Drug-Induced Liver Injury: Causes, Diagnosis, and What to Do

What Is Drug-Induced Liver Injury?

Drug-induced liver injury is hepatocellular, cholestatic, or mixed liver damage caused by a medication, herbal product, or dietary supplement rather than by an intrinsic liver disease. The US Drug-Induced Liver Injury Network (DILIN), a prospective multicenter cohort, enrolled 1,257 patients between 2004 and 2012 and found that antibiotics accounted for 45.5% of all DILI cases, with isoniazid, nitrofurantoin, and amoxicillin-clavulanate leading the list [1].

Two mechanistic categories matter clinically. Intrinsic or dose-dependent DILI is predictable and reproducible. Acetaminophen is the clearest example: doses exceeding 7.5 to 10 grams per day in healthy adults reliably cause centrilobular necrosis, while chronic alcohol use lowers that threshold to roughly 3 to 4 grams per day [2]. Idiosyncratic DILI, by contrast, occurs in a small fraction of exposed individuals regardless of dose, often reflects immune-mediated or metabolic susceptibility, and is far harder to predict before it happens.

The liver's first-pass metabolism of orally ingested drugs means it absorbs the highest concentration of reactive metabolites. Cytochrome P450 enzymes, particularly CYP2E1 and CYP3A4, generate oxidative intermediates that deplete glutathione, trigger mitochondrial dysfunction, and ultimately initiate hepatocyte apoptosis or necrosis. Polymorphisms in these enzymes partially explain why one patient tolerates a drug for years while another develops fulminant hepatic failure within weeks [3].

Which Drugs Cause DILI Most Often?

The ten agents responsible for the greatest burden of serious DILI cases are acetaminophen, isoniazid, nitrofurantoin, amoxicillin-clavulanate, diclofenac, phenytoin, minocycline, statins (as a class), anabolic steroids, and herbal and dietary supplements as a combined category. Supplements deserve special attention. A 2014 analysis from the DILIN cohort found that herbal and dietary supplements caused 20% of DILI cases, up from 7% a decade earlier, and that body-building supplements containing synthetic steroids carried a particularly severe clinical course [4].

Statins generate disproportionate concern among patients because transaminase elevations are measurable, but the clinical reality is reassuring in most cases. Symptomatic statin-induced DILI severe enough to require drug discontinuation occurs in roughly 1 per 10,000 to 100,000 exposed patients per year [5]. The American College of Cardiology and American Heart Association guidelines explicitly state that routine liver-enzyme monitoring in asymptomatic statin users is not indicated. Stopping a statin because of an asymptomatic ALT elevation below 3× ULN does more cardiovascular harm than hepatic good.

Methotrexate occupies a different position. Used long-term for rheumatoid arthritis or psoriasis, it causes hepatic fibrosis in a dose-cumulative manner. The American College of Rheumatology recommends baseline liver biopsy consideration when total cumulative methotrexate dose approaches 3.5 to 4 grams, especially in patients with pre-existing steatosis.

Amiodarone is particularly insidious. Its elimination half-life can exceed 100 days, meaning liver injury may appear months after dose reduction and persist long after the drug is stopped. Phospholipidosis and a pseudo-alcoholic hepatitis pattern on biopsy are characteristic findings.

How Does Pre-Existing MASLD or MASH Change DILI Risk?

Metabolic dysfunction-associated steatotic liver disease (MASLD, formerly called NAFLD) and its inflammatory form, metabolic dysfunction-associated steatohepatitis (MASH, formerly NASH), create a primed hepatic environment that amplifies idiosyncratic DILI risk. The renaming from NAFLD to MASLD by the international consensus group in 2023 better reflects the underlying metabolic drivers: obesity, type 2 diabetes, dyslipidemia, and hypertension [6].

In patients with MASLD, baseline oxidative stress is already elevated, glutathione reserves are partially depleted, and mitochondrial respiratory chain function is impaired. Experimental data from a 2020 paper in Hepatology showed that mice with diet-induced steatohepatitis exhibited roughly twice the degree of acetaminophen-induced hepatocyte necrosis compared to lean controls at equivalent doses. Human registry data from the DILIN cohort support a similar pattern: patients with diabetes, a common MASLD comorbidity, were significantly more likely to develop severe DILI and had longer time to recovery [1].

