Outsource Analytical Testing for QC

By Paul Smith Can save money, provide essential flexibility
Outsourcing your analytical testing can be one of the best decisions you ever make or cause sleepless nights for years to come if your outsourced results are scrutinized during a U.S. Food and Drug Administration (FDA) or other regulatory audit. The outcome will depend on your approach to selecting your outsourcing partner and the working relationship you develop with that company. All analysis must be carried out to meet a valid business decision or requirement; contract analysis is no different.

Analysis costs money, and laboratory analysis is outsourced to save money. During the lifetime of a drug product, the approach to costs changes—from research and development (R&D) to routine supply to product maturity (generic manufacture). For example, during the R&D phase, during which the process, chemistry, sources of materials (and therefore potential impurities), and formulations may change, the emphasis is on time and maximizing the analytical information available to make timely scientific decisions on an evolving platform. This can result in the use of hyphenated analytical techniques and technology that may:

• not routinely be available outside of an R&D facility;
• not be sufficiently robust for a routine quality control (QC) laboratory; and/or
• not be required to support routine current good manufacturing practice (cGMP).

In contrast, QC analysis carried out to support routine manufacture has a strong cost focus, which becomes stronger as a product matures into generic manufacture. The business and analytical requirements change during the lifetime of a drug. For pharmaceutical contract QC analysis, the drug company usually has to perform analytical technology transfer to allow the contract QC laboratories to perform the testing. Providing documented evidence of the successful transfer is fraught with potential problems, especially when any form of analytical technology transfer occurs or when there are problems with the core technology; for example, the contract laboratories may not have access to necessary analytical techniques, such as nuclear magnetic resonance.
Anything that is outside the core expertise of the contract laboratories will be a potential high-risk area. Therefore, even the transfer of gradient high performance liquid chromatography (HPLC) methods could be a problem if the method was not robust or if the laboratory is used to isocratic HPLC systems.
These examples are mentioned because the nature of the analysis requirements directly affects:

• the type of contract testing laboratories needed;
• the process used for deciding on a supplier;
• the response time required for the results;
• the approach to analytical quality normally used; and
• the time component of the decision.

In addition to QC testing, transfer of a manufacturing process to a different facility will require environmental waste stream monitoring as part of the process optimization and discharge consent limit compliance—depending on what the waste streams contain, local legislation, and the consent limits.
Additionally, unless a fully dedicated manufacturing facility is being used, analysis of cleaning validation samples is usually required. For contract facilities new to cGMP manufacture, it is critical that any “history of use” information covers use of the facility prior to cGMP use and a risk assessment should be performed. If this information is not available, how do you know the facility was not used to manufacture pesticides or other potential high-risk materials?

Click to Enlarge

Figure 1: Contrasting analytical requirements for quality control, environmental, and cleaning validation.

Different Types of Contract Analysis

Aspects of contract pharmaceutical QC analysis, environmental testing, and cleaning validation testing are highlighted in Figure 1 (see below), which shows some of the differences among these analytical sample types.
There is usually strong competition among service providers, along with well-established accreditation for laboratory collaboration opportunities to demonstrate compliance and conformance. As shown in Figure 1, however, the approach taken to sample analysis can be very different to the kind used in routine pharmaceutical QC analysis. Using HPLC as an example, in pharmaceutical QC testing, samples are tested in an analytical “run” or injection sequence, whereas pharmacopeial system suitability requirements (e.g. USP <621>) define standardization and criteria for demonstrating that the results are valid within the run. The run or injection sequence is repeated each time samples are tested, and the ratio of setup injections is high (e.g., it may take several hours to test one sample).
Additionally, analytical equipment may not be dedicated to a particular kind of testing, forcing the analyst to take time each day to set up the HPLC system for testing samples. In contrast, efficient environmental testing laboratories will dedicate analytical equipment to particular methods and testing that can run the analytical methods 24/7.
Generally, this approach provides greater stability for HPLC equipment, which is especially important for trace analysis; fewer analytical problems are experienced with dedicated use of HPLC. For environmental testing, this dedication and 24/7 operation means that as soon as a particular sample requires testing and sample preparation is complete, it is added to the run. Overall, this approach reduces the cost per sample and increases the return on investment for the equipment. The impact of instrument downtime for repair or maintenance becomes much more significant where equipment is dedicated, however.
Outsourcing saves money. Sometimes, more importantly, it provides essential flexibility in resource-constrained QC or a development department, allowing more projects to be supported. The outsourcing service provider’s flexibility is another important consideration; certainly, testing laboratories with limited flexibility would constrain any benefits they might offer.

