In November 2001 Janet Woodcock, MD, Director, Center for Drug Evaluation and Research (CDER), U.S. Food and Drug Administration, outlined the agency's Process Analytical Technology (PAT) Initiative to promote innovation in manufacturing process research and development to achieve drug quality improvements in the U.S. pharmaceutical industry (1). The PAT subcommittee of the Advisory Committee for Pharmaceutical Science was established and chaired by Ajaz Hussein, PhD, Deputy Director, CDER, Office of Pharmaceutical Sciences. This initiative is designed to address manufacturing-related quality problems within the industry by encouraging process innovation in a regulatory risk-adverse industry.

PAT has been defined as systems for analysis and control of manufacturing processes based on timely measurements during processing, and control of critical quality parameters and performance attributes of raw and in-process materials and processes to ensure acceptable end-product quality at the completion of the process.

At the 23 October 2002 PAT subcommittee meeting presentations (2) were made on the PDA Technical Report No. 33, The Evaluation, Validation, and Implementation of New Microbiological Testing Methods (Jeanne Moldenhauer (Vetech Pharmaceutical Consultants, Inc.) and the role of rapid microbiological methods in PAT (Dr. Robert Johnson and colleagues, GlaxoSmithKline). Also there was an afternoon breakout session moderated by Dr. Peter Cooney to discuss the role of Rapid Microbiological Methods (RMM) within the PAT Initiative. Drs. Michael Korcynzski and Scott Sutton, members of the USP Analytical Microbiology Expert Committee, participated in the breakout session. When Dr. Korcynzki summarized feedback from the breakout sessions, he emphasized that the overriding concerns were the acceptance of rapid methods by regulatory agencies and the complexity of the validation of new microbiological testing methods.

In the pharmaceutical and biotechnology industries, microbial testing may be divided into two areas: first, in-process monitoring and second, product release testing. Examples of in-process monitoring are incoming pharmaceutical ingredient testing, water for pharmaceutical purposes monitoring, intermediate monitoring, and microbial monitoring of personnel and the manufacturing environment. Product release testing includes microbial limit testing of nonsterile drug products, sterility testing of finished product, microbial assay of vitamins and antibiotics, and microbial identification. Typically, because of the extended testing cycle time for classical microbial tests, the results of both the in-process monitoring and product release testing are reviewed as part of the product release and do not immediately influence the manufacturing process.

Classical microbial tests require extended cycle times because they rely on the growth of microorganisms on microbiological culture media. Classical microbial detection is based on the development of turbidity or other indicators of microbial growth in liquid media and of colonies on solid media. The cultivation of microorganisms includes a pregrowth phase that includes the germination of bacterial and fungal spores and the resuscitation of stressed vegetative bacterial and fungal cells, a lag phase for the biochemical and physiological acclimatization of the cells to the media, and a logarithmic phase during which the cells are actively dividing. Rapidly growing bacterial cells may have a generation time of the order of 20 minutes, and most microorganisms isolated from pharmaceutical ingredients, intermediates, and products are stressed. Thus, the lag phase may be extended, and the time for microbial detection and/or enumeration typically ranges from 2 to14 days of incubation. Table 1 summarizes the classical microbial tests and their incubation times.

Table 1. Examples of in-process microbial monitoring and product release tests and their incubation times

Clearly the extended incubation times of the current microbial tests preclude them from being applied to PAT. Furthermore, optimization of the manufacturing process and the substitution of in-process chemical and physical monitoring to reduce the product manufacturing and release cycle will not be possible when the in-process and product release microbial testing are the most lengthy parts of the cycle. For example, the bulk solution preparation, sterile filtration, and aseptic filling of sterile pharmaceutical product may take 1–2 days, while the sterility test has an incubation time of 14 days. The advent of RMMs that are less reliant on microbial growth represent a major opportunity to reduce the risk of microbial contamination and shorten the product release cycle time.

The PDA Task Force conveniently classified microbiological testing methods as 1) growth-based, 2) viability-based, 3) cellular component– or artifact-based, and 4) nucleic acid–based methods (3). Growth-based methods rely on the measurement of biochemical, physiological, or physical parameters that reflect the growth of the microorganisms; hence they will never be real-time methods. Conventional examples of growth-based microbial testing technologies are plate count, most probable number multiple-tube, and membrane filtration methods. Rapid microbiology examples of growth-based microbial testing technologies are ATP bioluminescence, impedance, and colorimetric or radiometric CO2 detection methods. The commercialization of highly sensitive image analyzers has made it possible to count microbial colonies on plates before they develop into visual colonies, eliminating lengthy incubation time (updating Frost's Little Plate Count Method.)

