Acute fatty liver with lactic acidosis
Microvesicular hepatic steatosis ± other tissue involvement
Didanosine, Fialuridine, Valproate
Acute hepatic necrosis
Collapse and necrosis of liver parenchyma
Plasma cells and interface hepatitis with detectable autoantibodies
Balloon hepatocytes with minimal inflammation
Balloon hepatocytes with inflammation, predominance of serum alkaline phosphate elevation (phenytoin, amoxicillin-clavulanic acid)
Hepatic collagenization with minimal inflammation
Micro- or macroscopic liver nodules
Azathioprine, Oxaliplatin, TDM1
Nonalcoholic fatty liver
Macro- and microsteatosis, hepatocyte ballooning and periportal inflammation
Sinusoidal obstruction syndrome
Inflammation with obliteration of central veins
Vanishing bile duct syndrome
Paucity of interlobular bile ducts
Diagnosis of DILI
Making a diagnosis of DILI is often very challenging due to varied presentations in varied clinical scenarios with reliable diagnostic tests. One of the commonly used instruments in the clinical trials and by regulatory authorities is the Roussel Uclaf Causality Assessment Method (RUCAM ). It is a structured means of assigning points for temporal association, clinical, biochemical, serologic and radiological features. The overall score ranges from −9 to +14 with a higher score associated with a higher likelihood of DILI from the implicated agent. In the real world, however, first and foremost in the diagnosis of DILI, is to entertain the possibility of DILI in any patient with unexplained liver injury. Clinical judgment, often by exclusion of competing etiology, is probably the most frequent method used to identify a case of DILI (Fig. 25.1). In judging the likelihood of DILI from an implicated agent, a process called causality assessment primarily based on six features are crucial for a definitive diagnosis.
Flow chart for evaluation of a case of apparent DILI. The diagnosis of DILI requires pertinent clinical information and adequate evaluation for exclusion of competing etiology
Latency: The duration between the exposure to the offending agent and the onset of liver injury is typically called the latency period and could be variable ranging from 5 days and up to 3 months. Drugs such as sulfonamides and macrolide antibiotics have short latency period of few days where as drugs such as INH, nitrofurantoin or amiodarone have long latency period of up to several months.
Dechallenge: Improvement in liver tests is usually observed within a few days (acetaminophen) or few weeks (macrolide antibiotics or sulfonamides) of discontinuation of the offending agent. Certain drugs, however, cause persistent elevations in liver tests and often associated with bile duct damage or chronic hepatitis.
Pattern of injury: Drugs often have a signature pattern in their clinical presentation. For example, anabolic steroids cause bland cholestasis, or amoxicillin-clavulanic acid (Augmentin) causes mixed pattern, or INH causes hepatocellular. A few drugs can cause immune-allergic hepatitis with features such a rash, facial edema, myalgia, arthralgia, eosinophilia and atypical lymphocytes. For example, allopurinol causes these symptoms with short latency. Telaprevir is another recent example of a drug that has been associated with Drug Rash Eosinophilia and Systemic Symptoms (DRESS).
Exclusion of competing etiology: A careful history and work up for exclusion of competing etiology is crucial as DILI is primarily a diagnosis of exclusion. For exclusion of acute hepatitis C, the routinely ordered viral hepatitis panel that include hepatitis C antibody is inadequate since it takes approximately 4 weeks to develop positive antibodies. A hepatitis C virologic assay is necessary to exclude acute hepatitis C. A liver biopsy is not required to make a diagnosis of DILI. However, rarely a liver biopsy should be considered if autoimmune hepatitis remains a competing etiology or when liver enzymes fail to improve with discontinuation of the offending drug.
Prior reports of the implicated agent: Some drugs are well known, well described, and well reported to cause DILI and have a characteristic signature with more than 50 cases including case series that have been described. Causality assessment of such cases in general is easier when adequate information is available. Alternatively, certain drugs despite extensive use have no evidence that they cause liver injury. Although single case reports may have been published, they are largely unconvincing. Implicating such an agent as the cause for DILI would raise concerns about concomitant drugs that could have been missed, or surreptitious use of herbal and dietary supplements often under the perception of being safe due to natural ingredients.
Re-challenge: Although included in several causality instruments, a premeditated re-challenge with the suspected drug is not recommended. However, inadvertent re-challenge followed by a repeat liver injury is very convincing for DILI related to that drug.
