AASLD-FDA-NIH-PhRMA Hepatotoxicity Steering Group Meeting, 2006 Presentations: Criteria for causality assessment in drug-induced liver injury (DILI): results from the Acute Liver Failure (ALF) Study Group
William M. Lee, MD
University of Texas Southwestern
Criteria for causality assessment in drug-induced liver injury (DILI): results from the Acute Liver Failure (ALF) Study Group [PDF]
Assessing causality in ALF is essential to determining treatment and prognosis. For example, a diagnosis of acetaminophen (APAP) or mushroom poisoning leads to specific antidote therapy. Diagnosing ALF due to viral hepatitis is based on standard serological tests that are quite definitive, at least for acute hepatitis A and B, and a diagnosis of hepatitis B-related ALF nowadays leads to use of a nucleoside analogue. There is no specific antidote for idiosyncratic drug-induced liver injury (DILI) which now comprises 12% of all US ALF; however, determining causality in DILI is important so that likely offending agents can be stopped and their future use interdicted in the affected patient. Causality assessment is not evidence-based. Committees of experts determine the likelihood of a given reaction in the liver being related to a suspect agent but often cannot agree on the criteria used or the likelihood in a given situation due at least in part to lack of adequate data. The unique features of DILI-related ALF (severe, associated with encephalopathy, 80% fatal) make causality assessment even harder. An adequate history is impossible in many instances and the de-challenge (recovery) period, considered crucial to most CAMs, is unavailable. However, since patients with ALF are invariably hospitalized, adequate testing to rule out other causes can usually be performed. Data from a series of 307 patients with severe liver injury (SLI), 50% of whom suffered from ALF, have demonstrated that alcohol and age are not indicative co-factors of DILI-related SLI and should probably be removed from future CAM considerations. In our series, significantly fewer DILI patients gave a history of any alcohol use (25%) as compared to those with acetaminophen (55%) and viral hepatitis (49%, p values <0.05). Alcohol abuse was also less frequent in the DILI group (5%) than in the acetaminophen (18%) or viral hepatitis group (12%), but the differences were not statistically significant. In this series, DILI patients were older than the other groups (median age 49 years vs. 35 and 34 years for APAP and viral hepatitis cases) but this likely relates to the association of high-risk behavior (e.g., substance abuse) observed in a younger population and specifically associated with APAP and viral hepatitis. The DILI group took a median of 6 medications compared to 2 and 3 respectively for the other groups. This may simply relate to their increased age, since increased age is associated with greater numbers of medications used in the NHANES study. It is possible that the association of age as a risk factor for DILI relates only to the increased probability of toxicity provided by increased numbers of medications in play.
Causality assessments also function to help determine the overall burden of drug-induced liver damage from a given product. This is vital to knowing the relative safety of a drug from the standpoint of the population as a whole. In this context, APAP toxicity represents the biggest drug-related issue in the US accounting for ~50% of all US ALF. While many patients readily admit to taking an overdose in a suicidal gesture, other patients are unable, unknowing or unwilling to give a history of APAP use. In this setting the unique features of APAP toxicity and its high acuity (very high aminotransferases, low bilirubin levels) signal its presence. Better than these clinical clues is a recently described assay that measures acetaminophen adducts which represent the toxic byproduct of hepatocyte necrosis and remain in serum roughly as long as aminotransferase are elevated, days after the parent compound is gone. This kind of assay represents a breakthrough in providing irrefutable proof of causality and will be a powerful tool in determining whether APAP is involved in settings where it has not been suspected. For example, our data suggests that many indeterminate and viral hepatitis cases suffer from acetaminophen hepatotoxicity.
Dr. Lee was educated at Amherst College and received his medical degree from Columbia University College of Physicians and Surgeons, AOA, completing his training in Internal Medicine at the Presbyterian Hospital, New York City (Columbia-Presbyterian Medical Center). He served as Honorary Registrar and Research Fellow to Dr. Roger Williams at the Liver Unit, Kings College Hospital, London from 1973-74, then was on the faculty of Columbia-Presbyterian and later served as Chief of Gastroenterology at the MUSC, Charleston, SC. Since 1990 he has been Professor of Internal Medicine at UT Southwestern Medical Center at Dallas, and holder of the Meredith Mosle Chair in Liver Disease and Director, Clinical Center for Liver Disease. UT Southwestern Medical Center at Dallas.
Dr. Lee is Principal Investigator for the NIH-sponsored Acute Liver Failure Study Group comprising 24 adult and 24 pediatric sites. In addition, he is Principal Investigator for one of the 10 sites participating in the NIH HALT-C study, a long-term treatment program for patients who are non-responders to interferon therapy. His group currently is conducting more than 10 trials of new agents for the treatment of viral hepatitis and conducts basic research in hepatitis B and C.