A decision tree model for predicting high mono-N-desethylamiodarone concentrations and reducing tissue toxicity in patients with low-dose amiodarone therapy: A multicentral retrospective cohort study.

ABSTRACT

OBJECTIVE: High plasma levels of mono-N-desethylamiodarone (MDEA), an active amiodarone metabolite, may be associated with tissue toxicity in heart failure (patients with heart rhythm disturbances); therefore, a tool that can identify patients for whom therapeutic drug monitoring (TDM) of MDEA is required. This multicenter study aimed to develop a decision tree (DT) model that can identify patients with heart rhythm disturbances at high MDEA concentrations. MATERIALS AND METHODS: A multicenter retrospective cohort study was conducted, including 157 adult patients with heart failure who received oral amiodarone treatment. A χ2 automatic interaction-detection algorithm was used to construct a DT model. In the DT analysis, the dependent variable was set as an MDEA trough plasma concentration of ≥ 0.6 µg/mL during the steady-state period. Explanatory variables were selected as factors with p < 0.05 in multivariate logistic regression analysis. RESULTS: The adjusted odds ratios for the daily dose of amiodarone and body mass index were 1.01 (95% coefficient interval 1.008 - 1.021, p < 0.001) and 0.91 (95% confidence interval 0.834 - 0.988, p = 0.025), respectively. For DT analysis, the risk of reaching plasma MDEA concentrations ≥ 0.6 µg/mL was relatively high, combined with a daily dose of amiodarone > 100 mg and body mass index ≤ 22.3 kg/m2 at 69.0% (20/29), and its trend was also detected in the sensitivity analysis. CONCLUSION: Patients taking a daily amiodarone dose > 100 mg and with a body mass index ≤ 22.3 kg/m2 warrant TDM implementation for MDEA to minimize the risk of MDEA-induced tissue toxicity.