19th May 2025

Drug-induced Liver Injury

Introduction

The term drug-induced liver injury (DILI) refers to all forms of hepatotoxicity that result from the administration of a medicinal agent. It encompasses a broad range of presentations of hepatic disease and may manifest as either acute or chronic disease. DILI is the commonest cause of acute liver failure (ALF) in the UK. 

A diagnosis of DILI is one of exclusion, once liver injury has been identified clinically and/or biochemically. Nearly all cases of DILI are idiosyncratic, with the notable exception of paracetamol overdose. Idiosyncratic drug reactions affecting the liver are generally understood to arise from immune-mediated processes. 

The treatment of DILI involves cessation of the causative therapeutic agent. ALF may benefit from treatment with N-acetylcysteine early within the disease course and liver transplantation should be considered early.

DILI classification

DILI can be classified into one of several groups, on the basis of either its clinical manifestations, or the underlying pathophysiological mechanism of hepatotoxicity.

When the classification of DILI is made on the basis of its clinical presentation, three distinct classes emerge: hepatocellular injury, cholestatic injury, or mixed disease. The features of each of these are outlined in the table below. When classifying DILI based on the clinical manifestations, the ‘R’ value is used; this is calculated by the following equation: R = (ALT/ULN)/(ALP/ULN) where ALT = alanine transferase, ALP = alkaline phosphatase, and ULN = upper limit of normal.

Table 5: The Features of DILI

Examples of causes of hepatocellular DILI: diclofenac, fenofibrate, isoniazid, rifampicin, pyrazinamide, disulfiram, minocycline, nitrofurantoin

Examples of causes of cholestatic DILI: co-amoxiclav, cephalosporins, flucloxacillin, terbinafine, erythromycin, chlorpromazine

Examples of causes of mixed DILI: carbamazepine, sulphonamides, lamotrigine, phenytoin

When the pathophysiological mechanism is used to distinguish DILI types, it may be separated into either predictable or unpredictable (i.e. idiosyncratic) types. The majority of cases of DILI are idiosyncratic, with the exception of paracetamol overdose. The differences between predictable and idiosyncratic DILI are summarised in the table.

Table 6: Predictable and Idiosyncratic DILI

Idiosyncratic DILI can be further subdivided on the basis of the underlying pathophysiology, i.e. immune-mediated vs non-immune mediated. The features of each are outlined in the table below. The lion’s share of instances of idiosyncratic DILI are immune-mediated.

Table 7: Features of Idiosyncratic DILI

Some drugs which result in DILI that does not fit into the above classifications include:

• Amiodarone, methotrexate, valproate, tamoxifen – can induce hepatic steatosis
• Androgens, contraceptive steroids, anti-neoplastic agents – peliosis hepatis (multiple foci of blood-filled cavitations within the hepatic parenchyma). 
• Anabolic steroids – hepatocellular adenomas
• Thiopurine therapy (including azathioprine, mercaptopurine) – nodular regenerating hyperplasia 

Epidemiology

It is likely that the incidence of DILI is underestimated due to underreporting and variable diagnostic criteria, however it is thought to be in the region of 14-19 cases per 100 000 people per annum. DILI is the most common reason for withdrawal of medications from the marketplace.

Aetiology

Risk factors

Factors which increase the risk of developing idiosyncratic DILI include: 

• Female gender – in addition, women are more likely to be of younger age and have hepatocellular DILI, whereas affected males are more likely to be older and demonstrate cholestatic DILI. Women are also more likely to demonstrate severe hepatotoxicity.
• Increased age
• Drug factors – increased duration of use, higher dose
• Genetic factors – certain HLA genotypes appear to confer greater risk of DILI caused by certain drugs such as flucloxacillin
• Race and ethnicity – the commonest agent causing DILI in African Caribbean population groups is co-trimoxazole, while for Caucasian population groups it is co-amoxiclav
• Pre-existing liver disease – however the evidence is as yet unclear as to whether this is a true risk factor
• Methotrexate induced liver injury is 3-4x more likely in patients with a history of alcohol excess, chronic viral hepatitis or those with risk factors predisposing to non-alcoholic fatty liver disease (diabetes, obesity).

Causative Agents 

Paracetamol is the leading cause of drug-induced acute liver failure (ALF) in Western civilisation. Antibiotics are the commonest cause of idiosyncratic DILI, with co-amoxiclav as the most commonly implicated individual drug. Complementary and herbal medicines and dietary supplements, antiepileptics, cardiovascular drugs, anti-neoplastic agents and analgesics are among the other most commonly involved drug classes. On a global scale, over half of the total cases of DILI are attributable to antibiotics and antiepileptics alone. 

Drugs which are associated with relatively high incidences of severe liver injury include isoniazid and methotrexate. 

Drugs which have been seen to cause idiosyncratic DILI include isoniazid, methyldopa, phenytoin, and diclofenac.

Drugs which have been seen to cause asymptomatic progression to cirrhosis include methotrexate, methyldopa and amiodarone.

