Ashwagandha in South Africa: evidence, SAHPRA wording and prices

Dis-Chem and Clicks both sell ashwagandha. The Dis-Chem Gold page puts the product in Category D, Western herbal medicine, and prints the line that matters: this unregistered medicine has not been evaluated by SAHPRA for its quality, safety or intended use. Clicks marketing for GNC still talks about stress. Low-risk complementary-medicine wording can cover general health and minor symptoms. It does not cover treating an anxiety disorder, and 'clinically proven' is the sort of phrase SAHPRA's own roadmap treats as high risk. The page does not repeat it.

The regulator's frame

SAHPRA's complementary-medicine page explains that Category D covers discipline-specific medicines and health supplements, and that sale can continue while call-up and registration catch up, if the indication stays low risk and the other licence and labelling conditions are met. The Advertising Regulatory Board watches the advert. HPCSA-registered practitioners have their own conduct rules and are not a shop. A price on Clicks is not a SAHPRA registration.

What human trials suggest, and what is still early

A 2022 meta-analysis of randomised trials (PubMed 36017529) pooled 12 papers and 1,002 adults, ages roughly 25 to 48. Anxiety scores fell more than placebo, standardised mean difference −1.55, and stress scores fell, standardised mean difference −1.75. The heterogeneity was very high (I² about 94% for anxiety and 83% for stress), and the authors graded the certainty as low. A large average effect with low certainty means the trials point the same general way and are not yet solid enough to treat as settled medicine. The same paper's dose-response note associated stress changes with about 300–600 mg a day. One upper figure in that abstract, for anxiety, is far above ordinary supplement servings. That figure is the authors' model, not an amount to take.

A 2021 sleep meta-analysis (PubMed 34559859) pooled five randomised trials and 400 adults. The overall sleep effect was a standardised mean difference of −0.59 (interval −0.75 to −0.42). The effect looked larger in people with insomnia, at 600 mg a day or more, and at eight weeks or more. Four of the five trials used KSM-66 root extract. Doses in the underlying trials included 120 mg a day of a root-and-leaf extract, 250 mg and 600 mg a day, and 600 mg a day in three trials. Durations were six to twelve weeks. No serious side effects were reported in that review, and the authors said serious-harm data were too thin to call long-term use safe. That last sentence matters more than the average sleep score.

A 2021 narrative review on PubMed Central (PMC8762185) tabulates individual KSM-66 trials. Several used 300 mg twice a day, which is 600 mg a day, for eight to twelve weeks, in stressed adults or in people with insomnia. Those tables are why 600 mg of root extract appears so often on labels. Appearing in a trial is not proof that every 600 mg capsule matches the extract, the withanolide content, or the people who were studied. Sensoril, used in the Doctor's Best US product, is a root-and-leaf extract. The sleep meta-analysis's 120 mg trial used a root-and-leaf extract as well. Root-only and root-and-leaf are not interchangeable just because both say ashwagandha.

What looks repeated is a short-term change in stress and sleep questionnaires, often with KSM-66 around 600 mg a day for about two months, in adults. What is not established is a cure for an anxiety disorder, a thyroid treatment, a fertility treatment, or long-term safety. US structure/function wording can be lawful and still outrun the certainty of the evidence.

Amounts used in studies are not a dose for you

The figures above, 120 mg, 250 mg, 300 mg twice daily and 600 mg a day, are what investigators put in capsules. They are not advice. A label that says 600 mg per capsule, one a day, matches the total daily amount of several trials. A label that says 600 mg and tells you to take two capsules may be twice that, or the 600 mg may already be the two-capsule serving. Where the page did not separate those, the table says so.

Safety: liver, thyroid, pregnancy, and a ban that is not universal

Denmark's food authority says ashwagandha may not be sold as a food, including supplements. The Danish Veterinary and Food Administration page cites the Technical University of Denmark. The 15 May 2020 DTU assessment looked at root and root extracts and said it could not set an intake below which harm was unlikely. The concerns it listed were effects on sex hormones and reproduction in animal studies, possible effects on thyroid hormones, and effects on metabolism, the immune system and the nervous system. DTU also summarised a case series, Björnsson and colleagues 2020, of liver injury after ashwagandha supplements in Iceland and the United States. That is a case series, not a trial that measures how often the injury happens. Denmark's ban is Danish law. It is not a UK, US, UAE or South African ban. It is the reason a serious page cannot call the herb risk-free.

Australia's Therapeutic Goods Administration published a safety advisory on 22 February 2024. It describes sudden gastrointestinal symptoms, sometimes after one or two doses, and a very rare risk of liver injury. Up to 5 February 2024 the TGA had 12 reports of liver problems, of which seven had enough detail to suggest liver injury, and four people were in hospital. Most people recovered after they stopped. The advisory says people with a current or past liver problem should avoid the herb, and it lists warning symptoms: yellow skin or eyes, dark urine, nausea, vomiting, unusual tiredness, weakness, abdominal pain, and loss of appetite. The figures below are from that 22 February 2024 advisory.

Thyroid effects are why several labels say to ask a doctor if you take thyroid hormone. DTU judged that animal and some human observations pointed that way, without a dose it considered safe. The practical point is on the NOW US label: do not use it if you are pregnant or nursing, and ask a physician if you take thyroid medicine or have a thyroid, liver or gastrointestinal condition.

Pregnancy and breastfeeding are a reason to stay away, not a lifestyle choice the trials cleared. DTU's reproductive concerns come largely from animal work, and the authority's public page states them as the basis of the Danish ban. Ancient+Brave, Igennus and Holland & Barrett labels tell pregnant and breastfeeding people not to use the product or to ask a doctor first. That is label text.

Gold's capsule is 200 mg of a 15:1 root extract, described as 3,000 mg root equivalent, one or two capsules a day. Two capsules would be 400 mg of extract, under the 600 mg a day used in several KSM-66 trials, and the extract is not identified as KSM-66. GNC's 60-capsule page says 600 mg in the name and says take two capsules daily, without a facts panel that splits the 600 mg across one capsule or two. The page does not split that 600 mg across one capsule or two. The 30-tablet GNC page also says 600 mg and does not give a direction. Pregnancy and liver disease are reasons not to start this because of the DTU assessment and the TGA advisory, not because South Africa has copied Denmark's ban. Call 10177 for an ambulance, or 112 from a mobile.