Banaba for Blood Sugar

Banaba is the common name of Lagerstroemia speciosa, a Southeast Asian tree whose leaves have long been taken in the Philippines for diabetes. The leaf compounds studied most are corosolic acid and a group of tannins. In people, banaba has been tested in only a handful of small, short trials, several of them with banaba mixed into a product with other ingredients.

Evidence base: thin. The research records behind this page hold four small human trials of banaba extract or corosolic acid on their own (one of two weeks, one of a single dose and two of 12 weeks), three randomized trials and one uncontrolled study of banaba combined with other ingredients, and animal and laboratory work. One trial enrolled only people with type 2 diabetes, and it lasted two weeks. One of the combination trials found no significant effect.


Table of Contents

  1. What the Plant Is and How It Has Been Used
  2. Active Compounds and How They Act
  3. What Human Trials Found
  4. What Animal and Laboratory Studies Found
  5. Amounts Used in the Trials
  6. Safety Findings and Interactions
  7. Limits of the Evidence
  8. Key Research Papers
  9. Connections
  10. Featured Videos

What the Plant Is and How It Has Been Used

Banaba is a tree of the loosestrife family (Lythraceae). A 2024 paper that mapped its genome describes it as a well-known medicinal and ornamental plant with large flowers and a flowering period of 3 to 5 months (PMID 39543103).

The part used is the leaf. A 2007 review records that Filipinos have traditionally consumed banaba leaves in various forms for diabetes and for kidney-related diseases, and that the herb began to draw the attention of scientists worldwide in the 1990s (PMID 18227906). A decoction, which is leaves boiled in water, is one traditional form (PMID 35280577). A 2012 review dates the first published research study on banaba to 1940 (PMID 22095937).

Today banaba is mostly sold as a leaf extract in capsules. Many extracts are standardized, meaning the maker adjusts each batch to hold a set percentage of one marker compound. For banaba that marker is corosolic acid (PMID 39338524). The trials described below used extracts standardized to 1% or to 0.3% corosolic acid. Banaba is also combined with other ingredients. The studies on this page include it with a cinnamon bark fraction (PMID 27099473), with myo-inositol (PMID 41340221) and, in one study named by a 2025 review, with chromium picolinate (PMID 40613623).

A long history of use shows what people did. It does not show whether the leaf lowers blood sugar. That question is what the trials below were set up to answer.

Back to Table of Contents

Active Compounds and How They Act

The compounds

Researchers do not agree on which compound matters most

Corosolic acid is the compound extracts are standardized to, but two records question whether it is the main active one. The 2007 review reports that when a banaba water extract was split into fractions and each was tested on fat cells in the laboratory, the gallotannins raised glucose uptake and corosolic acid did not account for the activity (PMID 18227906). A 2024 study of a fermented leaf extract argues that asiatic acid is an alternative to corosolic acid as the most important active compound (PMID 39338524). Neither point comes from a trial in people. They matter to a reader because a label that states a corosolic acid percentage describes only one of the candidate compounds.

Three proposed ways of acting

The 2012 review, which covered animal, laboratory and human studies, lists the mechanisms that appear to be involved (PMID 22095937):

  1. In the gut: impaired hydrolysis of sucrose and starches. Hydrolysis is the step in which digestive enzymes split table sugar and starch into simple sugars that can be absorbed. Slowing it slows the arrival of glucose in the blood after a meal.
  2. In the liver: decreased gluconeogenesis, which is the liver's production of new glucose.
  3. In body cells: enhanced cellular uptake of glucose, meaning more glucose moves out of the blood and into cells.
Three places where a 2012 review reports that banaba compounds act: on the arrow from the gut to the blood (slower breakdown of starch and sucrose), on the arrow from the liver to the blood (less new glucose made), and on the arrow from the blood to muscle and fat cells (more glucose taken up). WHERE BANABA IS REPORTED TO ACT Three mechanisms listed in a 2012 review LIVER 2 GUT 1 BLOOD GLUCOSE 3 MUSCLE AND FAT CELLS 1 Slower breakdown of starch and sucrose in the gut 2 Less new glucose made by the liver 3 More glucose taken up by muscle and fat cells

The same review adds that these effects may be mediated by cell signalling factors, among them the peroxisome proliferator-activated receptors, a family of switches inside the cell that control how fat and sugar are handled (PMID 22095937). The review's wording is "appear to involve" and "may be mediated". The abstract does not say which mechanism was seen in animals, which in cells and which in people.

