evipedia.ai
← Back

Evodia rutaecarpa for Health & Longevity

Evidence Review created on 08/25/2026 using AI4L / Opus 5

Also known as: Tetradium ruticarpum, Euodia rutaecarpa, Evodiae Fructus, Euodiae Fructus, Wu Zhu Yu, Wuzhuyu, Evodia Fruit, Goshuyu

Motivation

Evodia rutaecarpa is a small tree in the citrus family whose dried, unripe fruit has been a fixture of Chinese and Japanese herbal practice for roughly two thousand years. The fruit is sharply pungent and was classified as a “warming” remedy, given for headache, cold-type stomach pain, vomiting and long-standing morning diarrhoea. Modern interest centres on two of its plant compounds, evodiamine and rutaecarpine, which act on the same heat-and-pain sensor in nerve endings that chili pepper compounds act on.

That shared target is why the fruit reappeared in Western supplement catalogues as a fat-burning ingredient, frequently blended with stimulants. In parallel, the fruit has long carried a formal “slightly toxic” designation in Chinese herbal texts, and laboratory work has repeatedly flagged liver injury at higher intakes. Almost all of the published biology comes from cells and rodents; human data on the fruit itself remain very thin.

This review examines what the evidence shows about the effects attributed to Evodia rutaecarpa, how strong that evidence is, what safety signals accompany it, and how the fruit is dosed, sourced and monitored.

Benefits - Risks - Protocol - Conclusion

A short, curated set of high-level sources that explain what this fruit is, what it does, and where its safety limits lie.

No material from any of the six priority platforms could be found. Evodia rutaecarpa is a niche traditional Chinese herb with almost no Western clinical footprint, and none of these outlets has covered it, its alkaloids, or its heat-receptor mechanism in a dedicated piece.

Grokipedia

Tetradium ruticarpum

Covers the botany, the accepted binomial and its synonymy, the traditional Chinese uses, and the alkaloid chemistry — useful for untangling the several Latin names attached to this one plant.

Examine

Evodia rutaecarpa

States plainly that no human studies exist on evodia berries, gives the traditional decoction (simmered water extract) dose, and explains why isolated alkaloid capsules are far weaker than whole fruit.

ConsumerLab

No ConsumerLab article exists for Evodia rutaecarpa. ConsumerLab has never run a product review or dedicated report on this fruit, which is not sold widely enough in North America to enter its testing queue.

Systematic Reviews

The systematic-review literature on this fruit is almost entirely preclinical, and the five qualifying papers below define its outer limits.

Both sides of the trade-off are represented: Yin et al. and Zheng et al. pool the claimed antitumour effect, while Zhou et al. and Sun et al. systematically review the hepatic and cardiac toxicity that is the principal risk. The one human-outcome pooling, Li et al., covers a multi-herb formula rather than the fruit alone, so no meta-analysis of the fruit itself in people exists.

Four of these five papers, and most of the primary literature cited throughout this review, come from Chinese universities of traditional Chinese medicine and their affiliated hospitals — institutions whose standing is bound up with the herbal tradition they study. That is a structural interest on the supporting side of the ledger, and it sits alongside the two industry-funded studies identified later in this review.

Mechanism of Action

Activity is carried by three related indole alkaloids — evodiamine, rutaecarpine and dehydroevodiamine — alongside bitter limonoids such as limonin.

The central target is TRPV1 (transient receptor potential vanilloid 1, the heat and chili-pepper sensor sitting on sensory nerve endings). Evodiamine occupies the same binding site as capsaicin but only partially opens the channel, and after an initial burst it desensitizes the nerve instead (Wang et al., 2016). Channel opening releases CGRP (calcitonin gene-related peptide, a powerful blood-vessel-relaxing nerve signal), which lowers blood pressure and blocks platelet clumping in hypertensive rats (Li et al., 2008).

Rutaecarpine separately switches on the aryl hydrocarbon receptor (a cellular sensor that turns up drug-clearing enzymes), strongly inducing CYP1A2 (cytochrome P450 1A2, the liver enzyme that clears caffeine, theophylline and several psychiatric medicines) (Estari et al., 2021).

Inside cells, evodiamine blocks topoisomerases I and II (the enzymes that untangle DNA during cell division), suppresses NF-κB (nuclear factor kappa B, the master switch for inflammatory genes) and activates PPAR-γ (peroxisome proliferator-activated receptor gamma, the nuclear receptor that improves insulin sensitivity) (Rebhun et al., 2015).

