Drug-induced hepatitis

Author: Dr Tim Davern California Pacific Medical Center 2009-05-14

What is drug-induced hepatitis?

Drug-induced hepatitis, also known as hepatoxicity, refers to liver injury caused by drugs or other chemical agents and represents a special type of adverse reaction. The condition is called many things, but the one that experts currently favor is drug-induced liver injury (DILI); this is the term that will be used throughout this article. In addition, the generic term “drugs” shall be used to encompass the entire range of potentially injurious agents, also called hepatoxins.
Severe DILI is a life-threatening illness, but many patients who experience mild DILI are completely asymptomatic and the disease usually causes no significant liver problems. When diagnosed, the standard treatment involves quickly stopping and avoiding the implicated drug.  For mild to moderate drug-induced liver injury this is usually sufficient to resolve all signs and symptoms of a liver injury.
The problem is that to date there is a growing list of nearly a thousand prescription and over-the-counter drugs suspected of causing liver injury, which can be severe, even life threatening. Herbal and other complementary and alternative medications – as well as illicit drugs such as anabolic steroids and amphetamines – have also been implicated.     
DILI can be categorized in several ways based on its clinical features, the pattern of associated liver biochemistries - often  called "liver function tests" or "LFTs" [see table below], the changes observed under a microscope during a liver biopsy, and by the suspected pathogenic mechanisms underlying the development of the DILI event. 

Table: What are liver function tests (LFTs)?

• Aminotransferases -> includes alanine aminotransferase (ALT) and aspartate aminotransferase (AST); enzymes released from liver cells (heptocytes) in response to liver injury/inflammation. 
• Alkaline phosphatase (AP) –> an enzyme synthesized and released from injured bile duct cells (biliary epithelia)
• Albumin –> a major serum protein produced by the liver; with liver damage and dysfunction, the serum concentration of albumin falls.
• Prothrombin time (PT) –> a measure of blood clotting that relies on several proteins (clotting factors) made by the liver; with liver damage and dysfunction, the concentration of these proteins decreases, blood takes longer to clot and the PT gets longer.  (The international ratio (INR) is a standardized measurement of the PT with generally less lab-to-lab variation.)

One broad classification of DILI separates drugs into “predictable” and “unpredictable” hepatotoxins.  Predictable hepatoxins, such as acetaminophen (Tylenol®), cause dose-dependent liver injury.  That is, if any given individual takes enough acetaminophen they will develop severe liver injury, possible even acute liver failure [see ALF knol].  Indeed, acetaminophen poisoning is now the single leading cause of life-threatening acute liver failure in the United States and Europe.      Unlike acetaminophen, most of the other drugs that cause DILI do so in an unpredictable or so-called idiosyncratic fashion.  The word “idiosyncrasy” comes from Greek (idios "one's own" and sun-krasis "mixture”) and refers to traits peculiar to an individual.  A drug that causes idiosyncratic liver injury does so in relatively rare individuals due to genetic and environmental characteristics unique to those individuals.  In contrast to liver injury from predictable hepatoxins, idiosyncratic reactions tend to be difficult if not impossible to predict and are usually not dose-dependent. 
DILI is also classified by the pattern of liver injury observed.  There are two main patterns. Acute “hepatocellular injury” is caused by injury primarily to hepatocytes and characterized by elevated levels of serum alanine aminotransferase [see Table above], with minimal elevations of the serum alkaline phosphate. Severe hepatocellular injury may evolve into acute liver failure with hepatic synthetic dysfunction (characterized by an increased time needed for blood clotting [prothrombin time]) and brain dysfunction(hepatic encephalopathy) [see hepatic encephalopathy knol], which carries a very poor prognosis and often requires liver transplant for survival.
The second pattern, “cholestatic injury,” is caused by liver injury affecting either the cells lining the bile ducts or the molecular mechanisms involved in the flow of the pigment bilirubin from the blood into the bile. It is characterized by a disproportionately elevated level of alkaline phosphatase, which is an enzyme synthesized and released by injured bile ducts.  Severe cholestatic liver injury, while characterized by troublesome and often lingering, severe itching (pruritus) and jaundice, rarely results in death. Occasionally, however, it can evolve into permanent bile duct injury with so-called "vanishing bile duct syndrome”.
Liver injury that has both hepatocellular and cholestatic features is called “mixed liver injury.”  Of note, serum bilirubin may be elevated in all three forms of liver injury – hepatocellular, cholestatic and mixed – and, if so, signifies more severe liver injury. 
In the context of clinical trials, detection of milder liver injury which may herald a problem for the drug, is conventionally defined as an alanine aminotransferase (ALT) level of more than three times the upper limit of the normal range, a serum alkaline phosphatase (ALP) level of more than twice the upper limit of normal, or a total bilirubin (TB) level of more than twice the upper limit of normal if associated with any elevation of either the ALT or ALP.

