Licorice root has been widely used as herbal medicine dating back to at least 500 BC.1 Traditional applications across diverse cultures include as both a demulcent and an antiinflammatory, often used to soothe respiratory or gastrointestinal (GI) symptoms. Modern botanical applications of the herb continue this tradition, with recommendations including the treatment of gastric ulcers, bronchitis, cough, and dyspepsia.2 While licorice is indispensable in these and many other herbal applications, it comes with a slight but measurable risk of side effects when used as a whole-root extract.
One of the constituents in licorice root is glycyrrhizin, a triterpene glycoside that has been shown to have aldosterone-like effects.3 High doses or chronic consumption of glycyrrhizin from licorice root can create side effects that mimic a hypermineralcorticoid state. Clinically, this presents as electrolyte imbalances, hypertension, and possible cardiac abnormalities. While side effects are not common and risk is further minimized with monitoring of the patient and knowledge of proper dosing, removal of the glycyrrhizin compound eliminates the risk altogether. Today, deglycyrrhizinated licorice (DGL), which is specifically marketed for GI complaints, is a popular over-the-counter supplement.
While DGL is popular for relieving gastric irritations, clinical trial support is weak. DGL exemplifies a widely accepted natural agent that “works” when assessed empirically but has yet to meet criteria for inclusion in an evidence-based, conventional medical paradigm.
Modern medical history of licorice root begins in 1946, with the first article about “licorice juice” for gastric ulcers, published by Dutch physician F. E. Revers.4 Two years later, he published a report on the use of licorice juice in the treatment of duodenal ulcers.5 While efficacy was demonstrated, side effects of edema, headaches, and cardiac complaints were also observed in these early applications of crude licorice extract. In 1950, J. G. Borst and colleagues proposed that the side effects were due to retention of sodium and water along with increased excretion of potassium and suggested the constituent glycyrrhizin (glycyrrhizinic acid) was to blame.6 By 1952, Revers had tried an extract of licorice with glycyrrhizic acid removed in a patient with peptic ulcer.7 He reported that removal of glycyrrhizin from the licorice extract resulted in the eradication of observed side effects. Reintroduction of whole root consumption in the same patient confirmed the hypermineralcorticoid side effects were indeed due to glycyrrhizin.
Support for the Beneficial Effects of DGL
Much of the in vitro data on licorice root is on glycyrrhizin, with less information available on the other constituents of licorice. For example, glycyrrhizin and its derivatives have been shown to effectively kill Helicobacter pylori strains, including those that are clarithromycin-resistant or metronidazole-resistant.8 Various flavonoids found in DGL have also been shown to be bactericidal toward H pylori and may account for some of the antiulcer activity preserved with the removal of glycyrrhizin.9 In addition, DGL has demonstrated the ability to increase mucin, possibly by increasing the number of mucous secreting cells in the stomach.10
Deglycyrrhizinated licorice exemplifies a widely accepted natural agent that “works” when assessed empirically but has yet to meet criteria for inclusion in an evidence-based, conventional medical paradigm.
Perhaps the most suggestive evidence supporting DGL’s healing effect of GI mucosa is demonstrated through protection from aspirin-induced GI bleeding. Rats given a large dose of DGL (2,000 mg/kg orally) demonstrated a reduction in aspirin-induced gastric lesions when the DGL was coadministered but not when it was given before aspirin ingestion. In this same study, DGL delivered intraperitoneally showed reductions of the gastric lesion scores, implying there may be a systemic component to the therapeutic effect. One human, placebo-controlled crossover study demonstrated that fecal blood loss induced by aspirin (375 mg 3 times/d) was significantly reduced with coadministration of DGL (175 mg/dose aspirin).11
Clinical Evidence to Date
Four different licorice medicaments are used in clinical studies on ulcerative diseases, and differentiation among them is essential to an accurate review of the literature. The 4 include crude extracts of the whole plant, DGL combined with other agents (ie, Caved-S), DGL alone, and drug analogs of glycyrrhizin derivatives (eg, carbenoxolone, enoxolone). Many reviews, references, and marketing materials on DGL report on these 4 forms interchangeably. A closer look at this indiscriminate use of clinical trial data shows that they are not equivalents, and extrapolation of studies using agents containing glycyrrhizin or glycyrrhizin-like compounds are not relevant when compiling trial data on DGL.
