(568d) Lab Scale Experimental Workflow to De-Risk Production Equipment Change for Filtration and Drying of Active Pharmaceutical Ingredients

As an Active Pharmaceutical Ingredient (API) moves through development, the scale and site of manufacture is subject to change. At AstraZeneca we see significant production scale equipment changes between our pilot plant and commercial scale facilities. Particularly for the API stage, a change in the filtration and drying equipment used at production scale can result in changes to the quality attributes of the API. These quality attributes include particle size and shape, and the range of acceptability on secondary processing is often difficult to quantify. Therefore the preference is to maintain the particle properties within the range we have manufactured before. However, this leads to uncertainty whether we can achieve this goal when making substantial changes to the filtration and drying equipment, for example moving from an agitated filter dryer to centrifuge and spherical dryer.

One approach to de-risk equipment change has been to manufacture large scale trial batches using the new equipment and use the material in formulation manufacture to check that quality attributes are met and no operability issues are encountered. However, this requires the use of expensive intermediates, often with restricted availability, as well as requiring significant manufacturing time and effort in both the API and formulation manufacturing plants.

There is a range of filtration and drying equipment used at production scale and it would not be practical to have a scaled down version of each type of filter and dryer in our pilot plant.

Another significant consideration is that once the large scale trial has been performed, timelines are often too short to make an equipment change based on the results as the accommodation process for that compound into the particular plant has already begun.

Therefore the first step is to identify the key changes in the characteristics of the type of equipment being used. For example when moving from a pressure filter to a centrifuge the key change would be the increase in applied pressure across the cake due to the centrifugal force, potentially leading to cake compression and lump formation. Or, when moving from agitated filter dryer to spherical dryer, the type and duration of agitation is changed which could lead to agglomeration or attrition.

An experimental workflow has been designed to assess the impact of different filtration and drying equipment types at lab scale, using 200-500g of API. The material can then be blended with excipients using a formulation process such as roller compaction or wet granulation, followed by compaction using a tablet simulator. This allows characterisation of the API, the blend and final product form and any changes to product quality attributes can be identified.

One example of the use of this workflow is an API that had previously been isolated on an agitated filter dryer. The API was known to be prone to attrition so was minimally agitated during drying in the pilot plant. The manufacture of this API was then transferred to a commercial facility where a pressure filter followed by spherical dryer would be used so there was concern that significant attrition would be observed and therefore the particle size would be smaller than any of the previously manufactured batches. It was also not known if agglomeration would be observed.

We do not have a lab scale spherical dryer, but the key characteristics of rapid drying with continuous agitation were tested in a lab scale agitated filter dryer and also a conical dryer. It was found that some attrition was observed during the drying stage, and this was broadly comparable between the continuously agitated pressure filter and the conical dryer. Despite agitating for an extended period of time, the particle size did not drop below the size of smallest batch manufactured so far. It was also noted that there was a particle breakage in the pressure filtration - there was a decrease in particle size between the sample taken from the crystallised slurry and a sample taken from the wet cake. The material was then filled into capsules and dissolution testing performed. This experiment allowed us to show that the quality attributes of the API and product were maintained despite changing the drying equipment type. Samples taken from the manufacturing campaign showed that the particle size decreased during pressure filtration and again during spherical drying and the particle size of the dried material was within the range previously observed. This experimental protocol was repeated for a second API but no attrition was observed at lab scale and this was matched by observations at commercial scale, again in the spherical dryer.

This approach is currently being embedded into late stage development of API processes and will lead to more robust technical transfers of processes to our commercial facilities.