Pharmacologically, evodiamine is cleared in human liver preparations mainly by CYP3A4, CYP2C9 and CYP1A2 (drug-metabolising liver enzymes) into hydroxylated and demethylated products (Sun et al., 2013). Oral absorption is poor: dehydroevodiamine reaches about 15% in rats, evodiamine far less, with distribution favouring gut, liver and kidney and rodent half-lives of a few hours. No human pharmacokinetic data exist. Competing readings hold that nerve desensitization, not activation, drives most whole-animal effects.

Historical Context & Evolution

Evodiae Fructus entered the written record in the Han-dynasty Shennong Bencao Jing, classed among the middle-grade drugs and explicitly flagged as slightly toxic. Its original indications were narrow and symptomatic: cold-type abdominal and upper-stomach pain, vomiting of clear fluid, headache at the crown, and chronic early-morning diarrhoea. It was almost never given alone. The two canonical vehicles are Wuzhuyu Tang, combining it with ginger, ginseng and jujube for headache with vomiting, and Zuojin Wan, pairing it with Coptis chinensis for acid reflux. Japanese Kampo (traditional herbal medicine) adopted the same fruit as goshuyu.

Interest widened from the 1990s as three findings arrived from separate directions. Korean pharmacologists isolated dehydroevodiamine, showed it inhibited acetylcholinesterase (the enzyme that clears the memory-related signalling chemical acetylcholine) and reversed drug-induced amnesia in rats more potently than tacrine, then the only approved Alzheimer’s drug (Park et al., 1996). Researchers at Kyowa Hakko, an ingredient manufacturer with a direct commercial interest in the result, reported that evodiamine behaved like capsaicin at the heat receptor and cut visceral fat in rodents (Kobayashi et al., 2001). Chinese oncology groups reported broad tumour-cell killing.

Only the fat-loss claim reached people, where a trial of an evodia extract found no metabolic-rate change (Kim et al., 2008). The rest has never been tested in humans; what accumulated instead was safety evidence, as repeated rodent work identified the fruit’s own alkaloids as liver toxins. Current opinion treats the pharmacology as genuine and the clinical case as unmade rather than disproved.

Expected Benefits

High 🟩 🟩 🟩

No benefit of Evodia rutaecarpa reaches this evidence level. There is no meta-analysis, and no adequately powered randomized human trial, showing a clinical benefit for any endpoint.

Medium 🟩 🟩

No benefit reaches this evidence level either. The single human trial of an evodia extract was neutral, and the remaining human signals come only from trials of multi-herb formulas rather than the fruit alone, graded low-quality; every other efficacy claim rests on animal or cell work.

Low 🟩

Preclinical Antitumour Activity

Evodiamine kills many tumour cell lines by blocking DNA-untangling enzymes and triggering programmed cell death, and shrinks implanted tumours in mice. Two pooled animal analyses agree on direction. Nothing has been tested in humans, and poor absorption remains the obstacle (Yin et al., 2021; Zheng et al., 2025).

Magnitude: Pooled mouse tumour volume fell with a standardized mean difference (the effect expressed in units of the studies’ own spread, so results measured differently can be combined) of −5.99, with a 95% confidence interval (the range most likely containing the true value) of −8.89 to −3.10; pooled tumour weight fell with a standardized mean difference of −3.51. The studies disagreed with one another to an extreme degree, and no human tumour-response figure exists.

Headache and Migraine Relief in Traditional Formulas ⚠️ Conflicted

Wuzhuyu decoction, in which this fruit is the lead herb, is the classical prescription for headache with vomiting, and rodent work traces the effect to serotonin signalling. The one placebo-controlled trial was positive on pain intensity but negative on attack frequency, hence the conflicted flag (Liu et al., 2018).

Magnitude: In the 78-patient trial, pain-intensity scores separated from placebo at week 8, while attack frequency, attack days and analgesic consumption did not differ between groups; no effect-size figure was reported.

Gastrointestinal Symptom Relief

The fruit’s oldest use is settling cold-type stomach pain, reflux and vomiting. Animal and cell work supports it: evodiamine reduces ulcer area, dampens inflammatory signalling and modulates gut motility through the same heat receptor; human data cover only multi-herb formulas (Zhou et al., 2024; Cui et al., 2024).

Magnitude: Direction is consistent across rodent ulcer, colitis and reflux models, holding at oral doses in the tens of milligrams per kilogram; human symptom-score data exist only for multi-herb formulas containing the fruit, so the literature gives no outcome figure for the fruit alone.

Blood Pressure Lowering via Nerve-Peptide Release

Rutaecarpine triggers sustained release of a vessel-relaxing nerve peptide, producing a durable fall in blood pressure in genetically hypertensive rats. The mechanism is abolished by blocking either the heat receptor or the peptide receptor. No human study has been run (Li et al., 2008; Tian et al., 2019).

