Chromatography is an integral step in downstream bioprocessing, and buffered mobile-phase properties such as ionic strength and pH are often critical process parameters for protein purification. The interactions between the mobile and stationary phases can affect chromatographic performance, and thus the separation profile between product and impurities. Product recovery, purity, and throughput are often marred by the variability in buffer preparation during production-scale bioprocessing. The popular approach to biological production typically relies on postchromatographic adjustments (e.g., fractionation controlled by ultraviolet (UV) signal and segregation of out-of-specification product through quality-control testing) to achieve the desired product quality. Although the US Food and Drug Administration accepts this traditional practice, this post-process approach to quality control is time-consuming, increases production cost, and does not provide the process understanding and control from the quality-by-design (QbD) approach that FDA promotes. 

Pharmaceutical CGMPs for the 21st Century

 initiatives, companies should employ process monitoring and control strategies to streamline biopharmaceutical manufacturing. Real-time process knowledge can help manufacturers avoid excessive rework and discarded product. To achieve this standard of quality, FDA highly recommends the implementation of process analytical technology (PAT) to facilitate a QbD approach. According to FDA, the working definition of PAT is "a system for designing, analyzing, and controlling manufacturing through timely measurements (i.e., during process) of critical quality and performance attributes of raw and in-process materials and processes with the goal of ensuring final product quality" (1). 

Many modern technniques for process analysis such as pH measurement, near-infrared spectroscopy, Raman spectroscopy, refractive-index changes, humidity movement, UV monitoring, conductivity detection, and dissolved-oxygen analysis are commercially available for different applications. These analytical tools can measure physical and chemical properties of raw materials directly in real time, in contrast with techniques such as flow-meter monitoring, which only makes indirect measurements. 

Applications of PAT in downstream bioprocessing

In the past 10 years, protein-expression titers have increased 1000 times from mg/L to g/L. As upstream yields continue to increase, downstream bioprocessing involving buffer preparations and delivery also must increase proportionally to keep pace with demand. Therefore, conventional practices of preparing and delivering buffers have to be modified or replaced with new manufacturing sciences (see Figure 1). This principle is especially true for industrial-scale quantities of buffers that must be reproducibile and meet tight specifications. Scientists at Genentech (South San Francisco, CA) proposed using mass-flow meters to dilute buffers from concentrates, which is a great improvement (2). The scientists described the dilution of concentrated acetone, which could be monitored by an in-line optical-density meter. Additional monitoring through feedback control of conductivity and pH can complement the design and reduce the variability of buffer feedstock, preparation, and delivery frequently encountered in bioprocessing. 

 Figure 1: Conventional system for buffer preparation and delivery. CGMP is current good manufacturing practices, CIP is clean-in-place, and QC is quality control. (ALL IMAGES ARE COURTESY OF THE AUTHORS)			

This article examines the operational requirements and design of a process-ready PAT-based in-line buffer-dilution (IBD) system, which is capable of making reproducible and accurate in-line buffer dilutions in concentrations as high as 100 × of the product using conductivity or pH feedback control. The technology can be used within stand-alone packaged equipment or engineered into a process-scale liquid chromatography skid. In the latter option, mobile phases are made on demand and sent directly to a process column for a sequence of chromatography steps in a purification regime. The comparison of a PAT-based and a mass-flow control system includes conductivity gradients, which are typical for biochromatographic processes, and the generation of linear pH gradients with good acidity control. 

Figure 2: A process analytical technology-based in-line buffer-dilution skid equipped with conductivity and pH process analyzers linked with a process feedback-control loop. 			

The literature includes several examples of diluting buffers from concentrates in processes consistent with QbD (3). Some biopharmaceutical companies currently use a patented PAT-based IBD skid for buffer preparation and dilution (see Figure 2). To understand how IBD assists the QbD paradigm recommended by FDA, one should examine the basic technology platform. Figure 3 shows a sample piping and instrumentation diagram (P&ID) from the process-control screen of an IBD system. 

Figure 3: Piping and instrumentation diagram screen of a process analytical technology-based in-line buffer-dilution system. 			

A typical PAT-based IBD skid consists of three pumps labeled P001, P002, and P102. Deionized water for injection (WFI) is connected to P001, and buffer at concentrations as high as 100 × of the product is connected to P002. WFI and buffer concentrate are pumped into a blending loop, where the buffer concentrate is diluted to 1 × the product. The diluted buffer has a specific conductivity that can be analyzed by an in-line process conductivity sensor linked to a process-control feedback loop. The sensor continuously analyzes the conductivity of the buffer blend in the loop and sends an output signal to the process-control feedback loop. The loop then compares the current process conductivity value with a user-defined conductivity target set point for the buffer diluted to 1 ×  the product. Based on the feedback signal, the concentrate pump speeds can be increased or decreased. Whenever the acidity of the buffer solution must be adjusted, alkali or acid can be delivered from P102. In this instance, a pH probe would be installed in the blending loop and linked with its own process-control feedback loop to adjust the solution's pH to meet the user's specifications.  

To assess the ability of a PAT-based IBD system to monitor a process and to actively manipulate it to maintain a desired state, the authors performed experiments to compare simple mass-flow and PAT-based control blending, and to reveal the true benefits of directly measuring the quantity of raw materials present during runs. The first experiment was a buffer-preparation scenario in which errors in the formulation of the concentrate raw material can cause an out-of-specification buffer-dilution product. The second comparison was made in the context of producing a linear pH gradient with pH adjustment and control not commonly available in a mass-flow-based dilution system. 

				 The conductivity and pH analyzers used in these experiments were standard components on a 5-L/min IBD skid from Asahi Kasei Bioprocess (Glenview, IL). The pH probe was an ABB Limit model #AP121 21000 electrolyte-filled glass electrode from ABB Instrumentation (Cary, NC). The conductivity sensor was an Optek CF60-45 with six electrodes and a Pt1000 platinum resistance-temperature device used for temperature measurement and compensation from Optek-Danulat (Germantown, WI). 			

Both probes were mounted in a 1-in. F40 Optek PEEK Flow cell with a PF12 pH electrode adapter. An Optek Control 200 controller was installed on the skid and used a dual-input electrochemical converter to monitor pH, conductivity, and temperature. Conductivity, pH, and temperature output signals from the electrochemical converter were processed and trended in a run chart using Proficy HMI/SCADA-iFIX, Proficy Historian Version 3.1a, and SP1 process-control software from GE Fanuc (Charlottesville, VA) on a personal computer housed in the skid. Oakton Instruments's (Vernon Hills, IL) pH and conductivity standards were used for sensor calibration. 

