Quality--by design: pharmaceutical regulators are moving to encourage new methodologies that should make it easier to improve manufacturing processes once drugs are licensed.

[ILLUSTRATION OMITTED]

If out-of-specification drugs reach the market, the consequences for patients are potentially disastrous. This is true whether there is too much or too little active ingredient in a tablet, the active is contaminated or not what it should be, or the properties of the dosage form mean that it isn't absorbed properly by the body. Historically, the regulatory environment has reflected this by focusing on identifying out-of-spec product through careful inspection, and keeping process parameters the same for every batch to minimise the chances of something untoward happening.

But now there are moves to focus less on inspection and allow more leeway in changing and improving processes after a product is on the market, as long as the process itself is well understood. Compared with many other industries. pharma manufacturing procedures have advanced very little over the years, partly because of the sheer complexity. 'If there are 10 stages to produce a pharmaceutical, the numbers of potential variables and parameters that affect those variables are huge,' says Britest's director of innovation, Derek Lindsay. 'And this is for something that will likely be a fairly small volume product, unless it's one of the big blockbusters; even there the scale is not massive.'

While only good quality products are allowed to enter the supply chain, this is often at great expense because many batches of failed product are also made. Indeed, Lindsay believes that at the formulation stage, the reject rate can be up to 30%. 'There could be seven stages of chemistry to produce the active pharmaceutical ingredient (API), and then a three or four stage formulation process,' he says. 'Variability could arise anywhere in that whole process, and it's important to understand where it might be coming from as it could be several steps back from the final product. It's important to understand all the parameters, and then monitor and control them.'

Because the manufacturing route forms part of the product licence, any alterations to the reaction conditions, however minor, have to be reported to the regulators, adding both cost and time. This has been a major disincentive to finding better manufacturing procedures. Moving from a 'quality by inspection' approach to this 'quality by design' (QbD) strategy would simplify the introduction of better processes. Although no deadlines have yet been set for submissions to have to include QbD data, the technique is being actively encouraged by regulators such as the US Food and Drug Administration (FDA). Much more work is done to understand the process at the outset, so the likely effects of such changes on product quality are predictable.

QbD is, essentially, an approach to process development that emphasises the need for a good understanding of both the product and the process, based on sound science and quality risk management. 'It is not intended to replace an inspection process, but is intended to provide a higher level of process understanding throughout a project's development timeline or lifecycle, both pre- and post-commercialisation,' explains Tim Watson, associate research fellow, development API, at Pfizer in Groton, US. 'A QbD scientific approach, compared with traditional development, significantly expedites improved process understanding. Sound scientific principles generate data that stimulate creativity to continually modify and improve a process. The end result is a more robust process that reduces unexpected events.'

'The FDA is very keen,' says consultant Trevor Laird of Scientific Update. 'They want people to be able to improve their manufacturing processes without necessarily having to come back to them if there are changes that won't alter the product licence. While major changes, such as a completely different manufacturing route, would still have to be approved, if the synthetic route remains the same you can make changes within the design space you develop without having to notify them.'

[ILLUSTRATION OMITTED]

Rather than providing exact parameters, such as temperature, pressure and reaction times, the initial submission to the regulator would give a range of parameters for which the process was understood. 'You would be free to change the parameters within the range, providing you had a good rationale for the change,' says Guy Villax, chief executive at Portuguese custom manufacturer Hovione.

The key question for Villax is: when should the time, effort and money be put in to developing a greater process understanding? There is a conflict that FDA wants to see the process is well understood and well defined, he says, but that means starting early on in the development process, when there is no guarantee that the product will make it onto the market. 'We can make [APIs for clinical trials] that could be described as "quick and dirty" but where the quality is still reasonable, even though the yields may be terrible or there may be multiple purifications involved,' he says. 'The benefits of QbD are only seen after approval. But we are being told that the approval will be quicker, and if we have a good understanding of the process then they will allow changes much more easily to enable constant improvements.'

Good process understanding

The fundamental basis of QbD is to have a good process understanding in terms of the chemistry, Villax says. And then you need to be able to characterise it well, and know what parameters have an impact on the process--what ranges are safe and do not result in out-of-spec product. It is then vital to execute it without errors: it's no good having a perfect understanding and a clear plan on paper if mistakes are made in the process. And finally, constant improvement of the process is essential if it is to be as cost and time effective as possible. 'Britest has excellent methodologies where it gets people to think about the process, and come up with questions about what isn't known about it that could be relevant,' Villax says.

Several tools can help with process characterisation. PARs, or proven acceptable ranges, involve carrying out experiments at the edge of failure. 'If you say that the reaction will run well at 65[degrees]C, you have to show that even at the edge of failure--75 or 85[degrees]C--the process still ran OK,' Villax says. 'And then there's PAT, or process analytical technologies, to characterise the process, and FMEA (failure mode and effects analysis), which is a fast way to question a process and determine what can go wrong.'

Increasing complexity

Continuing advances in drug discovery mean that pharma products are becoming ever more complex to develop and manufacture. While the inspection process ensures that it is extremely rare for defective products to reach the supply chain, Britest's Lindsay says that manufacturing problems are often a reflection of unsolved development issues.

He adds that the vastly reduced bureaucrocacy with QbD will make developing better processes much easier. 'One of the carrots the regulators are offering the industry is that if they can demonstrate this understanding within an "envelope" of parameters, they will be able to move the process around within that envelope.' He says that one of the reasons for this is the pressure from governments to reduce the drugs bills. 'There is a disincentive to do any sort of continuous process improvement because of the regulatory costs of any changes. Yes, you should design the process right first time through QbD, but if you also design an envelope for it, you will have the flexibility to react to manufacturing changes.'

[ILLUSTRATION OMITTED]

Taking a QbD approach really does reduce process failure. 'By taking a [QbD approach] to understanding a process before implementing it, I have seen significant reductions in unexpected events,' says Pfizer's Watson. 'From my experience, the QbD approach facilitates cost reductions and time savings, minimises impacts to the environment and, most importantly, improves quality, safety and confidence of the process and product.'

'The old-fashioned attitude was to keep everything the same, and then we know we won't have a problem,' concludes Trevor Laird. 'You're now allowed to make changes, as long as you understand the process. This better understanding will lead to better process control, and fewer failed batches and reworks, and therefore better economics.'

In brief

* Quality by design (QbD) focuses on a better understanding of the manufacturing processes used to make drugs

* It allows drug manufacturers more leeway in changing and improving processes after a product is on the market

* QbD should lead to better process control, fewer failed batches and lower manufacturing costs

Sarah Houlton is a freelance writer based in London