Feasibility of effect-controlled clinical trials of drugs with pharmacodynamic hysteresis using sparse data

To explore, by simulation procedures, the feasibility of characterizing, from sparse data, the concentration-effect relationship of drugs with pharmacodynamic hysteresis.

For computer simulations, the concentration-effect relationship was assumed to be describable by the Sigmoid-Emax equation, the site of drug action was located in a distinct effect compartment (keo = 10 x kelim), and the pharmacokinetics were those of either a linear one- or two-compartment system. In view of the poor estimability of the parameters of the Sigmoid-Emax model under the usual clinical conditions, central compartment post-distributive drug concentrations required to elicit various intensities of effect within the therapeutic range were used as data descriptors. Effect intensities of 5 and 25, or 25 and 50 units (with the "unknown" Emax = 100 units) were targeted in multiple-dose (steady state) trial designs. From these data, drug concentrations required to produce effect intensities of 15 and 50 units were estimated by both log-linear and linear interpolation and the actual effect intensities produced by these concentrations were calculated. These simulations were performed over a wide range of Hill coefficient values (0.5 to 4.0) and dosing intervals (0.1 to 1.5 x elimination t1/2.

Acceptable results could be obtained by measuring drug concentrations and effect intensities at or near the end of a dosing interval. The largest deviations of effective concentration estimates (in terms of effect intensity) occurred at a Hill coefficient value of 0.5 and the results were very little affected by changing the dosing interval.

Our results demonstrate that effect-controlled clinical trials, with sparse data, of drugs with pharmacodynamic hysteresis for determining concentration-effect relationship in the therapeutic range are feasible in principle.