Forms, Dosing, and the Herb vs Drug Question

Almost every argument about Artemisia annua comes down to one practical question: how much artemisinin is in the thing you are actually holding? A Nobel Prize was awarded for isolating a molecule from this plant. A supplement capsule of the plant is not that molecule, and a mug of tea is not either. This article is about the size of that gap, measured rather than asserted.

It also carries the full cautions list for the whole Benefits set — neurotoxicity, pregnancy, liver, allergy, and the drug interactions that catch people out.

One thing this article will not do is give you a therapeutic dose. There isn’t one to give. For malaria there is a proper drug regimen a clinician prescribes; for everything else there is no dose with credible human evidence behind it. What follows tells you how to understand the numbers on a label, not how to treat yourself with one.

Table of Contents

1. The Four Different Things Sold Under This Name
2. How Much Artemisinin Is Actually in the Leaf
3. What a Cup of Tea Delivers — the Measured Number
4. Cold Water Versus Boiling in a Real Kitchen
5. The Whole-Plant Argument, and Why It Has Not Won
6. Why the Pharmaceutical Derivatives Exist
7. Doses People Actually Use, and What Backs Them
8. Reading a Supplement Label
9. Drug Interactions: CYP2B6, CYP3A4 and Autoinduction
10. Safety and Cautions in Full
11. Cost and Sourcing
12. The Bottom Line
13. Key Research Papers
14. Connections

The Four Different Things Sold Under This Name

Conversations about this plant go wrong because four quite different products share a vocabulary.

The first three are supplements. Only the fourth has been tested in the trials that made this plant famous. When someone cites “the artemisinin research” to justify the first three, that is the substitution to watch for.

How Much Artemisinin Is Actually in the Leaf

Artemisinin is a minor constituent of A. annua, produced in glandular trichomes on the leaf surface, and its concentration is one of the most variable numbers in medicinal plant chemistry.

• Genetics dominates. Breeding programmes exist precisely because wild-type plants are so poor. Classic work on the genetics of artemisinin content and the breeding of high-yielding cultivars established that content is heritable and that selected lines vastly outperform wild collections.
• Geography and season matter. A study of A. annua from three regions of Uganda found substantial variation in antimalarial components between regions. Content also shifts across the growing season, peaking around flowering in most cultivars.
• Handling matters. Artemisinin degrades with heat and time. Drying temperature, storage duration and light exposure all reduce it.

Put concretely: two visually identical jars of dried sweet wormwood can differ several-fold in the amount of active compound they contain, and nothing on either label will tell you which is which. For a culinary or aromatic use that is irrelevant. For anything where the dose determines whether a parasite lives or dies, it is disqualifying.

It is worth noting one contrast, because it appears in the retracted-trial literature: Artemisia afra, the African species frequently sold and promoted alongside A. annua, contains negligible artemisinin. Products and protocols that treat the two as interchangeable are not making a small error.

What a Cup of Tea Delivers — the Measured Number

This has been measured properly, in people, and the number is the single most useful fact on this page.

Fourteen healthy male volunteers each drank one litre of tea prepared from 9 g of dried A. annua leaves, with artemisinin quantified in both the tea and their blood by HPLC:

• The litre contained 94.5 mg of artemisinin — about 19 percent of the usually recommended daily dose.
• Peak plasma concentration: 240 ± 75 ng/mL. Area under the curve: 336 ± 71 ng/mL·h.
• Absorption from tea was faster than from solid oral dosage forms, with similar bioavailability.
• The authors’ conclusion, verbatim in spirit: concentrations are sufficient for clinical effects but insufficient to recommend such preparations as equivalent substitutes for modern artemisinin drugs.

Some arithmetic, because it clarifies what “a bigger cup” would mean. If 9 g of good-quality leaf yields roughly 95 mg of artemisinin in a litre, then matching a typical adult daily artemisinin dose would take on the order of 45–50 g of dried leaf and around five litres of tea per day — assuming your leaf is as good as theirs. Nobody drinks that, the tea is bitter, and the leaf-quality assumption is the one you cannot check.

