Power of Parametric: Methods to Validate Ethylene Oxide Sterilization Parametric Release

Abstract

When approaching an ethylene oxide (EO) sterilization validation, medical device manufacturers traditionally have two choices. They can use biological indicators (BIs) to monitor each production run or establish a parametric release process in which sterile release is based on the monitoring and control of physical process parameters that ensure process specifications are met. In ISO 11135:2014, parametric release was brought to the forefront as an acceptable release method; however, a perception exists that implementing parametric release is challenging and time consuming. This article will demonstrate that the opposite is true. It presents a streamlined approach in which parametric release is addressed through the various stages of validation: product definition, process definition, performance qualification, routine release, and process control. Considerations for establishing specifications directly from validation versus “run and record” and trending critical process parameters (e.g., relative humidity, temperature, EO concentration) are discussed. In addition, the benefits of parametric release (active monitoring) over BI release (passive monitoring), including improvements to turnaround time, process control, risk mitigation, reduction of resource investment, and elimination of microbiological release testing, are highlighted. With multiple benefits, parametric release should be the gold standard for EO sterilization processes. It is not novel and has been widely accepted by regulatory agencies globally and notified bodies. The article further describes how the data collection and process capability that is central to process control and parametric release is more powerful than the information provided by a BI, which is merely a catastrophic indicator when used in routine processing.

Parametric release is a practice that was first recognized in 1985¹ to release finished drug product terminally sterilized by moist heat, based on measurement and documentation of critical physical process parameters. For ethylene oxide (EO) sterilization, parametric release became more prominent with the release of ISO 11135:2014.² Subsequently, ISO/TS 21387:2020 provided additional guidance on its application for EO sterilization.³

As described in ISO 11135, the validation of an EO sterilization process consists of performance qualification (PQ)—both microbiological PQ (MPQ) and physical PQ (PPQ)—and is performed with the equipment used to routinely sterilize the product. Parametric release can be incorporated as part of an original validation or can be implemented for an existing process.

This article first addresses the benefits of parametric release, then discusses in detail the stages of validation for parametric release, conversion of an existing cycle from biological indicator (BI) release to parametric release, and implementation of the parametric release process, including measurement of chamber temperature, relative humidity (RH), and EO gas concentration (subsequently referred to as parametric release critical parameters).

By providing a streamlined approach that addresses parametric release throughout the various stages of validation (product definition, process definition, and PQ), this article demonstrates that the process of validating and implementing parametric release is not more challenging or time consuming than validating for BI release.

Benefits of Parametric Release

Compared with BI release, parametric release involves many benefits, including the following:

1. Increased resolution to detect potential risks of a nonconforming cycle, through the measurement and documentation of additional physical process parameters.

2. Quicker turnaround times for product release. Traditional BI incubation requires multiple days, whereas a review of parametric data can be automated or completed immediately.

3. Reduction of resource investment compared with BI release, which includes BI purchasing and storage, placement, and retrieval; shipping; testing; and analysis. These steps are not needed for parametric release processes.

4. Eliminates the risk of false-positive BI results.

5. Process trending of routine runs allows earlier identification of trends that could lead to process failures.

Validation Approach for Parametric Release

Figure 1 outlines example approaches based on the different ways in which parametric release can be applied.

Figure 1.

Pathways for establishing parametric release. Abbreviations used: PQ, performance qualification; SAL, sterility assurance level.

To establish parametric release, a prerequisite is that the sterilization facility and chamber must be capable of measuring the critical process parameters and that the data are being collected. Parametric release can be established and implemented directly from the PQ with supporting process data from the equipment qualification, or it can be established during the PQ and implemented after running routine loads with BIs for a defined number of runs. Although the latter has been commonly used, it is not a requirement of ISO 11135.

This article presents an approach that does not require running routine loads with BIs, but instead implements a parametric release process directly from PQ, with supporting process data from the equipment qualification. Regardless of the approach, the ability to monitor and verify equipment performance across the operational range for each critical process parameter is essential in establishing and maintaining a parametric release process. Two key objectives need to be planned for in establishing release criteria:

1. Sterility assurance level (SAL)/functionality: Specifications will need to consider if all values within the specified range deliver the desired lethality and maintain product/packaging functionality.

2. Process capability: Specifications will need to consider the inherent process variability of a given parameter.

Product Definition

As with any sterilization process validation, attention to product definition is necessary. In accordance with ISO 11135, product loads should be configured “to allow removal of air and penetration of heat, humidity and EO during the sterilization process, and removal of EO at the end of the process so that the sterilization process applicable for the defined product shall be established.”² Consideration should be given to the load variance and its impact on critical process parameter measurement due to the load absorption characteristics. The following attributes and their impact on the parametric release specifications should be considered during product definition: product and packaging materials, product weight and density, pallet load volume, and maximum and minimum load configurations. Further considerations are provided in Table 1.

