Club Moss

What is Club Moss?

Huperzine A, the alkaloid sold as club moss, is a single purified molecule, C15H18N2O, molecular weight 242.32, CAS 102518-79-6, and a licensed prescription medicine in China. The thing in the bottle is not a herb. It is a reversible acetylcholinesterase inhibitor that works by the same enzyme mechanism as donepezil and rivastigmine, and in China it is dispensed for Alzheimer's disease under the name Shuangyiping. In the United States the identical molecule sits on a shelf next to fish oil, at 200 micrograms per capsule, with a label that says club moss.

Two unrelated plants answer to that common name. Lycopodium clavatum L., described by Linnaeus in 1753, is the running or stag's horn club moss of Europe and North America, source of the homeopathic preparation Lycopodium and of the spore powder once sold as photographic flash powder; it is not a commercial source of huperzine A. Huperzia serrata (Thunb.) Trevis. is the Chinese club moss, Qian Ceng Ta, and the only club moss of commercial importance for huperzine A. Liu and colleagues isolated the alkaloid from it in 1986.

The taxonomy moved underneath the common name too. Lycopodium serratum Thunb. is now generally treated as a synonym of Huperzia serrata, and the Huperzia group was for a period raised to its own family, Huperziaceae, which the older huperzine A literature still says; the 2016 Pteridophyte Phylogeny Group classification folds it back into Lycopodiaceae as the subfamily Huperzioideae.

Common Names

• Chinese club moss, toothed clubmoss, Qian Ceng Ta (Huperzia serrata, the huperzine A source)

• Common club moss, running club moss, stag's horn club moss, wolf's claw (Lycopodium clavatum, no commercial huperzine A)

• Shen Jin Cao or Lycopodii Herba, in Chinese materia medica Lycopodium japonicum Thunb., a wind damp herb for joint pain that vendors sometimes mislabel as Lycopodium clavatum

• Fir club moss (Huperzia selago, also written Lycopodium selago, which does contain huperzine A and has poisoned people)

• On labels: Huperzine A, Hup A, huperzia extract, toothed clubmoss extract

Primary Active Compounds

• Huperzine A, a Lycopodium sesquiterpene alkaloid and a reversible, competitive acetylcholinesterase inhibitor. The natural (minus) enantiomer is the active one, and the (plus) enantiomer is substantially weaker at the enzyme

• Huperzine B, a weaker congener alongside it

• Lycopodine, clavatine and annotinine, which dominate Lycopodium clavatum and Lycopodium annotinum and are not huperzine A

Key Note

Content in the plant is tiny and so is the dose, and those two facts define the supply chain. Measured Huperzia serrata tissue runs 0.012 percent huperzine A in root to 0.056 percent in leaf, so a 1 percent standardized extract is an 18 to 85 fold concentration of what it is made from. Nobody eats this as food, and the one documented oral exposure that was not a supplement was a poisoning.

Primary Functions & Benefits

The primary function is reversible, competitive inhibition of acetylcholinesterase: enzyme activity returns once the compound is removed, unlike the irreversible phosphorylation produced by organophosphates. The 2008 Cochrane review describes it as a linearly competitive, reversible inhibitor of acetylcholinesterase with both central and peripheral activity. That places it among the pharmaceutical cholinesterase inhibitors rather than among herbs, at a potency in the same range as the licensed drugs rather than orders of magnitude below them. It is also substantially more selective for acetylcholinesterase than for butyrylcholinesterase, which matters for the anesthesia question below.

The secondary effects, all preclinical, are NMDA receptor antagonism, nerve growth factor upregulation, a shift in amyloid precursor protein processing and reduced brain iron. None has been shown in a human being. What randomized trials measured is a change in cognitive test scores in people with Alzheimer's disease, and the best designed of those trials missed its primary endpoint.

Forms & Standardization

Huperzia serrata whole plant or whole herb extract, standardized to 1 percent huperzine A. The dominant retail form, and a 200 microgram dose is delivered by 20 milligrams of it. Measured Huperzia serrata tissue runs from 118.35 micrograms of huperzine A per gram in root to 562.15 in leaf, while a 1 percent extract holds 10,000 micrograms per gram. That is an 18 to 85 fold concentration, and a 200 microgram dose represents roughly 0.4 grams of leaf or 1.7 grams of root. A 60 capsule bottle at 200 micrograms holds 12 milligrams of huperzine A, tracing back to roughly 21 grams of leaf or 100 grams of root. Which end of that range a batch sits at depends on the tissue that went into it, and a label saying whole plant does not say.

Higher percentage extracts and near pure powder. Suppliers list the same ingredient at 98 to 99 percent purity, at which a 200 microgram dose is about 204 micrograms of powder. That is why the 1 percent grade exists: it is a dilution that makes the material weighable in a blending operation, and it explains a failure mode. If a 99 percent grade is weighed into a formula written for a 1 percent grade, the dose is about 99 times the label, and nothing about the finished capsule looks different.

