Derek Duncan

Container closure integrity (CCI) plays an important role in maintaining the sterility and stability of sterile injectable products. The defects that cause a sterile vial to leak are not necessarily detectable by a visual inspection process. Examples of such defects are those that are hidden by the crimp, microscopic cracks and scratches in the glass, or temporary defects such as stopper pop-up that result in temporary container leakage. 

	New regulatory guidance has triggered changes in industry best practices in the area of CCI testing (CCIT). This article summarizes the current state of container closure integrity testing in the pharmaceutical and biopharmaceutical industries and outlines possible approaches for developing a CCIT strategy.

	Historically, good CCI has been linked to the maintenance of sterility. A container that loses, or does not have, good closure integrity is at risk for microbial contamination. However, the context of CCI has become broader over the years. 

	An increasing number of formulations have significant sensitivity to oxygen and need to be packaged under an inert atmosphere. Freeze-dried product requires protection against water vapor and is often packaged at a partial vacuum to help with reconstitution and/or seating of the stopper. In these cases, good CCI is necessary not only for the maintenance of sterility but also to maintain critical headspace gas conditions. 

	Note that, quite generally, a container that is gas-tight will also be tight against microbial ingress. Therefore, the requirement to maintain headspace gas conditions imposes higher standards on CCI than the requirement to maintain sterility.

	In light of the importance of CCI for product sterility and stability, regulatory guidance has placed an increasing emphasis on CCI concepts. The current 

) <1207> chapter titled Package Integrity Evaluation-Sterile Products was implemented in late 2016 and represents the most thorough guidance document to date on CCI concepts for sterile injectable product (1). 

	The chapter gives an overview on CCI testing technologies and approaches for CCI control over the product lifecycle. Traditional CCIT methods, such as microbial challenge tests or blue dye ingress tests, are described as methods associated with probabilistic outcomes having some uncertainty in the results which, in turn, makes such methods difficult to quantitatively validate for the detection of critical leaks (1). The chapter also makes clear that CCIT should be performed throughout the product lifecycle. Deterministic CCIT methods based on non-destructive analytical measurements can be used to generate science-based CCI data that, coupled with a risk-based approach, enable informed decisions about a CCIT strategy in commercial manufacturing.

	A draft revision of the European Union’s Annex 1 requirements for sterile product manufacturing was released at the end of 2017 (2). CCIT was a popular discussion topic for the revision, and the draft text contains new requirements for CCIT in manufacturing. Other world regulatory bodies, Russia and South Korea for example, have also been putting increasing emphasis on CCI control for finished sterile products. It is clear from these developments that regulators are wanting to see improved industry practices in the area of CCIT.

<1207> provides an overview of CCIT technologies and categorizes them as being deterministic or probabilistic (see 

). The chapter emphasizes that this overview of CCIT technologies is not exhaustive but is a summary of technologies that have been implemented for CCIT in the pharmaceutical industry and that are described by a body of peer-reviewed literature.

	It is important to distinguish between CCI technologies and CCI test methods. Once a leak testing technology has been chosen as the basis for a test method, the chapter emphasizes the need to perform method development studies generating data that demonstrate detection of a critical leak for a specific product container configuration using defined test method parameters (1): “After a methodology has been selected for use, the test equipment operation and performance is qualified. Test method parameters are optimized during method development and confirmed during validation. Thus, a final leak test method is specific to a particular container-closure or product-package system.”  

	Another point emphasized in the chapter is that “no one test is appropriate for all packages or for all leak testing applications.” The chapter and its three sub-sections describe a framework in which appropriate CCI test methodologies are chosen, optimized per product configuration, and a robust validation of the method for detecting a critical leak is performed. In selecting a methodology, “deterministic leak test methods are preferred over probabilistic methods when other key method selection criteria permit.” 

