Designing An Effective PAT-Driven Scale-Up Of Lyophilization Processes

Published on: 

Pharmaceutical Technology Europe

Pharmaceutical Technology Europe, Pharmaceutical Technology Europe-11-01-2010, Volume 22, Issue 11

This article examines the difficulties in designing lyophilisation processes that can be faithfully scaled up to production volumes and suggests the most effective ways in which this can be achieved.

Producing a freezedried cake with the desired characteristics — moisture content, stability and reconstitution — often requires a number of trials, each one requiring the necessary facilities, time, personnel and media (power, WFI, etc.). The media is frequently very expensive and for this reason it is vitally important to ensure that scaleup can be achieved smoothly and effectively to avoid unnecessary expense and delays in time to market.

IMAGE: WJONES

In some cases, it is possible to scaleup from a laboratory-sized lyophiliser to small production equipment in a single stage. However, scaleup processes are often required in two stages: laboratory to pilot production, and pilot production to full commercial volumes. The pilot stage is required to produce sufficient product under relevant GMP conditions for clinical validation. A rule of thumb for size categorisation is:

  • laboratory lyophiliser (<0.3 m2)

  • pilot (development) lyophiliser (from 0.1 m2 to 8 m2)

  • production lyophiliser (from 8 m2 to 55 m2).

The process design depends on the requirements of each application. Therefore, a flexible approach is required to ensure the best security of outcome.

Definition of the scale-up process

In theory, the scale-up process is very simple: the outcome from the pilot (or laboratory) production must be recreated in the production lyophiliser. For this to be achieved, however, the conditions within the product and on the product's surface must be identical. There are several influencing factors that have an enormous impact on the process:

  • Finding a formulation (soup or cocktail) with sufficient interaction between different excipients — buffers, bulking agents, preservatives, tonicity agents, stabilisers, surfactants, solubilising agents and antioxidants — is the most difficult step in lyophilisation. The intention is to avoid changing the formulation during scaleup as much as possible, but this is often difficult to achieve.

  • The vacuum profile during primary and secondary drying influences the sublimation and desorption processes. Some influencing factors are:
Advertisement

– the vapour pressure difference between the ice drying front

– the vapour pressure at ice surface in the condenser

– the methods of vacuum control used

– the instrumentation used in achieving comparable conditions for pressure regulation.

  • During drying, the product temperature can be controlled by pressure; however even then the heat transfer fluid in the shelves and cooling of the condenser has a huge impact on the performance of the process.

Other influencing factors are heat input and vapour transport out of the product; the different contact conditions from containers to the shelves owing to tolerances and surface variations; radiation and heat convection of surrounding areas; temperature gradient from product to the ice condenser; and the temperature difference within the shelves. Any one or combination of these conditions can lead to differences in the final outcome.1

As a temperaturerelated parameter, the freezing rate is also a major influencing factor on pore diameter, product quality and drying time. 2 Figure 1 shows the influence of freezing rate on the pore diameter of a dextrin solution.

Figure 1: The influence of freezing rate on the pore diameter of dextrin solution.

The influence of design parameters and dimensions

The design of the separation valve between chamber and condenser has a significant influence on the flow characteristics of the vapour. The kind of valve (mushroom, plate or butterfly), size of opening and position of this valve are all important (Figure 2).1 In addition to the separation valve, other vapour flow influencing factors have to be considered, including the position of the condenser, the design of the evaporator (pipes or plates),1 the size and the inter-distance of the shelves.

Figure 2: Design of a mushroom valve.

A number of other design parameters and factors must also be considered during the scaleup process:

  • Smaller lyophilisers use doors made partly from acryl. This material has a strong influence on the temperature at the front of the shelves caused by different heat radiation characteristics compared with stainless steel doors. To reduce this influence, aluminium foil is often placed on the inside of the door, but this has only a limited corrective effect.3

  • Pilot and development equipment is not usually subject to the same clean room conditions as production lyophilisers. This leads to a potential problem as particles contaminating the containers can affect crystallisation points and lead to different freezing behaviour.

  • The ratio of heat transfer media flow to the amount of product has an influence on the outcome; the amount of thermal energy required is dependent on the area and the amount of product to be brought to the required temperature (heating or cooling). A different ratio in mass flow will lead to a different freezing period and different temperature set points.4 Different flow conditions also have an impact on the implementation of thermal energy into the system especially within ramps (during static conditions this influence is negligible).

  • The ratio of stainless steel mass and surface area within the lyophiliser leads to different process conditions as the walls and shelves affect the conductive, radiation and convective heat transfer. Varying distances from the shelves to the chamber walls also has an effect.5

  • Behaviour during temperature and pressure ramping, the overshoots and control accuracy of set values all have an impact on product characteristics. It is necessary, for example, to establish whether the heat transfer system for the shelves is controlled by shelf inlet or outlet temperature. Similarly, the condenser can be controlled by direct expansion (which both may be cooled by LN2 or compressors) or by circulation heat transfer fluid.

  • The type, amount and sequence of collected variables mean that the data and reports generated from the control systems may not be directly comparable.

  • The instrumentation used for process control, such as product temperature probes, may be different for pilot equipment and production size lyophilisers. Product probes are used easily in laboratories for small projects with a small amount of containers and good access. Under production conditions, however, it is difficult to handle probes especially when using isolators, or if the ratio of values is not comparable (e.g., 1 probe/50000 vials instead of 1 probe/50 vials).6 The type of vacuum control, whether capacitance or pirani, has a huge influence as well.1 The use of mass flow controllers (calibrated leaks) compared with the option of opening and closing the vacuum valve must also be considered.

