January 15, 2009 (Vol. 29, No. 2)

Waters Applies Its UPLC Technology to Obtain QC Results on the Biomanufacturing Floor

With guidance from the FDA and the ICH, the pharmaceutical industry is seeking to accelerate the pace of manufacturing innovation to meet ever-increasing cost, efficiency, and time-to-market demands.

Since the FDA issued its process analytical technology (PAT) guidance in 2003, PAT has received a considerable amount of attention. Its goal is to assure final product quality through process design, measurement, and analysis of key product-quality attributes, and dynamic control of the manufacturing process. 

Pharmaceutical companies have, historically, taken a conservative approach to implementing process changes and upgrading manufacturing technology. But pharmaceutical business models are rapidly changing, and the importance of manufacturing’s role in overall financial performance has become a focal point as demonstrated by the increased adoption of Six Sigma and Lean Strategies throughout the industry. 

While the cost of restructuring and potentially retooling production lines is significant, the long-term savings gained from the more efficient use of existing human and capital resources, reduced in-process scrap and waste, greater assurance of product quality, and mitigating the risk of product recalls outweigh the cost of implementing a PAT Program. 

Traditional approaches to assuring product quality in pharmaceutical manufacturing depend heavily upon post-process off-line laboratory analyses. This model typically causes intermediate batch materials and the manufacturing line to idle for long periods awaiting the results of the off-line QC tests. The cost of this off-line lab-centric approach was documented in a recent MIT study that demonstrated over 80% of the manufacturing processes evaluated spent more time analyzing the product than manufacturing it.

The reasons that companies sometimes put off plans to address manufacturing inefficiencies are due to tradition, the cost of change, and inertia. Yet, most industry professionals willingly admit the consequences of maintaining this traditional approach include higher costs, higher levels of rejected product and rework, and ultimately slower time to market.

PAT enables manufacturers to consistently meet or exceed their product specifications by giving them the ability to dynamically adjust and control the manufacturing process based on real-time feedback information on critical product quality attributes. Equally important is having the informatics and automated process controls in place to make production adjustments based on analysis results. In other words, the impact of real-time analysis of the production stream is minimal if you are unable to quickly respond to the analytical results.

Implementing PAT is likely to have the biggest immediate impact on products with recurring quality issues because process deviations or exceptions often result in lost or poor product quality—leading to lower yields and higher costs, especially with expensive and hard-to-acquire raw and intermediate materials. Other good candidates are new products that have yet to come on-stream and for which regulatory submissions are in the formative stages.

Room for Improvement

Numerous sensor technologies can be employed throughout the manufacturing process to measure critical quality attributes of the in-process material. Typically, process steps such as reaction monitoring are assessed by spectroscopic sensors, which include near-infrared spectroscopy  or Raman spectroscopy. These techniques can provide real-time information about the reaction progression but lack the ability to effectively resolve and quantify product variants or multiple components in a sample, particularly if the amounts of some of the components are at low levels. These analytical techniques rely heavily on chemometrics and are greatly challenged when differentiating chemically similar compounds, while at the same time their sensitivity and dynamic range is relatively poor and their ability to deliver absolute quantification is impractical.

The performance of these sensors is typically benchmarked and confirmed against an analytical reference standard, which in most instances is liquid chromatography (LC). As the predominant analytical reference standard in the QC lab, LC has demonstrated exceptional resolving power, is ultrasensitive, and can detect small amounts of impurities as low as 0.001% even in the presence of main components.

As the gold standard for off-line in-process sample analysis it would seem a likely technology for PAT implementation.  The biggest issue with using LC as a PAT sensor has been that analysis time is too slow and existing LC platforms require a high degree of expertise to operate. These primary drawbacks have prevented LC from being routinely deployed for use in at-line or online in-process analysis.

Manufacturing Efficiency

With the introduction of Waters’ UltraPerformance LC® (UPLC®) technology it is now possible to achieve near real-time chromatographic analysis for in-process samples. The Patrol™ UPLC Process Analyzer brings reference-standard methodology and performance to the manufacturing floor, eliminating the need to send in-process samples to an offline QC laboratory.

