Is herbal medicine safe?

It depends. "Herbal" does not mean "safe" — some of the most toxic substances known are plant-derived (hemlock; ricin from castor beans; pyrrolizidine alkaloids). The key factors for safety: regulation and quality control (EU-regulated herbal medicines are considerably safer than unregulated products from uncontrolled sources); correct identification and standardisation; awareness of drug-herb interactions; appropriate doses; avoiding herbs in contraindicated situations (pregnancy, specific medical conditions). The most important safety principle: always tell your doctor or pharmacist what herbal products you take — particularly if you are on warfarin, the contraceptive pill, HIV drugs, ciclosporin, or any medication metabolised by CYP3A4.

Does St John's Wort work for depression?

Yes — for mild to moderate depression specifically. The Cochrane review (2008, updated 2009) of 27 randomised trials including 3,000+ patients found St John's Wort was: significantly more effective than placebo for mild-moderate depression; as effective as standard antidepressants (TCAs and SSRIs); better tolerated than TCAs; with fewer side effects than SSRIs. The evidence does NOT support its use for severe depression. Critical caveat: St John's Wort is a potent CYP3A4 and CYP2C9 inducer and P-glycoprotein inducer — it reduces blood levels of the contraceptive pill, HIV antiretrovirals, warfarin, ciclosporin, SSRIs and many other drugs. Patients must disclose use to their prescribers.

What about echinacea for colds?

The evidence is disappointing. Cochrane 2015 review (24 RCTs): some Echinacea preparations may reduce cold incidence by approximately 10-20% and duration by approximately half a day — but the evidence is inconsistent across different preparations (different species: E. purpurea, E. angustifolia; different parts: root vs aerial; different preparations: juice, extract, tincture). Some preparations showed no benefit vs placebo. The inconsistency makes evidence-based recommendations difficult. NICE and most national formularies do not recommend echinacea. The immune-stimulating mechanism is plausible but not definitively demonstrated.

What herbal products should I avoid?

The most important ones to avoid or use with extreme caution: Kava — anxiolytic evidence exists, but associated with severe hepatotoxicity; banned in multiple countries. Comfrey — pyrrolizidine alkaloids cause hepatic veno-occlusive disease; do not take internally. High-dose green tea extract — multiple hepatotoxicity cases reported. Pennyroyal — abortifacient at high doses; hepatotoxic. Herbal products containing ephedrine (ma huang) — cardiovascular risk; banned in most countries. Any product you cannot verify as standardised and from a regulated manufacturer should be treated with caution.

How is herbal medicine regulated?

Regulation varies enormously by country: EU: Directive 2004/24/EC requires Traditional Herbal Medicinal Products to demonstrate quality (standardised manufacturing) and safety (30+ years traditional use, 15 years EU use), but NOT efficacy. Products carry the THR (Traditional Herbal Registration) mark. UK: MHRA Traditional Herbal Registration — similar to EU standards post-Brexit. USA: DSHEA 1994 classified herbal supplements as food — no pre-market approval needed; FDA must prove harm after marketing. This is the weakest regulatory framework. Canada: Natural Health Products Directorate requires a Natural Product Number (NPN) — intermediate between EU and US standards. Most LMICs: unregulated — significant quality control concerns.

How does GMJ News cover herbal medicine?

GMJ News covers evidence by herb (Grade A-D), St John's Wort evidence and drug interaction warning, EU/US regulatory frameworks, the DSHEA regulatory gap, drug-herb interactions (CYP450), hepatotoxic herbs, quality control and adulteration, and the pharmacological origins of modern drugs from plants. GMJ position: herbal medicine contains both genuinely effective remedies and unsupported products — the evidence must guide clinical decisions, and drug-herb interactions are the most critical patient safety issue.