Why is DILI important?

The problem of DILI is clearly important to patients, physicians, and the pharmaceutical industry; its importance will likely increase as more drugs are used to prevent and treat disease in the future. 

• DILI is obviously important to the individuals affected and their families because severe DILI is a catastrophic, life-threatening illness.  Another often unappreciated consequence of DILI for patients is that it may also prevent otherwise promising drugs from entering the marketplace, or result in their withdrawal from the market (see below), thus affecting the population as a whole.
• DILI is important to the physician because it represents an unintended, but often severe side effect of therapy that flies in the face of our creed – Primum non nocere  ("first do no harm”).  On a more mundane level, DILI also is important for physicians because affected patients, their families, and their lawyers may consider DILI an egregious mistake, a personal assault and betrayal, and DILI may thus have serious legal consequences. 
• DILI is very important to the pharmaceutical industry because it is a leading cause of drug withdrawal from the marketplace. Of the 548 drugs approved by the Food and Drug Administration between 1975 and 1999, 10 received "black box" warnings for their potential to cause liver injury.  Four additional drugs were withdrawn from the market.  As discussed below, DILI is often not fully recognized until a drug enters the market and many thousands of people are exposed to the drug, meaning that the substantial costs of drug development, preclinical and clinical testing, and marketing can be jeopardized when a drug is associated with DILI.  These substantial financial losses combined with the associated liability risk can be financially devastating for the pharmaceutical companies involved.

What causes DILI?

In most cases, the development of DILI is poorly understood, though given the way it functions in the body, it’s not surprising that the liver is subject to injury by drugs. Virtually all of the major intra-abdominal organs drain into the liver, so it is exposed to an unusually high concentration of drugs absorbed from the intestine.  The liver also extensively metabolizes most of these drugs and indeed, in most cases, liver injury appears to be initiated by reactive substances (metabolites) that are generated by processing (metabolism) of the drugs in the liver rather than by the drugs themselves. These metabolites can bind to and injure various critical components within liver cells and begin a series of events culminating in liver injury. In short, the liver may produce the very compounds that damage it.  Liver injury from the common analgesic acetaminophen is a classic example that is discussed further below. 
Given these facts, the real question probably should not be why drugs injure the liver, but rather why such injury is not observed more frequently.  The answer probably reflects the liver’s impressive array of mechanisms to protect, repair and regenerate itself.  These protective mechanisms probably evolved long ago and enabled our ancestors to survive the numerous threats of prehistoric life, including food borne botanical toxins. 
Deciphering the mechanisms underlying idiosyncratic DILI is challenging in part because most idiosyncratic reactions cannot be reproduced in animal models. In contrast, acetaminophen poisoning seems to be quite similar in humans and mice, and experiments in mice have provided important mechanistic insights into liver injury from acetaminophen in man. When acetaminophen is ingested, a small fraction is metabolized in the liver to a reactive metabolite (N-acetyl-p-benzoquinoneimine (NAPQI) that is normally rapidly bound to and detoxified by an abundant intercellular compound called glutathione.  However, with excessive ingestion of acetaminophen (typically > 10 grams at a single point in time), the production of the reactive metabolite overwhelms and rapidly depletes the available glutathione.  With glutathione depleted in liver cells, NAPQI is left free to bind to critical intercellular proteins, possibly leading to liver cell injury and death.  Though it was once thought this was the whole story, it is now very clear from a series of elegant experiments in various animal models that the innate immune system and several cellular signaling pathways also play important roles in modulating acetaminophen toxicity.  For example, experimental depletion of natural killer cells, critical to innate immunity, results in significant lessening of acetaminophen toxicity in mice.  Likewise, interruption of signaling by interferon gamma and the transcription factor, jun N-terminal kinase (JNK), also markedly attenuate acetaminophen toxicity in mice.  These experiments clearly demonstrate that even liver injury from acetaminophen toxicity is far more complicated than we had previously appreciated. 
Compared with liver injury from acetaminophen poisoning, much less is known about the development of idiosyncratic DILI, in large part because the reactions cannot be reproduced in animal models. We do know that like acetaminophen, many drugs that cause idiosyncratic DILI are metabolized in the liver into reactive compounds. These may bind to cellular components disrupting their function and initiating a cascade of further liver injury by acting as immune targets and triggering cell death pathways.  Indeed, programmed cell death (apoptosis) appears to be a major mechanism of liver injury from at least some drugs.  Some idiosyncratic drug reactions appear to be primarily immune mediated based on commonly associated clinical features such as rash, the presence of anti-drug antibodies, and the rapid onset of liver injury with rechallenge of the drug.  In addition, some of these reactions have been associated with specific human leukocyte antigens (HLA types) that have also been associated with autoimmunity.  These associations aside, however, many of the underlying genetic and environmental factors that account for idiosyncratic DILI have remained elusive.

