Drug-induced Liver Injury Remains Under-reported, Poorly Understood
Understanding of the causes and outcomes of drug-induced liver injury remains low not only among clinicians and patients, but also among the experts who study this phenomenon. There are emerging data showing that antibiotics and herbal and dietary supplements, or HDS, are the most important culprits of hepatotoxicity in the U.S. Beyond that, less frequent associations with other drug classes have been reported, but interesting and provocative genetic signals and immune factors are starting to emerge from ongoing studies that require further follow-up.
At the heart of the problem is that most clinicians still do not report DILI, according to Robert J. Fontana, MD, Medical Director of Liver Transplantation and professor of medicine at the University of Michigan. Fontana is co-chair of the Drug-Induced Liver Injury Network (DILIN), which aims to tackle the problem head-on. “Our network is important because these injuries are both under-recognized and under-reported,” he said. “We currently have six sites right now studying the problem in the U.S., but awareness is increasing.”
DILIN, along with the Livertox website, where clinicians can report injuries and search the database for information on DILI, are key steps to learning about potential genetic and immune factors influencing these injuries. “We are doing a lot of detective work,” Fontana said. “But we’re just starting to scratch the surface. Although we are making slow and steady progress, we likely won’t have the ability to prevent DILI in the near future.”
Lily Dara, MD, assistant professor of medicine in the division of gastrointestinal and liver disease in the department of medicine at the USC Research Center for Liver Disease at the Keck School of Medicine, went deeper. “The main reason it’s under-reported is because it is under-diagnosed,” she said. “We have not done a great job of educating and raising awareness. Hepatologists will always consider DILI on the differential diagnosis of abnormal liver tests, but primary and other providers don’t always think of it.”
One hurdle is that DILI is a diagnosis of exclusion, according to Dara. “When working a patient up, you need to rule out immune disorders, genetic disorders, fatty liver, viral hepatitis, etc. You need a thorough medication history, and you need to be clever enough to ask about herbal, dietary, and weight loss supplements,” she said. “It is also important to consider that DILI can manifest after the patient is finished taking the offending agents, like a course of antibiotics. They don’t feel good after the course is over, but it may be some time before they go to the doctor. Clinicians need to be aware of these complications.”
The lack of an objective confirmatory lab test to confidently diagnose DILI is also an obstacle, according to Fontana. “It’s largely guilt by association after other more common causes of livery injury have been excluded and a careful review of medication use in relationship to injury onset,” he said. “We also don’t know if clinicians are always considering alcohol use, HIV or hepatitis C, gall stones. You have to look at the medications these patients are taking, and whether the patterns and clinical presentation are similar to what has been previously reported. Because this takes a high degree of suspicion (and time) on behalf of the busy clinician, a diagnosis of DILI is frequently delayed or overlooked altogether.”
Fortunately, more information is becoming available, largely in the form of the Livertox website, which has been assembled by the NIH in collaboration with the National Library of Medicine. “We have almost 2,000 patients in the DILIN database, from 2003,” Fontana said. “Before 2003, you had to go into textbooks or online, and information wasn’t easy to find. With Livertox, the world’s literature has been collated into an easily searchable database of all human hepatotoxicity.”
Livertox includes a catalogue of some 700 drugs. It is free and supported by PubMed, with the DILIN network and the NIH updating and curating the site. DILIN is also publishing reviews and case series, and there is biopsy information on the LiverTox site.
“The Livertox website has been a very helpful reference for the busy clinician,” Michael D. Leise MD, associate professor of medicine and consultant in the division of gastroenterology and hepatology at the Mayo Clinic in Rochester, Minn., said in an interview. “Additionally, the publications by the DILI Network have also been pivotal for advancing our understanding of the epidemiology and outcomes of DILI in the U.S. The actual reporting of DILI could probably be made easier. For cases of possible DILI, forms can be submitted through the FDA website, through the drug manufacturer, through Livertox, or through case reports or case series. A streamlined and systematic application that might be used for multiple sources would certainly help the busy clinician.”
