Estimation of relative potency in a 4-point, parallel line assay

For inhaled asthma medications, the current EMA guideline requires equivalence in respect of efficacy to be demonstrated also through a comparison between drugs in terms of relative potency. Differently from the more common comparison of the magnitude of responses observed following administration of different products, the doseresponse relationship is taken into account by this approach. In this paper, we discuss the rationale of a two-step analysis to estimate relative potency and its confidence limits in a 4-point assay, when the same two dose levels of each product being compared are administered. This is the typical design for the comparison of different formulations of the same medication. The analysis is based on the assumption of a parallel line assay, i.e. linear and parallel doseresponse curves on log-dose scale. The development of a SAS® macro to handle this analysis in both parallel group and cross-over designs will also be described. INTRODUCTION In the current EMA Guideline on the Requirements for Clinical Documentation for Orally Inhaled Products [1], two different approaches for the demonstration of therapeutic equivalence in respect of efficacy of inhaled asthma drugs are described: demonstration of equivalence for at least two dose levels on the pharmacodynamic endpoint and analysis of relative potency. The relative potency of the test product to the reference product is defined as the dose of the test product that produces the same biological response as one unit of the dose of the reference product. The first approach is based on the usual comparison on the response axis, while the second one focuses on the dose axis, allowing the dose-response relationship to be taken into account. For either approach to be acceptable a minimum requirement is that the study has assay sensitivity, therefore at least two non-zero dose levels of both the test and reference products need to be studied and one dose level needs to be show to be superior to the other. In this paper, the analysis of relative potency will be illustrated in a 4-point assay, when the same two dose levels of each product being compared are administered. This is the typical design for the comparison of different formulations of the same medication, for example a dry powder inhaler and a pressurized metered-dose inhaler. The analysis is based on the assumption of a parallel line assay, i.e. linear and parallel dose-response curves on log-dose scale. A continuous response is also assumed. A two-step approach to estimate relative potency and its confidence limits in both parallel group and cross-over designs will be discussed and a SAS macro to implement this method will be presented. ESTIMATION OF RELATIVE POTENCY All the following considerations regarding the point and interval estimation of relative potency apply to all the parallel line assays and are not limited to the 4-point design.