A related detail from analytical work on A. annua herbal teas: infusions extract a range of flavonoids and other constituents as well as artemisinin, and in-vitro work suggests artemisinin is not the only antiplasmodial compound in the infusion. That is scientifically interesting and it is the seed of the whole-plant argument below. It does not change the dose problem.

Cold Water Versus Boiling in a Real Kitchen

Tu Youyou’s breakthrough came from Ge Hong’s fourth-century instruction to soak the herb in cold water and wring out the juice rather than boil it — because heat degrades artemisinin. People reasonably ask whether that means a cold infusion at home is better.

The honest answer has two parts, and they point in opposite directions.

The chemistry is real. Researchers tested ancient Chinese preparation methods directly and found that the traditional approaches produced artemisinin-rich extracts with potent antimalarial activity in vitro — the pressed-juice method genuinely preserves more of the compound than prolonged boiling does.

But artemisinin is barely water-soluble at any temperature. Hot water extracts more of it into solution than cold water does, even while degrading some. The German pharmacokinetic study used a hot infusion and still recovered 94.5 mg per litre. So “cold is better” is not a straightforward kitchen rule; it was a decisive clue about a heat-labile molecule in a laboratory extraction, not a recipe optimisation.

Either way, the answer to a five-fold dosing shortfall is not a better steeping technique. It is a measured product.

The Whole-Plant Argument, and Why It Has Not Won

The serious scientific case for the plant over the molecule deserves to be stated properly rather than dismissed.

Researchers at Worcester Polytechnic Institute and collaborators have shown that in rodent malaria, dried whole-plant A. annua outperformed an equivalent dose of pure artemisinin, associated with roughly 40-fold greater artemisinin bioavailability in mouse blood. The proposed explanation is elegant: flavonoids and other leaf constituents inhibit the cytochrome P450 enzymes that would otherwise clear artemisinin, so the same milligrams last longer and reach higher levels. Later work reported that whole-plant material also slowed the evolution of resistance in rodent malaria and retained activity against artemisinin-resistant parasites, and independent in-vitro work found that Artemisia extracts affect human CYP2B6 and CYP3A4 differently from artemisinin alone.

That is a coherent, testable hypothesis: the plant as a natural combination therapy with a built-in pharmacokinetic booster. Why has it not changed practice?

1. Rodent malaria is not human malaria. The bioavailability finding is a mouse finding.
2. The two large human trials that claimed to confirm it were retracted in 2020. The whole-plant case currently has no standing human efficacy trial.
3. The dose problem is unsolved even if the hypothesis is right. A pharmacokinetic advantage per milligram does not help when the milligrams themselves are unmeasured.
4. The CYP-inhibition mechanism is a two-edged sword. A plant that inhibits CYP3A4 well enough to boost its own artemisinin will also boost other CYP3A4 substrates — statins, calcium-channel blockers, immunosuppressants, some anticoagulants, many chemotherapy agents.
5. The asymmetry of being wrong. If the hypothesis is right and unused, we keep a therapy that works. If it is wrong and adopted, artemisinin resistance accelerates.

The right test is a properly powered, independently monitored randomised non-inferiority trial with molecular genotyping. Until that exists, this belongs in the research column.

Why the Pharmaceutical Derivatives Exist

It is easy to read the semisynthetic derivatives as corporate meddling with a natural product. They are actually solutions to three specific defects of the natural molecule.

• Solubility. Artemisinin dissolves poorly in water and in oil, which makes both an injection and a reliable tablet difficult. Artesunate (water-soluble) and artemether (oil-soluble) were built to fix exactly that. Being able to give a full, exact dose intravenously is what makes artesunate the treatment for severe malaria.
• Potency and absorption. All the derivatives convert to dihydroartemisinin, which is more potent than the parent compound and more consistently absorbed.
• Predictability. A pharmacopoeial tablet contains a stated number of milligrams, batch-tested. That is the entire difference between a therapy and a hope.

Note what the derivatives did not change: the peroxide bridge. The active chemistry is still the plant’s. What chemists added was deliverability.

Doses People Actually Use, and What Backs Them

For completeness, here is what circulates — with what supports it stated plainly.

