Spices and health: what the evidence actually supports

Almost every headline about a spice and health rests on a study that did not use the spice. It used a solvent extract, or a purified compound, or a compound formulated with an absorption enhancer, at a quantity chosen to produce a measurable effect. The distance between that and a teaspoon of something in a pan is not a detail; it is usually the entire finding.

Curcumin is the clearest case. It is poorly soluble in water, unstable, badly absorbed, and extensively metabolised before it reaches the bloodstream — which is why the supplement industry sells it with piperine or in lipid formulations, and why the trials do too. A curry delivers a small fraction of the curcuminoids in one capsule and almost none of that fraction gets into circulation. Reporting a curcumin trial as a fact about turmeric is not a simplification. It is a different claim.

SpiceHQ records the exposure on every health assessment for exactly this reason, and states separately whether the finding transfers to eating the spice. Most of the time the honest answer is that it does not, or that nobody has checked.

Search any spice with "benefits" and you will find thousands of papers. Volume of publication reads as weight of evidence and is very nearly the opposite here: the reason there are thousands of curcumin papers is that the results keep failing to replicate, and a finding that replicated would have stopped the field rather than fuelled it.

Three specific problems recur. Publication bias means trials that found nothing are less likely to appear. Many trials are small, short, and conducted by groups with an interest in the extract being tested. And a great deal of what circulates is not clinical at all — it is cell culture or animal work, where a compound at a concentration no human tissue reaches does something interesting, and the press release does the rest.

The distinction SpiceHQ draws hardest is between "this has been shown in cells" and "this has been shown in people". Both are real research. Only one of them is about you.

Ginger for nausea in pregnancy is the strongest positive finding about any spice. Randomised trials consistently show it reduces the severity of nausea against placebo, and the US National Institutes of Health summarise the research as supportive. The effect on vomiting frequency is far less clear, and the two are habitually reported together in a way that overstates the second.

Topical capsaicin for pain is the best-evidenced result in the whole field, and it concerns skin rather than food. A Cochrane review of fourteen randomised trials covering around 2,050 participants found topical chilli preparations reduced pain more than placebo, with the strongest evidence in low back pain. Eating cayenne is a completely separate question with no comparable evidence behind it.

Garlic supplements produce small reductions in total and LDL cholesterol in people who already have raised cholesterol, and possibly a small reduction in blood pressure in people who already have hypertension. Both qualifications matter, and neither has been shown for garlic in food.

Cassia cinnamon is a real dietary source of coumarin, for which regulators have set a tolerable daily intake. That is a genuine consideration for someone eating cinnamon heavily every day and not for anyone else.

A result showing that something does not work is a result, and in this field it is often the more valuable one, because the claim it refutes is the one being marketed.

Fenugreek is recommended for milk supply more widely than any other spice, and a meta-analysis of randomised trials found that fenugreek seed capsules did not increase milk volume. Several of the studies usually cited in its favour lack randomisation, blinding or a placebo control, which is enough on its own to explain a positive result.

Ginger does not appear to help with motion sickness, despite helping with nausea in pregnancy — a useful reminder that nausea is not one condition. Garlic does not appear to reduce stomach cancer risk. Star anise does not treat influenza: it is an industrial feedstock for the synthesis of an antiviral drug, which is a fact about a chemical plant rather than about a cup of tea.

SpiceHQ records these at the same evidence level as positive findings, and its health section colours them the same way, because strength of evidence and desirability of result are different things.

The best-established health effect of any spice product is an adverse one. The US LiverTox database classes turmeric as a well-documented cause of clinically apparent liver injury — its highest likelihood category — with a hepatocellular pattern appearing one to four months after starting a product and a strong association with the HLA-B*35:01 immune variant. The cases cluster in high-absorption formulations, including those combining curcumin with piperine, so the property being advertised as a benefit is the one concentrating the risk.

The other genuine risks in this domain are supply-chain rather than pharmacological. Lead chromate adulteration of turmeric has been documented in specific supply chains and lead has no safe exposure threshold. Star anise adulterated with the toxic Japanese species has caused illness, and infusions are the preparation in which it happened.

And there is one class of question SpiceHQ can now answer that has nothing to do with people: garlic and onion are seriously toxic to dogs and cats, and the dried and powdered forms are the most concentrated. A pet fed seasoned leftovers is a more likely exposure than one that ate a raw onion.

Are spices good for you? Almost certainly yes, and for a reason that has nothing to do with any of the above. Spices are how a bowl of lentils becomes something you want to eat twice a week. They are how a plate of vegetables stops being a duty. They carry flavour into food that would otherwise need salt, sugar or fat to be worth eating, and they cost almost nothing per portion.

That is a claim about diets rather than about compounds, and it is the one that survives inspection. Nobody has run a trial on it because nobody needs to.

What SpiceHQ will not do is dress that up as pharmacology. If a spice does something specific and a competent source has assessed it, the profile says so with the evidence level and the exposure attached. If it does not, the profile says that too. There is no third option in which a page hints.