Fits and Starts
Some PAT projects appear promising during process development, but never make it past that point. Patricia Andres, Ph.D., director of drug development at contract development firm SSCI, uses measurements of turbidity, temperature, and Raman analysis in designing crystallization processes for active pharmaceutical ingredients, but is unaware of analytic technology being passed down to manufacturing. “I believe that technology transfer to manufacturing may be an issue,” she admits. “It would take a lot of effort for something that would only be used a few times.”
Emil Ciurczak, who heads Doramaxx Consulting, a firm specializing in near-infrared analytics, says that even success of individual projects does not guarantee that PAT becomes part of the pharma/biotech DNA. He observes that even companies that reduce manufacturing times substantially by implementing analytics often drop the ball with later projects.
“They’re afraid of seeing what they’re going to see, of what’s really wrong or can go wrong with their process,” he explains. “They’d rather shoot craps and hope nothing shows up.”
Then there’s the regulatory uncertainty angle, which has plagued PAT since its original promulgation in 2003. “Manufacturers are still quite wary about adopting new technology and having to document it to regulators. It happened during the 1970s with HPLC, and it’s happening now with PAT,” Ciurczak says.
Speaking of HPLC, the workhorse analytic tool of pharma/biotech labs has barely made inroads onto the manufacturing floor despite several articles per year demonstrating its feasibility. The knock on chromatography is its slowness and complexity. John Hanrahan, Ph.D., CTO at silica particle maker Glantreo, believes these criticisms are less and less justified as HPLC systems become faster.
“We now have the capability for manufacturing small, porous particles that could be used in high-pressure, and rapid, HPLC,” Dr. Hanrahan is talking about particles that are significantly smaller than the sub-2-micron silica used in rapid HPLC techniques today, for example smaller than 1 micron. These will require pressures well above the 20–25 kpsi that the best pumps deliver today and could conceivably return results in a matter of seconds.
But as Dr. Hanrahan notes, no significant process today employs HPLC in a PAT environment. Manufacturers who are unimpressed with generating a diagnostic HPLC trace every few minutes are unlikely to be won over by a 50% or even 90% reduction in analysis cycle time.