Fentanyl-induced Ventilatory Depression: Population Pharmacokinetic-Pharmacodynamic Framework for Evaluation of...

Opioids are associated with potentially severe ventilatory depression. However, in the perioperative setting, when opioids are combined with anesthetics, there is little information about their "isolated" effects on ventilatory control.

In 12 healthy volunteers of either sex, this study tested the effect of five clinical doses of fentanyl (initial dose of 100 μg, followed by 75 μg at t = 30 and 60 min and 50 μg at t = 90 and 120 min). The study measured minute ventilation, end-tidal CO 2 concentrations, and arterial fentanyl concentrations during the sham procedure and for another hour. Pharmacokinetic-pharmacodynamic modeling was conducted using three distinct models: a model describing just ventilation, one describing just end-tidal CO 2 concentrations, and a physiologic model in which ventilation and end-tidal CO 2 concentrations were modeled simultaneously and closed-loop conditions and CO 2 kinetics were considered. The hypothesis was that the physiologic model yields the most reliable estimation of fentanyl's ventilatory potency. The parameter values are median ± standard error of the estimate.

The value of the fentanyl potency parameter (the steady-state fentanyl concentration causing 50% depression of ventilation) derived from the physiologic model, was less than one third of those derived from the separate models for : 2.3 ± 0.5 versus 7.5 ± 1.3 ng/ml (typical value ± standard error of the estimate). Few other differences in mutual parameter values between models were observed. Parameters exclusive to the physiologic model were the gain and time constant of the ventilatory controller, with values of 5.3 ± 1.4 l · min -1 · kPa -1 and 2.4 ± 1.4 min, respectively, and tissue volume (6.1 ± 1.2 l).

In this exploratory and hypothesis generating study, the pharmacodynamic model that incorporates CO 2 kinetics and a ventilatory controller produced a fentanyl potency estimate that was lower than that estimated from simpler models lacking these components, a value that can be considered pharmacologically realistic and clinically relevant.