Drug-Induced Liver Injury: ACG and EASL Diagnosis with R-Ratio Hy’s Law RUCAM Causality and N-Acetylcysteine Management
Executive Summary
S1. Pathophysiology and Drug Classes
Drug-induced liver injury arises through two fundamentally different mechanisms. Intrinsic (direct, dose-dependent) hepatotoxicity occurs predictably above a threshold dose in nearly all exposed individuals — the prototype is acetaminophen, where excess substrate overwhelms glutathione conjugation pathways and the toxic metabolite N-acetyl-p-benzoquinone imine (NAPQI) covalently binds hepatocyte proteins, producing centrilobular necrosis within 24–72 hours. Idiosyncratic hepatotoxicity occurs unpredictably in 1 in 1,000 to 1 in 100,000 exposed individuals, generally without strict dose dependence, after a latency period of 5–90 days; mechanisms include reactive metabolite formation by polymorphic CYP enzymes, mitochondrial injury, immune-mediated drug-protein adduct presentation via specific HLA alleles, and adaptive immune failure. Immune-mediated idiosyncratic DILI may be hapten-driven (drug or metabolite covalently binds liver protein → MHC class I/II presentation → CD4/CD8 T-cell hepatitis) or may resemble systemic hypersensitivity (DRESS) with rash, eosinophilia, fever, and lymphadenopathy.
Most common implicated agents
| Class | Examples | Typical pattern | Notable features |
|---|---|---|---|
| Analgesic — intrinsic | Acetaminophen (paracetamol) | Hepatocellular (ALT often >1000–10,000 U/L) | Most common cause of US ALF; treat with NAC |
| Antibiotics | Amoxicillin-clavulanate (most common idiosyncratic in West), isoniazid, nitrofurantoin, minocycline, fluoroquinolones, macrolides, sulfa drugs | Cholestatic (amox-clav, macrolides) or hepatocellular (INH, nitrofurantoin) | Amox-clav latency 1–8 weeks; INH 2–6 months |
| Antiepileptics | Phenytoin, carbamazepine, valproate, lamotrigine, phenobarbital | Hepatocellular ± hypersensitivity (DRESS) | HLA-B*15:02 with carbamazepine in Asians; valproate mitochondrial toxicity in children <2 y |
| Statins | All HMG-CoA reductase inhibitors | Hepatocellular (mostly mild ALT elevation) | Severe DILI rare (<1/100,000); routine LFT monitoring not recommended |
| Antifungals | Ketoconazole, itraconazole, terbinafine, fluconazole | Hepatocellular or mixed | Ketoconazole highest risk; check baseline + monthly LFTs for terbinafine |
| Antituberculous | Isoniazid, rifampin, pyrazinamide | Hepatocellular (synergistic with INH + RIF) | Risk increases with age >35, HBV/HCV, alcohol; baseline + monthly LFTs |
| Immune checkpoint inhibitors | Pembrolizumab, nivolumab, ipilimumab, atezolizumab, durvalumab | Hepatocellular ± cholestatic ± immune-mediated | 3–9% incidence; highest with combo (CTLA-4 + PD-1); steroid-responsive |
| NSAIDs | Diclofenac, ibuprofen, sulindac, nimesulide | Hepatocellular or autoimmune-like | Diclofenac risk highest; check 2–6 month latency |
| Antineoplastic — chemo | Methotrexate, busulfan, cyclophosphamide, oxaliplatin (sinusoidal obstruction) | Variable — fibrosis (MTX), VOD/SOS, steatohepatitis (oxaliplatin) | Methotrexate cumulative-dose threshold; FibroScan monitoring |
| Anesthetic | Halothane (rare now), isoflurane, sevoflurane | Hepatocellular — repeat exposure increases risk | Halothane hepatitis once classic; halt all halogenated anesthetics |
| Herbal & dietary supplements | Green tea extract (EGCG), kava, Garcinia cambogia, anabolic steroids, Ma huang, comfrey, He shou wu, Hydroxycut | Variable — anabolic androgens cause cholestasis; green tea hepatocellular | ~20% of US DILIN registry; underreported; review ALL OTC products |
| Antiplatelet/anticoagulant | Ticlopidine, clopidogrel, warfarin (rare), DOACs (very rare) | Cholestatic (ticlopidine) or hepatocellular | Ticlopidine latency 1–3 months |
| Cardiovascular | Amiodarone (steatohepatitis), methyldopa, hydralazine | Steatohepatitis (amiodarone — months to years); autoimmune-like (methyldopa, hydralazine) | Amiodarone cumulative dose >100 g; baseline + 6-month LFTs; phospholipidosis |
S2. Clinical Presentation, R-Ratio, and Hy’s Law
The clinical phenotype of DILI is highly variable, ranging from asymptomatic enzyme elevation discovered on routine labs (most common) to fulminant acute liver failure. Symptoms — when present — include fatigue, anorexia, nausea, right upper quadrant discomfort, pruritus (cholestatic), jaundice, dark urine, and clay-colored stools. Hypersensitivity features (rash, eosinophilia, fever) are uncommon but specific when present. Latency between drug exposure and biochemical abnormality is one of the most useful clues: most idiosyncratic DILI presents within 5 days to 12 weeks of starting a new medication, although some agents (nitrofurantoin, amiodarone, methotrexate) have latencies of months to years.