	Package integrity data are generated over the product lifecycle and serves as input for an ongoing database of CCI data (the package integrity profile), which then serves as a risk management tool to ensure that CCI of finished product meets the product quality requirements. The framework described in the chapter is currently driving changes in industry best practices for CCI testing, including:

		Implementation of a ‘toolbox’ of CCI test methods optimized and chosen on a per product configuration basis rather than the application of a single legacy test method in a one-size-fits-all approach

		Generation of science-based CCI data in robust method validation studies, which demonstrate the detection of a critical leak represented by various types of positive controls.

Statistical sampling and generating science-based CCI data

	A big topic of current discussion is how much CCIT is required, especially for commercial batches of finished sterile product. Despite the general consensus that CCI is a critical quality parameter for finished sterile product, the industry has historically expended much more effort on testing for particle contamination than for CCI. 

	Visual inspection to detect particulate contamination has been a requirement for many years with 100% inspection of finished parenteral product being done manually or by automated inspection platforms. In the context of risk to the patient, a loss of CCI would, in general, be assessed as being just as critical as particle contamination.

	The current EU Annex 1 guidelines require 100% leak testing for certain types of product containers. “Containers closed by fusion, e.g., glass or plastic ampoules, should be subject to 100% integrity testing” (3). This requirement is a result of the fact that the inherent failure rate of the sealing process for these types of containers cannot be sufficiently controlled. 

	The ongoing draft revision of the EU Annex 1 guidelines again states the requirement of 100% integrity testing for fused containers and adds the following requirements for all other types of containers. “Samples of other containers should be checked for integrity utilizing validated methods and in accordance with QRM, the frequency of testing should be based on the knowledge and experience of the container and closure systems being used. A statistically valid sampling plan should be utilized. It should be noted that visual inspection alone is not considered as an acceptable integrity test method” (2). If finalized in this form, these CCIT requirements will require the evolution of best practices for CCIT in the manufacturing environment. 

	Currently, a small percentage of the industry performs statistical CCIT of finished commercial product. Most companies point to the 100% visual inspection process to justify meeting current CCIT guidance, such as the following from the FDA (4). “A container closure system that permits penetration of microorganisms is unsuitable for a sterile product. Any damaged or defective units should be detected, and removed, during inspection of the final sealed product.” The language of the draft EU Annex 1 revision makes clear that visual inspection is not considered an acceptable integrity test method; in other words, the CCI test methods that enable the testing of larger amounts of samples will need to be implemented.

	To demonstrate statistical confidence in the process requires the generation of statistical CCI data. However, an argument could be made that a better place to do this in the product lifecycle is in process development and scale-up rather than in manufacturing. The guidance provided in 

<1207> to collect package integrity data throughout the product lifecycle so that a package integrity profile database is built up implies an approach in which a significant amount of CCI data are generated outside of the manufacturing environment. The generation of robust CCI data providing knowledge of the container and closure system (which then gives guidance to a CCIT strategy in manufacturing) is also implied in the text of the draft revised EU Annex 1, “the frequency of testing should be based on the knowledge and experience of the container and closure systems being used.” 

outlines a possible approach to generating CCI data that enables the design of an appropriate CCI testing program in manufacturing.

	After validation of the fundamental closure system, data need to be generated to understand if the process introduces risk to CCI. To gain statistical confidence in the process, it would be necessary to perform testing on statistical sample sets. This in turn will require the use of non-destructive deterministic test methods because the probabilistic legacy test methods (blue dye and microbial ingress testing) have limited throughput capability. Testing could be done on either a pilot scale or with test and engineering batches from the manufacturing environment. Once a baseline failure rate has been established, process controls could be implemented to improve the process, if necessary. 

	Product from the improved process would be tested to quantify the residual risk to CCI after which a decision could be made for an appropriate testing strategy in manufacturing. Packages and processes having a high inherent failure rate that is difficult to control would require a heavier inspection process and vice versa. In this way, the decision for an inspection process design is driven by science-based statistically relevant data.

