Crazy Like a Fox

Feb. 7, 2011
New paradigms are needed if PAT’s promise is to be realized for biopharma process control.

As many of you know, I’ve spent years advocating the use of real-time process monitoring and control in the pharmaceutical industry. And, as any of you who have read this column may have noticed, I’ve been disappointed by the fact that process analytical technology (PAT) has, so far, failed to become routine for most drug manufacturers.

Even though they continue to make newspaper headlines every day, pharmaceutical cost and quality issues have failed to drive change, and to establish a true mindset of process control, within the drug industry.

In the small-molecule world, some groups dabble with PAT tools for manufacturing, others venture a bit farther, also applying PAT within the R&D environment, others use PAT, but keep quiet about the applications. Very few, if any, use PAT to control an entire process. 

So how about biopharmaceutical development and manufacturing, where it is often said that “the process is the product and the product is the process?”

Biopharma manufacturers already use analytics to a greater degree, and have more control over their processes, than their small-molecule counterparts. 

For these and other reasons, Ajaz Hussain, my former boss at FDA, has predicted that biopharma will initiate and champion PAT’s adoption for the entire drug industry.

At this point, though, biopharma PAT has a long way to go before its promise is fulfilled.

Despite its use of twenty-first century analytics and control, biopharma continues to struggle with PAT because it is still rooted in the same twentieth-century methods that drive (and impede) small-molecule manufacturing.

Consider the relatively straightforward process of manufacturing therapeutic proteins. It is at the scaleup stage that we see most clearly the dichotomy of biopharma. 

Cell hosts need to be grown in a sufficiently large bioreactor, under “controlled” conditions, but, instead of adjusting the controls to organisms’ current needs in real time, this control is still achieved by using static parameters—the temperatures or pH levels, for instance, which, in the past, have been shown to produce the “right” proteins. 

The proteins that are isolated from the cultures, purified, and tested so stringently are produced by living organisms in one of the most dynamic process environments imaginable.

The health and behavior of these organisms directly affect the proteins being produced, and their performance is, in turn, affected by even the slightest change in process conditions or raw materials.

Yet, today, the entire operation is still focused on the same static “validation batches” and “testing of finished product” that rule small-molecule manufacturing, and limit potential improvements in cycle time, process performance, and supply chain efficiency.

We have come to understand but some of the organisms’ processes and their relationship and interactions with such materials as media, cells, and nutrients. 

Yet, most of the material attributes that directly indicate the performance of the cells, allowing proactive adjustments of the process parameters—in other words, real-time process control—not only are not monitored but are completely ignored.

Only when we monitor those attributes, and adjust the processes to facilitate real-time control, will real savings and improvements in performance and efficiency be achieved.  At this point, we are still validating bioprocesses at the edge of failure. 

The technology and knowhow required to do this already exist, and the concepts were spelled out in FDA’s PAT guidance document six years ago, and reaffirmed in FDA’s recently published Process Validation guidance, which I am proud to have played a part in drafting.

But more than science and technology are involved. There is economic reality. The drug industry is still more profitable than other manufacturing sectors, so, perhaps its managers don’t see the need for control. (Although J&J’s management team must have gotten a hint recently, with news of the 12% drop in profits and 5% fall in sales resulting from its product recalls and cGMP noncompliance issues.)

Over the winter break, my daughter and I were riveted by the movie based on Roald Dahl’s classic children’s book, “Fantastic Mr. Fox.” As bullets ricocheted around him, his tail painfully shot off and his wife and children starving, the hero developed a winning strategy, but only after admitting that none of the solutions he’d come up with up to that point would work.

Is it time for the drug industry to do the same thing, to own up to its past failures and, using the PAT and Process Validation guidances, come up with a solution rooted in dynamic, real-time process control?

Culture and tradition make it especially difficult for pharma to bulldoze its silos and dig its way to fundamental process understanding, but this radical approach is the only way for process control to take root and for pharma to come closer to achieving the ideals expressed so many decades ago by Dr. Deming.

Can biopharma, arguably the most dynamic of pharma’s sectors, throw off the static shackles of the past? 

The biosimilars industry, which will depend upon demonstrating that one product’s biological activity is the same or similar to that of another’s, cannot take root without PAT and advanced analytics. But PAT cannot exist in an environment that is limited by traditions. Will biopharma, like Dahl’s hero, take a dynamic approach to ensure its survival, and its future?

About the Author

Ali Afnan | Principal