Momordica charantia (bitter melon/bitter gourd) is an annual climbing cucurbit widely cultivated and consumed as a vegetable.

Momordica charantia is used ethnomedically, especially for glycemic control. Its accepted botanical identity and broad native range across the tropical/subtropical Old World to the South Pacific are well established in authoritative plant databases, with pantropical introduction beyond the native range.[1]

Phytochemically, bitter melon is rich in cucurbitane-type triterpenoids and triterpene glycosides (e.g., momordicosides, charantosides), along with proteins/peptides (e.g., polypeptide‑p; insulin-receptor-binding peptides such as mcIRBP-19 in certain standardized extracts), phenolics, and micronutrients; “charantin” is best treated analytically as a mixed marker comprising two steroidal glucosides (β-sitosteryl glucoside and stigmasteryl glucoside).[2]

Mechanistic pharmacology is plausible for antihyperglycemic effects through multiple convergent pathways (e.g., modulation of insulin signaling and AMPK, inhibition of carbohydrate-digesting enzymes, effects on pancreatic β-cell function and glucagon dynamics), but much of the mechanistic evidence remains in vitro/animal-based and depends strongly on plant part, cultivar, and extraction/standardization.[3]

Human evidence for glycemic benefit is mixed but, when pooled in recent meta-analyses of randomized controlled trials (RCTs), suggests a modest reduction in fasting blood glucose and HbA1c. A 2019 meta-analysis reported mean differences versus placebo of −0.72 mmol/L fasting plasma glucose, −1.43 mmol/L postprandial glucose, and −0.26% HbA1c, while emphasizing low-to-very-low quality evidence and moderate-to-high risk of bias across included trials.[4] A 2024 meta-analysis reported HbA1c reduction of ~−0.38% with significant heterogeneity for fasting glucose, again rating evidence quality as low/very low for key endpoints.[5] Individual RCTs range from null findings (e.g., a capsule trial with no significant HbA1c effect) to statistically significant subgroup effects for a standardized mcIRBP‑19-containing extract (600 mg/day for 3 months) in participants failing to meet glycemic targets on medications (HbA1c drop from 8.0% to 7.5% in a prespecified subgroup).[6]

Safety signals in trials are generally mild (gastrointestinal complaints, dizziness), but clinically relevant risks include additive hypoglycemia when combined with antidiabetic drugs, uncertainty about long-term use, and pregnancy avoidance given reported miscarriage risk and insufficient safety data.[7] From a regulatory perspective, bitter melon is widely consumed as food; as a supplement (in jurisdictions such as the United States), products are not preapproved by regulators for effectiveness and generally are not reviewed for safety prior to marketing, which elevates the importance of quality control, standardization, and third-party testing.[8]

Botanical identity, taxonomy, morphology, and diversity

Taxonomically, Momordica charantia is placed in the family Cucurbitaceae and is accepted under the authority of major plant databases, with Linnaean publication (1753) widely cited.[9] Its taxonomy records in curated resources (e.g., GRIN-Global and NCBI Taxonomy) document its placement and common-name usage (bitter melon, bitter gourd, balsam pear) and are suitable anchors for unambiguous botanical identity in research and commerce.[10]

Morphologically, bitter melon is a climbing annual vine with tendrils and yellow flowers; the fruit is characteristically warty/ridged and typically harvested unripe (green) for culinary use, with ripening associated with orange coloration and arillate red seeds in many cultivars. These descriptive traits are consistent with botanical and applied reviews, including recent scholarly summaries that emphasize its herbaceous vine habit and distinctive bitter fruit.[11]

Intraspecific diversity matters for both agronomy and bioactivity. Authoritative flora-style treatments and taxonomic notes recognize forms often treated as “wild” or less-domesticated within the species complex (e.g., forms that are widespread in the Old World and long naturalized in parts of the tropics), underscoring that “bitter melon” in commerce can refer to materials with different domestication histories and, potentially, different chemistry.[12] This diversity is a recurring theme in phytochemical literature (large numbers of cucurbitane triterpenoids and glycosides reported) and is one reason why “standardization to marker compounds” is frequently recommended yet inconsistently implemented.[13]

Distribution, cultivation, and food uses

Major plant databases converge on a broad tropical/subtropical distribution: authoritative records describe a native range spanning the tropical/subtropical Old World to the South Pacific and widespread introduction beyond that native range, yielding a pantropical cultivated/naturalized footprint.[14] In practical terms, this pattern fits a crop adapted to warm-season production and widely integrated into tropical and subtropical food systems, with historical and ongoing spread through cultivation.[15]