Can DILI be prevented?

Several strategies can potentially be used to prevent severe liver injury from drugs:

• Patient education:  Patients taking drugs associated with liver injury should be warned about the symptoms commonly associated with severe liver injury, and should be instructed to stop the medications immediately upon developing these symptoms and to contact the prescribing physician.  Patients should also be educated regarding the possible interaction of other drugs and alcohol with potentially toxic drugs.  Patients who suffer significant DILI should be warned to avoid re-exposure to the implicated drug and, if appropriate, the reaction should be noted in the medical record as an allergy.  
• Liver test screening: Periodic screening of liver biochemistries, particularly serum ALT, is recommended for many drugs that have been associated with liver injury.  However, whether serum ALT monitoring during drug treatment effectively prevents severe DILI is controversial.  For example, some patients with acute liver failure from troglitazone (an anti-diabetic drug pulled from the US market in 2000) developed this life threatening complication despite undergoing the recommended monthly monitoring.  Furthermore, ALT monitoring on a frequent basis is costly and cumbersome for both patients and clinicians; consequently, compliance with such surveillance is poor even with drugs where formal recommendations for surveillance exist.  Adding to this problem, the significance of a mildly elevated serum ALT is not always clear, although this situation often causes great angst among both clinicians and patients. Indeed, anxiety caused by the detection of abnormal, but asymptomatic, elevated aminotransferases may ultimately result in inappropriate drug withdrawal in some patients. 
• Legislation: Another way of decreasing the problem of DILI is through effective legislation aimed at restricting or eliminating potentially toxic drugs from the marketplace.  Consider, for example, acetaminophen-opiate drugs. Nearly half of all cases of acute liver failure in the US are due to acetaminophen toxicity.  Many of these cases are from intentional acetaminophen overdose in suicidal patients, but a recent, multi-center US study suggests that as many as 50% of patients accidentally overdose with acetaminophen in an effort to relieve pain.  Many of these patients are prescribed acetaminophen-opiate combinations such Vicodin®, a combination of hydrocodone (an opiate) with acetaminophen.  Over time, many such patients appear to become tolerant to the analgesic effect of the opiate and thus take more than prescribed in order to achieve pain relief.  Eventually, some inadvertently consume so much acetaminophen that severe liver injury and even liver failure can ensue.  This problem is compounded by the presence of acetaminophen in many different cold and flu products. To the author, the concept of combining a highly addictive drug (an opiate) and a dose dependent liver toxin (acetaminophen) into a single tablet defies logic and is analogous to mixing “candy and poison.”  There is no reason, other than convenience, that these drugs are combined together in a single pill; they can be taken as separate tablets, an opiate (e.g., hydrocodone) and acetaminophen, with the same analgesic affect.  In the author’s opinion, patients should not be prescribed acetaminophen-opiate combinations for chronic pain and all patients taking these drugs should be warned about the potential for severe liver injury. Although patient and physician education is important, the most effective solution will probably be a legislative action on the part of the FDA to uncouple acetaminophen and other drugs, including opiates as well as cold remedies.  Such decisive legislation would help save many of the lives lost each year in the US from unintentional acetaminophen overdoses.   

