The influence of drug properties and host factors on delayed onset of symptoms in drug-induced liver injury.

Abstract

Most patients with drug-induced liver injury (DILI) manifest clinical symptoms while on therapy, while some patients manifest days or weeks after drug cessation (delayed onset). This challenges DILI causality assessment and diagnosis. Factors contributing to the delayed onset phenotype are currently unknown. We explored factors contributing to delayed onset of DILI by analysing culprit drug properties, host factors and their interactions in a large patient population from the Spanish DILI Registry. Clinical information from 388 patients (69 presented delayed onset) and drug properties of 43 causative drugs (45 active ingredients) were analysed. A two-tier regression-based model was used to assess host/drug interactions affecting the probability of delayed onset. Antibacterial and anti-inflammatory drugs accounted for the delayed onset cases. Drug property of <50% hepatic metabolism (odds ratio [OR] 11.06, 95% confidence interval [95% CI]: 4.4-32.2, P=0.0003), daily dose ≥1000mg (OR: 2.77, 95% CI: 1.3-6.1, P=0.0063) and the absence of pre-existing conditions in a patient (OR: 2.55, 95% CI: 1.3-4.9, P=0.0043) were independently associated with delayed onset. The findings were consistent when externally validated using Latin American DILI Network cases (N=131). Likewise, drug properties of mitochondrial liability and Pauling electronegativity were associated with delayed onset, but dependent on specific host factors such as age, sex and pre-existing cardiac diseases. This study demonstrated that delayed onset, a specific DILI phenotype, is explained by complex interactions among drug properties and host factors and provided mechanistic hypotheses for future studies. These findings can help improve the diagnostic capability and causality assessment.