Cultivation requirements are frequently summarized as “warm-climate vine crop” production with trellising to support climbing growth and improve fruit quality/harvest efficiency; propagation is typically by seed, and harvest is usually targeted at the immature-green stage for culinary markets. While these broad principles are stable, specific agronomic parameters (temperature optima, precise soil pH ranges, planting densities, cultivar-specific days-to-harvest) vary materially by cultivar and region and are often not reported in clinical papers because interventions are based on processed extracts rather than whole-crop production systems. Consequently, many clinical sources leave these agronomic specifics unspecified.[16]

Nutritionally, bitter melon is a low-energy, high-water vegetable. A scholarly review figure explicitly attributes per-100 g composition to U.S. nutrient database identifiers (FDC ID 168393; NDB 11024), reporting approximately 94% water, ~3.7% carbohydrate, ~1% protein, and ~0.17% fat, alongside meaningful micronutrient presence (notably vitamin C and folate in the cited figure) and minerals including potassium and others.[17] Because many nutrition values are reported through secondary renderings or figures, granular nutrient quantities (complete vitamin/mineral panels) are often unspecified in the primary clinical RCT reports themselves.[18]

Culinarily, bitterness drives both preparation strategy and product selection. Common techniques include salting and rinsing, blanching, stuffing, and cooking with fat/protein-rich components to balance bitterness; these practices reflect an implicit “dose modulation” of bitter principles (triterpenoids and glycosides) that may differ substantially from standardized extract dosing in trials.[19]

Traditional and ethnomedical uses are geographically broad and often diabetes-centered, but also include gastrointestinal complaints, infections, and inflammatory conditions. Reviews of traditional use emphasize diabetes management across multiple cultural contexts and also mention non-glycemic uses (e.g., wound applications).[20]

Phytochemistry and analytical markers

Bitter melon’s phytochemistry is dominated by cucurbitane-type triterpenoids and their glycosides, which are repeatedly identified as major bitter principles and as a central chemical class underlying bioactivity claims. Reviews and primary isolation papers document both chemical diversity and the continued discovery of new glycosides.[21]

A key cornerstone in the glycoside literature is the isolation and structural elucidation of momordicosides A and B from seeds, with structures determined by spectral/chemical evidence and X-ray analysis. This early, high-quality natural products work underpins later analytical and mechanistic research on triterpene glycosides.[22] Subsequent studies and reviews expand the catalog substantially (multiple momordicoside series, charantosides, goyaglycosides), illustrating why extraction method and plant part can shift the “active profile.”[23]

“Charantin” is frequently cited in supplement marketing and some scientific discussions as an antidiabetic principle, but analytically it should be treated as a mixed marker: a 2021 analytical paper states charantin is a mixture of stigmasterol glucoside and β-sitosterol glucoside and develops quantitative methods based on this interpretation.[24] This matters for quality control because “charantin content” depends on whether both components are quantified consistently and whether the botanical material/extraction enriches these sterol glucosides relative to triterpene glycosides.[25]

Protein/peptide constituents add a second axis of complexity. A subset of products and mechanistic papers focus on insulin-receptor binding peptides (mcIRBP-19) and historical “polypeptide‑p” (sometimes described as plant insulin) with clinical research emphasizing subcutaneous administration in earlier work and oral administration for certain standardized preparations in more recent RCTs.[26]

Major constituents, reported activities, and evidence type

Pharmacology and mechanisms of action

Mechanistic proposals for bitter melon’s antihyperglycemic effects fall into three overlapping domains: (1) effects on carbohydrate handling (intestinal enzyme inhibition and postprandial glucose blunting), (2) endocrine modulation (insulin secretion/sensitivity and glucagon dynamics), and (3) cellular energy and lipid signaling (e.g., AMPK-related pathways). This pluralism is supported by compound discovery and signaling-focused studies on triterpenoids and glycosides, as well as by clinical observations (e.g., changes in post-load glucose and glucagon dynamics in prediabetes).[35]

