Pandan: Blood Sugar and Metabolic Claims Examined

“Pandan tea for diabetes” is by a wide margin the most-searched thing about this plant. It is also the least supported. That combination — heavy demand, thin evidence — is exactly the situation in which health writing tends to go wrong, so this page is deliberately structured around what the studies actually did rather than what they are said to have shown.

The short version, stated plainly up front: the pandan blood-sugar literature consists of test-tube enzyme assays, cell-culture work, rodent experiments, and one small human study whose glucose result went the wrong way. There has never been a randomised controlled trial of pandan in people with diabetes or prediabetes. There is no established dose, no known duration of effect, no data on drug interactions, and no evidence that drinking pandan tea changes anyone's HbA1c. Pandan tea is not a treatment for diabetes and must not replace prescribed medication.

That is not a reason to avoid pandan. It is a reason to keep it in the category it belongs to — a superb aromatic leaf — and to be sceptical of anyone selling capsules of it.

Table of Contents

1. Where the Claim Comes From
2. A Species and Plant-Part Problem First
3. The 1998 Rat Study, Read Carefully
4. The Only Human Study — and Its Surprise
5. Alpha-Glucosidase and Alpha-Amylase Inhibition
6. Why In-Vitro Enzyme Inhibition Rarely Survives Contact With a Human
7. The Wider Rodent and Cell Literature
8. What a Convincing Study Would Look Like
9. The Real Risk: Interaction and Substitution
10. If You Want to Drink Pandan Tea Anyway
11. What Actually Moves Blood Sugar
12. Evidence Tiers for This Page
13. Key Research Papers
14. Connections

Where the Claim Comes From

The pandan-for-diabetes idea has three separate roots, and they get tangled together in most retellings.

1. A genuine traditional thread — but a narrow one. In Thai household medicine, the root of the pandan plant has a small traditional reputation as a diuretic and for “heat” conditions including diabetes. Note the part: root, not leaf. Traditional Malay and Vietnamese use of the leaf is overwhelmingly as a pleasant cooling drink and a kitchen aromatic, not as an antidiabetic remedy. The traditional record for pandan leaf and blood sugar specifically is much weaker than the internet suggests.

2. One purified compound in normal rats. A 1998 Thai paper isolated a hypoglycaemic constituent from pandan root and showed it lowered glucose in healthy rats. This is the single most-cited primary source in the whole edifice, and it is discussed in detail below.

3. Laboratory enzyme assays. Pandan leaf extracts inhibit alpha-glucosidase — and in some reports alpha-amylase — in the test tube. This is a real, reproducible in-vitro observation, and it is the mechanism most often quoted. It is also the finding most often overstated, because in-vitro enzyme inhibition is one of the easiest results to obtain from almost any plant extract and one of the hardest to translate into a person.

Stack those three and you get a story that sounds well-founded. Look at each one and the picture changes considerably.

A Species and Plant-Part Problem First

Before any result can be interpreted, two questions have to be answered about it: which plant? and which part? In the pandan literature these are not pedantic questions — they are the difference between studies.

• Pandanus amaryllifolius Roxb. — the fragrant culinary leaf of Southeast Asia. Effectively sterile, propagated only from offshoots, spineless-margined, intensely aromatic from 2-acetyl-1-pyrroline. This is the plant people mean by “pandan” and the one in the tea bags.
• Pandanus odorus Ridl. — a name used in the older pharmacology literature and generally treated as a synonym of P. amaryllifolius. If you search only the current name, you will miss the foundational papers. This is a live problem for anyone trying to find the literature.
• Pandanus tectorius — the Pacific hala or beach screwpine, a large fruiting coastal tree with spiny leaves. Its fruit is a traditional food across Micronesia and Polynesia and has its own separate composition and its own separate small literature. Findings about hala fruit are frequently and wrongly cited as findings about pandan leaf.
• Pandanus odorifer (also seen as P. fascicularis) — the South Asian screwpine whose male flowers are distilled into kewra water. A different plant, a different part, a different aroma chemistry, and a different literature again.
• Pandanus utilis — ornamental and fibre use. Not a food or medicine plant.

