Drug-induced liver injury and paracetamol overdose

• Drug-induced liver injury is not one disease. It is any clinically significant liver injury caused by a drug, herbal or dietary supplement, and it can look like almost any liver disease. ^[1]
• Most cases fall into three types: direct, predictable, dose-related injury (paracetamol is the classic example); idiosyncratic injury that is unpredictable and usually not dose-related, but still needs a dose threshold; and indirect injury, driven by the drug's effect on immunity (checkpoint inhibitors, infliximab). ^[3]
• There is no specific blood test for it. The diagnosis rests on a careful medication history, a pattern of liver enzymes, and excluding everything else. ^[2][3]
• Paracetamol (acetaminophen) in overdose causes a direct, zone III hepatocellular necrosis. Acetylcysteine is almost completely protective when given within 8 hours, and a single Rumack-Matthew treatment line decides who needs it. ^[4][5][6]
• The rule for every DILI is the same: stop the offending agent. Most patients recover, supportively, without any antidote. ^[1]

By the end of this page you should be able to:

1. Define DILI and separate direct, idiosyncratic and indirect injury, giving the classic example of each. ^[3]
2. Recognise the patterns of enzyme derangement (hepatocellular, cholestatic, mixed) and calculate the R value. ^[2][3]
3. Assess causality with the CIOMS (RUCAM) scale and know its limitation. ^[2]
4. List the common culprits worldwide and explain why the list differs by region. ^[2][3]
5. Plot an acute paracetamol level on the Rumack-Matthew nomogram and state the UK treatment line. ^[5]
6. Prescribe intravenous acetylcysteine, including loading and maintenance doses, infusion times and stopping rules. ^[6][7]
7. Apply the King's College Criteria and arrange referral when liver failure is developing. ^[4]

DILI was once treated as a single entity, but current guidance separates three mechanisms. The distinction matters, because it predicts dose relationship, latency, and whether an antidote exists. ^[2][3]

The three classes are useful rather than absolute. There is considerable overlap between idiosyncratic and indirect injury, and the same drug (for example an antibiotic after a long course) can produce more than one clinicopathological picture. ^[3]

• Direct hepatotoxins harm the liver in almost everyone once a threshold dose is exceeded, and they can be reproduced in animals. Paracetamol is the direct hepatotoxin every medical student meets first. ^[2]
• Idiosyncratic hepatotoxins injure only susceptible people, at ordinary doses, after variable latency, and they cannot be reproduced reliably in animals. Most approved drugs that cause DILI do so in roughly 1 in 1000 to 1 in a million exposed people. ^[2]
• Indirect hepatotoxins injure by the drug's biological action rather than by a toxic metabolite (for example immune-mediated hepatitis from checkpoint inhibitors). ^[3]

Idiosyncratic DILI is rare but not negligible. Its estimated annual incidence in the general population is about 14 to 19 events per 100,000 inhabitants, roughly 60,000 cases per year in the United States. ^[2] The incidence rises to 33 to 40 per 100,000 in studies that use electronic medical records, and hospital-based cohorts report even higher rates. ^[2]

Paracetamol (acetaminophen) is the dominant cause of acute liver failure in the English-speaking world: it is responsible for about 65.4% of cases of acute liver failure in the United Kingdom and 45.7% in North America. ^[4] In the United States acetaminophen toxicity has replaced viral hepatitis as the most common cause of acute liver failure. ^[4] The Annual Report of the National Poison Data System recorded 64,263 single-substance exposures to acetaminophen in the United States, of which 25,104 were in children aged 5 years or younger and 21,554 were in adults aged 20 years or older. ^[4]

Women experience more frequent and more severe drug-induced hepatotoxicity. A French population-based study found the standardised incidence more than twice as high in women than in men after the age of 50, with no sex difference below that age. ^[2]

Idiosyncratic DILI is usually independent of the total dose, but most implicated drugs are given at a daily dose above 50 to 100 mg, and over 80% of drug-related liver transplants in the United States involved drugs with daily doses above 50 mg. ^[2] Drugs with high lipophilicity and extensive hepatic metabolism (over 50%) carry a higher hepatotoxic potential, especially at daily doses above 100 mg, and drugs that form reactive metabolites, exert mitochondrial toxicity or inhibit bile acid transporters are more likely to injure the liver. ^[2] Concurrent administration of several hepatotoxic drugs increases risk, and dose escalation may do the same (for example with azathioprine). ^[2]

