Drug-Induced and Toxic Liver Disease

More than 670 drugs are currently listed in the LiverTox database (with >1000 compounds responsible if HDS and industrial chemicals are included).

Nine of the top 10 causes of acute nonacetaminophen DILI in the U.S. DILIN are antimicrobial agents; the three leading agents are amox-clav, INH, and nitrofurantoin. Collectively, these 10 drugs represent more than one third of all DILI cases. The most frequent 25 agents (including HDS) that cause DILI account for about 50% of the 190 individual causes.

Although liver injury caused by industrial chemicals is much less of an issue today than in the past, due in part to improved occupational safety procedures, liver injury due to certain household agents continues to be reported. For example, iron toxicity from inadvertent ingestion remains an issue with young children. Exposures to other heavy metals (e.g., phosphorus, copper) are largely a problem in the developing world.

Among botanical hepatotoxins, mushroom poisoning is one of the most common and requires prompt treatment if ALF is to be avoided.

HDS that cause liver injury are being recognized increasingly and are now the third leading class of drugs that cause DILI in the United States.

The low overall incidence of DILI is out of proportion to its potential severity and clinical significance. The risk of death or need for LT from drug-induced ALF is about 10%, and hepatotoxicity in clinical trials represents one of the most frequent reasons that new drugs are withdrawn during drug development.

Most agents that cause liver injury do so in an unpredictable, dose-independent manner (so-called metabolic idiosyncrasy). Host factors that convey an increased risk for DILI have included female gender, older age, obesity, underlying chronic hepatitis B or C or human immunodeficiency virus (HIV) infection, and chronic alcoholism. More recent studies suggest that the pharmacologic properties of the drug (i.e., higher daily dose, high degree of hepatic metabolism, high degree of lipophilicity), the patient’s innate immune system, and the presence of certain genetic markers (human leukocyte antigen [HLA] and cytochrome P-450 [CYP] polymorphisms) are likely more important factors.

Clinical presentations of DILI range from asymptomatic elevations in serum levels of ALT to ALF. Some drugs are associated with characteristic clinical syndromes that can point the differential diagnosis in the direction of DILI and away from other causes of acute liver injury and CLD.

Diagnosing DILI can be extremely challenging because drug injury mimics all other forms of acute, chronic, benign, malignant, vascular, and granulomatous liver diseases, and there is no specific biomarker or pathognomonic histologic feature. The diagnosis, therefore, remains one of exclusion of other causes. Knowledge of the biochemical and clinicopathologic “signature” of known cases is essential when a drug suspected of causing liver injury is encountered. The extent of the evaluation is directly proportional to the ability to eliminate other more and less common etiologies.

Several causality assessment methods are available to help diagnose DILI. Roussel-Uclaf Causality Assessment Method (RUCAM) scoring is specifically designed to diagnose DILI but requires clinical expertise to calculate and interpret. Until a specific biomarker is found that can accurately diagnose DILI, the opinion of a clinician skilled in the management of DILI will remain the most important diagnostic method.

Identification of suspected acute DILI requires prompt discontinuation of the putative culprit. Typically, biochemical and clinical abnormalities should start to resolve soon after its discontinuation. In cases in which hypersensitivity or autoimmune features fail to improve, treatment with glucocorticoids may be necessary. NAC is generally reserved for an acute acetaminophen overdose when the patient presents within 16 hours of ingestion. NAC can be given for nonacetaminophen drug-induced ALF but appears to be of benefit only in adults with early-stage hepatic coma (see Chapter 2 ).

Although severe DILI is a relatively rare event, drugs are the most common cause of ALF in the United States and other Westernized countries, with acetaminophen implicated in 40% to 50% of the estimated 2000 cases that occur annually in the United States.

• ▪

  Other drugs (including HDS) are responsible for 11% to 12%, a percentage equal to that of ALF from acute viral hepatitis and greater than that seen with all other individually identified causes.

• ▪

  Drug-induced ALF accounts for approximately 15% of liver transplants performed for ALF. Acetaminophen accounts for most of these drug-induced cases, with all other idiosyncratic causes of drug-induced ALF accounting for a relatively small fraction.

