Drug-induced liver injury: recognition and management in AU general practice

Drug-induced liver injury (DILI) is hepatic injury caused by exposure to a prescription medicine, over-the-counter product, herbal preparation, or dietary supplement — established by a compatible temporal relationship and the exclusion of competing causes. It is more common than most clinicians realise: DILI accounts for approximately 10% of acute hepatitis presentations referred to hepatology and up to 50% of cases of acute liver failure in Australia and Western countries, with paracetamol responsible for the largest share of the acute liver failure burden.

Herbal and dietary supplement-induced liver injury (HILI) — now an accepted subcategory — accounted for approximately 20% of cases in the most recent Drug-Induced Liver Injury Network (DILIN) prospective cohort, a figure that has tripled over two decades, driven by bodybuilding stacks, weight-loss formulations, green tea catechin extracts, and a rapidly rising turmeric/curcumin signal. GP-level recognition depends on asking — explicitly and without judgement — about every medicine, supplement, and herbal product taken in the last six months, not just what is currently on the prescription pad.

A. Core clinical — the AU general-practice framework

Classification: mechanism and biochemical pattern

DILI is classified by mechanism and by biochemical pattern at presentation.

Mechanism:

Mechanism	Key features	AU examples
Intrinsic (dose-dependent)	Predictable; reproducible in animals; latency hours to days	Paracetamol (NAPQI metabolite), methotrexate (cumulative fibrosis), high-dose niacin
Idiosyncratic (dose-independent)	Unpredictable within therapeutic range; immune-mediated or metabolic; latency days to months	Amoxicillin-clavulanate, flucloxacillin, isoniazid, statins, phenytoin, nitrofurantoin
Indirect	Drug alters immune tone or metabolism rather than directly injuring hepatocytes	Immune-checkpoint inhibitor hepatitis (pembrolizumab, nivolumab), HBV reactivation with rituximab

Biochemical pattern — R ratio (Danan, J Clin Epidemiol 1993; EASL 2019):

Calculate R = (ALT ÷ ALT-ULN) ÷ (ALP ÷ ALP-ULN) at the time of first presentation:

R ratio	Pattern	Common culprits
R > 5	Hepatocellular	Paracetamol, isoniazid, statins, NSAIDs, green tea extract, methotrexate
R < 2	Cholestatic	Amoxicillin-clavulanate, flucloxacillin, erythromycin, anabolic steroids, oestrogens
R 2–5	Mixed	Phenytoin, sulfonamides, carbamazepine, nitrofurantoin

Hy’s Law — the critical threshold

Hy’s Law predicts approximately 10% mortality or liver transplant when all three criteria are met simultaneously (Zimmerman 1978; FDA 2009):

1. ALT >3× ULN, AND
2. Total bilirubin >2× ULN, AND
3. No biliary obstruction, no Gilbert syndrome, no haemolysis as the explanation.

When Hy’s Law is met: stop all non-essential drugs immediately, admit, and contact the regional hepatology or liver transplant unit. This is a medical emergency.

History — the 6-month medication review

The critical step is a complete medication and supplement review covering the preceding six months (AASLD 2014; EASL 2019):

• Prescription medicines — including short courses already finished. Amoxicillin-clavulanate and flucloxacillin typically present two to eight weeks after the course ends, when the drug is no longer visible in the medicine cabinet.
• Over-the-counter products — total daily paracetamol across all products (including cold-and-flu combination tablets), NSAIDs, antihistamines, PPIs.
• Complementary and herbal products — use open-ended, non-judgemental language: “Do you take any teas, vitamins, gym supplements, weight-loss capsules, traditional medicines, turmeric capsules, or anything from a naturopath or health food shop?”
• Bodybuilding and weight-loss stacks — anabolic 17α-alkylated steroids, SARMs, Hydroxycut, Garcinia cambogia.
• Traditional medicines — Traditional Chinese Medicine herbal mixtures, Ayurvedic compounds (heavy-metal contamination risk), bush teas (comfrey, chaparral, pyrrolizidine alkaloid teas).
• Recreational substances — MDMA, cocaine, anabolic agents, nitrous oxide.

Ask patients to bring their bottles to the next appointment. The exact brand, dose, and start date matter for RUCAM scoring and TGA reporting.

Symptom inventory: anorexia, nausea, vomiting, right upper quadrant ache, fatigue, jaundice, dark urine, pale stools, pruritus, rash, fever, arthralgia, lymphadenopathy.