What are the symptoms of DILI?

Many patients who experience mild DILI are completely asymptomatic and are only diagnosed because they have liver tests performed for other reasons.  Indeed, DILI is an important cause of abnormal liver biochemistries. 
When symptoms of DILI are present, usually in cases of more severe hepatocellular DILI, they are often similar to viral hepatitis and include malaise, anorexia, nausea and vomiting, right upper quadrant abdominal pain, jaundice, acholic (light or clay colored) stools, and dark (tea colored) urine.  Patients with cholestatic DILI may also have intense pruritis (itching).  Fever and rash are the hallmarks of hypersensitivity that may be present with DILI from some drugs (e.g., anticonvulsants such as phenytoin, sulfa drugs such as sulfamethoxazole-trimethoprim).  If a patient with severe hepatocellular DILI develops hepatic encephalopathy – a hallmark of acute liver failure that is characterized in its severe form by severe memory loss, confusion, and even coma – the prognosis is grim without a liver transplant [see knols on ALF and hepatic encephalopathy].

What are the risk factors for DILI?

For most drugs, adults are usually at higher risk for liver injury than children; a notable exception is the anti-seizure medication valproate, which appears to cause liver injury in children more commonly than adults.  Women as a group may also be somewhat more susceptible to DILI from many drugs than men, although this may also reflect more frequent drug exposure in woman, or their generally smaller size.  For some drugs, including acetaminophen, alcohol abuse and malnutrition appear to predispose to liver injury. 
Whether preexisting liver disease predisposes to DILI is unclear with most drugs, although there are notable exceptions.  For example, patients infected with both human immunodeficiency virus (HIV) and hepatitis C virus (HCV) appear to be at higher risk for liver injury from antiretroviral drugs than those infected with HIV infection alone.  Patients with chronic hepatitis may also be more susceptible to liver injury from the anti-tuberculosis drug, isoniazid (INH). While underlying liver disease may not predispose to DILI in most situations, common sense dictates that patients with chronic liver disease who suffer significant superimposed DILI will have a worse outcome than patients who are otherwise healthy.  In addition, the presence of abnormal baseline liver tests may make screening for DILI in the patient with underlying chronic liver disease more challenging.  Thus, drugs associated with a significant risk of liver injury should be used with great care, if at all, in patients with chronic liver disease.

How is DILI diagnosed?

Making a diagnosis of DILI is often very challenging because there are generally no reliable diagnostic tests.  A diagnosis of DILI is thus usually made through a process called causality assessment, which is quite similar to a criminal investigation, particularly one in which there are no witnesses to the crime.  In such cases, the judgment of guilt or innocence hinges on the weight of carefully collected, but ultimately circumstantial evidence.  During the process of causality assessment, many features of the drug reaction are considered, including:

• Was the drug in the "right place at the right time?”  Obviously if the drug was started after the patient developed signs and symptoms of liver injury, the drug essentially has an “alibi” and it must be exculpated. 
• Does the drug have a "prior record?"  That is, has it been previously reported to cause liver injury in similar patients? 
• Are the characteristics of the DILI event (the "crime") consistent with the known signature, or "modus operandi" of the implicated drug?  For instance, isoniazid (INH), a drug used to treat tuberculosis, characteristically causes hepatocellular injury with elevated serum aminotransferases and, in severe cases, jaundice as well.  The alkaline phosphatase, a marker of cholestasis, is typically only minimally elevated.  Anticonvulsants, such as phenytoin (Dilantin®), can cause liver injury as part of a hypersensitivity reaction so that rash and fever are often prominent features.  A peculiar signature of amoxicillin-clavulanic acid (Augmentin®), which can cause liver injury in a cholestatic or mixed pattern with jaundice, is that it typically only does so only after the drug has been stopped, usually within two to three weeks.  Finally, nitrofurantoin (Macrobid®) characteristically causes a chronic hepatitis after many weeks, months, or even years of therapy, and is often associated with serum antinuclear antibodies (ANA).  Signature features such as these can be helpful in a DILI diagnosis, though for most drugs there are no such features.
• Are there other explanations for the liver injury?  Critical to the diagnosis of DILI is the exclusion of other possible (so-called “competing”) causes of severe liver injury, such as viral hepatitis, bile duct obstruction, hepatic ischemia (e.g., secondary to shock), autoimmune hepatitis, and the like.  In essence, DILI is a diagnosis of exclusion.  More about this critical point below.
• Is the clinical course of the reaction consistent with DILI?  Most cases of drug-induced liver injury, particularly when not very severe, resolve relatively quickly when the drug is discontinued; the tempo of this improvement in the liver tests is called the “de-challenge”.  If the drug is reintroduced, the liver injury often returns, sometimes more rapidly than with the initial exposure; however, “re-challenge” of the patient with the suspect drug is not generally advised as a diagnostic test because the liver injury that follows re-challenge can sometimes be quite severe, even life-threatening.  Indeed, re-challenging the patient with suspect drug is analogous to letting the suspect criminal out of custody to see if he commits the crime again – risky, to say the least!