Magnitude: Direction is consistently downward in spontaneously hypertensive rats given oral rutaecarpine, and the fall persists for hours after dosing; no human blood-pressure figure has been published.

Blood Lipid and Atherosclerotic Plaque Reduction

Evodiamine lowered triglycerides, total and low-density lipoprotein cholesterol in high-fat-fed mice by switching on a cholesterol-export pathway, and cut plaque formation in a second mouse model. Both are rodent-only; no human lipid measurement exists (Hou et al., 2021; Zha et al., 2023).

Magnitude: Direction is consistently downward for triglycerides, total cholesterol and low-density lipoprotein cholesterol in mice given 10–20 mg/kg evodiamine for 8 weeks, alongside less liver fat and smaller plaques; no human lipid figure has been published.

Fat Loss and Heat Production ⚠️ Conflicted

Evodiamine mimics capsaicin at the heat receptor, raising heat dissipation and fat breakdown in rodents. The one human trial found no change in resting metabolic rate and no weight advantage over placebo, contradicting the animal picture and undercutting the fat-burner marketing (Kobayashi et al., 2001; Kim et al., 2008).

Magnitude: In rats fed an extract supplying 0.02% evodiamine for 21 days, kidney-region and testicular fat pads and liver lipids fell significantly; in 45 women given evodia extract for 8 weeks, resting metabolic rate was unchanged and body-mass index fell no more than on placebo.

Speculative 🟨

Cognitive Protection

Dehydroevodiamine inhibits the enzyme that clears the memory-related signalling chemical acetylcholine and reversed drug-induced amnesia in rats. The basis is rodent and test-tube only; no human cognitive study of any kind has been conducted.

Insulin Sensitization

Evodiamine binds and activates the nuclear receptor targeted by glitazone diabetes drugs, switching on the same gene programme in liver cells. The basis is cell-culture work funded by a supplement manufacturer; no glucose outcome exists.

Benefit-Modifying Factors

  • Drug-metabolising enzyme variants: CYP1A2 fast-metaboliser variants (a common gene variant that speeds caffeine clearance) and CYP3A4 activity govern how quickly the alkaloids are destroyed, so identical doses can produce very different exposure between individuals.

  • Baseline blood pressure: The vessel-relaxing peptide mechanism produces its largest signal in rats with high blood pressure and little effect in rats with normal pressure, so any blood-pressure benefit is likely confined to those starting elevated.

  • Baseline body fat and cold tolerance: The heat-receptor mechanism acts partly by blunting the perception of cold rather than by burning more energy, so people who already run warm are unlikely to notice a metabolic effect.

  • Sex-based differences: The one human trial enrolled only premenopausal women and found no benefit; rodent fat-loss studies used male animals almost exclusively, so no sex-specific efficacy comparison is possible.

  • Pre-existing conditions: Reduced stomach-acid output, reflux and cold-pattern digestive complaints are the classical responder profile; existing liver disease shifts the balance sharply against any expected benefit.

  • Age-related considerations: Liver clearance and kidney filtration both decline with age, raising alkaloid exposure from an identical dose in adults over about 65 and narrowing an already thin margin between effect and toxicity.

Potential Risks & Side Effects

High 🟥 🟥 🟥

No risk of Evodia rutaecarpa reaches this evidence level, because no controlled human safety study large enough to establish adverse-event rates has been conducted.

Medium 🟥 🟥

Dose-Dependent Liver Injury ⚠️ Conflicted

The fruit’s own alkaloids become reactive metabolites that bind liver proteins, drive oxidative stress and damage mitochondria; the fruit also carries a formal slightly-toxic designation in Chinese herbal texts. Evidence is conflicted: injury is reproducible in mice, zebrafish and human liver cells, yet a systematic review calls the toxicity question unsettled and a subchronic rat study found none. Cultivar explains much of the split — large-flowered material, poorer in the protective compound limonin, is hepatotoxic (liver-damaging) where smaller grades are not (Yan et al., 2023; Zhang et al., 2021).

Magnitude: Direction is consistent and cultivar-dependent — large-flowered fruit produced overt liver injury in mice while small- and medium-flowered grades did not, and a 13-week rat study of evodia fruit powder placed the no-observed-adverse-effect level above 2,000 mg/kg/day (Kim et al., 2014); no human incidence figure has been published.

Loss of Medication Efficacy from Enzyme Induction

Rutaecarpine is among the most potent known plant inducers of CYP1A2, acting through the aryl hydrocarbon receptor. Any drug cleared by that enzyme can drop to sub-therapeutic levels while evodia is taken — and rebound to toxic levels when it stops. The effect appears within hours of a single dose and is not predicted by measurable rutaecarpine blood levels, which makes it easy to miss (Estari et al., 2021; Ueng et al., 2005).