Click to Enlarge

Figure 2: Components of the data quality triangle based on USP <1058>.

Analytical Data Quality

In the June/July edition of Pharmaceutical Formulation & Quality, Lori Valigra discussed analytical instrument qualification and introduced the quality data triangle from USP general chapter <1058> as part of her article, “Qualifying Analytical Instruments: General chapter <1058> clarifies terminology, classifies instruments.”¹ This critical data triangle from USP <1058> is reproduced in Figure 2 (see p. 18). The principles are fundamentally applicable to all analytical testing laboratories.
A few differences exist among laboratories in relation to the data triangle shown in Figure 2, however:

• the level where the greatest emphasis is placed;
• the detail applied at each level; and
• how the information is documented.

Failing to ensure the performance of the analytical instrument (by instrument qualification) or the method (by method validation) introduces a risk that the results may be invalid. The deceptively simple question, “How do you know your analytical results are valid?” almost always relies on all aspects of the data triangle for a full and robust defense of the question (evidence that all the levels are important for valid data). The different ways in which the data triangle is interpreted can be both a risk and an opportunity. Change can be slow in large pharmaceutical companies, particularly when there is an opportunity to relax standards to appropriate GMP levels.
Therefore, contracting out services can be used to facilitate and accelerate change; for example, a company may adopt the standards of work that the contract lab uses, standards that are appropriate to the work, rather than the company’s own default standards. There are risks: Any change can become an area of focus in a regulatory audit, and adopting different standards can mean an increased risk from the standards of work applied in the contract laboratories.
Contract environmental testing laboratories will place significant emphasis on the results obtained from the analysis of control samples (the top of the triangle). The results of these control samples are plotted in control charts such as Shewhart charts, both to demonstrate the quality of the results and to demonstrate and monitor the ongoing performance of the analytical system, including equipment, method, and analyst. Additionally, contract environmental testing laboratories are often ISO 17025 accredited, so the results are reported with uncertainty figures. For contract environmental testing laboratories, the ongoing system suitability is particularly significant in defending the question “How do you know your results are valid?” For cleaning validation samples, good design of the analytical method, run injection sequence, and sample testing can mean that the results are reported as a pass or fail against the limit test being applied, rather than numerically as in pharmaceutical QC testing.
This type of “self-contained” screening analysis, where results are reported as pass or fail, rather than numerical values, can be much more efficient. Emphasis is focused on the samples that fail or are close to the limit. These are the high-risk samples, where further cleaning of the process equipment is required. However, this approach needs careful consideration at the method development and validation stage, because it is fundamental to how the method will be used.