	The current environment for CCIT of sterile injectable product is evolving. New regulatory guidance recognizes CCI as a quality parameter that is critical for the maintenance of both the sterility and the stability of finished sterile product. New concepts introduced in the regulatory guidance are changing industry best practices and include the following:

		Generate science-based CCI data throughout the product lifecycle to build up a package integrity profile database that can be used as input for risk management. 

		When possible, use deterministic CCI test methods that have been validated to detect a critical leak.

		There is no one-size-fits-all CCI test; a toolbox of CCIT technologies that can be optimized on a per-product package configuration is necessary for a robust CCIT program.

	Because industry best practices will be evolving as the impact of new guidance becomes clearer, a certain amount of uncertainty in CCIT best practices is to be expected in the near term. However, a general approach that includes the implementation of validated deterministic CCIT methods and the increased generation of science-based CCI data to enable informed risk assessments will help prepare the industry for the future.

40 <1207> “Sterile Product Packaging-Integrity Evaluation” (US Pharmacopeial Convention, Rockville, MD, 2017).
	2. European Commission, 

EU GMP Annex 1 Revision: Manufacture of Sterile Medicinal Products (Draft) 

(Brussels, December 20, 2017). 
	3. European Commission, 

EudraLex, Volume 4, EU Guidelines to Good Manufacturing Practice-Medicinal Products for Human and Veterinary Use

, Annex 1, Manufacture of Sterile Medicinal Products (Brussels, 2009).
	4. FDA, 

Guidance for Industry: Sterile Drug Products Produced by Aseptic Processing-Current Good Manufacturing Practice 


	Vol. 43, No. 9
	September 2019
	Pages: 38–41

	When referring to this article, please cite it as D. Duncan, “Container Closure Integrity Testing of Finished Sterile Injectable Product,” 

Articles

Changes in CCIT testing have followed the evolution of regulatory guidance, and this article examines different possible CCIT strategy approaches.

Updates on Container Closure Integrity Testing for Sterile Injectables

As regulatory guidance has evolved, changes in CCIT testing have also become apparent. In this article, possible CCIT strategy approaches are outlined.


Container Closure Integrity Testing of Finished Sterile Injectable Product

Lyophilization, or freeze-drying, is a process used to stabilize a pharmaceutical formulation and increase the shelflife by removing water from the drug product. During lyophilization, the drug formulation is first frozen and then the ice is removed by sublimation under vacuum during a primary drying phase. A secondary drying phase is then used to remove unfrozen water molecules at a temperature higher than that used for primary drying. Pharmaceutical freeze drying cycles are designed to remove most of the loosely bound water and to achieve a pharmaceutically elegant cake. For biological materials, it is important to retain a high level of activity in the final product.


	Determining the residual moisture content of a lyophilized pharmaceutical product is important for several reasons. First, the amount of residual moisture content is related to the stability of the formulation over the shelflife of the product. Small-molecule formulations can have direct degradation pathways triggered by water, and it is crucial that all final product is below a defined residual moisture specification. In general, the degradation pathways for large-molecule formulations are more complex, with water often playing an indirect role. Second, moisture analysis of a statistically relevant sample set can give insight into the freeze-drying process itself. Residual moisture determination can be used as a tool in process studies to confirm the efficiency, consistency, and robustness of a specific freeze-drying cycle that has been designed for a particular drug formulation.

	Typical pharmaceutical freeze-drying cycles usually target residual moisture contents in the range of 1% to 3% water by weight. Historically, a strategy that can be described as “the drier, the better” was often followed. For small molecules having a direct degradation pathway triggered by water, this approach was an appropriate strategy. However, in the world of large biopharmaceutical molecules, it is possible to over-dry. Studies have shown that even in the lyophilized state, proteins depend on small quantities of water to help maintain higher-order structure. Other types of products, such as certain lyophilized blood plasma formulations, need a minimum amount of water to achieve efficient dry-heat viral inactivation. It is therefore sometimes necessary to design a freeze-drying cycle that keeps all product vials within a certain moisture range, having both minimum and maximum specifications.