LiverTox: A Clinical and Research Information Resource on DILI
A significant development in this field is LiverTox (livertox.nih.gov), a novel online resource with content developed by Liver Disease Branch of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) with help from National Library of Medicine. It is in the public domain and freely available to use. It includes a complete and accurate summary of clinical features, diagnosis, cause, frequency, patterns, and management of DILI associated with prescription and nonprescription drugs such as HDS. LiverTox is a single-destination shortest path for easy accesses to up-to-date, accurate, and comprehensive information for the health care provider facing challenging cases of apparent DILI. The RUCAM worksheet is also available through the LiverTox website (www.livertox.nih.gov/rucam.html).
Drug-Induced Autoimmune Hepatitis
Certain drugs can cause autoimmune hepatitis with presence of autoantibodies, hyper gammaglobulinemia, and characteristic liver histology similar to de novo autoimmune hepatitis . Resolution of liver test abnormalities has also been reported with discontinuation of the offending drug but often in real world, steroids are initiated due to difficulty in differentiating from de novo autoimmune hepatitis (AIH). When liver tests improve with steroid therapy, consideration could be given for cessation of steroid therapy through a slow taper especially in drugs that have a definite association with AIH such as nitrofurantoin, minocycline, methylodopa, hydralazine, or halothane. Latency period with these agents is typically long and could be several months. For example, nitrofurantoin (Macrobid), characteristically causes a chronic hepatitis after many weeks, months or even years of therapy with some cases presenting with cirrhosis associated with serum antinuclear antibodies (ANA) .
Herbal and Dietary Supplement-Related Liver Injury
A dietary supplement is any product intended to supplement diet and may contain vitamins, minerals, herbs, amino acids or extracts thereof. Unlike most prescribed pharmaceuticals, attribution of DILI to herbal and dietary supplements (HDS ) is confounded by product variability, complexity, and contamination or adulteration . These factors notwithstanding, the problem of DILI attributable to HDS remains an important concern for Americans, as HDS are the second most common cause of DILI among enrollees into the DILIN. The widespread use of HDS, the current lack of regulation, and the uncertainty of the manufacturer or provider-initiated reporting of adverse events preclude an accurate estimation of the frequency of use and the scope of attributable hepatotoxicity. That said, approximately 15 % of the DILIN cohort have experienced liver injury caused by HDS . In the recent report from DILIN, liver injury caused by HDS increased from 7 to 20 % (P < 0.001) during the study period between 2004 and 2013 . Bodybuilding HDS caused prolonged jaundice (median, 91 days) in young men, but did not result in any fatalities or liver transplantation. Non-bodybuilding HDS cases presented more with hepatocellular injury, predominantly in middle-aged women, and, more frequently, led to death or transplantation, compared to injury from conventional medications (13 % vs. 3 %; P < 0.05) .
DILI in Patients with Disease State
Certain disease conditions may predispose individuals to DILI due to alterations in host immune function, drug metabolizing enzyme activity, lower serum albumin, glutathione reserve in the liver and overall ability to scavenge chemically reactive metabolites that can induce hepatotoxicity. One classic example is the increased incidence of trimethoprim–sulfamethoxazole hepatotoxicity (up to 20 %) in patients with acquired immune deficiency syndrome compared to the general population . Another, well-known example is the increased incidence of DILI from antituberculosis treatment in patients who have advanced disease and lower pretreatment serum albumin (<3.5 g/dl) . A recent study examined the association between on-treatment weight loss and risk of DILI from tuberculosis treatment and reported that weight loss of 2 kg or more within 4 weeks, concomitant hepatitis C, older age and multi-drug resistant TB were independently associated risk factors . Among these, the strongest risk factor was on-treatment weight loss (OR 211, 95 % CI: 36–1232) . Furthermore, patients with concomitant alcohol intake had threefold higher odds of developing hepatotoxicity . Alcohol consumption is one of the criteria in the RUCAM causality assessment instrument  although the risk of liver injury has only been reported in select medications such as INH, methotrexate or halothane . Chronic alcohol abuse might increase the hepatotoxicity of antituberculosis treatment through the induction of hepatic CYP2E1 [42, 43]. It has been suggested that alcoholism in conjunction with patient’s poor nutritional status may result in poor glutathione reserve and increased risk of DILI [17, 43].
