Purpose: API bulk properties can have a significant effect on the formulation of drug product and an understanding of quality attributes (QAs) such as particle size, surface area, morphology, and powder flow is essential to development. Implementation of a new crystallization process and/or a change in processing equipment can often cause changes to the API bulk properties. In this work, we demonstrate the importance of correlating particle size, surface area, morphology (via scanning electron microscopy), and powder flow to gain a comprehensive understanding of API bulk properties. Three case studies (GS-X, GS-Y, and GS-Z) are presented where all four of the bulk property characterization techniques played a critical role in identifying API QAs that impacted the manufacturability of the drug product.
Methods: Particle size distribution (PSD) was determined using a Malvern 3000 in the wet dispersion mode with 0.1% Span 85 in heptane as the dispersion fluid. A stir speed of 2000 rpm was used with no sonication. The obscuration was between 10-20% and the sample was stirred for 10 minutes to 1 hour until the change in obscuration level stabilized (pre-measurement delay was sample dependent) before data collection. Fraunhofer approximation method was used to analyze the data. Surface area was performed using a Tristar 3000 instrument from Micromeritics. Nitrogen was used as the inert gas and the measurements were performed at 77 K and at 5 partial pressures. Scanning electron microscopy (SEM) was performed using a Hitachi Tabletop Microscope. Samples were sputtered with gold to increase conductivity and images were taken at various magnifications. Powder flow was investigated using a Schulze Ring Shear Tester RST-XS with normal load pre-shear of 1000, 3000, and 5000 Pa and normal loads at shear of 1250, 1500, 2250, 3000 Pa.
Results: GS-X. Heavy sticking and bridging was observed in the roller compaction feeder while formulating API from the current Generation 3 (G3) process at manufacturing site B. This had not been seen using API produced at manufacturing site A and manufacturing site C, both running the same G3 process. Furthermore, there had been no previous issues when formulating with API from the Generation 2 (G2) process manufactured at all three sites. Comparison of API lots derived from each process and each site indicated that differences in bulk properties were linked to the manufacturing site and not to the API process. The G3 API from manufacturing site B was found to have poor flow compared to G2 API, as well as other G3 lots manufactured at other sites. This poor flow was attributed to its small PSD, large surface area, and resulting morphology.
GS-Y. When several drug product batches failed to meet dissolution specifications for GS-Y, failure was correlated to API lots generated at manufacturing site C. Previous campaigns using API from manufacturing site A and manufacturing site B (running the same process) showed no issues. API from manufacturing site C was found to have poor flow compared to those of API from manufacturing site A and manufacturing site B, and also qualitatively showed poor wettability as a function of dispersion during PSD measurements. The poor flow of API from manufacturing site C was attributed to its smaller PSD and larger surface area when compared to API from manufacturing site A. However, although API batches from manufacturing site B and manufacturing C were found to have similar PSD and surface area values, they had significantly different flow properties and this was attributed to morphology differences based on SEM images.
GS-Z. Several lots of GS-Z had been characterized by particle size and surface area, but the data resulted in unexpected trends. Crystallizations from three different solvent systems were examined with and without milling. Early data from polarized light microscopy indicated that the particle morphology from two of the solvent systems looked similar, but when milled the increase in surface area was dramatically different. Imaging the particles with SEM showed that one solvent system produced porous crystals while the other solvent system resulted in smooth faced crystals. Investigation of both lots indicated that the porous crystals displayed a smaller D10 and D50 after milling with a corresponding increase in surface area attributed to increased fines. Although differences in the processing conditions caused dramatic changes in the surface area, particle size and morphology for the resulting materials, they were all found to be easy flowing to cohesive and thus suitable for formulating drug product.
Conclusion: Particle size, surface area, morphology and powder flow are bulk properties that can be measured to gain insight into API performance during drug product formulation. When the data from all four techniques are correlated, it provides more understanding of the API quality attributes than by using any of the individual techniques alone. These quality attributes can thus help to proactively guide decision making and provide a range of values to improve robustness in drug product manufacturing.
Henry Morrison– Gilead Sciences, Inc.
Henry Morrison– Gilead Sciences, Inc.
Olga Lapina– Gilead Sciences, Inc.
Sean Ritchie– Gilead Sciences, Inc.
Ernest Lee– Gilead Sciences, Inc.
Philippa Payne– Gilead Sciences, Inc.