Coming soonCannabis EducationCannabis Education at the Institute: a professional course and a free public program.Register your interest

Not-for-profit research & education institute · Cassowary Coast, Far North Queensland


Bitter melon (Momordica charantia)

Bitter melon (Momordica charantia) - Australian Institute of Pharmacognosy
In this article
  1. 0. Free-teaser block
  2. 1. Nomenclature
  3. 2. Definition
  4. 3. Source and Australian supply
  5. 4. Macroscopy
  6. 5. Microscopy
  7. 6. Chemistry
  8. 7. Named markers and assay limits
  9. 8. TLC fingerprint
  10. 9. HPLC fingerprint
  11. 10. Identity, purity, adulterants and contaminants
  12. 11. Traditional use
  13. 12. Current practice: use, preparations and doses
  14. 13. Clinical evidence
  15. 14. Pharmacology
  16. 15. Safety
  17. 16. Discussion
  18. 17. Conclusions
  19. 18. Research still needed
  20. 19. References
  21. 20. Document control
Share this articleFacebookXLinkedInWhatsApp

AIP-PH-MOMCHA · Version 1.0 draft · First published 2026 · Last modified 2026 · Version history · Report an error

Bitter melon fruit (Momordica charantia L.), Cucurbitaceae

Australian Herbal Pharmacopoeia (working title) · Monograph AIP-PH-MOMCHA · Version 1.0 draft
Date: 2026
Status: working document, version 1.0. Not a statutory standard.
Prepared by: the Australian Institute of Pharmacognosy, Far North Queensland, Australia. Every author and contributor is named at the end of the monograph.

Figure F1. Momordica charantia growing wild at Bulgun, near Tully, Queensland, within 50 km of Cardwell, 2022: a ripenin
Figure F1. Momordica charantia growing wild at Bulgun, near Tully, Queensland, within 50 km of Cardwell, 2022: a ripening warty fruit, yellow-orange and spindle-shaped, hanging among deeply lobed leaves. Photograph J.C. Frost, iNaturalist observation 123796169, CC BY 4.0.

