The evidence · Aspalathus linearis

Side effects, and how weak each signal is

Rooibos has been drunk in enormous quantities for a century and exported to fifty-odd markets. Against that, the entire documented safety record in humans is two case reports. This page is each signal that exists, what it actually consists of, and why almost none of it should change what you do.

The record
Human liver case reports
two, worldwide
First
2010, Finland
Second
2016, as “red bush tea”
Review's conclusion
“seems to be safe”with caveats
Oestrogenic compounds
3 of 25 isolated
Caffeine
none
The short answer

Very little is documented, and the one thing that is — liver injury — amounts to two case reports in the whole medical literature.

  • Liver Two published human cases, ever. A 2023 review still concludes rooibos seems safe on this
  • Medicines That review's real warning: rooibos may affect the enzymes that clear some drugs. Ask a pharmacist
  • Hormones Three of 25 compounds isolated from the leaf showed oestrogenic activity in a lab assay
  • Contamination The review flags the plant material itself, not the plant, as a route to harm
  • Iron Unlike ordinary tea, rooibos did not block iron absorption in the one study that tested it
  • Caffeine None, so no caffeine side effects. That much is settled

Rooibos and medicines · How to read a case report

I — The record

Everything that is actually documented

Liver injury

Strength: two case reports, worldwide, ever

In 2010 the European Journal of Clinical Pharmacology published a short report titled “Possible hepatotoxic effect of rooibos tea: a case report”, from a Finnish group. In 2016 the Journal of Clinical Anesthesia published a second, in which raised liver enzymes and low platelets were found during pre-operative assessment of a patient who drank red bush tea, complicating the work-up for a suspected appendicitis.

A 2023 review in the International Journal of Environmental Health Research gathered the preclinical and human safety data and concluded that rooibos consumption seems to be safe in terms of hepatotoxicity, while noting that there may be a subgroup of consumers at higher risk of liver irritation. In rodents, it adds, rooibos is protective or neutral for the liver.

The same review makes a point worth repeating: contamination of the plant material may contribute to herb-induced liver injury. That is a claim about supply chains rather than about the species.

Interference with medicines

Strength: a mechanism, and a reviewer's warning

Rooibos can modulate cytochrome P450 enzymes — the liver machinery that clears a great many prescription drugs. The 2023 review's closing recommendation is that caution should be exercised in patients being treated with allopathic medicines, to avoid unintended changes in drug plasma concentration.

Set against that, a 2024 study testing fermented and unfermented rooibos extracts for CYP inhibition found no clinically significant interaction potential at the concentrations it used.

Those two are not in contradiction. One is a review flagging a plausible mechanism and asking for caution; the other is one experiment at particular concentrations finding nothing. Neither settles it.

Oestrogenic activity

Strength: a laboratory assay, no human data

A 2006 study at the University of Shizuoka isolated twenty-five compounds from rooibos leaves and tested each for oestrogenic activity. Three showed it. The paper is titled, plainly, Phytoestrogens from Aspalathus linearis.

Weakly oestrogenic plant compounds are present in many ordinary foods and their presence is not a finding of harm. It is, though, why this site does not say that rooibos is inert. If you are managing a hormone-sensitive condition, this is a real question for your doctor rather than a website, and it is not one the internet's rooibos pages mention at all.

Adrenal and ovarian effects in cells

Strength: cells in dishes

Rooibos and two of its signature compounds cut total steroid output roughly fourfold in a human adrenal cell line under stimulated conditions. Separately, rooibos extract suppressed proliferation in pig ovarian cells and blunted their response to follicle-stimulating hormone; the authors wrote that rooibos's potential anti-reproductive effects should be taken into account in animal and human nutrition.

Cells, in dishes, at concentrations chosen by researchers. This is the weakest kind of evidence for a human effect and the strongest reason to stop calling rooibos biologically inactive. The same evidence, in the pregnancy context.

Kidney stones

Strength: tested in 8 people, nothing found

Because rooibos is a plant infusion, people ask about oxalate and stones. It has been tested, once: eight calcium-oxalate stone formers drank rooibos for thirty days, and none of the urinary risk factors measured changed significantly in either direction.

That is neither a warning nor a reassurance. Eight people is eight people, and the researchers said so themselves.