And the parts: leaf, root, and fruit are not interchangeable. The best-known hypoglycaemic finding in the genus came from a root extract. Nobody drinks pandan root tea. The tea bags, the bundles at the market and the leaf juice in a chiffon cake are all leaf.

Throughout this page, the species and part used are stated for every study described. If a health article about pandan does not tell you which species and which part a study used, it has not read the study.

The 1998 Rat Study, Read Carefully

The foundational pharmacology paper is Peungvicha and colleagues, “4-Hydroxybenzoic acid: a hypoglycemic constituent of aqueous extract of Pandanus odorus root,” published in the Journal of Ethnopharmacology in 1998. A group at Mahidol University in Bangkok used activity-guided fractionation — splitting an extract, testing each fraction, following the activity — and arrived at 4-hydroxybenzoic acid as the compound responsible. Given orally at 5 mg/kg, it lowered blood glucose and raised serum insulin and liver glycogen.

That is a competent, honest piece of pharmacology. It is also nearly the opposite of what people cite it for. Four things to notice:

1. The plant part was the root. Not the leaf. The root is not what anyone consumes.
2. The agent was a single purified compound. Not a tea, not a leaf infusion, not a whole extract. Activity-guided fractionation exists precisely because whole extracts are usually too weak or too messy to be informative.
3. The animals were normal rats, not diabetic ones. A glucose-lowering effect in a healthy animal tells you a compound is pharmacologically active; it does not tell you it helps a disease of insulin resistance. Many things lower glucose in a normal rat.
4. 4-hydroxybenzoic acid is not distinctive to pandan. It is a widely distributed simple phenolic acid found across the plant kingdom, including in many ordinary foods. If its presence made a plant antidiabetic, an enormous number of plants would be.

Evidence tier: preliminary (animal, purified compound, non-diabetic model, wrong plant part). This study is a legitimate starting point for research. It is not evidence that pandan-leaf tea lowers blood sugar in a person, and it has never been followed by the human work that would be needed to make it one.

The Only Human Study — and Its Surprise

There is one published human experiment, and its result is the part almost never reported: Chiabchalard and Nooron, “Antihyperglycemic effects of Pandanus amaryllifolius Roxb. leaf extract,” published in Pharmacognosy Magazine in 2015.

The human arm was small and simple. Thirty healthy volunteers received either pandan-leaf tea or hot water fifteen minutes after a standard 75 g oral glucose load — the same sugar challenge used in a glucose tolerance test. The paper's own reported outcome was that mean blood glucose was higher in the pandan group (about 6.16 mmol/L) than in the hot-water control group (about 5.55 mmol/L), a difference the authors described as statistically significant.

Read that again, because it is the single most important sentence on this page. In the only human data that exists, pandan tea did not lower post-glucose blood sugar — the pandan group's readings were higher than the control group's. The paper's positive findings came entirely from its laboratory arms: the leaf extracts inhibited alpha-glucosidase, and increased insulin release from RINm5F rat insulinoma cells in a dose-dependent manner.

How should a small result like that be weighted? Cautiously, in both directions. Thirty healthy volunteers is a small sample; a single unreplicated finding in either direction is weak evidence; the participants were not diabetic; the tea's strength and 2AP content were not the sort of standardised variable a drug trial would control. It would be as wrong to declare that pandan raises blood sugar as it is to declare that it lowers it. But the honest reading is unavoidable: the one time anyone tested the actual claim in actual humans, the claim did not replicate. A field with a positive-results bias produced a negative human result, and that deserves more attention than it gets.

The 2024 Fitoterapia review of the plant by Wang and colleagues — the most comprehensive modern survey, cataloguing more than a hundred compounds — likewise describes the hypoglycaemic evidence as pharmacological, meaning preclinical. It does not report clinical trial evidence, because there is none to report.

Alpha-Glucosidase and Alpha-Amylase Inhibition

The mechanism offered for pandan is inhibition of the intestinal carbohydrate-digesting enzymes. It is worth understanding properly, because the mechanism is real even though the clinical claim is not established.

What these enzymes do. Alpha-amylase, from saliva and pancreas, breaks starch into shorter chains. Alpha-glucosidase, on the brush border of the small intestine, chops those into absorbable single sugars. Slow either one and glucose enters the bloodstream more gradually, blunting the post-meal spike.