Risk factors for severe DILI: older age, female sex, polypharmacy, prior DILI episode, chronic liver disease (paracetamol threshold substantially lower), malnutrition or fasting (depleted glutathione), chronic alcohol use, slow-acetylator CYP2D6 status (isoniazid, sulfonamides).

Examination

• General: jaundice, fever, lymphadenopathy, facial oedema, morbilliform or pustular rash (DRESS, AGEP, Stevens-Johnson spectrum).
• Abdominal: tender hepatomegaly (acute), splenomegaly, ascites (decompensation).
• Neurological: asterixis, altered consciousness, fetor hepaticus — any encephalopathy mandates ICU-level escalation.
• Stigmata of chronic liver disease if pre-existing: spider naevi, palmar erythema, Dupuytren’s contracture, gynaecomastia.

Investigations

First-line:

• LFTs (serial) — ALT, AST, ALP, GGT, total and direct bilirubin, albumin. Calculate the R ratio at first presentation. Recheck within 24–72 hours to establish trajectory.
• INR / coagulation — INR ≥1.5 signals hepatic synthetic failure; escalate.
• FBC with differential — eosinophilia ≥5% supports immune-allergic DILI / DRESS.
• Glucose, electrolytes, lactate, ammonia if encephalopathy is possible.
• Paracetamol level — obtain in every acute hepatocellular DILI presentation, including when the patient does not volunteer recent use. Plot on the Rumack-Matthew nomogram for single acute ingestion ≥4 hours post-ingestion.
• Viral hepatitis exclusion — HBsAg, anti-HBc IgM, anti-HCV (with HCV RNA if antibody-positive), HAV IgM, HEV IgM and HEV RNA (HEV is a commonly missed mimic of DILI in Australia), CMV/EBV PCR in atypical cases.
• Autoimmune panel — ANA, anti-smooth muscle antibody, anti-LKM1, anti-SLA, AMA, total IgG to exclude autoimmune hepatitis and distinguish AIH-like DILI.
• Iron studies, caeruloplasmin, α1-antitrypsin — rule out haemochromatosis, Wilson disease, and α1-AT deficiency.
• Liver and abdominal ultrasound with Doppler — mandatory in cholestatic pattern to exclude biliary obstruction and Budd-Chiari syndrome.

RUCAM causality scoring (Danan, J Clin Epidemiol 1993) assigns a score across seven domains — time to onset, course after cessation, risk factors, concomitant drugs, exclusion of non-drug causes, previous hepatotoxicity information, and rechallenge response. Scores of 3–5 indicate possible DILI, 6–8 probable, and ≥9 highly probable. Use the hepatocellular or cholestatic/mixed column to match the R ratio pattern. The NIH LiverTox database provides a published likelihood score (A = well-established to E = unlikely) for individual agents and is an invaluable point-of-care resource.

Pitfall — latency: The latency trap is the commonest reason DILI is missed. Amoxicillin-clavulanate, flucloxacillin, isoniazid, and nitrofurantoin regularly present weeks to months after exposure has ended. Asking only about current medications will miss these cases.

Differential diagnosis

Differential	Key discriminator
Viral hepatitis A / B / C / E	Serology and RNA; HEV is under-tested and commonly missed
Autoimmune hepatitis	ANA/ASMA, elevated IgG, biopsy; may be serologically indistinguishable from AIH-like DILI — drug withdrawal first
Ischaemic hepatitis	ALT often >50× ULN with rapid fall; preceding hypotension, sepsis, or arrhythmia
Biliary obstruction	Dilated ducts on US or MRCP; gallstones; malignancy
Wilson disease	Age <40; low caeruloplasmin; haemolytic anaemia; Kayser-Fleischer rings; low ALP
Haemochromatosis	Ferritin >1,000, transferrin saturation >45%, HFE genotype
MASLD/MASH flare	Metabolic syndrome; imaging steatosis; no temporal drug association
Alcohol-related hepatitis	AST > ALT ratio >2:1; recent heavy drinking history
HELLP / acute fatty liver of pregnancy	Third trimester; haemolysis; thrombocytopaenia; hypoglycaemia
Hepatic congestion	Right heart failure; Budd-Chiari on Doppler ultrasound
Sepsis-associated cholestasis	Sepsis source; mild ALT, marked bilirubin elevation

GP management

Immediate steps (all grades):

1. Stop the suspected agent and all non-essential medicines and supplements. In polypharmacy, stop everything reasonably stoppable and reintroduce essentials one at a time with LFT monitoring.
2. Apply Hy’s Law — if met, admit to hospital and contact hepatology or the regional liver transplant centre.
3. Paracetamol nomogram — if any possibility of paracetamol exposure exists, check a level and plot it on the Rumack-Matthew nomogram. If at or above the treatment line, start IV NAC without delay using the standard three-bag regimen per AMH.
4. Supportive care — IV fluids, antiemetics, glucose monitoring, electrolyte correction.
5. No rechallenge — document clearly in the patient record and My Health Record allergy field. Rechallenge is associated with rapid and severe recurrence; it is not acceptable outside narrow specialist-supervised scenarios involving genuinely life-saving therapy with no alternative.
6. Report to the TGA DAEN for serious DILI.