• Older age raises hepatotoxic risk for some agents (amoxicillin-clavulanate, isoniazid), whereas younger people are more prone to DILI from anticonvulsants and minocycline. ^[2]
• Women are over-represented in diclofenac, macrolide, flucloxacillin, halothane, ibuprofen, interferon beta-1a and nitrofurantoin hepatotoxicity; men in azathioprine, anabolic steroid and amoxicillin-clavulanate hepatotoxicity. ^{[2] [3]}
• Obesity and diabetes are associated with steatosis and fibrosis in drug-treated patients (for example tamoxifen, methotrexate). ^[2]
• Most data do not link alcohol consumption to clinical outcome in DILI cohorts, although alcohol does induce CYP enzymes and a poor nutritional state is a clear risk factor for paracetamol injury. ^[2][4]
• Chronic liver disease such as non-alcoholic fatty liver disease increases the risk of poor outcome once a DILI episode occurs, and subjects with pre-existing liver disease are at increased risk of injury with selected drugs (methotrexate, anti-tuberculosis therapy). ^[2]
• Paracetamol-specific risks: chronic alcohol use, fasting and reduced oral intake, older age, underlying liver or kidney disease, CYP inducers (ethanol, tobacco smoking, isoniazid, rifampin, phenytoin, phenobarbital, barbiturates, carbamazepine) and concurrent use of acetaminophen with non-steroidal anti-inflammatory drugs. ^[4]

A polymorphism in PTPN22 is a genetic risk factor across multiple drugs and major ethnic groups. Several human leucocyte antigen alleles have been associated with increased susceptibility to individual agents (for example HLA-B*57:01 with flucloxacillin), but pretreatment HLA testing is of limited clinical usefulness because of the low incidence of DILI and the high negative predictive value of the alleles. ^[2]

Although hundreds of drugs can injure the liver, a small group of classes dominates worldwide. ^[2][3] Antibiotics are still the commonest class in most registries, with amoxicillin-clavulanate the most frequently implicated individual agent in Western countries. ^[2][3] In Asian countries anti-tuberculosis drugs and traditional Chinese medicines or herbal products dominate. ^[2][3] Herbal and dietary supplements accounted for about 16% of cases in the Icelandic and the DILIN studies. ^[1][3]

In a recent 5-year prospective Spanish study, anticancer drugs (mainly checkpoint inhibitors) became the most frequent cause, followed by antibiotics, analgesics and recreational drugs. ^[3] A prospective European DILI study found the following most common causes: amoxicillin-clavulanate 12%, flucloxacillin 11%, atorvastatin 8%, nivolumab and ipilimumab 8%, infliximab 5%, nitrofurantoin 4.5%, disulfiram 2% and ibuprofen 2%. ^[3]

The epidemiology in the table also shows striking geographic differences for the same disease. Herbal and dietary supplements account for 27 to 62% of DILI in China, Korea and Singapore, but only a minority in Japan, the United States and Spain, although the proportion is rising. ^[2] In India, anti-tuberculosis drugs are the leading culprits. ^[2]

At therapeutic doses paracetamol is a familiar and generally safe analgesic and antipyretic. The danger comes from a metabolite, not the parent drug. ^[4]

Paracetamol is rapidly absorbed from the stomach and small intestine, with serum concentrations peaking at 1 to 2 hours and, in an overdose of an immediate-release preparation, peak plasma levels within 4 hours. ^[4] The elimination half-life is about 2 hours (range 0.9 to 3.25 hours). ^[4] Ordinarily the liver conjugates most paracetamol to non-toxic water-soluble compounds, and a smaller fraction is oxidised by cytochrome P450 (CYP2E1, 1A2, 2A6 and 3A4) to a reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI). NAPQI is normally inactivated by conjugation with glutathione and excreted by the kidneys. ^[4]