This is a validated drug target. Acarbose, miglitol and voglibose are prescription alpha-glucosidase inhibitors used clinically, mainly for post-meal glucose control. Acarbose has real trial evidence behind it and real, well-known side effects — flatulence, bloating and diarrhoea, caused directly by undigested carbohydrate reaching the colon and being fermented there. Those side effects are a useful diagnostic, as the next section explains.

What pandan extracts do in a test tube. They inhibit alpha-glucosidase. This has been reported in the 2015 Pharmacognosy Magazine work and appears in the broader screening literature on Southeast Asian plants. Reports on alpha-amylase are less consistent. The inhibition is generally modest by the standards of these assays and is attributed to the leaf's ordinary phenolic content — catechin, epicatechin, rutin, kaempferol, naringin and the phenolic acids quantified in the 2013 Malaysian profiling study by Ghasemzadeh and Jaafar.

Evidence tier: preliminary (in vitro). Solid as an observation about enzymes in a cuvette. Not evidence about a person.

Why In-Vitro Enzyme Inhibition Rarely Survives Contact With a Human

This section is the crux of the whole page, and it generalises far beyond pandan — the same reasoning applies to cinnamon, to countless “blood-sugar support” botanicals, and to most plant-extract enzyme-inhibition claims you will encounter.

Concentration. An in-vitro assay applies a concentrated extract directly to purified enzyme. In a person, a cup of tea is diluted into the gut's contents, and the polyphenols responsible are present at a small fraction of the assay concentration. Plant polyphenols also bind indiscriminately to proteins in general — which is a well-known source of false-positive enzyme inhibition in these assays.

Site of action. Alpha-glucosidase sits on the intestinal brush border, downstream of the stomach. An inhibitor has to survive gastric acid, arrive at the right place at the right time relative to the meal, and be present at a sufficient local concentration. The 2015 study gave the tea fifteen minutes after the glucose load. Timing is not a detail here; it is most of the pharmacology.

Bioavailability and metabolism. Tea polyphenols are extensively metabolised by gut bacteria and by the liver. What reaches a target is often not the molecule that was tested.

And the diagnostic tell. Effective alpha-glucosidase inhibition has a characteristic, unmistakable consequence: undigested carbohydrate reaches the colon and ferments, producing gas, bloating and loose stools. That is why acarbose is poorly tolerated by many people. Nobody reports this from pandan tea. Hundreds of millions of people drink pandan-flavoured drinks and eat pandan-scented rice without any such effect. If pandan were meaningfully inhibiting intestinal carbohydrate digestion at culinary or tea-strength doses, the population would already know — the side effect is that noticeable. The absence of that signal is, in its own way, evidence.

The Wider Rodent and Cell Literature

Beyond the two studies above there is a scattered body of preclinical work on pandan and metabolic endpoints, mostly from Southeast Asian groups, and it is worth describing as a body rather than paper by paper — which is itself the honest characterisation.

• Cell-culture insulin secretion. The 2015 study's finding that leaf extract increased insulin release from RINm5F rat insulinoma cells is the clearest example. Insulinoma cell lines are a standard screening tool; a great many extracts stimulate them, and the step from a cancerous rat beta-cell line to human islet function is a long one.
• Rodent diabetes models. Various extracts have been tested in streptozotocin-induced and diet-induced rodent models with generally favourable reported outcomes. These studies are typically small, use varied extraction methods and doses, are not standardised to any marker compound, and are rarely replicated by independent groups. They belong in the “interesting, unproven” column.
• Antioxidant and lipid endpoints. Reported improvements in oxidative-stress markers and lipid profiles appear in some animal work. The same limitations apply. There is no human data.
• Weight, appetite and metabolic syndrome. Occasionally claimed online. There is no credible human evidence for pandan and body weight, appetite regulation, or any metabolic-syndrome endpoint. Claims of this kind should be treated as marketing.

The methodological problem running through all of it is extract heterogeneity. As covered on the aroma chemistry page, fresh leaf, dried leaf, aqueous juice, ethanol extract and essential oil are chemically different materials. There is no accepted standardisation marker for pandan — no equivalent of the curcuminoid percentage on a turmeric label. Two studies both reporting on “pandan leaf extract” may have used preparations with little in common. That alone makes pooling the literature impossible.

