Bitter Gourd (Pavakkai, Pavakka) for Diabetes Excellent Vegetable, Unreliable Medicine

Answered by Dr. Dante Ruskin · Diabetologist & Intensivist

Eat pavakkai; do not prescribe it. As a vegetable it is close to carbohydrate-free — about 0.9 grams of available carbohydrate per 100 grams (USDA FoodData Central) — so a thoran or kichadi costs your meter almost nothing. As medicine, pooled trials are inconsistent and Cochrane called the evidence insufficient.

The reason is dose and consistency: trial doses are grams of standardised extract, not a curry, and the pooled effects flip with the analysis method. Keep pavakka thoran, theeyal and kichadi on the plate as the low-carbohydrate half of the meal — and swap the daily juice glass for that thoran.

Dr. Dante Ruskin consulting a patient at PPK Hospital, Marthandam
On this page

The claim, and what a diabetologist will concede

No plant in South India carries a heavier reputation. Pagarkai in written Tamil (பாகற்காய்), pavakkai in the way this district actually says it, pavakka in Malayalam (പാവയ്ക്ക, കൈപ്പക്ക in the north of Kerala) — all one plant, Momordica charantia, and half the district believes a morning glass of it does the work of a tablet. Here is the concession worth making before the argument: the belief is not baseless. Bitter gourd genuinely contains charantin, polypeptide-p and vicine, compounds with measurable glucose-lowering activity in cells and animals. Extracts do things in laboratories. What has never followed is the part that matters clinically — a consistent, repeatable drop in human HbA1c at a dose anyone actually eats. That gap between mechanism and outcome is the whole story of this vegetable, and it is why a diabetologist will happily put pavakkai on your plate and refuse to put it on your prescription.

What the trials actually show

Four bodies of evidence, in the order they were published, and they do not agree — which is itself the finding.

The Cochrane review (Ooi and colleagues, 2012) found only four randomised trials of usable quality, together about 479 people with type 2 diabetes, treatment lasting four weeks to three months. Three showed no significant difference against placebo or against glibenclamide and metformin. Its verdict: insufficient evidence to recommend the plant as treatment.

The head-to-head trial most often quoted is Fuangchan’s 2011 Thai study in 129 newly diagnosed patients. Bitter melon at 2,000 mg a day did lower fructosamine from baseline (−10.2 µmol/L) — a real, modest effect — but metformin 1,000 mg a day lowered it more (−16.8 µmol/L). Bitter melon was the weaker drug, at a capsule dose.

A 2023 systematic review in Frontiers in Nutrition pooled nine trials, 414 patients, four to sixteen weeks. Analysed as change scores, HbA1c moved −0.12 percent (95% CI −0.35 to 0.11) and fasting glucose −0.03 — nothing. Its title says the rest: the metabolic effect of Momordica charantia cannot be determined from the available evidence.

A 2025 GRADE-adherent meta-analysis in Metabolism Open pooled 25 trials and did find reductions in fasting glucose and HbA1c — then graded that evidence low certainty for fasting glucose and very low for HbA1c, citing publication bias. When the most favourable pooling downgrades itself, you have a supplement, not a therapy. The ADA’s Standards of Care in Diabetes—2026 reaches the same place in one line: supplementation with herbs or spices is not recommended for glycaemic benefit.

How much carbohydrate is actually in a pavakkai dish?

Almost none, and this is the genuinely good news that gets lost in the argument about medicine. Raw bitter gourd carries about 17 kcal, 3.7 g carbohydrate and 2.8 g fibre per 100 g (USDA FoodData Central) — leaving roughly 0.9 g of available carbohydrate. To assemble the 50 g carbohydrate portion used in a standard glycaemic-index test you would need over five kilos of it, which is why a meaningful GI for pavakkai does not exist and never will. Everything below is our arithmetic from those per-100-g figures plus ordinary Kanyakumari recipe quantities — estimates, not measurements.

Pavakkai on the plateServingAvailable carbohydrateThe number that matters
Raw or steamed pavakkai100 g~0.9 g (USDA)17 kcal — effectively free
Pavakka thoran, coconut + 1 tsp oil~100 g~2 g (est.)~110 kcal, mostly the oil
Pavakka kichadi, curd-based~80 g~3 g (est.)the rice beside it
Pavakka theeyal, roasted coconut + jaggery1 ladle~6–7 g (est.)the jaggery, ~1 tsp
Pagarkai varuval / kondattam, deep-fried~50 g~4 g (est.)8–12 g absorbed oil (est.)
Pavakka juice1 glass~2 g (est.)not carbohydrate — the medication

Is bitter gourd juice safe?

As a vegetable, pavakkai is about as safe as food gets. The juice fad is where the honest cautions sit, and there are four.

Stacking. Bitter preparations taken habitually and in concentrated form can add to what your medication is already doing. On a sulfonylurea (glimepiride, gliclazide) or on insulin, that arithmetic ends in the sweating, trembling afternoon low that gets blamed on a skipped lunch. Read low sugar and hypoglycaemia once, properly, if you drink it daily.

The gut. Vomiting and diarrhoea after bitter-gourd juice are common enough to be documented, and large volumes of concentrated juice have been associated with gastric irritation and bleeding (Frontiers in Pharmacology safety review, 2022).

Seeds and leaves, not fruit. The same review is precise about where the toxicity lives: the seeds carry vicine, which triggers favism in G6PD deficiency; a mildly toxic lectin sits in seed and rind; two children aged three and four went into hypoglycaemic coma after drinking a tea of the leaves and vines. Bitter-gourd preparations are advised against in pregnancy. No serious adverse events are reported for the ordinary unripe-fruit juice.

