From four Centella asiatica triterpenes and gut biotransformation to the GABA shunt, neuroplasticity, and vascular filtration. A review of clinical studies, extract standardization, research doses, combinations, and liver risk, distinguishing mechanistic hypotheses from outcomes in people.
Ankles that swell after a long day and difficulty concentrating seem like problems for entirely different medical specialties. In studies of Asian pennywort, or Centella asiatica, however, they appear side by side: its constituents are investigated both in the connective tissue surrounding microvessels and in nerve cells. That juxtaposition is attractive to advertising, which can readily turn it into a promise to “improve blood flow to the brain and calm the nerves.” The scientific story is more complicated. Several distinct evidentiary steps lie between a change in cell signaling, a measure of skin microcirculation, and a meaningful improvement in how a person feels.
Gotu kola, or Centella asiatica, is a perennial creeping plant in the Apiaceae family. In Ayurvedic tradition it is known as mandukaparni and has been used among plants associated with supporting mental activity. This history helps explain the research interest, but does not establish a modern medical indication. It is especially important to distinguish the whole, multicomponent leaf, a dry extract, a purified triterpene fraction, and an individual molecule: findings for one of these cannot automatically be transferred to the others.
The name “brahmi” creates further confusion. In different traditions it has referred both to Centella and to Bacopa monnieri, although these are distinct plants with different chemical profiles. Bacopa belongs to the Plantaginaceae family and contains mainly bacosides with a dammarane triterpene scaffold; Centella is characterized chiefly by pentacyclic triterpenes of the ursane series. They can be compared as research subjects, but one plant cannot be substituted for the other on the basis of a shared vernacular name.
This article follows Centella from extract composition to clinical outcomes. The vascular research is considered through fluid filtration and the extracellular matrix; the nervous-system research through GABA metabolism, plasticity, and cellular protection. In each case, one question matters: what exactly was shown, in which system, and how far can the conclusion reasonably be extended?
The main Centella triterpenes studied form two pairs: asiaticoside and asiatic acid, and madecassoside and madecassic acid. The first member of each pair carries a sugar moiety and is a saponin glycoside; the second is the corresponding aglycone. The term “centellosides” is generally used for the plant’s triterpene glycosides; analytical profiles also include the free acids alongside them. When reading a label, a list of named compounds is more useful than relying on a single umbrella term.
| Compound | Formula | Approximate molar mass | Relevant feature |
|---|---|---|---|
| Asiaticoside | 959.12 g/mol | Sugar moiety is linked to the aglycone carboxyl group at C-28 | |
| Asiatic acid | 488.70 g/mol | Aglycone with hydroxyl groups at positions 2α, 3β, and 23 | |
| Madecassoside | 975.12 g/mol | Glycoside with an additional hydroxyl group at position 6β | |
| Madecassic acid | 504.70 g/mol | Corresponding aglycone: 2α,3β,6β,23-tetrahydroxyurs-12-en-28-oic acid |
In asiaticoside and madecassoside, the sugar chain contains two glucose units and one rhamnose. It can be represented as an α-L-rhamnopyranosyl-(1→4)-β-D-glucopyranosyl-(1→6)-β-D-glucopyranosyl moiety attached through an ester bond at C-28. One additional hydroxyl group distinguishes the madecassic pair from the asiatic pair, but even this small structural difference can affect solubility, protein interactions, and metabolism. The claim that “all triterpenes act alike” is therefore chemically too broad.
The plant also contains minor triterpenes, including asiaticoside B, terminolic acid, and madecassic acid derivatives, as well as compounds described as centelloside, centellazide, and brahmoside. Their nomenclature and relative abundance vary among analytical studies. In addition to triterpenes, aqueous and alcoholic extracts contain other constituents; the effects of a whole extract cannot be attributed to asiaticoside alone without fractionation. This becomes important when comparing experiments on neurons with studies of purified vascular preparations.
Vascular studies often refer to TECA, a titrated extract of Centella asiatica, and TTFCA, the total triterpenic fraction of Centella asiatica. Classical purified fractions have been described as containing about 40% asiaticoside and 60% free acids, usually around 30% each of asiatic and madecassic acids. This describes particular preparations; it is not a universal rule for every product carrying these initials. The abbreviation does not remove the need to find out the product’s actual specification.