0. Free-teaser block

0.1 Identity card

Series number AIP-PH-MOMCHA (MediHerb A–Z list No. 16)
Title Bitter melon fruit (Momordica charantia L.), Momordicae charantiae fructus
Version / year 1.0 draft, 2026. Next review 2029, or sooner if a registered trial listed in section 13.3 reports, or if a pharmacopoeia adopts an assay for the fruit
Accepted binomial Momordica charantia L., Sp. Pl. 2: 1009 (1753). Accepted. IPNI 293413-1. Checked in September 2026 and again in October 2026 against the World Checklist of Vascular Plants (the data behind Plants of the World Online, dataset version 2026-09-23, unchanged between the two checks) and IPNI [1,2]
Family Cucurbitaceae (APG IV), order Cucurbitales [1,3]
Plant part The fruit, fresh (as juice) or dried, usually harvested green and unripe; whole fruit with or without seeds, or the pulp without seeds [4,5]. The MediHerb tablet uses a dry extract of the fruit [6]
Latin drug name Momordicae charantiae fructus; WHO heads its monograph Fructus Momordicae [4]
Principal synonyms Momordica indica L., M. elegans Salisb., M. muricata Willd., M. zeylanica Mill., M. charantia var. abbreviata Ser. and 22 other names, 27 in all (24 in Momordica, 3 in Cucumis and Sicyos), placed under subsp. charantia [1]. WHO also lists M. balsamina Blanco and M. chinensis Spreng. [4]; see section 1 on why M. balsamina L. is a different species
Common names English: bitter melon, bitter gourd, balsam pear, African cucumber, wild cucumber [4,7]. Other vernacular names include karela (India), kāravella, kuguazi, ku gua, goya, nigauri, ampalaya, cundeamor and cerasee [4,5,7]. Trade and regulatory names: TGA ingredient name MOMORDICA CHARANTIA [8]; the MediHerb Metabol Complex label names "Momordica charantia fruit Extract dry concentrate" [6]
Pharmacopoeial status elsewhere WHO: yes, vol. 4 (2009), Fructus Momordicae [4]. Ayurvedic Pharmacopoeia of India: yes, Part I vol. II (1999), the source of the WHO purity limits, as cited by WHO [4]. Ph. Eur., BP, USP, EMA/HMPC, BHP: no monograph was found. Status in the Chinese Pharmacopoeia was not checked for this version
Australian status Permissible Ingredients Determination (No. 2) 2026: item 3392 MOMORDICA CHARANTIA, purposes A (active) and H (homoeopathic), no specific requirement [8]. Poisons Standard (June 2026): not named [9]. ARTG: 33 entries name the plant (public search September 2026), among them MediHerb Metabol Complex, ARTG 300236, listed 2/03/2018 [6,10]. Native status: the Australian Plant Census gives its distribution as Christmas Island and Queensland, naturalised in New South Wales [3]. Not named in the Queensland biosecurity legislation [11]
Named markers No pharmacopoeia sets an assay limit. Purity limits only (WHO, from the Ayurvedic Pharmacopoeia of India): total ash not more than 8.5 per cent, acid-insoluble ash not more than 0.6 per cent, water-soluble extractive not less than 28 per cent, alcohol-soluble extractive not less than 6 per cent [4]. Proposed identity and assay markers: a five-compound set of cucurbitane triterpenoids [5]. Safety marker needing an upper limit: vicine (CID 135413566), in the seeds [5]
Evidence snapshot Type 2 diabetes and prediabetes pooled, glucose and HbA1c: Tier B, no reliable effect in blinded trials of the fruit alone, 9 RCTs pooled (414 participants), low certainty. Prediabetes, fasting glucose: Tier C, inconsistent (a small fall in one 12-week trial, none in two others), 3 RCTs read in full, very low certainty. Blood lipids, weight, blood pressure: Tier B, no effect. Metabolic syndrome: Tier C, 1 open uncontrolled trial. Digestive, bitter-tonic, anthelmintic and fever uses: Tier T. Antiviral, anticancer and insulin-sensitising actions: Tier P
Suggested citation Ridley T, Iggulden L. Bitter melon fruit (Momordica charantia L.), Momordicae charantiae fructus. AIP-PH-MOMCHA, version 1.0 draft. Cardwell: Australian Institute of Pharmacognosy; 2026. DOI to be assigned
Provenance Original AIP data: none. Every limit, method, analytical figure, trial result and case report in this monograph belongs to the source cited beside it. Sections 4, 5, 7, 8, 9 and 10 carry AIP bench data pending: AIP has not vouchered a sample, made a section or powder mount, run a plate or run a chromatogram of this drug. Photographs and plates are licensed or public-domain works credited in each caption; F6b, F8 and F10c are journal figures reused under CC BY 4.0; F3b and F5 are AIP schematics redrawn from cited descriptions; F6c and F6d are AIP replots of published tables; F9 and the other F10 charts are AIP charts drawn from published data
Compiler and review Compiled by Dr Thomas Ridley (Education and Research). Technical review: Dr Fiona MacAllister (Botanical, sections 1-5 and the adulterant species in 10), reviewed in 2026; Dr Hana Suzuki (QA, sections 6-10), reviewed in 2026; Dr Eleanor Whitcombe (Medicine, sections 12, 13, 15), reviewed in 2026; evidence and citation audit, Dr Thomas Ridley (Education and Research, all sections, references and figure sources), reviewed in 2026; layout, figure placement and figure credits, Jesse Kirby (Community Outreach, whole document), reviewed in 2026. Published: not yet. Last modified: 2026. Version 1.0 draft
Canonical URL /pharmacopoeia/momordica-charantia/ (on publication)
Figure F1b. A yellow, five-petalled unisexual flower of the same wild Bulgun plant, 28/06/2022. WHO describes the flower
Figure F1b. A yellow, five-petalled unisexual flower of the same wild Bulgun plant, 28/06/2022. WHO describes the flowers as yellow to orange and about 3.5 cm across (ref 4). Photograph J.C. Frost, iNaturalist observation 123796169, CC BY 4.0.