A practical implication is that “mechanism” likely depends on formulation: whole-vegetable intake emphasizes mixed dietary fibers/micronutrients and bitter glycosides; ethanol extracts may enrich specific glycoside families; peptide-standardized preparations explicitly target insulin-receptor-related signaling; and injected protein preparations (polypeptide‑p in historical reports) bypass oral bioavailability constraints.[36]

flowchart TD
A[Ingestion: bitter melon food or extract] –> B{Bioactive fractions}
B –> C1[Cucurbitane triterpenoids & glycosides] B –> C2[Sterol glucosides marker: “charantin”] B –> C3[Peptides/proteins: mcIRBP-19, polypeptide-p] B –> C4[Phenolics & micronutrients]

C1 –> D1[↓ α-glucosidase/α-amylase activity] D1 –> E1[↓ postprandial glucose excursions]

C1 –> D2[AMPK-related signaling modulation] D2 –> E2[↑ glucose uptake / ↓ hepatic glucose output (hypothesized)]

C3 –> D3[Insulin receptor binding / insulin-like signaling (product-specific)] D3 –> E3[Improved glycemic indices in some RCTs]

C1 –> D4[β-cell function & insulin secretion (reported in some trials)] D4 –> E4[↑ insulin secretion indices (select endpoints)]

C4 –> D5[Antioxidant/anti-inflammatory milieu (supportive)] D5 –> E5[↓ oxidative stress markers (mostly preclinical)]

E1 –> F[Net effect: modest glycemic improvement (variable)] E2 –> F
E3 –> F
E4 –> F

Mechanistic confidence varies by endpoint. For antihyperglycemic activity, confidence is moderate for “biological plausibility” but low-to-moderate for “predictable clinical effect,” because RCTs show heterogeneity and meta-analyses rate evidence quality as low/very low for several outcomes.[37] For antioxidant and anti-inflammatory effects, confidence is low-to-moderate for preclinical activity but low for demonstrated clinical benefit in cardiometabolic endpoints, given mixed metabolic syndrome results and limited biomarker-driven RCTs.[38]

Claims for antiviral and anticancer benefits are supported mainly by in vitro and animal models: antiviral activity has historical protein-level in vitro evidence (e.g., MAP30), and anticancer narratives cite apoptosis, AMPK/mTOR modulation, and lipid-metabolism effects in models; however, rigorous human clinical efficacy trials for antiviral or anticancer indications are largely unspecified in the accessible clinical evidence base and are not established by diabetes-focused RCTs.[39]

Lipid-lowering effects appear, at best, selective and modest in pooled analyses: one meta-analysis of T2DM RCTs reported significant total cholesterol reduction (WMD ≈ −0.38 mmol/L) but nonsignificant effects on triglycerides, HDL, and LDL overall, with evidence graded very low for total cholesterol and very low/low for several lipid endpoints.[5]

Gastrointestinal effects occupy two roles: (1) traditional indications (e.g., GI complaints) in ethnomedical narratives and (2) tolerability/adverse events (e.g., dyspepsia) reported in trials; the net clinical direction is therefore context-specific and not uniformly beneficial.[40]

Human clinical evidence

Human studies evaluate heterogeneous interventions: dried whole-fruit powder tablets, capsule extracts with unreported marker standardization, add-on use in medicated T2DM, and more standardized peptide-containing products designed to target insulin receptor signaling. This heterogeneity is repeatedly flagged as a central limitation in evidence synthesis.[37]

Summary of pooled effects from meta-analyses

Across T2DM trials, a 2019 systematic review/meta-analysis reported statistically significant pooled improvements versus placebo for fasting plasma glucose (−0.72 mmol/L; 95% CI −1.33 to −0.12), postprandial glucose (−1.43 mmol/L; 95% CI −2.18 to −0.67), and HbA1c (−0.26%; 95% CI −0.49 to −0.03), with follow-up durations 4–16 weeks and overall moderate-to-high risk of bias; the authors explicitly downgraded the evidence to low/very low and emphasized sparse safety data.[4]

A 2024 meta-analysis restricted to RCTs in T2DM similarly reported significant reductions in fasting blood glucose (WMD −0.85 mmol/L; 95% CI −1.44 to −0.26), postprandial glucose (WMD −2.28 mmol/L; 95% CI −3.35 to −1.21), and HbA1c (WMD −0.38%; 95% CI −0.53 to −0.23), plus a significant reduction in total cholesterol (WMD −0.38 mmol/L; 95% CI −0.70 to −0.07); however, evidence quality was judged low for HbA1c and very low for total cholesterol, with heterogeneity remaining nontrivial for some endpoints.[5]