These factors make the transfer of lyophilisation processes difficult from one type of equipment to another. Experience allows some factors to be estimated with sufficient accuracy; however, all too often it is only through trial and error that production is successfully transferred from development to production.

Strategic design of the scaleup lyophiliser

Small lyophilisers achieve acceptable values, but are often not designed in the same way as production machines. To collect comparable values, lyophilisers used for the development and scaleup of new products should be designed to be as much like the production equipment as possible. This design principle needs to be applied to both the process-related conditions and the handling of the equipment itself. Specific requirements, such as placement of probes and container loading, must be taken into account from the beginning.

The lyophilisation process is classically controlled by time, pressure and temperature (Figure 3). As described above, these parameters depend significantly on the design of the lyophiliser because of the influencing factors. The current scaleup approach is a trial and error process of turning the traditional screws of time, pressure and temperature to align the product attributes independent of lyophiliser design.

Figure 3: The classical parameters.

The current discussion about PAT is mainly orientated towards the increase of process understanding during manufacturing. The design is specifically mentioned in the definition of PAT: "Process Analytical Technology (PAT) is a system for designing, analysing, and controlling manufacturing through timely measurements of critical quality and performance attributes of raw and inprocess materials and processes with the goal of ensuring final product quality."7

The aim of getting deeper into the process should cover the current gap between laboratories, development areas and the manufacturing facilities. This will allow the process to be controlled more by product rather than designrelated parameters, such as moisture content, desorption rate or ice temperature (Figure 4). The outcome of this approach will have a lot of positive advantages besides the reduced scaleup time. For example, the definition of recipes will be more accurate, avoiding lengthy recipes or excessive ramping caused by lack of process knowledge. This will lead to a shorter overall turnaround time.

Figure 4: Possible PAT parameters.

The focus for defining product requirements should be based on PAT. Therefore it is necessary to use sufficient appropriate instrumentation to measure and analyse values through the entire process.

Some examples of PAT to be used as scaleup tools include:

  • Pressure Rise Measurement (PRM) — this is not a new tool but belongs to the PAT family. By closing the separation valve the pressure rise within the chamber provides information about the moisture rate within the product.1

  • Barometric Temperature Measurement (BTM) — this measurement uses the pressure rise by closing the separation valve to calculate the ice temperature as deduced from saturation pressure.6

  • Wireless temperature sensors — this method is more related to handling than to process values. It has the benefit of monitoring the entire process (including loading and unloading) with limited interruption.

  • Near infrared spectroscopy (NIR) — this may be used to measure the moisture content of the lyo cake. A 100% verification of the moisture content can be easily done at the end of the process during automated unloading, which provides immediate process verification.1

  • Humidity Monitoring Techniques, such as Lyotrack — this sensor uses the diffraction of the light emitted by the plasma to determine the humidity of the atmosphere during lyophilisation cycle. This allows the end of primary drying to be identified.8

  • Measurements of gas compositions such as mass spectrometer and infrared (IR) spectrometry — this method may also be used for the evaluation of water vapour.

  • It is important to note that some of these tools may be difficult to use during real production owing to validation requirements and the need to maintain sterile conditions. They are, however, helpful during the transfer process to set the required parameters for traditional lyophilisation processes.

Conclusion

For the time being, scaleup will remain a trial and error process, but measures can be taken to improve its accuracy and reduce the time taken. Each scaleup run avoided will bring the system into production faster and help save resources. Importantly, scale-up should consider production conditions from the beginning; i.e., scale-up should start with scale-down using the characteristics in a production lyophiliser as a template for pilot and laboratory equipment. Additionally, PAT thinking should be used as a scaleup tool. Currently, discussions concerning the introduction of new features to control and to understand the lyophilisation process have been met with a positive reaction. The intended outcome is to gain a process control focused on actual product conditions instead of traditional external parameters, such as silicone oil temperature and duration. It's possible, therefore, that PAT will turn out to be the missing link between pilot and production scale lyophilisers. This will, however, require pharmaceutical companies to use PAT tools more consistently as part of everyday operations.

The author says...

Manufacturers of lyophilisation equipment have invested heavily in the design and control of their equipment to streamline the process of scaleup from the laboratory, through the pilot phase and on to full production. Pharmaceutical companies could leverage this experience, involving their supplier partners, at the earliest stage of new product development, to minimise development time, reduce costs, save resources and help themselves to bring their products to market more quickly. Through this type of cooperation it will be possible to smooth the way to faster, cheaper, more effective production in the future.

Maik Guttzeit is Head of Quality Assurance at GEA Pharma Systems, Germany.Tel: +49 2233 6999 401 maik.guttzeit@geagroup.com

References

1. G-W. Oetjen and P. Haseley, FreezeDrying (Wiley-VCH, Weinheim, Germany, 2004).

2. FDA Guidance — Guide to Inspections of Lyophilisation of Parenterals, 2001.

3. P. Cameron (Ed), Good Pharmaceutical Freeze-Drying Practice (Interpharm Press, Englewood, CO, USA, 1997).

4. L. Rey, J.C. May, Freeze-drying/lyophilisation of pharmaceutical and biological products (Informa Health Care, London, UK, 2004).

5. T.A. Jennings, Lyophilisation: introduction and basic principles (Interpharm Press, Englewood, CO, USA, 1999).

6. W.Jorisch, Vakuumtechnik in der chemischen Industrie (Wiley-VCH, Weinheim, Germany, 1999).

7. FDA Guidance for Industry PAT — A Framework for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance, 2004.

8. S.M. Patel, Process control of heat and mass transfer in freezedrying (ETD Collection for University of Connecticut. Paper AAI3383922, 2010).