Monitoring after drug cessation:

• Acute: LFTs ± INR at 1 week, 2 weeks, then four-weekly until normalisation.
• If LFT abnormality persists beyond 12 weeks, refer to hepatology for assessment of chronic DILI, AIH-like DILI requiring biopsy, or fibrosis staging.

B. Evidence appraisal — key controversies

NAC for non-paracetamol acute liver failure: The Lee 2009 randomised controlled trial demonstrated that IV NAC improves transplant-free survival in early-stage (coma grade I–II) non-paracetamol acute liver failure. This is an adjunct used at tertiary hepatology / ICU level; it is not routinely recommended for mild idiosyncratic DILI in the general-practice setting.

Corticosteroids in cholestatic DILI: DILIN prospective cohort data (Chalasani 2021) and AASLD 2021 guidance do not support corticosteroid use in routine cholestatic DILI — no clear benefit has been demonstrated, and the risks of immunosuppression in this population are real. Corticosteroids are appropriate for DRESS with hepatic involvement, AIH-like DILI where drug withdrawal has failed, and immune-checkpoint-inhibitor hepatitis at grade ≥2 severity (specialist-supervised).

Turmeric/curcumin hepatotoxicity: Halegoua-DeMarzio et al. (Am J Med 2023) and Lukic et al. (Eur J Gastroenterol Hepatol 2018) document an accelerating international case series of hepatocellular injury from turmeric and curcumin supplements, particularly piperine-enhanced formulations. The NIH LiverTox curcumin chapter identifies an HLA-B*35:01 genetic association. Dietary turmeric in cooking remains safe; the risk is concentrated in supplement formulations. The TGA DAEN reflects an increasing Australian signal.

RUCAM versus RECAM: The original Danan-Bénichou RUCAM instrument (1993) remains the international standard and is embedded in LiverTox. The updated RECAM (Hayashi, Hepatology 2022) improves inter-rater reliability and performance in digital workflows. RUCAM is appropriate for use in general practice; RECAM is emerging at specialty level. Either tool is acceptable; applying one systematically is what matters clinically and medico-legally.

Rechallenge: Rechallenge is effectively closed-door for idiosyncratic DILI. AASLD 2014 and EASL 2019 both state that rechallenge produces more rapid, more severe, and sometimes fatal recurrence. The narrow exception — a life-saving drug such as an anti-tuberculosis regimen with no alternative — requires specialist hepatology supervision with liver function monitoring at a frequency determined by the prior course severity.

C. Australian-specific culprits and the HILI signal

The DILIN prospective cohort (Chalasani, Gastroenterology 2015 and 2021) provides the most robust data on DILI causation in Western healthcare systems. Amoxicillin-clavulanate accounts for approximately 10% of cases in registry data — the most commonly implicated single agent — followed by nitrofurantoin, trimethoprim-sulfamethoxazole, isoniazid, and azithromycin. Several patterns are clinically important:

Amoxicillin-clavulanate — cholestatic or mixed pattern; latency two to eight weeks after a course ends; predominant in older patients; generally resolves after cessation but can be prolonged. Prescribers should be aware when seeing an elderly patient with new jaundice or pruritus and a recent sinusitis or dental infection treated by another clinician.

Flucloxacillin — cholestatic; HLA-B*5701 genetic association makes it unpredictable in carriers; prolonged cholestasis (weeks to months); rare but severe.

Isoniazid — hepatocellular pattern; risk increases with age, concurrent rifampicin, underlying liver disease, and alcohol use. Baseline LFTs before starting TB treatment and monthly monitoring for the first two months are standard.

Statins — mild transaminase elevation is common and does not require cessation unless symptomatic or progressive; statins are not contraindicated in stable chronic liver disease and the evidence supports their cardiovascular benefit in MASLD. Severe statin hepatotoxicity is rare.