In an overdose the conjugation pathway saturates and more drug is pushed down the oxidative pathway. ^[4] NAPQI is produced in amounts that exhaust the hepatocellular glutathione store (depleted by roughly 70%), after which NAPQI binds covalently to hepatocellular cysteinyl sulfhydryl groups, forming adducts that drive oxidative damage, mitochondrial dysfunction and an inflammatory cascade. ^[4] Necrosis is not random: it is centrilobular, in zone III, because zone III hepatocytes carry the highest concentration of CYP enzymes and therefore generate the most NAPQI. ^[4] Acute renal injury can follow by the same mechanism in the kidney (acute tubular necrosis). ^[4]

Hepatocellular necrosis can develop after as little as 7.5 g of paracetamol, and single supratherapeutic exposures of 150 mg per kilogram or more have been associated with hepatotoxicity. ^[4][6]

Acetylcysteine works because it is a precursor of glutathione and an alternative sulfhydryl donor, sustaining the conjugation of NAPQI. It also enhances sulfate conjugation of unmetabolised paracetamol and has anti-inflammatory and antioxidant effects with improved microcirculatory blood flow. ^[4][6]

Idiosyncratic DILI arises in susceptible individuals through complex host-drug-environment interactions. Aberrant adaptive immunity is implicated in most cases, which is why immune-modulating drugs (checkpoint inhibitors, anti-TNF antibodies) can produce a distinct "indirect" phenotype. ^[2][3] The enzymatic derangement that defines the bedside pattern (hepatocellular, cholestatic or mixed) reflects whether the dominant injury is to the hepatocyte (transaminases) or to the biliary epithelium (alkaline phosphatase). ^{[2][3] [1]}

DILI can be silent or dramatic. It may surface as asymptomatic enzyme abnormalities on routine bloods, or as symptomatic hepatitis with malaise, anorexia, nausea, jaundice and dark urine. Cholestatic DILI presents with jaundice, pale stools and pruritus. Idiosyncratic injury often declares itself with a systemic hypersensitivity picture (fever, rash, eosinophilia) several weeks after starting the drug, although latency can range from days to years. ^[1][3] Direct hepatotoxins classically present within days. Immune-mediated (indirect) injury has a delayed onset (weeks to months) and may smoulder on even after the drug is stopped. ^[3]

Drug-induced autoimmune hepatitis can look identical to sporadic autoimmune hepatitis, with positive autoantibodies and a hepatitis on the biopsy; the clinical distinction rests on the drug history and on whether the disease persists after withdrawal. ^[2] Immune checkpoint inhibitors can cause a panlobular or centrilobular hepatitis and, occasionally, secondary sclerosing cholangitis. ^[2][3]

In adults the minimum toxic dose after a single ingestion is roughly 7.5 to 10 g. ^[4] In children the minimum toxic dose after a single acute ingestion is 150 mg per kg, although healthy children aged 1 to 6 years are less susceptible and medical toxicologists often use 200 mg per kg in that group. ^[4] Young febrile children with poor oral intake, given repeated analgesic doses, are at particular risk of a repeated supratherapeutic (chronic) overdose. ^[4]

Patients are often well on the first day after a single large ingestion, because enzyme leakage lags behind the biochemical injury. Hepatotoxicity becomes visible over the following days, and acute kidney injury may appear in parallel. ^[4] For this reason the absence of symptoms early on is misleading and should not reassure. ^[4]

DILI is a diagnosis of exclusion. The biochemical pattern (R value, see below) helps organise the possibilities. ^[2]

• Liver enzymes and bilirubin: alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP) and gamma-glutamyl transferase (GGT), with total and direct bilirubin. ALT is more specific for hepatocyte injury; ALP with an elevated GGT confirms a cholestatic (hepatic) source. ^[1][2]
• Severity and synthetic function: albumin, international normalised ratio (INR) and prothrombin time. A rising INR or ratio suggests impending liver failure. ^[1][2]
• Kidney function: creatinine and urea, since DILI (particularly paracetamol) injures the kidney too. ^[1][4]
• Viral and autoimmune screening: as listed in the differential table. ^[2]
• Imaging: abdominal ultrasound in all patients with suspected DILI, to exclude biliary, parenchymal and vascular disease; CT or MRI when indicated. ^[1][2]