What a Convincing Study Would Look Like

It is more useful to say precisely what is absent than to gesture at “more research is needed.” For pandan and blood sugar, the missing items are:

1. A randomised, placebo-controlled trial in people with type 2 diabetes or prediabetes. Not healthy volunteers. Zero such trials exist.
2. A defined, standardised preparation. Stated species, plant part, extraction method, and quantified marker compound, so the study could be replicated and a product could be matched to it.
3. A dose-finding study. There is currently no established medicinal dose of pandan for anything. Any number on a supplement label is invented rather than derived.
4. Meaningful endpoints over a meaningful duration. HbA1c over three months, or continuous glucose monitoring across real meals — not a single reading after a sugar drink.
5. Human pharmacokinetics. Nobody knows what pandan's constituents do in a human body: absorption, distribution, metabolism, elimination. This is completely unstudied.
6. Formal interaction studies with metformin, sulfonylureas, insulin and SGLT2 inhibitors. None exist.
7. Toxicology for concentrated extracts. Very little exists, which is why the safety page treats capsules quite differently from cooking.

Until at least the first three exist, the correct label for pandan and blood sugar is UNKNOWN, not “promising.” “Promising” is what a supplement marketer says about an absence of evidence.

The Real Risk: Interaction and Substitution

Pandan leaf in food is not dangerous. The risk here is not toxicity, and being clear about which risk is which is part of giving useful advice.

Risk 1 — substitution. This is the serious one. Someone with type 2 diabetes reads that pandan tea is a natural blood-sugar remedy, feels reassured, and lets a prescription lapse or skips glucose monitoring. Uncontrolled hyperglycaemia damages kidneys, retinas, nerves and blood vessels quietly over years. A pleasant tea with no demonstrated glucose-lowering effect cannot do the work of metformin, insulin, dietary change or clinical follow-up. If pandan tea makes a person feel their diabetes is handled, it has caused harm without any pharmacological action at all.

Risk 2 — additive hypoglycaemia, theoretical but worth respecting. Suppose pandan does have a modest glucose-lowering action that the single small human study missed. Someone on insulin or a sulfonylurea — drugs that can cause hypoglycaemia on their own — who starts drinking strong pandan decoctions several times a day has added an unquantified variable. There are no interaction studies. The practical answer is not fear but monitoring: if you change a daily routine, check your numbers more often for a couple of weeks and tell your prescriber. That is ordinary good practice with any new herbal habit.

Risk 3 — concentrated extracts and capsules. The exposure that has never been studied. Pandan's long culinary safety record supports culinary amounts of leaf. It says nothing about a standardised extract in a capsule taken daily, and pandan's poorly characterised alkaloids — pandamarilactones and relatives — are the concrete reason the two are not equivalent. Marketed pandan blood-sugar supplements have no human trial behind them, no standardisation, and no toxicology. We would not spend money there.

Risk 4 — the sugar the pandan arrives with. Worth stating because it is the most common real-world metabolic effect of pandan consumption. Pandan chiffon cake, kaya, sweetened pandan drinks, pandan bubble tea and pandan condensed-milk desserts are sugar-delivery vehicles. Someone managing blood sugar who increases pandan intake through the dessert route has made their glycaemic control worse, not better, whatever the leaf does. The savoury pot is the better route: a knotted leaf in brown rice, congee, curry or coconut-milk vegetables adds aroma and nothing else.

If You Want to Drink Pandan Tea Anyway

You may well want to, and there is no reason not to. Pandan tea is a pleasant, aromatic, calorie-free hot drink with a long record of casual use. Just hold it in the right category.

• Make it as a drink, not a dose. Two or three fresh leaves, chopped, steeped in about 500 mL of just-off-the-boil water for 10–15 minutes, covered to keep the aroma in. Fresh or frozen leaf; dried is much weaker.
• Do not concentrate it. Boiling large quantities of leaf down into a strong decoction moves you from the well-tolerated culinary range into the unstudied one, for no demonstrated benefit.
• Do not sweeten it and call it healthy. If the point is metabolic, the sugar defeats it.
• Keep taking your medication and keep monitoring. Nothing here justifies changing either. Tell your prescriber about any herbal routine, including this one.
• Skip the capsules. No trial, no dose, no standardisation, no toxicology.
• Pregnancy: culinary use is customary and unremarkable across the region; concentrated extracts and supplements are unstudied and should be avoided.

