(T0930-02-12) A PBPK Model to Simulate Intranasal Naloxone Rescue from Fentanyl Overdose

Purpose: To construct a physiologically based pharmacokinetic model of dispositions of fentanyl given IV and naloxone given IV and IN (intranasal), their interactions with mu-opioid receptors (MORs) that mediate adverse fentanyl effects such as respiratory depression and simulate various overdose and rescue scenarios for a hypothetical patient population.

Methods: Preliminary assessments of published clinical data for fentanyl and naloxone administered IV were made using PKSolver® to verify both drugs followed 2-compartment models. A third compartment was added for the brain and included concentration of each drug vs. time for the free and bound (to MORs) forms. The kinetics of binding and release of each drug at the MORs was modeled. A relation between the fraction of MORs occupied by fentanyl and the ventilation depression was constructed by fitting the overall model to clinical respiratory depression data vs. time and the associated fentanyl concentration vs. time in the central compartment. The model equations were written and evaluated using the R programming platform. Estimates of “typical” values for 28 parameters (compartment volumes, rate constants, etc.) were obtained by fitting the models to published data. A hypothetical patient population was constructed by varying the 28 parameters about these typical values using a truncated log normal distribution, and simulations for the population were performed to obtain 23 outputs (compartment concentrations, fraction of MORs occupied by fentanyl, ventilation suppression, etc.) were obtained as average and standard deviations at each outputted time. Simulations were done for various IV doses of fentanyl (300-2000 mcg) and 4mg naloxone IN rescue doses at 5 minutes, 5 and 8 minutes, and 5, 8 and 11 minutes.

Results: A ventilation fraction of 0.4 or more was taken as successful rescue level for comparison purposes. The simulations indicated that naloxone rescue was not needed in almost all simulated patients for fentanyl 300 mcg IV, while rescue was indicated for some patients after fentanyl 500 mcg IV. For fentanyl 1000 mcg IV, multiple doses of naloxone 4mg were indicated but not sufficient for some patients in the population, while rescue was ineffective even after three 4mg doses of naloxone following fentanyl IV 2000 mcg (sometimes taken as a typical lethal dose). It was also found that the time from the first naloxone dose to when the ventilation reached 40% of its full value was critically dependent on the ventilation suppression at the time of the first naloxone dose. The data indicated that the fentanyl release from the MORs was high enough to facilitate reasonably rapid enough fentanyl release from the MORs (release rate constant ~200 h^-1), and the binding rate constant at the MOR for naloxone was greater than that of fentanyl (~10⁴  M^-1h^-1 vs. 10² M^-1h^-1) so the sites opened by release of fentanyl were preferentially filled with naloxone.

Conclusion: The calculated kinetic parameters for binding at the MOR sites indicated that naloxone has a higher binding rate constant than fentanyl for MOR sites, and favorably competes for the binding sites left open as fentanyl is released.  It was concluded that the model and simulations can be successfully used to predict rescue from various fentanyl overdose scenarios resulting from various intranasal naloxone rescue regimens.


Figure 1. 500 mcg fentanyl IV and IN naloxone 4 mg at 5 minutes


Figure 2. 2000 mcg fentanyl IV and IN naloxone 4 mg at 5m 8 and 11 minutes