• Plasma or serum paracetamol concentration drawn at least 4 hours after ingestion (earlier levels are misleading). ^[6]
• ALT, AST, bilirubin, INR, creatinine, urea, blood glucose and electrolytes. ^[6]
• Extended-release formulations and co-ingestions that delay gastric emptying may push the serum peak beyond 4 hours. ^[4]

A biopsy is not required to diagnose DILI. It becomes useful when the diagnosis is uncertain, when the illness is severe or protracted, when there is a possibility of autoimmune hepatitis, or when the liver tests fail to improve after the suspect drug is withdrawn. Histology can also help with prognosis (eosinophils and granulomas suggest a better outcome; necrosis and fibrosis a worse one). ^[1][2]

A commonly used working definition is any one of the following: ^[2]

1. Serum AST or ALT above 5 times the upper limit of normal, or ALP above 2 times the upper limit of normal (or above the pre-treatment baseline, if the baseline is abnormal), on two separate occasions at least 24 hours apart. ^[2]
2. Total serum bilirubin above 2.5 mg/dL together with an elevated AST, ALT or ALP. ^[2]
3. INR above 1.5 together with an elevated AST, ALT or ALP. ^[2]

Mild aminotransferase abnormalities occur in up to 20% of the general population, so the definitions deliberately set a threshold. ^[2]

At presentation divide the ALT by its upper limit of normal, divide the ALP by its upper limit of normal, and take the ratio. ^[2][3]

The pattern guides the search for competing causes and the expected prognosis: cholestatic DILI tends to last longer, and hepatocellular DILI with jaundice carries the worse outlook. ^[2][3] A cut-off of ALT under 5 times the upper limit of normal with an R value of 5 or more was the EASL hepatocellular criterion, while ALP under 2 times the upper limit of normal with an R value of 2 or less defined cholestatic injury. ^[1]

DILI causality is organised, not intuitive. Several instruments exist: RUCAM (also called the CIOMS scale), the Maria-Victorino clinical diagnostic scale, the Japanese DDW-J score and the newer Revised Electronic Causality Assessment Method (RECAM). ^[2] RUCAM is the most widely used. It awards points for latency, dechallenge, competing causes, rechallenge, the drug's known track record, and risk factors, on a range of minus 9 to plus 14. ^[2] The updated 2016 version is the one in current use. ^[2]

The resulting categories are: ^[1][2]

• Definite (highly probable): 9 or more.
• Probable: 6 to 8. ^[2]
• Possible: 3 to 5. ^[1][2]
• Unlikely: 1 to 2. ^[2]
• Excluded: 0 or less.

A less structured alternative is expert opinion (the DILIN semiquantitative scale), which assigns probability bands from less than 25% to above 95%. ^[2]

Two cautions apply. First, no causality score is a gold standard, and scoring the same case with two tools can give different answers. Second, pharmacovigilance signals and causality scores describe populations and individual cases, not proof of mechanism. The strongest individual proof of causation is a positive rechallenge: hepatocellular injury with an ALT above 3 times the upper limit of normal on re-exposure to the drug. ^[1] This is also dangerous, since re-exposure can produce a faster, more severe injury with fulminant failure. ^[1] Rechallenge is only considered for essential, irreplaceable drugs (for example anti-tuberculous or chemotherapy regimens), under specialist supervision and meticulous monitoring. ^[1]

The single most important bedside rule is older than any score. If drug-induced hepatocellular injury is accompanied by jaundice (hyperbilirubinaemia) the prognosis worsens sharply and the patient is at high risk of death or transplantation. ^[2] This is the reason hepatic necrosis with jaundice is reported urgently. In clinical trials the analogous observation is an elevated aminotransferase with a concurrent bilirubin rise in the absence of a marked alkaline phosphatase rise. Do not let a causality score postpone recognition of a patient who is visibly deteriorating. ^[2]

DILI is usually acute and self-limiting. Liver damage persisting beyond 6 months is labelled chronic DILI by the DILIN (the Spanish group uses 12 months as the better cut-off). A liver biopsy, when performed, may show fibrosis or even cirrhosis. ^[2]