Estimates of nonacetaminophen-related DILI range from 10 to 20 cases per 100,000 per year in Western nations. Among hospitalized patients, DILI is a relatively rare cause of jaundice.

Among hospitalized patients in Ireland with a serum ALT value >1000 U/L, ischemic liver injury was found to be the most common etiology (61%), and DILI was the next most common cause (16%), followed by acute viral hepatitis (12%), an undetermined cause (5%), and acute choledocholithiasis (4.4%). Acetaminophen toxicity accounted for almost one half of the DILI cases, with others caused by antituberculosis agents and other antibiotics.

The most common causes of DILI in the DILIN are listed in Table 10.1 . The 10 most frequent causes represent about 35% of the 190 individual agents in the registry. The top 25 agents account for nearly one half of all cases.

The most common causes of ALF and drug-induced ALF from the U.S. ALF Study Group are listed in Table 10.2 (see also Chapter 2 ).

• ▪

  The definition of drug-induced ALF includes all of the following:

  • –

    Encephalopathy: Any degree of mental alteration (e.g., day/night confusion, disorientation, sleepiness)

  • –

    Coagulopathy (international normalized ratio [INR] >1.5)

  • –

    Absence of preexisting cirrhosis

  • –

    Injury of <26 weeks in duration

• ▪

  Nonacetaminophen drug-induced ALF has a poor prognosis, with a mortality rate of up to 75% without LT.

Hepatotoxicity due to HDS is increasingly recognized and reported. The DILIN reports HDS as a cause in 16% of cases, second only to antimicrobial agents as the most common reason for all-cause DILI. Similarly, HDS are a significant cause of drug-induced ALF. The more widespread use of HDS has also been associated with an increase in safety alerts issued by the U.S. Food and Drug Administration (FDA), as well as foreign regulatory bodies, for several products—most recently, the weight loss and muscle building compounds Hydroxycut and OxyELITE Pro. Among patients with HDS liver injury reported in the DILIN between 2003 and 2013, bodybuilding HDS accounted for 35% of the total. Many of these were anabolic steroids, nitric oxide boosters, and slimming aids. In addition, most patients were taking more than one compound, making causality difficult to ascertain in some cases. More than 200 products have been implicated ( Table 10.3 ).

DILI may present as asymptomatic elevations of serum ALT and aspartate aminotransferase (AST) levels. The usual threshold for a clinically significant elevation is >3 times the upper limit of normal (ULN).

• ▪

  Depending on the specific laboratory, the ULN for serum ALT and AST can vary considerably.

• ▪

  The true normal values for serum ALT in healthy subjects without liver disease are <30 U/L for men and <19 U/L for women (see Chapter 1 ).

If an ALT elevation does not worsen, and the patient remains free of any associated hepatitis-related symptoms despite continuing the drug, the term “drug tolerance” is applied, implying hepatic adaption to subclinical injury.

• ▪

  A list of drugs and the percentage of patients in whom tolerance occurs is shown in Table 10.4 .

  TABLE 10.4

  Drugs Associated with Hepatic “Tolerance”

  Examples	Frequency of Drug Tolerance (%)
  Tacrine	>25
  Amiodarone, chlorpromazine, phenytoin, valproate	20–25
  Androgens, disulfiram, erythromycin estolate, isoniazid, leflunomide, ketoconazole	10–20
  NSAIDs (e.g., diclofenac)	5–10
  Statins, sulfonamides, sulfonylureas, tricyclic antidepressants	<5

  NSAIDs, Nonsteroidal antiinflammatory drugs.

• ▪

  It is important to note that in some cases, ALT levels may continue to rise and overt hepatic injury can occur. As a result, careful biochemical and clinical monitoring is needed.

Recognition of biochemical injury patterns (hepatocellular, cholestatic, mixed) and how the absolute values (fold-elevation above normal) and the specific ratios of ALT, AST, alkaline phosphatase (AP), and bilirubin correlate with the various forms of DILI remain essential to the diagnosis of DILI. Alcoholic liver disease (ALD), in particular, has a unique biochemical profile that can often help distinguish it from DILI.