The clinician is typically faced with the daunting task of carefully weighing these various bits of circumstantial evidence to make the DILI diagnosis.  Moreover, that diagnosis can never be definitive because, with very few exceptions, we lack specific diagnostic testing as it really is a diagnosis of exclusion. 
Serological testing is done to exclude viral hepatitis A, B, C and sometimes even E (very rare in the United States) in most patients, as are serological markers of autoimmunity (e.g., an antinuclear antibody (ANA), smooth muscle antibody (SMA), and gamma globulins). Others tests exclude metabolic and genetic problems, including iron and copper overload (hemochromatosis and Wilson disease, respectively), as well as other disorders (e.g., alpha-1 antitrypsin deficiency, celiac disease).  In some cases, a careful history is needed to exclude severe hypotension preceding the onset of liver injury, which would suggest "shock liver" as a possible diagnosis if the liver tests suggest this possibility.  Alcohol abuse must also be excluded by careful interrogation of the patient.   Fatty liver (particularly non-alcoholic steatohepatitis or NASH) is a common cause of low level abnormal liver tests that can further evaluated by liver imaging, although liver biopsy is often needed for definitive diagnosis. Finally, most patients will also require some type of liver imaging, either by ultrasound, CT scan, MRI scan or, in cases of suspected biliary obstruction, possibly a test known as endoscopic retrograde cholangiopancreatography (ERCP).  Liver biopsy is also sometimes very useful in the diagnosis of drug-induced liver injury, but it is rarely diagnostic, because under the microscope DILI can mimic virtually the entire spectrum of other causes of liver injury, including viral and autoimmune hepatitis, biliary tract disease, and, in some cases, even alcohol abuse.  Liver biopsy also carries a small but significant risk of bleeding and other complications.
Cases of possible DILI in patients with underlying liver disease or who are taking multiple drugs are particularly vexing.  The diagnosis of DILI in such cases is almost always tentative.

What is the prognosis of DILI?

Although they may cause patients understandable anxiety, mild, asymptomatic DILI usually causes no significant liver problems.  Indeed, in many cases the liver tests improve over time despite continued use of the drug due to a poorly understood process called “adaptation,” which probably reflects the liver’s robust response to injury.  We believe that the relatively rare cases of more severe idiosyncratic DILI may reflect the affected patient’s liver failing to undergo this process in response to injury. 
Over two decades ago, Dr. Hyman Zimmerman, a pioneer of DILI research, observed that an elevated serum ALT (> 3 x ULN) accompanied by jaundice due to drug toxicity carried a significant mortality ranging from 10-50%, at least 10 fold higher than icteric viral hepatitis.  This important observation has been termed "Hy's law" in honor of Dr. Zimmerman and is regarded by regulatory agencies such as the FDA as a sign of serious DILI. 
A rise in serum bilirubin, which reflects impairment in the liver's ability to excrete bilirubin from blood into bile, is associated with a worse prognosis.  Likewise, a rise in the prothrombin time (or its international normalized ratio (INR), and a fall in serum albumin reflect a severe impairment of the liver's protein synthetic function and are also markers of poor outcome.  Finally, DILI severe enough to cause hepatic encephalopathy [see HE knoll], thus signifying acute liver failure [see ALF knoll], has a dire prognosis and is associated with a 70-80% chance of death without prompt liver transplantation.
Rarely, an episode of severe acute DILI can appear to lead to a smoldering chronic liver disease.  Likewise, there are rare drugs that have been associated with severe liver scarring; methotrexate, used to treat rheumatoid arthritis and psoriasis, can cause cirrhosis in rare individuals, particularly with long term use. 