“The NIH plays a tremendous role and has done its part by building the DILIN registry,” Dara added. “There is only so much we can do, or individual clinicians can do.”
The AASLD is also helping, according to Leise. “They have been very supportive of fostering additional research and collaboration in the area of DILI through promotion of an annual Drug Induced Liver Injury Conference,” he said. “It engages interested stakeholders from all aspects of medicine including clinicians, scientists, pharmaceutical and regulatory bodies.”
Dara said that the AASLD and the FDA also conduct workshops and conferences on DILI, while the FDA is also ramping up post-marketing studies to see how drugs and drug classes behave in real-world settings. “The problem is that there is little information on new drugs,” she said. “Doctors don’t even know what to monitor. The good news is that the FDA has become more sensitive about DILI. When they see a new family of drugs, they put them through the ringer. Often the severe cases of idiosyncratic DILI from new drugs is noted after the drug has been approved and consumed by a large number of patients. Therefore, there is a need for post marketing surveillance.”
It is difficult, though, to conduct this kind of phase 4, comprehensive follow-up non-intervention and post marketing surveillance on every single new drug, Dara said. “It’s up to us to educate and raise awareness, and report cases of DILI when the drug is used in the real world.”
Leise put a different spin on the under-diagnosis of liver injury. “There are perhaps a couple of reasons for this,” he said. “One, the liver biopsy is being used less often in current practice given the advent of other modalities such as transient elastography and MR elastography. The liver biopsy can be a very important clue to the diagnosis of DILI. Second, NAFLD is increasingly common in our GI and hepatology clinics. It is estimated that approximately 20% of American adults have NAFLD. I suspect that from time to time, cases of DILI are misclassified as NAFLD because of the high prevalence of NAFLD, the lack of a gold standard blood test for either diagnosis, and the waning use of the liver biopsy.”
Specific Drug Classes
Dara pointed to several non-FDA-approved treatments as culprits of liver injury. She mentioned The Dietary Supplement Health and Education Act of 1994 stating that herbal products are not drugs and do not have to be approved by the FDA. “We see patients who have taken Chinese medicines, dietary and herbal supplements that present with liver toxicity and even liver failure,” she said. “None of it is regulated. These treatments don’t go through the steps required for FDA approval. As a society, we have to decide if we want to regulate these.”
Fontana built on this point. “Patients often don’t reveal to doctors that they are taking dietary supplements because they can be bought over the counter,” he said. “The sum total amount of money we in the U.S. spend on over-the-counter medications and supplements exceeds that of all prescription drugs. We as a society are obsessed with these products. People think they are safe because they’re at the health food store or corner market and they’re selling it without a doctor being involved. However, the safety of most of these products is completely unknown and we are seeing an increasing incidence of HDS cases in the DILIN study. The public needs to be aware.”
Chalasani and colleagues from DILIN followed 1,257 participants with suspected DILI for 6 months or longer from 2004. Clinicians determined the causality of DILI in 1,091 of those patients, with 899 considered to have definite, highly likely, or probable DILI, according to the findings. The rate of mortality or liver transplantation was 10%, while chronic liver injury was reported in 17%. DILI was found to trend toward more severe among the 10% of patients with DILI who had pre-existing liver disease. Mortality rates were 16% among those with pre-existing liver disease and 5.2% for those without (P < .001). Azithromycin was the causative agent in 6.7% of patients with pre-existing liver disease and 1.5% of those without (P = .006). Among 41 cases with latency 7 days, antibiotics caused DILI in 71%. The study also included 60 cases with latency longer than 365 days. Nitrofurantoin was the causative agent in 25% of those cases, while minocycline caused 17%. Outcomes were similar among patients with short-term and long-term latency. Cholestatic injuries were more common among individuals aged 65 years or older compared with those younger than 65.
“After herbal supplements and antibiotics, it’s a potpourri of other drugs,” Fontana said. He stressed that many clinicians are surprised that antibiotics are the leading cause of DILI, as they are usually only given for a few days, whereas statins or anti-hypertensive medications can be taken for months or years. “It’s important for doctors to know this, because this will help them ask the right questions of their patients. Once we have more incidence information, we can look in the lab to find the active constituents.”
