Houttuynia as Food and in Traditional Medicine

State this plainly, because hedging it would be inaccurate: Houttuynia cordata is an ordinary food. In Vietnam it is eaten raw by the handful as diếp cá, on the herb platter that arrives with grilled pork and rice crepes. In southwestern China the white rhizome, zhé'ěrgēn, is a market vegetable dressed cold with chilli oil and black vinegar. In Japan and Korea the dried aerial parts are a common household tea — dokudami-cha, eoseongcho — sold in every supermarket in tea bags. Millions of people eat or drink this plant weekly, and have for a very long time.

That matters, and it is not a hedge or a consolation prize. Long dietary use at customary amounts is the single strongest safety datum this plant has. It is stronger than any rodent toxicology study, because the exposure is human, the duration is generational, the population is unselected, and the preparation is the one people actually use. On a site that spends most of its energy deflating overstated herbal claims, the honest move here is the opposite one: do not manufacture a hazard where the record is reassuring.

The discipline required is at the boundary. Long use establishes a great deal about tolerability at food doses and almost nothing about efficacy for any disease — and the tradition itself makes claims that a modern reader will translate into disease claims it never made. This page keeps those apart, and looks closely at the two traditional uses that are actively marketed today: haemorrhoids and diuresis.

Table of Contents

  1. Food First: What the Culinary Record Establishes
  2. And What It Does Not Establish
  3. Three Traditions, Three Different Claims
  4. What the Names Imply, Checked
  5. The Haemorrhoid Claim and Borrowed Flavonoid Evidence
  6. The Diuretic Reputation, and Potassium
  7. Nutrition: Portion, Not Per 100 g
  8. Where Food Ends and Medicine Begins
  9. The Taste Barrier Is the Real Practical Issue
  10. Key Research Papers
  11. Connections

Food First: What the Culinary Record Establishes

Four separate food traditions use this plant in four different forms, which is unusual and analytically useful — the safety record is not resting on one preparation in one place.

What that gives, taken together:

  1. No recognised food-borne toxicity syndrome. There is no described illness attributed to eating this plant as food. For a plant eaten this widely, an acute or subacute toxicity of any prominence would have surfaced — that is how the toxicity of many other traditional foods was identified.
  2. Multiple organs, multiple preparations. Leaf and rhizome, raw and cooked, fresh and dried. A hazard confined to one form would have a chance to show itself in a population using another.
  3. Both extremes of the volatile fraction covered. The Vietnamese exposure is maximal in aldehydes; the Japanese tea is nearly free of them. Neither is associated with a described harm.
  4. Children and pregnancy included in customary practice, without a recognised signal — which is a real observation, even though it is not a substitute for the formal safety data that does not exist.

And What It Does Not Establish

The same record, read carefully, is silent on several things, and a reader is entitled to have the silences named rather than glossed.

  1. It says nothing about efficacy. That a food is eaten for centuries tells you it is tolerable and available and liked. It tells you nothing about whether it treats disease. Every widely eaten plant on earth has a folk medical reputation; the reputations cannot all be right.
  2. It does not cover concentrated extracts. A capsule of standardised extract is not the vegetable. It concentrates the plant by an unstated multiple, is taken daily and away from meals, and has no traditional record at all — the tradition is a few decades old at most for that form.
  3. It does not cover long-term daily high-dose tea. Ordinary household use of dokudami-cha is a cup or a few a day. That is well covered. Litres a day for years is a different exposure that nobody has studied.
  4. It cannot detect a slow, silent effect. Traditional surveillance detects poisoning, rashes and stomach upsets. A small increase in a chronic risk decades later is invisible to it — which is precisely why the aristolochic-acid question was worth asking analytically rather than settling by appeal to long use. See the safety page.
  5. It says nothing about drug interactions. The populations who established this food record were not on statins, anticoagulants, immunosuppressants or lithium. "No known interactions" here means no interaction study exists — absent data, not reassurance.

