Grayanotoxins in Mad Honey: Therapeutic Potential, Regional Comparison, and the 2025 Research Frontier

This 2025 review from Tribhuvan University in Kathmandu, Nepal, is the most recent peer-reviewed synthesis of grayanotoxin research in the CMHI library. It is also the first paper in the library authored by someone from Nepal. That provenance matters. 

The review is not a narrow technical paper; it draws together the chemical composition of mad honey, the toxicological profile from animal studies, a regional comparison of concentration data between Nepal and Turkey, documented human case records from both countries, the molecular mechanism of grayanotoxin action, evidence for therapeutic applications, and a policy framework for what the field needs next.

What distinguishes this paper from the earlier three CMHI entries is scope and orientation. Jansen (2012) established the mechanism and clinical case record. Ahn (2022) documented batch-level concentration variability. Ullah (2018) built the diagnosis and treatment protocol. This paper situates all of that within the current state of research in 2025 and provides an honest accounting of what still needs to happen before therapeutic potential can move from animal evidence to clinical application.

What the Paper Covers

The paper is organized around six areas. It begins with the sources and chemical composition of mad honey, documenting how different plant species, altitudes, climates, and harvest seasons drive variation in grayanotoxin content. It then reviews the toxicological effects of GRAYs across animal studies, a body of literature that now includes hepatotoxicity, genotoxicity, ovarian toxicity, and cardiovascular effects in addition to the better-known bradycardia and hypotension. 

The third section presents human intoxication cases from Nepal and Turkey side by side. The fourth section covers the molecular mechanism of GRAY-induced voltage-gated sodium channel dysfunction, building on and extending the mechanistic record established in earlier literature. 

The fifth section addresses what has become the most clinically and commercially significant question in the field: whether grayanotoxins have genuine therapeutic applications, and what the current evidence base actually supports. The sixth section lays out the policy and research agenda, which standardization, regulatory oversight, and study design the field requires before therapeutic applications can be responsibly pursued.

Nepal and Turkey: What the Regional Concentration Data Shows

One of the most practically significant contributions of this paper is Table 1, the first peer-reviewed compilation in a single review of GRAY concentration data from both Nepal (Himalayan region) and Turkey (Black Sea region). Until now, that comparison has required cross-referencing multiple individual studies. This paper assembles the data in one place.

The table below draws from the concentration records compiled in the 2025 review. All values are in µg/g (micrograms per gram of honey) unless noted.

What the comparison does and does not show

Several things emerge from reading this data together. First, both regions produce honey with substantial concentration variability, not just between regions but within them. The Nepal data from Ahn et al. (2022) alone spans a factor of 86 for GRAY-I and 255 for GRAY-III across 60 samples. Turkish data shows comparable internal variation. No single figure characterizes either region’s honey.

Second, the regions do not have clearly ranked profiles. Turkish honey, in some studies, shows higher GRAY-III dominance. Aygun et al. (2018) found GRAY-III at 27.60 µg/g versus GRAY-I at 0.39 µg/g, while Nepal honey from Ahn et al. (2022) shows both isoforms reaching similar high-end values. Turkish honey also consistently shows GRAY-II presence, which is less documented in Nepal samples.

Third, and most importantly, the comparison is not methodologically controlled. The studies in Table 1 used different analytical methods, different sample sources, and were conducted at different times. Aggregating them gives a useful empirical picture but not a head-to-head comparison. 

The paper is explicit about this. Environmental factors, altitude, climate, soil composition, harvest season, and specific plant species, are as likely to drive concentration differences as geography alone.

The Sodium Channel Mechanism: What the 2025 Review Adds

The foundational mechanistic account of how grayanotoxins work on voltage-gated sodium channels is covered in depth in the Jansen (2012) CMHI entry. The 2025 paper builds on that record with a more detailed structural description of the channels themselves and some additional mechanistic pathways that were not part of the 2012 treatment.