• ▪

  R value to determine the biochemical injury pattern:

  • –

    This classification system uses the ratio of elevation above the ULN of ALT to AP to classify liver injury as hepatocellular, cholestatic, or a mixed pattern.

  • –

    First proposed as part of an International Consensus Meeting Criteria in 1990

  • –

    The serum bilirubin level is not a component of the pattern definition; if jaundice is present, the pattern is referred to as hepatocellular jaundice or cholestatic jaundice.

    $R\mathrm{value}=\mathrm{ALT}/\mathrm{ULN}÷\mathrm{AP}/\mathrm{ULN}:$

  • –

    Hepatocellular injury (R ≥5 and ALT >2× ULN)

  • –

    Cholestatic injury (R ≤2 and AP >2× ULN)

  • –

    Mixed injury (5 > R > 2 and ALT, AP both >2× ULN)

• ▪

  R values can help determine whether a specific drug is the cause of liver injury ( Table 10.5 ).

  • –

    Example: ALT 500 U/L (ULN = 40 U/L); AP 230 U/L (ULN 115 U/L)

    $\begin{array}{c}R\mathrm{value}=\left(\left[\mathrm{500}/\mathrm{40}\right]=\mathrm{12}.5\right)÷\left(\left[\mathrm{230}/\mathrm{115}\right]=2\right)\\ \mathrm{equals}6.\mathrm{25},\mathrm{signifying}a\mathrm{hepatocellular}\mathrm{pattern}\mathrm{of}\mathrm{injury}\end{array}$

  TABLE 10.5

  Causes of Drug-Induced Liver Injury According to Predominant Injury Patterns and R Value ^a

  Drug Class	Hepatocellular (R ≥5)	Acute Cholestatic (R ≤2)	Chronic Cholestatic/VBDS (R ≤2)	Granulomatous (R ≤2 or mixed)	AIH (May be Hepatocellular, Cholestatic, or Mixed)
  	Examples
  Anesthetics	Halothane
  Antibiotics	Sulfonamides Ketoconazole Dapsone	Erythromycin estolate Amoxicillin-clavulanic acid Flucloxacillin	Clindamycin Amoxicillin-clavulanic acid Thiabendazole SMX-TMP	—	Nitrofurantoin Minocycline
  Anti-TB drugs	INH PYZ Rifampin	—	—	—	—
  Anticonvulsants	Phenytoin VPA Carbamazepine	—	—	Phenytoin	—
  Antiinflammatories/analgesics	Diclofenac AAP	Sulindac	—	Allopurinol Gold salts	Diclofenac
  Psychotropics	TCAs	CPZ Haloperidol	CPZ Haloperidol Imipramine	Carbamazepine	—
  Miscellaneous causes	Disulfiram Labetalol PTU Nicotinic acid	Anabolic and contraceptive steroids Captopril TPN	Tolbutamide Ethinyl estradiol Terbinafine	Hydralazine Quinidine Mineral oil	Fenofibrate Infliximab Ipilimumab Methyldopa Procainamide Statins

  AAP, Acetaminophen; AIH, autoimmune hepatitis; CPZ, carbamazepine; INH, isoniazid; PTU, propylthiouracil; PYZ, pyrazinamide; SMX-TMP, sulfamethoxazole-trimethoprim; TCAs, tricyclic antidepressants; TB, tuberculosis; TPN, total parenteral nutrition; VPA, valproic acid; VBDS, vanishing bile duct syndrome.

  a R = ALT/ULN ÷ AP/ULN, where ALT is serum alanine aminotransferase level, AP is serum alkaline phosphatase level, and ULN is upper limit of normal.

The absolute values (as times elevations above ULN) of ALT and AST and the ratio of ALT to AST can be helpful in differentiating idiosyncratic DILI from other causes of liver injury. Mean peak ALT values for acute idiosyncratic DILI are generally much lower than those seen with acute viral hepatitis, ischemic hepatitis, and acute poisoning from toxic mushrooms, acetaminophen, or chemicals. Certain non-DILI etiologies are associated with an AST > ALT ( Table 10.6 ).