Magnitude: A single 100 mg/kg oral dose of rutaecarpine in rats tripled liver CYP1A2 activity within 3 hours and cut oral caffeine bioavailability by up to 75%, with clearance also increased after intravenous dosing.

Low 🟥

Kidney Injury

Evodiamine drives calcium overload in kidney cells through the heat receptor, triggering programmed cell death. Blocking that receptor or binding up the calcium prevented both weight loss and kidney damage in mice, confirming the pathway (Yang et al., 2024).

Magnitude: Direction is consistently harmful at above-therapeutic doses in mice, and fully reversible by receptor blockade; no human incidence figure exists.

Cardiotoxicity

Pharmacology reviews consistently list cardiac toxicity alongside liver toxicity as evodiamine’s two dose-limiting problems, based on rodent and zebrafish work. No human cardiac endpoint has ever been measured (Sun et al., 2020; Lin et al., 2024).

Magnitude: Direction is consistently harmful in rodent and zebrafish models, appearing only at the high, dose-limiting exposures also used to demonstrate antitumour activity; the literature reports no outcome figure, as no controlled trial has measured cardiac endpoints with this fruit in humans.

Increased Bleeding Tendency

Rutaecarpine strongly inhibits collagen-driven platelet clumping in washed human platelets and reduces clot formation in mice. Combined with the vessel-relaxing peptide effect, this creates a plausible additive bleeding risk with anticoagulants (Huang et al., 2021).

Magnitude: Rutaecarpine at 1–5 µM strongly inhibited collagen-induced human platelet aggregation while barely affecting aggregation driven by thrombin (the enzyme that converts fibrinogen into a clot); no bleeding-event rate has been reported in people.

Airway Narrowing and Sensory Irritation

Evodiamine contracts isolated airway tissue by releasing nerve peptides at the heat receptor, and produces burning and pain on injection before the nerve desensitizes. Rutaecarpine does not share this effect (Kobayashi et al., 2000; Kobayashi, 2003).

Magnitude: Evodiamine contracted guinea-pig airway tissue maximally at 3 µM, roughly threefold the capsaicin concentration required; no human airway figure exists.

Gastrointestinal Irritation

The pungency of this fruit is also its commonest complaint. Burning, mouth dryness, nausea and stomach discomfort are the classical dose-limiting effects and appear throughout the toxicology literature (Li & Wang, 2020). The mechanism is the same heat-receptor activation that carries the intended effects.

Magnitude: Direction is consistent and dose-related, appearing at the upper end of traditional decoction doses; no controlled incidence figure has been published.

Speculative 🟨

Uterine Stimulation in Pregnancy

Evodiamine has documented uterus-contracting activity, and classical texts treat the fruit as contraindicated in pregnancy. The basis is isolated-tissue pharmacology and traditional prohibition only; no clinical case series exists.

Excess Nerve Signalling at High Doses

Dehydroevodiamine crosses into the brain and inhibits the enzyme clearing acetylcholine, so very high intakes could in principle cause sweating, cramping or a slow heart rate. The basis is mechanistic; no case reports exist.

Risk-Modifying Factors

  • Drug-metabolising enzyme variants: CYP3A4 activity determines how much reactive metabolite is generated; inducers worsen liver injury and inhibitors reduce it, so anyone carrying high-activity variants or taking inducers sits at the exposed end.

  • Baseline liver enzymes: Elevated ALT (alanine aminotransferase, a liver enzyme released when liver cells are damaged) at baseline marks pre-existing liver stress and is the single strongest reason to avoid this fruit.

  • Baseline kidney function: Reduced filtration slows alkaloid clearance and compounds the calcium-overload mechanism behind kidney injury, so impaired function raises exposure and vulnerability together.

  • Sex-based differences: Rodent toxicity studies show comparable liver effects in both sexes but higher ketone excretion in male rats; no human sex-difference data exist, so this remains unresolved.

  • Pre-existing health conditions: Chronic liver disease, chronic kidney disease, bleeding disorders, reflux oesophagitis (acid-driven inflammation of the food pipe) and asthma each amplify a documented mechanism and shift the balance toward harm.

  • Age-related considerations: Adults past about 65 clear the alkaloids more slowly, are more often on CYP1A2-cleared medicines, and are likelier to be taking anticoagulants, stacking three risk channels at once.