How is DILI treated?

The mainstay of therapy for DILI most common treatment for DILI is quickly stopping and avoiding re-exposure to the implicated drug.  Typically with mild to moderate drug-induced liver injury this is sufficient and results in relatively rapid and complete resolution of all symptoms and signs of liver injury, although cholestatic liver injury may linger, with complete resolution potentially taking many weeks, even months. 
In some DILI cases, specific therapies are available.  Prompt institution of n-acetylcysteine (Mucomyst®) for acetaminophen poisoning is critically important and may be life-saving.  Some studies suggest that the use of intravenous L-carnitine may be effective for ameliorating the toxicity associated with the anti-convulsant valproate.  The use of cholestyramine (Questran) is recommended for severe toxicity from leflunomide (Arava®) probably because this drug has a prolonged circulation within the liver, which makes it possible for it to be effectively bound and eliminated by cholestryamine. 
For cholestatic drug-induced liver injury, ursodeoxycholic acid (Actigal®) is safe and may possibly hasten resolution of jaundice and pruritus, although this has not been well studied.  Antipruritic agents (e.g., cholesytramine) are also used in this setting as well to relieve itching. 
Some controversy surrounds the role of corticosteroid therapy in the treatment of drug-induced liver injury.  For DILI from most drugs, corticosteroid therapy is probably ineffective, and does carry some risk.  Potential side effects of corticosteroids include glucose intolerance (diabetes) and infection, and with prolonged use, accelerated osteopenia, cataracts and skin thinning.  However, a short course of steroids is sometimes used for DILI, particularly when there are so-called hypersensitivity features such as rash, fever and eosinophilia. 
When hepatocellular liver injury is more severe, acute liver failure may evolve, and the only effective therapy may be liver transplantation.  Because of this, patients with severe DILI liver injury should be referred promptly to a specialist, typically a hepatologist or gastroenterologist, for further evaluation and care.

Why are potentially toxic drugs put on the market?

The finding of unanticipated liver injury during experiments in preclinical models probably results in the termination of many drugs at an early stage in development; drugs that possess intrinsic, dose-dependent liver toxicity are the ones most likely to be identified in this fashion. 
Unfortunately, such experiments do not detect all potentially injurious agents, particularly those that cause liver injury in an idiosyncratic fashion.  These agents may enter clinical trials where they again may go unrecognized as possibly toxic.  There are a number of reasons for this. First, severe idiosyncratic DILI is uncommon; for most drugs the reported incidence is between 1 in 10,000 and 1 in 100,000 patients exposed, although this number is difficult to determine accurately due to under reporting.   Most clinical trials, which are routinely limited to a few thousand volunteers, are simply not large enough to detect these very rare events.  Unfortunately, larger trials that might capture these rare cases are more expensive and take longer to complete, both strong financial disincentives for the pharmaceutical industry.  As a result, it is often only when the drug is marketed and many thousands of patients are exposed that the cases of severe DILI tend to appear.  This was the case with troglitazone (Rezulin®), the first in a class of novel diabetic drugs called peroxisome proliferator-activated receptor-gamma agonists.  In clinical trials of troglitazone, the incidence of ALT greater than three times the upper limit of normal in patients treated with the drug was 1.8% versus 0.6% in placebo treated controls.  Overt jaundice occurred in two troglitazone subjects, but there were no episodes of acute liver failure observed during clinical trials of the drug.  However, after approval and 2 million patients were treated with troglitazone, nearly 100 cases (1 in 20,000 patients) of acute liver failure were reported to the FDA, which withdrew the drug from the US market in March of 2000. 
Not all potentially hepatotoxic drugs escape clinical trials unrecognized.  For example, the FDA recently failed to approve ximelagatran (Exanta®), which was poised to replace coumadin as a convenient, effective, and presumably safe oral anticoagulant for deep blood clots, atrial fibrillation, and other conditions.   The drug was associated with a serum ALT of greater than three times the upper limit of normal in almost 8% of treated patients in clinical trials.  More importantly, there were several cases of acute liver failure observed in the clinical trial population of approximately 7,000 patients. As a result, the drug was never marketed, much to the chagrin of the pharmaceutical company that invested considerable time and resources into the development of this “block buster” drug.  
On the other hand, some drugs that are well-characterized idiosyncratic liver toxins remain on the market because of practical considerations.  Isoniazid (INH) is one of the mainstays for treatment of tuberculosis (TB) but is associated with significant hepatocellular injury in 0.3 to 4.6 percent of treated adult patients, an increasing risk associated with advancing age.  Despite this risk, INH presumably remains on the market because there are no other effective and safe drugs to fill its niche for the treatment of TB, a major public health threat. 