In contrast, a 2024 systematic review of metabolic syndrome parameters concluded that bitter melon supplementation could not be confirmed to significantly improve common metabolic syndrome indices, reporting largely null standardized mean differences across glycemia and lipid outcomes and highlighting contradictory findings across studies.[41] Differences in populations (metabolic syndrome vs diagnosed T2DM), interventions, and analytical choices (e.g., endpoint selection and heterogeneity handling) plausibly explain part of this discordance, but the practical takeaway is that confidence in consistent cardiometabolic benefit remains limited.[42]

A 2025 systematic review/meta-analysis using GRADE (prediabetes and T2D combined) reported small-to-moderate improvements across glycemic indices (e.g., effect sizes on fasting glucose and HbA1c), but the underlying limitation—trial heterogeneity and variable risk of bias—remains a constraint on high-confidence recommendations.[43]

Key RCTs and human studies

Interpretation of clinical magnitude. The HbA1c lowering magnitude seen in meta-analyses (~0.26–0.38%) is clinically modest relative to first-line pharmacotherapy in many contexts and may fall below thresholds that would change management for many patients unless targeted to specific subgroups or used as adjunctive lifestyle support. The evidence base also contains clear null or borderline-significant trials, reinforcing that response is variable and likely depends on formulation, baseline glycemia, concomitant therapy, and adherence.[54]

Confidence levels by claimed benefit (human data).
For glycemic control in T2DM (fasting glucose, postprandial glucose, HbA1c), the direction of effect is supported by meta-analyses, but quality is low/very low and heterogeneity is meaningful; confidence is low-to-moderate for modest benefit.[55] For lipid lowering, evidence suggests a possible small reduction in total cholesterol with no consistent effect on TG/HDL/LDL; confidence is low.[56] For antiviral, anticancer, and broad anti-inflammatory clinical outcomes, human efficacy evidence is largely unspecified; confidence is low despite preclinical plausibility.[57]

Safety, contraindications, interactions, regulation, and research gaps

Safety and toxicity

Across diabetes-focused RCTs and meta-analyses, serious adverse events are uncommon in reporting, but adverse-event capture is often sparse, and many trials are short (weeks to a few months), limiting inference about long-term safety.[58] In a 12-week prediabetes trial, dizziness occurred significantly more often in the bitter melon extract group; other events (fatigue, pruritus, dyspepsia) were observed with variable frequency, and one acute hepatitis case was described but attributed to other herbal medicine use shortly before final blood sampling.[53] These patterns support a general tolerability profile that is “often acceptable,” but not “side-effect free,” particularly at multi-gram daily dosing.[59]

A clinically important risk is hypoglycemia, especially when bitter melon products are combined with glucose-lowering medications. Some trials reported hypoglycemia events (e.g., in the capsule trial requiring further study), and meta-analyses emphasize the need for standardized formulations and better safety outcomes reporting.[60] Because supplement products can vary widely in composition, real-world hypoglycemia risk is not reliably predictable from any single trial’s extract.[61]

Contraindications and special populations

Pregnancy avoidance is commonly recommended in clinical and review literature due to miscarriage risk signals and insufficient safety data; a systematic review focusing on pregnancy and lactation highlights gaps and the need for careful risk assessment.[62] For lactation and pediatric use, robust, formulation-specific safety data are often unspecified, and some reviews caution against use in children based on reports of serious hypoglycemia events.[40]

Interactions

Potential interactions of greatest practical relevance involve additive glucose lowering with antidiabetic drugs; therefore, monitoring and medication adjustment may be needed if bitter melon extracts are introduced. Because many supplement products are multi-ingredient or variably standardized, interaction risk is frequently unspecified at the product level in trials.[63]

Regulatory status and quality control

In the United States supplement framework, regulators generally do not preapprove dietary supplements for effectiveness and typically do not evaluate safety prior to marketing (with limited exceptions such as “new dietary ingredients”), making post-market surveillance and manufacturer responsibility central.[64] This regulatory reality increases the importance of: (1) standardization to chemically meaningful markers, (2) validated analytical methods, and (3) independent testing for identity, contaminants, and label accuracy.[65]

For bitter melon specifically, the chemistry strongly argues for multi-marker or profile-based quality control rather than reliance on a single named principle. Even for “charantin,” the literature frames it as a two-component mixture and discusses its use as an HPLC target, underscoring that “charantin mg” on labels can be analytically ambiguous unless methods and reference standards are specified.[25] For peptide-characterized products (mcIRBP-19-BGE), quality control must also include peptide identity/quantity and batch consistency, because clinical evidence is tied to the specific standardized preparation rather than to bitter melon generically.[66]