Leaf powder, an extract with a 10:1 herb-to-extract ratio, and a preparation with quantitatively specified triterpenes answer different questions. A 10:1 ratio describes a manufacturing relationship but does not reveal the asiaticoside content; a label stating “20% triterpenes” without an analytical method and list of components is also incomplete. Analytical standardization, for example by HPLC, improves comparability of composition, but does not by itself demonstrate clinical efficacy. Applying study findings requires a match in chemical profile, regimen, and research question.
The sugar moiety makes asiaticoside and madecassoside large molecules with many polar groups. This limits simple passive diffusion across lipid membranes, but does not mean that intact glycosides are never absorbed. Exposure depends on how the preparation dissolves, the extract’s composition, transformations in the gut, and the experimental model. A universal claim that “less than three percent is absorbed,” without specifying the compound, preparation, and species, is misleading here.
In the intestine, the sugar moiety may be removed, yielding the corresponding acids. Microbial biotransformation is one pathway that helps explain the appearance of aglycones after glycoside intake, but the entire process should not be rigidly assigned to one gut region or three named bacterial genera. β-glucosidases and α-rhamnosidases cleave suitable glycosidic bonds; the ester linkage attaching the sugar fragment to the C-28 carboxyl group requires separate chemical consideration. It is inaccurate to portray any glycosidase as a universal enzyme that cleaves every bond in this molecule.
The overall conversion can be written as hydrolysis without implying a single enzyme or sequence of steps:
The first reaction yields asiatic acid and the second madecassic acid; the products of complete hydrolysis are two glucose molecules and one rhamnose. These equations describe the chemical outcome, not the obligatory complete conversion of an ingested dose in the body. Some of the substance may remain unabsorbed, circulate in its original form, or undergo other metabolic steps.
The free acids are less polar than their glycosides, but lipophilicity does not equal high oral bioavailability. Poor water solubility may limit delivery of a compound to the membrane even when the membrane itself is permeable to it. After absorption, portal transport, interaction with plasma proteins, and hepatic biotransformation matter. Conjugation and biliary excretion of metabolites are possible parts of this pathway; their quantitative contribution cannot be inferred from molecular structure alone.
Grimaldi and colleagues showed that asiatic acid could be measured in healthy volunteers after single and repeated doses of a triterpene fraction. This supports discussion of systemic exposure, but does not justify assigning the same , elimination half-life , or duration of effect to every product. is the time to the peak measured concentration, while is the time for that concentration to fall by half over the relevant part of the curve. Neither parameter by itself establishes a “therapeutic concentration for 12–16 hours” when a target level and clinical relationship have not been defined.
The blood–brain barrier consists of specialized endothelium and associated cellular structures that regulate the entry of substances into the brain. Molecular mass and lipophilicity are among the factors that affect permeability, but they do not replace measurements: ionization, protein binding, transporters, and free concentration also matter. The formula “under 500 Da means it passes freely” is not a pharmacokinetic rule.
Centella’s experimental neuroactivity makes brain exposure a reasonable question, but an effect in cultured neurons or animals does not tell us the concentration in the human hippocampus after taking a capsule. A change in the human startle reflex likewise does not show which molecule crossed the barrier or in what amount. The “30–60 minutes” mentioned in a clinical test therefore cannot be turned into a measured time for asiatic acid to reach the brain. That conclusion would require a separate pharmacokinetic study.
GABA, or γ-aminobutyric acid, is the main inhibitory neurotransmitter in the mature central nervous system. It is formed from glutamate, which also participates in metabolism and excitatory signaling. The conversion is catalyzed by glutamate decarboxylase (GAD); its GAD65 and GAD67 forms differ in cellular organization, and the cofactor is pyridoxal 5′-phosphate, the active form of vitamin B6. Even at this level, it is clear why the phrase “increase GABA” is too broad: synthesis in the cell, packaging into vesicles, release, receptor binding, reuptake, and breakdown are distinct processes. That is too broad: synthesis inside a cell, packaging into a vesicle, release, and action at a receptor are different events.
