Bitter melonBotanical

• Blood sugar in type 2 diabetes and prediabetesPossibly helps

  This is the best-studied use, and the evidence pulls in both directions.
  
  The 2012 Cochrane review of 4 RCTs (479 patients) found no significant difference versus placebo, and none versus metformin or glibenclamide [15]. Its verdict was insufficient evidence [15].
  
  A 2014 meta-analysis in Nature's Nutrition and Diabetes likewise found no significant pooled effect on HbA1c or fasting plasma glucose [17].
  
  On the other side, a 2020 Korean RCT gave 2,380 mg/day for 12 weeks and saw fasting glucose fall significantly versus placebo (p=0.014), with HbA1c unchanged [2].
  
  Fuangchan's 2011 trial found 2,000 mg/day lowered fructosamine, though less than 1,000 mg/day of metformin [6]. A review of 13 trials (892 participants) found reductions in A1c and fasting and post-meal glucose in several of them [8]. ^[15]

• Cochrane versus the individual trials — an open disagreementPossibly helps

  We are flagging this rather than resolving it.
  
  Two pooled analyses, Cochrane 2012 and a 2014 meta-analysis, found no significant glycemic effect, while several individual RCTs did [15][17].
  
  Reviewers attribute the gap to non-standardized preparations, short trials, and small samples rather than one dataset being wrong [8]. The literature reads it as an open evidence gap, not a settled question [8]. ^[15]

• Weight loss and body compositionNot shown to help

  A 2024 systematic review and meta-analysis of 10 RCTs (448 participants) found no significant improvement in BMI, body weight, waist circumference, or body fat [9].
  
  A subgroup analysis hinted at a BMI reduction at doses of 2,000 mg/day or less. That is a signal inside a null result, not a finding on its own [9]. ^[9]

• Knee osteoarthritisPossibly helps

  One small randomized trial suggested symptom improvement in knee osteoarthritis [1].
  
  That is a single small study with no replication behind it [1]. ^[1]

• Anticancer and antiviral activityPossibly helps

  Both come from laboratory and animal work — cell cultures and rodents, not patients [1].
  
  No human trial has tested either use, so there is nothing here that carries over to people [1]. ^[1]

• Fresh fruit and juiceCulinary use. The serious case reports involve concentrated juice or homemade extract, not ordinary cooking

  Elemental content

  Whole food. Charantin sits highest in the flesh, about 0.16 mg per gram [11]

  Absorption

  Dosed at 50-100 mL/day of juice in diabetes studies [10]

  Best for

  Culinary use. The serious case reports involve concentrated juice or homemade extract, not ordinary cooking [1]

• Standardized encapsulated extractThe form most clinical trials used.

  Elemental content

  Sold as standardized, but brands are not standardized to each other

  Absorption

  Dosed at 100-200 mg three times daily, up to 2,000-2,500 mg/day in glycemic trials [10][15]

  Best for

  The form most clinical trials used. Note that no head-to-head human comparison exists across forms [8]

• Dried fruit and dry powder capsulesTraditional preparations. Hard to line up against extract milligrams

  Elemental content

  Whole dried fruit, ground; no standardized active content

  Absorption

  Dosed at 3-15 g/day of powder, or 5 g of dried fruit three times daily [10]

  Best for

  Traditional preparations. Hard to line up against extract milligrams [10]

• Tea from sliced fresh or dried fruitCommon in practice, absent from the trial record — no reviewed trial tested tea

  Elemental content

  Water infusion; active content undefined

  Absorption

  Not characterized in the literature reviewed

  Best for

  Common in practice, absent from the trial record — no reviewed trial tested tea [8]

• Hypoglycemia and rare serious events

  Case reports include hypoglycemic coma and convulsions in children [5].
  
  Others describe an atrial fibrillation episode in a young man after days of drinking the juice, and acute gastric ulceration requiring transfusion after a concentrated homemade extract [1].
  
  Two further case reports describe acute interstitial nephritis — one after an Ayurvedic formulation containing bitter melon, one after the extract alone [1]. ^[5]

• Common effects

  Gastrointestinal upset, diarrhea, abdominal pain, headache, and dizziness are the routine complaints [14].
  
  They are the reason dosing guidance splits the daily amount across meals [10]. ^[14]

• Pregnancy, breastfeeding, and childbearing age

  Animal studies show bitter melon and isolated fruit proteins can cause developmental abnormalities [14]. It also has documented emmenagogue and abortifacient effects [14].
  
  Government and clinical sources say avoid it in pregnancy, and in women of childbearing age not using effective contraception [14]. Breastfeeding falls under the same advice, on insufficient data [14]. ^[14]

• G6PD deficiency

  The seeds contain vicine, which can set off a favism-like hemolytic reaction in people with glucose-6-phosphate dehydrogenase deficiency [1].
  
  That group should avoid bitter melon altogether [1]. ^[1]

• Liver

  The NIH LiverTox database gives bitter melon a likelihood score of E — unlikely to cause clinically apparent liver injury [14].
  
  Despite wide use, no published case has been attributed to it. One trial recorded a single patient's ALT rising from 82 to 102 U/L, which resolved after stopping [14]. ^[14]

• Insulin and oral diabetes drugs

  The clearest risk on this page. Bitter melon lowers blood glucose on its own, so stacking it with insulin, sulfonylureas, or metformin can push blood sugar too low [1].
  
  Memorial Sloan Kettering's integrative-medicine reference carries an outright do-not-take warning for people on insulin or diabetes medication [1]. ^[1]

• CYP2C9 and P-glycoprotein (laboratory only)

  Bitter melon extract inhibited the CYP2C9 enzyme and P-glycoprotein transport in vitro [1].
  
  In principle that could raise levels of drugs cleared by those routes. Clinical significance in humans has not been established [1]. ^[1]

Bitter melon is unusual among supplements here: the fruit is both the food and the source of the extract [11].

Fruit and seeds carry the highest concentration of charantin, polypeptide-p, and vicine, and the fruit shows the most pronounced antidiabetic activity in research so far [11]. Leaves and vines hold related compounds such as kuguacins, far less studied in people [11].

In the kitchen it is eaten raw in Vietnam, and more often stir-fried, curried, or stuffed and cooked with meat [20].

Cooks usually pre-salt the slices and squeeze out the juice, or parboil briefly, to take the edge off the bitterness [20]. Strong partners — garlic, black bean sauce, chili, coconut milk — do the rest [20].

Store the whole melon loose in a paper or plastic bag in the fridge, and use it within about 3-5 days [19]. Slice it only right before cooking [19].

No trial has compared timing strategies for bitter melon — with food versus fasting, morning versus evening [13].

What trial protocols do consistently is split the daily amount. A 2020 study of the bitter melon peptide mcIRBP-19 used 300 mg capsules twice daily, before lunch and before dinner [13].

Other designs used two capsules three times a day, and professional dosing guidance follows the same split-with-meals pattern [10].

Taking it with food is commonly suggested to blunt the abdominal discomfort and nausea it can cause [13]. Read that as inference from protocol design, not the result of a timing study [13].