Causality assessment of liver injury after chronic oral amiodarone intake.

摘要

BACKGROUND/AIM: The number of patients receiving amiodarone will increase in future years. As clinically significant hepatotoxicity associated with oral amiodarone is infrequent and difficult to predict, a new Bayesian-developed model is proposed to help in the causality assessment of amiodarone-induced liver injury. METHODS: Incidence of abnormal liver enzymes in patients receiving amiodarone was obtained from placebo controlled clinical trials. Published case reports of amiodarone-induced hepatotoxicity were identified through a literature search. Maximum number of expected hepatotoxicity cases in amiodarone and placebo-treated patients was calculated using Poisson distribution. The calculated odds ratio was used as a Prior Odds (PrO) to subsequent quantification, using a Bayesian-approach, of individual amiodarone-induced hepatotoxicity likelihood. RESULTS: PrO of amiodarone-induced hepatotoxicity was 0.48. Thirty nine amiodarone-associated hepatotoxicity case reports were retrieved. Half of published case reports developed an irreversible damage. The amiodarone Bayesian model combining information about latency period and period of remission, together with analytical parameters properly defines the toxicity profile shown in published case reports. The analytical pattern defined by this model is different from the one expected if liver injury in published cases was caused by other etiologies. CONCLUSIONS: A method based on a Bayesian-approach, which links information from clinical trials with clinical hepatotoxicity profile from published case reports can be a useful tool for amiodarone-induced liver injury causality assessment. At present, this method is limited due to scarcity and quality of available data. Further efforts are needed to improve model ability in order to identify amiodarone-induced liver injury.