Home › Topics › Herbal Medicine

Herbal Medicine

GMJ News knowledge hub · last reviewed September 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

Herbal medicine — the use of plant-derived preparations for therapeutic purposes — underpins virtually all of modern pharmacology (digitalis, aspirin, morphine, artemisinin, metformin and penicillin all originate from natural product traditions), yet today’s commercial herbal market combines genuinely evidence-based remedies with poorly studied, misbranded and occasionally dangerous products under the EU Traditional Herbal Medicinal Products Directive 2004/24/EC, which requires quality and safety data but NOT proof of efficacy — a dramatically lower bar than pharmaceutical regulation (WHO). The most critical clinical fact: St John’s Wort (Hypericum perforatum) — Grade A evidence for mild-moderate depression — is also one of the most dangerous drug interaction risks in all of medicine, inducing CYP3A4 and P-glycoprotein to reduce serum levels of the contraceptive pill, HIV antiretrovirals, warfarin, ciclosporin and SSRIs — a drug interaction that has caused treatment failures, HIV resistance mutations and unwanted pregnancies.

Key messages

Modern pharmacology began with plants — aspirin, morphine, artemisinin

Digitalis (foxglove → digoxin); aspirin (willow bark → salicylic acid); morphine (opium poppy); artemisinin (Artemisia annua → 2015 Nobel Prize); metformin (Galega officinalis); penicillin (Penicillium mould). The boundaries between herbal medicine and conventional pharmacology are artificial — the difference is isolation, standardisation and clinical trial evidence.

St John's Wort — Grade A evidence for depression, but dangerous interactions

St John's Wort (Hypericum perforatum): Cochrane 2008 (27 RCTs, 3,000+ patients) — comparable to standard antidepressants for mild-to-moderate depression; better tolerated than TCAs; similar to SSRIs. BUT: a potent inducer of CYP3A4, CYP2C9 and P-glycoprotein — reduces serum levels of the contraceptive pill (unintended pregnancies); HIV antiretrovirals (treatment failure, drug resistance); warfarin (thrombosis); ciclosporin (organ rejection); SSRIs (serotonin syndrome); digoxin. Must be disclosed to all patients on these medications.

EU Directive 2004/24 — quality without efficacy proof

The EU Traditional Herbal Medicinal Products Directive requires: demonstration of quality (standardised manufacturing, content); safety data from 30+ years of traditional use (15 years EU); but does NOT require proof of efficacy from clinical trials. This creates a lower regulatory bar than pharmaceuticals — products can legally be sold in EU pharmacies without evidence they work.

Drug-herb interactions — the critical safety issue

The most important clinical issue in herbal medicine is interactions with conventional drugs: St John's Wort (CYP3A4 inducer); ginkgo, garlic, ginger (antiplatelet — bleeding risk with anticoagulants); valerian, kava (CNS depression with sedatives/alcohol); licorice root (potassium loss, hypertension — interacts with antihypertensives, diuretics); echinacea (theoretical immune suppression — avoid in autoimmune disease or immunosuppressive drugs). Always ask patients about herbal use alongside conventional medications.

Quality control — the unregulated market

In unregulated markets (USA — DSHEA 1994; many LMICs), herbal products require no pre-market approval, no proof of content or purity. Common problems: products containing less (or none) of the stated herb; products containing unlisted pharmaceutical drugs (sildenafil in "herbal" erectile products; prednisolone in "natural" anti-inflammatory); heavy metal contamination (particularly in imported Ayurvedic and TCM products); species substitution (wrong plant in the bottle); microbial contamination.

Evidence by herb — what works

Grade A: St John's Wort (mild-moderate depression — with interaction warning). Grade B: valerian (sleep quality — modest, inconsistent); ginkgo (dementia symptoms — marginal benefit); Boswellia (OA pain). Grade C: echinacea (cold prevention/duration — inconsistent); saw palmetto (BPH — Cochrane 2012: not effective vs placebo). Insufficient/negative: Aloe vera (most claims); evening primrose (eczema — Cochrane 2013: not effective); kava — anxiolytic evidence but hepatotoxicity risk.