0.2 Summary

Bitter melon is the fruit of Momordica charantia, a scrambling annual vine of the tropics that is eaten as a vegetable across Asia, Africa and the Americas and grows wild on the Queensland coast around Cardwell. The drug is the green, warty fruit, used fresh as juice or dried in slices, powders and extracts. No European, British or United States pharmacopoeia has a monograph. The World Health Organization gives a botanical description and four purity limits taken from the Ayurvedic Pharmacopoeia of India, and leaves microscopy and assay "to be established". The chemistry is dominated by some two hundred cucurbitane-type triterpenoids, which vary seventeen-fold between cultivars supplied by a single research station; the older marker, charantin, is a pair of common plant sterol glucosides and a poor guide to quality. The fruit is best known as a traditional remedy for diabetes. Blinded trials of the fruit alone, pooled, do not show a reliable effect on glucose or HbA1c in type 2 diabetes; in prediabetes, one 2025 trial found a small fall in fasting glucose after twelve weeks, and two other trials found none. The fruit is widely eaten as food, and mild digestive upset is the usual adverse effect. The main safety points are additive glucose-lowering with diabetes medicines, avoidance in pregnancy and in children, and the seeds’ vicine, a risk in glucose-6-phosphate dehydrogenase deficiency. Anyone considering bitter melon medicinally should consult a degree-qualified herbalist with competencies and adequate training in pharmacognosy.

0.3 Contents with extent

Section Content Pages
0 Free-teaser block: identity card, summary, contents, contributors 4
1 Nomenclature 1
2 Definition 1
3 Source and Australian supply, with 3.1 origin, climate and chemistry 2
4 Macroscopy 1
5 Microscopy 1
6 Chemistry 1
7 Named markers and assay limits 1
8 TLC fingerprint 1
9 HPLC fingerprint 2
10 Identity, purity, adulterants and contaminants 2
11 Traditional use 1
12 Current practice: use, preparations and doses 2
13 Clinical evidence 5
14 Pharmacology 1
15 Safety 3
16 Discussion 1
17 Conclusions 1
18 Research still needed 1
19 References 3
20 Document control 1
Totals 21 sections · about 59 printed pages · 47 figures (16 licensed or public-domain photographs and plates, 3 journal figures reused under CC BY 4.0, 11 PubChem structure panels, 2 AIP schematics redrawn from published descriptions, 2 AIP replots of published tables, 11 AIP charts, 2 captioned placeholders) · 0 AIP photomicrographs · 0 AIP chromatograms · 15 tables · 5 studies graded from full text (4 randomised: Mes 2025 studies 1 and 2, Kim 2023, Yang 2022; 1 uncontrolled: Tsai 2012), 7 further blinded trials reported through a meta-analysis read in full · 43 references; every PMID and DOI resolved in September 2026, and the full text of every cited journal article read

Known gaps in this version, stated plainly.

  1. F4, the photomicrographs, and F7, the TLC plate, are captioned placeholders. AIP holds no vouchered sample. Europe PMC open-access literature and Wikimedia Commons were searched in September 2026, and again for the photomicrograph in October 2026, for a reusable photomicrograph of the fruit or its powder, and for a reusable TLC plate of the fruit; none was found. WHO itself leaves microscopy "to be established" [4].
  2. No pharmacopoeial identity or assay test exists for the fruit in any text read for this monograph. Sections 7 to 9 set out published methods and an interim AIP proposal; none is a validated AIP method.
  3. Several trials reach this monograph only through a meta-analysis. Cortez-Navarrete 2018 and 2022, Dans 2007, John 2003, Kim 2020, Kinoshita 2018 and Trakoon-osot 2013 are reported as Laczkó-Zöld and colleagues tabulated them [12]; Fuangchan 2011, Krawinkel 2018 and Guarneiri 2025 as other papers read in full describe them. Their primary papers are on the library fetch list and have not had the integrity screen. The 2012 Cochrane review is reported as Braun and Cohen and Çiçek summarise it [5,7].
  4. Searches not completed: ANZCTR refused automated access in September 2026, so its records were reached only through WHO ICTRP (searched in September 2026); Cochrane CENTRAL and Embase were not searched; the TGA safety alerts were searched, but the TGA Database of Adverse Event Notifications search could not be completed and needs a manual search (section 15); the full texts of the pazopanib interaction case and the osteoarthritis trial named in sections 13 and 15 are on order and unread; DAFF BICON import conditions, the IUCN Red List and CITES were not queried; PubPeer was not checked for the cited papers.
  5. Queensland collection law was not read. The Australian Plant Census treats the species as native in Queensland [3], so wild collection falls under the Nature Conservation Act 1992 (Qld); the regulation that governs taking protected plants was not read for this version.
  6. The Chinese Pharmacopoeia and the Ayurvedic Pharmacopoeia of India were not read directly. The Indian limits are quoted from WHO [4].