Key Interactions & Contraindications

  • CYP1A2 substrates (theophylline, clozapine, olanzapine, tizanidine, duloxetine, melatonin, caffeine): Absolute contraindication for narrow-margin agents. Enzyme induction drives levels down, risking loss of seizure, psychiatric or airway control; levels rebound on stopping. Separation of weeks, with drug-level monitoring, is the stated mitigation.

  • CYP3A4 inducers (rifampicin, carbamazepine, phenytoin, St John’s wort, dexamethasone): Caution. Inducing CYP3A4 increased reactive-metabolite formation and worsened liver injury in mice (Zhang et al., 2021). The combination is avoided in practice.

  • Anticoagulants and antiplatelet drugs (warfarin, apixaban, clopidogrel, aspirin): Caution. Additive platelet inhibition raises bruising and bleeding risk. Mitigation is bleeding-sign and clotting-time monitoring, with discontinuation 10–14 days before surgery.

  • Antihypertensives (amlodipine, lisinopril, losartan) and nitrates: Monitor. The vessel-relaxing peptide mechanism is additive and can produce dizziness on standing. Seated and standing blood pressure over the first two weeks is the usual check.

  • Hepatotoxic drugs (paracetamol, methotrexate, isoniazid, high-dose statins): Caution. Overlapping liver injury pathways compound risk. Liver enzymes before and 4 weeks after starting, with paracetamol capped at 2 g/day, is the usual mitigation.

  • Liquorice-containing herbal formulas: Monitor. Liquorice inhibits CYP3A4 and lowers reactive-metabolite formation, shifting alkaloid exposure unpredictably in combined products (Chen et al., 2025). Dose is set by the finished formula, not by adding components.

  • Additive supplements — capsaicin, ginger, garlic, fish oil, ginkgo, nattokinase, berberine: Caution. Capsaicin and ginger stack on the same heat receptor and gut irritation; the rest add antiplatelet effect; berberine plus evodiamine is synergistic in animal work. Doses are reduced or separated.

  • Over-the-counter medicines — caffeine tablets, pseudoephedrine, NSAIDs (non-steroidal anti-inflammatory painkillers such as ibuprofen and naproxen), antacids: Monitor. Caffeine effect is blunted by enzyme induction, stimulant blends amplify heart rate, NSAIDs add bleeding and gut irritation, and antacids alter absorption.

  • Other interventions — surgery, alcohol, thermal stress: Caution. Alcohol compounds liver injury; sauna, heat training or intense heat exposure compound the heat-receptor and fluid-loss effects; anaesthesia interacts with the enzyme-induction profile.

Populations who should avoid Evodia rutaecarpa:

  • Pregnancy at any stage, and breastfeeding — documented uterus-contracting activity and no safety data
  • Any chronic liver disease, including Child-Pugh Class A cirrhosis (the mildest grade of liver scarring) or worse, hepatitis B or C, and fatty liver with ALT above the reference limit
  • Chronic kidney disease with an estimated filtration rate (eGFR, an estimate of how much blood the kidneys clean per minute) below 60 mL/min/1.73 m²
  • Anyone taking a narrow-margin CYP1A2-cleared drug such as theophylline or clozapine
  • Inherited or drug-induced bleeding disorders, and the 14 days before any planned surgery
  • Active reflux oesophagitis, peptic ulcer, or asthma with sensory-irritant triggers
  • Children and adolescents under 18 — no data at any dose

Risk Mitigation Strategies

  • Cultivar and species verification before purchase: Prevents the single largest liver-injury variable. Large-flowered material is hepatotoxic in mice where small- and medium-flowered material is not; only suppliers stating botanical origin and grade are verifiable.

  • Baseline liver panel before starting: Prevents unmasking existing liver injury. Protocols obtain ALT, AST (aspartate aminotransferase, a second liver-damage enzyme), ALP (alkaline phosphatase, which rises when bile flow stalls) and bilirubin, excluding anyone whose ALT exceeds the reference limit.

  • Low starting dose within traditional limits: Limits dose-dependent liver, kidney and gut toxicity. Protocols start near 1.5 g of dried fruit as a decoction daily and stay at or below 9 g/day in divided doses.

  • Liver-enzyme recheck at week 4: Detects injury before symptoms appear. Any doubling of ALT from baseline, or bilirubin above 1.2 mg/dL, is treated as an immediate stop signal with a repeat panel two weeks later.

  • Medication audit for CYP1A2 clearance: Prevents silent loss of drug efficacy. Where a substrate cannot be avoided, drug levels at baseline, week 2 and two weeks after discontinuation replace blind dose adjustment.

  • Dosing with food, stopping at first burning: Reduces the gut irritation, reflux and mouth dryness that are the commonest complaints. Upper-abdominal burning persisting beyond a week is treated as a discontinuation signal.

