PHILADELPHIA – A registry created by the National Institutes of Health (NIH) to track idiosyncratic drug-induced liver injury (DILI) largely confirms the significant risks suggested by small studies and textbook summaries, according to newly summarized data presented at the American College of Gastroenterology (ACG).
According to the registry, antimicrobials dominate cases of DILI, a substantial minority of DILI patients die of this complication or require liver transplant, and the risks of adverse outcomes from DILI are far greater in the presence of hepatitis or another preexisting liver disease, reported Dr. Naga P. Chalasani, chief of the division of gastroenterology and hepatology, Indiana University school of medicine, Indianapolis.
Dr. Naga P. Chalasani
Since 2004, when the Drug Induced Liver Injury Network was established by the NIH, 1,257 patients have been enrolled of which 899 were found to have probable, highly likely, or definite DILI by an adjudication committee charged with determining causality. The mortality rate overall was 6.2% with an additional 4% undergoing liver transplant.
The most commonly implicated causal agents by class were antimicrobials (45.4%), herbal and dietary supplements (16.1%), cardiovascular agents (9.7%), and CNS agents (9.1%) with a wide variety of other classes falling in at much lower representation. Nine of the top 10 individual agents implicated in DILI (amoxicillin-clavulanate led the list) were antimicrobials. The exception — and in the 10th spot — was diclofenac.
“What may be of interest to the clinician is how long it takes on average to recover,” Dr. Chalasani observed. For those who present with jaundice, which represented about 70% of patients in the registry, the average was about 70 days. This information may be useful “if a patient asks you how long I am going to be yellow.”
Characterizing DILI as hepatocellular, cholestatic, or mixed may have value in assessing risk. According to Dr. Chalasani, the rates of DILI-related death were higher in cholestatic than hepatocellular presentations, but the risk of liver transplant was lower. In the mixed group, liver-related death was uncommon, and there were no liver transplants even though these patients more often developed chronic DILI.
The average time from exposure to DILI in drugs with a short latency was 36 days. Drugs with a long latency, defined as more than 1 year, included nitrofurantoin, statins, amiodarone, and mercaptopurine.
“The pattern of long latency is very interesting. Patients are often on a stable dose of a given drug for a length of time and doing well and then the dose is escalated and, boom, they present with DILI,” Dr. Chalasani reported.
Despite the hypothesis that there might be differences in characteristics and outcomes from long versus short latencies, none were seen.
Stevens-Johnson Syndrome (SJS) was observed in about 1% of the DILI patients in the registry. Antimicrobials were again implicated in patients who developed SJS, but the list of most common drugs in this subset also included lamotrigine and carbamazepine.
Patients with preexisting liver disease, such as a form of hepatitis or nonalcoholic fatty liver disease, represented 10% of the registry population. The causative agents were similar with the exception of azathioprine, which appeared to impose an even greater risk in those with liver disease than in those without.
“Although it has been written in the textbooks, this is the first set of data to show DILI in patients with preexisting liver disease imposes a three times higher likelihood of death than in those without liver disease,” Dr. Chalasani reported.
DILI is a rare disease, occurring in only about 20/100,000 patients. As a result, the NIH registry is considered an essential tool for gathering data on this complication in order to assess its causes and prognosis.
Dr. Chalasani reports that he and his coauthors have multiple financial relationships with industry but none that are relevant to this study.