Natural versus synthetic. Huperzine A has been synthesized repeatedly since Xia and Kozikowski's 1989 route, and catalogs offer both plant extracted and synthetic material. Beyond the regulatory question, chirality is why this matters: the natural (minus) enantiomer is the active one at the enzyme, the (plus) enantiomer is substantially weaker, and a synthesis delivering racemic rather than single enantiomer material therefore delivers less active drug than its weight suggests. The Munich poisoning report of 2000 measures that gap from the other direction: the anticholinesterase isolated from Lycopodium selago had a 2 to 3 fold higher inhibitory potency than the racemic standard run against it. No Supplement Facts panel has a field for enantiomeric purity, and this issue found no published survey of commercial material.

What the trials actually used. Purified huperzine A in a tablet or capsule at the doses below, the isolated alkaloid rather than a crude extract or decoction in every trial this issue read in full. So the trial preparation and the retail preparation are, on their face, the same molecule at the same dose. That is unusual among herbal supplements, and it is why the quality data matter: the dose is right only if the content is right.

Plant part and geography change the number. The western Hunan assays above, with stem at 411.09 micrograms per gram between leaf and root, are a 4.75 fold spread across tissues of one plant, and published surveys report wide variation between wild populations as well. A label that says whole plant extract describes a raw material whose starting assay can vary several fold before any extraction step, which is why the finished extract is standardized to a percentage rather than to a plant weight.

Food Sources

There is no dietary source of huperzine A in the commercial food supply, and no population has a background dietary intake of it. The one exception on record is not a food: a club moss tea brewed by mistake. What the plant was used for in Chinese practice was contusions, strains, swelling and fever, as a crude whole plant preparation rather than a purified alkaloid at a measured microgram dose. Schizophrenia and myasthenia gravis belong to the modern record of the purified drug, not to the herbal tradition.

The one documented route of oral exposure outside supplements is a poisoning. Felgenhauer and colleagues reported in 2000 that two patients in Munich brewed tea from dried herb that was in fact Lycopodium selago, the fir club moss, a plant easily confused in Europe with the common Lycopodium clavatum. They developed sweating, vomiting, diarrhea, dizziness, cramps and slurred speech. The major anticholinesterase the investigators isolated co-chromatographed with authentic huperzine A, in an amount sufficient to explain clinically relevant enzyme inhibition, and they concluded the syndrome should respond to atropine.

A plant alkaloid concentrated tens of fold out of a slow growing moss, delivered in micrograms and acting on the same enzyme as a prescription drug, is not a food and has no history of being eaten as one.

Who Should Take Club Moss

The group with a defensible reason to take huperzine A should be taking it under a prescriber's supervision, which in the United States is not how it is sold.

People with diagnosed Alzheimer's disease, in a system where it is a licensed medicine. In China a clinician selects it, monitors heart rate and gastrointestinal tolerance, knows the medication list, and can stop it. That is a different proposition from a buyer choosing it off a shelf, and the one US trial powered to detect a cognitive effect in this population failed its primary endpoint.

People already on a prescription cholinesterase inhibitor: no. This is the group most likely to reach for it and the group for whom the pharmacology is worst. Adding a second inhibitor of the same enzyme compounds the effect without adding a mechanism, and at least one widely sold 200 microgram product names donepezil on its own caution panel.

Healthy adults seeking a cognitive edge: the evidence is negative, not absent. Wessinger and colleagues ran a randomized double blind crossover trial in 15 exercise trained adults, published in the International Journal of Exercise Science in 2021, giving a single 200 microgram dose or placebo one week apart. No measure of cognitive function differed from placebo, all p values at or above 0.296, and nothing differed on neuromuscular or exercise performance either. The authors concluded that its inclusion in pre-workout supplements warrants reconsideration. One small crossover trial is not the last word, but it is the direct test of the use case most buyers have in mind and it went the wrong way.

People with cognitive complaints but no diagnosis. The Cochrane review by Yue and colleagues, searching to May 23, 2011, found no eligible randomized placebo controlled trials in mild cognitive impairment at all. There is nothing to extrapolate from.

Nobody choosing it off a shelf. That is the honest answer for every remaining group. Several common conditions make huperzine A actively hazardous rather than merely unproven: bradycardia and conduction disease, asthma and chronic obstructive pulmonary disease, a seizure history, peptic ulcer disease, and any scheduled anesthetic. The next section sets out each and the reason it is there.

Who Should AVOID or Use Caution

Contraindications

• Pregnancy and breastfeeding. In 2024 the Dutch National Institute for Public Health and the Environment (RIVM) published letter report 2024-0028 advising consumers not to use preparations containing Huperzia serrata or huperzine A, and specifically not during pregnancy. The developmental data behind that advice: in mice dosed on gestation days 6 to 15, resorbed fetuses and stillbirths rose significantly, giving a derived no observed adverse effect level of 0.08 milligrams per kilogram; in rabbits dosed on gestation days 7 to 18, stillbirths rose, giving a no observed adverse effect level of 0.04 milligrams per kilogram. Both studies used non-oral routes and did not meet OECD standards for group size and fetal examination. Weak studies showing harm are a reason for caution, not for reassurance.

