The use of advanced monitoring technologies in four key aspects of EtO sterilization can yield substantial business and regulatory benefits.

Paul J. Sordellini, Frank R. Bonanni, and Gregory A. Fontana

The use of ethylene oxide (EtO) gas has long been a dominant mode of terminal sterilization. Today, close to half of all medical devices produced in the United States are processed with EtO. But while it remains a popular method of sterilization, the pressures of global competition demand greater cost-effectiveness and flexibility of the process. At the same time, compliance with regulatory requirements must be maintained. Fortunately, the application of spectroscopic and electronic analytical equipment during process development, validation, and operation is allowing EtO users to increase the efficiency and efficacy of the process while reducing turnaround times, labor requirements, and raw material use. To gain these improvements, EtO users are focusing on four key areas:

The DataTrace system from Mesa Labs measures process temperature, pressure, and humidity.

• Releasing product parametrically.
• Managing product load configuration within validated parameters.
• Adopting new products into already-validated product families.
• Determining process equivalency across multiple sterilizers.

Correctly implemented, programs addressing these areas will increase productivity, reduce costs, and yield a more controlled process. At the same time, manufacturers will have increased flexibility to adjust their sterilization practices quickly in response to changing market conditions.

PROCESS MONITORING TECHNOLOGY

When a chemical process is developed and validated, the resulting data are used to establish and verify the acceptable range for each of the critical process parameters. Process repeatability is therefore predicated on the routine control of each parameter within the validated ranges. The simplest and most accurate way of verifying process conformance is to directly monitor each critical parameter and then compare the data collected during and after processing to the validated specifications.

Direct analytical technology using nondispersive mid-infrared (NDIR) spectroscopy allows the real-time simultaneous monitoring of headspace water vapor and EtO concentrations during an EtO sterilization process. An analyzer with the ability to measure 0­300 mg/L of water vapor and 0­1200 mg/L of EtO, and to operate at 0­70°C whether above or below atmospheric pressure, can be mounted on an EtO sterilizer to confirm process conformance with minimum validated parameters. Proper design and installation of a direct EtO sterilizer gas analyzer requires selection of an appropriate sample point from which analyte gases are drawn.1 Provided that a sterilizer has ample internal recirculation to minimize stratification, a single sample point is sufficient to obtain a gas stream representative of the process conditions. The current standard makes no mention of a minimum number of either sample points or a sampling frequency. As a result, the user can be asked to justify the selected locations and sampling plan.

The EOS-200 from Spectros Instruments Inc. (Whitins-ville, MA) uses a nondispersive infrared analyzer.

The optimum position from which to sample is toward the top of the sterilizer. Stratification, if it occurred, would result in the heavier gas settling toward the bottom of the sterilizer. Therefore the first signs of stratification caused by recirculation obstruction or failure would appear as lower-than-expected EtO gas concentrations beginning in the top of the headspace. By attaching a length of flexible tubing to the sample port, the port can effectively be moved around the sterilizer during a series of identical cycles. In this way different areas of the sterilizer may be analyzed for comparison and the most challenging location identified and documented for regulatory review. A report can then be generated showing details of gas distribution throughout the chamber.

External sample lines are sealed and heated to prevent leakage and condensation. In proximity to the sample point, the analyzer may have independent pressure and temperature sensors. Besides providing data for the gas analysis, these extra sensors can be used as a redundant process monitoring system. The results can be compared with the pressure and temperature data recorded by the sterilizer's main control system.

The user who reduces the definition of parametric release to simply replacement of biological indicators (BIs) with a gas analyzer is throwing away the opportunity to achieve an unprecedented level of process control, process optimization, and final assurance of sterility. With appropriate modifications to the sterilizer hardware and control software, the scope of process-gas analysis can be expanded from simply monitoring headspace gases to actually controlling the addition and maintenance of process-gas concentrations.

A data feedback loop, where concentration data from the analyzer are ported directly to the command logic of the sterilizer controller, allows the user to deliver accurate levels of water vapor and EtO to the process. Using this approach, processes can be developed and maintained to within ±4 mg/L. Direct gas analysis eliminates the concern of EtO depletion during sterilant dwell. EtO gas makeup is automated and controlled to add compensatory sterilant whenever the concentration drops due to load absorption. It must be noted that while this approach works with pure EtO systems, EtO-diluent systems (EtO/CO₂, for example) safety features to avoid over pressurization during makeup.