Table 1.

Interrelationships of additional parametric release parameters and their effect on the process. Abbreviations used: EO, ethylene oxide; PQ, performance qualification.

Process Definition and MPQ

Performing process definition and MPQ runs at one or more minimum tolerances for specified process parameters, along with the abbreviated EO dwell time, is one approach to validate the minimum requirements needed for a selected SAL. Due to the absorption characteristics of EO and humidity into the load, using injection differentials for validation that are less than the minimum tolerances of such parameters may be necessary. This will allow for a lower minimum tolerance for parametric release parameters that achieves both the selected SAL and a robust process that minimizes the potential for nonconformances. One method to establish the minimum tolerance for RH is decreasing the steam injection pressure to or slightly lower than the minimum tolerance. This same approach can be used when evaluating the minimum tolerance for EO gas concentration (i.e., decreasing the EO injection pressure to or slightly lower than the minimum tolerance).

In addition, the cycle calculation method can be useful because it uses growth from internal process challenge devices to establish the lethality rate and minimum time to reach the selected SAL. It has the potential of providing the shortest exposure dwell time to achieve a selected SAL, while also reducing the overall cycle time compared with the half-cycle approach. When assessing reproducibility and consistency to satisfy the objective of establishing a process that captures the inherent variability of the critical process parameters capability, critical process parameter data from all validation runs should be reviewed. Additional considerations are detailed in Table 1.

PPQ

Performing PPQ cycles at one or more of the maximum tolerances for the specified process parameters is one approach for establishing the maximum specifications for EO gas concentration and RH and may also be used to validate whether the product meets required EO residual limits of ISO 10993-7:2008/Amd 1:2019.⁴ Adjusting critical process parameters using an approach similar to that described in the MPQ, but to the maximum tolerances or just above those maximum tolerances, should be considered. Sterilant residual analysis following exposure in a higher EO gas concentration cycle will demonstrate that in routine production if the maximum tolerance is reached, the product can still successfully off-gas to the residual patient safety limits of ISO 10993-7. In addition, product and packaging functionality can also be evaluated following exposure in the higher EO gas concentration cycles to verify whether the design requirements are met and if the device materials are compatible.⁵ Additional considerations are provided in Table 1.

Use of Sterilization Equipment Data to Support Parametric Release

To gain confidence in the specifications established from the sterilization validation and understand the variability of the critical process parameters without performing additional cycles, reviewing the sterilizer equipment trending data is useful. Equipment variability is critical to any sterilization process, and these data should be reviewed to understand appropriate tolerances for the critical process parameters. This is especially important for capturing sensor drift, which may not be observed in a few PQ runs over a short period of time. A generic cycle could be used to understand sensor process capability for a given dunnage material range. The cycles can be used to demonstrate the chamber achieves critical process tolerances consistently and in a reproducible manner.

Transitioning from BI to Parametric Release

If an existing process was running with BI release and critical process parameter data were being recorded (including measurement of chamber temperature, RH, and EO gas concentration), an opportunity exists for leveraging historical data to establish parametric release tolerances for each critical parameter. This approach uses the actual load configurations variability and cycle data (e.g., parameter set points, parameter variability, tolerances) for establishing the parametric release requirements (as outlined in ISO/TS 21387).³

The parametric release temperature specification is established from that which already is in place for the chamber temperature specification of the existing BI release process. Statistical analysis can be performed on historical RH and EO gas concentration data from BI release runs, such that a set number of standard deviations (SDs) from the average can be used to establish the parametric release specifications. Performing a confirmatory MPQ cycle at or below the minimum tolerances may be necessary to ensure that the process can still deliver the selected SAL. Conversely, a confirmatory PPQ cycle followed by sterilant residual analysis may be needed to ensure that at the maximum EO gas concentration specification to be established, the product can still off-gas to the required patient safety limits.

The data from existing BI release cycles can also demonstrate the consistency and reproducibility of the equipment running a specific cycle. It can be used to assess process capability and, in conjunction with product definition, help determine the extent of testing needed to complete the product family process characterization (i.e., less frequent product loading configurations not equally represented in historical data) and be incorporated into implementation of process specification tolerances ³,⁶

Although routine cycles of the actual product provide great insight into process and parameter variability and allow for the establishment of parametric release specifications, only MPQ cycles should be used to assess whether such processes are providing the selected SAL.