• Concurrent use of any prescription cholinesterase inhibitor: donepezil, rivastigmine, galantamine. Also pyridostigmine, taken by people with myasthenia gravis.

• Children, including as a study aid. The only trial in schoolchildren dates from 1999, and this issue found no replication of it in the indexed literature.

Use Caution

The first four items below are class concerns stated in the prescribing information for donepezil containing products. Huperzine A acts on the same enzyme by the same mechanism.

• Bradycardia, sick sinus syndrome, atrioventricular block, or a rate lowering drug such as metoprolol, atenolol, carvedilol, diltiazem, verapamil or digoxin: cholinesterase inhibitors may have vagotonic effects on the sinoatrial and atrioventricular nodes, producing bradycardia or heart block.

• Asthma and chronic obstructive pulmonary disease: the labeling directs care in patients with a history of either, because cholinergic tone drives bronchoconstriction and secretions.

• Seizure disorders: cholinomimetics are believed to have some potential to cause generalized convulsions.

• Peptic ulcer disease, or regular use of ibuprofen or naproxen: the labeling directs monitoring for gastrointestinal bleeding in patients at increased ulcer risk.

• Any scheduled surgery or procedure requiring anesthesia. See the Critical Safety Point.

• Urinary outflow obstruction, since cholinergic stimulation increases bladder detrusor activity.

• Parkinson's disease, where added cholinergic tone can worsen tremor.

Critical Safety Point

Tell the anesthesiologist. Donepezil labeling states that it is likely to exaggerate succinylcholine type muscle relaxation during anesthesia, and that cholinesterase inhibitors can interfere with anticholinergic medications while acting synergistically with cholinergic agonists. Huperzine A has a reported terminal half life of roughly 6 to 12 hours in humans, so a dose taken the day before surgery is still present. Nobody reviews a Supplement Facts panel in a pre-operative assessment unless the patient volunteers it, and a product labeled club moss does not read as a drug. Of the ways a 200 microgram capsule turns into a clinical event, this is the one with a scheduled date attached. One qualification, so the hazard is not overstated in the wrong direction: succinylcholine is cleared by butyrylcholinesterase, and huperzine A is substantially more selective for acetylcholinesterase, so its effect on succinylcholine should in principle be weaker than that of a non-selective inhibitor. The selectivity is not absolute, the interaction has not been studied in humans, and the correct action is still to disclose it.

Recommended Dosages

Every number here comes from one of three places: a supervised clinical trial, a product licensed as a prescription drug abroad, or a retail label quoted so you can see what it directs. None is a recommendation, none was established in unsupervised use by people without a diagnosis, and the retail numbers are evidence about labeling rather than doses to follow.

Doses used in the US Alzheimer's trial. Rafii and colleagues randomized 210 people with mild to moderate Alzheimer's disease to placebo, 200 micrograms twice daily or 400 micrograms twice daily, 70 per arm, for at least 16 weeks. The 200 microgram twice daily arm was the primary comparison and did not change ADAS-Cog at 16 weeks.

Doses used in Chinese Alzheimer's trials. 200 micrograms twice daily for 8 weeks in a 103 patient trial; 400 micrograms per day for 12 weeks in a 202 patient trial, 100 on huperzine A and 102 on placebo.

Dose used in the only trial in adolescents. 200 micrograms per day for 4 weeks in 34 matched pairs of junior middle school students. This is a 1999 trial in a population that should not be given a cholinesterase inhibitor for school performance; it belongs in the research record, not a dosing table.

Doses used in the TBI trial. In the HUP-TBI Phase 2 pilot, participants were titrated up to 600 micrograms per day and treated for 12 weeks. The trial found no memory benefit, and the placebo arm improved significantly where the treated arm did not.

What US labels actually direct. Retail products are predominantly 100 and 200 micrograms per capsule or tablet. One widely sold 200 microgram product directs one capsule one to four times daily, which is 200 to 800 micrograms per day, so following the top of a retail label delivers the same daily amount as the highest arm of the Rafii trial. The nine Huperzia serrata preparations in RIVM's Dutch market survey carried recommended doses of 0.05 to 0.8 milligrams per day.

Regulatory ceilings and reference values. There are none in the United States: no tolerable upper intake level, no acceptable daily intake, no monograph limit. RIVM examined the question in 2024 and concluded that no safe use level could be established, declining to derive either an acceptable daily intake or an acute reference dose because the toxicological dataset was inadequate and because genotoxicity could not be properly evaluated against current test guidelines.

Duration

The Rafii trial treated participants for at least 16 weeks, which is the longest exposure this issue can document from a report it read in full. Chinese trials in the Cochrane and systematic review datasets ran 6 to 16 weeks, and the 12 schizophrenia trials pooled by Zheng averaged 11.7 weeks. No published controlled trial located for this issue ran beyond roughly six months, so the safety of continuous use for a year or more in a person without a diagnosis is unstudied, and reviewers have flagged that gap explicitly.

Timing & Administration

What follows describes how the molecule behaves and how it was given in trials. It is not a schedule being recommended.