  • Time-limited courses: Avoids the entirely uncharacterised chronic-exposure risk. Traditional use is symptomatic and episodic; continuous use is capped at 4–8 weeks with a washout rather than left open-ended.

  • Discontinuation 10–14 days before surgery or invasive procedures: Prevents additive bleeding from platelet inhibition, allowing full platelet turnover before any procedure carrying bleeding risk.

Therapeutic Protocol

  • Standard traditional protocol: A decoction of 3–9 g of dried unripe fruit daily, divided into two or three servings taken morning, midday and evening — the dosing used in classical Chinese practice and reported on Examine.

  • Chinese Pharmacopoeia protocol: A more conservative 2–5 g/day, reflecting the fruit’s formal slightly-toxic classification. Neither approach is framed here as the default; the conservative range is the safer starting point.

  • Whole fruit versus isolated alkaloid: Isolated evodiamine capsules are absorbed far more poorly than fruit or ethanolic extract, so the traditional decoction and the modern capsule are not interchangeable at equal alkaloid content.

  • Multi-herb formula approach: Classical practice never uses this fruit alone. Wuzhuyu Tang for headache with vomiting and Zuojin Wan for reflux are the two formulas most cited by Chinese practitioners.

  • Best time of day: Taken with meals, splitting doses across morning and evening. Late-evening dosing is avoided because the enzyme-induction effect on caffeine and melatonin clearance can disturb sleep.

  • Half-life and dose splitting: Rodent half-lives run to a few hours and human data do not exist, so twice- or thrice-daily splitting is used rather than a single dose, mirroring traditional practice and limiting peak gut irritation.

  • Genetic polymorphisms: CYP1A2 and CYP3A4 activity variants (gene variants governing how fast the liver clears drugs) shift both exposure and reactive-metabolite load, so fast metabolisers of CYP3A4 warrant a lower starting dose.

  • Sex-based differences: No dosing difference is established. The only human trial enrolled premenopausal women exclusively, and rodent efficacy work used male animals, so neither sex has a validated dose.

  • Age-related considerations: Protocols for adults over about 65 start at the low end of the conservative range, given slower liver and kidney clearance and more interacting medication.

  • Baseline biomarkers: Normal liver enzymes and normal kidney filtration are the entry conditions; elevated baseline ALT predicts poorer tolerance and is the most useful single pre-dose measurement.

  • Pre-existing conditions: Cold-pattern digestive complaints and headache with nausea are the classical responder profile; reflux oesophagitis, ulcer or asthma predict poor tolerance of the same pungency.

Discontinuation & Cycling

  • Short-term, not lifelong: Traditional use is symptomatic and episodic. No study has examined continuous use beyond 13 weeks in animals or 8 weeks in humans, so open-ended daily use is entirely uncharacterised.

  • Withdrawal effects: None documented. The one relevant rebound is pharmacological, not physiological: drugs cleared by the induced liver enzyme rise back toward — and can overshoot — their previous levels within days of stopping.

  • Tapering: Not required for the fruit itself. Tapering matters only for co-administered CYP1A2-cleared medicines, whose doses may need downward adjustment as enzyme induction fades over one to two weeks.

  • Cycling: Commonly practised as 4–8 weeks on with a 2–4 week washout. The rationale is limiting cumulative liver exposure rather than preserving efficacy; no tolerance to any effect has been demonstrated.

  • Sensory desensitization: The heat receptor desensitizes with repeated exposure, so the warming and irritant sensations fade over days. This is expected and is not evidence that the underlying pharmacology has stopped.

Sourcing and Quality

  • Botanical identity verification: The most important single check. Several Tetradium species and the historical spellings Evodia and Euodia are conflated in trade; only suppliers confirming species by DNA barcoding or authenticated reference material are verifiable.

  • Cultivar and flower size: Large-flowered fruit carries more alkaloids and less protective limonin and was hepatotoxic in mice where small- and medium-flowered grades were not. Grade is rarely disclosed, and most suppliers cannot answer when asked.

  • Third-party testing: A current certificate of analysis from an independent laboratory covering alkaloid content, heavy metals, pesticide residues and microbial limits is the minimum standard. United States Pharmacopeia or NSF marks are rare on this herb but preferred where present.

  • Processing method: Traditional processing with liquorice or ginger measurably lowers hepatotoxic constituents. Raw, unprocessed fruit carries the highest alkaloid load and is the form most linked to liver injury in animal studies.

  • Formulation choice: Whole dried fruit or ethanolic extract reaches the bloodstream far better than isolated evodiamine capsules, which stay largely confined to the gut. Standardisation to evodiamine content alone can therefore mislead.

