Acute liver injury associated with the use of herbal preparations containing glucosamine: three case studies

Abstract

The use of complementary and alternative medicines is becoming increasingly popular in Western society. As a result the number of reported adverse reactions is increasing. Glucosamine is a herbal remedy commonly used to ease joint pain in osteoarthritis, and only two previous reports of hepatotoxicity have been published in the scientific literature. The present report describes three patients who developed acute liver injury following exposure to glucosamine; one patient made a complete recovery on cessation of ingestion, the second developed chronic hepatitis and the third died following progression to fulminant hepatic failure. A diagnosis of glucosamine-induced hepatotoxicity was made based on the temporal relationship between onset of liver injury and glucosamine ingestion, exclusion of all other potential aetiologies and, in the two surviving cases, improvement in condition on withdrawal of the supplement.

BACKGROUND

The use of complementary and alternative medicine in the UK is becoming increasingly popular; surveys estimate 10% to 20% of the population consult a complementary practitioner each year and 33% have used some form of alternative therapy.¹ Herbal medicine is one of the most popular complementary disciplines; over £40 million a year is spent on herbal products in the UK, mainly from self-prescription of products available over the counter.²

Most herbal preparations are classed as dietary supplements and not subject to the regulations and safety standards applied to conventional medicines. Current European Union (EU) regulations require demonstration that herbal medicinal products meet acceptable standards of quality, safety and efficacy before a product licence can be issued.³ The UK has alternative regulatory routes for herbal products and not all products are licensed.

Although there have been clinical trials to study the efficacy of several herbal medicines, most of the evidence is from systematic reviews and meta-analyses.⁴ There have been few prospective randomised control trials to date raising the concern that present data may not highlight potential adverse effects. Information documenting adverse reactions is currently based on case reports and review articles.⁵ As the use of herbal medicines increases, a number of idiosyncratic hepatotoxic reactions to herbal medicines have been reported (table 1). Adverse reactions to conventional medicines are reported to the Medicines Control Agency (MCA) and the Committee on Safety of Medicines (CSM) via the yellow card scheme, the system (used by clinicians and pharmacists) is limited by under-reporting of adverse events. The National Institute of Medical Herbalists have begun a similar yellow card scheme for herbal practitioners and the National Poisons Unit has set up a database to record adverse events. However, without a more systematic reporting scheme, involving medics and herbalists, the incidence of adverse reactions to herbal preparations will remain unknown.

Table 1.

Herbal preparations associated with hepatotoxicity

Herbal preparation	Indications	Hepatic disease	References
Atractylis gummiferia	Antipyretic, emetic, diuretic	Hepatitis, FHF	Georgia6
Black cohosh	Menopausal flushes	Hepatitis, FHF, idiosyncratic AIH	Whiting et al5, Cohen et al7, Lontos et al8
Chaparral	Antioxidant, antiageing, snake bites	Cholestasis, hepatitis, cholangitis, cirrhosis	Sheikh et al9, Batchelor et al10, Gordon et al11
Comfrey	Health tonic, poultice	Veno-occlusive disease, hepatocellular carcinoma	Ridker et al12, Miskelly and Goodyer13
Germander	Antipyretic, weight loss	Hepatitis, cirrhosis	Larrey et al14, Mostefa-Kara et al15
Glucosamine	Joint pain, osteoarthritis	Hepatitis, cholestasis	Ossendza et al16, Fujii et al17
Greater celandine	Dyspepsia, IBS, gallstones	Hepatitis, cholestasis, cirrhosis	Benninger et al18
Jin Bu Huan	Sedative, analgaesic	Cholestasis, hepatitis, fibrosis	Piccioti et al19, Woolf et al20
Kava	Anxiolytic, sedative	Hepatitis, FHF	Stickel et al21
Ma Huang	Weight loss	Hepatitis, idiosyncratic AIH	Nadir et al22
Mistletoe	Sedative, antihypertensive	Hepatitis	Hyde23, Farnsworth and Loub24, Stirpe25
Noni juice	Health tonic	Hepatitis, FHF	Stadlbauer et al26
Pennyroyal oil	Abortion, pesticide	Hepatic necrosis, FHF	Bakerink et al27, Sullivan et al28
Sassafras	Herbal tea	Hepatitis, hepatocarcinogen	Stickel et al29, Larrey30
Senna	Laxative	Hepatitis	Beuers et al31
Skullcap/valerian	Sedative	Hepatitis	Hullar et al32, MacGregor et al33

AIH, autoimmune hepatitis; FHF, fulminant hepatic failure; IBS, irritable bowel syndrome.