R-ratio: pattern of injury
The R-ratio classifies DILI into hepatocellular, cholestatic, or mixed pattern based on the relative elevations of ALT and alkaline phosphatase at presentation. Calculate as: R = (ALT / ULN of ALT) ÷ (ALP / ULN of ALP). The pattern dictates differential diagnosis priorities and informs prognosis and treatment.
| R-Ratio | Pattern | Typical offenders | Differential to exclude |
|---|---|---|---|
| R ≥ 5 | Hepatocellular | Acetaminophen, INH, NSAIDs, methotrexate, statins, halothane, ICIs | Viral hepatitis (A, B, C, E, EBV, CMV, HSV), AIH, ischemic, Wilson, Budd-Chiari |
| R 2–5 | Mixed | Amox-clav (early), trimethoprim-sulfa, phenytoin, carbamazepine | AIH, biliary obstruction, sepsis, infiltrative |
| R ≤ 2 | Cholestatic | Amox-clav (late), erythromycin, anabolic steroids, oral contraceptives, ticlopidine | Biliary obstruction (US/MRCP), PBC, PSC, infiltrative (granulomatous) |
Hy’s Law — the 10% mortality marker
Hy’s Law was derived by Hyman Zimmerman from observations that hepatocellular jaundice from a drug carries far worse prognosis than enzyme elevation alone. The bilirubin requirement reflects the failing synthetic and excretory capacity of an injured liver — once jaundice appears, the threshold for transplant evaluation is low. Always recalculate Hy’s Law when LFTs progress: a patient who initially had isolated transaminitis but later develops bilirubin >2× ULN has crossed into Hy’s Law territory and is at high risk.
S3. Diagnostic Workup and RUCAM Causality Score
DILI remains a diagnosis of exclusion — there is no confirmatory test. The workup must rule out viral, autoimmune, vascular, biliary, and metabolic causes before attributing injury to a drug. ACG (2021) and EASL (2019) recommend the following baseline panel in any patient with suspected DILI.
