Deglycyrrhizinated licorice (DGL): mucus, secretin and clinical evidence

DGL is licorice extract processed to reduce glycyrrhizin. The rationale is to retain selected constituents while lowering the exposure responsible for licorice-induced sodium retention, potassium loss and hypertension. This is a change in composition, not proof that an extract cannot influence mineralocorticoid physiology or cause other adverse effects.

The gastric mucus–bicarbonate barrier provides a plausible biological context for studying licorice extracts. It does not establish that DGL increases human MUC5AC, stimulates secretin, heals ulcers or replaces proton pump inhibitors (PPIs). Human evidence is mixed: several older controlled ulcer trials were negative, whereas short trials of a specific flavonoid-rich extract reported improvement in functional dyspepsia and mild to moderate reflux symptoms. Product, indication, outcome and study duration matter.

What is removed, and what can remain?

Licorice roots from species such as Glycyrrhiza glabra and G. uralensis contain triterpenoid saponins, flavonoids and polysaccharides. Their proportions depend on species, growing conditions and extraction. DGL describes glycyrrhizin reduction; it does not define a uniform flavonoid mixture, a universal residual threshold or a pharmaceutical formulation.

The molecular formulas above were checked in PubChem. Confusing glycyrrhizin with glycyrrhetinic acid obscures both extraction and metabolism. Claims that processing preserves every useful compound unchanged also require analytical evidence. Solvent choice, purification and drying can alter the fraction that reaches the finished product.

In Raveendra and colleagues’ dyspepsia trial, GutGard was specified as glabridin ≥3.5%, total flavonoids ≥10% and glycyrrhizin ≤0.5% by weight. The methods of Raj and colleagues’ 2025 reflux trial specify glycyrrhizin ≤3.0%, alongside the same minimum glabridin and total flavonoid percentages. These are reported specifications from different studies, not a universal definition of DGL and not measurements of every commercial batch.

Why licorice can produce apparent mineralocorticoid excess

Intestinal microbial metabolism converts glycyrrhizin into glycyrrhetinic acid and related metabolites. The key mechanism is inhibition of renal 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2), which normally converts cortisol to cortisone. This conversion protects the mineralocorticoid receptor (MR) from cortisol.

When that protection is weakened, cortisol can activate MR in renal tubular cells. Increased sodium reabsorption and potassium loss can cause fluid retention, hypertension, hypokalemia and metabolic alkalosis. Volume expansion suppresses renin and aldosterone; this is why low aldosterone does not exclude a licorice-related mineralocorticoid phenotype. It is not a mechanism in which glycyrrhizin must directly bind an “aldosterone receptor.” Stewart and colleagues’ 1987 human study linked licorice exposure to altered cortisol metabolism and sodium retention.

Reducing glycyrrhizin should reduce exposure to this pathway, but the word DGL alone does not quantify it. In a 2024 randomized crossover study of 28 healthy participants, licorice providing 100 mg glycyrrhizic acid daily for two weeks increased home systolic blood pressure and suppressed renin and aldosterone. This was ordinary licorice, not DGL. It shows that a commonly quoted exposure limit is not a guarantee of no physiological effect, not that every DGL product produces the same response.

A simple mass calculation clarifies the issue:

$$m_{\mathrm{glycyrrhizin}} = m_{\mathrm{extract}}\times w_{\mathrm{glycyrrhizin}}$$

Here, both masses use the same unit and the mass fraction is a decimal. As an arithmetic example, 1,500 mg of extract containing 1% glycyrrhizin corresponds to 15 mg glycyrrhizin. This is neither a recommended dose nor a proven safe threshold. Actual exposure also depends on the analytical result, daily intake, duration, metabolism and other licorice-containing foods or supplements.

The gastric barrier: mucus is more than a coating

Normal gastric surface and foveolar epithelial cells produce mucus associated with MUC5AC; MUC6 is characteristic of glandular mucus. They should not be described as ordinary intestinal goblet cells. Mucins form a hydrated gel that helps organize the protective environment next to the epithelium. Human gastric organoid work by Lyon and colleagues in 2026 identified MUC5AC, MUC6 and additional stomach-specific proteins, while also showing that an engineered mucus preparation did not reproduce every physical property of native mucus.

Bicarbonate secreted toward the epithelial surface buffers hydrogen ions within the mucus. The simplified neutralization reaction is:

$$\mathrm{H}^{+}+\mathrm{HCO}_{3}^{-}\longrightarrow\mathrm{CO}_{2}+\mathrm{H}_{2}\mathrm{O}$$

The resulting local pH gradient can coexist with an acidic gastric lumen. It depends on secretion, mucus organization and blood flow; it is not a fixed mucus thickness or a universal surface pH. Phillipson and colleagues measured juxtamucosal pH with microelectrodes in rats and demonstrated the importance of epithelial bicarbonate transport. That experiment did not test DGL in people.

The older DGL experiment by van Marle and colleagues studied renewal of rat stomach epithelium. It found changes in labeled mucus-secreting gland cells and proposed altered differentiation and mucus formation. It did not establish a 1.8–2.4-fold rise in human MUC5AC, a 50–70% extension of epithelial lifespan or a guaranteed 200 μm barrier. Forestomach and glandular stomach findings also differed, limiting a simple “more proliferation everywhere” interpretation.

