All Times EDT
Keywords: COVID, Bioequivalence, Estimand
COVID-19 public health emergency has posed challenges to the conduct of bioequivalence (BE) studies in human participants due to travel limitation, study site closures, batch availability, subjects missing sampling times, protocol revisions, and other issues. The BE studies can be suspended at various stages ranging from protocol development, recruitment, screening, during treatment, or in between study periods for a crossover study. All of these pose statistical challenges to support demonstration of bioequivalence and approval of an Abbreviated NDA (ANDA) study. As a result, FDA/CDER (2020) recommends alternative study design and statistical approaches to assess BE. For example, applicants are encouraged to prospectively identify how the study will handle participant illness (either from COVID-19 or other reasons) during the conduct of the study, and prospectively state how missing data will be handled and the statistical analyses as well as justifying the methods chosen; Applicants are also encouraged to contact the Agency early for any protocol or statistical analysis plan (SAP) amendment, which should be made prior to data lock and unblinding; Adaptive design is encouraged in order to adapt to protocol amendment; In order to reduce the number of visits, CDER also recommends alternative PK approaches such as partial area under the curve (pAUC) rather than full AUC without compromising quality of the data. In particular, in this talk a de-identified COVID-impacted PK BE ANDA study is given as an example to showcase what is the impact on BE assessment in a four-way replicated crossover pharmacokinetic (PK) BE study prolonged by COVID-19 lockdown, how the clinical question was re-defined in this COVID-impacted PK BE study, and how the estimand and statistical approaches were revised in order to avoid inflation of Type 1 error and bias in BE determination.