The recurring “three days on, four days off” style of supplement cycling deserves a comment, because it is sometimes justified by real pharmacology — artemisinin’s autoinduction of its own metabolism. It is true that blood levels fall over consecutive days of dosing. It is not true that anyone has established a cycling schedule that fixes this for any human indication.

Reading a Supplement Label

If a product will not tell you how many milligrams of artemisinin are in a capsule, you cannot compare it to any study in the literature. That is not pedantry — comparison to the literature is the only reason to care about the number.

Drug Interactions: CYP2B6, CYP3A4 and Autoinduction

This is the most under-appreciated risk with artemisinin supplements, and it runs in both directions depending on what you are taking.

Artemisinin induces drug-metabolising enzymes. Human studies identified CYP2B6 as the principal enzyme metabolising artemisinin, with CYP3A4 contributing, and showed that artemisinin induces its own metabolism through CYP2B6 — blood levels drop measurably over several days of continued dosing. Further work in human liver microsomes and hepatocytes confirmed that artemisinin antimalarials both inhibit and induce various P450 enzymes.

Whole dried-leaf material can inhibit the same enzymes. The flavonoid fraction of the plant inhibits CYP2B6 and CYP3A4 — the mechanism proposed for the whole-plant bioavailability advantage. In-vitro comparison found that Artemisia extracts differ from artemisinin alone in their effects on human CYP2B6 and CYP3A4.

Practically, that means a purified artemisinin capsule and a whole-leaf capsule may push a co-administered drug in opposite directions. Categories worth a pharmacist conversation:

• CYP3A4 substrates with narrow margins — tacrolimus, ciclosporin, sirolimus, some statins, certain antiarrhythmics, many kinase inhibitors and chemotherapy agents.
• Anticoagulants — warfarin above all; also the direct oral anticoagulants, several of which are CYP3A4 substrates.
• Antiretrovirals — a class where induction can cause virological failure. Efavirenz and nevirapine also interact at CYP2B6 themselves.
• Anticonvulsants — carbamazepine, phenytoin, phenobarbital induce CYP2B6/3A4 and will lower artemisinin levels; the reverse is also possible.
• Hormonal contraceptives — CYP3A4 induction is the classic mechanism of contraceptive failure with herbal products.
• Other antimalarials — taking a supplement alongside a prescribed ACT is exactly the uncontrolled dosing that guidelines exist to prevent.

Tell any prescriber that you are taking this. The interaction is not theoretical and the direction is not obvious.

Safety and Cautions in Full

Neurotoxicity — the animal signal

In the early 1990s, high, repeated intramuscular doses of the oil-soluble derivatives arteether and artemether produced a distinctive and fatal brainstem lesion in dogs and rats, affecting auditory and vestibular nuclei. This was a serious finding that shaped the development of the entire class, and later work characterised the pattern in detail.

The reassuring counterweight is that this has not been reproduced in humans at therapeutic doses. A review asking directly whether currently deployed artemisinins are neurotoxic found the human evidence largely negative; case-control auditory evaluations of patients treated with artemisinin derivatives for multidrug-resistant falciparum malaria found no clear hearing loss; and a neuropathological assessment of patients who died despite artemether treatment for severe malaria did not show the animal lesion.

The practical reading: short, correct antimalarial courses have a good human safety record. The animal data are still the reason nobody should improvise high daily doses of an oil-soluble artemisinin derivative for months. The ototoxicity questions raised by the long-term artesunate breast-cancer study sit in the same territory.

Pregnancy — the clearest contraindication for supplement use

Artemisinins are embryotoxic in animal studies, causing embryo death and malformation in a narrow, defined window of early gestation, apparently through depletion of embryonic red-cell precursors. That work is the basis for caution in the first trimester.

Human data are more reassuring than the animal work. A meta-analysis of observational studies of first-trimester artemisinin exposure found no increase in miscarriage or major malformation compared with quinine, and a later individual-patient-data meta-analysis in The Lancet comparing first-trimester artemisinin derivatives with non-artemisinin antimalarials was similarly reassuring — which is why WHO now permits ACT in the first trimester when a pregnant woman has malaria. Untreated malaria in pregnancy is itself dangerous to both mother and fetus.