We describe three patients with hepatotoxicity associated with the use of glucosamine, a herbal preparation commonly taken to ease joint pain associated with arthritis. All three patients were referred to liver clinic for investigation of deranged liver function tests (LFT), two of the three presented with jaundice. None of the patients had a history of alcohol excess, intravenous drug use, blood transfusion, history of hepatobiliary disease or family history of liver disease. Alternative aetiologies were excluded by ultrasound scanning, iron studies, viral serology (hepatitis A, B, C, Epstein–Barr virus (EBV), cytomegalovirus (CMV)) and immunology screening tests (anti-smooth muscle, anti-nuclear and anti-mitochondrial autoantibodies and IgA, IgG and IgM).

CASE PRESENTATION

Patient 1

Patient 1 was a 64-year-old man who presented with a 3-week history of nausea and vomiting. At 1 week after the onset of illness he developed jaundice with pale stool, dark urine and associated night sweats. He was on no prescription medication but had been taking glucosamine and chondroitin sulfate for 4 weeks. On examination the patient was jaundiced, abdomen was soft with no detectable masses or organomegaly and there were no stigmata of chronic liver disease. Abdominal ultrasound showed normal liver echotexture, normal spleen, gallbladder and bile ducts. Derangement in liver function was predominantly hepatocellular, alanine aminotransferase (ALT) was 1461 U/litre, bilirubin 419 μmol/litre and alkaline phosphatase (AP) 141 U/litre. Ferritin was markedly elevated at 10 000 μg/litre raising the suspicion of underlying haemachromatosis but genotyping was negative. ALT improved on cessation of glucosamine/chondroitin sulfate but bilirubin continued to rise. At 2 weeks after stopping the herbal medication the patient developed acute renal failure (creatinine 294 μmol/litre) and coagulopathy (prothrombin time 32.2 s) and was admitted for urgent treatment including renal replacement therapy. Examination revealed right upper quadrant tenderness and marked icterus, there were no signs of hepatic encephalopathy. Repeat ultrasound showed a coarse liver echotexture but was otherwise unchanged from the previous examination. Liver function tests on the day of admission were: ALT 597 U/litre, bilirubin 695 μmol/litre and AP 135 U/litre. The patient was not encephalopathic on admission but hepatic and renal failure progressed over the following 2 days and he developed grade I/II encephalopathy and a metabolic acidosis. At this stage he met the Kings College criteria for emergency liver transplantation and was transferred to the liver unit in Edinburgh, UK, where he was listed for super-urgent transplantation. Unfortunately his condition continued to decline and was complicated by peritoneal and blood-borne sepsis. Disseminated sepsis precluded transplantation and he died the morning after transfer, postmortem examination of the liver revealed extensive tissue necrosis.

Patient 2

Patient 2, a 57-year-old woman, presented with a 10-day history of jaundice and a 24-h history of an itchy rash. The jaundice was preceded by 2 weeks of anorexia and general malaise. She had been taking capsules containing glucosamine and cod liver oil for 1 month but had changed to a herbal preparation comprising methylsulfonylmethane (MSM) and glucosamine 5 days before presenting to hospital. The patient was not on any prescribed medication and took no other herbal preparations. On examination she was icteric with a macular, excoriated rash affecting the upper and lower limbs. Her abdomen was soft, non-tender with no masses or organomegaly, the patient had no stigmata of chronic liver disease. Liver, spleen and biliary system appeared normal on ultrasound examination. ALT was 1130 U/litre, bilirubin 198 μmol/litre and AP 198 U/litre. Prothrombin time (PT) was slightly prolonged at 13.4 s (control 12). The patient ceased taking glucosamine and was monitored closely over the first few weeks of her recovery. The rash and anorexia resolved within 1 month and the jaundice cleared after 3 months, prothrombin time returned to normal within 2 months. Liver enzymes gradually improved (ALT fell to 548 after 4 weeks, 167 after 8 weeks and 109 after 12 weeks) but failed to return to within normal parameters, AP remained slightly elevated between 120–190 U/litre (normal range 40–130) and ALT remained elevated, ranging from 80–100 U/litre (normal range 13–43). At 6 months after initial presentation LFT derangement persisted; the patient remained clinically well though she reported symptoms of fatigue. MRI scanning revealed gross macronodular cirrhosis, splenomegaly and ascites, oesophageal varices were evident at upper gastrointestinal (GI) endoscopy. Liver biopsy demonstrated chronic portal based hepatitis with moderate interface activity. The inflammatory infiltrate comprised predominantly plasma cells and eosinophils, therefore, it was not possible to distinguish on histopathological grounds between autoimmune hepatitis or drug-induced hepatitis. The patient received a trial of steroid but her condition remained unchanged, this excluded the possibility of seronegative autoimmune hepatitis and added further weight to the diagnosis of glucosamine-induced chronic hepatitis.

