AZD-1305

The objectives were to estimate and compare, in silico and in vivo, the effects of a strong and a moderate CYP3A4 inhibitor on AZD1305 pharmacokinetics. In silico, simulations were performed with the computer software Simcyp, and the predicted outcome was compared with the results observed in healthy male subjects. In silico, the geometric mean plasma exposure of AZD1305 + ketoconazole showed a 7.1‐fold higher AUC and a 4.4‐fold higher Cmax compared with AZD1305 alone. Coadministration with verapamil gave a 1.9‐fold higher AUC and a 1.7‐fold higher Cmax compared with AZD1305 alone. In vivo, the plasma exposure of AZD1305 + ketoconazole showed a 7.7‐fold higher AUC and a 4.8 ‐fold higher Cmax compared with AZD1305 alone. Coadministration with verapamil gave a 2.2‐fold higher AUC and a 2.0‐fold higher Cmax compared with AZD1305 alone. The mean maximum QTcF increase from baseline was 407, 487, and 437 milliseconds for AZD1305, alone and in combination with verapamil or ketoconazole, respectively. Simcyp predicted the effects of ketoconazole and verapamil on the sensitive CYP3A4 substrate AZD1305 pharmacokinetics well. Both the in vivo study and the Simcyp predictions suggest a contraindication for strong CYP3A4 inhibitors and AZD1305 when given in combination.

Pharmacokinetic-pharmacodynamic (PKPD) modelling can improve safety assessment, but few PKPD models describing drug-induced QRS and PR prolongations have been published. This investigation aims to develop and evaluate PKPD models for describing QRS and PR effects in routine safety studies.

Exposure and telemetry data from safety pharmacology studies in conscious beagle dogs were acquired. Mixed effects baseline and PK-QRS/PR models were developed for the anti-arrhythmic compounds AZD1305, flecainide, quinidine and verapamil and the anti-muscarinic compounds AZD8683 and AZD9164. RR interval correction and circadian rhythms were investigated for predicting baseline variability. Individual PK predictions were used to drive the pharmacological effects evaluating linear and non-linear direct and effect compartment models.

Conduction slowing induced by the tested anti-arrhythmics was direct and proportional at low exposures, whilst time delays and non-linear effects were evident for the tested anti-muscarinics. AZD1305, flecainide and quinidine induced QRS widening with 4.2, 10 and 5.6% μM(-1) unbound drug. AZD1305 and flecainide also prolonged PR with 13.5 and 11.5% μM(-1). PR prolongations induced by the anti-muscarinics and verapamil were best described by Emax models with maximal effects ranging from 55 to 95%. RR interval correction and circadian rhythm improved PR but not QRS modelling. However, circadian rhythm had minor impact on estimated drug effects.

Baseline and drug-induced effects on QRS and PR intervals can be effectively described with PKPD models using routine data, providing quantitative safety information to support drug discovery and development.