Crude extracts of licorice are a traditional use of the plant and led to the initial interest of Revers and others in proving its therapeutic use for ulcers. Since the crude extract contains glycyrrhizin, and glycyrrhizin has been shown to heal ulcerative lesions as a single agent, the efficacy of licorice crude extract can be assumed to be at least partly due to the glycyrrhizin content. DGL by definition is lacking glycyrrhizin; therefore its efficacy in ulcerative complaints must be proven independently from data on crude extracts.
Carbenoxolone (Biogastrone™), a synthetic analogue of the biologically active metabolite of glycyrrhizin, β-glycyrrhetinic acid, is an approved drug in the United Kingdom and Europe for ulcers and aphthous stomatitis. Carbenoxolone and the newer analogue, enoxolone, both have a well-documented risk of inducing symptoms of a hypermineralcorticoid state. The induction of side effects with a glycyrrhetinic acid analogue is to be expected, as glycyrrhizin and its metabolites have long been identified as the agents inducing these side effects.5 As synthetic analogues of the very constituent that was removed from DGL, there should be no extrapolation of clinical trial data derived from carbenoxolone or enoxolone as support or repudiation of DGL’s efficacy in ulcerative complaints.
That leaves clinical trials with DGL as a single agent or DGL in combination with other compounds as the only relevant trials in an assessment of the clinical evidence. While trials of DGL as a single agent are ideal, the majority of published trials are on a combination product, Caved-S, in which DGL comprises 42% of the weight of each capsule. The components of Caved-S per tablet are DGL (380 mg), bismuth subnitrate (100 mg), aluminium hydroxide gel (100 mg), magnesium subcarbonate (200 mg), sodium bicarbonate (100 mg), and frangula bark (30 mg). It is essential to note that the term DGL is often used synonymously with Caved-S in the primary publications as well as in current reviews and marketing materials. For this reason, the form of the intervention used is included in the table summarizing the available trial data (Table).
There have been no new clinical trials on the use of DGL for ulcers in more than 20 years. The trials that have been published on Caved-S and DGL, as a whole, have been criticized for their lack of rigorous design. Most trials were small (the largest trial evaluated 96 patients). Not all trials were blinded, and many did not include a placebo control group or the placebo was not adequately matched to the intervention group. Further, the natural course of ulcerative disease is one of exacerbation and remission, leading to a significant number of results in which participants experienced improvements regardless of the intervention. This makes a statistically significant improvement in the rate of healing difficult to obtain in trials with small numbers of patients.
A comparison of the results derived from these trials provides conflicting results without trend or obvious conclusions able to be drawn. A cursory look at the summary of clinical trials (Table) shows no standard dosage, delivery, trial duration, or methodology used in the majority of trials. How efficacy was assessed varied between the trials as well with questionnaires, clinical observation, barium radiography, or endoscopic assessments used.
Despite these many shortcomings, the results of some trials are intriguing enough to merit deeper exploration. The most impressive trial demonstrating benefit, published in 1973, enrolled 40 patients with more than 6 relapses of duodenal ulcer in the prior 12 months. The study clinic was also a surgical clinic, and all patients had been referred to the clinic for consideration of duodenal resection from a community physician. All patients had symptoms of unrelenting vomiting and/or pain on enrollment. Since these patients were severely ill, no placebo was offered; instead 2 levels of moderate and intense intervention with Caved-S was used. The moderate intervention included 8 tablets 8 times daily for 8 weeks, and intense intervention was 12 tablets 8 times daily for 16 weeks. While none of the patients required surgical intervention in the ensuing year, the cohort consuming the higher dosage for longer duration had significantly fewer symptomatic events during that year as well. Both groups used a dosage of Caved-S significantly higher than most trials, with a total DGL consumption of 4,560 mg per day in the high-dose group.12
In another trial by W. Larkworthy and colleagues, the intervention closely matched the above intervention of over 4 g daily, except it used DGL as a single agent and there was a mastication group to see if chewing conferred any benefit.13 This trial, which assessed patients endoscopically, did not show statistical difference in the healing of ulcers between the intervention and placebo groups. This may be due to the suggestion of 5 meals per day for all study participants. This frequency of eating increased the salivation of all participants, which has been independently proven to have an antiulcer effect.14 Further trials assessing the efficacy of these high doses are warranted to definitively answer the question of statistically significant effects of DGL.