That clinical nuance does not extend to supplements. Avoid A. annua herb, tea and artemisinin capsules entirely in pregnancy and while trying to conceive. There is no benefit to weigh against the risk, the dose is unknown, and Artemisia species also carry a traditional emmenagogue reputation. Breastfeeding: not adequately studied; avoid.

Liver

Published case reports link artemisinin-containing products to cholestatic liver injury: acute cholestatic hepatitis after A. annua tea; artemisinin-induced cholestatic injury with intrahepatic ductopenia (loss of small bile ducts, a serious pattern); and a case assessed by updated RUCAM as immune-mediated herb-induced liver injury associated with herbal artemisinin. A separate report documented hepatotoxicity in a glioblastoma patient taking artesunate with temozolomide and Chinese herbs.

Stop and seek care for jaundice, dark urine, pale stools, persistent itching, right-upper-quadrant pain, or unusual fatigue with nausea. If you use artemisinin for more than a few weeks, get liver enzymes checked. Avoid entirely with existing liver disease.

Allergy

Artemisia is one of the most important weed-pollen allergen genera in the northern hemisphere, with well-documented cross-reactivity between mugwort, sage and other Artemisia species, and between mugwort and ragweed at the molecular level. If you have mugwort or ragweed hay fever, or react to chamomile, echinacea, feverfew or chrysanthemum, treat sweet wormwood as likely to cross-react. Reactions can include rash, rhinitis, asthma exacerbation and, rarely, anaphylaxis.

Other

• Gastrointestinal upset — nausea, abdominal pain and diarrhoea are the commonest complaints.
• Post-artesunate delayed haemolysis — relevant to people treated with intravenous artesunate for severe malaria: a fall in haemoglobin one to three weeks afterwards, usually self-limiting but worth a follow-up blood count.
• Children — no established supplement dose. Antimalarial dosing is weight-based and prescribed.
• Never use as malaria prophylaxis or self-treatment. Travellers who did have developed imported falciparum malaria. See the malaria article.

Cost and Sourcing

Sweet wormwood is one of the cheapest medicinal plants there is — it is a vigorous annual weed that grows easily in temperate gardens, and dried herb sells for a few dollars an ounce. “Artemisinin” supplements cost far more, typically tens of dollars for a month, and the price reflects extraction rather than any verification of what you receive.

Two sourcing notes worth knowing:

• A. annua is a biomass accumulator. Fast-growing leafy crops can concentrate heavy metals and pesticide residues from soil. Third-party contaminant testing matters more here than for a root harvested from a controlled plot.
• Species substitution is common. “Wormwood” on a label may be A. absinthium (thujone, no artemisinin) or A. afra (negligible artemisinin). Insist on the binomial.

Growing your own is easy and makes a fine aromatic plant for dried arrangements — “sweet Annie” is a craft-trade staple for its scent. It does not make you a source of medicine, for the reasons this article has spent several thousand words on.

The Bottom Line

Artemisia annua gave the world one of its most valuable medicines, and the story of how — a fourth-century instruction to use cold water, read carefully by a chemist in 1971 — is as good as science history gets. Honour that by being clear about what the plant is and is not.

• The molecule, purified, dosed and paired with a partner drug, is a frontline treatment. Millions of people are alive because of it.
• The leaf delivers roughly a fifth of a daily antimalarial dose per litre of tea, from a starting material that varies severalfold, with no way to check.
• Under-dosed artemisinin is not merely ineffective — it is the specific condition that breeds resistance, and that cost is paid by other people.
• Outside malaria, human evidence is one small osteoarthritis pilot, one negative cancer pilot, and a phase I study. Everything else is cells and mice.
• If you might have malaria, get tested today and take the full ACT course. That is the whole of the practical advice, and it is the part that saves lives.

Key Research Papers

Every PMID was verified live against NCBI E-utilities before publication — first author, title, journal and year all had to match.