  • Reputable suppliers: Established traditional Chinese medicine dispensaries with granule or raw-herb testing programmes — Plum Flower, Nuherbs, Kan Herb Company and E-Fong — publish identity and contaminant testing that generic fat-burner brands do not.

  • Proprietary stimulant blends: Evodia is most often sold inside proprietary fat-burner stacks alongside ephedra-type stimulants and synephrine, where dose is undisclosed and the combination has never been tested for safety.

Practical Considerations

  • Time to effect: Digestive and headache effects in traditional use are described within days. The trial of the traditional headache formula needed 8 weeks to separate from placebo, so 4–8 weeks is the earliest reasonable assessment point.

  • Pitfall — treating it as a fat burner: The rodent fat-loss data did not replicate in the only human trial. Buying evodia for weight loss means paying for an effect that controlled human evidence does not support.

  • Pitfall — ignoring the enzyme interaction: The most consequential error is taking evodia alongside a CYP1A2-cleared medicine. The failure is silent, showing up as loss of drug control rather than as an obvious side effect.

  • Pitfall — buying unlabelled cultivar or proprietary blends: Both remove the two variables that matter most for safety — alkaloid load and total dose — and make an adverse event impossible to attribute.

  • Regulatory status: In the United States it is a dietary supplement under the Dietary Supplement Health and Education Act, not reviewed for efficacy by the U.S. Food and Drug Administration. In China it is a Pharmacopoeia-listed medicinal with a slightly-toxic designation.

  • Cost and accessibility: Inexpensive and unpatentable, which is informative: no manufacturer can recoup trial costs, and neither insurers nor health systems have any financial incentive to fund comparisons against patented drugs. That structural gap, not negative findings, best explains the missing human evidence.

Interaction with Foundational Habits

  • Sleep: Indirect and potentially disruptive. Inducing CYP1A2 speeds clearance of both caffeine and melatonin; caffeine tolerance appears to improve while melatonin exposure may fall. Morning and midday dosing rather than late evening is the standard mitigation, and any new difficulty falling asleep is treated as attributable.

  • Nutrition: Direct and potentiating. The alkaloids are poorly water-soluble, so a meal containing fat measurably improves absorption, while food also buffers gut irritation. Grapefruit inhibits the same enzyme that clears evodiamine, and alcohol compounds the liver injury pathway, so both are excluded.

  • Exercise: Indirect, with no evidence of blunting. Nothing suggests interference with muscle growth or endurance adaptation, unlike antioxidant supplements. The heat-receptor effect raises heat dissipation, so training in hot conditions raises the importance of fluid and electrolyte replacement.

  • Stress management: No direct interaction demonstrated. No study has measured cortisol or stress-axis endpoints with this fruit. The one indirect link runs through sleep: if enzyme induction degrades sleep quality, stress resilience falls with it, which is the practical reason to protect dose timing.

Monitoring Protocol & Defining Success

Because the dominant safety signal is liver injury and the dominant interaction is enzyme induction, monitoring is built around those two axes rather than around any efficacy marker. Baseline work-up comprises a full liver panel, kidney function with electrolytes, a complete blood count, seated and standing blood pressure, and — where a relevant medication is in use — a baseline drug level. Anyone with a liver enzyme above the reference limit is excluded at that point. Ongoing monitoring runs at 4 weeks, again at 12 weeks, then every 6 months while use continues, with an extra liver panel two weeks after any dose increase and a repeat drug level two weeks after stopping. Any doubling of a liver enzyme from an individual’s own baseline is a stop signal rather than a reason to recheck later.