As an aside, Larkworthy and colleagues suggest that further trials must include endoscopic assessment, as symptoms do not always correlate with healing of ulcers. This suggests that licorice’s pain-relieving abilities should be assessed separately from its ability to visibly increase the rate of healing in future trials.
Why is there such a lack of trials in the past 20 years? With the advent of H2 blockers in the late 1970s and early 1980s, the medical management of ulcers changed dramatically, followed by the even more dramatic shift toward use of antibiotics after the widespread acceptance of the role of H pylori in ulcer etiology in most patients in the 1990s. DGL was no longer being studied for up-front application in the healing of ulcers but as maintenance therapy alone or with H2 blockage agents. Caved-S was demonstrated to be equivalent to cimetidine in the prevention of relapse in a 2-year follow-up study undertaken by A. G. Morgan and colleagues.15 Another study in 1987 showed DGL decreased the rate of relapse in patients taking ranitidine for gastric ulcers.16 The introduction and widespread usage of the proton pump inhibitors (such as omeprazole) further clouds the issue. This drug class has a more powerful acid suppression effect than the older class of H2 blockers, and there is no data on the use of DGL with proton pump inhibitors. Nevertheless, cumulatively the data suggest DGL may have a role in protecting the erosion of mucosa and decreasing relapse; therefore it deserves continued consideration for adjunctive management in these patients.
Table. Chronological Listing of Clinical Trials on DGL and Ulcersa
|Journal Citation||Form of DGL Used||Dosage Used|
|Summary of Design ||Summary of Outcome ||Authors’ Conclusions|
|Gut. 1968;9(1):48-51||Caved-S ||2 tablets, chewed 3x/d after meals||8 wk, with 4-wk crossover||Double-blind RCT of 6 gastric and 48 duodenal ulcer patients. Crossover was done in all gastric ulcer patients and 15/48 patients with duodenal ulcers. All patients were classified as “good,” “improved,” or “not improved” based on reported pain levels. Radiological imaging was also used before and after intervention. Usage of antacid was recorded in all patients.||All patients with gastric ulcers achieved symptomatic and radiographic improvements. There was less use of antacid in the Caved-S group and a trend toward improvements in observed symptoms in the duodenal ulcer intervention group. Follow-up imaging of duodenal ulcer patients suggests a spasmolytic effect of DGL at the level of the duodenal bulb.||“Duodenal ulcer patients showed marked improvements, with radiographical improvements in a few cases.” All gastric ulcer patients showed healing of ulcers. There appears to be a spasmolytic effect of DGL in patients with duodenal ulcers.|
|Gut. 1969; 10(4):299-302||DGL 380mg/ capsule||2 capsules, 3x/d after meals||4 wk||Double-blind RCT of 33 patients with radiographic evidence of gastric ulcerations greater than 10 mm2. Healing of ulcers was assessed through radiography. ||The group receiving DGL had an average reduction in ulcer size of 76%, vs 34% for the placebo group. No side effects were reported.||DGL represents a therapeutic alternative to carbenoxolone without the risk of mineral corticoid side effects. |
|Gut. 1971;12(6):449-451||Caved-S||2 tablets, chewed 3x/d after meals ||4 wk||Double-blind, placebo-controlled study of 47 patients with active duodenal ulcer for a minimum of 6 mo. Inclusion included baseline radiographic evidence of deformed duodenal bulb or ulcer niche as well as abdominal pain. Assessment was through symptom questionnaire. Intervention consisted of 4 wk of treatment.||There was no significant difference in symptom reports between the 2 groups. Electrolytes, complete blood count, blood urea nitrogen, and urinalysis were unchanged in both groups. No side effects were reported. ||There is no advantage to the addition of DGL in symptom management in patients with duodenal ulcers. |
|Br Med J. 1971;3(5773): 501-503||DGL, 380 mg/ capsule||2 capsules, 3x/d after meals||6 wk||Multicenter RCT in the United Kingdom. 90 men with recurrent duodenal ulcers. Symptoms assessed through patient reports and practitioner observation, fortnightly. Frequency of alkali medication for symptom relief was also tracked. Serum pepsinogen measured in both groups.||No measurable difference in subjective pain reports. No difference in the amount of alkali medication consumed for symptomatic relief. No difference in practitioner reported observations of patient’s relief of symptoms. No change in serum pepsinogen levels.||The difference between carbenoxolone and DGL is statistically not significant. As it has no side effects, DGL represents a safe alternative to carbenoxolone.|