What Actually Moves Blood Sugar

Since people arrive at pages like this one looking for help with real glucose problems, it would be evasive to critique pandan and stop. The interventions with genuine human trial evidence behind them are not mysterious, and none of them is a tea:

• Carbohydrate quality and quantity. The single largest dietary lever. Whole, intact grains rather than refined flour; whole fruit rather than juice. See brown rice — and note that pandan's real contribution to a diabetic diet may simply be making unpolished rice taste good enough to eat regularly.
• Physical activity, especially after meals. Muscle contraction takes up glucose by a route that does not require insulin. A walk after dinner is a genuine intervention with real evidence.
• Weight loss where relevant. Substantial trial evidence for remission of type 2 diabetes with significant weight reduction.
• Sleep and stress. Both measurably affect insulin sensitivity and both are usually ignored.
• Prescribed medication, taken as prescribed. Metformin in particular has decades of outcome data. Herbal teas do not compete with that and should not try.
• Monitoring. HbA1c and, where appropriate, fasting insulin tell you what is actually happening, rather than what a supplement label claims.

Pandan can sit alongside all of that as a flavouring. It cannot substitute for any of it.

Evidence Tiers for This Page

• Randomised clinical trial evidence for pandan and blood sugar: none. Zero trials in people with diabetes or prediabetes.
• Human evidence that exists: one small study in 30 healthy volunteers, in which the pandan group's post-glucose readings were higher than the control group's. Weak, unreplicated, and pointing away from the claim.
• Preliminary (in vitro): alpha-glucosidase inhibition by leaf extracts; dose-dependent insulin release from a rat insulinoma cell line. Real observations, minimal human relevance.
• Preliminary (animal, purified compound): 4-hydroxybenzoic acid from pandan root lowered glucose in normal rats at 5 mg/kg (1998). Wrong plant part, purified compound, non-diabetic model.
• Traditional use only: pandan root in Thai household medicine for “heat” conditions including diabetes, and as a diuretic. Leaf tea is traditionally a refreshing drink, not an antidiabetic remedy.
• UNKNOWN: effective dose; duration of any effect; human pharmacokinetics; interaction with metformin, sulfonylureas, insulin or SGLT2 inhibitors; safety of concentrated extracts; any effect on HbA1c, weight or appetite.
• Not supported by any evidence: pandan as a treatment for diabetes; pandan for weight loss; pandan as a substitute for glucose-lowering medication.

Key Research Papers

Links are PubMed topic searches, which return current literature rather than a single fixed record. Named papers have their title, journal and year given in the text so you can identify them in the results. Note the Pandanus odorus synonym — searching only the current name misses the foundational pharmacology.