The vesicular transporter VGAT loads GABA into synaptic vesicles, after which stimulation of the nerve terminal releases the transmitter into the synaptic cleft. Ionotropic receptors regulate anion conductance, chiefly for chloride; metabotropic receptors act through G proteins and alter ion-channel activity. The direction of the cellular response depends, among other things, on the chloride gradient. Thus, even an accurate description of the receptor cannot be used to predict the clinical degree of sedation.
The signal ends in part through reuptake by GAT transporters into nerve and glial cells. Catabolic GABA transaminase (GABA-T) acts inside the cell, mainly in a mitochondrial context; it does not break down the transmitter directly in the synaptic cleft. Its reaction is followed by oxidation of succinic semialdehyde to succinate, a metabolite of the tricarboxylic acid cycle. This links neurotransmitter metabolism with energy metabolism and forms part of what is known as the GABA shunt.
Awad and colleagues studied the effects of traditionally used plants on enzymes in this system under laboratory conditions. For Centella, stimulation of GAD has been discussed; an extract’s result in an enzyme experiment should not automatically be attributed to pure asiatic acid or to concentrations reached in people. The claim that Centella’s clinical action is based on a proven two-part mechanism, “GAD up, GABA-T down,” goes beyond the available evidence. Both branches matter for understanding the hypothesis, but GABA-T inhibition and its contribution to the effects of a particular preparation require separate confirmation.
Even if a preparation changes intracellular GABA metabolism, this does not directly establish an increase in the “lifetime of transmitter quanta” in the cleft. That depends on release, diffusion, reuptake, and synaptic organization. Likewise, a proposed action at does not identify a specific binding site between its α and β subunits. Such detail requires receptor experiments; comparisons with medicines require clinical evidence on effects and adverse events.
Benzodiazepines are positive allosteric modulators of certain receptors. Their effects and risks depend on the medicine, dose, duration of use, and the person’s condition; drowsiness, impaired coordination or memory, and dependence should not be portrayed as equally inevitable for every user. Centella has been studied far less. The absence of these events in a small study does not prove that it is guaranteed to cause no sedation, ataxia, amnesia, tolerance, or withdrawal syndrome.
Proposed links between Centella and stress systems are also best considered separately. The locus coeruleus is an important source of noradrenaline, the amygdala participates in threat processing, and the hypothalamic–pituitary–adrenal axis links CRH, ACTH, and cortisol. A change in animal behavior may be consistent with an effect on these circuits, but does not demonstrate suppression of locus-coeruleus firing or CRH expression in people after taking an extract. These remain research explanations unless measured directly.
The acoustic startle reflex is a rapid flinch in response to a sudden sound. Brainstem circuits, including the pons, contribute to it, while other brain structures shape its emotional context. It is a useful experimental measure of reactivity, but not a “gold standard” for establishing therapeutic efficacy in an anxiety disorder. Clinical relevance requires reliable evidence about symptoms, functioning, how long the outcome lasts, and safety in addition to a laboratory signal.
In chronic venous insufficiency, the problem does not begin with “thick blood,” but with impaired venous return and elevated pressure, including pressure caused by valve dysfunction. This pressure is transmitted to the skin’s microcirculation, changing the balance between fluid filtration and its removal. With a prolonged course, inflammation, leukocyte activation, changes in surrounding connective tissue, and impaired skin trophism may follow. Purified Centella triterpene fractions have been studied against this background.
The capillary wall does not have a substantial smooth-muscle layer: the endothelium, basement membrane, pericytes, and surrounding matrix are important. Postcapillary venules have a somewhat different structure, so it is inaccurate to treat every microvessel as the same kind of tube. The basement membrane contains mainly network-forming components, including collagen IV and laminins, whereas fibrillar collagens I and III are important in the surrounding connective tissue. Increased collagen I synthesis in a fibroblast does not amount to direct “repair” of the basement membrane or a venous valve.
Maquart and colleagues found that a Centella triterpene stimulated collagen synthesis in fibroblast cultures. Lee and colleagues showed that asiaticoside increased collagen I synthesis in human dermal fibroblasts and was accompanied by Smad2/3 phosphorylation, formation of a Smad3/4 complex, and its translocation to the nucleus. A key detail of that study is that the observed response was independent of TGF-β receptor I kinase in that experimental system. It would therefore be inaccurate to replace the finding with the familiar arrow “asiaticoside activates the TGF-β receptor.”