Voltage-gated sodium channel structure

GRAYs disrupt voltage-gated sodium channels (VGSCs), complex heteromultimeric structures consisting of a pore-forming α subunit and auxiliary β subunits. The α subunit is a polypeptide of approximately 2,000 amino acids organized into four homologous transmembrane domains (DI–DIV), each containing six helical segments (S1–S6). 

The S1–S4 segments form the voltage-sensing domain, characterized by positively charged lysine and arginine residues along the S4 helix that enable voltage detection and channel activation. The adjacent S5–S6 segments form the re-entrant P-loop, which constitutes the ion-selective pore through which sodium ions pass.

Site-directed mutagenesis studies on Nav1.4 sodium channels have identified two residues critical to grayanotoxin binding: phenylalanine (Phe-1579), which enhances GRAY accessibility to the receptor site, and tyrosine (Tyr-1586), which governs binding affinity. GRAYs specifically bind to the IS6 and IVS6 segments, segment 6 of domains DI and DIV, and induce a conformational change that locks the channel in its open state, preventing the normal inactivation process.

Downstream effects beyond vagal stimulation

The consequence of persistent channel activation is sustained sodium influx into excitable cells. Neurons experience increased energy demands as a result of continuous depolarisation, and if the resting potential is not restored, cellular damage can follow. 

The accumulation of intracellular sodium also triggers secondary calcium influx through voltage-dependent calcium channels, a pathway that impacts neurotransmitter release and muscle activity, and that underlies some of the cardiac and neurological symptoms seen in poisoning cases.

The paper also documents a CNS pattern not fully captured in earlier reviews. Grays initially stimulate the central nervous system, producing tremors and convulsions. Prolonged exposure shifts this to CNS depression, drowsiness, impaired motor function, and, in severe cases, seizures. 

In the heart, sustained VGSC activation disrupts conduction, creating the conditions for arrhythmias, including atrial fibrillation and ventricular tachycardia. Gray’s also induces vasodilation through mechanisms not fully characterized, which contributes to the hypotension documented in clinical cases.

For the full mechanistic account, including M2 muscarinic receptor data, atropine reversal of bradycardia, and the Nav1.4 S6 domain findings, see the Jansen et al. (2012) CMHI entry.

Animal Study Evidence for Therapeutic Applications

The therapeutic potential of mad honey and grayanotoxins is one of the most discussed and least clearly evidenced topics in the field. This paper addresses it directly, synthesizing the available animal study data across several application areas. The discipline required in this section is the same as in the clinical signs sections: document what the evidence shows, attribute it to the right study, and state clearly what level of evidence that represents.

The table below organizes the documented therapeutic findings from the 2025 review and the studies it cites.

The fracture healing finding in context.

The most striking therapeutic finding in this paper is that Sahin et al. (2018) demonstrated that mad honey enhanced bone fracture healing in female Sprague-Dawley rats at 80 mg/kg, outperforming both normal honey and propolis at both 15-day and 30-day assessments. Propolis is widely regarded in ethnopharmacology for its wound healing properties, which makes this a meaningful comparison rather than a trivial one.

What it is not is clinical evidence. The study used a rat fracture model, an oral dose of 80 mg/kg, and a 30-day observation window. Translating that to human therapeutic use would require dose-escalation trials, pharmacokinetic data in humans, and a safety assessment that accounts for the genotoxicity findings documented elsewhere in the same literature. None of those studies exists yet.

Blood glucose and blood pressure: the folk medicine connection

The blood glucose reduction documented by Oztasan et al. (2005) and the blood pressure reduction documented by Turkmen et al. (2013) are the two findings that most directly connect to the historical folk use record. 

Mad honey has been used for centuries in Turkey and Nepal for hypertension and diabetes management. These animal studies provide a pharmacological mechanism for those uses: parasympathetic nervous system stimulation leading to increased insulin secretion, and direct vasoactive effects from VGSC modulation. They do not, by themselves, establish therapeutic efficacy in humans. The paper is explicit that clinical trials are the next required step.

Genotoxicity: Evidence the CMHI Library Has Not Addressed Until Now

None of the previous three CMHI articles, Ahn (2022), Jansen (2012), or Ullah (2018), documents genotoxic effects of grayanotoxins or mad honey. The 2025 review is the first paper in the CMHI library to synthesize this body of evidence, and it warrants its own section rather than a footnote to the animal toxicity data.

