Increasing impact of drug‐induced liver injury

Watch a video presentation of this article

Watch the interview with the author

Answer questions and earn CME

Abbreviations

DILI

drug‐induced liver injury

DILIN

Drug‐Induced Liver Injury Network

EMR

electronic medical record

HDS

herbal and dietary supplements

HLA

human leukocyte antigen

NIH

US National Institutes of Health

NLM

National Library of Medicine

Idiosyncratic drug‐induced liver injury (DILI) represents one of the most challenging consultations for a hepatologist, and it continues to be a leading cause of acute liver failure. The true incidence is difficult to determine, but the population incidence is likely on the rise, and for certain medications, the risk may be higher than previously recognized. Therefore, DILI will be an increasing part of clinical practice. Ongoing research will change how we prevent, recognize, and diagnose DILI.

Rising Incidence and Under‐Reporting

A recent population‐based study from Iceland reported the rate of DILI to be 19.1 per 100,000, representing an increase over results from a study in France, which reported 13.9 per 100,000 in 2006.^1,2 The Iceland study was remarkable for robust data on nationwide prescriptions and reporting of DILI cases. It is also noteworthy that the population incidence increased with age from 8.5 per 100,000 for those less than 25 years old to 40 per 100,000 for those 70 years and older. Therefore, more DILI will be seen as the US population ages.

Certain medications had remarkably higher risk than expected, including amoxicillin‐clavulanate (1 in 2350), azathioprine (1 in 133), and infliximab (1 in 148). These data need verification in larger and more diverse populations. However, under‐reporting is well known in DILI, and may explain the lower incidences estimated by prior non–population based studies that relied on reports to registries and government agencies. Preliminary data from the US Drug‐Induced Liver Injury Network (DILIN) suggests that as few as 4% of isoniazid hepatotoxicity cases may be reported to the isoniazid adverse events surveillance program of the US Centers for Disease Control and Prevention (DILIN unpublished data).

Hepatotoxicity due to herbal and dietary supplements (HDS) may also be on the rise. Half the US population use HDS products, and this proportion may be growing. There was also a tripling of the percentage of HDS cases enrolled in the DILIN from 2004 to 2013³ (Figure 1). Therefore, clinicians will need to take an assiduous history of HDS use and keep abreast of a myriad of hepatotoxic products.

Figure 1.

Percentage of drug‐induced liver injury cases due to herbal or dietary supplements (HDS) enrolled in the Drug‐Induced Liver Injury Network (DILIN) over time. (Adapted from Navarro et al.3)

Future Advances in the Field

There are several areas of research that will bear clinical applications: (1) increasing use of DILI registry cases and electronic medical records (EMRs) for clinical and pharmaco‐epidemiologic studies, (2) better diagnostic aids such as online resources and diagnostic modeling, and (3) mechanistic studies that will lead to blood tests for diagnosis and risk assessment.

Registries and Electronic Medical Records

There are now several DILI registries enrolling cases worldwide. As these registries mature, they will provide growing opportunities for clinical and translational research. The registries are already producing robust clinical descriptions to provide clinicians with injury patterns and prognosis information. Registries can identify the most common classes and individual medications, thus raising clinician awareness of such agents. For example, antimicrobials and HDS products are the two most commonly implicated classes reported in the Drug‐Induced Liver Injury Network's Prospective Study (Table 1). In a preliminary analysis, there was also a higher mortality rate for those with preexisting liver disease (16% versus 5%, P < 0.01) (DILIN unpublished data). Registries that follow patients prospectively support a growing concern for chronic DILI with 12% or more having possible persistent liver injury 12 months after presentation.4

Table 1.

The Five Most Common Therapeutic Classes and Individual Agents to Cause Idiosyncratic DILI in the United States (in Order of Most Frequent to Less Frequent)^a

	Therapeutic Classes		Individual Agents
1	Antimicrobials	1	Amoxicillin‐clavulanate
2	Herbal and dietary supplements	2	Isoniazid
3	Cardiovascular agents	3	Nitrofurantoin
4	Central nervous system agents	4	Sulfamethoxazole/trimethoprim
5	Antineoplastic agents	5	Minocycline

^aGeneral observation by the authors from their experience in the DILIN Prospective Study.

As hospitals and clinics consolidate and buy larger EMR systems, the opportunities for pharmaco‐epidemiologic studies and DILI surveillance will blossom. Such “big data” coming from multiple EMRs that can merge clinical information, including laboratory values and prescriptions, will enhance adverse events tracking efforts such as the FDA's Sentinel Initiative.5 These efforts may then recapitulate the data from Iceland but on a larger and more diverse scale. Finding cases for study could increase dramatically. One could also envision computerized notifications. Any new prescriptions found 90 days prior to onset of elevated liver biochemistries could be flagged through an EMR, thus automatically alerting the clinician to consider DILI.

Diagnosis

DILI diagnosis relies on a careful history and diagnostic tests which focus on pattern of injury, defining drug exposure and ruling out competing etiologies. For now, the RUCAM (Roussel Uclaf Causality Assessment Model), which is a diagnostic scorecard system, remains the best available tool for the clinician.6, 7 Although not a stand‐alone diagnostic, it provides a useful starting point for the clinician. Again, growing registries of bona fide cases will allow development of better diagnostic tools. Such efforts are already underway.

Liver biopsy will remain a valuable tool when the diagnosis of DILI remains unclear. Recently, the various histologic patterns in DILI were described and some had prognostic significance.8 Severe necrosis and microvascular steatosis were associated with worse injury, whereas eosinophils and granulomas were associated with mild injury.

DILI evaluation was greatly enhanced by LiverTox, a free, online textbook for clinicians to find information on hepatotoxicity of specific agents.9 Sponsored by the NIH and National Library of Medicine (NLM), it was launched in 2012. The site already averages 115,000 unique visitors per month and contains more than 750 agents with plans to expand to more than 1000. Information is concisely structured around risk, latency, pattern of injury, and outcome for each agent. Robust references from the NLM are provided along with illustrative cases.

Mechanistic and Genetic Studies

A myriad of genes and metabolic enzymes have been associated with hepatotoxicity, particularly human leukocyte antigen (HLA) genotypes.10 The many HLA associations clearly point to an immune component in DILI and will open new avenues for both diagnostic and therapeutic studies. Few markers thus far are specific or common enough to be used diagnostically or for risk stratification. However, the combination of growing registries and next‐generation sequencing technologies is expected to hone in on more clinically useful markers.

Summary

DILI incidence will continue to grow as more medications are prescribed to an aging population, and more HDS products are used. It will remain a clinical challenge, but advancements are poised to accelerate as registries grow. LiverTox represents the newest tool for the clinician and is being used extensively. Pharmaco‐epidemiologic and registry studies will enable elucidation of DILI pathophysiology and genetic risk factors. Such understanding will lead to better diagnostic tools, and eventually, prevention and therapeutics.