The genes COL1A1 and COL1A2 encode the chains of collagen I, while COL3A1 encodes collagen III. After procollagen is synthesized, intracellular processing, secretion, and extracellular maturation are required; lysyl oxidase contributes to cross-link formation between collagen molecules. This is general matrix biology, not evidence that every Centella extract enhances each step in a patient’s blood vessel. Nor can collagens I and III be simplistically divided into “rigid” and “elastic”: mechanical behavior depends on fiber organization, cross-links, elastin, and the overall tissue structure.
Increased matrix synthesis may be helpful in one context and undesirable in another. A clinical promise to “normalize the collagen ratio without fibrosis” requires direct evidence of a tissue outcome. In the case of Centella, cell experiments offer a biological explanation for interest in connective tissue, but do not prove that it eliminates pathological pores, prevents aneurysms, or restores venous elasticity through one established molecular cascade.
Water movement across the microvascular wall depends on pressure, its hydraulic conductivity, the properties of the endothelial barrier, and the interstitium. A simplified form of the Starling equation can help explain it:
Here is fluid flow, the filtration coefficient, hydrostatic pressure, and oncotic pressure; subscripts and refer to the vascular and interstitial compartments, and is the protein reflection coefficient. This is a teaching model; modern descriptions also account for the endothelial glycocalyx and subglycocalyx space. The practical point is that edema depends not only on “permeability,” but also on pressure and the lymphatic system’s ability to remove filtered fluid.
Vascular studies have measured capillary filtration, limb volume, laser-Doppler flowmetry parameters, and transcutaneous gas tensions. and describe oxygen and carbon dioxide at the skin measurement site; they depend on the method, local perfusion, and recording conditions. These measures are useful for assessing a microcirculatory model, but are not direct measurements of oxygen delivery to the brain. A reduction in skin edema is likewise not equivalent to preventing thrombosis or venous ulcers.
Vascular biology research examines VEGF, occludin, and ZO-1, which participate in angiogenesis and the organization of cell junctions. However, not every capillary has the same tight junctions, and fenestrations can be a normal feature of an organ. Suppressing VEGF is not universally beneficial: this signal is also needed for physiological tissue repair. Barrier mechanisms should therefore be treated as a research question, not as a proven “sealing” effect of Centella throughout the body.
In diabetes, microangiopathy is associated with chronic hyperglycemia, advanced glycation end products (AGEs), inflammation, and structural remodeling of microvessels. A thickened basement membrane does not guarantee proper barrier function. A study of a triterpene fraction using skin microcirculatory measures does not prove the removal of glycation, normalization of membrane thickness, a reduction in albuminuria, or prevention of diabetic retinopathy. Each of these outcomes requires its own measurement; diabetes care cannot rest on presumed equivalence among them.
A nerve cell continually changes the strength of synaptic connections and the organization of its processes. Neurite growth in culture, dendritic branching in an animal, and improved memory in a person are different levels of plasticity. Centella has produced interesting findings at the first two levels, but they cannot be combined into a conclusion that a supplement course has been shown to “build a neural network.” It is particularly important to know whether the study used an aqueous extract, a purified fraction, or a derivative of an individual triterpene.
A kinase cascade important in neuroscience is Ras → Raf → MEK1/2 → ERK1/2. It transmits intracellular signals and may participate in neurite growth, cell survival, and changes in gene expression. Calcium signals and CaMKII are linked to another set of plasticity-related responses. A change in these pathways in an experiment indicates biological activity of the substance being studied, but does not establish one universal Centella receptor or the same response in every type of neuron.
CREB is a transcriptional regulator involved in changes in the expression of several genes, including context-dependent regulation of BDNF. Phosphorylation of CREB at Ser-133 is an important event, but its outcome depends on other signals and the cell’s state. ERK’s connection to CREB often passes through intermediate kinases, so the direct arrow “ERK phosphorylates CREB” oversimplifies the mechanism. BDNF, in turn, acts through the TrkB receptor and may support synaptic function; this does not mean that an increase in any single marker guarantees better memory.