Absorption is fast. In 12 healthy volunteers given a single 0.4 milligram tablet, huperzine A appeared in plasma at 5 to 10 minutes and peaked at 58.33 plus or minus 3.89 minutes, with a terminal half life of 716 plus or minus 130 minutes, close to 12 hours. Published half life estimates vary with study and formulation across a range of roughly 6 to 12 hours. That range is why the trials dosed twice daily, and it is the number that matters for the surgery disclosure above.

Points drawn from the trial record:

• Nausea and vomiting were among the side effects recorded in the Rafii trial, described as mild to moderate, and taking a cholinesterase inhibitor with food is the standard mitigation for that class.

• An evening dose is a poor idea: insomnia and vivid dreaming are documented cholinergic effects, and raising central acetylcholine at bedtime works against sleep.

• Do not stack it with a second product that also contains huperzine A. Pre-workout and nootropic blends frequently include it, and that is the easiest way to double a dose without noticing.

Timeline of Effects

Within one hour. Enzyme inhibition follows the plasma curve, which peaks at roughly 1 hour, and that is the window in which cholinergic side effects show up: nausea, sweating, excess salivation, loose stools.

Days 1 to 3. Steady state arrives within about two days on twice daily dosing, when cholinergic side effects are most likely to appear. It is not an all clear: bradycardia, gastrointestinal bleeding and weight loss are later onset problems for this drug class, and the trials were too small and short to say when they emerge.

Weeks 6 to 8. This is where the Chinese trial data first separate from placebo. In the 103 patient trial at 200 micrograms twice daily for 8 weeks, 29 of 50 treated patients showed improvement in memory, cognitive and behavioral function against 19 of 53 on placebo, 58 percent against 36 percent, p less than 0.05. The 2008 Cochrane review found pooled benefit on ADAS-Cog at six weeks and on the Mini-Mental State Examination; the 2013 systematic review found Mini-Mental State Examination benefit from eight weeks and reported no significant change on ADAS-Cog.

Weeks 11 to 16. This is the interval that decides what you believe about this ingredient, and where the effect faded in the one trial designed to catch it. In Rafii's study the 400 microgram twice daily arm showed a 2.27 point ADAS-Cog improvement at week 11 against a 0.29 point decline on placebo, p equal to 0.001. By week 16 the same arm showed 1.92 points of improvement against 0.34 on placebo, p equal to 0.07. The placebo group had caught up. The 200 microgram twice daily arm, the prespecified primary analysis, showed nothing at 16 weeks.

What counts as a fair trial for a person without dementia. There is no timeline to give, because there is no positive trial in that population to draw one from. The authors of the 12 week TBI trial, in which the placebo arm improved significantly and the treated arm did not, wrote their paper as a caution against reading improvement in an uncontrolled setting as a drug effect. That is the warning for anyone judging this alone.

When to conclude it is not working. The Rafii trial is the only timeline evidence here, and it points away from patience. Waiting longer than that trial did is not a strategy the trial supports. Side effects in the first week are the expected onset window for this mechanism, not something to push through.

Benefits of Taking Club Moss

Symptomatic cognitive benefit in Alzheimer's disease: moderate evidence from low quality trials, and a failed primary endpoint in the best one. The 2008 Cochrane review by Li and colleagues pooled six trials with 454 patients and found benefit on Mini-Mental State Examination, ADAS-Cog at six and twelve weeks, Clinical Dementia Rating, CIBIC-plus and activities of daily living. The part usually left out is that the same review found no benefit on the Hasegawa Dementia Scale, weighted mean difference 2.78, confidence interval minus 0.17 to 5.73, or the Wechsler Memory Scale, weighted mean difference 6.64, confidence interval minus 3.22 to 16.50. The reviewers concluded that only one included study was of adequate quality and size and that the evidence was inadequate to make any recommendation about its use.

Cognitive deficits in schizophrenia, as an add-on to antipsychotics: positive pooled result, entirely from one country. Zheng and colleagues pooled 12 randomized trials with 1,117 participants, all conducted in China; memory quotient on the Wechsler Memory Scale Revised favored huperzine A with a weighted mean difference of 10.59, confidence interval 5.65 to 15.53. This is a prescription combination managed by a psychiatrist rather than anything a buyer can act on.

Memory in adolescents reporting memory complaints: one unreplicated trial from 1999. Sun and colleagues matched 34 pairs of junior middle school students complaining of memory inadequacy on sex, class and baseline memory quotient, then gave 200 micrograms per day for four weeks. Memory quotient reached 115 plus or minus 6 on huperzine A against 104 plus or minus 9 on placebo, p less than 0.01. Twenty-seven years later this issue found no replication of it in the indexed literature, and it is the single study most often cited to sell huperzine A to students. One trial of 68 adolescents is not a basis for giving children an acetylcholinesterase inhibitor.

Vascular dementia: no. The Cochrane review by Hao and colleagues in 2009 found one eligible randomized trial, with 14 participants, and a Mini-Mental State Examination weighted mean difference of 2.40, confidence interval minus 4.78 to 9.58, wide enough to include meaningful benefit and meaningful harm. The reviewers found no convincing evidence of value.