How much is in the plant, and what a tea delivers

1. Räth K, Taxis K, Walz G, Gleiter CH, Li SM, Heide L. Pharmacokinetic study of artemisinin after oral intake of a traditional preparation of Artemisia annua L. (annual wormwood). American Journal of Tropical Medicine and Hygiene. 2004;70(2):128–132. Human. The measured number: 94.5 mg per litre from 9 g of leaf.
2. Delabays N, Simonnet X, Gaudin M. The genetics of artemisinin content in Artemisia annua L. and the breeding of high yielding cultivars. Current Medicinal Chemistry. 2001;8(15):1795–1801.
3. Engeu PO, Omujal F, Agwaya M, et al. Variations in antimalarial components of Artemisia annua Linn from three regions of Uganda. African Health Sciences. 2015;15(3):828–834.
4. Carbonara T, Pascale R, Argentieri MP, et al. Phytochemical analysis of a herbal tea from Artemisia annua L. Journal of Pharmaceutical and Biomedical Analysis. 2012;62:79–86.
5. Mouton J, Jansen O, Frédérich M, van der Kooy F. Is artemisinin the only antiplasmodial compound in the Artemisia annua tea infusion? An in vitro study. Planta Medica. 2013;79(6):468–470.
6. Wright CW, Linley PA, Brun R, Wittlin S, Hsu E. Ancient Chinese methods are remarkably effective for the preparation of artemisinin-rich extracts of qing hao with potent antimalarial activity. Molecules. 2010;15(2):804–812. Tests Ge Hong’s cold-water method directly.

Whole plant versus purified drug

1. Elfawal MA, Towler MJ, Reich NG, Golenbock D, Weathers PJ, Rich SM. Dried whole plant Artemisia annua as an antimalarial therapy. PLoS One. 2012;7(12):e52746. Rodent.
2. Elfawal MA, Towler MJ, Reich NG, Weathers PJ, Rich SM. Dried whole-plant Artemisia annua slows evolution of malaria drug resistance and overcomes resistance to artemisinin. Proceedings of the National Academy of Sciences. 2015;112(3):821–826. Rodent.
3. Desrosiers MR, Mittelman A, Weathers PJ. Dried leaf Artemisia annua improves bioavailability of artemisinin via cytochrome P450 inhibition and enhances artemisinin efficacy downstream. Biomolecules. 2020;10(2).
4. Weathers PJ. Artemisinin as a therapeutic vs. its more complex Artemisia source material. Natural Product Reports. 2023;40(7):1158–1169.
5. Kane NF, Kyaw NN, Monera-Girona AJ, et al. Artemisia extracts differ from artemisinin effects on human hepatic CYP450s 2B6 and 3A4 in vitro. Journal of Ethnopharmacology. 2022;298:115587.
6. Czechowski T, Larson TR, Catania TM, et al. Editorial: artemisinin — from traditional Chinese medicine to artemisinin combination therapies; four decades of research on the biochemistry, physiology, and breeding of Artemisia annua. Frontiers in Plant Science. 2020;11:594565.

Pharmacokinetics and drug interactions

1. Svensson US, Ashton M. Identification of the human cytochrome P450 enzymes involved in the in vitro metabolism of artemisinin. British Journal of Clinical Pharmacology. 1999;48(4):528–535.
2. Simonsson US, Jansson B, Hai TN, Huong DX, Tybring G, Ashton M. Artemisinin autoinduction is caused by involvement of cytochrome P450 2B6 but not 2C9. Clinical Pharmacology and Therapeutics. 2003;74(1):32–43. Human.
3. Xing J, Kirby BJ, Whittington D, et al. Evaluation of P450 inhibition and induction by artemisinin antimalarials in human liver microsomes and primary human hepatocytes. Drug Metabolism and Disposition. 2012;40(9):1757–1764.
4. Navaratnam V, Mansor SM, Sit NW, Grace J, Li Q, Olliaro P. Pharmacokinetics of artemisinin-type compounds. Clinical Pharmacokinetics. 2000;39(4):255–270.
5. Gautam A, Ahmed T, Batra V, Paliwal J. Pharmacokinetics and pharmacodynamics of endoperoxide antimalarials. Current Drug Metabolism. 2009;10(3):289–306.