A useful way to read a mechanistic diagram is as a sequence of research questions: does ERK signaling change, then CREB, then neurotrophic expression, and ultimately synaptic function? PSD-95 reflects the organization of the postsynaptic density, synaptophysin is associated with presynaptic vesicles, and MAP-2 with dendritic microtubules. Tau is also a microtubule-associated protein, but its pathological regulation has separate significance in neurodegeneration. It is not safe to assume that increasing all these proteins at once is automatically desirable or means the cytoskeleton is being “built” beneficially.
Soumyanath and colleagues investigated nerve regeneration after oral Centella administration and neurite elongation in cell systems. The presence of several active fractions underscores that a whole-extract response may be more complex than the action of one molecule. This is preclinical support for further research, not proof that damaged nerves are restored in patients. Nor does it establish that the outcome necessarily resulted from the entire ERK → CREB → BDNF sequence.
Rao and colleagues studied dendritic branching in CA3 hippocampal neurons of adult rats after administration of fresh-leaf extract. Morphometry of stained neurons can measure process length and branch number—that is, a structural outcome. It is not equivalent to the birth of new neurons and does not prove improved human memory. Extending the finding to CA1, particular branch orders, mature mushroom spines, or inhibitory amygdala interneurons requires separate experiments rather than one generic “before and after” illustration.
Long-term potentiation (LTP) means a sustained increase in synaptic response after a specific stimulation protocol. AMPA and NMDA receptors participate in excitatory transmission and plasticity mechanisms, but dendritic morphology alone does not measure LTP. Gray and colleagues’ study in a mouse model of Alzheimer’s disease examined synaptic density and antioxidant responses. The finding is meaningful in the context of that model, but does not turn Centella into a dementia-prevention treatment.
The peptide can form aggregates associated with neurotoxic processes. A cell’s protective response to amyloid exposure may arise for different reasons: changes in oxidative stress, calcium homeostasis, mitochondrial function, or aggregate formation itself. Mook-Jung and colleagues’ work on asiaticoside derivatives supports discussion of experimental protection against amyloid toxicity. It does not establish universal direct binding of all triterpenes to amyloid or prove that they block the transition from α-helical to β-sheet structures in the human brain.
Membrane polyunsaturated fatty acids, including docosahexaenoic and arachidonic acids, are susceptible to lipid peroxidation. Malondialdehyde (MDA) and 4-hydroxynonenal are used as markers of related damage, although they do not capture the whole picture of oxidative stress. Superoxide dismutases SOD1/2, catalase, and glutathione peroxidases act at different points in antioxidant defense. Their reactions can be represented as follows:
Here GSH is reduced glutathione, GSSG its oxidized form, and ROOH an organic hydroperoxide. Glutathione reductase and NADPH help return GSSG to the reduced pool. These pathways explain why changes in enzyme activity or damage markers are biochemically interesting, but do not support a clinical promise to “neutralize free radicals.” An extract’s action may involve regulation of the cellular response rather than direct scavenging of every oxidant.
Subathra and colleagues’ studies concern age-related changes in nervous-tissue antioxidant status in an experimental setting. Discussing the mitochondrial permeability transition pore (mPTP), cytochrome c release, and activation of caspases 9 and 3 helps explain a possible route to cell death. However, a change in one antioxidant measure does not prove that a particular preparation suppresses this entire pathway. For clinicians, these data primarily generate hypotheses; decisions about neurodegenerative disease require outcomes in patients.
Clinical studies of Centella should not be treated as one set of equivalent trials. Some examined healthy volunteers’ immediate response to a sound, others assessed subjective symptoms without a placebo, and still others measured microcirculatory outcomes in patients with venous hypertension. Extracts, doses, durations, and endpoints differed. The doses below describe the interventions that were studied and are not a ready-made prescribing regimen for readers.