Secretin belongs to a separate intestinal–pancreatic circuit

Secretin is produced by duodenal S cells and participates in the response to acid entering the small intestine. Its endocrine signal reaches pancreatic ducts through blood and promotes bicarbonate-rich fluid entering the duodenum. Primary localization and secretion experiments, including Polak and colleagues and Magee and Naruse, support this physiological circuit. It should not be drawn as a direct arrow from gastric mucus to the pancreas.

A normal secretin pathway does not prove that oral DGL stimulates S cells in humans. The clinical trials discussed here did not establish increased secretin concentration or a DGL-dependent pancreatic bicarbonate response. Gastric surface buffering and duodenal neutralization are related protective processes with different compartments.

Pepsin, prostaglandins and epithelial restitution

Pepsin contributes to protein digestion under acidic conditions and may injure exposed tissue when protective barriers fail. Proposed experimental effects on pepsin or pepsinogen do not establish a predictable percentage reduction in human enzyme activity, protection of the larynx or an effect equivalent to acid suppression. The 2025 reflux trial omitted its planned 24-hour gastric pH measurements, so it cannot directly demonstrate that DGL leaves gastric acidity unchanged.

Prostaglandins such as PGE₂ and PGI₂ help regulate mucosal defense, perfusion and secretion. NSAIDs impair prostaglandin-dependent defenses, but this does not demonstrate that DGL reliably restores prostaglandin production during NSAID treatment. Likewise, epithelial restitution is rapid lateral migration across an intact basement membrane, distinct from later cell proliferation. Growth-factor experiments in primary gastric cultures show roles for EGF-related signaling and trefoil peptides. They do not establish a clinically demonstrated DGL→EGF→integrin α5β1→fibronectin repair sequence.

Flavonoids: mechanistic interest and its limits

Flavonoids are not one pharmacological substance. An isolated compound at an experimental concentration, a standardized extract and an oral supplement may differ in solubility, metabolism, tissue exposure and accompanying constituents. Glycosides can undergo intestinal hydrolysis; aglycones may then be conjugated. A local luminal hypothesis does not establish zero systemic absorption, and a blood concentration does not establish the concentration inside gastric mucus.

Liu and colleagues studied isoliquiritigenin in macrophages and a mouse model of acute lung injury. They found AMPK/Nrf2/ARE-related signaling, changes in oxidative stress and suppression of inflammatory pathways involving NF-κB and NLRP3. This is useful evidence about an isolated chalcone in experimental systems. It is not a clinical DGL study of gastric mitochondrial repair, nor proof of a particular Keap1 cysteine reaction in the stomach.

Glutathione-dependent lipid protection is discussed further in the article on GPX4 and ferroptosis. That biology helps interpret oxidative injury, but does not mean DGL has been shown to prevent gastric ferroptosis in people. Neither antioxidant labeling nor radical scavenging in a laboratory is a clinical healing outcome.

H. pylori: growth, adhesion and eradication are different endpoints

Fukai and colleagues tested isolated licorice flavonoids against H. pylori in vitro, including strains resistant to selected antibiotics. Activity in that setting does not establish clinical activity against all resistant strains or show that resistance can never develop.

Asha and colleagues studied a flavonoid-rich whole extract, GutGard. They reported growth inhibition and effects on protein synthesis, DNA gyrase and dihydrofolate reductase. The reported enzyme IC₅₀ values, 4.40 μg/mL for gyrase and 3.33 μg/mL for dihydrofolate reductase, belong to that experimental extract assay, not automatically to isoliquiritigenin or to oral tissue concentrations. Importantly, this study found no significant inhibition of adhesion at the tested concentrations.

Wittschier and colleagues investigated a different system: aqueous licorice extracts and polysaccharide fractions on human gastric tissue sections. An aqueous extract at 1 mg/mL reduced adhesion; purified polysaccharides were not directly cytotoxic to the bacterium. These results support a preparation-specific antiadhesive hypothesis. They do not demonstrate universal blockade of BabA/SabA or clinical eradication by DGL.

In Hajiaghamohammadi and colleagues’ 2016 trial, 120 patients were enrolled to compare an antibiotic/PPI regimen with the same regimen plus D-Reglis licorice. Reported eradication was 83.3% versus 62.5%, not 67.5%. The full text contains inconsistent completion counts, and the retrieved methods do not quantify residual glycyrrhizin. This is a limited adjunct study, not a demonstration of DGL monotherapy or a general antibiotic-sparing protocol. Current ACG guidance emphasizes susceptibility-aware treatment and post-treatment confirmation of cure. A licorice supplement should not substitute for those steps.

What controlled human trials actually found

The doses in this table describe studies. They are not a treatment schedule for ulcers, reflux, children or people taking aspirin.

