Tyramine

What is it?

Tyramine is a biogenic amine that builds up as food ages, ferments or is stored badly. Bacteria make it from the amino acid tyrosine, which is why it appears wherever protein has been broken down slowly: aged cheese, cured meat, fermented soy, sauerkraut. For most people it is of no consequence. For someone taking a monoamine oxidase inhibitor it can raise blood pressure sharply, and that is the whole reason this axis exists.

The gap between those two readers, in figures

EFSA's Panel on Biological Hazards (2011) gives the doses. No adverse effect at about 600 mg of tyramine for a healthy person on no medication, 50 mg for someone on a third-generation MAOI, and 6 mg on a classical one. A hundredfold difference, which is why this axis matters enormously to a few people and barely at all to everyone else. The same opinion states that the available information was insufficient for a quantitative risk assessment, so these are doses and not rates, and it publishes no figure for any single food.

The restricted list has narrowed twice, seven years apart

Shulman (1989) assayed tyramine in over a hundred disputed foods and concluded that only a very limited number require absolute restriction — all aged cheeses, concentrated yeast extracts such as Marmite, sauerkraut, and broad bean pods — with alcoholic drinks, Chianti included, appearing safe in moderation. That paper published its conclusion and none of the hundred values behind it. Gardner (1996) turned the same narrowing into a diet, and the useful part is what it permits rather than what it forbids: avoid aged cheese, aged or cured meats, any potentially spoiled meat, poultry or fish, broad bean pods, concentrated yeast extract, sauerkraut, soy sauce and soy condiments, and tap beer — while wine and bottled or canned beer are safe in moderation, and foods it does not mention are unrestricted. Those are categories, not numbers: the abstract carries no milligram figure for anything.

Why the old numbers should be read with care

McCabe-Sellers (2006) argues that the early published tyramine analyses no longer hold — analytical methods, packaging, storage and handling have all moved since — and that the original restrictions were drawn from case reports rather than from measurement, then over-extrapolated. McCabe-Sellers publishes no measurement of its own, so it can support a ranking and nothing finer. Read each level on this axis the same way: how fresh a food is, and how it was kept, moves it further than which category it belongs to.

Cheese, where the variation is the finding

Mayer (2010) measured twenty biogenic amines by UPLC in 58 cheeses bought at retail and found 22.4 per cent of them carrying more than 10 mg of tyramine per 100 g. That is a distribution across one country's shelf, not a value for any variety — two cheeses of the same name can sit at either end of it. Sadighara (2024), reviewing the literature on tyramine in cheese, puts extra-sharp cheddar at around 270 mg/kg; that figure is pooled from other studies rather than measured, so it describes a published range and not a cheese anyone can buy.

Dry fermented sausage, the one food with an exposure estimate

Suzzi (2003) explains why two salamis off the same shelf are not the same food: a fermented product is by definition one with high microbial counts, and those accumulate tyramine, phenylethylamine, tryptamine, cadaverine, putrescine and histamine — with great fluctuations reported within a single product type. That review publishes no figure. Latorre-Moratalla (2017) supplies them, from 474 dry fermented sausages and Spanish consumption data: a meal delivers on average 6.2 mg of tyramine, which is negligible for a healthy person and gives someone on an MAOI a 34 per cent probability of passing the 6 mg threshold. Those figures belong to Spanish portions and to this kind of sausage, and 474 samples make a distribution rather than a value for salami or chorizo.

Fish: the temperature decides, not the species

Křížek (2011) stored carp, rainbow trout and perch as fillets and mince at 3 °C and at 15 °C. Shelf life ran 11 to 16 days for fillets at 3 °C and 7 to 10 for mince, against 2 to 3 days at 15 °C — and samples of good or acceptable quality contained no toxicologically significant tyramine or histamine. In other words fresh fish is not the problem; the break in the cold chain is. That study covers three freshwater species and publishes shelf lives rather than amine concentrations, so it says nothing about tuna, mackerel, sardine or salmon.

Three foods where the expected answer turns out to be no

Shalaby (2000) followed broad bean, chick pea and lupine raw, through germination and through cooking, and found that on cooking the samples became free of biogenic amines — which matters because the broad bean pod is on every restricted list. Brückner (2012) extracted amines from 21 teas with boiling water and measured tyramine at 16 to 431 micrograms per litre of infusion, with no histamine detected at all; micrograms, not milligrams. It measures what is in the infusion and says nothing about DAO inhibition, which is the other reason tea appears on this axis. Pastore (2005) looked for dopamine, serotonin, tyramine, histamine and phenylethylamine in chocolate and found only the first three in the real samples — tyramine and phenylethylamine appeared only as standard additions. That is a method paper on a small set, so it unsettles the usual claim about chocolate without disproving it everywhere, and it publishes no concentration.

Why the vegetable rows sit low

Tyramine needs protein that bacteria have broken down. A vegetable offers little protein and goes through no fermentation step, so neither condition is met — which is the mechanism rather than a measured figure, and it is why these rows sit where they do.

Source: EFSA Panel on Biological Hazards, EFSA Journal 9:10 (2011) 2393, for the tyramine doses and for its own statement that a quantitative risk assessment could not be carried out; Shulman K.I. et al., J Clin Psychopharmacol 9 (1989) 397-402, for the narrowing of the restricted list — it publishes the conclusion and none of the hundred values; Gardner D.M. et al., J Clin Psychiatry 57 (1996) 99-104, for the diet verbatim, including that foods it does not mention are unrestricted — categories, not milligrams; McCabe-Sellers B.J., Staggs C.G. & Bogle M.L., J Food Compos Anal 19 (2006) S58-S65, for why the early published analyses no longer hold — it publishes no measurement of its own; Mayer H. et al., J Chromatogr A 1217 (2010) 3251-3257, for the distribution across 58 retail cheeses; Sadighara P. et al. (2024), for the pooled cheddar figure; Suzzi G. et al., Int J Food Microbiol 88 (2003) 41-54, for the fluctuation within one sausage type; Latorre-Moratalla M.L. et al., Food Chem Toxicol 99 (2017) 78-85, for the exposure per meal and the 34 per cent on an MAOI; Křížek M. et al., Acta Vet Brno 80 (2011) 365-372, for the shelf lives of three freshwater species at 3 and 15 °C; Shalaby A., Nahrung 44 (2000) 23-27, for the three legumes and the effect of cooking; Brückner H. et al., Amino Acids 42 (2012) 877-885, for the amines in 21 tea infusions; Pastore P. et al., J Chromatogr A 1098 (2005) 111-115, for what was and was not detected in chocolate

Frequently asked questions