| Study | What was studied | Intervention used | How to interpret the result |
|---|---|---|---|
| Bradwejn et al., 2000 | 40 healthy volunteers; randomized, double-blind, placebo-controlled comparison | Single 12 g dose of gotu kola | Reduced startle-response amplitude; therapeutic efficacy for an anxiety disorder was not established |
| Jana et al., 2010 | 33 participants with anxiety symptoms; open study | 500 mg extract twice daily for 60 days | Scores improved over time, but the extract’s effect could not be reliably separated from other causes |
| Wattanathorn et al., 2008 | Small placebo-controlled study in healthy older volunteers | 250, 500, or 750 mg daily for two months | Positive results on selected cognitive and emotional measures; limited study scale |
| Pointel et al., 1987 | Venous insufficiency of the lower limbs | TECA, 60 or 120 mg daily | Vascular symptoms and venous-function measures; not evidence of cerebral protection |
| Cesarone et al., 2001, series of studies | Venous hypertension, flight-associated edema, diabetic microangiopathy | TTFCA in specific protocols | Results should be read separately by population and measured outcome |
| Farhana et al., 2016 | 48 patients with post-stroke cognitive impairment; quasi-experimental comparison | 750 or 1,000 mg extract daily versus 3 mg folic acid for six weeks | Within-group improvements without a significant difference in overall MoCA-Ina change; not proof of equivalence |
In Bradwejn’s study, the startle response decreased 30 and 60 minutes after administration. There was no substantial change in self-rated mood, heart rate, or blood pressure. This is a preliminary signal of neuropharmacological activity, but does not confirm treatment of an anxiety disorder or a specific GABA mechanism. The large single amount of plant material studied is also not a practical recommendation.
Jana’s study used an extract prepared with a 70% hydroethanolic solvent. The figure 70% refers to extraction, not to the amount of asiaticoside or total triterpenes in a capsule. Anxiety, stress, and related symptom scores improved over the observation period, but the open design cannot rule out participant expectations or natural fluctuation in symptoms. Without a specified, validated scale and its data, this finding should not be turned into a universal promise to “reduce anxiety by 26%.”
Wattanathorn and colleagues reported positive changes in selected measures of cognitive function and mood in older volunteers. The study can help shape future hypotheses, but its small size limits the precision and reproducibility of its findings. Different tests of memory, attention, and reaction speed cannot be combined into a single claim of “improved intelligence.” An electrophysiological measure such as P300, if studied, characterizes particular processes involved in stimulus processing; by itself it does not demonstrate dendritic growth or greater independence in daily life.
In Farhana’s post-stroke study, there was no significant difference in overall cognitive-score change among the three groups. A separate result in the memory domain is of interest, but the design and active comparator limit causal conclusions. The absence of a significant difference does not prove equal efficacy; that requires a specific design and prespecified criteria. This study does not support using Centella in place of rehabilitation or secondary stroke prevention.
Pointel’s study assessed venous-insufficiency symptoms and measures of venous function. The Cesarone series examined microcirculatory models using capillary filtration, skin gas exchange, and edema measures. These results provide a more concrete basis for discussing peripheral microcirculation than for advertising “blood flow to the brain.” At the same time, older, small studies do not automatically establish a role for the preparation in modern treatment or make it a first-line option.
The study of edema during air travel should be distinguished from studies of venous hypertension and diabetic microangiopathy. Reduced swelling does not mean that deep-vein thrombosis is prevented, and a change in skin oxygen does not demonstrate protection of the retina or kidneys. Without direct evidence, one should not promise complete prevention of edema, resolution of microangiopathy, or a particular percentage of improvement. For practical decisions, a person’s symptoms and functional outcomes matter more than an isolated favorable laboratory measure.
For someone considering Centella, the first question should be the intended purpose. Venous edema, anxiety, memory difficulties, and post-stroke recovery require different assessments; they do not form one indication “for nerves and vessels.” If a symptom is new, worsening, or interfering with daily life, choosing a supplement does not replace finding its cause. In particular, sudden one-sided leg swelling, pain, and redness require medical assessment; if shortness of breath or chest pain also occurs, seek emergency care.
When choosing a product, identify the botanical name, plant part, extraction method, and quantitative standardization. Ideally, the manufacturer should specify marker compounds, the analytical method, and controls for contaminants such as heavy metals, microbial load, and residual solvents where used. Batch quality matters whether the product is a powder or a purified fraction. “Natural” does not mean inherently safer, and leaf powder is not necessarily poor quality.