Functional dyspepsia: a signal, not a universal effect

The 2012 trial used Rome III, not Rome IV, criteria. It evaluated a ten-symptom score and the short-form Nepean Dyspepsia Index, not the Gastrointestinal Symptom Rating Scale. Total scores favored GutGard at days 15 and 30, but early satiety was comparable to placebo. Several authors were affiliated with the extract manufacturer, and the randomization sequence was generated there. These features do not erase the result; they strengthen the case for independent replication and longer follow-up.

Reflux: distinguish symptoms from healing and mechanism

In the 2025 trial, the baseline-adjusted GERD-HRQL comparison at day 28 favored GutGard (p=0.014). Data at that time were available for 78 participants in the extract group and 87 in the placebo group: 165 of 200 randomized participants. Rescue famotidine use and gastric emptying did not differ significantly. Planned lower esophageal sphincter pressure and 24-hour gastric pH measurements were not performed because of pandemic restrictions.

The study was sponsored by Natural Remedies, which provided technical input during design and protocol development; the authors reported no sponsor role in data collection, analysis or reporting. It did not compare DGL directly with a PPI, establish noninferiority, document healing of severe erosive esophagitis, or assess prevention of Barrett-associated cancer. No clinically significant rebound was observed one week after stopping the study product in either group, which is not proof that DGL prevents rebound after long-term PPI withdrawal.

Formulation, quality and safety

Chewable tablets, capsules, powders and mouthwashes are different exposures. Saliva contributes to normal mucosal physiology, but there is no demonstrated rule that DGL must be chewed or that swallowed capsules are inactive: the two GutGard symptom trials used capsules. An oral mouthwash report cannot establish an effective swallowed dose, and the unpleasant or acceptable taste of a product does not determine its efficacy.

Useful product information includes botanical identity, extract amount, quantified residual glycyrrhizin, constituent standardization, batch-specific testing and excipients. HPLC can quantify constituents when validated, but a certificate must identify the analyte, method, units and actual lot. “DGL” or “natural” on a label is not that certificate. A sugar-free claim also requires the full ingredient list; sugar alcohols and other excipients can affect tolerability. Comparative adherence percentages or diabetes suitability should not be invented.

NCCIH states that licorice without glycyrrhizin might be safe for up to four months. That cautious statement does not establish safety of every product called DGL, long-term use, pregnancy, breastfeeding or pediatric dosing. Short trials with selected adults cannot exclude uncommon adverse events or clinically important interactions. Allergy, gastrointestinal symptoms, residual glycyrrhizin and other ingredients remain relevant.

For ordinary licorice, high blood pressure, heart or kidney disease and high salt intake can increase concern even at lower exposures; corticosteroid interactions have been reported. DGL cannot therefore be declared universally compatible with antihypertensive drugs, diuretics, cardiac therapy or corticosteroids. Decisions depend on the actual composition and medicines involved. A reduction in an established hazard is different from complete endocrine or cardiovascular safety.

Claims that the evidence does not currently justify

Bile acids also act as signaling molecules through FXR and TGR5, as explained in the gut–liver bile acid axis. This broader physiology is a reason to distinguish signaling from bile-acid tissue injury, not evidence that DGL controls bile transporters or prevents reflux injury.

Comparison with established treatments and PPI withdrawal

PPIs suppress acid production through the gastric proton pump; H₂-receptor antagonists reduce histamine-mediated secretion; antacids neutralize existing acid; mucosal agents have their own pharmacology and evidence. DGL should not be assigned all these actions. Misoprostol and rebamipide also have distinct indications and adverse-effect profiles. Comparisons require matching populations, clinically relevant endpoints and a direct trial, rather than a table declaring a botanical universally safer or cheaper.

ACG guidance retains PPIs as the principal medical treatment for GERD. Observational associations between PPI use and adverse outcomes do not by themselves show that a PPI caused every reported problem or that DGL is a safer effective substitute. The AGA de-prescribing update recommends reviewing the indication and bleeding risk. Severe erosive disease, peptic stricture, Barrett’s esophagus and high bleeding risk can be reasons to continue treatment. Transient symptoms from rebound acid hypersecretion can occur after withdrawal.

No verified DGL-supported six-week taper, universal monitoring interval or 85% withdrawal-success rate follows from the evidence reviewed here. Alarm symptoms such as difficulty swallowing, gastrointestinal bleeding, persistent vomiting or unexplained weight loss require clinical assessment, rather than escalating a supplement. These considerations also prevent a symptom improvement score from being mistaken for a diagnosis or ulcer healing.

How this article was checked

Targeted verification was performed on 8 October 2026 using primary research and official sources. Full texts were checked for the Raveendra dyspepsia trial, Raj reflux trial and Hajiaghamohammadi adjunct trial, including methods, results and relevant limitations. For the other primary papers, available indexed abstracts and bibliographic records were checked; PubChem chemical records and NCCIH guidance were consulted directly. Access to an abstract was not presented as access to the full paper.

This is a narrative, focused assessment, not a systematic review or a formal evidence-grade exercise. It does not establish that all relevant studies were found. The distinction between isolated compounds, experimental models, defined extracts and clinical outcomes is central to its interpretation. DGL remains an investigational, composition-dependent option with some symptom-trial signals and substantial unresolved questions about mechanisms, durability and broader clinical uses.