0.4 Contributors and paywall notice

Prepared by: the Australian Institute of Pharmacognosy, Far North Queensland, Australia. Every author and contributor is named at the end of the monograph.
Contributors: listed with the monograph and resolved from the Legal byline roster by the publication pipeline for each department that reviews the draft. Dr Luke Iggulden, Chief Scientific Officer, is senior author: he set the scope of the Pharmacopoeia in the directive of 22/09/2026 and made the research possible.
What is free and what is not. Everything above the free-sample line is free to every reader. The remaining sections, macroscopy and microscopy, chemistry and markers, the identity tests, purity and adulterants, tradition, doses, graded evidence, pharmacology, safety and references, are read by Scholar-level members and above under a revocable personal licence. Nothing behind the line contradicts anything in front of it.

The Australian Herbal Pharmacopoeia is the name of a reference series published by the Australian Institute of Pharmacognosy. It is not a statutory standard under the Therapeutic Goods Act 1989 and carries no official, government or national-standard status.

Free to every reader. The safety section of every monograph is open to everyone, signed in or not. In an emergency call 000. For poisoning, call the Poisons Information Centre on 13 11 26.

15. Safety

The cautions that matter most, stated first. (1) Anyone on insulin, a sulfonylurea or other glucose-lowering medicine who adds bitter melon may see extra glucose lowering; WHO and Braun and Cohen expect an additive effect and advise glucose monitoring, although no interaction study exists [4,7]. (2) Avoid in pregnancy [4,5,7]. (3) Do not give to children, and do not take while breastfeeding [4]. (4) Avoid in glucose-6-phosphate dehydrogenase (G6PD) deficiency, because the seeds contain vicine [5,7]. (5) A product sold in Australia as "Bittermelon Slimming Capsules" was found by the TGA to contain undeclared sibutramine [40]; bitter melon bought from unknown online sellers carries that risk.

Adverse effects in trials. No serious adverse effects were reported in the nine blinded trials; headache and gastrointestinal complaints were the most common, and 6 g pulp a day caused more diarrhoea and flatulence than placebo; liver enzymes and creatinine did not change, with wide confidence intervals [12]. Braun and Cohen describe epigastric discomfort, pain and diarrhoea that stop when treatment stops [7]. In the 12-week Mes trial, side effects were few and similar in both arms by the authors’ judgement (nausea in 4 participants on bitter gourd and 1 on cucumber; bloating or abdominal pain in 4 and 2), and liver and kidney markers did not worsen [16]. In Kim 2023, dizziness was reported by 8 of 39 participants on the extract and none of 37 on placebo (p = 0.005), and fatigue by 11 and 4 (p = 0.084). One participant on the extract developed acute hepatitis; the authors attributed it to another herbal medicine taken during the trial and excluded that participant from the analyses, and they report no critical side effects [24]. In the Yang trial one participant on the extract withdrew with a widespread papular rash, which persisted after the extract was stopped and which the investigators judged unrelated [19].

Case reports (causality by AIP, WHO-UMC wording).

  • Hypoglycaemic coma and convulsions in two children aged three and four after a tea of leaves and vines taken before food [4,5]: not assessable from the summaries, and the preparation was leaf and vine, not this drug.
  • Atrial fibrillation in a 22-year-old man after two days of juice of crushed bitter melon (part not stated), two tablespoons three times a day; sinus rhythm restored with amiodarone; no recurrence off the herb [41]: possible.
  • Acute interstitial nephritis, biopsy-proven, in a 60-year-old woman with type 2 diabetes and hypertension after seven months of a pure extract, taken with glimepiride, metformin, gemigliptin, pioglitazone, amlodipine, losartan and carvedilol; the extract and pioglitazone were stopped and prednisolone given, and creatinine returned close to baseline in three months [42]. The authors rated the link probable (Naranjo score 6; WHO-UMC probable) [42]; AIP rates it possible, because pioglitazone was withdrawn at the same time.
  • Herb-induced liver injury in a 38-year-old man taking a Haitian blend of asowosi (bitter melon), cactus and aloe vera twice daily for two to three weeks, with amoxicillin, nimesulide and two to three alcoholic drinks a week; liver histology supported herb-induced injury [43]: unassessable for bitter melon alone, given the co-exposures.
  • Headache with ingestion of seeds [7].

