Drug-Induced Liver Injury Network Causality Assessment: Criteria and Experience in the United States

Abstract

Hepatotoxicity due to drugs, herbal or dietary supplements remains largely a clinical diagnosis based on meticulous history taking and exclusion of other causes of liver injury. In 2004, the U.S. Drug-Induced Liver Injury Network (DILIN) was created under the auspices of the U.S. National Institute of Diabetes and Digestive and Kidney Diseases with the aims of establishing a large registry of cases for clinical, epidemiological and mechanistic study. From inception, the DILIN has used an expert opinion process that incorporates consensus amongst three different DILIN hepatologists assigned to each case. It is the most well-established, well-described and vigorous expert opinion process for DILI to date, and yet it is an imperfect standard. This review will discuss the DILIN expert opinion process, its strengths and weaknesses, psychometric performance and future.

Keywords: causality; diagnosis; expert opinion; hepatotoxicity.

Figure 1

Drug-Induced Liver Injury Network (DILIN) 2015. NIDDK = National Institute of Diabetes and Digestive and Kidney Diseases.

Figure 2

Drug-Induced Liver Injury Network (DILIN) consensus causality assessment. (a) Six months of clinical information are collected at the enrolling center and communicated to the data coordination center. Clinical narrative created and approved by enrolling center; (b) Narrative and data forms created by the coordinating center; (c) Finalized narrative and data forms uploaded to a secure website for the enrolling investigator and two other DILIN hepatologists, chosen at random, to independently review. Reviewers are given two to three weeks to render a causality score on the website. In this example, two reviewers give the case a “2”, but one reviewer renders a “3”; (d) Scores are released to the group and email communication between the three reviewers occurs over four days to resolve differences in scoring. All DILIN hepatologists have the opportunity to comment on the case at this stage; (e) If consensus amongst the three reviewers is obtained by email, then finalized score is sent back to the data coordinating center to be logged; (f) If consensus is not obtained via email, then the case is formally presented and discussed with all DILIN hepatologists by conference call. All attendees have the opportunity to ask questions, make comments and give opinions. If agreement amongst the three reviewers is still not reached, then consensus is obtained by majority vote (one vote per center). Finalized score sent back to the data coordinating center to be logged.