Key statistics

Grade A

St John's Wort for mild-moderate depression (Cochrane 2008, 27 RCTs)

Cochrane

CYP3A4 inducer

St John's Wort reduces blood levels of OCP, HIV drugs, warfarin, ciclosporin

BNF/MHRA

EU Directive 2004/24

quality + safety required; efficacy proof NOT required for EU herbal approval

EU/EMA

~80%

of global population uses herbal medicine as part of primary healthcare (WHO)

WHO

Adulteration

pharmaceutical drugs found in herbal products — sildenafil, prednisolone, etc.

FDA/MHRA warnings

2015

Nobel Prize (Tu Youyou): artemisinin from Artemisia annua — TCM to WHO medicine

Nobel Committee

Herbal medicine evidence grade by product — Cochrane/EMA assessment

Source: Cochrane/EMA. Few herbal products have Grade A evidence; most are Grade C or insufficient.

Glossary of key terms

CYP450 system — drug metabolism
PharmacologyCytochrome P450 enzymes (CYP3A4, CYP2C9, CYP2D6 etc.) metabolise approximately 70-80% of all drugs. Herbal inducers (St John's Wort) increase CYP activity → faster drug clearance → sub-therapeutic drug levels → treatment failure. Herbal inhibitors (grapefruit — not a herb, but similar mechanism; kava) decrease CYP activity → higher drug levels → toxicity. St John's Wort is the most dangerous herbal CYP inducer — it induces CYP3A4, CYP2C9, CYP1A2 and P-glycoprotein simultaneously.

Standardised extracts vs whole plant
PhytochemistryStandardised extracts contain a guaranteed minimum % of the active constituent (St John's Wort: standardised to 0.3% hypericin or 4-5% hyperforin). Whole plant preparations contain variable amounts. Clinical trial evidence was obtained with standardised extracts — non-standardised products cannot be assumed to have the same efficacy or safety. EU Directive requires standardisation; US DSHEA does not mandate standardisation.

Hepatotoxicity from herbal products
WHO/HepatologyHepatotoxicity (drug-induced liver injury, DILI) from herbal products is significantly underreported. Major herbal hepatotoxins: Kava (Piper methysticum) — severe hepatotoxicity; banned in Germany, France, Canada; available in some countries for anxiety. Pyrrolizidine alkaloids — found in comfrey, coltsfoot, borage; cause hepatic veno-occlusive disease. Green tea extract (high dose supplements) — hepatotoxicity reported. Chinese herbal formulations — multiple cases of herb-herb interaction-related hepatotoxicity. The WHO Collaborating Centre for Traditional Medicine maintains a pharmacovigilance database.

DSHEA (1994) — the regulatory gap
FDA/US lawThe US Dietary Supplement Health and Education Act (DSHEA, 1994) classified dietary supplements (including herbal products) as food — not drugs. Consequence: manufacturers do not need to prove safety or efficacy before marketing; FDA must prove harm AFTER a product is on the market (post-market surveillance only). This created the world's largest unregulated supplement market ($60+ billion). Contrast with EU Directive 2004/24, which at least requires quality and safety data (but not efficacy proof).

The entourage effect — whole plant vs isolated compound
PhytochemistryThe hypothesis that whole plant extracts may have synergistic effects from multiple active constituents — greater than any single isolated compound. St John's Wort: both hypericin AND hyperforin appear to contribute to antidepressant effect through different mechanisms. Cannabis: CBD may modulate THC effects (supporting THC:CBD combined preparations over isolated THC). The entourage effect is biologically plausible but difficult to prove rigorously — it is also used to justify unsubstantiated claims about "whole food" supplements.

Traditional use ≠ evidence
EpistemologyA critical distinction: 2,000 years of traditional use does not constitute clinical evidence of efficacy. Traditional use demonstrates: cultural persistence (people continued using it across generations); some degree of safety (if used as directed, the preparation did not kill users); plausibility of biological activity. Traditional use does NOT prove: that the treatment works for the claimed indication; that it is safe in modern drug-interaction contexts; that it is safe for all populations (e.g. pregnant women).