Required exclusion workup
| Category | Tests | Trigger |
|---|---|---|
| Viral hepatitis | HAV IgM, HBsAg + anti-HBc IgM, anti-HCV + HCV RNA, HEV IgM/IgG (especially >50 years or immunocompromised), EBV/CMV PCR if atypical lymphocytes or fever | All patients with new ALT/AST elevation |
| Autoimmune hepatitis | ANA, ASMA, anti-LKM-1, IgG, AMA (for PBC); consider liver biopsy if AIH suspected | Female, hepatocellular pattern, no clear drug temporality |
| Biliary obstruction | RUQ ultrasound — assess CBD diameter, gallstones, masses; MRCP if US non-diagnostic and cholestatic | Cholestatic or mixed pattern, jaundice |
| Metabolic | Ferritin + transferrin saturation (hemochromatosis), ceruloplasmin + 24h urine copper (Wilson if <40 y), alpha-1 antitrypsin | Young patient or chronic enzyme elevation without obvious cause |
| Vascular | Doppler ultrasound (Budd-Chiari, portal vein thrombosis); MRI/CT venography if abnormal | Acute hepatocellular pattern with hepatomegaly, ascites, or thrombophilia |
| Toxic/ingestion | Acetaminophen level (immediately, even if denied); urine drug screen, alcohol level, herbal/supplement history | Every patient; especially altered mental status or unintentional overdose |
| Liver biopsy | Indicated when diagnosis unclear, AIH suspected, severe ongoing injury despite drug withdrawal, or cholestatic pattern persists >3 months | Selected cases — confirms necroinflammation, eosinophilic infiltrate, granulomas, or AIH features |
RUCAM causality score (Roussel Uclaf 2016 update)
The RUCAM score systematically grades the likelihood that a specific drug caused the observed liver injury, with separate algorithms for hepatocellular and cholestatic/mixed patterns. Score components include:
- Time to onset (5–90 days suggestive, >90 days possible) — score +1 to +2
- Course after drug discontinuation (dechallenge) — >50% drop in ALT within 8 days highly suggestive (+3); <50% drop within 30 days possible (+2); no change or worsening unfavorable
- Risk factors — age >55, alcohol use, pregnancy
- Concomitant medications — exclude known hepatotoxins or score against if competitor exists
- Exclusion of competing causes (viral, biliary, metabolic) — full exclusion +2
- Previous information on the drug (LiverTox database, label warnings) — known hepatotoxin +2
- Rechallenge — positive rechallenge = +3 (rarely available; ethically discouraged unless inadvertent)
| RUCAM total | Causality |
|---|---|
| ≤ 0 | Excluded |
| 1–2 | Unlikely |
| 3–5 | Possible |
| 6–8 | Probable |
| ≥ 9 | Highly probable |
S4. Management — Drug Withdrawal, NAC, Corticosteroids, Transplant
Step 1 — Immediate drug withdrawal
The single most important intervention in suspected DILI is prompt withdrawal of the offending agent. Stop the suspected drug and any non-essential potentially hepatotoxic concomitant medications. For Hy’s Law-positive patients, hospitalize and place a hepatology consult immediately. The expected pattern after dechallenge is gradual normalization of LFTs over 4–12 weeks for hepatocellular DILI and 2–6 months for cholestatic DILI; failure to improve within this window suggests an alternative diagnosis or progression to chronic DILI/vanishing bile duct syndrome.
Step 2 — Supportive care
- Avoid all non-essential medications and herbal supplements
- Adequate hydration; correct any volume depletion
- Treat coagulopathy with vitamin K (Vitamin K 10 mg SC daily × 3 days) only if response confirms vitamin K deficiency rather than hepatic synthetic failure; FFP if active bleeding
- Treat hepatic encephalopathy with Lactulose 30 mL PO q6h titrated to 2–3 soft stools/day ± Rifaximin 550 mg PO BID
- Treat pruritus in cholestatic DILI: Cholestyramine 4 g PO BID–QID, Ursodeoxycholic acid 13–15 mg/kg/day PO divided BID, or Rifampin 150–300 mg PO daily
Step 3 — N-acetylcysteine (NAC)
NAC is the antidote for acetaminophen-induced hepatotoxicity and has been shown to improve transplant-free survival in non-acetaminophen acute liver failure with early-stage encephalopathy (ALFSG study, NEJM 2009). NAC replenishes hepatic glutathione, scavenges free radicals, improves microcirculatory oxygen delivery, and modulates inflammatory cytokines.