Traumatic brain injury, cognitive enhancement in healthy people, and mild cognitive impairment: no. The HUP-TBI Phase 2 pilot, titrated to 600 micrograms per day for 12 weeks, found no difference from placebo on the California Verbal Learning Test second edition; no measure of cognition, neuromuscular performance or exercise performance differed from placebo in the 15 person exercise crossover trial, in which subjective difficulty after exercise was worse on huperzine A; and the Cochrane review published in 2012, searching to May 2011, found zero eligible trials in mild cognitive impairment.

Where it ranks against the drugs. A 2020 Bayesian network meta-analysis of 37 randomized trials and 14,705 participants ranked doses of donepezil, galantamine, rivastigmine, memantine, huperzine A and Ginkgo biloba extract EGb761 in mild to moderate Alzheimer's disease. Huperzine A topped none of the rankings for cognition, acceptability, safety, daily living or global impression.

Potential Negatives & Side Effects

Most reported side effects of huperzine A are the predictable consequence of raising acetylcholine, and those scale with dose. That does not make the list below complete. The developmental findings above are not cholinergic excess in the person taking the capsule, RIVM reported that genotoxicity could not be properly evaluated, and the trials were too small to surface anything uncommon.

Reported in human trials and surveillance

• Nausea and vomiting, recorded as mild to moderate in the Rafii trial. In that trial 177 of 210 randomized subjects completed the treatment phase, so 33 did not, and attrition at that rate over 16 weeks is itself a tolerability signal.

• Diarrhea, sweating, excess salivation and lacrimation

• Dizziness, blurred vision, insomnia and vivid dreaming

• Reduced heart rate, the expected consequence of vagotonic action on the sinoatrial node

• Muscle cramps and twitching

• Worse subjective difficulty after exercise, which is a separate measure from perceived exertion: 6.8 on huperzine A versus 5.7 on placebo in the exercise crossover trial, p equal to 0.002, with Borg ratings of perceived exertion and heart rate unchanged and no gain in strength, endurance or aerobic capacity

Reported in overdose

The Dutch Poisons Information Center recorded two huperzine A overdose cases in which patients developed nausea, tremors, dysarthria, diarrhea and blurred vision; slurred speech was also a presenting finding in the two Munich tea poisonings. Dysarthria from a supplement marks this as a drug effect rather than a digestive upset.

Documented for the drug class, not quantified for huperzine A

The bradycardia, bronchoconstriction, gastric acid and seizure threshold effects in the caution list above are class concerns stated in donepezil labeling. They have not been quantified for huperzine A, because the trials were small and short, and treating that absence as evidence of absence would be the error here.

The structural negative

This is a compound with a 6 to 12 hour half life, a steep dose response in animals, and no monitoring attached to its sale: no pharmacist checking it against a medication list, no prescriber to notice a resting pulse of 48, no route by which an adverse event gets connected back to the product. In the three published US trials located for this issue, the Rafii Alzheimer's trial, the HUP-TBI pilot and the exercise crossover, it either failed its primary endpoint or showed nothing at all.

Deficiency Symptoms

There is no huperzine A deficiency. Huperzine A is not a nutrient, has no dietary source, and has no reference intake set by any authority this issue could identify.

What huperzine A addresses

It addresses a pharmacological target, not a nutritional gap: the breakdown of acetylcholine in a brain where basal forebrain cholinergic neurons are dying, which is the rationale for donepezil, rivastigmine and galantamine as well. The target exists because neurons have been lost, not because a substance is missing from food.

Bottom line

Contrast a genuine trace element. Iodine has a recommended dietary allowance of 150 micrograms per day for adults, a defined deficiency disease, a biomarker in urinary iodine, and food sources with known content. Huperzine A has a label dose of 200 micrograms and none of the rest. That the two figures land in the same order of magnitude is a coincidence of potency, and a label that puts huperzine A in a micrograms column next to vitamins borrows a visual grammar that does not apply.

Toxicity Symptoms

At high intake

Cholinergic excess is the toxic syndrome. In male Wistar rats given single oral doses, the lowest observed adverse effect level was 0.3 milligrams per kilogram body weight, at which tremors and increased salivation appeared, worsening with dose and reversing within hours. In mice, 100 micrograms per kilogram produced about 2.5 percent inhibition of brain acetylcholinesterase and roughly a 5.5 fold rise in acetylcholine, while 500 micrograms per kilogram produced about 21.5 percent inhibition, nearly 13 times basal acetylcholine, and chewing, salivation, fasciculation and tremors. A fivefold dose increase moved the animals from a modest enzyme effect to visible cholinergic excess, and other mouse work adds muscular weakness, lacrimation, diarrhea, paralysis and blurred vision.

Those animal doses do not transfer directly to a human dose. The 0.3 milligram per kilogram rat level corresponds to 21 milligrams in a 70 kilogram person, roughly 100 times a 200 microgram capsule. The rabbit developmental no observed adverse effect level of 0.04 milligrams per kilogram corresponds to 2.8 milligrams, roughly 14 times a capsule, and came from a non-oral study that did not meet OECD standards. A margin you cannot calculate is not a large margin.