Three Traditions, Three Different Claims

The three medical traditions using this plant are usually presented as agreeing. They do not, quite, and the differences are informative.

Chinese medicine: yu xing cao, a heat-clearing herb for the lung. Classified as cold and acrid, grouped with the herbs that "clear heat and resolve toxin," and aimed at cough with thick yellow phlegm and the pattern called lung abscess. It is a mainstream pharmacopoeial herb, not folklore. Note the framing: it treats a pattern, within a formula, prescribed by a practitioner, for a limited period, and it comes with a traditional contraindication — long-term use is discouraged in people with what the tradition calls spleen-and-stomach deficiency cold, which in plain terms describes people who reliably get loose stools and bloating from cold raw food. A tradition that names a contraindication is doing something more careful than "natural and safe."

Japan: dokudami, a household detoxicant and daily tea. One of the folk-medicine staples of the domestic tradition, drunk continuously rather than prescribed for an episode, and credited broadly — skin clearing, mild laxative and diuretic action, general "detox." This is the most diffuse claim set of the three, and the least falsifiable. It is also the exposure that most resembles a modern supplement habit, which is worth noticing: the Japanese daily-tea pattern is the closest thing this plant has to long-term data, and it involves the preparation with the least of the interesting chemistry.

Vietnam: diếp cá, a food that doubles as a home remedy. Within thuốc nam, the leaves blended into a green juice are given for heat rash, boils, constipation and haemorrhoids, and to children as a cooling drink in hot weather. Two features distinguish it: the preparation is the food itself rather than a decoction, and the claims are for common self-limiting complaints rather than for serious disease.

Two observations follow. First, the three traditions do not converge on a single indication. A herb whose independent traditions agreed tightly — as several do for, say, sedation or for wound-healing — would offer a stronger prior. Lung heat, general detoxification and skin-and-bowel complaints are not the same claim, and the overlap between them is thin. Second, none of the three claims to treat a specific infection in the modern sense. The retrofit of "clears lung heat" into "antiviral" is done by contemporary marketing, not by the tradition, and the tradition should not be blamed for it.

What the Names Imply, Checked

A common name sometimes smuggles in a claim or a safety implication that the plant does not carry, so it is worth running the check explicitly and reporting the answer either way.

So the check comes back mostly negative, which is itself information: four of five names describe the smell or place the plant among vegetables, and only the Japanese name embeds a claim. And the honest caveat on that one — the etymology of dokudami is contested, with more than one proposed derivation, so this page reports the standard gloss without asserting that the gloss is the true origin. Where the name-check matters most is in the opposite direction from the usual case: unlike a name such as "motherwort," which implies an action on the uterus and therefore a pregnancy contraindication, none of Houttuynia's names implies a safety hazard that the plant then turns out to have.

The Haemorrhoid Claim and Borrowed Flavonoid Evidence

In Vietnam this is the plant's most commercially prominent medical use. Diếp cá juice is a standard home remedy for haemorrhoids and constipation, and packaged diếp cá products — powders, capsules, tablets — are sold specifically for it. A reader outside Vietnam will meet the claim through those products.

The proposed rationale is a flavonoid one, and it is worth taking seriously before taking it apart. Haemorrhoidal disease is partly a venous problem, and there is a real class of drugs called phlebotonics or venoactive agents used for it — most prominently micronised purified flavonoid fraction (MPFF), which is about 90% diosmin and 10% hesperidin. MPFF has randomised controlled trials in acute haemorrhoidal disease and in chronic venous insufficiency, and phlebotonics as a class have been assessed in systematic reviews. So "flavonoids can do something for veins" is not a fantasy claim.