A serum gamma-glutamyltranspeptidase (GGTP) elevation can reflect inflammation or enzyme induction from multiple causes (GGTP is not considered a true liver injury–associated biochemical measure).

• ▪

  GGTP rises in parallel with AP in cholestatic injury.

• ▪

  Mild-to-moderate elevations of GGTP are seen in idiosyncratic DILI, nonalcoholic fatty liver disease (NAFLD), and other CLD.

• ▪

  GGTP is elevated out of proportion to AST and ALT in ALD (because of enzyme induction).

• ▪

  Isolated causes of a GGTP elevation are seen with anticonvulsants (e.g., phenytoin, barbiturates), alcohol ingestion, fatty liver, and heart failure.

Clinical syndromes associated with idiosyncratic DILI:

• ▪

  Many instances of mild DILI are asymptomatic with self-limited ALT and AST elevations.

• ▪

  Clinical features of severe DILI are often nonspecific and include nausea in 60%, abdominal pain in 40%, jaundice in 70%, and pruritus in 50% of cases.

• ▪

  Clinical syndromes associated with DILI are shown in Table 10.7 .

  TABLE 10.7

  Clinical Syndromes Associated with Acute Drug-Induced Liver Injury (DILI)

  Clinical Syndrome	Manifestations	Drug Cause(s)
  Acute viral hepatitis-like	Absence of hypersensitivity symptoms; presents with malaise, fatigue, anorexia, nausea, vomiting, RUQ pain	INH
  Acute hypersensitivity	Fever, rash, and/or eosinophilia, usually with short latency and prompt rechallenge response	Allopurinol, amoxicillin-clavulanic acid, carbamazepine, halothane, phenytoin, SMX-TMP
  Sulfone reaction	Fever, exfoliative dermatitis, lymphadenopathy, atypical lymphocytosis, eosinophilia, hemolytic anemia, methemoglobinemia	Dapsone
  Pseudomononucleosis	Hypersensitivity syndrome with atypical lymphocytosis, lymphadenopathy, and splenomegaly	Phenytoin, dapsone, sulfonamides
  DILI associated with severe skin injury	Stevens Johnson syndrome, toxic epidermal necrolysis	Phenytoin, carbamazepine, dapsone, nevirapine, allopurinol, SMX-TMP
  AIH	Fatigue, anorexia, lethargy, arthralgias, positive autoantibodies (antinuclear and smooth muscle/anti-actin antibodies)	Nitrofurantoin, minocycline, methyldopa
  Immune-mediated colitis with autoimmune hepatitis	Similar to AIH	Ipilimumab
  Acute cholecystitis-like	Biliary pain	Erythromycin estolate, ceftriaxone
  Reye syndrome−like	Hepatocellular injury, acidosis, hyperammonemia, encephalopathy, abdominal pain, nausea, vomiting, paradoxical worsening of seizure activity, microvesicular steatosis on biopsy	Valproic acid
  Budd-Chiari−like	Acute onset ascites and jaundice due to SOS	Myeloablative treatment for bone marrow, stem cell transplant
  Liver injury with atypical seizures	May present with worsening and more frequent seizures (including status epilepticus), and severe abdominal pain	Valproic acid
  Noncirrhotic portal hypertension	Variceal bleeding and/or ascites due to nodular regenerative hyperplasia	Azathioprine, 6-mercaptopurine

  AIH, Autoimmune hepatitis; INH, isoniazid; RUQ, right upper quadrant; SOS, sinusoidal obstruction syndrome; SMX-TMP, sulfamethoxazole-trimethoprim.

Clinicopathologic presentations of the most common causes of idiosyncratic DILI are shown in Table 10.8 .

Clinical presentations of common chemical hepatotoxins are shown in Table 10.9 .

Risk of DILI in patients with CLD and cirrhosis

• ▪

  10% of U.S. DILIN series had CLD (mostly chronic hepatitis C, NAFLD, or unexplained baseline liver biochemical test abnormalities).