Withdrawal of the offending agent is the mainstay of treatment, and in the majority of cases spontaneous recovery follows without any specific therapy. ^[1] This dechallenge response is itself a criterion in causality assessment. ^[1] Do not wait for confirmatory serology before stopping a clearly implicated drug in a deteriorating patient. ^[1][2]

• N-acetylcysteine is not routinely indicated for DILI except in paracetamol toxicity and in acute liver failure. ^[1]
• Corticosteroids do not have a routine role. They are reserved for presumed autoimmune hepatitis, or for immune-mediated hepatotoxicity such as checkpoint inhibitor injury, and for the allergic features of a systemic hypersensitivity reaction. ^[1][2]
• Ursodeoxycholic acid has no established role in cholestatic DILI. ^[1]
• Document the dechallenge (enzyme trends after stopping) and avoid reintroducing the drug. ^[2]

Immune-related hepatitis is increasingly common. Immunosuppression with corticosteroids is considered for persistent severe hepatotoxicity alongside stopping the checkpoint inhibitor. ^[1]

Apply risk thresholds and the nomogram, and treat early. ^[4][5][6] The following approach matches the UK MHRA single-treatment-line nomogram and the US acetylcysteine label.

1. Take a careful history of amount and timing. A reported quantity is often inaccurate and is not a reliable guide to therapy - clinical context matters. ^[6]
2. Note risk contexts: healthy children aged 1 to 6 years are less susceptible, and chronic (repeated supratherapeutic) ingestions are managed differently from a single acute overdose. ^[4][6]
3. If the patient presents less than 8 hours after a single ingestion and the time is reliable, and an acetaminophen concentration is available, use the revised Rumack-Matthew nomogram to decide. ^{[6] [7]}
4. The UK (MHRA) line joins 100 mg per litre at 4 hours and 15 mg per litre at 15 hours after ingestion, regardless of risk factors. ^[5]
5. If the level is at or above the line, treat with intravenous acetylcysteine. If it is below the line but the ingestion time or sample was doubtful or under 4 hours, still give the loading dose and reassess. ^[6]
6. If the level is below the line, the time of ingestion is known and the sample was taken more than 4 hours after ingestion, acetylcysteine is not needed. ^[6]
7. If the ingestion was staggered over more than one hour or the timing is unknown, give acetylcysteine without waiting for a level; the nomogram does not apply. ^[5][6]
8. If the acetaminophen concentration cannot be obtained within 8 hours, or there is clinical evidence of toxicity, give the loading dose immediately and continue for three doses over 21 hours. ^[6]
9. The critical window is 0 to 8 hours: acetylcysteine is virtually 100% effective in preventing liver injury within that window; after that it falls substantially. Still, if the patient presents late (15 to 24 hours or beyond) it should not be withheld, since the reported history may be wrong and treatment does not worsen outcomes. ^{[5][6] [7]}
10. Acetylcysteine regimen: 300 mg per kg intravenously in three sequential infusions, and a total of 21 hours, given by the weight-based tables in the label: ^[6]

• bag 1, 150 mg per kg over 1 hour (loading dose); ^[6]
• bag 2, 50 mg per kg over 4 hours;
• bag 3, 100 mg per kg over 16 hours. Dilute in sterile water for injection, 0.45% saline or 5% dextrose, and infuse the initial loading dose over 60 minutes to limit early reactions. ^{[6] [5]}

11. Effervescent or oral acetylcysteine formulations are used in some local protocols; intravenous preparation is the one specified above and in the label. ^[6]
12. A regional 12-hour protocol (the Scottish SNAP regimen) administers the same total dose over 12 hours in a two-infusion schedule. It is widely used in UK practice though not yet MHRA-approved; arrange dosing through your local or poisons centre. ^[7]
13. After a larger overdose, repeated supratherapeutic ingestion, co-ingested drugs or pre-existing liver disease, absorption or half-life may be prolonged; recheck acetaminophen levels, ALT, AST and INR after the last maintenance dose and continue past 21 hours if the drug is still detectable or enzymes are rising. ^[6]
14. Do not withhold acetylcysteine because of allergy: hypersensitivity is no longer a contraindication, and reactions are usually anaphylactoid. ^[5]