Absorption depends on the form

A 2026 review of corosolic acid, written about its anticancer research, states that the compound has poor solubility in water and limited oral bioavailability, meaning that only part of a swallowed amount reaches the blood (PMID 41612741). This fits a finding from a human trial described below, in which the same extract lowered blood glucose more in a soft gel capsule than as a dry powder (PMID 12787964).

Back to Table of Contents

What Human Trials Found

The trials fall into two groups: banaba extract or corosolic acid given alone, and banaba given as one ingredient of a combination. The second group cannot show what banaba itself did.

Banaba extract or corosolic acid given alone

Two weeks in people with type 2 diabetes (randomized clinical trial, 2003). People with type 2 diabetes took a daily oral dose of a banaba leaf extract standardized to 1% corosolic acid. Daily amounts of 32 mg and 48 mg for 2 weeks produced a significant reduction in blood glucose. The soft gel capsule gave a 30% decrease in blood glucose, against a 20% drop with a hard gelatin capsule filled with dry powder (P<0.001). The abstract does not state how many people took part or what the comparison group received (PMID 12787964). Hemoglobin A1C (HbA1c) is the blood test that reflects average glucose over roughly the previous three months. A two-week trial is too short to show a full change in it, and the abstract reports none.

A single dose before a glucose drink (double-blind, placebo-controlled crossover trial, 2006). Thirty-one people swallowed a capsule of 10 mg of purified corosolic acid or a placebo, on different occasions, 5 minutes before a 75-g oral glucose tolerance test. In that test a person drinks a measured amount of glucose and blood glucose is tracked for two hours. Nineteen of the 31 had diabetes, seven had impaired glucose tolerance, one had impaired fasting glucose and four had normal glucose tolerance. Glucose levels did not differ before the drink or at 30 minutes. With corosolic acid they were lower from 60 minutes until 120 minutes, and the difference reached statistical significance at 90 minutes (PMID 16549220). This trial tested one dose on one day. It did not test daily use.

Twelve weeks in metabolic syndrome (randomized, double-blind, placebo-controlled trial, 2022). Metabolic syndrome is a cluster of a large waist, raised blood pressure, raised blood sugar and abnormal blood fats. Twenty-four patients took part: 12 received banaba 500 mg twice a day, before breakfast and dinner, and 12 received placebo. After banaba, fasting plasma glucose went from 5.9 ± 0.4 to 5.7 ± 0.4 mmol/L (P = .034), systolic blood pressure from 121.5 ± 12.9 to 116.3 ± 9.8 mmHg (P = .050) and triglycerides from 2.3 ± 0.4 to 1.7 ± 0.5 mmol/L (P = .021). The amount of insulin released after a glucose drink also fell. Eight patients (67%) in the banaba group no longer met the definition of metabolic syndrome (PMID 34726501). Three points limit this result. The fall in fasting glucose was 0.2 mmol/L. The abstract does not list insulin sensitivity or the glucose curve after the drink among the measures that changed significantly. And it does not report how many people in the placebo group left the metabolic syndrome definition.

Twelve weeks in overweight adults with mildly raised fasting glucose (randomized, placebo-controlled trial, 2014). This trial was designed mainly to test a different leaf extract (soybean leaf) and used banaba extract as its comparison arm. Overweight adults with fasting blood glucose of 100 to 125 mg/dL took placebo capsules, or banaba extract 300 mg per day (0.3% corosolic acid), with regular meals. Compared with placebo, the banaba group ended with significantly lower blood glucose, HbA1c, liver transaminase levels and homeostatic model assessment of insulin resistance (HOMA-IR), a score calculated from fasting glucose and fasting insulin in which a higher number means more insulin resistance. Body fat content, waist circumference and triglycerides were also lower. Body weight, body mass index, total cholesterol and low-density lipoprotein cholesterol did not differ from placebo, and banaba did not raise high-density lipoprotein cholesterol (PMID 24873894). The abstract gives neither the number of participants nor the size of the changes.