Genotoxicity refers to the capacity of a substance to damage genetic material, DNA, chromosomes, or the replication machinery that copies them. The concern is not theoretical. Multiple independent animal studies, using different methods and species, have documented genotoxic signals from mad honey or GRAY compounds at various doses.

What the animal studies document

• Eraslan et al. (2018) administered mad honey to Wistar male rats at three dose regimens: acute (12.5 mg/kg single dose), subacute (7.5 mg/kg daily for 21 days), and chronic (5 mg/kg daily for 60 days). The Comet assay, which measures DNA strand breaks in individual cells, showed dose-dependent increases in tail intensity across all three exposure regimens, indicating oxidative DNA damage at each level of exposure.
• Rasgele et al. (2021) found increased chromosomal aberrations and micronucleated erythrocytes in mice exposed to Rhododendron honey containing GRAYs, indicating bone marrow toxicity. The micronucleus test is a standard genotoxicity biomarker; elevated micronucleated erythrocytes reflect chromosomal damage during cell division.
• Cakmak-Arslan et al. (2020) documented lipid peroxidation and altered membrane fluidity in mice exposed to GRAY-III and toxic honey, consistent with systemic oxidative stress, a mechanism through which indirect DNA damage can occur even without direct compound-DNA interaction.
• GRAY-III at doses as low as 0.01 mg/kg-bw did not induce chromosomal breaks but did increase micronucleated polychromatic erythrocytes (MNPCE) and reduce the PCE/NCE ratio in bone marrow, indicating hematological toxicity even at sub-threshold doses for other effects.

The ovarian toxicity finding

A specific genotoxicity-adjacent concern documented in this paper involves female reproductive tissue. Yeşil et al. (2024) administered mad honey to female Sprague-Dawley rats at 80 mg/kg oral gavage over 30 days. The treated animals showed moderate to strong caspase-3 immunostaining, a marker of apoptosis, along with increased expression of inducible nitric oxide synthase (iNOS) in Graafian follicles. These findings indicate that mad honey may promote ovarian follicular atresia: accelerated cell death in developing ovarian follicles.

Follicular atresia is a normal physiological process, but its acceleration can affect fertility. The clinical relevance of this finding in humans at typical mad honey intake amounts is not established. No human study has examined ovarian function in relation to mad honey consumption. The paper documents the finding; it does not extrapolate it to human clinical risk.

The scope of what is and is not known

Blood Concentration Thresholds: A New Clinical Reference Point

One of the most practically significant pieces of data in the 2025 paper is the blood concentration threshold information compiled from Choi et al. (2017). Earlier CMHI articles, particularly Ullah (2018), document intake-weight thresholds for poisoning: approximately 15–30 grams of mad honey are associated with intoxication onset in documented cases. What Choi et al. (2017) add is a pharmacokinetic reference: what GRAY blood concentrations actually correspond to clinical outcomes.

Choi et al. (2017) analyzed GRAY-I and GRAY-III blood concentrations in six patients who consumed Rhododendron liqueur. Measured levels were 1.44 ng/mL for GRAY-I and 16.9 ng/mL for GRAY-III. The estimated blood concentrations associated with clinically significant hypotension are 2.52–4.55 ng/mL for GRAY-I and 17.5–27.3 ng/mL for GRAY-III.

Why this matters beyond the case report record

The intake-weight data from case reports, 15–30 grams, reflect what patients reported consuming before symptoms appeared. That figure is useful but limited, because the same gram weight of honey can contain radically different GRAY concentrations depending on the batch. A jar with 0.75 µg/g of GRAY-I and a jar with 64.86 µg/g, both documented in the Ahn (2022) Nepal sample set, represent an 86-fold difference in actual GRAY content for the same physical amount of honey.

The blood concentration data bridges that gap. It establishes the pharmacological threshold, not the intake weight that corresponds to the clinical effect. That is a more precise reference point for understanding dose-response relationships, even if it comes from a small dataset.