| Indication | IV protocol (preferred) | Oral protocol |
|---|---|---|
| Acetaminophen overdose (any pattern) | NAC 150 mg/kg IV in 200 mL D5W over 60 min → 50 mg/kg over 4 h → 100 mg/kg over 16 h (21-h Prescott protocol; total 300 mg/kg) | NAC 140 mg/kg PO loading → 70 mg/kg PO q4h × 17 doses (72 h total) |
| Non-acetaminophen ALF with grade I–II encephalopathy | Same Prescott IV protocol — continue 21 h or longer until improvement | Not validated |
| Non-acetaminophen DILI without ALF | Routine NAC not recommended | — |
Step 4 — Corticosteroids in selected DILI
| Indication | Regimen | Notes |
|---|---|---|
| Drug-induced autoimmune hepatitis (DI-AIH; minocycline, nitrofurantoin, methyldopa) | Prednisone 30–40 mg PO daily taper over 4–8 weeks based on LFT response | Liver biopsy should confirm AIH features; consider azathioprine if steroid-dependent |
| DRESS syndrome with hepatic involvement | Prednisone 1 mg/kg/day PO with slow taper over 3–6 months | Monitor for HHV-6 reactivation; refer to dermatology + hepatology |
| ICI-induced hepatitis grade 2 (ALT 3–5× ULN) | Hold ICI; Methylprednisolone 0.5–1 mg/kg/day IV → Prednisone 1 mg/kg/day PO taper over 4 weeks | Resume ICI when grade ≤1; use ASCO/SITC algorithm |
| ICI-induced hepatitis grade 3–4 (ALT >5× ULN or bilirubin >3× ULN) | Permanently discontinue ICI; Methylprednisolone 1–2 mg/kg/day IV; if no improvement in 3–5 days add Mycophenolate mofetil 500–1000 mg PO BID | Avoid infliximab (hepatotoxic); consider tacrolimus for refractory cases |
| Idiopathic idiosyncratic DILI without AIH features | NO routine steroids | Not supported by trial data; risk of infection outweighs benefit |
Step 5 — Liver transplant evaluation
S5. Special Scenarios
Acetaminophen toxicity
Acetaminophen is the leading cause of acute liver failure in the US (~46% of ALF cases). Single-dose ingestion ≥7.5 g (or ≥150 mg/kg in children) produces significant hepatotoxicity in most adults; chronic supratherapeutic dosing (≥4 g/day with malnutrition or alcohol) can cause injury at much lower doses. Plot the 4-hour serum level on the Rumack-Matthew nomogram (treatment line: 150 μg/mL at 4 h, 75 μg/mL at 8 h, 37.5 μg/mL at 12 h). For staggered ingestions, unknown timing, or chronic toxicity, do not rely on the nomogram — treat empirically with full NAC course if ALT is elevated or any concern exists. The peak ALT typically occurs at 48–96 hours; a falling ALT with rising INR and bilirubin is the early warning of evolving ALF.
Herbal and dietary supplements (HDS)
HDS account for approximately 20% of cases in the US Drug-Induced Liver Injury Network (DILIN) registry — a fivefold increase over the past 15 years. The most common implicated agents are anabolic androgenic steroids (cholestatic, often prolonged), green tea extract (high-dose EGCG, hepatocellular), Garcinia cambogia (mixed), Hydroxycut formulations, kava (banned in EU), Ma huang (ephedra), He shou wu, comfrey (pyrrolizidine alkaloid VOD/SOS), and various multi-ingredient bodybuilding products. Always ask explicitly about ALL supplements, teas, weight-loss products, and bodybuilding aids — patients commonly do not consider these “drugs” and will not volunteer the history.
Immune checkpoint inhibitor (ICI) hepatitis
ICI hepatitis occurs in 3–9% of patients receiving anti-PD-1/PD-L1 monotherapy and 13–25% with combination CTLA-4 + PD-1 therapy (typically nivolumab + ipilimumab). Onset is usually 6–14 weeks after the first dose but can occur as early as 1 week or as late as 1 year. The injury pattern is most often hepatocellular but can be cholestatic or mixed. Liver biopsy shows lobular hepatitis with sinusoidal lymphocytic infiltrate (predominantly CD8+ T cells), endothelialitis, and granulomas — distinguishable from AIH (which has plasma cell-rich portal infiltrate and interface hepatitis). Management follows ASCO/SITC grading: grade 1 (ALT ≤3× ULN) — continue ICI with weekly monitoring; grade 2 — hold ICI + corticosteroids; grade 3–4 — permanent discontinuation + high-dose steroids ± MMF or tacrolimus for steroid-refractory cases.
Pregnancy
DILI in pregnancy is uncommon but the differential includes acute fatty liver of pregnancy (AFLP), HELLP syndrome, intrahepatic cholestasis of pregnancy (ICP), and viral hepatitis E (high mortality). Workup must include LFTs, CBC, coagulation, fibrinogen, glucose (AFLP causes hypoglycemia), bile acids (ICP >10 μmol/L), HEV IgM, and obstetric ultrasound. NAC is safe in pregnancy. Most other DILI treatments are reasonable except for valproate (replacement only when essential) and certain hepatotoxic chemotherapy.