Signs to reduce or stop

• Resting heart rate falling, lightheadedness on standing, fainting

• Sweating, drooling, watery eyes, or loose stools that start after beginning the product

• Muscle twitching or cramping

• Slurred speech or tremor, which were presenting findings in both the Dutch overdose cases and the Munich tea poisonings

• Wheezing or increased inhaler use in anyone with asthma

• Vomiting that does not settle, or any black or tarry stool

General note

The Munich investigators noted the syndrome should respond to atropine. That is a decision for an emergency department, not an instruction for home use. Two features make this ingredient harder to manage than a bitter digestive herb: the effective unit is a microgram, so a weighing error multiplies the dose rather than nudging it, and the symptoms arrive within an hour, before most people connect them to a capsule swallowed with breakfast. If cholinergic symptoms appear, stop the product, keep the bottle, and take it with you to be seen.

How Club Moss Works

Acetylcholinesterase hydrolyzes acetylcholine at the synapse so the signal ends. Huperzine A occupies the enzyme's active site and stops that hydrolysis, so each released molecule of acetylcholine acts for longer. That is the whole of the primary mechanism, and it is the same mechanism the three approved Alzheimer's cholinesterase inhibitors use.

The binding is harder to explain than its strength would suggest. Raves and colleagues solved the crystal structure of acetylcholinesterase complexed with optically pure (minus) huperzine A in 1997 and reported an unexpected orientation, with surprisingly few strong direct interactions with protein residues to account for its high affinity; what the affinity tracks is hydrophobic contact with aromatic residues lining the active site gorge. That structure used Torpedo californica enzyme, and later work with human acetylcholinesterase mutants found the orientation in solution differs significantly from it, so the contact map is not settled. The compound crosses the blood brain barrier, unlike pyridostigmine, which protects only peripheral enzyme.

The secondary mechanisms listed earlier are cell culture and animal findings with no human counterpart. Reviews sympathetic to that work concede the point: with study durations under six months and no pathological endpoints, it is premature to conclude that huperzine A slows the progression of Alzheimer's disease. The mechanism demonstrated in people is enzyme inhibition, and the clinical consequence of that inhibition in the best powered US trial was nothing at the prespecified dose.

Synergistic Supplements

This issue found no human trial of huperzine A combined with another supplement. The only combination with controlled data is a prescription one, the schizophrenia add-on described above.

Pairings commonly sold, with no trial support

Citicoline, alpha-GPC, bacopa, ginkgo, phosphatidylserine and acetyl-L-carnitine are routinely co-formulated with huperzine A, and nobody has tested whether any of them changes the effect. The theoretical case for a choline donor, supplying substrate while slowing degradation, has not been tested at these doses in people.

The pairing argument that runs the other way

In the 22 product analysis by Crawford and colleagues, products listing huperzine A also carried undeclared noopept, vinpocetine, sulbutiamine, higenamine, halostachine, hordenine, beta-phenylethylamine, demelverine and two dimethylhexylamine isomers. A single ingredient capsule at a stated microgram dose is a more controllable product than a proprietary blend, though that is reasoning rather than something the analysis tested: it did not report content accuracy separately for blends. The unhelpful pairing is anything else cholinergic, and a second product containing huperzine A is the easiest way to double a dose in a stack without noticing.

Interactions & What NOT to Take

Additive: other cholinesterase inhibitors. Donepezil, rivastigmine, galantamine and pyridostigmine all inhibit the same enzyme, so adding huperzine A compounds the effect without adding a mechanism, and the risk is cholinergic excess: bradycardia, vomiting, diarrhea, sweating, cramping. This is the most important combination to avoid, and the people most motivated to try it are people already taking a prescription for memory.

Additive: cholinergic agonists. Bethanechol, pilocarpine and cevimeline. Donepezil labeling states that a synergistic effect may be expected when cholinesterase inhibitors are given with cholinergic agonists.

Additive: drugs that slow the heart. Metoprolol, atenolol, carvedilol, bisoprolol, diltiazem, verapamil and digoxin. The vagotonic effect described above stacks with every drug on this list.

Opposed: anticholinergics. Oxybutynin and tolterodine for bladder, diphenhydramine and hydroxyzine as antihistamines, scopolamine for motion sickness, benztropine and trihexyphenidyl for parkinsonism, dicyclomine and hyoscyamine for gut spasm, amitriptyline, nortriptyline, paroxetine, olanzapine and clozapine. Donepezil labeling states that cholinesterase inhibitors have the potential to interfere with anticholinergic medications. The consequence runs both ways: huperzine A can blunt a drug the patient needs, and stopping the anticholinergic can unmask cholinergic excess.

Neuromuscular blockers and anesthesia. Succinylcholine, plus the non-depolarizing agents rocuronium and vecuronium, which cholinesterase inhibition antagonizes; that antagonism is the principle behind reversing them with neostigmine. The selectivity caveat set out under the Critical Safety Point above still applies, and so does the instruction: disclose it.