Safety

1. Brewer TG, Peggins JO, Grate SJ, et al. Neurotoxicity in animals due to arteether and artemether. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1994;88 Suppl 1:S33–S36. Animal.
2. Brewer TG, Grate SJ, Peggins JO, et al. Fatal neurotoxicity of arteether and artemether. American Journal of Tropical Medicine and Hygiene. 1994;51(3):251–259. Animal.
3. Toovey S. Are currently deployed artemisinins neurotoxic? Toxicology Letters. 2006;166(2):95–104. The human counterweight.
4. Van Vugt M, Angus BJ, Price RN, et al. A case-control auditory evaluation of patients treated with artemisinin derivatives for multidrug-resistant Plasmodium falciparum malaria. American Journal of Tropical Medicine and Hygiene. 2000;62(1):65–69. Human.
5. Hien TT, Turner GD, Mai NT, et al. Neuropathological assessment of artemether-treated severe malaria. The Lancet. 2003;362(9380):295–296. Human.
6. Clark RL. Embryotoxicity of the artemisinin antimalarials and potential consequences for use in women in the first trimester. Reproductive Toxicology. 2009;28(3):285–296. Animal.
7. Dellicour S, Sevene E, McGready R, et al. First-trimester artemisinin derivatives and quinine treatments and the risk of adverse pregnancy outcomes in Africa and Asia: a meta-analysis of observational studies. PLoS Medicine. 2017;14(5):e1002290. Human, observational.
8. Saito M, Mansoor R, Wilairisak K, et al. Pregnancy outcomes after first-trimester treatment with artemisinin derivatives versus non-artemisinin antimalarials: a systematic review and individual patient data meta-analysis. The Lancet. 2023;401(10371):118–130. Human.
9. Ruperti-Repilado FJ, Haefliger S, Rehm S, et al. Danger of herbal tea: a case of acute cholestatic hepatitis due to Artemisia annua tea. Frontiers in Medicine. 2019;6:221. Human case report.
10. Thio J, Rahman A, Cheah D, et al. Artemisinin-induced cholestatic liver injury and intrahepatic ductopenia. Oxford Medical Case Reports. 2024;2024(7):omae070.
11. Mathavan A, Mathavan A, Reddy R, Jones K, Cavanagh Y. Immune-mediated herb-induced liver injury: a potential association with herbal artemisinin use as supported by the updated RUCAM. BMJ Case Reports. 2023;16(5).
12. Katial RK, Lin FL, Stafford WW, Ledoux RA, Westley CR, Weber RW. Mugwort and sage (Artemisia) pollen cross-reactivity: ELISA inhibition and immunoblot evaluation. Annals of Allergy, Asthma & Immunology. 1997;79(4):340–346.
13. Jahn-Schmid B, Hauser M, Wopfner N, et al. Humoral and cellular cross-reactivity between Amb a 1, the major ragweed pollen allergen, and its mugwort homolog Art v 6. Journal of Immunology. 2012;188(3):1559–1567.
14. Jaita S, Madsalae K, Charoensakulchai S, et al. Post-artesunate delayed hemolysis: a review of current evidence. Tropical Medicine and Infectious Disease. 2023;8(1).

Live PubMed Searches

1. Artemisinin content variation
2. Artemisia annua tea pharmacokinetics
3. Dried-leaf Artemisia bioavailability
4. Artemisinin and CYP enzymes
5. Artemether and arteether neurotoxicity
6. Artemisinin in the first trimester
7. Artemisia annua and cholestatic hepatitis
8. Artemisia allergy cross-reactivity
9. Derivative pharmacokinetics

Connections

• All Herbs
• Artemisia annua Benefits — the hub
• Artemisia annua (Sweet Wormwood)
• Artemisinin and Malaria — the discovery, ACT, and resistance
• Antiparasitic Uses
• Inflammation and Immune
• Malaria — what proper diagnosis and treatment involve
• Wormwood (Artemisia absinthium) — thujone, not artemisinin; a different safety profile entirely
• Mugwort (Artemisia vulgaris) — the cross-reacting allergen
• Andrographis

Back to Table of Contents