(M1130-10-67) Model Based Analysis for Determining Bioequivalence of Metronidazole Topical Cream

Purpose: There is an interest in the regulatory field to establish a reliable and validated approach to assess the bioequivalence (BE) of topical products to expedite regulatory and market approval. Tape stripping (TS) is a relatively non-invasive technique employed to study the dermatopharmacokinetics (DPK) of topically applied drugs and cosmetics within the stratum corneum (SC) which involve tedious experimental procedures and complicated calculations and assumptions. The objective of this study was to evaluate and propose a model-based approach based on DPK data for a more direct estimation of the rate and extent of “absorption” or input into the skin for bioavailability (BA)/BE assessment, thereby circumventing the need for numerous measurements and assumptions. This approach involves an improved skin-stripping procedure and a model-based assessment of the rate and extent of “absorption” or “delivery” of active ingredients through the skin using a compartmental PK model that aims at mimicking what happens following application to human skin.

Methods: The Reference (Ref) product, Metrocreme®, containing 0.75% metronidazole was compared to a Test product that was equal in strength. A total of 10 subjects were included in the study and received 15 mg of Ref and Test treatments in a fully replicate design. After each 60 min topical cream exposure (predetermined in a pilot dose-duration study to be associated with half of the total absorption), the remaining product was removed from each site, which were then sequentially stripped 20 times. Extraction and chromatographic analysis of each tape strip was conducted to determine the amount of drug that penetrated each layer of skin. A population PK (PPK) analysis was conducted with the maximum likelihood expectation maximization (MLEM) algorithm as implemented in ADAPT®5 V5.053. A 20-compartment model (Figure 1) was developed where the two DPK parameters that may be of importance for BE assessment would be K_in (i.e., rate) and F_s (i.e., extent), where K_in is the first-order absorption/input rate constant and F_s is the %absorbed/delivered amount of applied drug dose from the cream into the skin¹. BE was assessed using the average BE approach and T/R geometric mean ratios (GMR) and their associated 90% confidence intervals (CIs) were calculated.

Results: Multiple PPK models were investigated until the most appropriate one was identified (model in Figure 1 and goodness-of-fit in Figure 2). The selected model was verified by comparing the innovator formulation with itself (Ref vs. Ref) as a positive control (Table 1). Results suggests that, with a minimum of 12 subjects, the model would be successful at identifying BE in terms of extent of absorption/delivery between the Test and Ref formulations, and between the Ref and itself (F_s, Table 1). Rates of absorption/delivery were also estimated by the model for all formulations, but due to the high inherent variability of this PK metric, equivalence criteria of 50-200% (see K_in Ref vs. Ref 90% CI, Table 1) will be proposed to likely reflect clinical relevance.

Conclusion: The proposed model-based approach and DPK methodology were shown to accurately assess BE of an innovator product when compared with itself and when compared to an equivalent strength of a Test product with respect to the extent of exposure (i.e., F_s). These results are congruent with conventional methods of analysis using this tape stripping data².

References: 1. D. Ozdin, I. Kanfer and M. P. Ducharme, "Novel Approach for the Bioequivalence Assessment of Topical Cream Formulations: Model-Based Analysis of Tape Stripping Data Correctly Concludes BE and BIE," Pharmaceutical Research , vol. 37, no. 20, 2020.
2. S. Rath, A. Ramanah, C.Bon, I.Kanfer, “Application of dermatopharmacokinetic (DPK) method for bioequivalence assessment of topical metronidazole creams,”J Pharm Sci, vol. 23, pp. 437-450, 2020.

Acknowledgements: The authors gratefully acknowledge the funding received for this project from Learn and Confirm Inc., Saint Laurent, QC, Canada and Mitacs, Canada (Mitacs Accelerate Postdoctoral Fellowship) .

Figure 1. Final Population PK Model

Figure 2. Goodness-of-fit plots for the final population PK model for a) observations vs. individual fitted amount and b) for the population model predictions. From left to right for each panel i) when fitting Ref with Ref, ii) when fitting Test with Test and iii) when fitting Test with Ref

Table 1. Summary of BE assessment from Population Analysis Predictions When Comparing 0.75% Metronidazole