Methotrexate — cumulative hepatotoxic mechanism; risk stratified by cumulative dose and alcohol use; liver biopsy or FibroScan at defined cumulative dose thresholds under specialist protocols (typically rheumatology or dermatology-led).

Immune-checkpoint inhibitors (pembrolizumab, nivolumab, ipilimumab) — immune-mediated hepatitis managed per oncology immunotoxicity pathways; the GP’s role is to flag new hepatitis in a patient receiving checkpoint therapy and refer to the treating oncologist before initiating corticosteroids.

HILI culprits in Australia — the LiverTox database provides individual herb and supplement hepatotoxicity profiles. High-risk categories include: turmeric/curcumin supplements (especially piperine-enhanced), kava (documented acute liver failure; restricted or banned in several jurisdictions), green tea extract / EGCG capsules (particularly in weight-loss protocols), anabolic 17α-alkylated steroids and SARMs (cholestatic + peliotic injury), Hydroxycut and Garcinia cambogia (idiosyncratic, ALF case reports), and pyrrolizidine-alkaloid-containing herbal teas (comfrey, chaparral, bush teas).

D. Australian operations

MBS items

The following Medicare Benefits Schedule items are relevant to DILI work-up in general practice:

• GP consultation: items 23 / 36 / 44 — Level C or D appropriate for a comprehensive medication review and full work-up assembly.
• LFTs (biochemistry): item 66512.
• Abdominal ultrasound: item 55054 — mandatory in cholestatic pattern.
• MRCP (magnetic resonance cholangiopancreatography): item 63491 — for cholestatic pattern not explained by ultrasound.
• Percutaneous liver biopsy: item 30419 — specialist-performed; for diagnostic uncertainty or AIH-like DILI.
• Hepatology specialist attendance: items 132 / 133.
• GP Mental Health Care Plan: items 2715 / 2717 — for intentional paracetamol overdose, eating-disorder context, or chronic-illness adjustment.
• Chronic disease management (GPCCMP): items 965 / 967 — for chronic DILI or persistent post-DILI hepatic dysfunction requiring allied-health coordination.
• Telehealth GP: existing-relationship video items 91790 / 91891 for trajectory monitoring of mild-moderate DILI once stable.
• ATSI Health Assessment: item 715.

PBS medications

• N-acetylcysteine (IV) — hospital/emergency department supply for paracetamol overdose; not on the PBS general schedule.
• Prednisolone — PBS general schedule, unrestricted; for DRESS, AIH-like DILI, and immune-checkpoint hepatitis under specialist supervision.
• Cholestyramine — PBS general schedule; for cholestatic-DILI-related pruritus.
• Azathioprine — PBS general schedule; rarely used as steroid-sparing in chronic AIH-like DILI; TPMT genotype screening recommended before initiation.

TGA DAEN reporting

Serious DILI — defined as leading to hospitalisation, prolonged hospitalisation, life-threatening illness, congenital abnormality, or death — requires mandatory reporting to the TGA DAEN portal under the Therapeutic Goods Act 1989. Voluntary reporting is encouraged for all suspected DILI, including HILI from supplement products. Include the product brand name, batch number where available, dose, duration, and LFT trajectory. Consumer reports are also accepted via the DAEN portal.

Referral pathways

• Routine: persistent LFT abnormality >6 months after cessation; chronic DILI; fibrosis assessment.
• Semi-urgent (within days): ALT >5× ULN or ALP >2× ULN persisting after drug cessation; cholestatic DILI with jaundice; suspected AIH-like DILI needing biopsy; suspected DRESS.
• Urgent / Emergency: Hy’s Law met; INR ≥1.5; any hepatic encephalopathy; paracetamol overdose (for IV NAC and monitoring); severe drug rash with hepatic involvement (DRESS, Stevens-Johnson syndrome overlap).

Include with every DILI referral: the full medication and supplement list with dates, LFT trajectory, INR, FBC with eosinophil count, viral hepatitis and autoimmune serology results, paracetamol level, imaging, and RUCAM score if calculated.

E. Special populations

Pregnancy

Paracetamol overdose management is unchanged in pregnancy — IV NAC is safe and should not be withheld on grounds of gestation. Acute fatty liver of pregnancy (AFLP) and HELLP syndrome are important mimics of drug-induced hepatic injury in the third trimester; they require obstetric emergency management rather than drug cessation. The obstetric team should be involved in any pregnancy complicated by Hy’s Law-level hepatotoxicity.