But the transfer to Houttuynia fails on three counts, and they are worth labelling precisely at the point of transfer rather than in a preamble:

  1. Different compounds. The trial evidence belongs to diosmin and hesperidin — citrus flavanones. Houttuynia's flavonoids are quercetin and kaempferol glycosides: quercitrin, hyperoside, rutin, afzelin. These are a different subclass with different pharmacology. Rutin-family compounds have their own venotonic literature, but it is older, weaker and not interchangeable with the MPFF trials.
  2. Different dose and different material. MPFF trials use gram-scale daily doses of a micronised, purified, defined preparation — micronisation exists specifically because absorption of these compounds is otherwise poor. A glass of blended leaf juice is neither purified nor micronised nor dose-controlled.
  3. No trial of the herb. There is no randomised controlled trial of Houttuynia cordata in haemorrhoidal disease with a symptom or bleeding endpoint. The claim rests on tradition plus a borrowed mechanism.

The fair statement: the mechanism is not absurd, the compound class is adjacent to one with real evidence, and the herb itself has not been tested. If someone wants relief from haemorrhoids, the interventions with actual trial support are fibre, water, topical treatments, phlebotonics and, where warranted, procedures — and the plausible harmless benefit of drinking the juice may be that it is fluid and fibre-adjacent, which is the same first-line advice given by every colorectal service.

The Diuretic Reputation, and Potassium

All three traditions describe increased urine output, and quercitrin is the compound usually credited. This is one of the plant's oldest and most consistent reputations, and it deserves an honest and slightly deflating treatment.

What is actually known. There is no controlled human study measuring urine volume, sodium excretion or free-water clearance after a defined dose of Houttuynia. No diuretic mechanism has been established in people — not a loop, not a thiazide-like, not an aquaretic action. What exists is traditional description plus a general and rather weak association between flavonoid-rich infusions and mild diuresis, which is confounded by the obvious fact that drinking two or three cups of tea increases urine output regardless of what is in the cup. Nobody has controlled for the water.

Why the potassium question is worth raising anyway. If a substance genuinely increased renal sodium and water excretion, potassium handling would be affected — the direction depending entirely on which nephron segment was involved, which is unknown here. And separately, leafy greens and their infusions contribute potassium themselves. So the risk runs in two possible directions at once and neither has been measured. This page therefore does two things:

  1. It refuses a potassium figure. No milligrams-per-cup number appears here, because the content of an infusion depends on leaf quantity, water volume, steep time and whether the material was fresh or dried, and no reliable analysis of the brewed beverage is available to cite. A confident figure would be a guess.
  2. It names who should care. At ordinary culinary and household-tea amounts this is not a practical concern for a healthy person. It becomes one for people whose potassium handling is already precarious: chronic kidney disease, people taking ACE inhibitors, angiotensin-receptor blockers or potassium-sparing diuretics, and people taking prescription diuretics or lithium — where fluid-balance shifts change drug levels. Those readers should treat a several-litres-a-day tea habit as a change worth mentioning to their clinician, and should not use this plant as a diuretic in place of prescribed treatment.

For context on the mineral itself, see potassium.

Nutrition: Portion, Not Per 100 g

Herb pages routinely quote nutrient content per 100 grams, and for a garnish herb that is a misleading unit. Nobody eats 100 g of fish mint. A generous share of a Vietnamese herb platter is on the order of 5 to 15 grams of fresh leaf, eaten alongside several other herbs — so any per-100-g figure has to be divided by roughly seven to twenty before it means anything about a meal.

Three consequences, stated without numbers this page cannot source:

  1. As a vitamin or mineral source, the portion is negligible. Whatever the per-100-g content, a 10 g portion delivers a tenth of it. Fish mint is not a meaningful contributor of any nutrient at culinary amounts, and no herb eaten by the pinch is.
  2. The same applies to polyphenols, and it applies twice. First the portion divides the content; then poor absorption divides it again. Flavonol glycosides are hydrolysed, conjugated in the gut wall and liver, and degraded by colonic bacteria, so what reaches plasma is a small fraction of a small quantity.
  3. The tea is the larger polyphenol exposure, not the salad. A cup of dried-herb infusion extracts flavonoids from a gram or two of concentrated dried material, whereas the fresh leaf is mostly water. That reverses the intuition that the pungent raw herb must be the "stronger" form — it is stronger in aldehydes, weaker in everything stable.

