Drug Induced Liver Injury (DILI)

Nov 14, 2012

Pharmacovigilance, Drug Safety and Regulatory Affairs Author & Expert

The liver is the first stop on the path a drug takes after oral ingestion.  From the liver’s “point of view”, drugs are like other exogenous agents and are handled by the liver for excretion.  Absorption is largely into the portal system which brings intestinal blood to the liver where transformation and metabolism occur in preparation for elimination (e.g. in the urine, in bile etc.).  Metabolism may occur in two phases: Phase I is preparation by oxidation, reduction, hydrolysis and other reactions of the compound for action in Phase II.  The second phase occurs in the cytosol and involves reactions  catalyzed by transferase enzymes.    The cytochrome P-450 system plays a major role in this.  The study of the P-450 system allows some degree of prediction of possible hepatotoxicity.

Liver toxicity is usually thought of as producing damage to the liver parenchymal cells (hepatocytes) but toxicity can also be produced affecting biliary cells, kupffer cells, fat storing cells etc.

Adverse events can be classified as Type A which represents pharmacologic effects and are often (but not always) predictable.  There may be a dose response effect with worsening of the toxicity only after a particular blood level or threshold is reached of the drug or its metabolites.  The classic example of this is acetaminophen (paracetamol) toxicity which is toxic only if more than about 6-8 grams a day are taken.  Type B reactions are idiosyncratic and unpredictable.  Many drugs can do this.

The toxicity produced by drugs can be of various types.  Very common is inflammation (hepatitis) with cell necrosis and inflammation.  Viruses can produce hepatitis also of course.  Drug induced hepatitis may be acute (many drugs, alcohol) or chronic (e.g. methyldopa, alcohol).  It may be catastrophic or mild. Drugs may cause bile flow impairment with or without inflammation and with or without bile duct damage.  Oral contraceptives, allopurinol and chlorpromazine may produce cholestasis as can alcohol. Other drug toxicities include fatty liver (steatosis), granuloma production and injury to the blood vessels.  Finally long exposure to some exogenous agents including drugs can produce malignancies of the liver.

Many drugs may produce mild elevations of transaminases (enzymes) such as aspartate transaminase (AST also known as SGOT), alanine aminotransferase (ALT also known as SGPT), GGTP, and alkaline phosphatase which are often called “liver enzymes” or “liver function tests”.  This is something of a misnomer as other non-hepatic problems can raise these enzymes. Bilirubin can also be elevated with liver toxicity.

Mild reversible elevations of these lab tests are usually benign and reverse after the drug is stopped or sometimes decrease or return to normal even if the drug is continued.  The real issue is to prevent severe, irreversible and sometimes fatal liver toxicity.

The clinical spectrum of liver toxicity can range from no signs or symptoms at all to mild, moderate, severe and fatal.  There is no way to absolutely know that a drug induced liver toxicity.  Even liver biopsy may not prove it.

There are some risk factors including the underlying medical condition, the very old or young, bad nutritional status, nutritional status, alcohol use (amount, chronic, acute) and concomitant medications.  The frequency of severe liver toxicity may be less than one case per 10,000 which means that this will not be identified before marketing when only a few thousand patients at most are examined in clinical trials.

Unfortunately there is usually no way to predict which patients will get the severe or fatal types of liver injury.  The incidence and amount of AST/ALT elevation, the time to onset or the type of injury do not always predict the risk and outcome of liver injury.  Thus one cannot always take comfort from the fact that the liver enzymes were “only” twice normal.

FDA has published a guidance that is well worth reading and is entitled: Guidance for Industry: Drug-Induced Liver Injury: Premarketing Clinical Evaluation July 2009 at www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM174090.pdf . Many of the points here are applicable to the marketed drug world too.

So how then do we identify drug induced liver toxicity?  As FDA notes, AST/ALT elevation accompanied by jaundice is one of the strongest signs of hepatocellular damage and which can carry a 10 to 50 percent mortality.  The reasoning is that the liver has a large excess capacity to handle bilirubin excretion and when this is exceeded (jaundice) severe injury has occurred.

This has now become known as Hy’s Law named after Prof. Hy Zimmerman who made several of these initial observations.  The “law” states that in clinical trials one should suspect serious drug induced liver toxicity if there is/are:

  1. An excess of transaminase elevations to > 3 times the upper limit of normal (ULN) compared to a control group
  2. Marked elevations of transaminases to 5, 10 or 20 times ULN even in small numbers of the test drug group compared to control group and with increased total bilirubin to > 2 times normal without elevated bilirubin at study entry
  3. No other reason is evident for these elevations (e.g. viral hepatitis, alcohol etc.).

FDA also discusses the clinical evaluation and diagnosis of DILI in the premarketing phase but this can be extended into the postmarketing arena:

  • Underlying liver disease: Although patients with pre-existing liver disease may or may not be more susceptible to drug induced liver toxicity (this is controversial), patients with pre-existing liver disease of any sort should be carefully evaluated before new drugs are given and during any drug treatment.
  • Time to onset of liver signs and symptoms may be delayed some weeks but for some drugs a rapid onset is seen.
  • Although early signs and symptoms that are non-specific may be seen (e.g. anorexia, nausea, vomiting, fatigue, right upper quadrant pain) the first evidence is usually an elevation of ALT, AST or alkaline phosphatase.  Thus monitoring of these lab values is important to do periodically.
  • Repeated testing of these enzymes to see whether they are increasing or falling is important.  Changes may occur very rapidly (days to a couple of weeks).  Thus delay should be avoided.  Concern is high if the transaminases are greater than 3 times ULN and bilirubin greater than twice normal.
  • Other causes of liver problems should be ruled out (viral hepatitis, alcohol, ischemic hepatopathy, biliary disease, etc.).
  • Other lab tests of liver function (INR, direct bilirubin, albumin) should also be followed.
  • Dechallenge: If the drug is strongly suspected of being the cause of liver enzyme elevation, cessation of therapy is often appropriate.  Sometimes the values will not rapidly return to normal.  High levels may persist for days or weeks and may even rise further.  FDA, in its guidance on DILI in the premarketing area, suggests discontinuation of treatment if:
  • ALT or AST >8xULN
  • ALT or AST >5xULN for more than 2 weeks
  • ALT or AST >3xULN and (TBL >2xULN or INR >1.5)
  • ALT or AST >3xULN with the appearance of fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash, and/or eosinophilia (>5%)

Cases should be followed up to resolution.

  • Rechallenge may be cautiously considered if appropriate and the clinical condition warrants it.
  • Cases should be reported to the manufacturer or MedWatch (in the US or the equivalent body elsewhere).

A concise high level summary powerpoint presentation from FDA is available at www.fda.gov/downloads/Drugs/NewsEvents/UCM164456.pdf 

An attempt at causality assessment and the development of a method to do so has been developed by two scholars on drug induced liver injury: Danan and Benichou (see J Clin Epidemiol. 1993 Nov;46(11):1323-30 at https://www.ncbi.nlm.nih.gov/pubmed/8229110).

In summary, DILI is a complex topic.  The liver is a likely and frequent target of drug toxicity.  Most cases are mild and often undiagnosed (if transaminase levels are not sought) but some patients suffer severe and occasionally fatal liver toxicity.  Health care professionals and pharmacovigilance personnel should always keep this in mind when evaluating cases.

DILI can be particularly challenging for pharmacovigilance personnel in terms of coding.  Should one just code transaminase elevation and/or bilirubin elevation or should one code a specific liver problem?  Unfortunately there are no clear guidelines that work all the time.  Each case must be individualized.