Older adults

Older adults are the most common demographic affected by amoxicillin-clavulanate and flucloxacillin DILI. A new LFT abnormality in an older patient with polypharmacy and a recent antibiotic course within the last two months should trigger a structured 6-month medication review. Paracetamol thresholds are lower in the context of age-related reduction in hepatic glutathione reserves and reduced nutritional reserve.

Children and adolescents

Weight-based NAC dosing applies for paediatric paracetamol overdose per AMH protocols. Adolescents may use anabolic or performance-enhancing supplements without disclosure; direct, confidential questioning is important. HEV serology should be included in the paediatric DILI work-up.

Pre-existing liver disease (MASLD, viral hepatitis, alcohol)

Patients with pre-existing chronic liver disease have a lower threshold for DILI severity — paracetamol toxicity can occur at lower doses in the context of alcohol use, fasting, or malnutrition. The standard recommendation is to limit paracetamol to no more than 2 g/day in these patients. Statins are not contraindicated in well-compensated chronic liver disease; methotrexate and NSAIDs carry higher risk in pre-existing hepatic fibrosis or portal hypertension and should be avoided or used under specialist review.

Aboriginal and Torres Strait Islander peoples

Coordinate post-DILI care with ACCHO services. Closing the Gap PBS Co-payment applies for PBS medications. Culturally safe communication about traditional and herbal medicine practices should be approached in partnership rather than as prohibition — the goal is disclosure and harm reduction, not disengagement.

When to escalate

Call 000 / transfer to emergency immediately if:

• Hepatic encephalopathy (confusion, asterixis, drowsiness after paracetamol or other hepatotoxin exposure).
• INR ≥1.5 at any point after drug cessation.
• Hy’s Law criteria met (ALT >3× ULN + bilirubin >2× ULN without competing cause).
• Severe drug rash with mucosal involvement, skin blistering, or systemic features (possible DRESS or Stevens-Johnson).
• Known paracetamol overdose — the nomogram should be applied in an emergency setting with IV NAC available.

Safety-netting all patients with DILI: Provide written instructions — “Return urgently or call 000 if you develop yellow eyes or skin, very dark urine, confusion or unusual drowsiness, severe abdominal pain, high fever, bleeding or bruising easily, or a severe rash with sores in your mouth.”

What this article is and is not

This article is a clinical education resource for Australian-registered general practitioners, written by Dr Hoe Bing Lo (MBBS, FACRRM, AHPRA MED0001212640) and reflecting eTG, AASLD 2021 Practice Guidance, EASL 2019 Guidelines, and DILIN prospective cohort evidence as at the date of review.

This article does not constitute medical advice for patients. People with concerns about a medicine, supplement, or herbal product affecting their liver should speak with their doctor. If jaundice, dark urine, confusion, or severe rash develops, seek emergency care or call 000.

---

Sources cited

• eTG complete — Drug-induced liver injury
• RACGP — general clinical resources
• GESA — Gastroenterological Society of Australia: liver disease information
• TGA Database of Adverse Event Notifications (DAEN)
• Australian Medicines Handbook — NAC, paracetamol, methotrexate
• Fontana et al. AASLD Practice Guidance: Idiosyncratic DILI. Hepatology 2023
• Chalasani et al. ACG Clinical Guideline: Diagnosis and management of DILI. Am J Gastroenterol 2014
• Andrade et al. EASL Clinical Practice Guidelines: Drug-induced liver injury. J Hepatol 2019
• NIH LiverTox — Drug-Induced Liver Injury database
• NIH LiverTox — Curcumin / turmeric
• NIH LiverTox — Kava
• NIH LiverTox — Green tea extract
• Danan and Bénichou. RUCAM causality assessment. J Clin Epidemiol 1993
• Chalasani et al. DILIN prospective study causes, features, outcomes. Gastroenterology 2015
• Chalasani et al. DILIN 899 patients. Gastroenterology 2021
• Lee et al. IV NAC in early non-paracetamol ALF. Gastroenterology 2009
• Rumack and Matthews. Acetaminophen nomogram. 1975
• Zimmerman. Hepatotoxicity: adverse effects of drugs. 1978
• Halegoua-DeMarzio et al. Liver injury from turmeric. Am J Med 2023
• Lukic et al. Hepatotoxicity from turmeric. Eur J Gastroenterol Hepatol 2018
• O’Grady et al. King’s College criteria for fulminant hepatic failure. Gastroenterology 1989
• HealthDirect — Liver problems
• Better Health Channel — Liver
• LiverWELL (formerly Hepatitis Victoria)