How Scientists Are Catching Fake Manuka Honey: Inside the Detection Tech

Background

Manuka honey commands a price premium other honeys don't, and wherever there's a premium, there's an incentive to cut corners. Reports of blended or mislabelled product have pushed researchers to develop chemical tests that can, in principle, tell genuine manuka honey apart from honey that's been diluted or passed off under a false label. But it's worth being precise about what "catching fake honey" actually means in a lab. Most of the methods described here measure chemical composition — concentrations of specific marker molecules — not the honey's floral origin directly, and none of them amount to a single, universally agreed industry standard. This article looks at what the current analytical chemistry research actually shows, and where the open questions still sit.

The chemical fingerprint approach

The starting point for most authenticity testing is the idea that genuine manuka honey has a distinctive chemical signature. Laboratory chemical analysis has confirmed that genuine New Zealand manuka honey typically contains markedly higher methylglyoxal levels — up to roughly 100-fold — than conventional honeys, measured via HPLC. This establishes methylglyoxal as a testable chemical marker for laboratory verification. It's an important finding, but it comes from a single study that characterized a limited sample set and did not test for intentional adulteration or fraud specifically — it tells us the marker exists and is measurable, not that every fraud scenario has been modelled against it.

More recent work has tried to combine several markers at once rather than relying on methylglyoxal alone. In one analytical chemistry study of 264 New Zealand and Australian honey samples, researchers showed that 1H NMR spectroscopy paired with chemometric (statistical pattern-recognition) modelling could quantify three chemical markers associated with manuka honey — methylglyoxal, dihydroxyacetone, and leptosperin — and use them to estimate manuka proportion in a blend. That's a meaningfully more sophisticated approach than testing a single compound, since it can, in principle, flag blends where manuka honey has been diluted with other floral honey rather than only detecting outright substitution. Still, this is a single laboratory dataset rather than an independently validated industry-wide standard, and it measures chemical composition, not floral pollen content directly — an important distinction we'll come back to.

Why the target itself moves

One complication for any threshold-based test is that genuine manuka honey isn't chemically uniform to begin with. A controlled study of mānuka plant genotypes found that dihydroxyacetone content in nectar — the precursor to honey's methylglyoxal marker — varies naturally by plant genetics, flower age, and growing conditions. That matters for authenticity testing because it suggests marker levels used to distinguish real from fake honey must account for natural biological variability, rather than treating a single cutoff number as absolute. It's worth noting this study examined nectar directly from plants, not finished honey products or fraud scenarios, so it doesn't tell us how much this variability actually affects real-world detection accuracy — only that the underlying biology isn't as fixed as a simple pass/fail test might assume.

Chemistry versus pollen

Chemical marker testing isn't the only tool available, and researchers working on NMR-based methods have been careful not to present it as a full replacement for older techniques. A honey-authentication study notes that chemical marker profiling by NMR is intended to complement, rather than replace, traditional pollen analysis (melissopalynology) for confirming botanical origin. This corpus doesn't include a dedicated melissopalynology study, so we can't independently assess how well pollen analysis performs as a standalone fraud-detection tool from these papers — the honest summary is that chemistry and pollen analysis are described as complementary, not that either one alone has been shown to be sufficient.

What grade labels can and can't tell you

It's also worth separating the question of authenticity from the question of potency labelling, since consumers often conflate the two. A laboratory study testing over-the-counter manuka honeys of UMF 5+, 10+, and 15+ found that lower-graded honey sometimes showed equal or greater antibacterial activity in vitro than higher-graded products from the same manufacturer, suggesting UMF labelling alone may not consistently predict measured potency. This was a limited single-brand comparison using in-vitro bacterial inhibition assays, not a market-wide fraud audit — it doesn't mean UMF grading is meaningless, but it does support seeking independently lab-verified results rather than relying on a printed grade number alone.

Where this is heading

Looking forward, some researchers have floated the idea that machine learning could speed up or improve chemical authentication work. A 2024 review article on manuka honey research notes that artificial intelligence could, in principle, be applied to optimize chemical analysis and detect quality or authenticity issues, alongside other applications. It's important to be clear about what this is: a narrative review commentary flagging a future research direction, not a report of a validated AI or Raman-based screening tool already in use. Claims about AI-based fraud detection should, for now, be treated as exploratory rather than established.