One liter of sodium chloride (1 × the product) was carefully weighed out using a model JK-200 analytical digital balance from Chyo Balance (Komatsu, Japan) with ± 0.01 mg accuracy. Sodium chloride (~6 × the product) concentrate was weighed out using a model MK-2000B Petit balance (Komatsu) from Chyo Balance with ± 0.1 g accuracy and dissolved into deionized water in a polycarbonate drum. Acetic acid and ammonium hydroxide solution were diluted from their respective concentrations. Sodium chloride and glacial acetic acid were purchased from Amresco (Solon, OH), and ammonium hydroxide was obtained from Mallinckrodt Baker (Phillipsburg, NJ).  

Comparison of mass-flow and PAT-based control blending

 The carefully prepared 1-L standard of 0.167 M sodium chloride gave a conductivity reading of 17.8 mS/cm, which was used as the programmed conductivity set point in the process-control software. Initially, the 1-M  sodium-chloride concentrate delivered by the P002 pump was diluted in the blending loop by the WFI (P001) pump. Both pumps' frequencies were adjusted accordingly to obtain the programmed conductivity set point. The recorded WFI, concentrate, and blending-loop pump speeds were entered into a new method as pump preset start speeds. Starting from 1-M sodium-chloride concentrate, either mass-flow or PAT-based blending achieved the desired molarity with the targeted conductivity of 17.8 mS/cm (see Figure 4). Because of the presence of impurities, mixing inefficiency, and weighing errors, feedstock variability of 5% was expected. 

Figure 4: Conductivity of final products from mass-flow and process analytical technology (PAT)-based in-line buffer-dilution blending. 			

To reveal the state of a process, the authors prepared a sodium-chloride concentrate ± 5% from the original 1-M solution. Using the same pump frequency and method of mass-flow measurement determined earlier, the conductivity of the final product was obtained. The average conductivity of the dilution of 0.95-M sodium-chloride concentrate was about 16.8 mS/cm, which was 1 mS/cm or 5.6% below the set point. For 1.05-M sodium-chloride concentrate, the average conductivity was 18.5 mS/cm, or 5% above the set point. Data indicated that the variability of the starting raw materials determined the outcome, and mass-flow measurement could not adjust to the desired set point. In contrast, the PAT-based IBD used the in-line conductivity sensor in the loop to measure and correct the variability of the incoming salt concentrate through the process-control feedback loop. Whether the starting salt concentrate was 5% above or below the original 1-M concentrate, the conductivity of the product was well within the range of 17.8 ±  0.082 mS/cm. 

 The titration curve between acid and alkali, whether strong or weak, is not linear (4). Therefore, protein chemists seldom use pH gradients during a chromatographic run to resolve proteins with similar pI, especially at production scale. To determine whether the PAT-based IBD system could produce a linear pH gradient, the authors blended a weak acid with a weak alkali. Two experiments, one based on mass-flow, and the other PAT-based, were performed.  

Ammonium hydroxide (10 mM) was delivered from pump P001 to mix with acetic acid (20 mM) from pump P002. In mass-flow blending, the frequency of both pumps was tuned for the pH to start at around 5. A gradual increase of base over 30 min afforded a final pH of 10. Simply blending weak acid and weak base by mass-flow obtained a typical sigmoid curve (see Figure 5). The steep rise of pH from 6.5 to 9 in less than 2 min was typical of mass-flow control. The pH of the resulting solution followed the acid-and-base titration curve.			

The goal for this experiment was to generate a linear pH gradient from 5 to 9. To control the sigmoid change in pH and to obtain a linear pH gradient, acidity adjustment and control through a PAT-based system is necessary. Accordingly, ammonium-hydroxide solution (100 mM) was delivered from pump P102. Initially, three pumps were properly tuned to start at a pH of approximately 5.0. The authors programmed the process-control software with a linear pH gradient from 5 to 9 over 30 min. During the run, the acidity of the resulting solution in the loop was constantly adjusted to meet the set point in accordance with the signal from the process-control feedback loop. As a result, a linear pH gradient was obtained (see Figure 5).  

Figure 5: pH titration curve of 10 mM acetic acid and 10 mM ammonium hydroxide. PAT is process analytical technology.			

Pharmaceutical companies still suffer from excessive re-work and discarded product because of out-of-specification processes. PAT-based IBD has the potential to account for and to reduce the impact of raw-material variability when diluting concentrate solutions. Thus, IBD can reduce the amount of out-of-specification products that must be reprocessed or discarded. PAT-based control can be applied to chromatographic separations so that consistent and reproducible gradients are produced, regardless of the starting materials. Furthermore, PAT control also can enable the use of linear pH gradients for different applications normally not available from a mass-flow control system.  

The ability of an IBD system to monitor and manipulate the state of a process and actively manipulate it to maintain a desired state was supported by both the conductivity and the linear pH gradient experiments (1). FDA's QbD initiative incorporates PAT as a crucial tool for designing and optimizing manufacturing processes. The PAT-based IBD system described in this article successfully implements the principles of this guidance and minimizes variability by adaptively correcting the quantity of raw materials around a set point. 

 is manager of science and technology, and 

 is vice-president of science and technology, all at Asahi Kasei Bioprocess, 1855 Elmdale Ave., Glenview, IL 60026, tel. 847.556.9716, fax 800.293.5059, 

. Shree Jariwala is an undergraduate student intern at Northwestern University. 

*To whom all correspondence should be addressed. 

PAT: A Framework for Innovative Pharmaceutical Development, Manufacturing and Quality Assurance

www.pharmamanufacturing.com/articles/2009/046.html

 4. T. Brown, H. LeMay, Jr. ,and B. Bursten, 

 (Prentice Hall, Englewood Cliffs, NJ, 5th ed., 1991), pp. 607–615.  

Articles

Featured Content

The authors describe the operational requirements and design of a process-ready PAT-based IBD system.

Process Analytical Technology-Based In-Line Buffer Dilution In Downstream Bioprocessing

Detecting and measuring metal elements in pharmaceutical products is an important consideration in finished-product pharmaceutical manufacturing. Possible sources of metal-object contamination include ingredient contamination, mixing-process errors, and machinery failures in the tablet- and capsule-forming and filling process. In these instances, a food-processing metal detector can play an important role in detecting magnetic or conductive metals. The author discusses new detection-system technologies that can improve performance and provides key criteria to consider when selecting or upgrading a system for pharmaceutical use.  

