(M1430-06-35) Avoiding Pitfalls in Selecting Quality Control Concentrations in Surrogate Matrix Approach for Endogenous Drugs

Purpose: An adequate validation design and results are critical for demonstrating the accuracy and suitability of a bioanalytical method, thereby assuring the reliability of drug concentration data generated in bioequivalence (BE) studies, which support abbreviated new drug applications (ANDAs). In the case of drugs which also exist endogenously, the bioanalytical methods often utilize a surrogate matrix for calibration standards to avoid interference from the endogenous level of the analyte in blank authentic matrix. In such cases, the selection of appropriate concentrations of quality control (QC) samples especially low QC (LQC) can be challenging when the endogenous levels in the authentic matrix significantly exceed the lower limit of quantitation (LLOQ), resulting in a high adjusted LQC (LQC plus endogenous) concentration. Through a meta-analysis of bio-analysis part of in-house ANDA submissions for a model drug, Estradiol in dosage forms including Tablets, Transdermal Patch, Gel, Cream, Insert and Extended Release Film, this project develops recommendations for appropriate selection of QC concentrations that are reflective of study samples concentrations. Additionally, it identifies other avoidable deficiencies in the bioanalytical portion of the ANDA submissions.

Methods: ANDAs submitted to the Agency for drug products containing Estradiol from 1994 to 2021 were evaluated for their approaches to demonstrate bioequivalence. In the bioanalytical portions of these selected ANDAs, the analytical approaches (such as surrogate analyte or matrix) utilized to analyze estradiol in the study samples were collected along with other bio-analytical deficiencies issued by the Agency. In the method validation for surrogate matrix approach, which was the most common approach, the range of calibration curve standards and QC concentrations were compared to estradiol concentrations observed in BE samples. In addition, the validity of the selected surrogate matrix was assessed by evaluating further validation results, including matrix effect, recovery (from surrogate and authentic matrix) and parallelism.

Results: From the ANDAs collected, 96% utilized surrogate matrix approach. Of these, only 12% employed the artificial surrogate matrix (such as BSA) whereas the remining 88% used charcoal stripping of the authentic matrix to generate surrogate matrix.  The majority of the submissions lacked efforts on screening low endogenous estradiol levels in individual authentic matrix lots before preparation of QC samples, particularly LQC. When screening was performed, applicants pooled all lots screened instead of selectively pooling those with lower endogenous levels for QC sample preparation. Consequently, the adjusted LQC concentration was high relative to levels obtained in the study samples, resulting in high number of study samples having concentrations below adjusted LQC in authentic plasma. In the collected ANDA submissions, 91% had pre-dose concentrations lower than validated authentic LQC concentration, whereas 56% had total study sample concentrations below this level. Only 17% of the ANDAs conducted parallelism study to support the selection of authentic QC concentrations, despite the fact that pre-dose or post dose study samples were below LQC level. For ANDAs lacking parallelism studies between authentic and surrogate matrix to ensure consistent analyte response across the CS range, deficiency to demonstrate the appropriateness of selected QC concentrations was issued. In addition, other bioanalytical deficiencies observed included missing method details about derivatization of estradiol for quantification, lack of justification for use of surrogate matrix, and failure to submit recovery of analyte and internal standard in both matrices.

Conclusion: This project aims to provide guidance on selecting the appropriate matrix for establishing method accuracy and precision during the method validation of drugs that exist endogenously. It also aims to help ANDA applicants avoid unnecessary review cycles during the approval process. Based on the meta-analysis of in-house ANDA submissions for estradiol drug products, screening to find the blank plasma lot with the lowest endogenous estradiol level is crucial for adjusting the LQC in authentic matrix. This step appears pivotal for bioanalysis of in vivo BE study samples. When using surrogate matrix (stripped, diluted matrix or artificial) because of non-availability of matrix with lower endogenous levels, a parallelism study between surrogate and authentic matrix is highly recommended. Furthermore, this project also informs applicants about other commonly observed but easily avoidable bio-analytical deficiencies. The ultimate goal is to facilitate early access to safe and high quality generics to the American public.