MASH itself can progress to cirrhosis in 10 to 20% of affected individuals over 20 years without pharmacologic intervention [7]. If DILI occurs on top of compensated cirrhosis, the functional hepatic reserve is already compromised, and the clinical trajectory can shift rapidly from a manageable enzyme elevation to liver decompensation. Clinicians prescribing hepatotoxic drugs to patients with known MASLD or MASH should set ALT monitoring intervals at 4 to 6 weeks for the first 6 months rather than relying on annual checks.

The HealthRX Hepatic Risk Stratification Framework categorizes patients into three tiers before initiating potentially hepatotoxic therapy:

Tier 1 (Standard monitoring): No steatosis on imaging, ALT below 1× ULN, BMI <30, no diabetes. Baseline LFTs then annual review.

Tier 2 (Enhanced monitoring): MASLD confirmed by ultrasound or FIB-4 score 1.3 to 2.67, ALT 1 to 3× ULN, or BMI 30 to 40 with metabolic syndrome. LFTs at 4 weeks, 8 weeks, then every 3 months.

Tier 3 (High-risk, consider alternatives): Established MASH, FIB-4 >2.67, prior DILI episode, or known cirrhosis. If the drug has no safer substitute, LFTs monthly for 6 months with a prespecified stopping rule of ALT >3× ULN or any bilirubin rise above 1.5× ULN.

How Is DILI Diagnosed?

DILI has no single definitive biomarker. Diagnosis is one of exclusion combined with a causal assessment tool. The Roussel Uclaf Causality Assessment Method (RUCAM), validated across multiple international registries, scores seven domains: time to onset, course after stopping the drug, risk factors, concomitant drugs, exclusion of other causes, prior DILI information, and response to rechallenge. A RUCAM score of 6 to 8 indicates "probable" DILI; above 8 indicates "highly probable" [8].

Biochemical patterns guide the initial differential. The R-ratio, calculated as (ALT / ULN) divided by (alkaline phosphatase / ULN), separates hepatocellular injury (R >5) from cholestatic injury (R <2) from mixed injury (R 2 to 5). Hepatocellular DILI carries a higher risk of acute liver failure. Cholestatic DILI, seen classically with amoxicillin-clavulanate or chlorpromazine, tends to resolve more slowly but rarely progresses to fulminant failure.

Hy's Law, named after Hyman Zimmerman, remains the most cited prognostic criterion. The FDA's 2009 guidance document on DILI states: "Hy's Law cases... have an approximate 10% risk of fatal outcome or need for liver transplantation." [9] The two criteria are ALT or AST >3× ULN plus total bilirubin >2× ULN, after ruling out biliary obstruction and Gilbert syndrome. Any drug producing Hy's Law cases in a Phase II or III clinical trial is a serious candidate for withdrawal or black-box warning.

Liver biopsy is not routine in DILI but is indicated when autoimmune hepatitis cannot be excluded serologically, when the diagnosis is uncertain after 3 months of observation, or when the patient has underlying cirrhosis and a new hepatic decompensation occurs. Histologic patterns such as zone 3 necrosis, microvesicular steatosis, or granulomatous inflammation can narrow the causative drug class.

What Are the Symptoms of Drug-Induced Liver Injury?

Many DILI cases are asymptomatic and detected only through routine laboratory surveillance. When symptoms do appear, the most common are fatigue, nausea, right upper quadrant discomfort, anorexia, and jaundice. Dark urine and pale stools signal cholestatic involvement and a rising bilirubin.

Severe cases presenting with jaundice, encephalopathy, coagulopathy (INR >1.5), and ascites meet criteria for acute liver failure and require urgent transfer to a transplant center. The DILIN prospective registry reported that 17% of enrolled DILI cases were classified as severe, 3.5% required liver transplantation, and 2.5% died [1].

Fever, rash, and eosinophilia suggest a hypersensitivity-mediated DILI, most often associated with anticonvulsants (phenytoin, carbamazepine), sulfonamides, and allopurinol. This phenotype responds to corticosteroids in some cases, though randomized trial data on steroid use in DILI remain sparse.

Symptoms of MASH, when present, overlap with DILI and include similar fatigue and right-sided discomfort. A patient on a hepatotoxic drug who also has MASLD and reports new-onset fatigue deserves urgent LFT measurement rather than reassurance.