|Practitioner. 1973;210(260): 820-823||Caved-S||8 tablets, 8x/d for 8 wk, or 12 tablets, 8x/d for 16 wk ||1 y||Double-blind trial using 2 different doses of Caved-S in 40 duodenal ulcer patients with unremitting pain and greater than 6 relapses in 12 mo. Relapse was defined through the presence of pain, and relapse-free defined by its absence. All patients had been referred for surgical intervention. ||None of the patients required surgical intervention during the year of follow-up. Most of the patients receiving 8 tablets/d did have a relapse of symptoms within the y. Patients receiving 12 tablets/d did not suffer relapse. This difference was statistically significant.||Higher doses of Caved-S confer a greater protection from relapse of duodenal ulcer symptoms. Caved-S demonstrated efficacy even in patients with severe, relapsing duodenal ulcers who were referred for surgical intervention. |
|Gut. 1973;14(9):711-715||DGL, 380 mg/capsule||2 capsules, taken orally 3x/d||8 wk, with 4-wk crossover||Double-blind, placebo-controlled trial of 68 patients with radiographically confirmed gastric ulcer(s). Assessment was done through radiographic follow-up.||There was no statistical difference in the healing of the intervention group vs placebo. ||The location of the ulcer in the stomach, as well as the initial size, influences the healing time. With these considerations, previous studies showing efficacy may have had placebo groups unmatched to interventional groups.|
|Br Med J. 1977;2(6095): 1123||DGL 450 mg/ capsule (Ulcedal) or 450 mg/block chewing gum||2 capsules, 5x/d or 2 blocks, chewed for 30 min 5x/d between meals||8 wk||Controlled trial of 34 patients with duodenal ulcers. DGL consumed as capsules or chewing gum vs placebo of either capsules or chewing gum of inert substance. Endoscopy was used to assess efficacy.||There was no difference in healing of ulcers between the intervention group vs placebo.||The results do not support the concept of DGL usage in duodenal ulcer. Saliva introduced through chewing did not confer any benefit to healing either. Since symptoms do not correlate with healing, trials should be endoscopically controlled.|
|Br Med J. 1978;1(6106):148|
DGL (Ulcedal) 450 mg/capsule
|5 capsules/d||2 y||Double-blind, placebo-controlled trial of 33 patients with healed gastric ulcerations. Assessment was through gastroscopy and barium imaging every 6 mo. Patients were followed for up to 2 y.||Relapse rates were 45% and 59% for the DGL and placebo groups respectively. This was not statistically significant. ||Although the results did not reach statistical significance, the need for prophylaxis in this patient population lends itself to the use of DGL given the trend of less recurrence combined with a low toxicity profile. |
DGL (Ulcedal) 450 mg/capsule
|2 capsules 4x/d||4 wk||Double-blind RCT of 96 patients with endoscopy-verified gastric ulcers. Assessment was through repeat endoscopy and radiological assessment. Ulcer healing was defined as re-epithelialization of lesion on visualization through gastroscopy. Radiographical healing was defined as disappearance of the ulcer crater on repeat imaging.||Median percentage change in the ulcer area was comparable in both groups. Symptom improvement was also comparable in the 2 groups. Endoscopic and radiographic results had little concordance.||This trial was designed to show a statistically significant benefit of DGL if there was a doubling of the assumed placebo-healing rate of 30%. It did not show this; therefore there is no justification for continued use of DGL. |
Abbreviations: DGL, deglycyrrhizinated licorice; RCT, randomized controlled trial.
aOnly clinical trials published in English with the full text available in electronic or hard copy are included above.
DGL is specifically used to promote the healing of ulcers and reduce ulcer-related symptoms. While it appears to be efficacious in this regard, the only clinical trials published using DGL as a treatment are poorly designed and/or were conducted more than 20 years ago. Well-designed, larger clinical trials may be helpful to delineate optimal dosage of DGL, as well as to determine which patient populations can derive the most benefit. Meanwhile, the incentive for the continued use of DGL will remain the anecdotal relief of symptoms—a strong impetus for both the clinician and the patient experiencing pain.