A typical detector of metal foreign objects contains a transmitter antenna that sends out a radiofrequency signal ranging from 300 kHz to 1 MHz (see Figure 1). Two receiver antennas sit on each side of the transmitter at equal distance from the transmitter. When the system is balanced and there is nothing magnetic or conductive inside the aperture of the metal detector, the difference of the two signals is zero, signifying that no metal is present. When metal is present and traveling through the detector, a detectable imbalance is created. The accurate performance of this system depends on three factors:   

The closeness of the metal that is being detected to the antennas or coils (i.e., the aperture or opening size)

The effectiveness of the fields created by the transmitter (i.e., the transmit-antenna design) 

The frequency of the signals used (the higher the frequency, the better the detection of conductive, nonmagnetic metals). 	

 Figure 1: Pharmaceutical metal-detection system showing detail of the rejection system (APEX 500 Rx Metal Detection System, Thermo Fisher Scientific). The system is placed at the outlet of the tablet press, deduster, or capsule filler.  Every product coming off the production line passes through its aperture and is metal detected. ALL FIGURES ARE COURTESY THERMO FISHER SCIENTIFIC   

The receivers' signals are digitized and analyzed by a digital-signal processor (DSP) that filters the signals. The DSP uses signal-processing algorithms to increase the probability of an accurate detection. The signals have two components: one is magnetic (X), and one is conductive (R). These components enable the system to detect metal foreign objects that are mainly conductive and have a small amount of magnetism such as in 316-alloy stainless steel. As a result, detection of conductive metal objects relies on a different signal analysis compared with a ferrous metal containing iron. Most metals exhibit both magnetic and conductive behaviors; these behaviors can change with the size of the metal. An off-line laboratory instrument such as an X-ray fluoresence spectrometer is necessary to determine full metal composition. 

In some applications, the ability of the system to ignore the signals that can be caused by the uncontaminated product passing through the metal detector is crucial. For example, some products may have a chemical composition that appears to be slightly magnetic or conductive to the metal-detector fields. This type of product effect can be ignored by the system by first learning the magnitude of the product's X and R signals. During production, the system creates a region where any combination of X and R signals with the same ratio and similar magnitude are thereby ignored. This process, called phasing, is typically only required in pharmaceutical applications using products that have high concentrations of iron or other metallic elements. 

Figure 2: Multicoil architecture in a metal detector results in higher sensitivity. RF is radiofrequency.  

The pharmaceutical industry has used metal-detection systems for more than a decade to detect metallic contaminants, and the systems are evolving to offer improved performance and ease of use. One new technology is the inclusion of multicoil arrangements in metal detectors to improve the signal viewed by the receiver. Compared with a single transmitter with two receivers, multicoil arrangements can improve detection performance (as measured by the diameter of the metal sphere size that can be detected) of the instrument by up to 20%. Scientists at the Thermo Fisher Scientific  Minneapolis, Minnesota, site for example, uses electromagnetic field simulation software to optimize the number and placement of such coils or antennas, which make it easier to detect smaller pieces of metal foreign objects without seeing higher levels of false rejections (see Figure 2). 

There are many criteria to consider when evaluating metal-detection systems for tablet and capsule processing. Perhaps the most important is the system's sensitivity (i.e., how small a metal fragment can be detected [the typical range for pharmaceutical applications is less than 0.5 mm]) and how immune the system is to outside influences such as vibration and electromagnetic noise, which can cause false rejects (see Table I). When selecting the size of the system's aperture, one should consider performance as well as throughput and size to avoid product jams during production. 

Table I: Typical detectable metal diameter versus metal detector's aperture size.  

 One subtle, but important, factor to consider when evaluating different systems is what, if any, effect a product might have on the metal detector. Most tablets and capsules look like "dry" products to the metal detector, but it is possible that they may trigger a false reject if they contain significant concentrations of metal elements such as iron. If a false reject continually occurs, the detector's operating threshold can be desensitized, but this change  may reduce the system's performance. Before selecting a system, therefore, it is important to test the complete system on all of the pharmaceutical products that may be used to determine how the system will perform in a production setting.  

Once it is determined that a system can detect the smallest fragments of metal possible with no chance of false rejection, there are other aspects of the system to consider to ensure that it can be successfully used in a particular production process. Table II provides a list of key questions to ask when evaluating a system.  

Table II: Key criteria for selecting a pharmaceutical metal detection system.  

One key consideration is how the detection system is audited. Typically, manufacturers audit detection systems every 1 to 2 hours during production to check for detection sensitivity (see Table I). Most of these audits involve simple pass/fail checks. An additional automatic early-warning system has been developed (Thermo Scientific AuditCheck) that can be set to periodically (i.e., based on user-defined preferences) pass a metal piece through the detector's fields to measure the two primary signals previously defined (X and R) (see Figure 3). These signals are calibrated during set-up, and any changes from the expected values can generate warnings or alarms to the user. An early warning allows the user to take immediate action such as cleaning the detector of foreign material or adjusting the detection threshold. As a result, production volume and time can be maximized. In addition, the automation of the audit can help to reduce labor resources. 

Figure 3: AuditCheck (Thermo Fisher Scientific) block diagrams. The diagram on the left is a side view of the metal detector, displaying where the product and AuditCheck device passes through the electromagnetic field.  The diagram on the right shows how the magnetic (X) and conductive (R) signals are calibrated with AuditCheck and how warning and alarm limits are set.  

After selecting and implementing a new or upgraded metal-detection system, a short on-line trial test is advised to further determine how the system is performing during production. In addition to the criteria outlined in Table II, one should consider the following: 

1. Did the system reliably detect the metal sizes and types in question  

2. Were there false rejects and, if so, why? Can they be avoided or minimized? 

4. Can the system be adjusted easily to the production process with no loss of material? 

5. Was set-up easy and how much operator intervention is needed to run the system?  

6. How is the system audited and documented? 

7. Can the system be cleaned easily? Can it be disassembled and reassembled quickly? 

8. Did the system produce any errors? If so, how did it react? 

Foreign-metal-object detection systems can help to prevent product contamination. Selecting a system requires a trial-and-error approach that takes into consideration the system's sensitivity, ease of use, and audit ability. Once implemented on a production scale, a metal-detection system can help to maximize product volume and quality.  

 is a lead product manager at Thermo Fisher Scientific, tel. 763.783.2500, 

The author discusses new detection-system technologies that can improve performance and provides key criteria to consider when selecting or upgrading a system for pharmaceutical use.