How Is Drug-Induced Liver Injury Treated?

The primary treatment is stopping the offending agent. This sounds straightforward, but in practice it requires the prescribing clinician to weigh the benefit of continuing the drug against the risk of progressive liver injury. For isoniazid-induced DILI, for example, stopping the drug in active tuberculosis requires substituting effective antitubercular agents immediately to prevent disease reactivation.

N-acetylcysteine (NAC) is the only pharmacologic antidote with strong evidence, and it is specific to acetaminophen overdose. The Rumack-Matthew nomogram guides NAC initiation based on serum acetaminophen level and time since ingestion. NAC given within 8 to 10 hours of ingestion reduces mortality from acetaminophen-induced acute liver failure from roughly 58% to under 5% [10]. For non-acetaminophen DILI, a randomized controlled trial published in Gastroenterology in 2014 (N=173) found NAC did not improve transplant-free survival in adults with acute liver failure from other causes, though it showed a signal of benefit in patients with early-stage encephalopathy [11].

Ursodeoxycholic acid (UDCA) at 13 to 15 mg/kg/day is used by many hepatologists for cholestatic DILI, particularly in cases driven by amoxicillin-clavulanate or chlorpromazine, though the evidence base consists largely of case series rather than randomized trials. Corticosteroids are reserved for hypersensitivity phenotypes with clear immune features.

Patients with established MASLD or MASH who sustain DILI face an additional consideration: the resmetirom question. Resmetirom (Rezdiffra), a thyroid hormone receptor-beta agonist, received FDA approval in March 2024 for MASH with moderate to advanced fibrosis (F2 to F3). The MAESTRO-NASH trial (N=966) demonstrated that 80 mg resmetirom produced MASH resolution without fibrosis worsening in 29.9% of patients versus 9.7% on placebo at 52 weeks (P<0.001) [12]. Patients starting resmetirom who are also on other medications should have baseline LFTs confirmed normal before initiating, given that resmetirom itself carries a small risk of transaminase elevation.

Can DILI Progress to Cirrhosis?

Most DILI cases do not progress to cirrhosis. The natural history data from DILIN show that approximately 17% of cases develop a "chronic" course defined as LFT abnormalities persisting beyond 6 months, and a subset of those progress to fibrosis [1]. Drugs most associated with chronic DILI and fibrotic progression include methotrexate, amiodarone, and nitrofurantoin used for long-term urinary tract infection prophylaxis.

Cirrhosis from DILI is clinically indistinguishable from cirrhosis of other causes once established. Complications include portal hypertension, esophageal varices, hepatic encephalopathy, and hepatocellular carcinoma. The Baveno VII consensus (2022) recommends non-invasive liver stiffness measurement by transient elastography for all patients with suspected fibrosis, with a threshold of 10 kPa indicating clinically significant fibrosis and 15 kPa indicating probable cirrhosis regardless of etiology [13].

Patients with cirrhosis from any cause, including DILI, need twice-yearly liver ultrasound with alpha-fetoprotein measurement for hepatocellular carcinoma surveillance, per American Association for the Study of Liver Diseases (AASLD) guidelines updated in 2023.

When to See a Doctor and What Labs to Request

Any patient who develops jaundice, dark urine, severe fatigue, or right upper quadrant pain while on a new medication should have same-day liver function testing. The minimum panel includes ALT, AST, alkaline phosphatase, total bilirubin, and INR. If ALT exceeds 3× ULN or total bilirubin exceeds 2× ULN, the prescribing clinician should be contacted within 24 hours for a reassessment of ongoing drug use.

For patients with known MASLD or MASH who are starting a new potentially hepatotoxic drug, baseline labs should be obtained before the first dose. A FIB-4 index (calculated from age, ALT, AST, and platelet count) above 1.3 identifies patients at elevated fibrosis risk who need closer monitoring. The FIB-4 formula is: (age × AST) / (platelets × √ALT). Online calculators are available through the AASLD and the NIH.

Suspected DILI should be reported to the FDA through the MedWatch program at fda.gov/safety/medwatch. Reporting does not establish causality but contributes to post-market safety signal detection. The DILIN registry also accepts referrals for prospective evaluation of ambiguous cases through the NIH.