Re-exposure rules
Once a probable or highly probable DILI has been attributed to a specific drug, lifelong avoidance is recommended in nearly all cases. Cross-reactivity within drug classes (e.g., other beta-lactams after amox-clav DILI) is incompletely defined; consider class avoidance only for severe DILI, anaphylactic features, or DRESS. For essential medications (e.g., antiepileptics, antitubercular drugs), structured reintroduction with close LFT monitoring under specialist supervision may be considered, but inadvertent rechallenge has caused fatal hepatic necrosis and is generally contraindicated.
Practice Insights: DILI — R-Ratio Pattern Recognition, Hy’s Law Severity Flag, and Empirical NAC for Non-APAP
(The content below synthesizes published clinical patterns and expert consensus rather than first-person case reports.)Transaminases that begin to fall in drug-induced liver injury can mean recovery or a failing liver, and only the pattern and the bilirubin separate the two: (1) R-ratio (ALT/ULN ÷ ALP/ULN) calculated at presentation classifies the injury pattern and guides differential diagnosis; (2) Hy’s Law (ALT > 3× ULN + bilirubin > 2× ULN + no other cause) flags severe DILI requiring close monitoring and transplant readiness; (3) IV NAC has been extended from acetaminophen overdose to severe non-APAP DILI per Mendizabal 2009 evidence. The “stop the drug and watch” approach has been refined by structured causality assessment and biomarker-guided escalation.
R-ratio classifies injury pattern + narrows differential
R-ratio = (ALT/ULN) ÷ (ALP/ULN) calculated at presentation. R > 5: hepatocellular pattern (acetaminophen, anti-TB drugs, NSAIDs, statins, antiepileptics, immune checkpoint inhibitors). R < 2: cholestatic pattern (amoxicillin-clavulanate most common, anabolic steroids, chlorpromazine, sulfonamides). R 2-5: mixed (carbamazepine, phenytoin, sulfasalazine). Pattern recognition narrows the differential dramatically — for cholestatic DILI experienced clinicians think amox-clav first, for hepatocellular they think acetaminophen + isoniazid first. Pattern can also evolve over course of injury — initial hepatocellular may transition to mixed or cholestatic with resolution.
Hy’s Law flags severe DILI with ~10% mortality risk
Hy’s Law: ALT > 3× ULN PLUS bilirubin > 2× ULN in the absence of alternative cause = severe DILI with approximately 10% mortality or transplant risk. This is a clinical pearl from Hyman Zimmerman that has held up across decades of pharmacovigilance. When Hy’s Law is positive, seasoned clinicians admit for monitoring, daily LFTs and INR, hepatology consultation, and consider early transfer to transplant center if INR rising or encephalopathy develops. ALT level alone (even > 1000) without bilirubin elevation has better prognosis. The combination is the mortality predictor.
RUCAM causality scoring + LiverTox database
RUCAM (Roussel Uclaf Causality Assessment Method) provides standardized causality assessment: 7 components scored (time to onset, course after withdrawal, risk factors, concomitant drugs, exclusion of non-drug causes, prior literature, response to rechallenge). Total score: ≥ 9 highly probable, 6-8 probable, 3-5 possible, 1-2 unlikely, ≤ 0 excluded. Experienced practitioners use RUCAM in any case with ≥ 2 suspect drugs or ambiguous clinical picture. The NIH LiverTox database (livertox.nih.gov) provides drug-specific likelihood scores (A-E) and characteristic patterns — invaluable reference for any DILI workup.
IV NAC for severe DILI — extended beyond acetaminophen
The Mendizabal 2009 trial demonstrated IV N-acetylcysteine improved transplant-free survival in non-acetaminophen acute liver failure with early-stage encephalopathy. Experienced clinicians extend NAC empirically to severe non-APAP DILI with progressive hepatic dysfunction (rising INR, encephalopathy, ALT trending up despite drug discontinuation): 150 mg/kg over 1 hour loading, then 50 mg/kg over 4 hours × 4 doses, then 100 mg/kg over 16 hours continuous. Cost and toxicity minimal; potential mortality benefit substantial. For pure acetaminophen DILI, NAC protocol is the same. Corticosteroids ONLY for immune-mediated DILI patterns (immune checkpoint inhibitor hepatitis, autoimmune-like DILI confirmed by biopsy) — not routine.