Research gaps and future directions

The evidence base repeatedly identifies the same bottlenecks:

Standardization and comparability: Trials and meta-analyses emphasize that differing preparations (whole fruit, aqueous extracts, ethanol extracts, peptide-enriched products) hinder synthesis; future trials should report marker profiles (e.g., defined triterpenoid/glycoside panels and, where relevant, peptide identity) and include batch certificates.[67]

Dose–response and pharmacokinetics: Even when effects appear, dose-response relationships and exposure–response metrics are often unspecified, particularly for cucurbitane glycosides whose absorption and metabolism may be limited or highly variable.[68]

Long-term safety and clinically meaningful endpoints: Many trials are ≤12–16 weeks; long-term effects (sustained HbA1c changes, microvascular outcomes, incident diabetes prevention in prediabetes cohorts) remain largely unspecified.[69]

Population targeting and subgroup validation: The mcIRBP-19-BGE trial suggests subgroup responsiveness among those failing to respond to medications, but such subgroup claims require prospective replication with adequate power and prespecified hypotheses.[66]

Beyond glycemia: Anticancer and antiviral narratives remain mostly preclinical; translation requires careful phase-appropriate human research, with clear endpoints and safety characterization.[70]

Key references

Bara LV, et al. 2025. Functional health benefits and uses (review).
Çiçek SS, et al. 2022. Diabetes-related bioactivities and cucurbitane triterpenoids (review summary).[71] Dans AML, et al. 2007. Capsule preparation trial in type 2 diabetes.[45] Desai S, et al. 2021. Isolation/quantification of charantin by HPLC-DAD; charantin as sterol glucoside mixture.[24] Fuangchan A, et al. 2011. Fruit extract doses vs metformin; fructosamine endpoint.[47] Harinantenaina L, et al. 2006. Constituents and antidiabetic principles.[31] Jia S, et al. 2017. Functional components and activities (review).[72] Kim B, et al. 2022. Prediabetes RCT; OGTT and glucagon effects; adverse events.[53] Kim SK, et al. 2020. T2DM RCT; fasting glucose improved, HbA1c unchanged.[51] Khan MF, et al. 2019. Pregnancy/lactation safety review.[73] Okabe H, et al. 1980. Isolation and X-ray-based structural determination of momordicosides A and B.[22] Perera WH, et al. 2021. Cucurbitane-type glycosides; anti-inflammatory/antidiabetic activity.[74] Peter EL, et al. 2019. Systematic review/meta-analysis; modest glycemic reductions; low-quality evidence.[4] Sur S, Ray RB. 2023. Review including USDA-attributed nutritional summary and phytochemical overview.[75] Tan MJ, et al. 2008. Antidiabetic activities of triterpenoids; mechanistic signaling.[28] Yang Y‑S, et al. 2022. mcIRBP‑19-BGE RCT; subgroup HbA1c and FBG reductions.[32] Zhang X, et al. 2024. Meta-analysis of RCTs in T2DM; HbA1c and lipid effects with low/very-low evidence ratings.[5] U.S. Food and Drug Administration[76]. 2024. Q&A and consumer information on dietary supplements (DSHEA; no premarket approval).[77] National Center for Complementary and Integrative Health[78]. Using dietary supplements wisely; regulatory overview and consumer guidance.[79]

[1] [9] [14] https://powo.science.kew.org/taxon/urn%3Alsid%3Aipni.org%3Anames%3A293413-1/general-information

https://powo.science.kew.org/taxon/urn%3Alsid%3Aipni.org%3Anames%3A293413-1/general-information

[2] [22] Studies on the Constituents of Momordica charantia L. I. …

https://www.jstage.jst.go.jp/article/cpb1958/28/9/28_9_2753/_article/-char/ja/?utm_source=chatgpt.com

[3] [28] [35] Antidiabetic Activities of Triterpenoids – Bitter Melon

https://www.cell.com/fulltext/S1074-5521%2808%2900082-3?utm_source=chatgpt.com

[4] [37] [54] [55] [58] [67] [69] [78] https://pubmed.ncbi.nlm.nih.gov/30385422/

https://pubmed.ncbi.nlm.nih.gov/30385422/

[5] [56] [61] https://pmc.ncbi.nlm.nih.gov/articles/PMC11112315/

https://pmc.ncbi.nlm.nih.gov/articles/PMC11112315/

[6] [45] [48] [60] https://pubmed.ncbi.nlm.nih.gov/17493509/

https://pubmed.ncbi.nlm.nih.gov/17493509/

[7] [20] [40] [50] [53] [59] https://pmc.ncbi.nlm.nih.gov/articles/PMC10050654/

https://pmc.ncbi.nlm.nih.gov/articles/PMC10050654/

[8] [64] [65] [77] https://www.fda.gov/food/information-consumers-using-dietary-supplements/questions-and-answers-dietary-supplements

https://www.fda.gov/food/information-consumers-using-dietary-supplements/questions-and-answers-dietary-supplements