Banaba combined with other ingredients

With a cinnamon bark fraction, in impaired glucose tolerance (randomized, double-blind, placebo-controlled preliminary study, 2016). Eighty adults whose glucose two hours after a meal was 140 to 199 mg/dL took a combined fraction of Cinnamomum burmanii and banaba, 50 to 100 mg daily, or placebo for 12 weeks. HOMA-IR fell by 27.04% ± 29.41% with the combination and by 4.90% ± 41.27% with placebo (P = 0.013). Measures of early and late insulin release also improved more than with placebo (PMID 27099473). The authors call the study preliminary.

The same combination against metformin, in polycystic ovary syndrome (randomized, double-blind non-inferiority trial, 2024). A non-inferiority trial asks whether a treatment is no worse than a standard one by more than a margin set in advance. Here 124 women with polycystic ovary syndrome and insulin resistance took the combined fraction, 100 mg daily, or extended-release metformin 750 mg twice daily for 6 months. HOMA-IR improved by 1.03 ± 0.50 with the combination and by 1.19 ± 0.50 with metformin, a difference of 0.16 (95% confidence interval −1.24 to 1.56; P = 0.3168). The authors describe the two as comparable, and then state that non-inferiority to metformin "remains inconclusive". Fewer adverse events occurred in the combination group (PMID 39033369). These women had polycystic ovary syndrome, not diabetes.

The same combination, alone and with metformin: no significant effect (randomized, double-blind, three-arm trial, 2023). Sixty women with polycystic ovary syndrome and insulin resistance, 20 in each arm, took the combined fraction 100 mg once a day, extended-release metformin 750 mg twice a day, or both, for 6 months. With the combined fraction alone HOMA-IR was 3.55 at the start, 3.59 at 3 months and 3.80 at 6 months. With metformin alone it was 4.00, 2.21 and 4.40. With both it was 3.30, 2.86 and 3.12. Fasting plasma glucose, high-density lipoprotein cholesterol and triglycerides showed no apparent difference in any group. The authors concluded that the combined fraction, alone or with metformin, "showed no significant efficacy" (PMID 37424700). This is a negative result.

With myo-inositol, in postmenopausal women (retrospective observational study, no comparison group, 2025). Thirty-one overweight and obese postmenopausal women (mean age 51 ± 1.2 years) had taken myo-inositol 2 g with banaba extract 48 mg standardized to 1% corosolic acid every day for 12 weeks. Afterwards fasting insulin and HOMA-IR were significantly lower. During an oral glucose tolerance test the area under the glucose curve, a measure of total glucose exposure, was 23.5% lower, and the area under the insulin curve was 42% lower (PMID 41340221). The researchers looked back at records, and no group went without the product. A study of this kind cannot separate the product's effect from other changes over the 12 weeks, and it cannot separate banaba from myo-inositol.

What the reviews concluded

A 2012 review summarized the animal, laboratory and human literature and concluded that banaba extract, corosolic acid and other constituents "may be beneficial" for the features of metabolic syndrome. It notes that pure corosolic acid has been reported to lower blood sugar within 60 minutes in human subjects (PMID 22095937).

A 2025 narrative review of banaba and chromium picolinate for prediabetes states that clinical trials have confirmed glucose-lowering effects of banaba in people with type 2 diabetes and prediabetes. It also reports that only one study has tested banaba together with chromium picolinate, with positive outcomes in people with prediabetes, and that further research is needed to corroborate those preliminary findings (PMID 40613623). A narrative review describes studies in prose. It does not pool their numbers the way a meta-analysis does, and the records behind this page contain no meta-analysis of banaba trials.

What no trial here tested

No trial in these records gave banaba in place of a diabetes drug to people with diabetes. The only comparison with a drug was the combined cinnamon and banaba fraction against metformin in women with polycystic ovary syndrome, where one trial was inconclusive on non-inferiority and the other found no significant efficacy. No trial in these records followed people with diabetes for longer than two weeks, and none measured the long-term complications of diabetes in the eyes, kidneys, nerves or heart.

Back to Table of Contents

What Animal and Laboratory Studies Found

Everything in this section was found in mice, in rats or in cells grown in the laboratory. None of it has been shown in people by these studies.

In mice

Fasting glucose, insulin and body fat. KK-Ay mice are a genetically obese strain used as a model of type 2 diabetes. Six-week-old mice were fed a high-fat diet for 9 weeks with or without 0.023% corosolic acid. The mice given corosolic acid had 10% lower body weight and 15% lower total fat mass than control mice, and their fasting plasma glucose, insulin and triglyceride were lower by 23%, 41% and 22%. Their blood level of adiponectin, a hormone released by fat tissue, was higher (PMID 18379057).