Case reports: Sinisalo M, Enkovaara AL, Kivistö KT, Eur J Clin Pharmacol 2010;66(4):427–8; Reddy S et al., J Clin Anesth 2016;35:96–98. Review: Pyrzanowska J, Int J Environ Health Res 2023;33(2):129–142. CYP study: Kok AM et al., BMC Complement Med Ther 2024;24(1):152. Phytoestrogens: Shimamura N et al., Biol Pharm Bull 2006;29(6):1271–4. Adrenal cells: Schloms L et al., J Steroid Biochem Mol Biol 2012;128:128–38. Ovarian cells: Sirotkin AV et al., Physiol Res 2023;72(2):269–280. Stones: Rodgers A et al., Urolithiasis 2016;44(4):299–310.
II — The reading

How much should two case reports worry you

Almost not at all — and it is worth being clear about why, because the honest answer here runs against this site's usual direction.

A case report is one doctor writing up one patient. It cannot establish that a thing caused an outcome; it can only say that the two occurred together and the doctor thought it worth publishing. It has no control group and no denominator.

That last point is the one that matters. Rooibos is drunk by many millions of people. Two published cases against a denominator that large is close to the background rate at which anything gets written up. If rooibos routinely damaged livers, the literature would not consist of two short papers sixteen years apart.

The right conclusion is the review's own: it seems to be safe in this respect, and a small number of people may respond differently, and nobody can currently tell you who they are.

So why publish it at all? Because “rooibos has no side effects” is written everywhere and it is not quite true, and a reader who later stumbles on a case report they were never told about has been badly served by every page they read first.

What a case report is not
  • Not a trial
    no control group
  • Not a rate
    no denominator
  • Not causation
    only co-occurrence
  • Not nothing
    it is how signals are first spotted
III — The practical one

Medicines are the question worth actually asking

Of everything on this page, the interaction question is the one with a plausible route to mattering in an ordinary life, and it is the one people search least. Can you drink rooibos tea on blood thinners is a real query; so is the one about drinking it before a fasting blood test.

We are not going to answer either, because the answer depends on the drug and we are not pharmacists. What we can tell you is the shape of the concern: rooibos appears able to influence the liver enzymes that clear certain drugs, which would change how much of the drug is in your blood rather than cause any effect of its own. That is exactly the sort of thing a pharmacist answers in one minute and a website answers badly.

Two practical notes that are not advice:

  • A pharmacist is free and does not need an appointment. Bring the medicine and say how much rooibos you drink.
  • Say “rooibos” and also “redbush”. One of the two published liver cases is indexed under red bush tea, and the same split runs through medical records. The plant has two names in English and a clinician may only know one.
Interaction mechanism and the caution as stated by Pyrzanowska 2023, cited above. The naming point is our observation from the titles of the two case reports, not a claim from either paper.
IV — The blanks

Side effects people search for, with nothing behind them

Each of these is a real, suggested search. In each case we looked and found no human evidence either way.

Searched forWhat we found
Rooibos causes anxiety or insomniaNo study. Rooibos contains no caffeine, which is the usual cause in a tea
Rooibos raises oestrogen in menNo human study. Three compounds showed oestrogenic activity in a laboratory assay; nobody has measured a person
Rooibos affects testosteroneNo human study. A rat paper reports a change in a corticosterone-to-testosterone ratio
Rooibos and thyroidNothing found
Rooibos stains teethNothing found. It is an unstudied question, not a settled one
Too much rooibos is harmfulNo dose-response study in humans. The largest trial had people drink six cups a day for six weeks
Rooibos on an empty stomach, or with refluxNothing found
One thing to be careful with here

This table is a list of things nobody has studied. It is not a clean bill of health, and it is not a list of risks either. Where a row says nothing was found, that is a statement about the published literature on 18 September 2026, and nothing more.

Searches run against PubMed on 18 September 2026, each crossing rooibos with the outcome named. Rat corticosterone-to-testosterone ratio: Schloms L et al., Mol Nutr Food Res 2014;58(3):537–49. Six cups for six weeks: Marnewick JL et al., J Ethnopharmacol 2011;133(1):46–52.
The benefits, testedAll eight human studies, with their sample sizes and the three that found nothing. Rooibos and pregnancySix papers exist and none is a trial. What the NHS actually says. All the evidence pagesWhat has actually been tested in humans, and what has not.