Purpose: PEO-matrix, abuse-deterrent tablets prepared using direct compression can show non-homogenous drug distributions resulting in regions of high drug content within the tablet. This may compromise the abuse-deterrent properties of ADF tablets following manipulation or size reduction during attempts to recover the drug. To assess the potential for abuse of oxycodone-containing abuse-deterrent tablets (ADFs), a PBPK model was developed to describe the absorption and tissue distribution of oxycodone following nasal inhalation of granules formed following the grinding of a prototype ADF tablet. Simulated plasma and brain concentration-time profiles for the nasal inhalation of granules formed from manipulated tablets were used to predict the likelihood for rapid oxycodone absorption with sufficient amounts of drug absorbed to elicit euphoria or analgesia. Methods: A whole-body PBPK model was first developed in rats for i.v administration of oxycodone. Simulations of plasma, brain, unbound plasma, and unbound brain concentrations were conducted and compared to experimental results from literature reports. The rat PBPK model was scaled to humans using organ-specific parameters, and simulated plasma concentrations were compared to reported experimental results following i.v administration of oxycodone. The PBPK model for humans following i.v oxycodone administration was extended to include a compartment describing nasal administration and to describe absorption following the nasal inhalation of drug-containing ADF granules. Oxycodone nasal absorption was estimated for different sizes of granules formed following tablet grinding ( > 850 µm, 850-500 µm, 500-250 µm, and < 250 µm), and the resulting plasma and brain concentration-time profiles were used to identify granule fractions which may pose a risk for abuse. Results: The model simulations were able to predict observed rat and human plasma concentrations following i.v administration. The extended model (nasal absorption) was also able to capture observed plasma concentrations in humans following nasal inhalation of ADF granules. Predicted brain concentrations of oxycodone were higher than the corresponding plasma concentrations at each time point, a result of the estimated brain Kp values being >1. The plasma concentration-time profiles generated using the model showed a higher Cmax and shorter Tmax for smaller particles compared to larger granules. Although the amount of fine particles obtained following grinding an ADF tablet is much less than that obtained from the larger granules, the simulation results suggest that if an abuser attempts to grind ADF tablets and inhale only the fine particles, there is sufficient drug present to achieve euphoric effects. Conclusion: The PBPK model provides a quantitative understanding of oxycodone pharmacokinetics in the plasma and brain that will help to better understand the potential effects of oxycodone following nasal inhalation. The prototype tablets prepared for these studies showed extended-release properties with high mechanical strengths, but due to the higher drug content and more rapid dissolution of the smaller granules formed after tablet grinding, these particles were associated with significant drug absorption following nasal inhalation. These smaller, drug enriched particles provide a greater risk for abuse and reduce the abuse-deterrent features of the ADF tablets prepared by direct compression. Evaluation of granule particle sizes formed following ADF tablet grinding along with evaluating the drug content and release behaviors of the particles most likely to be inhaled ( < 500 µm) could improve the identification of drug products able to maintain or those failing to maintain their abuse-deterrent properties following tablet manipulation.