Nonsteroidal anti-inflammatory drugs. Ibuprofen, naproxen, diclofenac and aspirin. Increased cholinergic tone increases gastric acid secretion, and donepezil labeling names patients on these drugs among those to monitor for gastrointestinal bleeding.

Asthma and COPD therapy. Albuterol and formoterol oppose bronchoconstriction, and cholinergic stimulation promotes it; ipratropium and tiotropium are themselves anticholinergics and fall in the opposed category above.

Organophosphate and carbamate exposure. Insecticides act on the same enzyme, by different kinetics. The organophosphates chlorpyrifos and malathion inhibit it essentially irreversibly, by phosphorylation; the carbamate carbaryl inhibits it reversibly, with slow recovery. Anyone with occupational exposure to either class has no reason to add a third inhibitor on top.

One useful non-interaction. Huperzine A does not appear on the World Anti-Doping Agency Prohibited List, which does not make these products safe for a tested athlete. Crawford and colleagues found huperzine A labeled products containing higenamine, which WADA prohibits at all times as a beta-2 agonist, and 1,5-dimethylhexylamine, which WADA lists under specified stimulants as octodrine. The doping risk comes from the undeclared contents, not from the ingredient on the front of the bottle.

Quality, Testing & Adulteration

Crawford and colleagues analyzed 22 supplement products listing huperzine A by liquid chromatography quadrupole time of flight mass spectrometry, in Clinical Toxicology 2020, 58(10):991 to 996:

• 16 of 22 products, 73 percent, had at least one labeled ingredient that was not detected

• 16 of 22, 73 percent, contained compounds not declared on the label

• 9 of 22, 41 percent, listed ingredients that in the authors' assessment did not meet FDA's definition of a dietary ingredient

• Measured huperzine A ranged from detected but below the limit of quantification up to 267.1 micrograms per serving. A value reported as below the limit of quantification is not zero and should not be read as one

• Only 2 of 22 contained huperzine A within 10 percent of the declared amount

The undeclared compounds included vinpocetine, noopept, sulbutiamine, higenamine and hordenine, the last of which sits on FDA's Dietary Supplement Ingredient Advisory List.

What third party programs do and do not cover. NSF Certified for Sport and Informed Sport test for banned substances and verify label claim, but neither establishes that the huperzine A came from Huperzia serrata rather than a synthetic route, and neither establishes enantiomeric purity. Nor does either substitute for a pharmacopeial standard. Botanicals such as ginkgo leaf extract have USP monographs fixing identity tests, assay limits and impurity specifications; Huperzia serrata extract does not appear among USP's botanical monographs, which leaves each manufacturer writing its own specification and a verifier checking against that in-house number rather than a public one.

What the package itself can settle. Very little, which is the point. Check that the panel names Huperzia serrata rather than only club moss, that the dose is stated in micrograms of huperzine A rather than milligrams of an unspecified extract, that it is a single ingredient product, and that any certification mark corresponds to a certificate you can look up. You cannot verify species, enantiomeric purity, or whether the molecule grew in a plant. Radiocarbon methods of the kind standardized as ASTM D6866 separate recently fixed biobased carbon from fossil carbon, which bears on the last question without answering it: such a test cannot identify a species, nor separate plant extraction from a synthesis using biobased feedstock.

Special Considerations

Children and students. Giving a child an acetylcholinesterase inhibitor for exam performance is a pharmacological decision, not a nutritional one, and the 1999 trial is the whole evidence base.

Older adults. The population with the most trial data is also the one most likely to be on a drug from the interacting lists above, to have a conduction abnormality, and to have no one reviewing a supplement against a medication list.

Conservation and sourcing. Huperzia serrata grows very slowly from spore and has not been brought into cultivation at commercial scale, which is why the plant derived supply has depended on wild collection. The species is not among the plants listed in the CITES appendices, and this issue found no sustainability certification scheme covering wild collected club moss, so a buyer choosing plant derived over synthetic on environmental grounds is choosing the option with the wild harvest attached.

Regulatory status in China. The approval date is not cleanly citable. Secondary sources give 1994 and 1996, and several credit a State Food and Drug Administration, an agency that did not carry that name until 2003. What is solid is the drug name rather than the date: shuangyiping appears as a Chinese language search term for huperzine A in the published search strategy of the Cochrane vascular dementia review. Treat it as a mid 1990s approval and do not repeat a specific year without a primary regulatory document.

Regulatory status in the United States. Huperzine A is sold as a dietary supplement ingredient, and whether it lawfully qualifies as one is unsettled. The Department of Defense Operation Supplement Safety program states that it is unclear whether huperzine A can be legally marketed as a dietary supplement in the US. FDA's 2016 draft guidance, discussed above, has not been through final rulemaking, which is why the question stays open rather than settled against the ingredient. FDA has acted on claims rather than on the ingredient, and no FDA action has removed huperzine A from the market.

