[learn_more caption=”Jeffrey N. Hemenway”]

Gilead Sciences Inc., Foster City, CA, USA

Jeffrey Hemenway, Ph.D. is currently a Senior Scientist in Formulation and Process Development at Gilead Sciences Inc. He received his bachelor’s in Chemistry from University of Missouri, Kansas City, and his doctorate in Pharmaceutical Chemistry from the University of Kansas, where he was trained in the area of prodrug design and characterization. He joined the Biopharmaceutics Department at Bristol Meyers Squibb, where he worked in the areas of solid oral formulation and manufacturing process development, scale up and tech transfer for early and late stage programs. He is a member of AAPS and currently serves as the Chair Elect for the FDD Section. In his current position at Gilead Sciences, he leads a team of scientists responsible for discovery pharmaceutics, preformulation, formulation and manufacturing process development, scale up, outsourced GMP manufacturing and CMC regulatory documentation for preclinical, early and late stage programs.[/learn_more]

Abstract:

Presented here is an overview of the practical application of compaction simulation to assess fundamental mechanical properties and evaluate high speed press replication performance for final powder blends from designed experiments and other key development batches. This example is presented as a case study for a relatively high drug loading BCS class 2 compound in a tablet formulation that is manufactured using a roller compaction process. Compaction simulation was used to study the mechanical properties of the DoE final powder blends at commercially relevant speeds. Tabletability, compactability and compressibility profiles were generated using 10 mm flat-faced tooling and a single edge sign wave profile at a maximum punch velocity of approximately 680 mm/s and 10 ms dwell time. Results were analyzed using the Design Expert, version 9 (Stat-Ease Inc., Minneapolis, MN, USA). A stepwise regression by backward elimination was used to identify significant responses with ≥ 95% confidence (i.e., p ≤ 0.05). Compression replication was used to evaluate high-speed compression performance. Punch profiles were derived for a model commercial scale press and product specific tooling using Press Profiler software version 4.0.9 (Phoenix Calibration & Services Ltd, West Midlands, UK). Compression profiles were generated using product specific tableting parameters at simulated press speeds of 75 rpm with maximum punch velocities of approximately 800 mm/s and 6 ms dwell times.

A formulation optimization study was conducted as a five factor, fractional factorial 25-1 design of experiments using a fixed roller compaction process. The design factors studies were drug loading, API powder properties (e.g. mean particle size), lubrication level, excipient A level, and excipient B grade. A predictive model was identified for the effect of formulation design factors on tabletability. The statistical analysis showed main effects for lubrication level, excipient A level and API particle size and an interaction between drug loading and excipient B grade. Additional lubrication sensitivity studies were conducted to further assess the compression risk associated with a wider range of lubrication levels and final blending times for the optimized formulation. High speed compression replication was used to evaluate the final blends from the lubrication sensitivity batches. Higher lubrication levels and longer blend times result in decreased force/hardness profiles and capping on hardness testing (w/o precompression). The addition of between 10 and 25% precompression resolved issue with capping on hardness testing and resulted in acceptable compression profiles on the simulated press at speeds up to 75 RPM (>2500 tpm) for all lubrication levels at the longest blend times. The results predict coverage of a suitable range for the commercial scale final blending process.

The roller compaction process was developed using a four-factor central composite design to study the factors of roller compaction force, roll gap, roll speed, and screen size on process and product performance. Statistical analysis of the mechanical properties of the final blends resulted in relatively weak models, likely due to the presence of complex interactions, but the overall effects of the design factors on tabletability were clearly seen. High speed replication studies for the center point and extreme over/under compacted final blends demonstrated acceptable compression performance for the process parameter ranges studied. Approximately 10% precompression was required for all strengths to eliminate tablet capping and lamination at higher compression forces for undergrannulated and center point processing conditions. The results from these studies along with other key response factors resulted in the identification of a suitable design space for the roller compaction process.