Machine learning-based quantitative prediction of drug exposure in drug-drug interactions using drug label information

Many machine learning techniques provide a simple prediction for drug-drug interactions (DDIs). However, a systematically constructed database with pharmacokinetic (PK) DDI information does not exist, nor is there a machine learning model that numerically predicts PK fold change (FC) with it. Therefore, we propose a PK DDI prediction (PK-DDIP) model for quantitative DDI prediction with high accuracy, while constructing a highly reliable PK-DDI database. Reliable information of 3,627 PK DDIs was constructed from 3,587 drugs using 38,711 Food and Drug Administration (FDA) drug labels. This PK-DDIP model predicted the FC of the area under the time-concentration curve (AUC) within ± 0.5959. The prediction proportions within 0.8–1.25-fold, 0.67–1.5-fold, and 0.5–2-fold of the AUC were 75.77, 86.68, and 94.76%, respectively. Two external validations confirmed good prediction performance for newly updated FDA labels and FC from patients’. This model enables potential DDI evaluation before clinical trials, which will save time and cost.

Kratom Alkaloids: Interactions With Enzymes, Receptors, and Cellular Barriers

Parallel to the growing use of kratom, there is a wealth of evidence from self-report, preclinical, and early clinical studies on therapeutic benefits of its alkaloids in particular for treating pain, managing substance use disorder, and coping with emotional or mental health conditions. On the other hand, there are also reports on potential health risks concerning kratom use. These two aspects are often discussed in reviews on kratom. Here, we aim to highlight specific areas that are of importance to give insights into the mechanistic of kratom alkaloids pharmacological actions. This includes their interactions with drug-metabolizing enzymes and predictions of clinical drug-drug interactions, receptor-binding properties, interactions with cellular barriers in regards to barrier permeability, involvement of membrane transporters, and alteration of barrier function when exposed to the alkaloids.

Model-based comparative analysis of rifampicin and rifabutin drug-drug interaction profile

Rifamycins are widely used for treating mycobacterial and staphylococcal infections. Drug-drug interactions (DDI) caused by rifampicin (RIF) is a major issue. We used a model-based approach to predict the magnitude of DDI with RIF and rifabutin (RBT) for 217 cytochrome P450 (CYP) substrates. On average, DDI caused by low-dose RIF were twice more potent than those caused by RBT. Contrary to RIF, RBT appears unlikely to cause severe DDI, even with sensitive CYP substrates.

A Model for Predicting the Interindividual Variability of Drug-Drug Interactions

Pharmacokinetic drug-drug interactions are frequently characterized and quantified by an AUC ratio (Rauc). The typical value of the AUC ratio in case of cytochrome-mediated interactions may be predicted by several approaches, based on in vitro or in vivo data. Prediction of the interindividual variability of Rauc would help to anticipate more completely the consequences of a drug-drug interaction. We propose and evaluate a simple approach for predicting the standard deviation (sd) of Ln(Rauc), a metric close to the interindividual coefficient of variation of Rauc. First, a model was derived to link sd(Ln Rauc) with the substrate fraction metabolized by each cytochrome and the potency of the interactors, in case of induction or inhibition. Second, the parameters involved in these equations were estimated by a Bayesian hierarchical model, using the data from 56 interaction studies retrieved from the literature. Third, the model was evaluated by several metrics based on the fold prediction error (PE) of sd(Ln Rauc). The median PE was 0.998 (the ideal value is 1) and the interquartile range was 0.96-1.03. The PE was in the acceptable interval (0.5 to 2) in 52 cases out of 56. Fourth, a surface plot of sd(Ln Rauc) as a function of the characteristics of the substrate and the interactor has been built. The minimal value of sd(Ln Rauc) was about 0.08 (obtained for Rauc = 1) while the maximal value, 0.7, was obtained for interactions involving highly metabolized substrates with strong interactors.