The three-bag course delivers 300 mg per kilogram over 21 hours. ^[6] Weight-based dose tables in the label cover stepwise weight bands, starting at 5 kg; patients under 5 kg have not been studied. ^[6] For repeated supratherapeutic ingestion, the nomogram is not used; check a paracetamol level and liver, renal and electrolyte indices, and involve a poisons centre for dosing. ^{[6] [7]}

Levels may peak after 4 hours with extended-release preparations, so a second level at 8 to 10 hours is needed before declaring a concentration below the treatment line. ^[6] Presenting under 8 hours with normal enzymes and level under the line safely excludes toxicity. ^[6]

Fomepizole and calmangafodipir have been studied as adjuncts but remain investigational; the standard antidote is acetylcysteine. ^[4]

The King's College Criteria for poor prognosis after acute paracetamol overdose are: ^[4]

• arterial pH below 7.30 after fluid resuscitation; or
• prothrombin time above 100 seconds (or INR above 6.5), creatinine above 3.4 mg/dL and grade III or IV encephalopathy.

A creatinine above 3.4 mg/dL alone, or a PT above 100 seconds or INR above 6.5 alone with encephalopathy, should prompt urgent transplant consultation. ^[4] In one retrospective series, hypoglycaemia, coagulopathy and lactic acidosis predicted poor outcome better than the King's College Criteria, so do not rely on a single score. ^[4]

• Acute liver failure. The feared outcome of severe DILI and particularly of paracetamol hepatotoxicity, manifesting as coagulopathy (INR above 1.5) and encephalopathy. ^[2][4]
• Clinically significant DILI as defined above, including hospitalisation and death. ^[2]
• Chronic DILI: (drug-induced chronic hepatitis, fibrosis or cirrhosis), especially with agents such as methotrexate, amiodarone, nitrofurantoin. ^[2]
• Cholestatic complications: vanishing bile duct syndrome, prolonged cholestasis. ^[2]
• Immune-related hepatitis from checkpoint inhibitors, with a panlobular pattern. ^[2]
• Massive necrosis with hypoglycaemia and lactic acidosis as complications of ALF. ^[4]
• Rechallenge failure: faster, more severe recurrence. ^[1]

In case series, under 4% of patients with severe acetaminophen hepatotoxicity develop liver failure and less than half of those die or require transplantation; with timely treatment, the overall mortality is less than 2%. ^[4]

Most patients with DILI recover completely once the drug is withdrawn. ^[1] Chronic injury is the exception (beyond 6 months by the DILIN definition, 12 months by the Spanish group). ^[2] Paracetamol, treated in time, is usually reversible, although an early deterioration screen is needed: the King's or other prognostic criteria, phosphate dynamics and, in some studies, serum lactate. ^[4] Liver repair markers include histology: eosinophils and granulomas are associated with a better outcome, necrosis and ductular reaction with a worse one. ^[2]

There is no routine screening test for DILI before damage occurs, and HLA testing is not recommended for prevention except in organised programmes. ^[2] Prudent prescribing does most of the prevention:

• Know the hepatotoxic potential of the drugs you use; consult LiverTox for drug-specific details. ^[2]
• Keep daily doses of idiosyncratic hepatotoxins out of the danger zone where possible (under 50 mg per day where structurally feasible), avoid combining several hepatotoxic drugs and be cautious with lipophilic agents with heavy hepatic metabolism. ^[2]
• Avoid unnecessary repeated Courses of hepatotoxic drugs (for example frequent antibiotics and long-term methotrexate with monitoring).
• In countries where herbal and dietary supplements are common, ask about them specifically (they accounted for about 16% of DILI in the Icelandic and DILIN studies, and a much larger fraction in parts of Asia). ^[1][3]
• Paracetamol education: adults should not exceed the labelled daily dose (historically 4 g per day in the United States, with advisory committees having recommended 3.25 g, which has not become a mandate). ^[4] Keep children under 60 kg to 75 mg per kilogram per day or 4 g per day, whichever is less. ^[4] Highlight that many cough, cold and combination products contain paracetamol. ^[4]