Implementation

When implementing any validated EO process the two requirements for successful release per ISO 11135, section 11.1, are (1) confirmation that the data recorded during routine processing meet the sterilization process specification and (2) confirmation of no growth of the test organism from any BI (if used).²

With a parametric release process, the first requirement noted above involves additional parameters that need to be monitored and the second requirement is not applicable. Ultimately, meeting process specifications and therefore ensuring process control is always required, but BIs are not. Although the “run-and-record” approach of monitoring routine cycles with BIs prior to transitioning to full parametric release may be viewed as a more conservative approach, in reality, the full cycle BI result does not provide adequate information to support the achievement of a typical 10^–6 SAL when monitoring a validated, full cycle. If a BI with a population of 10⁶ was positive following full cycle processing, this would indicate a substantial failure in the cycle development or MPQ validation or that the actual cycle did not meet specification.

A full-cycle BI cannot be used to demonstrate the achievement of a 10^–6 SAL; instead, the cycle development, equipment capability, and PQ should validate that the process can reproducibly deliver the selected SAL and parametric release criteria can be established. Therefore, the BI is not needed to transition a cycle from BI release to parametric release.

Run and record sometimes is used to further assess the process variability as part of the capability assessment, but it is not required by ISO 11135. When referring back to the objectives, the usefulness of each data set is further highlighted: (1, SAL/functionality: can be assessed via process definition and PQ cycles; 2, process capability: can be assessed via all cycles performed).

Maintaining Process Effectiveness

Process capability of the established parametric release parameters can be assessed by real-time process trending of routine runs.

In this scenario, process trending is defined as monitoring trends and variation in critical process parameters. Although evaluating parametric data from each run for adherence to the established specifications is the requirement for release; trending the critical process parameters to assess continued process capability of the specifications used for parametric release is important. The benefits to trending routine processing runs for maintenance of parametric release are that the analysis:

1. Identifies unfavorable trends that may indicate the process is no longer in a state of control. An investigation that identifies a root cause would allow for mitigation. This mitigation would resolve an issue that could have escalated and caused the sterilization cycle to exceed tolerances, resulting in a nonconformance.

2. May aid in identifying when sterilization equipment is starting to fail or if variation in the sterilization load mix is affecting the sterilization process. Efficient sterilization processes include filling chambers, which may include more mixing of different product types in sterilization loads and more potential for process variability.

3. Would be an input to the annual review required by ISO 11135. It could be used in the decision process for the extent of requalification to be performed. Process trending is a recommendation of ISO/TS 21387, section 12.1.1: “It can be beneficial to also trend critical parameters, including those for humidity and EO concentration, to confirm the process is in a state of control. This will also provide valuable data in meeting the requirement for carrying out an annual review (see ISO 11135 Section 12.3.1) of the sterilization process.”³

4. Could be used to demonstrate continued process equivalence for sterilization chambers that have been validated for process equivalence, thus providing evidence of equipment performance for future cycles and products.

5. Will allow for a determination if greater process capability is needed with additional validation runs to expand critical parameters tolerances.

Parameters that Can Be Trended

Parameters to be considered for trending would be those that are critical, such as:

1. Vacuum depth/rate. (Duration to achieve vacuum set point can also be used to monitor rate.)

2. Pressure rise/rate due to steam injection.

3. Chamber humidity at the end of the conditioning phase.

4. Pressure rise/rate due to EO gas injection.

5. Chamber EO concentration during exposure.

6. Chamber temperature during exposure.

7. EO exposure time (consider also injection and removal).

Analysis

Statistical process control could be applied to identify negative trends indicating the process may no longer be in a state of control.⁷, ⁸ The use of control charting would identify shifts, drifts, and special causes. Alert and action levels could be established (e.g., mean ± 2 SDs and mean ± 3 SDs, respectively), along with criteria for identifying negative trends (e.g., x number of increasing/decreasing data points) that would trigger an investigation. The alert/action levels would fall within the specified tolerances to identify potential issues before they cause the sterilization process to exceed tolerances, resulting in a nonconformance.

Conclusion

This article described the advantages of parametric release and provided strategies for validating new cycles, characterizing critical process parameters needed to establish parametric release, converting existing cycles from BI to parametric release, and implementing a parametric release process without first performing the run-and-record approach. It is evident that data collection and process trending of critical parameters is far more valuable than the limited information provided by a routine-processing BI. For these reasons, parametric release is now considered best practice for EO sterilization processes.