Zoubek and colleagues investigated the Spanish and Latin American DILI registries. They found 21 cases of ibuprofen-induced idiosyncratic DILI in the Spanish registry and five ibuprofen-associated cases from the Latin-American registry. The data set also included 76 cases of idiosyncratic DILI caused by NSAIDs and 844 cases caused by non-NSAIDs. Ibuprofen caused 29% of the 76 NSAID cases in the Spanish registry, while diclofenac caused 18%. In the Latin American registry, nimesulide caused 38% of NSAID-associated hepatotoxicity cases, followed by diclofenac (34%) and ibuprofen (17%). Diabetes was reported in 27% of individuals with ibuprofen-associated toxicity, compared with a 7% diabetes rate among those with NSAID-associated toxicity and 12% in the non-NSAID group (P = .04). The researchers suggested that there was no characteristic drug signature associated with ibuprofen-associated hepatotoxicity compared with the other causative agents. Additionally, variability was reported among patients with ibuprofen DILI.
At this point, it may be useful to address the difference between direct and indirect, or idiosyncratic, toxicity. “Some drugs, like acetaminophen, for example, have a direct impact on hepatocytes,” Dara said. “Around 50% of liver failure cases in the U.S. are due to acetaminophen. If you take more than 3 g of acetaminophen, depending on your size and if you are fasting or fed, you are going to get some degree of liver injury.”
But acetaminophen can be found in products ranging from Nyquil (Vicks) to Vicodin (AbbVie), according to Dara. She noted that most cough and cold medications contain acetaminophen, as well. “Idiosyncratic toxicity is not dose-related, it does not have an on-target effect. It may be just part of the drug molecule, and it may depend on host-related and genetic factors.”
Fontana further defined idiosyncratic toxicity in a 2014 editorial, describing it as “not a single disease entity but rather a spectrum of rare diseases with varying clinical, histological, and laboratory features.”
Genetic, Immune-Mediated Response
Dara and colleagues wrote that associations have been reported between idiosyncratic DILI with certain drugs and single nucleotide polymorphisms in human leukocyte antigen (HLA). However, not all individuals with the possible HLA genotypes experience clinically significant liver injury when treated with those medications. The reason for this remains unknown, or poorly understood, at best. The investigators put forth the possibility that there is an adaptation process that occurs in certain individuals to prevent or protect from significant injury, or allow spontaneous resolution of injuries that occur. Further investigation of the immune microenvironment in the liver, along with mechanisms leading to idiosyncratic DILI, may yield answers, according to Dara.
“For example, it has been shown that single nucleotide polymorphisms such as DRB1*1501 or DRB5*01:01-DQB1*06:02, among others, are associated with amoxicillin-clavulanate (Augmentin, Dr. Reddy’s Laboratories) toxicity,” Dara said. “However, this HLA is common in the general population. People who experienced toxicity often had the HLA, but not everybody who has the HLA will experience toxicity.”
Dara suggested that liver toxicity may have nothing to do with the target of the drug, but how it alters T-cells. “This is not 100%, but it is what we are looking at,” she said.
Because of this association, the HLA target has become a hot area of research, Dara added. “DILI is happening once drugs hit the market,” she said. “Sometimes the signal will emerge when the drug is in development, but often since idiosyncratic DILI is not that common, it usually is reported once it’s taken by a large number of people, maybe at a prevalence rate of 10 or 15 per 100,000.”
Also in the 2014 editorial, Fontana wrote that new serum biomarkers are being investigated, such as “glutamate dehydrogenase, high mobility group box protein 1, and microRNA-122.” He suggested that associations in the HLA region may be associated with drugs including flucloxacillin, ximelagatran, lapatinib, and amoxicillin-clavulanate. “Further genome-wide studies are needed, as are serum proteomic, transcriptome, metabolome, and intestinal microbiome analyses,” he wrote.
In a new line of thinking, there may be what Dara calls a “second-level factor,” which is one associated with loss of immune-tolerance. “The liver is an immune tolerant organ,” she said. “But if you inhibit the immune checkpoint, or have an altered immune-tolerance due to genetic or environmental factors, you may be more susceptible to liver injury.”