Cholesterol. In the same strain of mice, a high-cholesterol diet containing 0.023% corosolic acid for 10 weeks gave a mean blood cholesterol 32% lower and a liver cholesterol content 46% lower than the same diet without it (both P<0.05). The authors suggest that corosolic acid acts on cholesterol absorption in the small intestine (PMID 20834178).

In cells and other laboratory systems

The 2007 review traces the laboratory search for banaba's active compound. Corosolic acid was isolated from a methanol extract of the leaf using tumor cells as the test system. Work on the water-soluble fraction then identified the ellagitannin lagerstroemin. In a third approach a banaba water extract made cultured fat cells (3T3-L1 adipocytes) take up glucose in an insulin-like way, and separating the extract pointed to gallotannins, not corosolic acid, as the components responsible. The most potent was penta-O-galloyl-glucopyranose, which also reduced the formation of new fat cells in the same laboratory system (PMID 18227906).

A 2024 study fermented banaba leaf extract with a strain of the bacterium Lactobacillus plantarum. The authors report that the fermented extract improved glucose uptake and insulin secretion in their test systems, and that its content of asiatic acid rose 3.8-fold (PMID 39338524). The abstract does not name the cells or animals used, and the study did not test the fermented extract in people.

Toxicity testing in rats

Sprague-Dawley rats were given a banaba preparation by mouth as a single dose of 2000 mg per kilogram of body weight and watched for 14 days, or 200 mg per kilogram daily for 28 days. No deaths or illness occurred. Organ weights, blood counts and blood chemistry did not differ significantly from control rats, and the liver showed no visible change under the microscope. The authors add that cytotoxicity and genotoxicity, which are damage to cells and damage to genetic material, remain to be investigated (PMID 35280577).

Back to Table of Contents

Amounts Used in the Trials

The table lists what the researchers gave in each human study. These are records of what was tested. They are not doses for the reader.

Study Who took part What was given Amount Length
Randomized clinical trial, 2003 People with type 2 diabetes (number not in the abstract) Leaf extract standardized to 1% corosolic acid 32 mg and 48 mg daily 2 weeks
Placebo-controlled crossover trial, 2006 31 people, 19 with diabetes Purified corosolic acid 10 mg once, 5 minutes before a 75-g glucose drink One occasion
Randomized placebo-controlled trial, 2022 24 patients with metabolic syndrome Banaba (preparation not described in the abstract) 500 mg twice a day, before breakfast and dinner 12 weeks
Randomized placebo-controlled trial, 2014 Overweight adults with fasting glucose of 100 to 125 mg/dL (number not in the abstract) Banaba extract, 0.3% corosolic acid 300 mg per day with regular meals 12 weeks
Randomized placebo-controlled preliminary study, 2016 80 adults with impaired glucose tolerance Combined cinnamon bark and banaba fraction 50 to 100 mg daily 12 weeks
Randomized non-inferiority trial against metformin, 2024 124 women with polycystic ovary syndrome and insulin resistance Combined cinnamon bark and banaba fraction 100 mg daily 6 months
Randomized three-arm trial, 2023 60 women with polycystic ovary syndrome and insulin resistance Combined cinnamon bark and banaba fraction, alone or with metformin 100 mg once a day 6 months
Retrospective observational study, 2025 31 overweight and obese postmenopausal women Myo-inositol with banaba extract standardized to 1% corosolic acid Myo-inositol 2 g and banaba extract 48 mg daily 12 weeks

Sources, in table order: PMID 12787964, PMID 16549220, PMID 34726501, PMID 24873894, PMID 27099473, PMID 39033369, PMID 37424700, PMID 41340221.

The amounts are not comparable with each other

The extracts differ in strength. By arithmetic from the stated percentages, 48 mg of a 1% extract holds about 0.48 mg of corosolic acid, and 300 mg of a 0.3% extract holds about 0.9 mg. These two figures are calculations, not numbers the abstracts state. The single-dose trial gave 10 mg of purified corosolic acid, more than ten times either amount. A result seen with the purified compound therefore does not show what a leaf extract does, and the reverse is also true.