[10] https://apg.pir.sa.gov.au/gringlobal/taxonomydetail?id=24520

https://apg.pir.sa.gov.au/gringlobal/taxonomydetail?id=24520

[11] [17] [23] [29] [34] [70] [75] pdfs.semanticscholar.org

https://pdfs.semanticscholar.org/2378/b7fc0971c1b512edf1539fbaa23a247d668b.pdf

[12] https://www.worldfloraonline.org/taxon/wfo-0000509811%3Bjsessionid%3D7396262442F07D7DD4A40F82C92AAF19

https://www.worldfloraonline.org/taxon/wfo-0000509811%3Bjsessionid%3D7396262442F07D7DD4A40F82C92AAF19

[13] [19] [21] https://www.sciencedirect.com/science/article/pii/S030881462504333X?dgcid=rss_sd_all

https://www.sciencedirect.com/science/article/pii/S030881462504333X?dgcid=rss_sd_all

[15] https://www.worldfloraonline.org/taxon/wfo-0000376916%3Bjsessionid%3D1462DC187D649E0778E06F266C9D0BBC

https://www.worldfloraonline.org/taxon/wfo-0000376916%3Bjsessionid%3D1462DC187D649E0778E06F266C9D0BBC

[16] [18] [26] [32] [63] [66] [76] https://pmc.ncbi.nlm.nih.gov/articles/PMC8788657/

https://pmc.ncbi.nlm.nih.gov/articles/PMC8788657/

[24] [25] https://www.sciencedirect.com/science/article/abs/pii/S0308814620325796

https://www.sciencedirect.com/science/article/abs/pii/S0308814620325796

[27] https://pmc.ncbi.nlm.nih.gov/articles/PMC12765105/

https://pmc.ncbi.nlm.nih.gov/articles/PMC12765105/

[30] [46] Charantoside L, A New Cucurbitane-Type Glycoside from …

https://pdfs.semanticscholar.org/d9f4/834f50d4f69c329e32f8849b5f69d73274fd.pdf?skipShowableCheck=true&utm_source=chatgpt.com

[31] Momordica charantia constituents and antidiabetic …

https://europepmc.org/article/med/16819222?utm_source=chatgpt.com

[33] [39] [57] https://pubmed.ncbi.nlm.nih.gov/1699801/

https://pubmed.ncbi.nlm.nih.gov/1699801/

[36] https://www.mdpi.com/1660-4601/11/2/2328

https://www.mdpi.com/1660-4601/11/2/2328

[38] [41] [42] https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2023.1200801/full

https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2023.1200801/full

[43] Effects of Momordica charantia L. supplementation on …

https://www.cell.com/heliyon/fulltext/S2405-8440%2824%2907157-3?utm_source=chatgpt.com

[44] [47] [68] https://pubmed.ncbi.nlm.nih.gov/21211558/

https://pubmed.ncbi.nlm.nih.gov/21211558/

[49] [52] https://pubmed.ncbi.nlm.nih.gov/29431598/

https://pubmed.ncbi.nlm.nih.gov/29431598/

[51] https://pubmed.ncbi.nlm.nih.gov/32951763/

https://pubmed.ncbi.nlm.nih.gov/32951763/

[62] [73] https://pmc.ncbi.nlm.nih.gov/articles/PMC3005383/

https://pmc.ncbi.nlm.nih.gov/articles/PMC3005383/

[71] https://europepmc.org/article/med/35656300

https://europepmc.org/article/med/35656300

[72] Recent Advances in Momordica charantia: Functional … – PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC5751158/?utm_source=chatgpt.com

[74] https://www.mdpi.com/1420-3049/26/4/1038

https://www.mdpi.com/1420-3049/26/4/1038

[79] https://www.nccih.nih.gov/health/using-dietary-supplements-wisely

https://www.nccih.nih.gov/health/using-dietary-supplements-wisely