Research Status & Evidence Quality

Tier 1, strong evidence from adequately powered randomized trials. Nothing. The largest positive randomized trial is Zhang and colleagues in 2002, 202 patients across 15 Chinese centers for 12 weeks, which is multicenter and positive but neither large nor long by the standards this tier requires, and the one trial built to Western regulatory scale failed its primary endpoint.

Tier 2, moderate evidence with a clear negative finding. The Rafii phase 2 trial is the best conducted study in the literature and it is negative on its primary endpoint. The prespecified primary analysis, ADAS-Cog change at week 16 on 200 micrograms twice daily, showed no effect, and the journal classified it as Class III evidence that 200 micrograms twice daily has no demonstrable cognitive effect in this population. The 400 microgram arm, whose week 11 numbers are given above, was a secondary analysis. Reporting that week 11 result without the week 16 decay and without the failed primary endpoint is the standard way this trial gets misrepresented.

Tier 2, moderate evidence from pooled low quality trials. The 2008 Cochrane review, six trials and 454 patients, searched to February 1, 2006. The 2013 systematic review by Yang and colleagues, 20 randomized trials and 1,823 participants, protocol CRD42012003249, searched to June 2013, found benefits on Mini-Mental State Examination at 8, 12 and 16 weeks and on activities of daily living at 6, 12 and 16 weeks, and concluded that the findings should be interpreted with caution because most included trials carried a high risk of bias. One citation point worth getting right: the 2013 Yang paper is a systematic review in PLOS ONE, not a Cochrane review. The Cochrane reviews here are three separate documents by different author teams, Li 2008 on Alzheimer's disease, Hao 2009 on vascular dementia and Yue 2012 on mild cognitive impairment.

Tier 3, single small trials. The adolescent memory trial is 34 matched pairs from 1999. The vascular dementia Cochrane review found one trial of 14 with a confidence interval spanning harm and benefit. The TBI trial was a small Phase 2 pilot and the exercise crossover trial enrolled 15, both negative.

Tier 4, absent. No eligible randomized trials in mild cognitive impairment, no published controlled trial located beyond roughly six months, no trial of any huperzine A supplement combination, no human study of the non-cholinergic mechanisms, and no human pharmacokinetic study of a retail supplement as opposed to a pharmaceutical tablet.

Tier 5, preclinical only. Amyloid processing, tau phosphorylation, brain iron, nerve growth factor, NMDA antagonism and nerve agent prophylaxis, all cell culture and animal work at non-human doses.

Research Limitations

The trials are small, short and geographically concentrated. The 2008 Cochrane review judged the methodological quality of most included trials to be not high; the 2013 analysis judged most to carry a high risk of bias and flagged publication bias. And the whole literature was generated with pharmaceutical grade material of known content, which is not what the 2020 survey found inside 20 of 22 retail products.

Summary & Key Takeaways

Bottom Line

Club moss on a label means huperzine A: a reversible acetylcholinesterase inhibitor, substantially more selective for acetylcholinesterase than for butyrylcholinesterase, with a reported 6 to 12 hour half life and a prescription license in China. It is sold in the US at 200 micrograms with no monitoring, and the best designed trial of it, 210 patients across three arms for at least 16 weeks, missed its primary endpoint. In exercise trained adults the direct test found no cognitive benefit and worse subjective difficulty afterward. Two of 22 surveyed products held huperzine A within 10 percent of what their labels claimed.

Key Safety Points

• Do not combine it with donepezil, rivastigmine, galantamine or pyridostigmine. Same enzyme, additive effect.

• Tell your anesthesiologist. Donepezil labeling warns that cholinesterase inhibitors exaggerate succinylcholine type muscle relaxation, and a dose from yesterday is still in you.

• Bradycardia, asthma, a seizure history, peptic ulcer disease or regular NSAID use all put you in the caution group, because increased cholinergic tone bears on every one of them.

• RIVM concluded in its 2024 report that no safe use level could be established and advised against use, with pregnancy singled out.

• Slurred speech, tremor, drooling, sweating, loose stools or a falling pulse after starting it are cholinergic signs. Stop, keep the bottle, get seen.

• Avoid proprietary blends. The undeclared stimulants turned up in products listing huperzine A, and a blend gives a manufacturer more places to put something you did not ask for.

Special Note

Both useful checks live on the panel rather than in the marketing. First, does it name Huperzia serrata, not just club moss? The common name is shared with Lycopodium clavatum, which is not a commercial huperzine A source, and with Lycopodium japonicum, the Shen Jin Cao of Chinese materia medica. Second, is it a single ingredient product with the dose stated in micrograms of huperzine A? If either answer is no, you do not know what you are holding. On the question that bears on its legal status, plant extracted or synthetic, no US label will tell you, and no test available to a buyer settles it either.

Researched and drafted with AI assistance. Every claim verified against primary sources and human-reviewed before publication. Last reviewed: [DATE]. Corrections: reply to any issue.

This issue is educational and is not medical advice. Huperzine A acts on the same enzyme as several prescription drugs, so if you take any medication or have a diagnosed heart, lung, neurological or gastrointestinal condition, talk with a clinician before you start, stop or change anything described here.