The capsule form changed the result

In the 2003 trial the same standardized extract lowered blood glucose by 30% in a soft gel capsule and by 20% in a dry-powder hard capsule (P<0.001). The authors read this as better absorption from the soft gel (PMID 12787964). Two products with the same amount of extract on the label can therefore act differently.

Back to Table of Contents

Safety Findings and Interactions

What the studies recorded

These records cover small groups for 2 weeks to 6 months. A trial of 24 or 80 people can miss a side effect that occurs in one person in several hundred, and a 12-week trial cannot show what happens after years of use.

Low blood sugar when combined with glucose-lowering drugs

Anything that lowers blood sugar adds to the effect of insulin and of sulfonylurea drugs, and low blood sugar (hypoglycemia) is the known result when glucose-lowering agents are stacked. None of the studies behind this page tested banaba together with insulin or a sulfonylurea, so the size of this effect for banaba has not been measured. The one trial that paired the combined fraction with metformin reported no harm to heart, liver or kidney function, and it also found no significant benefit (PMID 37424700).

The signs of low blood sugar are shakiness, sweating, a fast heartbeat and confusion. At lower levels, seizure or loss of consciousness can follow. More detail is on the page about hypoglycemia awareness.

Blood pressure

In the 12-week metabolic syndrome trial, systolic blood pressure, the upper number of a reading, fell from 121.5 ± 12.9 to 116.3 ± 9.8 mmHg in the banaba group (P = .050) (PMID 34726501). The trial did not study people taking blood pressure medication.

Type 1 diabetes

Type 1 diabetes is a state of absolute insulin deficiency. No herb, food or diet replaces insulin. Without insulin, diabetic ketoacidosis develops, and it is fatal if untreated. None of the studies on this page enrolled people with type 1 diabetes.

Pregnancy, breastfeeding and children

The records behind this page contain no study of banaba in pregnancy, during breastfeeding or in children. Its safety in those groups has not been measured in any of them.

Products differ from one another

Herbal products vary in plant part, extract and content from one product to the next, so a trial result for one preparation does not transfer to another. For banaba the studies show this directly: the extracts ranged from 0.3% to 1% corosolic acid, the capsule form changed the size of the glucose drop (PMID 12787964), and studies outside the human body point to tannins and to asiatic acid, which a corosolic acid percentage on a label does not measure (PMID 18227906, PMID 39338524).

Back to Table of Contents

Limits of the Evidence

Taken together, the human research shows short-term falls in blood glucose and in measures of insulin resistance in several small studies, one trial with no effect, and no evidence yet on long-term control of type 2 diabetes.