The biggest advantage of UPLC is its reliability, resolving power, sensitivity, and analytical speed. For example, UPLC is able to deliver analysis speeds up to ten times faster than HPLC. The improvements in speed, resolution, and sensitivity made possible by UPLC technology drew the attention of process engineers and support chemists seeking ways to improve the manufacturing efficiency and quality of biopharmaceuticals and small molecule pharmaceuticals.

The Patrol UPLC Process Analyzer has been designed and optimized to leverage and take advantage of new sub 2 µm hybrid particle technology. It features integrated hardware, chemistry, control and communication software, and a ruggedly engineered sample-introduction technology—all contained within a NEMA 3R- or NEMA 4X-rated enclosure purpose-built for the manufacturing environment.

The Patrol UPLC Process Analyzer is compatible for both online and at-line analysis of process streams or reactors providing fully automated real-time analysis of in-process samples with no need for user intervention. Results can be sent directly to a distributed control systems or a LIMS completing a closed-loop monitoring process. For at-line applications where it may be desired to have samples extracted from a reactor or the process stream by a technician, the system features a walk-up interface with internal barcode scanning capabilities eliminating the need for a technician to key in sample or analysis information.

Putting UPLC to the Test

Monitoring and quantifying all the components in a process reactor during a synthesis reaction enables the highest quality and yield of the target compound. The real-time monitoring capabilities of the Patrol UPLC Process Analyzer combined with Empower™ 2 Chromatography Data Software can analyze and graphically represent the amounts of identified components in the reactor to make decisions on reaction progress and forward process.

In a feasibility study conducted by Waters, process application specialists using the Patrol UPLC Process Analyzer monitored the conversion of acetylsalicylic acid to salicylic acid. Aliquots were automatically taken from the reactor vessel at repeated intervals and analyzed with the Patrol UPLC Process Analyzer as the reaction progressed.

The analyzer baseline-resolved all the components in each aliquot in two minutes with an injection-to-injection cycle time of four minutes (Figure 1). As the reaction progressed the consumption of the starting material and the formation of the API, intermediates, and impurities were monitored and quantified.

The Patrol UPLC Process Analyzer features a wide linear dynamic range that allows for the simultaneous quantification of both high- and low-level components within the same chromatographic run (Figure 2) even at levels below 0.01% of the major components.

Online spectroscopic sensors cannot simultaneously quantify such a complex matrix with such varying concentrations. By plotting the %Area of each of the components  in the chromatograms, a map of the reaction can be generated (Figure 3).  This reaction map can be used to determine the optimal time to quench the reaction based upon maximum yield of the API or even minimum formation of the specific low-level impurity.

With the Patrol UPLC Process Analyzer, manufacturers of biopharmaceutical and pharmaceutical products have a new tool for PAT. Just as LC is today’s gold standard for QC labs, bringing this technology to the manufacturing floor holds tremendous potential for realizing gains in the efficiency of the manufacturing process and the quality of manufactured products in number of ways.

To better insure that tactical and strategic objectives are met, quality control, PAT, and Six Sigma teams must work closely with their vendors and consultants. The solution chosen for PAT analysis should be thoughtfully designed and built for the purpose of at-line or in-process analysis.

Figure 1. Quantification of the API and process impurities for select aliquots

Figure 2. Quantification of both high- and low-level components within the same chromatographic run at levels below 0.01% of the major components

Figure 3. A summary plot for Area % of each component in the reaction vessel for determination of the reaction endpoint

The Patrol UPLC Process Analyzer increases the efficiency of the manufacturing process in a number of ways:

  • It reduces production cycle times by using online and/or at-line measurements and controls
  • In-process materials are managed more efficiently
  • Product variability is identified during manufacturing
  • Rejects, scrap, and reprocessing are reduced
  • The possibility of real-time release is closer
  • It increases automation to improve operator safety and reduce human errors

Tanya Jenkins (tanya_jenkins@waters. com) is senior applications chemist at Waters. Web: www.waters.com.

Previous articleJ.P. Morgan Healthcare Conference—Monday, January 12 Update
Next articleInvestigators Use DNA Repair Patterns to Determine Pancreatic Cancer Risk