as a backup food supply for when nectar isn’t readily available. When that honey has been removed from the hives, beekeepers often need to feed their bees sugar syrup during the winter months to keep them alive. But if this is done in excessive amounts, this sugar can wind up in the honey itself. Analytically it is hard to differentiate any feeding remains from intentional adulteration, says Martin Linkogel, a certified food chemist and the business manager of bee products at the contract analytical testing laboratory Quality Services International in Bremen, Germany.

The nomadic beekeeping style also confuses honey authentication. In nomadic beekeeping, hives are moved long distances regularly to extend the amount of time bees can access flowers in bloom. ‘The majority of the Chinese [honey] industry works in this way,’ says Hoyland. To allow for extra harvests per hive, nomadic beekeepers in China collect the honey from hives before it has had time to set. This runny honey is then industrially moisture-reduced under vacuum and heat, a process that can change the composition of some of its trace compounds.

How do scientists test if honey is real or fake?

Isotope ratio mass spectrometry (IRMS) to determine honey’s carbon-13 to carbon-12 ratio is the only internationally standardised method for detecting sugar syrups in this food. The amount of carbon-12 and -13 taken up during photosynthesis varies between plants. Notably, sugar cane and corn differ in their carbon-13 uptake from most other flowering plants. If a honey’s isotope ratio is outside the expected range, it indicates it has been spiked with syrup from one of those sources. This method is much less useful that it once was, as fraudsters have largely switched to using syrups from sugar beet and rice, which have the same carbon-13 to carbon-12 ratios are honey. In addition, excessive bee feeding with sugar cane syrup can cause honey to get a false positive for adulteration using IRMS.

Over the past decade or so, food scientists have developed several alternative methods for detecting sugar syrups in honey. These methods fall into two broad categories: non-targeted and targeted. See box Modern analytical tools for detecting fake honey  below for examples of the analytical tools being used for these methods.

Food scientists use the non-targeted methods to collect chemical fingerprints of honey that are then compared to authentic chemical fingerprints in databases. Most experts agree this approach has potential, but the limited scope of many of the current databases restricts its effectiveness. Honey is a natural product with huge variations in composition. If a database doesn’t hold details of the specific genuine honey the lab is analysing, the test can return a false positive for adulteration.

About six months after the From the Hives report, a UK protocol for the construction of honey authenticity databases was published. The protocol outlines the paper chain process needed to ensure the provenance of honey in databases. It also states that databases must contain reference samples with the same botanical and geographical origins as the honey being tested. ‘The scope of the database [must also be] clearly defined [with information on] what samples are in there and how well the samples were collected,’ says Charlton, who contributed to this protocol along with Walker and Hoyland.

‘One of the major issues with some of the commercial databases is they’re not being very transparent about what’s in there,’ Charlton explains. This means that to ‘the person commissioning the analysis, these untargeted systems are largely black boxes’ that give yes/no answers without explanation, agrees Thomas Linsinger, a scientific official at the JRC. Fera Science is currently building an exemplary database with honey on the UK market.

Building better databases

Databases are also becoming increasingly sophisticated. Maria Anastasiadi, a senior lecturer in bioinformatics at Cranfield University, UK, for example, is adding the Raman spectroscopy chemical fingerprints of authentic honey adulterated with various sugar syrups in different concentrations to her database. Anastasiadi also uses machine learning algorithms to improve the data analysis. ‘The model is tasked with identifying which features in the Raman spectrum differ between the different types of [honey], so then when we present it with unknown sample, it can give us an indication of whether it’s suspected of adulteration, or it looks like pure honey,’ she explains.

In targeted methods, food scientists look for markers of adulteration in honey. Currently, however, there is no known marker that alone definitively confirms adulteration. Mannose, for example, is a popular honey adulteration marker as it is found in many sugar syrups. But mannose is also naturally present in honeydew, linden and chestnut honey, as well as honey with high yeast content, meaning false positives for adulteration are common. ‘There’s never been a systematic study to determine what the range of natural components are in honey,’ says Charlton.

In addition, purified sugar syrups with mannose and other markers removed (to make them less detectable) are now readily available online. ‘The fraud has become more and more professional,’ says Linsinger. ‘You can buy the syrups on the internet [with] a certificate that [it] will pass a certain test.’

A lack of universally accepted thresholds for the markers further undermines the credibility of targeted approaches as it makes it difficult to interpret data meaningfully. ‘Lots of these tests have got no limits [and] no uncertainty range,’ Hoyland says. ‘It’s quite an unfortunate situation where the rules are being set by the laboratories that are doing the analysis, so one lab can tell you that something’s fake because of the way that they analyse it, and another lab can tell you something different,’ agrees Charlton. ‘There is a desperate need for harmonisation … and that work is going on,’ he adds.

What counts as honey?

With no single method able to consistently determine if honey has been adulterated with sugar syrup, food scientists have pinned their hopes on a combined approach. This means that increasingly a number of analytical results are being considered together to determine if a honey is authentic or not. ‘If you’ve got weak evidence [for authenticity] from NMR, weak evidence from high resolution mass spectrometer and weak evidence from DNA … it’s very likely that a whole pile of weak evidence adds up to quite strong evidence [for fraud],’ says Walker.

The combined approach doesn’t necessarily need multiple analytical tools. Instead, various independent markers can be looked for using the same instrument. The JRC, for example, is leading an international effort to validate the use of liquid-chromatography high-resolution mass spectrometry to quantify several specific markers in one go.

Once the consensus of analytical evidence points towards fraud, paperwork checks of product authenticity – from beehive to jar – should be carried out, says Walker. All findings should then be combined before it is decided if a honey is likely adulterated. This weight-of-evidence approach is well established in forensic science and growing in popularity for food authentication, Walker explains.

But while food scientists may have started pulling together to develop the analytical tools needed to combat modern-day honey fraud, more accurate analysis alone won’t stop the arguments between UK and EU beekeepers and the honey importers and packers. Their disagreements are multiple and complex but at the core there is a fundamental disagreement as to what can and can’t be called ‘honey’. The term is defined by UK and EU law, but its definition is open for interpretation as to whether moisture-reduced honey should be included. Attempts to gain an international consensus on this matter are ongoing, albeit very slowly. Charlton, however, is quick to note that he doesn’t think consumers would want to see their cheaper honey renamed. ‘Not everybody’s a connoisseur. My children wouldn’t go anywhere near a £15 jar of honey. They’re much more interested in something they can squeeze out of the bottle and onto their toast.’

Nina Notman is a science writer based in Salisbury, UK