Back to Table of Contents

Key Research Papers

  1. Judy WV, Hari SP, Stogsdill WW, Judy JS, Naguib YM, Passwater R. Antidiabetic activity of a standardized extract (Glucosol) from Lagerstroemia speciosa leaves in Type II diabetics. A dose-dependence study. J Ethnopharmacol. 2003;87(1):115-7. PubMed PMID: 12787964
  2. Fukushima M, Matsuyama F, Ueda N, Egawa K, Takemoto J, Kajimoto Y, et al. Effect of corosolic acid on postchallenge plasma glucose levels. Diabetes Res Clin Pract. 2006;73(2):174-7. PubMed PMID: 16549220
  3. López-Murillo LD, González-Ortiz M, Martínez-Abundis E, Cortez-Navarrete M, Pérez-Rubio KG. Effect of Banaba (Lagerstroemia speciosa) on Metabolic Syndrome, Insulin Sensitivity, and Insulin Secretion. J Med Food. 2022;25(2):177-182. PubMed PMID: 34726501
  4. Choi MS, Ryu R, Seo YR, Jeong TS, Shin DH, Park YB, et al. The beneficial effect of soybean (Glycine max (L.) Merr.) leaf extracts in adults with prediabetes: a randomized placebo controlled trial. Food Funct. 2014;5(7):1621-30. PubMed PMID: 24873894
  5. Manaf A, Tjandrawinata RR, Malinda D. Insulin sensitizer in prediabetes: a clinical study with DLBS3233, a combined bioactive fraction of Cinnamomum burmanii and Lagerstroemia speciosa. Drug Des Devel Ther. 2016;10:1279-89. PubMed PMID: 27099473
  6. Hestiantoro A, Permadi W, Tjandrawinata RR, Wiweko B, Ritonga MA, Ferrina AI, et al. The Efficacy and Safety of DLBS3233, A Combined Bioactive Fraction of Cinnamomum burmanii and Lagerstroemia speciosa Plants on The Endocrine-Metabolic Profile of Women with Polycystic Ovary Syndrome: A Randomized Clinical Trial. Int J Fertil Steril. 2024;18(Suppl 1):35-47. PubMed PMID: 39033369
  7. Hidayat ST, Mulyantoro I, Damas S, Tjandrawinata RR. The Effect and Safety Assessment of Metformin and DLBS3233 (A Bioactive Fraction of Lagerstroemia speciosa and Cinnamomum burmannii) on Improving Metabolic Parameters in Women with Polycystic Ovary Syndrome. Int J Womens Health. 2023;15:971-985. PubMed PMID: 37424700
  8. Battipaglia C, Vescovi V, Foschi M, Righi B, Sponzilli A, Setti V, et al. Combined myo-inositol and Banaba (1% Corosolic Acid) improve HOMA-IR and hepatic insulin extraction index in overweight and obese postmenopausal women. Gynecol Endocrinol. 2025;41(1):2596417. PubMed PMID: 41340221
  9. Stohs SJ, Miller H, Kaats GR. A review of the efficacy and safety of banaba (Lagerstroemia speciosa L.) and corosolic acid. Phytother Res. 2012;26(3):317-24. PubMed PMID: 22095937
  10. Derosa G, Maffioli P, D'Angelo A, Foscaldi V, Piazza R, Mangrella M, et al. Exploring nutraceutical solutions for prediabetes: a narrative review on the effects of banaba and chromium picolinate. Eur Rev Med Pharmacol Sci. 2025;29(6):324-338. PubMed PMID: 40613623
  11. Klein G, Kim J, Himmeldirk K, Cao Y, Chen X. Antidiabetes and Anti-obesity Activity of Lagerstroemia speciosa. Evid Based Complement Alternat Med. 2007;4(4):401-7. PubMed PMID: 18227906
  12. Yamada K, Hosokawa M, Yamada C, Watanabe R, Fujimoto S, Fujiwara H, et al. Dietary corosolic acid ameliorates obesity and hepatic steatosis in KK-Ay mice. Biol Pharm Bull. 2008;31(4):651-5. PubMed PMID: 18379057
  13. Takagi S, Miura T, Ishihara E, Ishida T, Chinzei Y. Effect of corosolic acid on dietary hypercholesterolemia and hepatic steatosis in KK-Ay diabetic mice. Biomed Res. 2010;31(4):213-8. PubMed PMID: 20834178
  14. An BC, Kwak SH, Ahn JY, Won HY, Kim TH, Ryu Y, et al. Identification of Bioactive Substances Derived from the Probiotic-Induced Bioconversion of Lagerstroemia speciosa Pers. Leaf Extract That Have Beneficial Effects on Diabetes and Obesity. Microorganisms. 2024;12(9):1848. PubMed PMID: 39338524
  15. Alkahtani S, Hasnain MS, Algamdy H, Aljarba NH, AlKahtane A. Acute and sub-acute oral toxicity Lagerstroemia speciosa in Sprague-Dawley rats. Saudi J Biol Sci. 2022;29(3):1585-1591. PubMed PMID: 35280577
  16. Uddin MB, Shipon MNH, Suma RB, Rahman MA, Alkhathami AG, Mia E, et al. Anticancer Activity of Corosolic Acid With Botanical Sources, Biopharmaceutical Profile, Mechanistic Insight, Toxicity, and Clinical Evidence. Phytochem Anal. 2026;37(3):362-374. PubMed PMID: 41612741
  17. Wan Z, Zheng T, Cai M, Wang J, Pan H, Cheng T, et al. Chromosome-level genome assembly of banaba (Lagerstroemia speciosa L.). Sci Data. 2024;11(1):1228. PubMed PMID: 39543103

PubMed Topic Searches

  1. PubMed: Lagerstroemia speciosa and diabetes
  2. PubMed: Banaba, corosolic acid and glucose
  3. PubMed: Corosolic acid clinical trials
  4. PubMed: Lagerstroemia speciosa randomized trials
  5. PubMed: Lagerstroemia speciosa safety and toxicity

Back to Table of Contents

Connections

Back to Table of Contents