Obesity, diabetes, and hepatic steatosis result in a chronic oxidant stress and mitochondrial dysfunction , which may enhance the toxicity of drugs that target mitochondria, e.g., acetaminophen . However, concern related to baseline liver test abnormalities from underlying nonalcoholic fatty liver disease (NAFLD) or chronic hepatitis C and risk of hepatotoxicity especially with drugs such as statins and thiazolidinediones (rosiglitazone and pioglitazone) did not really pan out and currently liver safety iof these agents in patients with underlying chronic liver disease is much less of an issue [46–50]. Minor fluctuations in aminotransferases upon initiating statin therapy are not uncommon but serious hepatotoxicity is quite rare and even when happens, it is almost universally reversible upon prompt recognition and withdrawal of the offending agent [46, 51]. Rare cases of statin-induced autoimmune hepatitis have been reported and in such cases immunosuppressive therapy may be necessary until sustained biochemical improved is achieved . Recommendations of the Liver Expert Panel to the National Lipid Association on Statin Safety states that current evidence supports the use of statins to treat hyperlipidemia in patients with nonalcoholic fatty liver disease and nonalcoholic steatohepatitis .
Treatment of DILI
The mainstay of treatment for DILI is prompt recognition, discontinuation, and avoiding reexposure to the offending agent. The recovery period is varied with relatively rapid improvement unless there is bile duct damage or drug-induced AIH. A low-fat diet is generally recommended in those with jaundice and anti-pruritic agents such a doxepin or hydroxyzine are prescribed to alleviate symptoms of itching. For some specific drugs, however, there are specific therapies available. These include N-actyl cysteine (NAC or mucomyst) for acetaminophen poisoning, intravenous l-carnitine for valproate toxicity, cholestyramine for leflunomide (Arava), and steroid or immunosuppressant therapy for drug-induced AIH. Close follow-up until a 50 % improvement in ALT or AlkP with subsequent follow-up to document normalization of liver test abnormalities is required to ensure that the patient has not progressed to chronic DILI. We generally recommend follow up until normalization of liver test abnormalities and then discharge from our clinic. Consideration should be given for referral of the patient with apparent DILI to one of the DILIN centers (dilin.dcri.duke.edu/).
Total Parenteral Nutrition-Related Liver Injury
Parenteral nutrition-associated liver disease (PNALD) is characterized by elevated liver tests, hepatic steatosis and progressive hepatic fibrosis that could evolve into a life-threatening disease with high mortality and morbidity . It is estimated that up 40 % of patients receiving long-term total parenteral nutrition (TPN) develop PNALD . Progression to cirrhosis and portal hypertension occurs rarely but appears to be more common in infants and neonates than in adults . Although the exact pathogenesis is unclear and probably multifactorial, there is accumulating evidence over the past few decades that the lipid component of TPN is the most probable cause of PNALD. Both animal and human research during the past decade has implicated phytosterols and ω-6 (n–6) fatty acids as the most likely components in the current lipid formulations as primary reasons for PNALD [56–58].
PNALD should be suspected in all patients receiving TPN in the hospital or at home. Pattern of injury is often mixed pattern or cholestatic and competing etiologies such as extra-hepatic causes of cholestasis and DILI should be excluded. In most cases, elevated liver tests are observed after the first 2 weeks of TPN, although some patients may develop abnormalities several months later [55, 59]. Histologic examinations reveal mild to moderate cholestasis with minimal inflammation or necrosis and no evidence of fat accumulation . With continued TPN and enteral fasting, severe cholestasis is associated with bile duct regeneration, portal inflammation, and fibrosis with cirrhosis developing within months [55, 60].
Vital information with regard to treatment aspect of PNALD comes from the pediatric world. Several nonrandomized human trials provide substantial data to the beneficial effects of ω-3-rich fatty acids in the treatment of PNALD [61–63]. Recent data suggest marked improvement in survival when PNALD is recognized at early stages and treated with lipid emulsions that contain fish oil rather than plant-based lipid emulsions [62–64]. Unfortunately, in the United States, only plant-based lipid emulsions are currently approved for use, with fish oil based lipid emulsions available only through compassionate-use program.