Patient 3

Patient 3 was a 55-year-old woman found to have deranged LFT results on routine blood testing. The adverse reaction was, in this case, cholestatic in nature with a peak ALT of 175 U/litre and AP 187 U/litre, bilirubin and PT remained within the normal range. In addition to her prescription medications, bendrofluazide and diclofenac, the patient took five herbal preparations, glucosamine, black cohosh, Kalms, cod liver oil and oil of evening primrose. She had been using these preparations long term (6 months) but stopped taking them approximately 2 months prior to attending clinic. Abdominal examination was unremarkable and ultrasound examination of liver, spleen and biliary tree was normal. Liver function tests were repeated during her clinic visit and showed improvement, ALT had fallen from 175 to 42 U/litre (normal range 13–43) and AP from 187 to 60 U/litre (normal range 40–130).

INVESTIGATIONS

Clinical and laboratory data for patients 1–3 is summarised in table 2.

Table 2.

Summary of patient clinical and laboratory data

Patient (age/sex)	Herbal preparations	Duration intake	Time to onset of illness	Symptoms	Peak	bilirubin (μmol/litre)/ALT (U/litre)	AP (U/litre)	Outcome
64/M	Glucosamine, chondroitin sulfate	4 weeks	1 week	Nausea, vomiting, jaundice, night sweats	725	1461	141	Fulminant hepatic failure, death
57/F	Glucosamine, cod liver oil, MSM	4 weeks	1 week	Anorexia, jaundice, malaise, itchy rash	290	1130	198	Jaundice resolved in 3 months, chronic hepatitis
55/F	Glucosamine, black cohosh, Kalms, cod liver oil, evening primrose oil	6 months	1 month	None	7	175	187	Complete recovery

ALT, alanine aminotransferase; AP, alkaline phosphatase; MSM, methylsulfonylmethane.

DISCUSSION

We present three cases of hepatotoxicity, one resulting in fulminant hepatic failure and death, associated with the use of glucosamine. With the exception of black cohosh, none of the other herbal preparations used by patients 1–3 have recognised hepatocellular toxicity. Glucosamine is the only preparation common to all three cases, with each case demonstrating a temporal relationship between consumption of glucosamine and the development of hepatotoxicity. A review of the current literature identified two previous case reports of glucosamine-associated hepatotoxicity; the first describes a patient presenting with a cholestatic reaction following 19 days of glucosamine sulfate ingestion,¹⁶ the second patient developed acute hepatitis after taking glucosamine sulfate, soybean extract and lutein supplements for 6 months.¹⁷ In both cases liver enzymes normalised on cessation of glucosamine ingestion, the clinical course was suggestive of herbal hepatotoxicity.

The type of hepatotoxic reaction for each patient was classified according to the Council for International Organisations of Medical Sciences definitions for acute liver injury.^34,35 The clinical diagnostic scale (CDS) was used to assess glucosamine causality of the hepatic adverse reaction.³⁶ Patients 1 and 2 presented with jaundice and marked LFT derangement, though both had features of cholestasis, the acute liver injury was hepatocellular in nature.³⁵ Patient 1 had a CDS score of 12 and patient 2 a CDS score of 13; a CDS score of >9 is strongly suggestive of drug-induced liver injury (DILI) providing all other aetiologies have been excluded.³⁶ In case 3 the hepatotoxic adverse reaction was cholestatic, the patient was asymptomatic and LFT derangement was detected on routine blood tests. For this reason it is more difficult to establish the CDS score. The precise temporal relationship between glucosamine intake and onset of LFT derangement, and time from glucosamine withdrawal to normalisation of LFT is unknown. It is possible that the other herbal remedies ingested may have contributed to the hepatic injury, exclusion of this possibility would require re-challenge with glucosamine and each preparation separately. Black cohosh is the only preparation previously associated with hepatotoxicity, the nature of the reaction being hepatocellular rather than cholestatic (5,7,8). A CDS score of 5 in patient 3 does not support a definite causative relationship between herbal ingestion and liver injury though the clinical course is highly suggestive of glucosamine-induced liver injury.

Glucosamine is classed as a dietary supplement rather than a medicinal product and is available in a variety of preparations (eg, glucosamine hydrochloride, glucosamine sulfate and N-acetylglucosamine) in tablet, capsule, powder or liquid form. In commercially available preparations it is commonly combined with chondroitin sulfate, MSM, manganese ascorbate and shark or bovine cartilage.³⁷ It is difficult to determine whether glucosamine itself, other ingredients within the herbal preparation or contaminants from its processing are the causative agent in adverse reactions. Some manufacturers extract glucosamine from the processed exoskeleton of shrimp, lobster and crab shells making it potentially allergenic for consumers with shellfish or seafood allergies.³⁷ Most synthetic glucosamine is produced in China, where some manufacturers use wastewater processing with the possibility of heavy metal or sewage contamination.³⁷