What is the future for DILI?

Clearly we are a long way from really understanding how idiosyncratic DILI develops. A critical question is: what are the traits of the individual – be they genetic or environmental – that increase the risk for idiosyncratic DILI?  If we can identify these traits then, perhaps, we can avoid prescribing certain medications to patients identified as being at particularly high risk for developing DILI.  Such testing, if sufficiently robust and accurate, may also allow use of medications that would otherwise be removed from the market.
In an effort to make inroads into this difficult but important problem, the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK) established in 2003 the Drug-Induced Liver Injury Network (DILIN).  DILIN consists of 5 clinical centers – the University of Michigan, Indiana University, University of Connecticut, University of California San Francisco, and the University of North Carolina – as well as a data collection center at Duke University [see Figure below].  The primary objective of the network is to advance our understanding of drugs that cause severe idiosyncratic hepatocellular and cholestatic liver injury.  Because there are currently few, if any, robust and accurate preclinical models of idiosyncratic drug-induced liver injury, the DILIN is developing a comprehensive registry of detailed clinical data, as well as serum, plasma, urine and DNA from patients who have suffered bona fide idiosyncratic DILI that investigators worldwide interested in DILI can exploit.  Hopefully, the DILIN registry, which currently includes several hundred DILI cases, will be used in pharmacogenomic and other studies to better understand the genetic and environmental factors contributing to the development of DILI and, furthermore, to identify novel biomarkers associated with this important problem.

References:

• Kaplowitz N. Idiosyncratic drug hepatotoxicity. Nat Rev Drug Discov. 2005 Jun;4(6):489-99
• Watkins PB, Seeff LB: Drug-induced liver injury: summary of a single topic clinical research conference. Hepatology 2006 Mar; 43(3): 618-31
• Watkins PB, Zimmerman HJ, Knapp MJ, et al: Hepatotoxic effects of tacrine administration in patients with Alzheimer's  disease. JAMA 1994 Apr 6; 271(13): 992-8
• Sgro C, Clinard F, Ouazir K, et al: Incidence of drug-induced hepatic injuries: a French population-based study. Hepatology 2002 Aug; 36(2): 451-5
• Ostapowicz G, Fontana RJ, Schiodt FV, et al: Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002 Dec 17; 137(12): 947-54
• Gordon DW, Rosenthal G, Hart J, et al: Chaparral ingestion. The broadening spectrum of liver injury caused by herbal medications. JAMA 1995 Feb 8; 273(6): 489-90
• Chalasani N, Aljadhey H, Kesterson J: Patients with elevated liver enzymes are not at higher risk for statin hepatotoxicity. Gastroenterology 2004 May; 126(5): 1287-92
• James LP, Mayeux PR, Hinson JA (2003). "Acetaminophen-induced hepatotoxicity". Drug Metab. Dispos. 31 (12): 1499–506.
• Björnsson E, Olsson R (2005). "Outcome and prognostic markers in severe drug-induced liver disease". Hepatology 42 (2): 481–9.
• Andrade RJ, Lucena MI, Kaplowitz N, et al (2006). "Outcome of acute idiosyncratic drug-induced liver injury: Long-term follow-up in a hepatotoxicity registry". Hepatology 44 (6): 1581–8.
• Reuben A (2004). "Hy's law". Hepatology 39 (2): 574–8.
• Bénichou C (1990). "Criteria of drug-induced liver disorders. Report of an international consensus meeting". J. Hepatol. 11 (2): 272–6.

Websites:

• Drug-Induced Liver Injury Network (DILIN):  https://dilin.dcri.duke.edu/
• National Library of Medicine Hepatotoxicity URL
• Food and Drug Administration, Center for Drug Evaluation and Research URL