All measurement and test equipment that may directly or indirectly affect the quality of process output must be routinely calibrated.^2,3 Analytical hardware for water vapor and EtO measurement must be calibrated to bracket the full gas concentration range specified in the sterilizer operational qualification manual. Calibration efforts can be aided by infrared (IR) detectors that automatically (e.g., at the start of each cycle) switch to a stable and constant blank reference gas in situ for background correction. Based on the user's selection, the analyzer automatically isolates itself from the sterilizer by closing the sample intake valve and then proceeds to flush the gas cell with nitrogen. Nitrogen, like all other diatomic molecules, has no IR absorption, making it the most secure method of rezeroing a gas analyzer. The nitrogen, making up 100% of the volume of the gas cell, is optically scanned and provides the instrument with a true zero reference for both water vapor and EtO detectors to read.

This type of self-referencing analyzer uses the spectrum from the blank signal as a zero in calculating the absorbency units, thus allowing for the automatic correction of drift in the system. Any technology that is unable to self-reference a blank reference gas will require a more-stringent calibration schedule and procedures to constantly correct drift. Drift is inherent to all IR, chromatographic, and electronic sensors.

During validation of all minimum and maximum parameters, the headspace EtO level is measured at intervals (e.g., every 1 to 5 minutes) throughout the entire phase of gas contact. This generates an EtO concentration profile. What is characterized, especially in those processes that use a single-charge method, is the EtO affinity of that particular challenge load. Typically, sterilant dwell will begin with EtO at its highest concentration. As the gas permeates all levels of packaging in the load, the concentration profile decays, reaching its minimum at the end of sterilant dwell. The EtO concentration decay is related to the process set-point program and the totality of the physical attributes of the load. Therefore, gas profiling during validation will yield an acceptable minimum-maximum range for EtO concentration.

If an EtO profile exhibits excessive decay during routine processing, post processing data review will detect the condition, quantify it, and possibly reject the cycle for falling outside validated parameters. A direct headspace-gas analyzer, when properly designed, can consistently measure both process gases with an accuracy of ±2% full scale or better. Even small variations in load absorbency can be detected in real time, allowing the user to quantify the impact that variations in load configuration of any magnitude have on the process. Sample charts and graphs showing the anatomy of a pure EtO process are presented in Figures 1­4.

For processes that add makeup EtO during sterilant dwell in an attempt to compensate for EtO that migrates into the load, direct gas analysis improves process control, safety, and product quality. Typically, in this approach, any decrease in headspace pressure triggers the addition of EtO. With continuous direct analysis, the user can distinguish between pressure loss due to temperature fluctuations, water-vapor loss through condensation on the load and sterilizer surfaces, and pressure loss due to true migration of EtO into the load. Interfacing the gas analyzer with the sterilizer control system allows the addition of EtO to take place only when the headspace EtO concentration truly decreases. Adding EtO to compensate for a drop in temperature or condensation of water vapor can increase EtO chemical residues on the product, which will require longer aeration times.

Figure 1. Example of an EtO set-point graph.

PACKAGE MONITORING TECHNOLOGY

The object of an EtO sterilizer is to heat, humidify, and expose all product surfaces to EtO gas for a specified time. This is achieved primarily by effecting pressure and temperature changes in the sterilizer headspace. Process elements introduced into the headspace act physically on the load to bring forth the conditions inside each primary package necessary to achieve a targeted sterility assurance level (SAL). In essence, the microenvironment inside every single primary package included in a load becomes an individual sterilization chamber. Direct characterization and documentation of the minimum required levels of heat, water vapor, and EtO concentration that must integrate inside the primary packaging in order to sterilize a device to a specific SAL are the keys to achieving advanced process control. Again, the interests of science, regulatory compliance, and business can all be well served through the implementation of technology, this time on a primary-packaging level.

Figure 2. Example of a graph showing percentage volume of air.

Early attempts at process monitoring practices were incomplete. Technology offered BIs that could be sealed inside the primary package to confirm lethality, and thermocouples that could be attached to the packaging to monitor the product temperature. Water vapor and EtO concentrations inside the packaging could be inferred to be sufficient only by negative BI growth, but were rarely measured in situ during the entire process.

In 1994 the United States adopted ANSI/AAMI/ISO 11135 as the new sterilization standard.⁴ For the first time, verification of product level humidity became a requirement during all validations.⁵ Because most users sterilized with pure EtO, temperature/humidity sensors were eventually developed to comply with the new practice. Because EtO inside the package was still practically impossible to measure, however, BIs continued to serve as the final process indicator.

The implementation of an additional package-level sensor, together with the previously mentioned water vapor sensor and BIs, now makes it possible to individually monitor each of the critical process parameters from within the load and create simultaneous profiles of headspace and load conditions. At the end of in-chamber conditioning dwell, the pressure inside the product packaging is at equilibrium with that of the sterilizer headspace. The next phase consists of the vaporization of EtO into the sterilizer.