Toxicology. Rats fed 0.02 to 0.5 per cent fruit in the diet for eight weeks showed no adverse effects; juice at 10 mL/kg for 30 days raised γ-glutamyl transferase and alkaline phosphatase [4]. Toxicity and death occurred in animals given high-dose extracts intravenously or intraperitoneally [7]. A seed extract was lethal to zebrafish embryos at 50 µg/mL, and fruit extract caused embryonic cardiac toxicity at 50 µg/mL [5]. Extracts were negative in the Ames test and positive in the comet assay [4].

Pregnancy. Avoid. Momorcharins from the seeds induced early and midterm abortion in mice and were teratogenic to cultured mouse embryos, and WHO contraindicates the seeds in pregnancy [4]. Fresh fruit juice at 6 mL/kg caused uterine bleeding in pregnant rats and rabbits, and Çiçek advises that bitter gourd preparations be avoided in pregnancy [5]. Braun and Cohen contraindicate bitter melon in pregnancy on animal evidence and traditional use (emmenagogue) [7]. Metabol Complex carries a pregnancy warning [6]. The trials read excluded pregnant women [16]. Grade: animal evidence; no human data.

Lactation. Avoid. WHO advises that the crude drug and its preparations not be taken during breastfeeding, reasoning from the severe hypoglycaemia reported in children [4]. No study of transfer into breast milk was found, and the trials read excluded breastfeeding women [16]. Grade: extrapolation from case reports; no lactation data.

Children. Do not give. WHO advises against use in children, citing hypoglycaemic coma and convulsions [4]. Grade: case reports (leaf-and-vine tea).

G6PD deficiency. Avoid, because the seeds contain vicine, which causes favism in G6PD-deficient people [5,7]. Çiçek suggests that seed-free fruit may reduce the risk [5]; certificates and labels should state whether seeds are included (section 2). Grade: mechanistic, with older human reports cited by Çiçek.

Men trying to father a child. Braun and Cohen advise against use, from antispermatogenic effects in animals [7]; seed and fruit extracts reduced spermatogenesis in rats and dogs [5]. Grade: animal.

Seeds and outer rind. Braun and Cohen advise avoiding them because of toxic lectins [7].

Warnings. WHO advises people with liver disorders to seek professional advice before use [4].

Herb-drug interactions.

Interaction Evidence Grade Source
Insulin, sulfonylureas and other glucose-lowering medicines: additive hypoglycaemia expected from the pharmacology; not studied as an interaction. In the Yang trial 20 people took the extract with metformin (18) and sulfonylureas or other secretagogues (14) for 12 weeks; the report describes no hypoglycaemic episode, although hypoglycaemia was not a stated outcome theoretical WHO [4]; Braun and Cohen [7]; Yang [19]
Drug-metabolising enzymes fruit juice reversed diabetes-related changes in hepatic cytochrome P450 and glutathione S-transferase activities in diabetic rats; no human pharmacokinetic study found theoretical, animal WHO [4]

WHO notes that added hypoglycaemic effects "need to be investigated" [4]. Braun and Cohen advise people with diabetes to monitor blood glucose, as medication may need changing [7]. A registered Indian trial adds bitter melon to metformin and glimepiride and lists adverse effects as an outcome (CTRI/2024/06/068929; section 13.3); no results were found. A case report of acute pancreatitis attributed to an interaction between bitter melon and the kinase inhibitor pazopanib (PubMed 34569860) has been identified; its full text is on order, it has not been read, and it is not graded here.