Chalasani and colleagues investigated liver injury severity in a cohort of U.S. patients. Results showed a higher mean severity score among black individuals compared with whites. Severe or fatal cases of DILI also occurred more frequently among blacks (34% vs. 24%; P < .001), as did an increased frequency of severe cutaneous reactions (2.1% vs. 0.36%; P = .048). A trend toward increased DILIN severity score was also observed among blacks after the researchers controlled for BMI (OR = 1.98; 95% CI, 1.42–2.75), while liver transplantation and death occurred more frequently among blacks in an analysis that controlled for age and BMI (OR = 2.12; 95% CI, 1.12–3.9).
“It is important to understand with papers like this that these are still rare events,” Fontana said. “They may occur in one in 10,000 or 100,000 people who get drug. Because they’re rare, they may get missed.”
Fontana and colleagues assessed short-term outcomes in a cohort of 660 individuals who demonstrated signs of definite, highly likely, or probable DILI. Liver transplants occurred in 4.5% (95% CI, 3.0-6.1) of the cohort within 6 months of DILI onset. The mortality rate at 6 months was 5% (95% CI, 3.2–6.5), with 53% of those deaths being liver-related, while 18.9% of the cohort had persistent liver damage at that time point. Several factors were independently associated with risk for faster time to transplantation or liver-related mortality, including Asian race, absence of itching, lung disease, low serum albumin levels, low platelet counts, and high serum levels of alanine aminotransferase and total bilirubin (C-statistic = 0.87), according to the findings. Chronic DILI was more likely to occur in blacks and in individuals with higher serum levels of alkaline phosphatase and a history of heart disease or malignancy requiring treatment (C-statistic = 0.71). “The profile of liver injury at presentation, initial severity, patient’s race, and medical comorbidities are important determinants of the likelihood of death/transplantation or persistent liver injury within 6 months,” the researchers concluded.
“When you look at the whole population, you might see more DILI among older adults than children, simply because they have experienced more medication use than children,” Fontana said. “Overall, though, we are not seeing an over-representation of older vs. younger adults, minorities vs. white patients. There is a slight increase in frequency of reporting in women than in men, but that may be because women generally receive more medical care and medications than men. They go to the doctor more, and therefore there is more exposure.”
Moving forward, with these complicating factors at hand, Dara put the onus on gastroenterologists and hepatologists, particularly those in academic centers, to spread the word about DILI. “We need to educate clinicians in the community and rural areas,” she said. “We need to help them ask the right questions and pay attention to the drugs their patients are taking. It is really important to recognize DILI early on so the culprit drug can be stopped in a timely manner.”
Fontana understands that busy clinicians may not have time to ask the specific and comprehensive questions needed to target DILI. “There are no easy solutions,” he said. – by Rob Volansky
- Chalasani N, et al. Am J Gastroenterol. 2017;doi:10.1038/ajg.2017.215.
- Chalasani N, et al. Am J Gastroenterol. 2015;doi:10.1053/j.gastro.2015.03.006.
- Dara L, et al. Liver Int. 2016;doi:10.1111/liv.12988.
- Fontana RJ. Gastroenterol. 2014;doi:10.1053/j.gastro.2013.12.032.
- Zoubek ME, et al. Clin Gastroenterol Hepatol. 2017;doi:10.1016/j.cgh.2017.07.037.
- For more information:
- Lily Dara, MD, can be reached at 2011 Zonal Ave HMR 101, Los Angeles, CA 90033; email: email@example.com.
- Robert J. Fontana, MD, can be reached at 3912 Taubman Center Ann Arbor, MI 48109; email: firstname.lastname@example.org.
- Michael D. Leise, MD, can be reached at 200 First Avenue SW, Rochester, MN 55905; email: Leise.Michael@mayo.edu.
Disclosures: Dara and Leise report no relevant financial disclosures. Fontana reports conducting research funded by Gilead Sciences, AbbVie Labs, and Bristol-Myers Squibb in viral hepatitis.