The Elements of an Effective Solid-Dosage Metal Detection Program

Jonathan Yourkin, global pharmaceutical product manager GE Analytical Instruments (Boulder, CO)

Total organic carbon (TOC) is a critical water-quality attribute. Reliance on periodic laboratory grab samples taken from the process is comparatively inefficient and less reliable than using on-line TOC analyzers located at critical locations within the water-distribution system. An important element of implementing on-line TOC is ensuring that the process analyzers use an analytical method that is equivalent to or better than the established laboratory method. Recognizing that most compendial methods were never intended for, nor designed to qualify continuous-process analyzers, the US Food and Drug Administration's guidance on process analytical technology (PAT) suggests that users consider any process analyzer used for real-time data as an "alternate analytical method" (1). Simply stated, the user must establish the suitability of the process analyzer for the intended use through prescribed method-validation procedures. Beyond establishing the suitability of on-line TOC methods, several important process-validation elements should be considered for complying to current good manufacturing practices, including evaluating variables within a measurement system.  

APOSTROPHE PRODUCTIONS, PHOTODISC, GETTY IMAGES  

Evaluating variables in a measurement system. 

A key challenge in any shift to on-line TOC analysis is ensuring that the measurement system that measures an analyte on-line provides the same or better data quality as the laboratory-based measurement system. Often, analytical methods used to produce on-line and laboratory data are different. These challenges can be solved when all analytical methods used are essentially the same or the analytical capabilities of the laboratory-based and on-line TOC methods have proven to be like-for-like. Once method comparability is established, any variability observed between laboratory and on-line TOC results can be attributed to different sampling processes. 

When comparing TOC data from similar samples analyzed by on-line and laboratory TOC instruments, observed variation can be attributed to two major sources: the sampling process used to introduce the sample to the measuring device and the inherent analytical capabilities of the analytical method. Given the ubiquitous presence of organics in the environment, grab-sample collection of TOC samples is subject to numerous error sources. With the on-line method, sampling-process variability is minimized or eliminated because the analyzer is physically integrated within the process-water stream for continuous quality-assurance monitoring. In addition, this evaluation quantifies sampling-process variation at TOC concentrations produced by the pharmaceutical water system, which are typically an order of magnitude lower than the TOC standard concentrations used to validate the analytical method. Sampling process variation is generally much more pronounced at these lower TOC concentrations.  

 Equivalency testing demonstrates the sameness of two measurement systems based upon the analytical results the methods produce. This type of testing differs from instrument-validation testing, which typically assesses various attributes using standards of known concentration. Validation protocols are not designed to identify differences in the test-sample data quality. Therefore, it is possible to validate two methods with the same protocols using the same criteria, yet have the methods yield nonequivalent analytical results under actual conditions of use. It is particularly important in on-line TOC implementations to augment validation protocols with equivalency testing. The fundamental hypothesis supporting the justification for measurment-system equivalency testing is that, given the sameness of on-line and laboratory analytical methods, any lack of equivalency observed between the two data sets can be attributed to aspects of the measurement system rather than the core analytical method.   

This type of analysis is typically performed using a Gage Repeatability and Reproducibility (Gage R&R) study to quantify measurement-system variation relative to variation of the process as a whole. The Gage R&R addresses several components affecting variability, including the gauge (the instrument), the operator, and the sample itself. Ordinarily, a full-scale Gage R&R would be used to assess variability contributed by all components. This example uses a simplified Gage R&R approach based on the degree of sameness of the analytical methods and the unique differences in the operational or environmental conditions associated with laboratory and on-line measurements. By pairing the on-line and laboratory data collection, process water variability remains common between the two systems. In addition, the inherent accuracy of the instruments remains common, based on like-for-like methods that have been validated appropriately. 

Figure 1 (TOC) is an example of a 30-day measurement-system equivalancy test protocol from a pharmaceutical company. In this protocol, the statistical parmameters of mean and standard deviation represent data-quality equivalency. The mean is determined from repeated measurements of the process. This average includes contributions from the pharmaceutical water itself, the sampling process, and the inherent accuracy of the measurement device. By pairing the on-line and laboratory data collection, the process water variability remains common between the two systems. In addition, the inherent accuracy of the instruments remains common, based on previous method validation results. The mean serves to quantify the added TOC contributions of the sampling process itself. Additionally, the standard deviation is used to assess the repeatability of the measurement system, taking into account the variation of the sampling process and the measurement device. 

 Figure 1 (TOC): A 30-day measurement-system equivalency test of off-line and on-line total organic carbon (TOC) analysis. (FIGURE 1 (TOC) IS COURTESY OF GE Analytical Instruments) 

Simon Robinson, HPLC product manager, Shimadzu Scientific Instruments (Columbia, MD)

Over-the-counter cold medicines often contain multiple active ingredients. These actives include combinations of decongestants, antihistamines, pain relievers, cough suppressants, and expectorants in addition to numerous vitamins and herbal extracts—all of which exhibit different chemical properties. A wide range of analyte polarities often makes developing chromatographic methods challenging. Separating ingredients in a cold medicine is important to understand the content of each medication and product quality. A high-speed liquid chromatographic system (Prominence UFLCxr with a SPD-M20A photodiode array detector, Shimazdu Scientific Instruments) was used to analyzed 20 common cold medcines.  

 A standard mixture of cold-medicine ingredients was prepared (dissolving them in Mobile Phase A/acetonitrile = 1/1 (v/v), 100 mg/L, each) and analyzed using the conditions shown in Table I (UF LC). The mixture consisted of thiamine, acetaminophen, caffeine, riboflavin, hesperidin, ethenzamide, chlorpheniramine, ambroxol, noscapine, isopropamide, ispropylantipyrine, dextromethorphan, glycyrrhizin, bromhexine, clemastine, and ibruprofen. The contents of a capsule of cold medicine was dissolved in 100 mL of Mobile Phase A/acetonitrile = 1/1 (v/v), or in the case of tablet form, a single tablet was dissolved in 50 mL of the previously described mobile-phase acetonitrile solution. The resultant solution was passed through a 0.22 µm pore membrane filter and analyzed using the conditions in Table I (UFLC). 