Anonymized Case: Amoxicillin-Clavulanate Cholestatic DILI with Hy’s Law
A man in his late fifties presented 4 weeks after completing a 10-day course of amoxicillin-clavulanate (875/125 mg BID for sinusitis) with progressive jaundice, pruritus, fatigue, and dark urine. No abdominal pain. No fever. Vital signs stable. Exam: scleral icterus, no hepatosplenomegaly, no peripheral edema, no asterixis. Labs: total bilirubin 8.5 mg/dL (predominantly direct), ALT 220 U/L (5.5× ULN), AST 180 U/L, ALP 480 U/L (4.0× ULN), GGT 580, albumin 3.8, INR 1.1, CBC normal. R-ratio: (220/40) / (480/120) = 5.5 / 4.0 = 1.4 — cholestatic pattern. Workup for alternative causes: HAV IgM negative, HBV surface antigen and core antibody negative, HCV antibody negative, HEV antibody negative, abdominal ultrasound with no biliary dilation or mass, autoimmune workup (ANA, ASMA, AMA) negative, IgG normal. Hy’s Law POSITIVE: ALT > 3× ULN + bilirubin > 2× ULN with no alternative cause. RUCAM score 8 (probable): typical time to onset 4 weeks after a known hepatotoxic drug, no concomitant drugs, no risk factors, exclusions negative, well-documented hepatotoxicity for amox-clav, +2 for resolution after withdrawal. LiverTox database confirmed amox-clav as one of the most common causes of cholestatic DILI (likelihood score A). Diagnosis: amox-clav cholestatic DILI. Treatment: supportive (no further amox-clav exposure, careful drug review for any other hepatotoxic exposures); given Hy’s Law positive, monitored as inpatient with daily LFTs and INR. Course: bilirubin peaked at 12.5 at week 2, ALT peaked at 280 at week 1, then both declined. INR remained stable at 1.0-1.2. No encephalopathy developed. Discharged hospital day 7. Resolution complete by week 8. Lifetime avoidance of amox-clav documented in chart.
Four Protocol Deviations Seasoned Clinicians Now Default To
- R-ratio (ALT/ULN ÷ ALP/ULN) calculated at presentation classifies pattern. R > 5 hepatocellular (APAP, INH, NSAIDs, statins, antiepileptics, ICI); R < 2 cholestatic (amox-clav most common, anabolic steroids, chlorpromazine); R 2-5 mixed. Pattern narrows differential.
- Hy’s Law (ALT > 3× ULN + bilirubin > 2× ULN + no alternative cause) triggers inpatient monitoring. ~10% mortality/transplant risk per Hyman Zimmerman observation. Daily LFTs + INR + early hepatology consult; transfer to transplant center if INR rising or encephalopathy.
- RUCAM standardized causality assessment + LiverTox database for drug-specific likelihood. RUCAM ≥ 9 highly probable, 6-8 probable, 3-5 possible. LiverTox (livertox.nih.gov) provides A-E likelihood for > 1000 drugs. Use RUCAM for cases with multiple suspects or ambiguous picture.
- IV NAC for severe non-APAP DILI per Mendizabal 2009 — empirical for progressive hepatic dysfunction. 150 mg/kg over 1 hr load → 50 mg/kg over 4 hr × 4 → 100 mg/kg over 16 hr continuous. Corticosteroids ONLY for immune-mediated DILI (ICI hepatitis, autoimmune-like) — not routine.
S6. Pearls, Pitfalls, and FAQ
Frequently Asked Questions
- Acetaminophen Hepatotoxicity: Rumack-Matthew Nomogram, N-Acetylcysteine Protocols, and Liver Transplant Criteria
- Liver Cirrhosis — Diagnosis and Complications Management
- Hemorrhoids: ACG/ASCRS Diagnosis, Goligher Grading, and Office vs Surgical Management
- Acute Liver Failure: Diagnosis, Etiology Workup, NAC, and King’s College Transplant Criteria
- Acute Diverticulitis: ACG Diagnosis, Antibiotic Selection, and Surgical Management
- Chronic Constipation: ACG/AGA Diagnosis, Rome IV Criteria, and Stepwise Medical Therapy
Primary guideline documents this article draws on. Links resolve via DOI to the publisher; access may require a subscription.
- ACG 2021 Clinical Guideline: Diagnosis and Management of Idiosyncratic DILI (American Journal of Gastroenterology · 2021)