Pathophysiological mechanisms

The pathogenesis of idiosyncratic DILI is unclear. Current schools of thought are that drug-related factors (dosage, metabolic pathways, duration of treatment) and patient factors (age, gender, genetic factors) and their interactions are together responsible for the development of hepatotoxicity.

Clinical features

There are no diagnostic tests or biomarkers for DILI. It therefore relies upon the identification of exposure to a drug which precedes the onset of liver injury, exclusion of alternative causes of liver pathology, and an improvement in markers of liver injury on cessation of the drug. Drug rechallenge is generally not advised due to the risk of associated morbidity and mortality. Recreational drugs and herbal and complementary medicines must be considered.

Clinical manifestations

Nearly three quarters of cases of DILI present with jaundice. The clinical spectrum of DILI includes:

• Asymptomatic, mild liver test abnormalities
• Significant elevation in liver enzymes with clinical features of acute hepatitis (malaise, abdominal pain, nausea and vomiting, jaundice, dark urine, pale stools)
• Cholestasis, with clinical features seen in acute hepatitis plus pruritus
• ALF; evidence of coagulopathy and hepatic encephalopathy may be present.
• Chronic liver disease with progression to cirrhosis 
• Chronic liver injury resembling autoimmune hepatobiliary disease; elevated serum markers of autoimmunity are seen e.g. raised anti-nuclear antibody.

A significant proportion of patients with DILI are asymptomatic. Some groups of patients develop signs of a hypersensitivity reaction, with fever and rash, and in some cases Stevens-Johnson syndrome, the mortality of which is higher than in patients without these features.

A significant proportion of patients with DILI are asymptomatic. Some groups of patients develop signs of a hypersensitivity reaction, with fever and rash, and in some cases Stevens-Johnson syndrome, the mortality of which is higher than in patients without these features. 

Histopathological findings

Liver biopsy may be useful to exclude alternative aetiologies for abnormal liver function. In the case of DILI, non-specific findings such as portal neutrophils and intra-hepatocellular cholestasis are seen. 

If other tests for alternative causes of liver injury are negative and the patient has been exposed to a drug known to be associated with liver injury in other patients, liver biopsy may not be necessary. 

Serological testing

As mentioned, there are no serological tests or biomarkers specific to DILI. 

Patients with DILI who meet at least one of the following criteria are deemed to have ‘clinically significant’ liver injury: 

• AST or ALT >5x ULN or ALP >2x ULN, on at least two occasions
• Bilirubin > 2.g mg/dL in association with raised AST, ALT or ALP 
• INR > 1.5 with raised AST, ALT or ALP

Modified Hy’s law states that patients with DILI with serum bilirubin 2 or more times greater than the ULN is associated with increased mortality, and drugs demonstrating this phenomenon within clinical trials are at high risk of causing DILI when introduced to a wider population. 

In cases of idiosyncratic DILI, AST in the realm of 18 times ULN and AST to ALT ratio greater than 1.5 indicates significant risk of progression to ALF.

Radiographic findings

Imaging studies are useful in DILI to exclude venous outflow obstruction as an alternative cause of liver injury.  If there is evidence of cholestasis, imaging can also help to exclude obstruction within the biliary tree.

Management

Immediate cessation of the offending therapeutic agent once DILI is suspected is vital. 

In some instances, there is a case for the use of corticosteroid therapy, particularly when an immune-mediated aetiology is suspected or in cases of DILI which demonstrate biopsy features of autoimmune hepatitis. 

Some specific treatment agents for DILI exist depending on the offending drug; among these are N-acetylcysteine for paracetamol overdose and L-carnitine for valproic acid overdose. N-acetylcysteine occasionally also has a role in the treatment of non-paracetamol induced ALF, particularly when identified early in the disease course. 

There is no evidence at present in support of the use of ursodeoxycholic acid in cholestatic DILI with pruritus but its use is common in practice. 

Acute Liver Failure

Approximately 4% of case idiosyncratic DILI progress to acute liver failure (ALF). Factors increasing the likelihood of developing ALF include hepatocellular DILI and female gender. 

The most common drugs responsible for DILI which progresses to ALF include anti-tuberculosis medications, herbal/complementary medicines and dietary supplements, co-amoxiclav, nitrofurantoin, sulfa-based drugs, phenytoin and valproate. 

In cases of drug-induced ALF, early discussion regarding liver transplantation is crucial as survival without transplant is poor (27% at 3 weeks). Evidence suggests steroids confer little survival benefit in ALF.

Prognosis

Prognosis is variable dependent on the classification of DILI. Studies have demonstrated that: 

• The majority of cases of DILI resolve after cessation of the causative agent / reduction in dosage.
• 17% progress to chronic DILI (abnormal liver function tests for greater than 6 months)
• 10% require liver transplantation and/or lead to mortality as a result of liver failure.  
• An underlying mechanism which is hepatocellular rather than cholestatic confers a greater likelihood of requiring liver transplant and shows higher mortality. 
• Conversely, cholestatic DILI is more likely than hepatocellular DILI to progress to chronic liver disease.