	Historical cycles have often been too conservative (i.e., too long), meaning that the final product was over-dried, because research and development efforts did not take the time to optimize both the formulation and the freeze-drying cycle. Although conservative cycles produce product that meets quality parameters (i.e., sufficiently low residual moisture), the same product quality could be produced with much shorter cycles if appropriate studies are done. Current scientific approaches use various tools to monitor the lyophilization process and analyze the finished product with the goal of defining optimum freeze-drying cycles on a per formulation basis. Data are generated to demonstrate that product quality parameters (such as stability, cake appearance, appropriate reconstitution, and bioactivity) are met in a targeted moisture range and that the defined cycle is robust and consistent in producing product in that moisture range.

Freeze-drying cycle, formulation, and equipment parameters


	The residual moisture content depends on both the cycle and the formulation. Focusing on the process side, it is important to control the endpoints of the three lyophilization stages--freezing, primary drying, and secondary drying. Key process parameters are the temperature and pressure gradients chosen for each stage. It is out of the scope of this article to fully treat all the factors that can affect the residual moisture content.

	As an example, there are several approaches that can be taken to optimize freezing conditions: annealing, supercooling, and controlled nucleation. The details of crystal formation during freezing have an effect on how efficient the sublimation process is in removing moisture during primary drying. In general, care should be taken in the freezing step to ensure that the product is fully frozen, that the ice crystal structure is open allowing for better sublimation, and that complete freezing is achieved at as high a temperature as possible to save time and energy.

	The formulation parameters also have an effect on the final residual moisture content. The exact ice crystal formation and eventual cake structure will depend on the type of pharmaceutical ingredients and buffers in the formulation as well as potentially varying product concentrations. All of these formulation parameters can influence the final residual moisture content achieved by a freeze-drying cycle.

	Finally, a third general factor influencing residual moisture content is the equipment and vial configuration used. Vacuum pump, condenser parameters, and the vial size should all be considered when implementing a freeze dryer for particular product batch sizes.

	In the end, the final residual moisture content depends on the interplay between the cycle, the formulation, and the equipment parameters. This interplay has motivated the development of approaches for efficient cycle development in recent years. Optimization of the drying cycle for a given formulation requires a balanced understanding of the fundamental science of freeze-drying, formulation characteristics, and equipment capabilities. The ability to make accurate measurements of residual moisture in statistical sample sets of finished product would be a useful tool in these efforts.

Traditional methods to determine residual moisture


	The traditional methods used in the pharmaceutical industry for residual moisture determination have been Karl Fischer (KF) titration and thermogravimetric analysis (TGA). The KF method is most widely used and is generally considered to measure the total water within the freeze-dried vial, assuming that the sample is wholly soluble in the KF medium. It should be noted here that residual water content may be present in a variety of forms--free, adsorbed, chemically bound, hydration shells, water of crystallization--not all of which may be directly linked to the activity or stability of the product in question. The KF method is time-consuming, requires operator expertise and careful sample handling to avoid contamination by environmental moisture, and is destructive (meaning that the sample is destroyed during analysis).

	Unlike Karl Fischer, which relies on a chemical reaction to detect water, TGA measures the weight loss as the sample is heated, driving off residual moisture. This method measures not only the water but also any other volatiles that are produced as a result of heating. Therefore, the composition of the lyophilized material must be well understood for this method to be used to accurately measure the residual water content. This method also suffers from potential artifacts due to environmental moisture and is also a destructive method.

	The destructive nature of the traditional methods is a limitation that has negative consequences for stability and process studies of freeze-dried product. The moisture content of a product vial in stability studies is usually inferred from that determined for sister vials. The degradation of the API cannot be directly correlated to the measured residual moisture because the moisture determination measurement destroys the sample. The moisture content of the vial in the stability study may not be identical to the reference sister vials due to vial-to-vial variability, and therefore, the stability results may be imprecise for defining moisture stability specifications.