The paper also notes that levels below 10 mg/kg in honey may still pose a significant toxicity risk for vulnerable populations, specifically toddlers and the elderly, whose pharmacokinetic profiles differ from those of healthy middle-aged adults (Schrenk et al., 2023). This is a relevant caution for public health guidance: the primary poisoning literature is dominated by cases in middle-aged men, which may not represent the full risk profile.

Human Case Records: Nepal and Turkey in the Same Frame

The paper’s Table 3 is the most comprehensive side-by-side compilation of mad honey case reports from Nepal and Turkey assembled in a single peer-reviewed review. It covers cases spanning decades, documenting intake amounts (where available), presenting vital signs, cardiac rhythm abnormalities, treatment protocols, and recovery times.

What the Nepal cases document

Several Nepal cases in the table are noteworthy for the specificity of their clinical details. Aryal et al. (2017), a paper also authored in Nepal, documented a patient who consumed 30–40 mL of wild honey. Blood pressure was not recordable on admission. The patient was treated with saline infusion, atropine, and adrenaline, and recovered within 48 hours. Thapa et al. (2024) reported a series of five cases from 2022–23, with intake ranging from 10–20 mL, treated with saline, atropine, and in some cases noradrenaline infusion or adrenaline. Recovery time was 24 hours for all five.

Across the Nepal cases in the review, intake amounts are frequently recorded in teaspoons or milliliters rather than grams, reflecting the way mad honey is typically consumed in Nepal: as a folk medicine taken in small, deliberate doses rather than as a food. This is a meaningful distinction. The cases in which patients consumed 200 mL (Adhikari & Bhandari, 2021) represent a different exposure scenario than someone taking two teaspoons.

Turkey and Nepal: the same clinical picture

Despite the different regional context and, in many cases, different Rhododendron species involved, the clinical presentation across Turkish and Nepalese cases follows the same pattern: bradycardia, hypotension, dizziness, nausea, and in severe cases cardiac rhythm abnormalities including AV block and atrial fibrillation. Treatment response is also consistent; saline and atropine resolve most cases within 24 to 72 hours. Temporary pacemaker use is documented in Turkish cases involving complete AV block, but the clinical course is not significantly worse or better by geography.

This consistency reinforces what the mechanism literature establishes: the toxin is the same, the channel target is the same, and the clinical presentation follows predictably from the pharmacology, regardless of where the honey was produced.

For the full clinical diagnosis protocol, differential diagnosis against organophosphate poisoning, and treatment hierarchy from IV saline to temporary pacemaker, see the Ullah et al. (2018) CMHI entry.

Policy and Future Research: What the 2025 Review Calls For

The final section of the paper does something none of the three prior CMHI entries do: it addresses what still needs to happen. This matters for a research index because it marks the current boundary of the literature, what is established, and what remains to be done.

The research agenda

The paper explicitly calls for three research priorities. First: pharmacokinetic studies to clarify how GRAYs are metabolized and excreted in humans. The rapid symptom resolution, typically within 24 hours, suggests efficient clearance, but the metabolic pathway is not characterized. Understanding it is the precondition for developing standardized dosing guidance. Second: diagnostic biomarkers for rapid clinical identification. No blood or urine test for GRAYs is commercially available. Diagnosis depends entirely on clinical history. A validated biomarker would significantly improve both the speed and accuracy of diagnosis, particularly in non-endemic regions where clinicians have limited experience with the presentation. Third: clinical trials to evaluate therapeutic efficacy. The animal evidence for wound healing, blood glucose reduction, and blood pressure lowering is sufficiently documented to justify well-designed human studies. None has been conducted.

The regulatory position

The paper cites the 2023 EFSA assessment (Schrenk et al., 2023) as the most recent regulatory framework, noting that it does not address standardized therapeutic dosing. The paper calls for enhanced regulatory oversight and public health awareness, particularly given that mad honey is increasingly available internationally through online retail, often without adequate labeling of grayanotoxin content or concentration.