The reason to eat it is that it tastes of something and cuts through fat. That is a complete and sufficient reason to eat a herb, and dressing it up as nutrition is unnecessary.

Where Food Ends and Medicine Begins

Most confusion about this plant comes from treating four exposures as one. They are not, and the risk-benefit reasoning differs at every step:

  1. Fresh leaf as food, with meals. Long, broad human record. Self-limiting quantities because it is pungent. The safest form and the one this page endorses without reservation for anyone who likes it.
  2. Dried herb tea, a few cups a day. Also long and broad, and the form with the least volatile chemistry. Effectively a beverage.
  3. Concentrated decoction as prescribed traditional medicine. Higher doses, within a multi-herb formula, for a defined period, under a practitioner. The tradition itself treats this as different from food, with its own contraindication. Reasonable within its own framework; not validated by trials.
  4. Standardised extract capsules taken daily and indefinitely. No traditional record, no long-term human data, unstated concentration factor, and the form for which the open questions on the safety page actually matter. This is the exposure that inherits none of the reassurance from the other three.

Anyone quoting "eaten safely for centuries" in defence of item 4 is performing exactly the transfer this page exists to block.

The Taste Barrier Is the Real Practical Issue

For most Western readers the operative question is not safety or efficacy but whether they can get it down. The aroma comes from medium-chain aldehydes — houttuynin, decanal, dodecanal, 2-undecanone — the same chemical family the nose reads as fish, coriander leaf and old frying oil. Reactions are strongly polarised, and while the soapy-coriander response has a documented genetic basis in an olfactory receptor gene, no comparable gene has been identified for this plant, so the split is best described as real, widely reported and unexplained.

Three things reliably reduce it, and they follow directly from the chemistry:

And a practical note that follows from the same instability: a weak smell means old stock, not a mild variety. Good fresh fish mint is firm, deep green and smells strongly from arm's length. Buy it from a grocery that sells it as food rather than from a garden centre — the variegated ornamental "chameleon plant" is the same species but is grown as a decorative plant and may carry pesticides not cleared for food crops.

Key Research Papers

Cited as PubMed searches rather than fixed identifiers, so the reader lands on the live record. Where metadata could not be verified, a topic search is given rather than a citation asserted.

  1. Houttuynia cordata Thunb.: a comprehensive review of traditional applications, phytochemistry, pharmacology and safety — Phytomedicine, 2024. The broad modern review, including the traditional-use record. PubMed search.
  2. Houttuynia cordata as a food and functional vegetable — topic search on the culinary and food-science literature, including composition of the leaf and rhizome. PubMed search.
  3. Volatile composition of Houttuynia cordata — topic search for the aldehyde and ketone profile responsible for the aroma, and how drying changes it. PubMed search.
  4. Micronised purified flavonoid fraction in haemorrhoidal disease — the trial literature that the Houttuynia haemorrhoid claim borrows from, and which belongs to diosmin and hesperidin rather than to this plant. PubMed search.
  5. Phlebotonics for haemorrhoids — systematic review — the class-level assessment, useful for judging how strong the borrowed evidence actually is. PubMed search.
  6. Quercetin glycoside bioavailability in humans — topic search establishing that intact quercitrin and rutin are poorly absorbed, which halves the nutritional argument twice over. PubMed search.
  7. Herbal diuretics — evidence in humans — topic search on how rarely traditional diuretic claims have been tested with urine-volume or natriuresis endpoints. PubMed search.
  8. Dietary potassium and chronic kidney disease — background for the group who should treat a heavy tea habit as a change worth mentioning. PubMed search.
  9. Genetic basis of cilantro soap taste and olfactory receptor variation — the documented case of a polarised herb response, offered as the contrast to Houttuynia, where no such gene has been identified. PubMed search.
  10. Houttuynia cordata anti-inflammatory activity — topic search for the flavonoid-fraction mechanism work that underlies most traditional-use rationalisations. PubMed search.

External Resources

Connections

Back to Table of Contents