| Form | What the label may tell you | Where interpretation can go wrong |
|---|---|---|
| Above-ground plant powder | Amount of raw material and botanical identity | Plant mass is not the same as the mass of active triterpenes |
| Dry aqueous or hydroalcoholic extract | Solvent, herb-to-extract ratio, sometimes marker content | Extraction ratio and alcohol percentage do not establish asiaticoside content |
| Standardized extract | Amounts of individual glycosides or total triterpenes | The same total percentage may conceal different glycoside and acid profiles |
| Purified TECA/TTFCA fraction | Specification of triterpene composition | Data on a specific fraction do not apply to every Centella extract |
If a label truly specifies 20% of a particular total of triterpenes, then 500 mg of that extract nominally contains 100 mg of the measured compounds. The calculation is , where is the mass fraction in the specification. This is an arithmetic estimate of composition, not equivalence to 100 mg of purified clinical fraction: component profiles, dissolution, and exposure may differ. One studied intervention cannot be substituted for another on the basis of this calculation.
In some studies of peripheral vascular outcomes, 60–120 mg per day of a purified fraction was used. Anxiety and cognition studies used hundreds of milligrams of particular dry extracts, while a single-dose startle test used grams of plant material. These are different units for pharmacological comparison. A convenient universal table saying “this much for anxiety, that much for memory” would obscure the fact that no single evidence-based regimen has been established.
A positive result after several weeks does not establish a required 8–12-week course or a benefit from continuous use. For venous symptoms, anxiety, and memory, the outcome to be monitored and the point at which to reassess should be defined in advance. If clinical benefit is unclear, increasing the dose to match a laboratory mechanism is not justified. Long-term use requires more confidence in safety than a brief study involving several dozen participants.
Taking the preparation after food appeared in some research protocols, but this does not prove that absorption strictly depends on unsaturated fats. The aglycone’s lipophilicity alone does not establish a need to take it with a fatty meal or at a particular time of day. Follow the instructions for the specific product and consider tolerability; a promise to “activate bile” is not a sound guide. The available evidence does not define a separate chronopharmacological regimen for Centella.
High values in animal experiments characterize certain forms of acute toxicity, but do not rule out drug-induced liver injury, interactions, or problems with long-term human use. Reports of liver injury associated with Centella products should not be dismissed as “a slight rise in enzymes from a crude herb.” Jorge and colleagues described three cases involving jaundice and histological changes; one patient experienced recurrent injury after re-exposure. This case series does not establish the frequency of risk, but it is a clinically meaningful signal.
If jaundice, dark urine, marked weakness, itching, or pain in the right upper abdomen occurs, stop taking the product and seek medical assessment. After suspected liver injury, restarting it independently to “test tolerance” is not appropriate. Extra caution is needed in people with existing liver disease or those taking other potentially hepatotoxic agents. There is no reason to wait an arbitrary three to six months before taking this risk into account.
ALT, AST, alkaline phosphatase, bilirubin, and, where appropriate, GGT help assess liver injury, but normal baseline tests do not guarantee that injury will not develop later. No universal monitoring schedule has been established for all Centella users. Whether and when to repeat laboratory tests depends on the product, duration of use, symptoms, and baseline health. Small studies reporting no adverse events cannot replace assessment of rare risks.
Hypersensitivity to product ingredients is possible; topical products and oral preparations have different exposure profiles. Available evidence is insufficient to recommend self-directed use during pregnancy or breastfeeding, or in children. Historical claims that extremely high doses affect the uterus should not be presented as an established human contraindication mechanism. The practical limitation is the lack of adequate safety data.
Possible increases in drowsiness when combined with sedatives should be treated as a precaution, not a quantitatively established interaction. There is no basis for promising that Centella has no effect on CYP1A2, CYP2D6, CYP3A4, or CYP2C9 in every clinical setting. Laboratory results, even when available, do not replace interaction studies of a specific product with a medicine. It is useful to discuss the product and a complete list of medicines with a physician or pharmacist; independently “titrating” prescribed benzodiazepines because of a supplement is inappropriate.
Centella’s popularity in combinations reflects its position between several areas of research. One ingredient is proposed “for synapses,” another “for blood flow,” and a third “for calm,” after which the combination is credited with an overall effect. Synergy, however, means an interaction that must be demonstrated experimentally or clinically. Simply putting plausible mechanisms side by side does not prove that a combination is better; adding components also makes outcomes and tolerability harder to assess.