Figure 3. Example of a volumetric profile.

As EtO gas is added to the sterilizer headspace, the sterilizer control system will immediately detect and monitor the headspace pressure rise. As the gas penetrates the load, the pressure sensors inside product packaging will detect and record the pressure rise. The time lag between the pressure rise in the sterilizer head-space and the resulting pressure rise inside the product packaging will depend on and characterize the physical load attributes.

By simultaneously monitoring the headspace pressure rise during the addition of EtO and the pressure rise inside select pallet locations, the user can determine the rate of penetration into specific products. The matrix of EtO pressure data will reveal the relationship between the headspace pressure rise and the rise occurring inside the selected combinations of products and packaging studied. During process development this information helps estimate the exposure time needed to inactivate qualified BIs placed in the most difficult product/packaging type. In addition, differences in delivered parameters due to pallet location within the sterilizer are also identified.

Figure 4. Example of a partial pressure profile.

PARAMETRIC RELEASE

Parametric release of EtO-sterilized product is defined as "declaring product as sterile, based on physical or chemical process data rather than on the basis of sample testing or biological indicator testing."⁴ The alternative--conventional product release--requires the routine placement of BIs throughout the load. The BIs are retrieved following processing and tested for up to 7 days before the load is released to market. The main difference between parametric and conventional release is the number of process parameters directly measured.

In conventional release, the user directly measures the time of each phase, the pressure throughout the process, and the headspace temperature. The remaining two critical parameters, headspace water vapor and EtO concentrations, can be quantified indirectly by thermodynamic calculation based on pressure rise and temperature.⁶ Acceptance of the two indirectly measured gas concentrations is supported by the negative growth of the exposed BIs. Thus the BI data serve as a process parameter integrator that confirms delivery of appropriate levels of heat, water vapor, and EtO concentrations.

Parametric release, in compliance with the current international sterilization standards, yields an immediate return on the investment by increasing the productivity of the manufacturer and the sterilization operation. Given a choice, good science would choose direct real-time monitoring of all critical process parameters. Business logic would agree: releasing product parametrically the moment it completes aeration saves both the time and the materials needed to routinely place, retrieve, and test BIs. In some cases this represents a reduction in turnaround time of as much as a week, bringing a significant decrease in inventory requirements for the manufacturer.

It is important to realize that from a scientific and regulatory point of view, the technology needed for parametric release also increases the quality of sterilization process monitoring and the degree of process control. The BI data, as used in conventional release methods, will detect only gross process failures. They will not reveal small drifts in the performance of the processing system or reductions of delivered lethality due to interfering factors, such as variations in load configuration. The achievement of an SAL cannot be empirically supported with BI test results. The way to scientifically confirm that each and every process delivers the required SAL is to directly measure all parameters that influence the SAL and then compare the data with those collected during validation. Routine processes, proven empirically to meet or exceed every minimum requirement set forth by the validation, will deliver the required SAL for release to market. The jump in the quality of EtO sterilization process control is what sets direct process analysis apart from biological process compliance monitors.

LOAD CONFIGURATION CONTROL

For device manufacturers with an extensive product catalog or customized multicomponent products, the master product family may include any combination of thousands of different items. During routine production, an almost infinite number of different sterilizer load configurations are possible. That is, loads may differ from one another in density, product material characteristics, packaging material type, and quantity.

All potential side effects originating from the attributes of every different load configuration must be identified and evaluated during process development in order for a maximum challenge load to be assembled for validation purposes. The logic is that once a process is validated for the most challenging load configuration, every different load configuration generated during routine production will be equally or less challenging and therefore will respect the conditions established by the validation.

Variations in load configuration can occur daily and may influence the efficacy of the sterilization process or cause a particular product lot to absorb a level of heat, humidity, or EtO that falls outside the validated ranges. Certain packaging materials are more difficult to permeate than others. Certain product materials are more difficult to heat or humidify. Certain loads have a higher density and require a greater sterilizer heating capacity. A welcome by-product of parametric release is that the user has the tools to scientifically verify that every different load configuration placed inside the sterilizer does not exceed the physical demands of the challenge presented during validation.

Once a direct gas analyzer is installed and programmed to routinely profile sterilant dwell, routine data monitoring and review simply involves comparing the EtO concentration profile for each routine process with the validated profile to verify that the minimum and maximum validated concentrations were within tolerance. Each time a different load configuration is presented for processing, the material constitution is compared with that represented in the validation. In cases where a load configuration appears to differ significantly from the validation load, product temperature, humidity, and pressure sensors are placed inside primary packages and then distributed throughout the load, occupying positions that were monitored during the validation.

