Decision support only — acute liver failure, acute severe AIH and severe DILI need immediate hepatology/transplant-centre discussion. Stop a plausible toxic agent, but do not diagnose DILI until competing causes have been sought.
1 Three pathways that overlap
Autoimmune hepatitisimmune-mediated; often chronic, sometimes acute
Hepatocellular enzymes, raised IgG, ANA/SMA/LKM/LC1/SLA ± other autoimmunity; biopsy with interface hepatitis supports diagnosis after exclusions.
Confirm active disease and treat. Predniso(lo)ne plus azathioprine or MMF is current first-line therapy. EASL 2025
DILI / HDS injurydiagnosis of exclusion
A plausible drug, supplement or dose change; compatible latency and phenotype; alternatives excluded; improvement after withdrawal.
Stop the suspect agent and grade severity. Most idiosyncratic DILI has no specific antidote; report serious/suspected cases to the TGA.
Drug-induced autoimmune-like hepatitisDI-ALH
Drug exposure plus AIH-like IgG, antibodies or histology; common culprits include nitrofurantoin, minocycline and some biologic/oncology therapies.
Withdraw drug ± short steroid course. Durable resolution without relapse after immunosuppression ends favours DI-ALH over idiopathic AIH.
Acute severe / liver failureINR rising, jaundice ± encephalopathy
Acute severe AIH without encephalopathy; or ALF defined by coagulopathy plus encephalopathy without established cirrhosis.
Discuss with a transplant centre now. In acute severe AIH without ALF, trial corticosteroid and judge response at 3–7 days; do not let a steroid trial delay listing. EASL 2025
Autoantibodies are supporting evidence, not ownership papers. They occur in DILI, MASLD and viral hepatitis; AIH can also be seronegative or have normal IgG in an acute presentation.
2 Initial work-up: pattern and severity first
Define the pattern
- R ratioAt presentation: (ALT/ULN) ÷ (ALP/ULN). ≥5 hepatocellular; ≤2 cholestatic; 2–5 mixed.
- SeverityBilirubin fraction, INR, glucose, lactate, albumin, platelets/creatinine; mental state and trajectory. INR and encephalopathy matter more than a dramatic ALT.
- ImagingUltrasound with Doppler first; MRCP/CT when cholestasis, obstruction, vascular disease or malignancy remains plausible.
Exclude common alternatives
- ViralHAV IgM, HBsAg/anti-HBc IgM/HBV DNA as needed, HCV RNA; HEV in compatible/travel or unexplained cases; CMV/EBV/HSV by host and severity.
- OtherIschaemia/sepsis/congestion, alcohol/MASLD, biliary obstruction, Wilson in young/ALF, Budd–Chiari, muscle injury (CK) and coeliac disease.
- OverdoseParacetamol level and history in acute severe hepatocellular injury, even when overdose is denied or timing is unclear.
DILI chronology
- EverythingPrescription, OTC, herbals, bodybuilding/weight-loss products, teas, traditional medicines and illicit drugs; record start, stop, dose change and prior exposure.
- LatencyUsually days to 6 months, but nitrofurantoin, amoxicillin–clavulanate, methotrexate and immune therapies can present late or after cessation.
- RechallengeAvoid deliberate rechallenge unless the medicine is essential and a specialist-led risk plan exists.
3 AIH diagnosis: assemble, do not cherry-pick
| Simplified IAIHG item | Points | Registrar note |
| ANA or SMA | ≥1:40 = 1; ≥1:80 = 2 | Or anti-LKM1 ≥1:40 or anti-SLA positive = 2. Maximum autoantibody score 2; assay/cut-off matters. |
| IgG | >ULN = 1; >1.1 × ULN = 2 | Polyclonal hypergammaglobulinaemia supports AIH; normal IgG does not exclude an acute presentation. |
| Histology | Compatible = 1; typical = 2 | Biopsy usually confirms, stages and tests alternatives. Ask pathology specifically about AIH, DILI, steatohepatitis and cholestatic overlap. |
| Viral hepatitis excluded | Yes = 2 | Score ≥6 probable, ≥7 definite. The score is less reliable in acute severe disease and was not designed to distinguish DI-ALH. |
Also check AMA and cholestatic imaging where PBC/PSC overlap is possible; screen thyroid disease and coeliac disease at AIH diagnosis.
4 Treat autoimmune hepatitis
| Setting | Regimen | Monitoring / caution |
| Active AIH; non-severe to severe | Predniso(lo)ne 0.5 mg/kg/day; up to 1 mg/kg/day for more severe disease. Add either MMF or azathioprine as below. Taper to biochemical response, not a fixed calendar. | Check glucose, BP, infection, mood, bone and eye risk. HAV/HBV vaccination; DEXA at treatment start; calcium/vitamin D as appropriate. EASL 2025 |
| MMF first-line option | MMF 1,000 mg/day with steroid, increase around week 2 to 1.5–2 g/day if tolerated. | Teratogenic. Pregnancy test and effective contraception; counsel patients who can conceive and male patients. Avoid in pregnancy. CAMARO showed better 6-month biochemical response/tolerability than azathioprine. |
| Azathioprine first-line option | Start 50 mg/day, preferably ~2 weeks after steroid and when bilirubin <6 mg/dL (103 μmol/L); increase by week 4 toward 1–2 mg/kg/day. | TPMT ± NUDT15 by local practice; FBC/LFT closely in first 6 weeks. Do not use alone for induction; avoid in acute severe AIH and generally in decompensated cirrhosis. |
| Acute severe AIH without ALF/ACLF | Predniso(lo)ne 0.5–1 mg/kg/day or equivalent IV methylprednisolone. | Assess bilirubin/INR/clinical response at days 3–7. Failure to improve → transplant-centre pathway. With ALF/ACLF, discuss transplant directly; evidence for steroid rescue is poor. |
| Budesonide | Not first-line in EASL 2025 and contraindicated in cirrhosis. | May be a specialist switch for steroid adverse effects in selected non-cirrhotic, prednisolone-dependent patients. Older sheets that lead with budesonide are now out of date. |
5 Response, maintenance and pregnancy
Measure the right endpoint
- TargetComplete biochemical response: normal aminotransferases and IgG, assessed within 6–12 months. Symptoms improving alone is insufficient.