Some examples of viability-based methods are:

Some examples of artifact-based methods are:

Nucleic acid–based methods that are most closely associated with microbial identification include DNA hybridization, ribotyping, and 16S rRNA sequencing techniques. Recent advances in Polymerase Chain Reaction (PCR) technology make it possible to use rapid cyclers and specific primers to amplify and use fluorescence-tagged probes to detect microbial nucleic acids in real time. It is possible to achieve up to 48 amplification cycles within 20 minutes—the same time required for a single generation for a dividing microbial cell. It is technically possible to semi-quantify specific microorganisms within a test sample using this technology. For example, the log magnitude of Gram-negative bacteria within a water sample can be determined using PCR technology based on the number of amplification cycles to reach a predetermined quantity of nucleic acid.

The challenge for a microbiologist within the PAT initiative will be identify critical processing steps in a manufacturing process for a particular pharmaceutical dosage form and to apply an RMM to monitor the bioburden within the product intermediate to control the process and lessen the risk of microbial contamination. Several risk assessment programs could be applied, including: Hazard Analysis and Critical Control Points (HACCP) from the food industry, and Failure Mode and Effects Analysis (FMEA), an engineering program. The system developed by the food industry has general applicability to microbial contamination in the pharmaceutical industry. In the 1960s, the Pillsbury Company, the U.S. Army, and NASA introduced a system for ensuring pathogen-free foods for the space program. This system, HACCP, focuses on critical food safety areas as part of total quality programs and has been implemented by FDA to prevent microbial contamination of foodstuffs at risk to cause food-borne microbial illness (4). It involves a critical examination of the entire food manufacturing process to determine every step at which there is a possibility of physical, chemical, or microbiological contamination of the food, which would render it unsafe or unacceptable for human consumption. These identified points are the critical control points (CCP).

There are seven principles to HACCP:

Process steps for the manufacture of a sterile injectable pharmaceutical drug are:

From a microbiological testing aspect the five critical process steps are a) incoming microbial testing of pharmaceutical ingredients (Step 3), b) preparation of Water for Injection (Step 5), c) bulk solution preparation and sterile filtration (Step 7), d) air, personnel, and facilities microbial monitoring (Step 14), and e) quality attributes testing (Step 16).

An example of the application of FMEA to the pharmaceutical industry may be found in a case study of risk associated with the routine use of sterility testing isolators (5). A numerical approach was employed based on 1) severity, i.e., the consequence of failure, 2) occurrence, i.e., the likelihood of failure based on past experience, and 3) detection, i.e., the monitoring system in place and likelihood that the monitoring system will detect a failure. The critical areas investigated were the room surrounding the isolator, the decontamination cycle, the frequency of decontamination, isolator integrity, transfer of material from the transfer isolator to the testing isolator, incomplete sample and equipment decontamination, and isolator operating parameter failure. The risk assessment identified glove leaks, loss of isolator integrity, and incomplete decontamination as the greatest risks, in descending order.

Because the formal application of risk assessment techniques to pharmaceutical processes is a new activity, the reader is cautioned to avoid the uncritical transfer of techniques from other industries to the pharmaceutical industry. The authors hope that unique intellectual frames for risk assessment in the pharmaceutical industry are developed and incorporate the specific requirements of pharmaceutical manufacturing and testing processes.