	For process studies, the traditional methods are not conducive to conducting studies with statistically relevant sample sets. The methods are resource intensive and sample analysis throughput is limited. Moreover, potentially valuable product is destroyed because of the destructive nature of the methods. The absence of statistically relevant data for residual moisture determination means that residual moisture stability studies are at risk of being imprecise and that limited insight can be gained into the consequence of process variability on final product quality. As a rapid non-destructive analytical method, headspace moisture analysis can potentially address the limitations of traditional methods.

Headspace moisture is measured by shining laser light through the vial headspace, tuning the laser to an absorption wavelength of the water molecule, and then analyzing the absorption signal to determine the headspace water vapor pressure inside a sealed vial. The measurement is rapid and non-destructive. Measurements can be made in less than a few seconds and the sample remains intact.

	Headspace water vapor levels measured inside of a freeze-dried vial have been directly correlated to KF measurements of those same vials (1). Headspace water vapor measurements have also been directly correlated to the degradation of the API (2), which means that the stability of a freeze-dried product could be defined in terms of the water vapor pressure inside the sealed vial. This correlation is not surprising as the equilibrated partial water vapor pressure is actually the water activity of the freeze-dried material inside the sealed vial. Water activity is a thermodynamic property and is an indication of how much free water is available for interaction with, and potential degradation of, the API. The rapid non-destructive nature of the headspace moisture technique enables a number of applications that are difficult, or even impossible, to achieve with the traditional residual moisture determination methods.

Applications of headspace moisture analysis


	The applications of headspace moisture analysis take advantage of the fact that it is straightforward to analyze the headspace moisture content of all vials in a statistically relevant sample set. In fact, it is even possible to perform 100% headspace moisture analysis of a commercial batch of freeze-dried pharmaceutical product using automated headspace moisture inspection machines.

 shows the measured headspace moisture levels of freeze-dried product produced on a shelf in a pilot freeze dryer (3). The numbers in the graphic are the measured water vapor levels in each vial in units of torr with the darker colors representing higher moisture values. By tracking the location of the vials on the lyo chamber shelf, a moisture map could be created that clearly shows a wet spot in the middle of the shelf.

	By performing 100% headspace moisture analysis of pilot batches in development, full insight can be gained into final product quality variability, which in turn enables efficient cycle development and optimization.

 shows the measured headspace moisture levels of an initial cycle and a subsequent optimized cycle. The initial cycle showed significant vial-to-vial variability in the headspace moisture levels as well as drying that seemed to be tray dependent. The insights provided by this analysis motivated changes to the cycle, and the optimized cycle was able to produce a product batch with little vial-to-vial variability in the headspace moisture levels and no drying dependency on tray position. The headspace moisture analysis also identified a few outlier vials that did not dry well even in the optimized cycle. The utility of the headspace moisture method is demonstrated by the fact that the total analysis time of each pilot batch was a few hours and the samples remained intact. Using traditional methods would have required a total analysis time of days and perhaps even weeks, and all the product would have been destroyed. These examples show how headspace moisture analysis can be used to characterize freeze dryer performance, optimize freeze-drying cycles, and identify lyophilized product vials that did not dry well and are potentially out of specification.

 65 (1) 2-11 (2011).
	2. M. Peeters, “Results FMS vs KF Analysis,” Lighthouse Instruments European User Symposium, June 2006.
	3. I. Cook and K. Ward, 

When referring to this article, please cite it as D. Duncan, “Headspace Moisture Analysis for Determination of Residual Moisture Content in Lyophilized Pharmaceutical Products," 

Headspace moisture analysis is a rapid non-destructive analytical method that may potentially address the limitations of traditional methods used for residual moisture determination.