Figure 1 (UFLC): Results of five serial ultra-high-speed liquid chromatographic analyses. (FIGURE 1 (UFLC) IS COURTESY OF SHIMAZDU SCIENTIFIC INSTRUMENTS) 

Figure 1 (UFLC) shows the results of five serial ultra-high-speed analyses of the cold medicine. Using UFLC, the autosampler cycle time becomes a critical factor. The injection speed of the autosampler was 10 s, and the total time for all five analyses took no more than 5 min. This cycle time, in conjunction with a reduction in carryover, increased sample throughput. The run time of a single analysis was shortened, and the total cycle time of the injection sequence and run time was optimized.  

Table I (UFLC): Analytical conditions. (TABLE 1 (UFLC) IS COURTESY OF SHIMAZDU SCIENTIFIC INSTRUMENTS) 

Lori Daane, PhD, vice-president of Celsis Rapid Detection, and Judy Madden, vice-president of Celsis International (Chicago) 

Microbiological tests may fall into several categories. Final product-release testing includes microbial limits testing of nonsterile drug products and sterility testing of sterile products. Environmental monitoring tests air, water, surfaces, and personnel for viable microorganisms. Bulk drug products, including active pharmaceutical ingredients and excipients, may be tested for bioburden (1). Rapid microbial screening is used to detect the presence or absence of microbial contamination within 24 h compared with 3–7 days using traditional microbial testing. For example, Celsis AKuScreen is a rapid microbial method that can detect contamination from bacteria, yeast, or mold definitively within 24 h.  

Rapid microbial methods can be based on adenosine triphosphate (ATP) bioluminescence, which is commonly known as the firefly luciferase reaction as follows, where AMP is adenosine monophosphate and PPi is inorganic phosophate (2): 

ATP drives the reaction, which emits light that is measured in a luminometer. ATP is present in all living cells. If microbial contamination is present, the reaction moves forward, and light is emitted. Because definitive detection is based on having a sufficient quantity of microbial ATP present to generate a light signal that is distinguishable from sample background, detection becomes more difficult if the sample contains nonmicrobial ATP. Enzyme and molecular-based technologies, such as adenylate kinase (AK) or AK-amplified assays (i.e., Celsis AKuScreen, Celsis), allow bacteria to be detected in 18 h and yeast and mold in 24 h. The AKuScreen assay uses microbial adenylate kinase (AK) to generate ATP in a reaction that is not constrained by the finite amount of metabolic ATP available in the standard bioluminescence assay. The generation or amplification of ATP beyond that inherent in the assayed sample is accomplished 

the enzyme-catalyzed and reversible reaction as follows:  

The AK reaction uses adenosine diphosphate (ADP) and microbial AK to catalyze the production of ATP. The produced ATP is detected and measured using the traditional firefly luciferase reaction. The longer the AK reaction is allowed to proceed in the presence of an adequate supply of ADP, the more ATP is generated and the greater the bioluminescence signal and the resulting assay sensitivity. Assuming the ratio of AK and ATP in a typical bacterium and a reaction turnover number of 40,000, approximately 40 times as much ATP as originally represented by microbial ATP can by generated per minute by the AK reaction (2, 3). Similarly, if the reaction is allowed to continue for 25 min, the amount of ATP can be 1000 times more than what was originally present, allowing  the test-enrichment period and time-to-result  to be reduced. 

Case study : in-process sterility testing

. Because of the ubiquity of AK in living organisms, AK-amplified assays can detect bacteria, molds, and yeasts, and this microbial testing is used in sterility testing to determine product release (4, 5). A company was testing its sterile, in-process materials using a 14-day sterility test. A comparability test was performed to see if equivalent results could be obtained using a rapid method. Spiking studies were conducted by inoculating samples at a level approximating 1 colony-forming unit (CFU). Following inoculation, the samples were enriched for 24 h, and contamination was detected using a rapid microbial system (Celsis AKuScreen). Microbial hold times and the corresponding time for quarantined inventory was 48 h, a savings of 12 days over traditional microbial testing. 

 Pharmaceutical companies use microbial specifications that are based on 

 guidelines, product history, and risk profile. With respect to mold, this typically means the ability to detect 

. Slow-growing, molds are often the rate-determining step for product release. AK-amplified assays (e.g., Celsis AKuScreen) can detect yeast and mold in 24 h compared with 48 h if standard bioluminescence is used or 4–7 days if using traditional methods. 

One company used standard ATP screening placed microbial specifications on certain products to include the detection of 

 The company modified its bioluminescence protocol using a rapid microbial system. A study was performed on a water-based lotion, examining how broth type, incubation temperature, and conditions affect 

 detection using standard bioluminescence and AK-amplified reageants (e.g., Celsis AKuScreen). Preliminary studies indicated that static incubation at room temperature optimized the growth of 

 Table 1 (rapid microbial testing) summarizes the study data in which the samples were inoculated with < 50 spores of 

 and incubated statically at room temperature (24–27 °C). The modifications of room temperature incubation and broth changes to letheen or malt extract broth (MEB) enabled detection with the AKuScreen reagents within 24 h. Detection with standard bioluminescence required 48 h using MEB and 72 h in letheen under the same incubation conditions.  

Table I (rapid microbial testing): Detection times for slow-growing mold.* (TABLE 1 (RAPID MICROBIAL TESTING) IS COURTESY OF CELSIS) 

Rapid microbial testing section references

 3. D.J. Squirrel and M.J. Murphy, "A Practical Guide to Industrial Uses of ATP-Luminescence" in 

Cara Technology, Lingfield, Surrey, UK, 1997).   

 ( Academic Press,  New York, 1973), p. 279–305.  

Cold Spring Harbor Symposia on Quantitative Biology

 6. L. Dane, "Study Results: Modifications to Detect

 in 24 Hours Using AKuScreen Reagents" (Celsis, Chicago, 2006). 