Following processing of this new load configuration, data detailing headspace temperature, water vapor, and EtO concentrations are compared with the resulting packaging levels of those concentrations. This data set, when compared with the headspace and product data obtained during validation, will allow the user to determine empirically whether the validated process continues to deliver the same minimum required process parameters to the product. If so, then the new load configuration is recorded in an amendment to the validation report and further monitoring is no longer necessary.

As more new load configurations are generated, monitored, and equated to the validated load configuration, significant historical data are created. Eventually, enough varying load configurations will have been tested and documented so that no further packaging-level monitoring of water vapor and EtO will be needed. Routine parametric release can then continue with the requirement of product temperature verification as prescribed in the requirement section of ANSI/AAMI/ISO 11135-1994.

With parametric release and the accompanying ability to monitor and evaluate load configuration influences, the sterilization operation upgrades the quality of process control. Manufacturing operations can vary load configuration according to market demands and regulatory bodies can be presented with a scientifically sound system for justifying the freedom exercised in building constantly changing product loads.

PRODUCT ADOPTION

Expansion of the medical device industry may bring frequent modifications to existing products and the development of new product lines. The most efficient way to deal with this growth is to allow new products (candidate products) to be quickly added to an existing sterilization product family (cycle group) already covered by a validated EtO process.

Following a program of product adoption, a candidate product is added to an existing cycle group after a thorough assessment of its physical and chemical characteristics and comparison with the other members of the cycle group. This is a documented study performed by a person with appropriate sterilization experience, and may include different degrees of physical, chemical, and microbiological testing to assess the product's suitability for the adoption. Provided that the candidate product is no more challenging to the penetration of heat, water vapor, and EtO than the original validation challenge, and that the product bioburden is no more difficult to sterilize than the indicator organism, product adoption into a cycle group is acceptable and becomes an important tool used to cope with business growth.

By distributing physical-microbiological test packs (PMTPs) throughout a validation load, a user can characterize the complete set of physical parameters delivered to the product site and correlate these conditions to the delivered lethality. A PMTP consists of a primary package containing a challenge product, a BI placed in the product location that is most difficult to sterilize, and a humidity/temperature and a pressure/temperature data logger both placed adjacent to the BI location. The PMTP is assembled and packaged under routine manufacturing conditions. Depending on the size of the load, a number of packs are seeded into the pallets so as to monitor an efficient selection of locations. The data loggers are programmed to monitor conditions at predetermined time intervals and store the data in nonvolatile memory.

Following completion of the process, the test packs are removed for analysis. The BI incubation results will map the delivered lethality throughout each pallet and the sterilizer. The humidity/temperature data will reveal the amount of water vapor and heat associated with the lethality achieved at each location. The user will see the amount of heat and humidity added as a result of preconditioning, the amount of heat and humidity lost during in-sterilizer air removal, the amount of heat and humidity added during the conditioning phase, and the heat added during sterilant dwell. The uniformity of the heat and water vapor distribution will also be charted. In this way, distribution variations can be attributed either to the product or packaging (in heterogeneous loads) or the physical pallet location within the sterilizer.

Monitoring pressure within the primary package to characterize gas penetration rates represents a novel and important aspect of this exercise. The user first identifies the start and stop times for EtO addition and sterilant dwell as recorded by the sterilizer. From this, a headspace EtO concentration profile can be derived, either by thermodynamic calculation using the EtO pressure rise and temperature, or through direct gas analysis in the case of parametric release.⁷ The pressure profile for the same two time periods is then extracted from each package-level pressure sensor. Package-level gas penetration can be revealed.

The headspace pressure and concentration profiles and the package pressure profiles can be overlapped, enabling the user to characterize the migration of gas from the moment it enters the headspace to the moment it penetrates across each pallet and enters the primary packaging. This type of headspace-and-product comparative study will reveal and quantify all the physical conditions inside the primary packaging that achieve the resulting level of sterility. For the first time, the user will be able to characterize with complete sets of data the degree of physical resistance that a particular product-package configuration and process-challenge device (or test pack) offers to the three physical elements of the sterilization process (heat, water vapor, and EtO).

Useful information derived from this study is headspace-product hysteresis, that is, the time lag exhibited by the load and process-challenge devices in reacting to physical changes made to the sterilizer atmosphere (addition and removal of heat and both process gases).