Older people. No specific data. The interaction and hypoglycaemia cautions apply with more force to people on several medicines, as in the nephritis case [42].

Overdose. No overdose reports were found in the sources read.

Australian regulatory signals.

  • Poisons Standard: the plant is not scheduled [9].
  • TGA safety alerts: a search of the TGA website for "momordica", "bitter melon", "bitter gourd" and "karela" in September 2026 returned one safety alert (a repeat search in October 2026 was stopped by the site’s automated-traffic check, so the September 2026 result stands). On 27/07/2017 the TGA warned that "Bittermelon Slimming Capsules", bought online by people in Australia, contained undeclared sibutramine, a prescription-only medicine withdrawn in 2010 after it was linked to major cardiac events; the product was not on the ARTG [40]. The risk came from the adulterant and says nothing about the fruit. The same search found a sponsor-requested cancellation of a "Bitter Melon Powder Capsule" listing on 14/03/2017, with no safety reason given [40].
  • TGA Database of Adverse Event Notifications (DAEN): searched in September 2026 but not completed. The DAEN medicines search did not filter on any term entered by the non-interactive browser, including a control ingredient known to be listed, so the absence of bitter melon reports could not be confirmed. A manual DAEN search is outstanding (gap 4 in section 0.3).

Referral. Anyone with diabetes, anyone pregnant, breastfeeding or trying to conceive, anyone with G6PD deficiency, and anyone giving bitter melon to a child should consult a degree-qualified herbalist with competencies and adequate training in pharmacognosy before use.

Figure F6. Vicine, CID 135413566, C10H16N4O7, M_r 304.3: the pyrimidine glycoside of the seeds that can cause haemolysis
Figure F6. Vicine, CID 135413566, C10H16N4O7, M_r 304.3: the pyrimidine glycoside of the seeds that can cause haemolysis (favism) in people with glucose-6-phosphate dehydrogenase deficiency (ref 5). Structure image from PubChem.
Figure F1h. Ripening and fully ripe fruit: when ripe the fruit turns orange and splits open, exposing seeds coated in a
Figure F1h. Ripening and fully ripe fruit: when ripe the fruit turns orange and splits open, exposing seeds coated in a red aril (left and bottom right); a green fruit of eating stage and a cut fruit (top right). Montage by RoRo from photographs by H. Zell, FlyingToaster, Sengai Podhuvan and Llez, CC BY-SA 3.0, Wikimedia Commons.
Figure F10e. Each safety statement in section 15 with what was seen and the grade of evidence behind it. Drawn by AIP fr
Figure F10e. Each safety statement in section 15 with what was seen and the grade of evidence behind it. Drawn by AIP from (refs 4, 5, 7, 12, 16, 19, 24, 40-43). AIP, 2026.

28 more pages follow · 14,018 words, 42 figures, 11 tables

Still to come in this monograph

  1. 1. Nomenclature
  2. 2. Definition
  3. 3. Source and Australian supply
  4. 4. Macroscopy
  5. 5. Microscopy
  6. 6. Chemistry
  7. 7. Named markers and assay limits
  8. 8. TLC fingerprint
  9. 9. HPLC fingerprint
  10. 10. Identity, purity, adulterants and contaminants
  11. 11. Traditional use
  12. 12. Current practice: use, preparations and doses
  13. 13. Clinical evidence
  14. 14. Pharmacology
  15. 16. Discussion
  16. 17. Conclusions
  17. 18. Research still needed
  18. 19. References
  19. 20. Document control

The rest of this monograph is open to members. One membership covers every members’ feature on the site.

Checking your membership…

Share this articleFacebookXLinkedInWhatsApp

Report an error or submit evidence

If something in this monograph is wrong, out of date or missing, tell us and show us the evidence. Every report is logged under the AIP Evidence Charter and assessed on the evidence alone; the outcome is kept on the record, and an upheld report is corrected with a dated notice on this page.

Open the report form

What kind of report is this?

A report without evidence is still logged, and closed as “no evidence lodged”.

We use your name and email only to acknowledge and answer this report. They are kept with the dispute record, are not shared, and are not published unless you tick the box above. The evidence you send may be published. See the Privacy Policy.

The Institute sets and publishes its own monographs. The catalogue is in Publications.