Kevin P. Menard, PhD, business manager of thermal analysis at PerkinElmer (Shelton, CT)

Differential scanning calorimetry (DSC) and Raman spectroscopy are well-known techniques in pharmaceutical analysis. DSC determines glass transitions, amorphous content, melting temperatures, and enthalpy and estimates the degree of crystallinity. In its fast-scanning mode, DSC also suppresses kinetic changes (i.e., polymorphism and decompositions). This allows measurements to be made before kinetic changes can occur, allowing one to determine the initial polymorphic form as well as measure melting  and heat capacity on materials before they decompose. Raman spectroscopy is able to detect changes in chemical composition as well as positional and stereochemical changes in a sample; this allows it to identify specific polymorphic states. Raman spectroscopy, however, has some difficulties in measuring temperature-dependent reactions such as dehydrations and polymorphic rearrangements. These reactions are sensitive to the temperature applied, and care must be taken so that the energy added by the Raman's laser does not affect the data by causing changes in the sample temperature  (1). A combination of dual furnace (i.g., power-compensated) DSC and a shuttered laser in an Eschelle Raman spectrometer were found to give minimal increases in bulk sample temperatures (2). 

As a hyphenated technique, however, DSC–Raman spectroscopy works well in detecting the changes in polymorphic materials as a function of temperature (3). For example, running a sample of acetaminophen in heating showed a series of peaks corresponding to changes in polymorphic forms, but the changes in the material corresponding to those thermal events are inferred (2). For example, an endothermic peak in the DSC themogram can be a melt, a water loss, or a polymorphic change. Raman spectroscopy allows one to confirm what the transition is; the chemical and structural changes detected from the Raman spectra clarify the DSC data. Because the DSC can performed under a wide range of  thermal conditions, including heating and cooling rates from 0.1 to 750 ° C/min, the combined techinques can be used to characterize the materials' kinetic behavior under a wide variety of conditions. For example, crystallization processes can be studied using DSC–Raman during cooling experiments. Other work has been reported on hydrates and pseudo polymorphs as measuring the reaction of materials by tracking changes in assigned bands in the Raman spectra and the energetic changes (4, 5). 

DSC–Raman spectroscopy section references  

Proceedings of the  North American Thermal Analysis Society (NATAS) Annual Conf. 

Proceedings of the Annual Technical Conf. of the Society of Plastic Engineers

(Albuquerque, NM, 2003), pp. 120–129;  

Several industry experts describe applications in pharmaceutical applications, including on-line total organic carbon analysis, ultra-fast liquid chromatography, rapid microbial testing, and differential scanning calorimetry-Raman Spectroscopy.

Analytical Applications

Routine, ongoing assessment of process performance and product quality is crucial to ensuring that high quality pharmaceuticals reach the patient in a timely fashion. In traditional solid-dosage pharmaceutical manufacturing, process data are routinely analyzed at two points in time to assess process stability and capability. During the production of each batch, process operators and quality-control departments collect data to ensure stability and capability and take appropriate remedial actions when needed. On a less frequent basis (i.e., monthly or quarterly), batch-to-batch variation is analyzed based on product parameters to assess the long-term stability and capability of the process. The author discusses this system, including key challenges, and describes a structured approach, including the role of quality by design (QbD), for operating the system effectively.  

The industry and regulatory focus on QbD places an even greater emphasis on the quality of pharmaceutical products and the performance of the pharmaceutical-manufacturing processes. Process and product control are major building blocks of QbD (1). The strategic structure of the International Society for Pharmaceutical Engineering's (ISPE) Product Lifecycle Implementation Plan (PQLI) lists the "process performance and product quality monitoring system" as one of its critical elements (2).  

 Central to any system is the assessment of process stability and capability. Manufacturing processes that are stable and capable over time can be expected to consistently produce product that is within specifications and thereby cause no harm to patients due to nonconforming product. Stability and capability are described as follows (3, 4): A stable manufacturing process is a process that is in a state of statistical control as each batch of tablets is being produced and as batches of tablets are produced over time. A process in a state of statistical control consistently produces product that varies within the process control limits; typically set at the process average (X-Bar) plus and minus three standard deviations (SD) of the process variation for the parameter of interest. Separate control limits are set for each parameter (e.g., tablet thickness and hardness). Any sample value that falls outside of these limits indicates that the process may not be in a state of statistical control.  

A capable process is one that consistently produces tablets that are within specifications for all tablet parameters (3, 4). A process-capability analysis compares the process variation to the lower and upper specification limits for the product. A broadly used measure of process capability is the Ppk index, or process performance index, which is discussed in greater detail later in this article. 

A production process can include any one of the four combinations of stability and capability: stable and capable (desired state), stable and incapable, unstable and capable, and unstable and incapable (worst possible situation). 

Process stability and capability are typically evaluated twice:  

1) During the production of each batch to ensure that the process is in control and to identify when process adjustments are needed. Some key questions that need to be addressed during this analysis include:  

 Is the batch production process stable during the production of the batch with no trends, shifts, or cycles present? 

Is the process capable of meeting specifications (i.e., are the process-capability indices acceptable)? 

Is the within-batch sampling variation small, indicating a stable batch production process?

2) Monthly or quarterly to ensure batch-to-batch control throughout a given year and between years. Some important questions that should be addressed during this analysis include:  

 Is the batch-to-batch variation stable from year to year and within years with no shifts, trends, or cycles present? 

These two analyses also help to assess the robustness of the process. 

Control limit versus specification limit. 

Control limits are calculated from process data and applied to the process. Control limits are used to assess the stability of the process and to determine the need for process adjustments when out of control samples are detected. On the other hand, specification limits apply to the product. Specification limits are used to assess the capability of the process to produce a product that has the desired properties and characteristics. 

It is generally agreed among industry that a systems-based approach enables operations to be more efficient and sustainable. Schematics of such systems are shown in Figure 1 for monitoring individual batches and in Figure 2 for monitoring batch-to-batch variation (5–7). The systems underlying Figures 1 and 2 have the following characteristics:  

Data are periodically collected from the process. Pharmaceutical manufacturing processes are often monitored using 30–60 minute samples. 

These data are used to monitor processes for stability and capability using control charts, process capability indices, analysis of variance, time plots, boxplots, and histograms. 

The analysis identifies when process adjustments are needed to get the process back on target.

Records are kept on the types of problems  identified. As significant problems are identified or problems begin to appear on a regular basis, the resulting issues and documentation are incorporated into process-improvement activities to develop permanent solutions.

 Figure 1: Framework example for monitoring process stability and capability. ALL FIGURES ARE COURTESY OF THE AUTHOR  

Process improvement can be effectively completed using the define, measure, analyze, improve, control (DMAIC) problem-solving and process-improvement framework (5, 6). The following section describes the tools typically used in this framework. 