To aid in adopting a new product, package, or load configuration into a validated cycle group, the user needs to include the new candidate product in a routine process. An appropriate number of PMTPs, using the candidate product, are distributed throughout the load. Following processing, the PMTP data can be compared with the same data collected during the original validation of the cycle group.

If the validated process is shown to successfully deliver to the candidate product the same levels of heat and water vapor, and gas penetration rates are similar to those in the validation, this lends support to the adoption. If the candidate product turns out to pose more of a physical challenge than validated products, the user needs to construct a microbiological validation to complete the adoption. Finally, the candidate product is tested to confirm product functionality, package integrity, and chemical residuals. It is then added to the cycle group.

PROCESS EQUIVALENCY

Increases in production volumes can exceed the capacity of a validated preconditioning room, sterilizer, and aeration room. Manufacturers need to efficiently expand the sterilization of a cycle group to additional rooms and sterilizers that can be proven equivalent to the validated site in their ability to deliver minimum validated parameters to the primary package environment. This is achieved through a scientific program of process reproducibility that is more commonly referred to as process equivalency.

By implementing the sensor technology described in this article, EtO process development will yield a complete set of physical parameters that are confirmed to achieve a goal at the primary packaging level. Validation then confirms the acceptability of the results and the repeatability of the process. The actual parameters programmed into the sterilizer (heat, water vapor, and EtO additions) as well as the physical characteristics of the equipment (recirculation, heat medium, capacity) have no importance independent of what resulting conditions are delivered inside the primary packaging.

Once a cycle group is validated in a particular sterilizer (the predicate sterilizer) using the two-tier process and package-level monitoring described herein, additional sterilizers (candidate sterilizers) can be certified equivalent. They must be able to deliver to the primary packaging the same minimum process parameters as detailed in the validation of the predicate sterilizer. To verify this capability, the same load configuration as employed in the original cycle group validation is used in the candidate sterilizer. PMTP packs are assembled and distributed throughout the load according to the validation pattern. Following processing, the BIs will confirm delivered lethality and the physical sensors will characterize the heat, water vapor, and gas penetration to the primary packaging.

Any candidate preconditioning room proven to deliver the same levels of heat and water vapor to the product site can be considered for equivalency. Each candidate sterilizer proven to deliver the same minimum levels of heat, water vapor, and EtO to the product site can also be considered for equivalency. In addition, each aeration room proven to deliver the minimum level of heat to the product site can be considered for equivalency. Whether or not candidate equipment is physically comparable and programmed identically to the validated predicate equipment is not always important. The final SAL of a processed load is based solely on the successful delivery of a specific set of physical conditions to all product surfaces.

CONCLUSION

Ethylene oxide has been used as a medical device sterilant for the better part of a century. Although advances have been made in computer-controlled automation and worker safety, achieving the increased process flexibility and improved process economics necessary to meet changing business demands has often conflicted with the requirements of both scientific and regulatory authorities. New technology that details process parameters as they are delivered to the sterilizer headspace and the primary packaging can assure the industry that EtO will continue to be the sterilant of choice.

Once a cycle group is validated using direct analysis of headspace gas and ample distribution of PMTPs, the resulting data can be used to implement scientifically sound programs of parametric release, load configuration control, product adoption, and process equivalency. The completeness of the data collected in the validation and the scientific nature of each optimization program enable users to respond with confidence to all regulatory bodies, as every conclusion is supported by integral sets of empirical data. The result is a harmonious relationship between sterilization, manufacturing, and the regulatory authorities. Management of EtO sterilization services can now respond more quickly than before to changes in market demands.

REFERENCES

1. Paul J Sordellini, "Speeding EtO-Sterilized Products to Market with Parametric Release," Medical Device & Diagnostic Industry 19, no. 2 (1997): 67­80.
2. Cheryl A Boyce, "Guidance on the New QS Regulation Calibration Requirements," The Validation Consultant 4, no. 7 (1997): 12­14.
3. Code of Federal Regulations, 21 CFR Part 820, "Quality System Regulations."
4. Medical Devices—Validation and Routine Control of Ethylene Oxide Sterilization, AAMI/ANSI/ISO 11135 (Arlington, VA: AAMI, 1994).
5. AAMI/ANSI/ISO 11135, (Arlington, VA: AAMI, 1994), sects. 5.3.4, 5.5.2.1, and 5.5.2.2.
6. Ethylene Oxide Sterilization Equipment, Process Considerations, and Pertinent Calculations, AAMI TIR No. 15-1997 (Arlington, VA: AAMI, 1998).
7. AAMI TIR No. 15-1997(Arlington, VA: AAMI, 1998), sect. 6.0.

Photo courtesy of Mesa Laboratories