Seven randomised, double-blind, placebo-control trials have been published evaluating the use of glucosamine for pain relief in osteoarthritis.^38–44 Trial duration ranged from 4 weeks to 36 months, subjects ingested glucosamine (sulfate or hydrochloride preparations) at the recommended dose of 1500 mg/day. The majority of the trials studied 50–60 patients in each treatment arm; 1 trial had only 10 subjects in each group and the largest trial recruited 120 volunteers in the placebo and glucosamine groups. Patients in the treatment arm of the trials reported adverse effects ranging from 0 (in a 6–8 week study) to 94% (in a 3-year study).^38,44 The commonest side effects involved the gastrointestinal system and included abdominal pain, abdominal cramps, constipation, diarrhoea, dyspepsia, nausea and vomiting. There was no significant difference in the incidence of adverse effects between subjects taking glucosamine or placebo.^38–44

The incidence of hepatotoxicity from herbal remedies has been estimated at 0.2% to 1%, the true incidence is unknown due to a lack of prospective trial data.⁵ The mechanisms causing hepatotoxicity are poorly understood, drug toxicity to the liver can manifest as any liver disorder (eg, hepatitis, cholangitis, cholestasis) with acute hepatitis occurring in 90% of cases.⁴⁵

Hepatitic drug reactions mimic acute viral hepatitis with clinical presentation ranging from asymptomatic to fulminant hepatic failure. The main biological feature is a marked elevation in aminotransferases (ALT). Liver biopsy specimens from such patients show liver cell necrosis with associated inflammatory infiltration. Features particular to drug toxicity include the presence of eosinophils in the infiltrate and centrilobular predominance of the inflammatory lesions.^45,46 The precise mechanism of injury is unknown but the hepatitic reaction may be triggered by the production of toxic metabolites when the herbal preparation is metabolised by the liver. The metabolites may be directly toxic to the hepatocytes (causing glutathione depletion similar to acetaminophen toxicity) or they may trigger an adverse immune response directed against the liver (observed in halothane and nitrofurantoin hepatotoxicity).^45,47

Acute cholestatic hepatitis is characterised by an elevation of serum alkaline phosphatase (AP), cholestatic reactions are divided into two subtypes: pure cholestasis and cholestatic hepatitis.^35,45 Pure cholestasis is characterised by jaundice, pruritis, pale stool and dark urine. Liver function tests show an obstructive pattern with elevated bilirubin, AP and γ-glutamyltranspeptidase (GGT), transmaninases are normal or slightly elevated. Liver histology shows dilated biliary cannaliculi containing biliary pigments and bilirubin deposited within hepatocytes, these changes predominate in the centrilobular area.⁴⁵ Pure cholestasis is most commonly associated with use of hormone derivatives (eg, oestrogens, androgens, tamoxifen), complete recovery follows discontinuation of the causative agent.⁴⁵

Acute cholestatic hepatitis may mimic acute biliary obstruction with features of cholestasis associated with abdominal pain and fever; liver biopsy shows cholestasis and inflammatory infiltrate within the portal tracts.^35,45 In rare cases, particularly those with features of cholangitis, chronic cholestasis mimicking primary biliary cirrhosis may develop. Drugs associated with acute cholestatic hepatitis include phenothiazines, macrolides, augmentin, tricyclic antidepressants and carbamazepine.⁴⁵

The majority of patients with DILI make a complete recovery within 1–3 months following discontinuation of treatment; herbal hepatotoxicity reactions follow a similar clinical course. In more rare cases however, the acute liver injury progresses rapidly, resulting in fulminant or subfulminant hepatic failure with 90% mortality. The only available treatment is emergency liver transplantation. The risk of developing rapidly progressive liver failure is significantly higher when drug administration is continued despite the occurrence of jaundice.⁴⁵

Drug reactions may also follow a more insidious course, as demonstrated by our second patient, with the development of chronic hepatitis eventually leading to cirrhosis. Until recently, it was assumed that if patients survived a severe DILI, complete resolution of hepatic injury would occur. Studies following DILI patients long term have challenged this precept, demonstrating that, although rare, severe DILI associated with jaundice may contribute to the development of chronic liver disease and cirrhosis.^48–50

These three cases provide evidence that glucosamine may be associated with the development of acute hepatocellular hepatitis and/or cholestasis. Further investigation is required to determine the causative agent within the glucosamine preparations that resulted in hepatotoxicity. The cases highlight the need for greater awareness, in the complementary and conventional medical communities, of the potential toxicity of glucosamine and other herbal preparations. It is essential to seek a history of exposure to herbal preparations in any patient presenting with unexplained acute liver injury, since early cessation of treatment may prevent progression to fulminant hepatic failure and death.

LEARNING POINTS

• It is important to consider drug reactions as a potential aetiology in any patient presenting with unexplained acute liver injury.

• A clear medication history must be obtained, including exposure to herbal preparations.

• Severe drug-induced liver injury (DILI) may contribute to the development of chronic liver disease and cirrhosis.