Liver Safety Assessment in Clinical Trials of New Agents for Chronic Hepatitis B

Robert J. Fontana; Mark I. Avigan; Harry L. A. Janssen; Arie Regev; Poonam Mishra; Anuj Gaggar; Nathaniel Brown; Cynthia Wat; Patricia Mendez; Ryan T. Anderson; Bruce Given; Veronica Miller; Maria Beumont


J Viral Hepat. 2020;27(2):96-109. 

In This Article

Idiosyncratic DILI With Investigational Agents

Cases of DILI have been reported with most of the 900 drugs approved by the Food and Drug Administration (FDA) and with a multitude of herbal and dietary supplements.[31,32] DILI may present with distinct clinical, laboratory and histological manifestations presumably due to differing mechanisms of hepatotoxicity.[33] Generally, the development of idiosyncratic DILI does not usually correlate with the dose or duration of suspect drug used or known effects of the drug identified during preclinical toxicology studies or early human pharmacokinetic (PK) studies. Nonetheless, heightened DILI risk may occur if a certain level of drug exposure or duration of treatment has been exceeded. Because these DILI reactions are idiosyncratic, they depend on increased host susceptibility that may reflect inter-individual variations in drug metabolism, immune responses or susceptibility to off-target toxicological effects.

In clinical trials of patients without pre-existing liver disease, hepatocellular DILI may be suspected whenever a drug-exposed patient experiences an increase in their serum ALT levels to >3× ULN or from their baseline values during or soon after treatment for which there is no other plausible cause identified. Given that CHB has a high underlying rate of serum ALT elevations, distinguishing DILI due to an investigational agent from an HBV-related cause of ALT increase is always challenging. In other instances, DILI due to an investigational agent may lead to a predominant increase in serum alkaline phosphatase (ALK) levels to >2× ULN and/or an increase in total bilirubin to >2× ULN or 2.5 mg/dL, consistent with a cholestatic form of injury. In patients without underlying liver disease, drug-induced acute cholestatic injuries typically do not progress to acute liver failure (ALF). Nonetheless, cholestatic injuries in certain patients with underlying liver disease or cirrhosis may be associated with substantial clinical worsening, and in some instances, hepatic decompensation or death[34] (see Text S1).

Establishing a diagnosis of DILI requires a methodical approach that must exclude other more common causes of liver injury. Currently, there is no objective laboratory biomarker to unequivocally establish that an episode of liver injury is due to a drug vs another more common cause of liver injury such as alcohol, viral hepatitis or pancreatobiliary disease.[35]

Causality Assessment in Clinical Trials

Causality assessment of DILI in clinical trials is most commonly based on expert opinion. Individuals with recognized expertise in the evaluation of DILI conduct a comprehensive review of liver injury cases of interest, scoring each of them on an ordinal scale of likelihood of causal association with the study drug that vary from >95% (definite) to unlikely (<25%). This approach permits accounting for extrahepatic features such as the presence of fever, rash or eosinophilia that make DILI more likely[36] whereas serum CPK, hepatitis serologies or liver imaging can help determine whether an alternative cause of liver injury is more likely. Expert opinion can also incorporate known pharmacological attributes of the study drug such as a prolonged serum or biological half-life of the suspect drug, as well as histopathologic correlates if biopsy data are available. Finally, experts will consider underlying complexities in the patient population being studied that may impact liver test findings such as the high prevalence of hepatic macrovesicular steatosis and abnormal liver biochemistry profiles in patients with diabetes mellitus receiving an investigational agent.[36]

The United States Drug-Induced Liver Injury Network (DILIN) has proposed a five-point categorical scale to grade the severity of a suspected DILI episode varying from 1) asymptomatic laboratory abnormalities to 5) representing death or the need for liver transplantation.[37] The late Hyman Zimmerman noted that the development of jaundice (ie total bilirubin >2.5 mg/dL) in patients with acute hepatocellular liver injury defined as a serum ALT > 3× ULN caused by a drug is associated with an estimated mortality of 10%.[38] This observation (known as Hy's law) has been retrospectively confirmed in post-marketing surveillance studies of patients with drug-induced hepatocellular injury associated hyperbilirubinemia.[39] When present, Hy's Law cases have proven to be of value to identify drugs that have substantial hepatotoxicity liability and a significant potential to cause ALF post-marketing.[40]