Biomarker Optimal Functional Range Why Measure It? Context/Notes
ALT < 25 U/L (men), < 20 U/L (women) Earliest and most sensitive signal of liver-cell injury ALT = alanine aminotransferase. Conventional labs flag only above 40–55 U/L, far too late here. Fasting not required
AST < 25 U/L Confirms liver injury and flags muscle sources AST = aspartate aminotransferase. Conventional labs flag only above about 40 U/L. Interpret alongside ALT; avoid measuring within 48 h of hard training
ALP and total bilirubin ALP 40–90 U/L; bilirubin 0.3–1.0 mg/dL Detects the bile-flow pattern of drug-induced liver injury ALP = alkaline phosphatase. Conventional labs flag ALP only above about 120–145 U/L and bilirubin above 1.2 mg/dL. Bilirubin rising with ALT is the most serious combination; recheck within 7 days
GGT < 20 U/L (men), < 15 U/L (women) Sensitive marker of liver stress and enzyme induction GGT = gamma-glutamyl transferase. Conventional upper limits run to about 55 U/L (men) and 38 U/L (women). Rises with alcohol; a rising GGT with normal ALT still warrants attention
eGFR and creatinine eGFR > 90 mL/min/1.73 m²; creatinine mid-reference Tracks the kidney-injury mechanism seen in mice eGFR = estimated glomerular filtration rate. Conventional labs flag only below 60 mL/min/1.73 m², far too late here. Pair with cystatin C in muscular individuals; hydrate normally before the draw
Blood pressure (seated and standing) < 120/80 mmHg seated, drop < 10 mmHg on standing Captures the vessel-relaxing peptide effect and dizziness risk Measure after 5 min seated, then 1 and 3 min standing; morning readings before dosing are most comparable
Complete blood count with platelets Platelets 200–300 ×10⁹/L; haemoglobin mid-reference Baseline for the antiplatelet and bleeding signal Conventional labs accept 150–450 ×10⁹/L. Platelet count does not fall with this mechanism; it is the reference point if unexplained bruising appears
Relevant drug level (e.g. theophylline) No established target for this interaction — track the individual’s own pre-evodia level and treat any fall below it as significant Detects silent loss of medication efficacy from enzyme induction Applies only to CYP1A2-cleared drugs. CYP1A2 = cytochrome P450 1A2. Sample at trough, same time of day each draw

Qualitative markers worth tracking alongside the labs:

  • Upper-abdominal burning, nausea or mouth dryness — the commonest early dose-limiting signals
  • Unusual fatigue, dark urine, pale stool or itching — the classic early warning signs of drug-induced liver injury
  • Easy bruising, prolonged bleeding from small cuts, or nosebleeds
  • Sleep-onset latency and subjective sleep quality, given the enzyme-induction effect on melatonin clearance
  • Dizziness or light-headedness on standing, particularly in the first two weeks
  • Headache frequency and pain intensity, where headache is the reason for use

Emerging Research

  • Wu-Chu-Yu Tang for reflux disease: A placebo-controlled trial of the evodia-lead formula in gastroesophageal reflux, 90 participants, registered as Phase 4 with a symptom questionnaire as the primary endpoint (NCT01822106). Status is listed as unknown, a recurring problem in this literature.

  • Mechanistic reflux follow-up: A 55-participant Phase 4 study of the same formula using 24-hour oesophageal acid measurement and impedance testing rather than questionnaires, which would give the first objective human physiological endpoint for this fruit (NCT04118647).

  • Delivery systems that could strengthen the case: Phospholipid complexes, nanoemulsions and solid dispersions are being developed specifically to overcome evodiamine’s poor absorption (Lin et al., 2024). If they succeed, the animal efficacy data become testable in people for the first time.

  • Combination pharmacology: Berberine plus evodiamine improved fat browning in mice and reflux markers in animal models (Zhang et al., 2025). Whether the pairing also multiplies liver risk is the open question.

  • Toxicity work that could weaken the case: Large-scale gene and protein profiling has now localised liver injury to one solvent fraction of the fruit (Li et al., 2026), while transporter work shows its indole alkaloids bypass the cation transporters that carry partner-herb alkaloids (Zhang et al., 2022). Both could tighten safety limits.

  • The decisive missing study: No human pharmacokinetic study of evodiamine or rutaecarpine has ever been published. Until absorption, half-life and clearance are measured in people, no dose in this review can be considered validated rather than inherited.

Conclusion

Evodia rutaecarpa is the dried unripe fruit of a small citrus-family tree, used for about two thousand years in Chinese and Japanese practice for headache with vomiting, cold-type stomach pain and long-standing diarrhoea. Its plant compounds act on the same nerve-ending heat sensor as chili pepper, which explains both its warming character and its later reinvention as a fat-burning supplement ingredient.

The honest summary is that the biology is real and the clinical case is unmade. Tumour-cell killing, blood-pressure lowering, stomach protection and fat loss are all reproducible in animals, and several have been confirmed across many separate animal studies. None has been demonstrated in people. The one controlled human trial of an evodia extract found nothing, and the one trial of the traditional headache formula was positive on pain intensity but not on how often headaches occurred. Against that thin benefit picture sits a reproducible liver-injury signal that varies with which grade of fruit is used, and a potent effect on a liver enzyme that can quietly strip other medicines of their effect.

Two of the most influential supporting studies came from companies selling ingredients or supplements, and the remaining literature is dominated by groups with an institutional stake in traditional Chinese medicine — an imbalance that cuts in both directions and leaves the evidence base narrower than its volume suggests.

Top - Benefits - Risks - Protocol