Development and Robustness Verification of an At-Line Transmission Raman Method for Pharmaceutical Tablets Using Quality by Design (QbD) Principles

Abstract

The purpose of the present study was to investigate the feasibility of developing a fast non-destructive at-line transmission Raman spectroscopy (TRS) method for core tablet potency and content uniformity (CU) as part of a real-time release testing (RTRt) control strategy. The effects of tablet hardness and weight (thickness), API particle size, and concentration were studied by using a novel experimental design called generalized subset designs (GSDs). A subset of 28 experiments plus three replicate center points were selected for a total of 31 experiments. Matrix effects included tablets with active pharmaceutical ingredient (API) at seven concentration levels (14 to 26%w/w), and API particle size (17 to 71 μm), tablet weight (275 to 328 mg), and tablet hardness (8 to 16 SCU) at two levels. Three calibration models were developed, by using partial least squares (PLS) and different preprocessing conditions (model nos. 1 to 3). In model no. 1, all matrix effects were excluded. This model showed high potency prediction errors (RMSEP of 10.0%). When all matrix variations were included in the multivariate calibration according to the GSD as shown in model nos. 2 and 3, the prediction accuracy was greatly improved (RMSEP 2.56 and 1.74% respectively). The statistical significance of the tablet weight, hardness, and API particle size in the %Recovery (TRS vs. the reference HPLC method) was investigated by using MODDE Pro (Sartorius Stedim Data Analytics). Statistically significant effects were identified if the calculated p value was ≤ 0.05. The main effect hardness, the cross-term hardness×particle size, and the quadratic term cal level×cal level showed to be statistically significant. However, these effects had a very small impact on tablet prediction accuracy (± 1.0%w/w) well within the intermediate precision of the method. A non-destructive TRS method for core tablet potency and CU was fully validated, following ICH Q2 and EMEA NIR guidelines. The applicability of the method to process development batches was demonstrated and compared to a previously developed and validated NIR method.