Process optimisation for faster drug development

Time is money in pharma development, with every day saved in the development process meaning, potentially, millions added to the bottom line. Sarah Houlton reports from the RSC Conference on possible ways of speeding things up

Subjects as diverse as astrochemistry and protein refolding were addressed at this year's RSC Conference, held at Birmingham's international convention centre at the end of July.

Developments in automated laboratory instruments and in situ spectroscopy techniques have continued apace, and are providing a wide variety of additional tools for the development of drug candidates. Joel Hawkins of Pfizer in the US explained how his company is using these techniques to increase productivity per person, per square metre of lab space, per gram of material, and over time.

Pfizer has identified three areas in which substantial savings in all these parameters can be made:

Machines commonly used at Pfizer include the Anachem SK233 workstation, which is used to screen catalysts, solvents and reaction temperatures to maximise yield of the desired product while minimising impurities. In situ yields can be plotted against time and reaction conditions to aid in the visualisation and interpretation of the huge amount of data that are generated. Another piece of equipment, the Endeavour, Argonaut's parallel pressure reactor, enables eight separate hydrogenations to be screened at once, measuring hydrogen uptake in real time.

Shari Taghavi from Cellular Process Chemistry Systems in Germany looked at the area of microreactors. These are miniaturised reactors that have been fabricated by microtechnology and precision engineering. Microreactors are particularly useful for toxic or exothermic processes, especially where mass transfer is important. Numerous different machines are available, mainly for research applications. However, they were designed for specific tasks, and were not intended to provide a universal system that could be used for the development and manufacture of pharmaceuticals under production conditions.

Taghavi explained that CPC has developed what it claims is the first commercial microreactor system. CPC's view of a microreactor is that it should be a continuous process using small volumes, of the order of millilitres, with enhanced control, and not rely on microlitre quantities. Its system offers the rapid equilibration of concentration gradients, a narrow temperature distribution (of the order of around 2K), and is said to be easy to use. The system is designed to give faster access to larger quantities by a continuous process and simple numbering up, with exact process control, safety and reproducibility. A huge range of reactions has already been investigated in the system, including Suzuki coupling, Wittig reactions, Pd-catalysed cross coupling, intramolecular Diels Alder reactions and BuLi chemistry.

Process intensification was addressed by Roshan Jachuck from the University of Newcastle upon Tyne's process intensification and innovation centre. The centre is involved in the development of improved technology that will give process enhancements by orders of magnitude, with 'greener' synthetic routes being applied if at all possible. Methods include using spinning disc reactors, which give rapid mixing and shorter residence time for the reaction, meaning higher reaction temperatures can be used.

supercritical fluids

The use of supercritical fluids in formulation and particle design was the subject of a talk by Michel Perrut from Lavipharm-Separex in France.

The fluids, particularly supercritical CO2, have applications in the solubilisation of products with very low solubility, like biologicals. But they can also be used for procedures as diverse as mixing and stabilisation of emulsions, and the development of sustained release aerosol formulations.

Several different approaches have been developed for optimising particle design using supercritical fluids, including micronisation, microencapsulation, impregnation, complexation and for water insoluble products. Both amorphous and crystalline products can be obtained. Of particular note is rapid expansion (RESS) methods. Another development is the use of sub-critical fluids for microencapsulation, particularly if the formulator wants to use no organic solvent at all, especially for fragile molecules like proteins. The active ingredient and the carrier are mixed in a suspension at low temperature (about 37°C) and subcritical pressure. This reduces stress on the protein.

One of the specialist symposia, entitled 'What's happening in process and control', included a couple of papers on process control and optimisation. Emerson Process Management's Chris Hawkins addressed the subject of real-time optimisation. The aim of production optimisation is to minimise or maximise, as appropriate, a plant or business objective by manipulating plant operation. If this is done in real time, then changes are made to the plant in an almost immediate response to changing process or business conditions. It is essentially the evaluation and implementation of the optimum process operation conditions — while they are still valid — so, for a large, fast-moving process, this will mean every few minutes and not every few hours.

Four stages are involved in real time optimisation, Hawkins explained:

If all criteria are taken into account, such as how the current process interacts with future procedures, and the process is kept on track, then the result should be an improved process.

Robin Brooks from Curvaceous Software presented an innovative way of visualising multiple variables on the same graph. Despite process plants only generating a relatively compact set of data, there are still too many to be able to visualise them easily. It is possible to reduce the number of variables, for example, by using dimensionless groups like Reynolds number, but Curvaceous wanted to develop a graph that enables more variables to be visualised directly. The solution proved to be the use of parallel coordinates, using a mathematically rigorous method to transform n dimensions to two. The result is a single line representing all the dimensions for each point, and these can then be overlaid onto one graph for comparison purposes.

AJM Consulting's Alan Mason spoke about recent developments in the integration of control systems with the other systems a business uses. He explained that there are numerous benefits to be gained from integrating control systems, including integrated batch reporting, which should prove invaluable in validation, and the availability of real-time operational cost information. It also enables dynamic production scheduling, and makes quality management under systems such as ISO9000 easier to achieve.

seveso directive

Perhaps the most important benefit, however, is in the area of regulatory compliance, notably meeting the requirements of COMAH (control of major accident hazards, also known as the Seveso directive). The regulations classify all chemical makers and users as high or low tier. Very large sites are automatically classified as high tier. But for medium-sized chemical plants, the COMAH score can move up and down between tiers, depending on the current operations. If it falls below the low-tier threshold, then there is no problem. Below the higher tier, a number of requirements come into play, incurring a cost of around £50k (e80k) pa for a small chemical factory. But if the operations at any time fall into the high tier, then this will cost another £250k (e400k). Accurate calculation of the tier score is thus essential. Each chemical's contribution to the tier score is the number of tonnes in stock divided by the threshold value. Data need to be validated, and all materials must be allocated a classification, even if it is not relevant, and the system should default to an error condition. The whole needs to be auditable and available for inspection.

A daily archive provides summary information on the overall trend for low and high-tier assessment. Each daily calculation can be analysed to show input information for all stock.

Benefits for the specialty chemical manufacturer include the integration of existing material management systems with existing process control systems, which provides cost effective implementation with little additional training. Instant automatic calculation of classification tier status can be done through SCADA. And, vitally, it can be validated. The future will see COMAH being stringent in its material classifications, with frequent inspections and severe punishments increased. There will also be more accurate definition of materials, notably taking into account what is actually going on within a reactor, instead of merely looking at the ingredients added and the product and by-products that remain at the end of a process.