• ▪

  No differences were seen in the classes or agents implicated in DILI; the one exception was azithromycin (which was nearly four times as likely to cause DILI in patients with than without CLD).

• ▪

  The severity was greater in patients with CLD, and mortality was significantly increased (16% versus 5.2%), than in patients without CLD.

• ▪

  Patients with CLD and DILI had a higher frequency of diabetes mellitus (38% versus 23%) as a possible risk factor for DILI than those without CLD.

• ▪

  Drugs that should be avoided in patients with cirrhosis are shown in Box 10.1 .

  BOX 10.1

  Drugs Contraindicated or Restricted for Use in Patients with Cirrhosis Because of an Increased Risk of Drug-Induced Injury

  • Agomelatine

  • Antituberculosis drugs

  • Azithromycin

  • Ketoconazole

  • Methotrexate (especially with leflunomide)

  • Pyrazinamide

  • Telithromycin

  • Tolvaptan

  • Trovafloxacin

  • Valproic acid

Effect of older age on DILI

• ▪

  16.6% of U.S. DILIN patients are >65 years of age.

• ▪

  Older patients have more cholestatic DILI (36% versus 21%) than younger patients; however, they have no greater need for LT, nor are they more likely to have a fatal outcome than younger patients.

• ▪

  The risk of overt INH injury is especially age dependent (frequency of <0.5% in those age <20 years; 0.5% age 20 to 35; 1% to 2% age 35 to 50; 3% age >50 in the United States).

Other risk factors

• ▪

  Gender: Women are more prone to develop acute and chronic DILI, as well as drug-induced ALF, possibly because of higher rates of exposure to DILI-causing drugs (e.g., sulindac, diclofenac, nitrofurantoin, azathioprine, leflunomide), but for unclear reasons with other agents (e.g., halothane, INH). Men are at higher risk of DILI caused by amox-clav than women.

• ▪

  HLA polymorphisms: Several drugs have been shown to have a possible genetic basis for causing DILI ( Table 10.10 ).

  TABLE 10.10

  Human Leukocyte Antigen Alleles That Predict Drug-Induced Liver Injury

  Drug	Risk Allele	Odds Ratio
  Class I Alleles
  Amoxicillin-clavulanic acid	A∗02:01 B∗18:01	2.2 2.8
  Flucloxacillin	B∗57:01	80.6
  Ticlopidine	A∗33:03	13
  Class II Alleles
  Antituberculosis agents (isoniazid, rifampin, pyrazinamide)	DQB1∗02:01 DQA1∗01:02	1.9 0.2 a
  Amoxicillin-clavulanic acid	DRB1∗15:01 DQB1∗06:02 DRB1∗07	2.3–10 0.18 a
  Flucloxacillin	DRB1∗07:01–DQB1∗03:03 DRB1∗15	7 a
  Lapatinib	DRB1∗07:01–DQA1∗02:01	2.6–9
  Lumiracoxib	DRB1∗15:01–DQB1∗06:02– DRB5∗01:01–DQA1∗01:02	5
  Nevirapine	DRB1∗01:02	4.72
  Ximelagatran	DRB1∗07–DQA1∗02	4.4

  a Risk decreased.

Latency periods

• ▪

  Short onset (days to few weeks) with immunoallergic DILI (one third of drugs) and more prolonged (1 to 9 months) with metabolic idiosyncrasy

• ▪

  Most DILI associated with ultrashort latency periods (<7 days) involves an antibiotic or occurs after a deliberate (or unintentional) rechallenge with an agent known to cause DILI through a hypersensitivity (immunoallergic) mechanism.

• ▪

  A few drugs have a delayed onset (up to 6 weeks) after the drug has been stopped (e.g., amox-clav, cefazolin).

• ▪

  It is quite rare for acute DILI to occur after a drug is taken >12 months with no sign of liver dysfunction during that time.

• ▪

  Prolonged latency (>1 year) is seen with drugs that cause a form of chronic autoimmune-like hepatitis (e.g., nitrofurantoin, minocycline, methyldopa, statins), often associated with positive antinuclear or anti-actin antibodies.

Causality assessment of DILI ( Box 10.2 )