Technology Transfer

When the analysis is performed with an in-house method that is not internationally accepted, the QC laboratory that developed the method is responsible for training and transferring analytical capability to test the samples to the contract laboratory. This is where many problems can occur. The robustness of the analytical methods and the capability and expertise of the scientists in the contract laboratories becomes significant. It is at this stage that the partnership relationship, essential for the success of the services being contracted out, is developed. This relationship is particularly important if analytical problems arise.
A technology transfer protocol documents and demonstrates how the analysis capability is transferred from the pharmaceutical company to the contract laboratories. However, a poorly designed protocol can prove that the contract laboratories’ results are different. Care must be taken when selecting acceptance criteria for subjective tests such as appearance and solution color and for low-level impurities. Over-reliance on student t-test comparisons can be troublesome—and not necessarily appropriate—for comparing low-level impurities. Analytical problems and out-of-specification (OOS) result investigations are well-known areas of high FDA compliance focus and, therefore, potential risk. Because of this high risk, OOS investigations, as well as the laboratory’s procedure for returning an instrument to routine use following maintenance or a breakdown, should be scrutinized when evaluating contract laboratories.
As part of the analytical technology transfer process, the analytical instrumentation used at both sites must be considered. If gradient HPLC instruments are from different suppliers, the “dead volume” and other performance attributes of the instruments can play an unseen problematic role in the results obtained. In my experience of technology transfer, there are generally fewer problems if both laboratories use the same analytical equipment for some tests, including HPLC.
Sometimes this is not possible. For example, the contract laboratory may be in a country where, for logistical reasons, another instrument is more popular. If using the same instrumentation is not possible, reduce analytical transfer risks by qualifying the full working range of instrument use. Testing all the instruments in the same rigorous way would establish a common baseline, allowing ready identification of any differences in performance between the instruments. These differences can then be taken into consideration when transferring methods between instruments.

Click to Enlarge

Figure 3: Options for approval of a contract laboratory.

Service Provider Selection

Each company uses a different approach when selecting contract service providers. The number of variables involved can result in a complex decision and, at the evaluation stage, companies often use a balanced score card or weighted Kepner Tregoe decision analysis model.² This balanced approach ensures a fair and unbiased decision and prevents jumping to a company for reasons that are not significantly important in the wider business and compliance context. The disadvantage of this approach is that a company with limited experience contracting out a particular service may base the weighting factors (what aspects of the decision are more important) on theoretical considerations. This can potentially result in a poor structure and, therefore, a poor decision. However, narrowing the selection to the top few and then considering the potential risks helps balance this out.
Because the requirements for testing QC pharmaceutical compounds are quite different from environmental samples or screening type analysis, the use of different selection processes, perhaps simply applying different weighting factors in the Kepner Tregoe model, would be ideal. In addition to the selection against the laboratory requirements, there is also the wider business requirement for approval of the contract laboratories or supplier. Some options that can be considered are summarized in Figure 3 (see p. 18). The greater the risk, the farther down the triangle the decision is made.
Complex pharmaceutical QC testing requires a physical audit with full technology transfer and ongoing analytical result monitoring. But for contract environmental analysis or any other widely available analytical testing, completion of a supplier questionnaire for a laboratory with the required ISO accreditation might suffice. Preferably, the laboratory will produce analytical data such as chromatograms or infrared along with the certificate of analysis, even though some charge more to supply this data. Receiving and reviewing the analytical raw data support reduces risks of compliance troubles. Finally, when evaluating the data, ask yourself if it looks real—like the kind of data you would normally observe in your laboratory. Such thinking can be critical to identifying potential problems and risks.
Smith is the validation program manager for Europe in Analytical Sciences and Laboratory Services at PerkinElmer. Reach him at paul.smith@perkinelmer.com.

References

1. Valigra L. Qualifying Analytical Instruments: General chapter <1058> clarifies terminology, classifies instruments. Pharmaceutical Formulation & Quality website. June/July 2010. Available at: http://www.pharmaquality.com/ME2/Audiences/dirmod.asp?sid=325598564E8C4B3EB736C7159241312D&nm=Browse+Articles&type=Publishing&mod=Publications%3A%3AArticle&mid=D3E3C719D8D44216836DCA4F4144BEC4&tier=4&id=C4030354EB4D4737BB47516C6D815CEB&AudID=5648A5C28C97462DBBDB309539B820EF. Accessed September 24, 2010.
2. Kepner CH, Tregoe BB. The Rational Manager: A Systematic Approach to Problem Solving and Decision-Making. New York: McGraw-Hill; 1965.