- CheckALT/AST, bilirubin, INR and IgG around weeks 4, 12 and 24; safety labs earlier/more often with azathioprine or severe disease. Then every 3–6 months in stable maintenance.
- Non-responseRevisit diagnosis, adherence, dose and toxicity. Thiopurine metabolites can distinguish underexposure/shunting from true failure.
Usually long-term treatment
- MaintainAzathioprine or MMF alone, or with predniso(lo)ne ≤5 mg/day, at the lowest regimen that sustains complete response.
- WithdrawalOnly a carefully selected patient with stable complete response for ≥2 years on low-dose monotherapy; taper stepwise and monitor closely. Most need lifelong therapy.
- RelapseRise in aminotransferases/IgG after response or withdrawal → confirm adherence/alternative cause, then promptly re-induce.
Pregnancy
- BeforeAim for complete biochemical response. Stop MMF at least 12 weeks before conception and switch under hepatology supervision.
- CompatiblePredniso(lo)ne and azathioprine are commonly continued when needed; uncontrolled AIH is more dangerous than appropriate maintenance therapy.
- PostpartumFlares are common: arrange close aminotransferase/IgG surveillance after delivery.
6 DILI: recognise, stop, support
| Decision | Threshold / action | What the evidence does not support |
| Clinically significant DILI | AST or ALT >5 × ULN, or ALP >2 × ULN, on two tests ≥24 h apart; or bilirubin >2.5 mg/dL (43 μmol/L) with enzyme elevation; or INR >1.5 with enzyme elevation. Stop plausible agent and investigate. AASLD 2023 | Milder abnormalities can still be DILI; thresholds help case definition, not permission to ignore a symptomatic patient. |
| Hy’s law signal | Hepatocellular injury with ALT/AST >3 × ULN and total bilirubin >2 × ULN, without major cholestasis or a better cause → high-risk specialist review. | It predicts population drug risk, not certain death in an individual. Check direct bilirubin, obstruction, sepsis and haemolysis. |
| Support / antidote | Stop culprit; fluids/nutrition/pruritus care; specific antidote when applicable. Use NAC for paracetamol toxicity; consider a 3-day NAC course in adults with DILI-related ALF, especially early encephalopathy. | Routine NAC for stable non-ALF idiosyncratic DILI is unproven. Ursodeoxycholic acid may help cholestatic symptoms but has no established outcome benefit. |
| Corticosteroid | Reserve for convincing AIH/DI-ALH, DRESS/hypersensitivity or immune-checkpoint/selected kinase-inhibitor hepatitis, with specialist plan and response date. | Steroids are not routine DILI treatment. In indeterminate ALF they can worsen infection and delay transplant decisions. |
| Escalate / report | Rising INR/bilirubin, encephalopathy, hypoglycaemia, AKI, acidosis or shrinking liver → ICU/transplant centre. Report suspected serious or unusual medicine/supplement injury to the TGA AEMS. | You do not need causal certainty to report a suspected adverse reaction. |
7 Follow-up and the AIH–DILI fork
After suspected DILI
- Serial testsFrequency follows severity and trajectory; hepatocellular injury can deteriorate quickly, while cholestatic injury can take months to resolve.
- ChronicityPersistent biochemical or imaging abnormality at 6–12 months needs reassessment for chronic DILI, vanishing bile-duct syndrome or a missed competing disease.
- RecordGeneric and brand name, product photo/ingredients for supplements, latency, phenotype, severity, outcome and explicit future avoidance advice.
Was it DI-ALH or idiopathic AIH?
- Favour DI-ALHClear drug latency, complete resolution after withdrawal ± short steroid, and no relapse during long follow-up after immunosuppression ceases.
- Favour AIHAdvanced fibrosis/cirrhosis at presentation, persistent disease after drug withdrawal or relapse after steroid-sparing therapy is removed.
- Time decidesHistology and scores overlap. Label uncertainty honestly and retain hepatology follow-up long enough to observe the disease course.
Guidelines & reviews. EASL Clinical Practice Guidelines on autoimmune hepatitis 2025; AASLD Practice Guidance on drug, herbal and dietary supplement-induced liver injury 2023; EASL DILI guideline 2019; International AIH Pathology Group criteria; IAIHG/EASL-associated expert report on drug-induced autoimmune-like hepatitis; Australian TGA adverse-event reporting guidance 2026. Trials. CAMARO (MMF vs azathioprine in treatment-naïve AIH), J Hepatol 2024; historic prednisolone/azathioprine trials; Lee et al. NAC in non-paracetamol ALF, Gastroenterology 2009. Caveats. AIH dosing/tapering must follow response and toxicity; simplified AIH scoring is supportive rather than definitive in acute severe disease or DILI. DILI causality tools assist structure but do not replace expert exclusion of competing causes.