Figure 2: Framework example for monitoring batch-to-batch variation over time. 

As a general principle, it is rare that a manufacturing process that is stable and capable will produce a product that is out of specification. The primary purpose of a process monitoring system is to address the question: Is this process capable of consistently producing product that is within specifications over time? The statistical analyses conducted to answer this question are briefly described below. These methods are generally accepted and well documented in the literature (4). 

 A control-chart analysis is used to assess the stability of a process over time. The Shewhart chart has been widely used to assess process stability since the 1930s. Other types of control charts are also useful for monitoring processes (4). 

A stable process is a predictable process; a process whose product will vary within a stated set of limits. A stable process is sometimes referred to as being in "a state of statistical control" (3, 4). A stable process has no sources of special-cause variation—that is, effects of variables are outside the process but have an effect on the performance of the process (e.g., process operators, ambient temperature and humidity, raw material lot).  

The most commonly used indicator of special-cause variation is a process that has product measurements outside of the control limits which are typically set at X-Bar plus and minus three SD of the process variation for the parameter of interest (see Equation 1).  

For example a process may be producing tablets with an average hardness of 4.0 kp and a standard deviation of 0.3 kp. The control limits are thus 4.0 +/–3(0.3) for a range of 3.1–4.9. Any tablet sample outside of that range is an indication that the process average may have changed and a process adjustment may be needed. Separate control limits are set for each parameter.  

Figure 3 shows a control chart for a total weight of 10 tablets manufactured by two tablet presses. The graphic demonstrates that both presses are stable and in control and producing tablets with the same average weight and variation. 

Figure 3: Control chart showing batch tablet weight produced using two presses (A and B). UCL is upper control limit and LCL is lower control limit. 

Figure 4 shows a control chart for assay values for batches produced over a 3-year period. The process is stable through the middle of Year 2 and begins to decrease in Year 3. When a process adjustment is made, the batch assay values return close to the values observed in Year 1. This example is interesting because a process shift is shown, but none of the batch assay values are close to the assay specifications of 90–110%. 

Figure 4: Control chart showing assay values of batches produced over three years. UCL is upper control limit and LCL is lower control limit. 

Out-of-specification (OOS) and out-of-control (OOC) values require investigation. These values are not always caused by manufacturing problems, and may be caused by sampling errors, testing errors, or human administration errors such as recording or data keying. The causes of OOS and OOC measurements should be carefully considered when interpreting the OOC and OOS values and deciding on appropriate action.  

A process-capability analysis is conducted to determine the ability of the process to meet product specifications. The Ppk value represents the ratio of the difference between the process average and the nearest specification divided by three times the process SD (see Equation 2).  

Where A is the upper specification minus the process average and B is the process average minus the lower specification. Two main statistics are used to measure process capability: percent of the measurements OOS and the process Ppk value. The interpretation of the Ppk value is summarized in Table I. 

Table I: Summary of process performance index (Ppk value) interpretation. 

Process capability indices of Ppk of 2.0 and higher are consistent with high-performance processes or robust processes. Figure 4 shows an example of process capability for tablet weight. In this case, the Ppk value is 1.91 (an excellent category) which is based on the distribution of tablet weights being a considerable distance from the lower and upper specifications for tablet weight.  

Figure 5: Process capability for a tablet weight of Ppk = 1.91. LSL is lower specification limit and USL is upper specification limit. 

When a process is robust, small process upsets will not create OOS product. Accordingly, small OOC signals do not result in OSS product. A process is said to be "robust" if its performance is not significantly uninfluenced by variations in process inputs (e.g., raw material lot), process variables (e.g., press force and speed), and environmental variables (e.g., ambient temperature and humidity). 

 Another way to assess process stability is to study the variation in process performance that is caused by potential special-cause variation (e.g., tablet presses, raw-material lots, process operating teams). Analysis of variance (ANOVA) enables one to identify variables that can increase variation in tablet parameters and that may produce OOS product (see Figure 6).  

Figure 6: Boxplots showing tablet hardness for two tablet presses (X and Y). 

The boxplot in Figure 6 shows the distribution of tablet hardness values for a batch of tablets produced by two different tablet presses (X and Y). Press X has a wider hardness distribution than Press Y, yet none of the hardness values are outside the hardness specification of 1–6 kp. With this data in mind, the process operators can determine whether to make process adjustments. 

After statistical significance of a comparison (e.g., average of Tablet Press X versus average of Tablet Press Y) is established, the practical significance of the difference in average values must be considered. This assessment is frequently carried out by expressing the observed difference in average values as a percentage of the overall process average. Subject-matter expertise is used to evaluate the practical importance of the observed percent difference.  

Nested analysis of variance is another form of ANOVA used to assess process stability. Nested ANOVA can estimate the portion of the total variation in the data attributed to various sources of variation. Typically, the larger the percent of the total variation attributed to a source of variation, the more important is the source of variation. Low amounts (< 30%) of long-term variation as determined by a nested ANOVA indicate a stable process. 

The author's experience has shown that there are important considerations to take into account when designing, implementing, and operating process monitoring systems. The first of these is process understanding—that is, a deep knowledge of the variables that drive the process, which enables the accurate prediction of process performance. Effective application of QbD will result in better process understanding. 

It is also crucial to understand the sources and magnitudes of measurement variation, in particular the repeatability and reducibility of the measurement of the process parameters. Gage R&R studies are an effective method for measuring the repeatability and reproducibility of the measurement methods used (6). Ruggedness studies are effective for determining method robustness for measuring typical variations that occur during the routine use of the method (8). 

A systematic method is necessary to keep track of special-cause variation and to determine whether a systemic problem exists. This information can then be used to improve the process. 

The author outlines key considerations for carrying out a structured approach to monitoring process performance and ensuring product quality.

Crucial Considerations in Monitoring Process Performance and Product Quality

Small-angle X-ray scattering (SAXS) offers various ways to characterize drug-delivery systems and large molecules. Understanding the structure of drug-delivery systems and large molecules at a molecular level is a crucial step in designing drugs and drug-delivery systems alike.  