The wrong gourd. The most serious Indian episode was not karela at all. ICMR’s expert committee investigated 26 hospitalisations and three deaths from bitter bottle-gourd (churakka, sorakkai) juice, sometimes mixed with karela juice — cucurbitacin poisoning, with blood in the vomit. All three who died were over 59 and had lived with diabetes for about twenty years. The committee’s rule is worth memorising: if a gourd juice tastes fiercely bitter, throw it away, and go to hospital if vomiting or diarrhoea follows.

Thoran, theeyal, kichadi or varuval — which form?

Cooking, not chemistry, decides what pavakkai does to your day. Pavakka thoran — sliced thin, turmeric, green chilli, grated coconut, a teaspoon of coconut oil — is the best version of this vegetable for a person with diabetes: bulk, fibre, fat you can count, nothing hidden. Pavakka kichadi, the curd-and-coconut sadya dish, is nearly as good, and is one of the few items on an Onam or Vishu leaf that costs a meter almost nothing; see Onam sadya for diabetics for the rest of the leaf. Pavakka theeyal is excellent food with two additions worth knowing about — roasted coconut and, in most household recipes, a small piece of jaggery to balance the bitterness. It is still a small number; it is simply not zero, which surprises people who came to theeyal believing it was medicine.

Then the fried forms. Pagarkai varuval and sun-dried kondattam are delicious and are, nutritionally, a delivery system for oil — a 2023 review of frying media found snack pieces taking up around 29 percent oil when fried at 180 °C. Nothing about being bitter protects a fried thing from being fried. One or two pieces beside a meal, not a mound.

What to say at the juice stall, at the sadya, and at your review

At a juice stall in Marthandam or Nagercoil. Vegetable juices are routinely sweetened to make them drinkable. Say it before they blend: Tamil, “சர்க்கரை வேண்டாம்” — sarkkarai vendam; Malayalam, “പഞ്ചസാര വേണ്ട” — panchasara venda. If the glass still tastes sweet, it was sweetened. And if it tastes savagely throat-catching rather than ordinarily bitter, do not finish it.

At the hotel or the mess. Ask for the poriyal, not the varuval: Tamil, “பாகற்காய் பொரியல், எண்ணெய் கம்மி” — pagarkai poriyal, ennai kammi. Malayalam, “പാവയ്ക്ക തോരൻ, എണ്ണ കുറച്ച്” — pavakka thoran, enna kurachu.

On sadya day. Take the pavakka kichadi twice if you like it; it is the cheapest item on the leaf. Ask for theeyal with “ശർക്കര കുറച്ച്” — sharkkara kurachu, less jaggery. Then spend your attention where the load actually is: the rice, the payasam, the second serving.

At your review. Bring the bottle, the churnam packet or the juice routine to your consultation and say how long you have been taking it. That one sentence lets your doctor read a low correctly instead of chasing it.

Sources

Everything numbered above is traceable, and where a figure is ours rather than a published measurement it says so in the table. The evidence base: Ooi CP et al., Momordica charantia for type 2 diabetes mellitus, Cochrane Database of Systematic Reviews 2012 (CD007845.pub3); Fuangchan A et al., Journal of Ethnopharmacology 2011, bitter melon versus metformin in 129 patients; the 2023 systematic review in Frontiers in Nutrition concluding the metabolic effect cannot be determined; and the 2025 GRADE-adherent meta-analysis in Metabolism Open. Safety comes from the 2022 Frontiers in Pharmacology review of M. charantia bioactivities, quality control and safety, and from the ICMR expert committee report on bitter bottle-gourd juice in the Indian Journal of Medical Research (2012). Composition is USDA FoodData Central; the guideline line is the ADA Standards of Care in Diabetes—2026, recommendation 5.16; cultivation and season come from the TNAU Agritech horticulture portal, which lists CO 1 and MDU 1 among recommended varieties and July and January as sowing months — the reason pavakkai is cheap and everywhere twice a year here.

Frequently asked questions

Not dependably. The Cochrane review found four low-quality trials in 479 people and concluded the evidence was insufficient. A 2023 pooling of nine trials showed no effect on HbA1c or fasting glucose once change scores were used. A 2025 pooling of 25 trials did find a signal, but graded it low to very low certainty. Good vegetable, unreliable drug.

No. Nothing on this page is a reason to change a prescription — your doctor adjusts doses against your readings, never against a juice habit. In the trial that compared them directly, bitter melon at 2,000 mg a day moved fructosamine less than metformin 1,000 mg a day (Fuangchan, 2011). Substitution is where people get hurt.

There is no studied culinary dose, because the trials used capsules and extracts rather than curry. Treat pavakkai the way you treat any non-starchy vegetable: one generous serving with lunch or dinner, most days, cooked in a little oil. Counting pieces the way you count tablets implies a job the vegetable is not doing.

For most people it is uncomfortable rather than dangerous, but two cautions are real. On a sulfonylurea or insulin, a bitter-juice habit can stack into hypoglycaemia. And any gourd juice that tastes fiercely bitter should be thrown away — ICMR's committee investigated 26 hospitalisations and three deaths from bitter bottle-gourd juice, all three in long-standing diabetics.

As a cooked vegetable, pavakkai is ordinary food. Concentrated preparations are the question. Reviews advise omitting bitter-gourd products in pregnancy on animal-toxicity grounds, and the seeds carry vicine, which can trigger favism in G6PD deficiency. Eat the curry; skip the seed decoctions, the capsules and the daily juice, and mention any product at your review.

Mostly yes, with one honest caveat: theeyal is built on roasted coconut and usually a knot of jaggery, so a ladle carries roughly six or seven grams of available carbohydrate by our estimate — still small. The gourd is not the problem and never was. The rice you serve beside it is what your meter is answering.