Both bacopa and Centella have been studied in relation to cognitive function, but they contain different compounds and do not form an established pair for “synaptic neurogenesis.” Even if cholinergic or serotonergic mechanisms are discussed for a particular bacopa extract, this does not show that Centella offsets its sluggishness or enhances memory. Recommending 300 mg of each plant without standardization does not define a reproducible intervention. It is more useful to distinguish their compositions and evidence than to repeat such a combination.
The article on bacopa, bacosides, and synaptic plasticity explores the separate topic that gotu kola does not replace. Reading the two side by side can clarify their differences, but does not turn the extracts into a recommended combined course.
The standardized ginkgo extract EGb 761 has its own composition and proposed mechanisms, including those related to PAF and vascular biology. This does not support a neat division of roles in which “ginkgo fixes arterioles and Centella fixes venules” to produce optimal brain perfusion. Skin vascular measures for Centella do not demonstrate a cerebral benefit from the combination. Risks and interactions associated with the second ingredient remain separate concerns even when both products are plant-derived.
For comparison, see the article on ginkgo, PAF, and cerebral blood flow. Combining products requires its own clinical rationale, not mechanical addition of research doses. In particular, such a combination should not replace assessment of vascular symptoms or standard secondary prevention.
L-theanine is studied for effects on stress measures, attention, and brain electrical activity; phosphatidylserine is studied for membrane functions and selected stress responses. These topics do not validate the formula “theanine blocks excitation, Centella raises GABA, phosphatidylserine stabilizes cortisol—therefore the result is high concentration with absolute calm.” An EEG alpha rhythm is not equivalent to treatment of anxiety, and cortisol varies physiologically over the day; it is not a measure that should always be lowered. The efficacy and safety of a three-part combination cannot be inferred from separate studies of each ingredient.
The articles on L-theanine and the glutamate system and on phosphatidylserine, membranes, and cortisol explain these distinctions. For an anxiety disorder, symptom severity, sleep, daily functioning, and available treatments matter more than informal labels such as “manager syndrome.” Adding several substances independently also makes it harder to identify which one caused drowsiness, gastrointestinal discomfort, or another adverse reaction.
Centella remains scientifically interesting precisely because its chemistry and experimental biology are substantive. The glycoside–aglycone pairs make it possible to study biotransformation; cell experiments examine collagen regulation and plasticity; vascular studies measure specific changes in peripheral microcirculation. These lines of research need not be dismissed because they have not established a universal therapeutic effect. Each should, however, remain at the level of evidence at which the finding was obtained.
For practical decisions, the most important considerations are identifying the preparation, recognizing the limits of clinical data, and accounting for safety. A small positive memory test does not prove dementia prevention; a reduced startle response does not establish treatment of an anxiety disorder; and reduced edema does not mean protection from thrombosis. Keeping these outcomes distinct gives professionals a more useful picture and gives users a basis for informed decisions without promises that current evidence does not support.
No. The shared traditional name “brahmi” can cause confusion, but Centella asiatica and Bacopa monnieri are botanically distinct and contain different triterpene complexes. Substituting one for the other based on a common name does not ensure comparable effects.
There are preliminary signals, including a startle test in healthy people and an open observation of symptoms. These study designs do not establish reliable therapeutic efficacy and do not justify replacing prescribed treatment with Centella.
The mass of plant powder, dry extract, and purified fraction describes different materials. The glycoside and acid composition, quantitative standardization, and dosage form matter; a milligram figure alone is not enough to compare them.
The table describes interventions used in particular studies, not a universal dosing regimen. They differed in form, composition, and purpose; a decision requires considering symptoms, safety, and whether the findings apply to the product in question.
The studies reviewed focused mainly on venous symptoms and peripheral microcirculation. Skin oxygen tension, fluid filtration, and edema are not direct measurements of blood supply to the brain.
Preclinical plasticity findings and selected cognitive outcomes do not prove dementia prevention. The post-stroke study has design limitations and does not justify replacing rehabilitation or secondary stroke prevention.
Absolute safety cannot be promised: clinical cases of liver injury have been reported. If jaundice, dark urine, or other concerning symptoms occur, stop taking it and seek medical assessment; the frequency of this risk is unknown.
An enzyme experiment shows a response in a research system, but does not measure exposure in the human brain or demonstrate improved everyday functioning. Those conclusions require separate pharmacokinetic and clinical evidence.