The question may be asked, if RMMs were available could they be beneficially deployed for process monitoring to establish CCPs? Clearly RMMs may be employed for the incoming microbial testing of pharmaceutical ingredients, for product release testing to reduce the manufacturing and release cycle time and control inventories, and to prevent product backorders. Because in-line conductivity and total organic carbon measurements are being employed for real-time monitoring of water for pharmaceutical purposes, it would be attractive to add in-line microbial monitoring as a control for these important pharmaceutical ingredients. Fluorescence flow cytometry may be a potential candidate for this function. The presterile filtration bioburden is a critical parameter for the maintenance of a high level of sterility assurance of aseptically filled pharmaceutical products. A real-time measurement of the number and size of bacteria within the bulk solution could control the bacterial challenge to the sterilizing filter. However, first an appropriate bioburden level for a specified bulk solution volume and sterilizing filter surface area should be determined. It is assumed that the alert and action levels would be related to the recommended microbial count for bulk Water for Injection and the sterilizing filter rating of the retention of 107 colony-forming units of the challenge organism, Brevundimonas diminuta, per square cm of filter surface (6). An RMM that can produce microbial counts within 1–3 hours could be employed as a process control providing information prior to starting the sterile filtration process. An RMM could be employed as a cleaning verification step for critical equipment and facilities. In response to bioterrorism threats, research organizations and commercial companies are working on optical laser microbial detection systems that could conceivably be used for real-time monitoring of air and surfaces in aseptic processing areas. Time will tell whether these technologies will have utility in the pharmaceutical industry.

Performing routine microbial monitoring to demonstrate a satisfactory level of process control and identify adverse trends during sterile product manufacturing takes considerable effort. Much of this effort is wasted because of the delay in obtaining microbial testing results and our inability to analyze the data and recognize adverse trends in a timely manner. Major benefits would arise from combining information technology and instrumentation for microbial detection, enumeration, and identification. These would include electronic data capture and the ability to analyze the data in real time to quickly identify adverse trends. Most microbial tests are limit tests. Examples of limit tests are: an absence of E. coli in 10 g of a pharmaceutical ingredient, less than 10 colony-forming units in 100 mL of Water for Injection, or less than 1 colony-forming unit in 10 cubic feet of air in a Grade A aseptic processing area. Unlike chemical assays for potency in which simple control charts can identify changes in the manufacturing process, the vast preponderance of microbial test results are below the sensitivity of the test. This makes the real-time capture and analysis of the test results more important with the employment of trend rules that are not immediately obvious on inspection of the data. Examples of these rules may be the frequency of out-of-limit results within an extended time period or the mean time between occurrences of rare events.

Given the potential benefits of RMMs in PAT what are the major obstacles to their development, validation, and implementation?

Obstacles include:

How are these obstacles to be met? Companies developing RMMs must work closely with opinion leaders in the pharmaceutical industry to develop applications that can be implemented. Microbiologists need to revisit the concepts contained in the PDA Technical Report No. 33 The Evaluation, Validation, and Implementation of New Microbiological Testing Methods (3) and the USP In-Process Revision Validation of Alternative Microbial Methods (7) as they may be too confined to USP–NF á1225ñ Validation of Compendial Methods (8). The acceptance criterion of same or better recoveries than the compendial method may be too rigid. RMMs will differ in their selectivity in the recovery of a microbial population and thus will have a high bias with respect to microbiological growth-based, methods. RMMs that have different units of measurement should be evaluated to determine if new control criteria (for pass/fail or alert/action levels) are warranted. An example is the difference between classic culture methods reported as colony-forming units and viable but not culturable microbial cells with solid phase laser scanning fluorescence microscopy. The results obtained by different enumeration methods should be compared to the existing requirements and evaluated against microbial risk for a particular pharmaceutical dosage form. For example, water microbiological standards are based on microbiological culture methods. The health impact of a potable water system would not change if higher microbial counts were obtained using a microbial enumeration method that counted microorganisms that do not grow on microbiological media unless there was compelling epidemiological evidence that higher microbial counts obtained by a RMM had public health consequences. The same argument can be made for water for pharmaceutical purposes.

Appropriate planning and detail can begin to develop an approach that is scientifically sound. One approach, using comparability (FDA Guidance, February 2003, Comparability Protocols—Chemistry, Manufacturing, and Controls Information) may be an acceptable way to justify changing test methodology from a classical method to an RMM. Key details that could be included in the comparability protocol are: instrument validation (design qualification, installation qualification, and operational qualification), tests to perform and criteria to meet to indicate equivalency within the capability of the two methods compared, and plans for periodic reporting to regulatory authorities if changes occur after implementation of the RMM.

Microbiologists need to convince their companies that an investment in RMMs will result in inventory reductions, fewer product failures, faster resolution of laboratory and manufacturing investigations, and better product quality.

There are now movements within industry and the regulatory agencies to collaborate and to develop acceptance criteria and to open avenues for changing to RMMs in an environment where technology can enhance analytical measurement in real time, providing increased assurance of product quality.