The SAXS technique can provide insights into structures in the 1–100 nm range. SAXS requires little or no sample preparation and enables scientists to run experiments at or close to 

Röntgen discovered X-rays in 1895. In 1912, Laue discovered the diffraction of X-rays by crystals (1). Guinier's work in late 1939 led to the main principles of SAXS (2). In the 1940s and 1950s, Otto Kratky investigated X-ray diffraction at small angles as a technique for the structural analysis of macromolecules. He developed the SAXS method into a powerful tool for structural research, particularly in the field of polymers and molecular biology (3). Considered one of the fathers of SAXS, Kratky founded the Institute for Physical Chemistry in Graz, Austria, which became an early center for this technique. The institute led to many advances in SAXS such as the first commercial instrument for SAXS.  

Early SAXS experiments took place in laboratories. In the 1970s, the availability of synchrotrons and high-intensity synchrotron radiation helped bring the technique to prominence. In recent years, technical advances have made laboratory-based SAXS instruments attractive again. 

SAXS is a form of X-ray diffraction that focuses on small scattering angles. Scattering intensities at large angles (i.e., wide-angle X-ray scattering) contains information about small objects such as crystalline structures. Small angles contain information about large objects such as particles, macromolecules, and micelles. Figure 1 shows this inverse relation. This article will use the terms "particles," "macromolecules," and "micelles" interchangeably.  

Figure 1: Positive interference of two spherical waves from an electron pair can be seen (left) at larger angles if the electrons are close to each other (e.g., in a crystalline structure) and (right) at smaller angles if the distance is greater (e.g., in a macromolecule). (ALL IMAGES ARE COURTESY OF THE AUTHOR) 

Although its name contains the word "angle," the scattering vector 

 is common for SAXS. Also known as momentum transfer, 

 reflects the process in which the X-ray photons transfer energy to the electrons with which they interact. The following equation describes the relation of the length of the scattering vector 

 and the scattering angle 2θ commonly used in other X-ray diffraction methods (4):  

) in Equation 2 is the geometrical representation of the object in the beam. 

) maps the distances of all electron pairs inside the particle. The scattering intensity 

However, representing a three-dimensional (3D) object with a one-dimensional distribution function necessarily omits some information. Converting 

) into a three-dimensional object becomes difficult and requires additional constraints by the scientist such as connectedness or compactness. 

The main technical challenges of SAXS are to separate the strong primary X-ray beam from the weak scattering information close to the primary X-ray beam and to increase the intensity of the scattered X-rays. 

Scientists have used different concepts to achieve these goals. However, all SAXS instruments feature an X-ray source, a collimation system focusing optics, a sample holder, a beam stop, and a detector (see Figure 2).  

Figure 2: A typical setup for a small-angle X-ray scattering instrument includes a source, collimation system with optics (i.e., mirror), sample holder, and detector. The area that the beam enters after passing the mirror is typically evacuated to reduce unwanted scattering from gas molecules. 

To date, copper anodes, producing Cu–Kα radiation, have been the best compromise between sensitivity and absorption. The following three technologies are currently in use:  

In sealed-tube sources, the electrons are accelerated toward a fixed anode. Part of the energy is emitted as X-rays, but the majority is converted to heat. In sealed-tube sources, the anode is fixed, so a limited amount of energy can be dissipated before the copper melts. Rotating-anode sources exceed this limit and avoid local overheating (5). 

Microfocus sources use electromagnetic lenses to focus the electron beam to a fine point, thus producing a high-brilliance X-ray beam. 

Scientists typically consider operating costs, energy consumption, and flexibility when choosing an X-ray source. They also must determine whether the source will match with other components of the SAXS system. 

Focusing optics increase the photon flux through the sample. Visible-light lenses and electromagnetic fields such as electron beams cannot focus X-rays. Multilayer mirrors reflect X-rays on a layered surface under the Bragg condition (see Equation 3). Depending on the layers' thickness, a certain wavelength positively interferes (i.e., reflects) at a specific angle (5).  

By reflecting certain wavelengths at certain angles, multilayor mirrors can produce monochromatic X-rays. The mirror's curvature produces the desired beam focus, typically parallel or converging.  

The monochrome and focused beam then travel through the collimation system to produce a precise beam profile. The following two collimation methods are common:  

Point collimation (i.e., a pinhole system)

Point-collimation systems send a point-shaped beam through a series of three pinholes. The first two pinholes produce a parallel beam, and the third pinhole blocks unwanted scattering produced by previous pinholes. 

Line-collimation systems use a line-shaped beam and employ slits instead of pinholes. The block collimator, a compact and effective design to eliminate parasitic scattering by the collimation system itself, was suggested by Kratky (4). 

Line collimation increases the photon flux through the sample by replacing a single point-shaped beam with the equivalent of many such points arranged in a line. The increase in intensity comes at the cost of a smearing effect on the scattering image (4). Figure 3 shows the effects of line collimation.  

Figure 3: Scattering patterns of silver behenate. The liquid crystalline structure with clear concentric rings in the case of a point-shaped beam profile (i.e., point collimation) (right) becomes smeared when using a line-shaped beam (i.e., line collimation) (left). The line-collimated experiment yields higher intensity. Reversing the smearing effect requires knowledge of the beam profile and position. 

Desmearing data is a mathematical operation that requires knowledge of the beam profile that causes the smearing effect (6–8). New technologies report the beam profile and position so that desmearing becomes a simple and precise routine, thus alleviating earlier concerns (9, 10).

X-ray detectors must be able to capture faint scattering signals. Scientists currently use the following technologies (5):  

The pixel resolution determines the required distance between the detector and the sample. High-resolution detectors can be brought closer to the sample, but low-resolution detectors should be positioned further away from the sample. 

The intensity of the scattered X-rays decreases with as distance increases, according to the following quadratic law:  

 to the sample receives only 25% of the intensity of a detector with a distance 

Typically, a beamstop protects the detector from exposure to the direct beam, which is several magnitudes more intense than the scattered X-rays. 

Dilute systems allow the analysis of particlate and intraparticlate structure. Interaction between particles can be observed with highly concentrated samples. 

The author describes the development of small-angle X-ray scattering and analyzes its advantages in the characterization of drug-delivery systems and large molecules. This article is part of a special Analytical Technology issue.

Pharmaceutical Technology-10-01-2010

Process Analytical Technology-Based In-Line Buffer Dilution In Downstream Bioprocessing

The Elements of an Effective Solid-Dosage Metal Detection Program

Analytical Applications

Crucial Considerations in Monitoring Process Performance and Product Quality

Small-Angle X-ray Scattering for Pharmaceutical Applications

Quantitative Open-Access HPLC Analysis

Investigating the Influence of Glycosylation on Protein Conformation and Dynamics