• Children: paracetamol toxicity in children aged 1 to 6 years is less likely after a single large ingestion (threshold often 200 mg per kg). Repeated supratherapeutic dosing in febrile children with poor intake is a distinct and preventable problem. ^[4]
• Pregnancy: paracetamol remains the analgesic of choice. Acetylcysteine crosses the placenta and is used when overdose is diagnosed; delay in treatment increases the risk to mother and fetus. ^[6] Decisions in pregnancy should be made with poison-centre and obstetric input.
• Older adults: more susceptible to several hepatotoxins and to liver failure; lower thresholds for referral. ^[2]
• Pre-existing liver disease: patients are not necessarily more likely to develop DILI with most drugs, but have worse outcomes when it occurs. Antituberculous and methotrexate injury are exceptions, where abnormal baseline enzymes matter. ^[2]
• Immunocompromised and ICI patients: immune-related hepatitis needs prompt recognition, grading and immunosuppression. ^[1][2]
• High and low body weight in paracetamol poisoning: Glasgow protocols cap the calculated dose weight at 110 kg and use pre-pregnancy weight for the toxic dose in pregnancy. ^[7]

• A simple, aminotransferase-based screen (AST, ALT, ALP, bilirubin) plus an abdominal ultrasound reaches almost all the clinically relevant information in most DILI. ^{[1] [2]}
• Stopping the drug is free and should not be delayed for a causal score. ^[1][2]
• Where acetylcysteine or antidotes are unavailable, supportive care (fluids, nutrition, antiemetics, treatment of infection) and early referral to a centre with transplant expertise remain the priorities. ^[4] Oral and intravenous regimens differ, so follow local formulary and poisons guidance.
• Paracetamol nomograms can be plotted on paper; the UK single line (100 mg per litre at 4 hours, 15 mg per litre at 15 hours) and weight-based dosing tables are designed for hard-copy use. ^[5][6]

Abnormal liver function tests and jaundice; acute liver failure; autoimmune hepatitis; drug eruptions and severe cutaneous adverse reactions; drugs and teratogens in pregnancy; common drug interactions and reactions; opioid, sedative, tricyclic and cardiac drug overdose; major drug side-effects table.

• DILI is liver injury caused by a drug, herbal or dietary supplement, diagnosed after competing causes are excluded. ^[2]
• Three mechanisms: direct (predictable, dose-related; paracetamol), idiosyncratic (unpredictable, usually not dose-related; amoxicillin-clavulanate) and indirect (immune-mediated; checkpoint inhibitors). ^[3]
• Most patients recover once the drug is stopped. N-acetylcysteine is not a routine antidote for DILI; it is for paracetamol (and acute liver failure). ^[1]
• Idiosyncratic DILI: incidence 14 to 19 per 100,000/year; most drugs injure fewer than 1 in 1000 exposed. ^[2]
• Enzyme pattern (hepatocellular, cholestatic, mixed) is captured by the R value; hepatocellular is R above 5, mixed R 2 to 5, cholestatic R below 2. ^[2]
• Paracetamol accounts for 45.7% of acute liver failure in North America and 65.4% in the United Kingdom. ^[4]
• Risk rises at high daily doses, with lipophilic, heavily metabolised drugs, with combination of hepatotoxins, and with genetic susceptibility. ^[2]
• Corticosteroids are not routine; they are for autoimmune-like or immune-related injury. ^[1][2]
• Causality uses structured tools (RUCAM score range -9 to +14), with categories definite (9 or more) through excluded (0 or less); positive rechallenge is the strongest proof and the most dangerous. ^[1][2]
• Paracetamol: single dose of 150 mg/kg or more can injure; minimum toxic adult dose 7.5 to 10 g; be careful with the 75 mg/kg/day paediatric daily limit. ^[4][6]
• Acetylcysteine is virtually 100% effective within 8 hours of ingestion and falls thereafter; the UK line is 100 mg/L at 4 hours and 15 mg/L at 15 hours. ^[5]
• The intravenous regimen is 300 mg per kg in three infusions over 21 hours; give it without waiting for a level in staggered or timed-unknown overdoses. ^[6]
• King's College Criteria (arterial pH below 7.30 after resuscitation, or PT above 100 seconds, creatinine above 3.4 mg/dL and grade III/IV encephalopathy) define likely need for transplantation. ^[4]
• Watch for chronic DILI, biliary complications and rechallenge reactions in follow-up. ^[2]