Dose finding by concentration-response versus dose-response: a simulation-based comparison

The investigations reported here aimed to evaluate the incremental benefit for dose finding by concentration-response analysis versus dose-response analysis.

Trials were simulated using an Emax model for a range of scenarios of drug properties, trial design options and target response levels. The simulated data were analysed by concentration-response and dose-response modelling; a dose was then chosen to target a specific response level in a confirmatory trial. The two approaches were compared in terms of the quality of model parameter estimation and the success rate for the confirmatory trial.

While the accuracy for ED50 estimation was comparably good with both approaches, the precision was up to 90 % higher with concentration-response approach. The difference was most notable when clearance was highly variable between subjects and the top dose was relatively low. The higher precision by the concentration-response analysis lead to better dose selection and up to 20 % higher success rate for the subsequent confirmatory trial. The relatively small difference in success rate translated into a remarkable difference in sample size requirement.

By customising these parameters, the approach and the findings can be applied to assessing the value of pharmacokinetic sampling in particular trial situations.