1. Chiabchalard A and Nooron N (2015). “Antihyperglycemic effects of Pandanus amaryllifolius Roxb. leaf extract,” Pharmacognosy Magazine. The only human data: 30 healthy volunteers, 75 g glucose load, pandan tea versus hot water — and post-load glucose higher in the pandan group. Positive findings were in vitro. Search: Pandanus amaryllifolius antihyperglycemic leaf extract
2. Peungvicha P and colleagues (1998). “4-Hydroxybenzoic acid: a hypoglycemic constituent of aqueous extract of Pandanus odorus root,” Journal of Ethnopharmacology. Purified compound, 5 mg/kg orally, normal rats, root not leaf. Search: 4-hydroxybenzoic acid, Pandanus odorus root
3. Wang W and colleagues (2024). “Botany, phytochemistry, pharmacology, and applications of Pandanus amaryllifolius Roxb.: a review,” Fitoterapia. The comprehensive modern survey; describes the hypoglycaemic evidence as pharmacological (preclinical), and reports no clinical trials. Search: Pandanus amaryllifolius pharmacology review
4. Ghasemzadeh A and Jaafar HZE (2013). “Profiling of phenolic compounds and their antioxidant and anticancer activities in pandan (Pandanus amaryllifolius Roxb.) extracts from different locations of Malaysia,” BMC Complementary and Alternative Medicine. Quantifies the catechin, epicatechin, naringin, rutin and kaempferol content credited with the enzyme inhibition, and shows how much it varies by growing site. Search: pandan phenolic profiling and antioxidant activity
5. The pandan diabetes literature as a whole. Small enough that a single search returns essentially all of it — which is itself the finding. Search: Pandanus amaryllifolius and diabetes
6. Pandan and alpha-glucosidase inhibition. The in-vitro mechanism, in context alongside the wider screening literature on Southeast Asian plants. Search: Pandanus amaryllifolius alpha-glucosidase inhibition
7. Plant extracts, alpha-glucosidase inhibition and postprandial glucose in humans. The general question of whether in-vitro enzyme inhibition by botanicals translates — the essential context for reading any such claim. Search: botanical alpha-glucosidase inhibitors and postprandial glucose trials
8. Acarbose and the alpha-glucosidase inhibitor drug class. What validated inhibition looks like clinically, including the gastrointestinal side effects that pandan conspicuously does not produce. Search: acarbose in type 2 diabetes
9. 4-hydroxybenzoic acid across the plant kingdom. Context for why the presence of this simple phenolic acid is not distinctive to pandan. Search: 4-hydroxybenzoic acid distribution and metabolism
10. Herbal supplements and diabetes medication interactions. The general safety literature that ought to exist for pandan and does not. Search: herb–drug interactions with antidiabetic medication
11. Cinnamon for glycaemic control. The closest well-studied parallel — a kitchen spice with a widely marketed blood-sugar claim, tested repeatedly in humans with inconsistent and generally small results. Useful for calibrating expectations about pandan. Search: cinnamon and glycaemic control meta-analyses
12. Pandanus tectorius — a different species entirely. Frequently and wrongly cited as pandan-leaf evidence. Search: Pandanus tectorius fruit

External Resources

• PubMed — the database behind the search links above.
• NIDDK (National Institute of Diabetes and Digestive and Kidney Diseases) — US government reference on diabetes management.
• NCCIH (National Center for Complementary and Integrative Health) — evidence summaries on herbs and supplements.
• American Diabetes Association — patient-facing standards of care.
• Diabetes UK — patient information and evidence reviews.

Connections

• All Herbs
• Pandan (Pandanus amaryllifolius) — the main topic page: identification, species, traditional use and cautions.
• Pandan Benefits Deep Dive — the hub for this set of four articles.
• Pandan: Species, Safety and What Is Genuinely Unknown — why a capsule is not a knotted leaf.
• Pandan Aroma Chemistry — the part of pandan science that is actually settled.
• Type 2 Diabetes — what genuinely works for glucose control.
• Prediabetes — the stage where diet and activity have the most leverage.
• Insulin Resistance — the underlying mechanism most blood-sugar claims gesture at.
• Hemoglobin A1C — the endpoint no pandan study has ever measured.
• Fasting Insulin — a more sensitive early marker than fasting glucose.
• Cinnamon — the same kind of claim, with far more human data and still modest results.
• Bitter Melon — another Asian botanical with a blood-sugar reputation and a mixed trial record.
• Fenugreek — a seed with genuine, if limited, glycaemic trial data, for comparison.
• Gymnema — a third traditional antidiabetic botanical worth comparing on evidence quality.
• Berberine — the botanical-derived compound with the strongest human glycaemic evidence, and a useful benchmark.
• Brown Rice — where pandan's real dietary contribution probably lies.

Safety note and disclaimer. Pandan leaf as a culinary flavouring has a long record of ordinary safe use; that record does not extend to concentrated extracts or supplements, which have no human safety or efficacy data. Pandan is not a treatment for diabetes or prediabetes and must never replace prescribed medication, dietary change or glucose monitoring. If you take insulin or a sulfonylurea and start drinking strong pandan preparations regularly, monitor your glucose more closely for a few weeks and tell your prescriber. Avoid concentrated pandan extracts in pregnancy and breastfeeding. This article is general information for education, not medical advice, and it is not a substitute for individual assessment by a qualified clinician.

Back to Table of Contents