The paper’s concluding recommendation is that scientists, policymakers, and local communities, specifically naming Nepal and Turkey, need to work together to reconcile traditional knowledge with evidence-based medicine. That is not a hedge. It is a statement that neither the folk medicine tradition nor the toxicological risk profile alone gives an adequate picture of what mad honey is and how it should be managed.

What We Don’t Know Yet

The 2025 paper is explicit about its limitations in a way that strengthens rather than weakens its credibility. A research review that accurately maps the boundaries of what is known is more useful than one that overstates its conclusions. The limitations below are drawn directly from the paper’s own statements and from the structural gaps in the studies it cites.

• No clinical trial for any therapeutic application: every therapeutic finding documented in this paper is animal study evidence or case report level. The paper explicitly calls for clinical trials as a research priority, not as a hedge, but as an acknowledged gap in the current literature.
• Pharmacokinetics are incompletely characterized: rapid absorption is documented and clearance within 24 hours is inferred from symptom resolution, but the full metabolic pathway for GRAYs in humans is not established. Without this, standardized dosing is not possible.
• No diagnostic biomarker exists: the 100% clinical history diagnosis model (established in Ullah 2018) remains unchanged in 2025. The paper identifies biomarker development as a research priority.
• Genotoxicity data are animal-only at chronic doses: documented in rats and mice at doses that may not correspond to typical human consumption in GRAY exposure terms. No equivalent human data exists. Cucer and Eroz (2010) found no genotoxic effects on cultured human lymphocytes, but that is an in vitro model under different exposure conditions.
• Ovarian toxicity is not characterized in humans: the Yeşil et al. (2024) finding in female rats at 80 mg/kg over 30 days has no human equivalent study. Clinical relevance is unknown.
• The Nepal vs Turkey comparison is not methodologically controlled: Table 1 aggregates studies using different analytical methods, different sample origins, and different time periods. It gives a useful empirical picture but not a direct comparison. Environmental factors drive at least as much of the variation as geography.
• No standardized dosing framework exists: variable toxin concentrations between batches, different susceptibility profiles across populations, and the absence of pharmacokinetic data make a universal safe threshold impossible to establish with the current evidence base. This is the most consequential gap for anyone seeking to use mad honey for therapeutic purposes.

Index Verdict

This 2025 review from Tribhuvan University, Kathmandu, Nepal, is the most current synthesis of grayanotoxin research in peer-reviewed literature and the anchor article for the CMHI library’s therapeutic framing.

What it confirms is nuanced. The pharmacological basis for several traditional therapeutic uses, blood pressure reduction, blood glucose lowering, and wound healing acceleration, is documented in animal models at specific doses. The fracture healing finding, in which mad honey outperformed propolis and normal honey in a rat model (Sahin et al., 2018), is the most commercially relevant individual result. None of these findings has been tested in controlled clinical trials in humans. The paper is direct about this: clinical trials are not a future refinement of the evidence base; they are the missing layer without which therapeutic recommendations cannot responsibly be made.

What the paper also documents is a genotoxicity and ovarian toxicity concern at chronic high doses in animals that has not been previously addressed in the CMHI library. These findings are not grounds for categorical rejection of therapeutic interest; they are animal evidence at doses that may not correspond to typical human consumption. But they are also not grounds for dismissal. They belong in the record, accurately scoped.

The regional comparison confirms what the Ahn (2022) entry established at the batch level: neither Nepal nor Turkey produces honey with a uniform concentration profile. Both regions show wide variability. Batch-level testing remains the only reliable basis for knowing what is in any specific jar.

The paper’s conclusion positions mad honey at the intersection of ethnopharmacology and contemporary toxicology, two frameworks that have historically been treated as opposites. The 2025 review argues they need to be reconciled. CMHI will update this entry when any of the research priorities the paper calls for, pharmacokinetic studies, diagnostic biomarkers, or clinical trials, produce peer-reviewed results.

Anyone experiencing symptoms consistent with grayanotoxin poisoning, bradycardia, hypotension, or significant dizziness after honey consumption should seek emergency medical attention immediately.