Speeding EtO-Sterilized Products to Market with Parametric Release

The difficulty of measuring EtO and water vapor during processing has been a major obstacle to adoption of parametric release by EtO sterilizers, but the latest gas analysis techniques are making in-process monitoring and control possible.

Article Contents

Introduction

According to industry experts, ethylene oxide (EtO) is used for more than half of all medical devices sterilized. Despite its popularity, EtO sterilization, as it is performed by most companies today, has one major disadvantage when compared with steam or radiation methods. Unlike these methods, EtO typically involves more than just in-process monitoring to determine that products are adequately sterilized.

Most EtO sterilizers still rely on incubating biological indicators (BIs) to confirm sterility after processing, but retrieving the BI samples and waiting for them to incubate can add several days to product distribution. This final testing also requires additional handling of the sterilized products and introduces the potential for human error.

In theory, by simply making sure that the critical parameters of the process are correct, a technique called parametric release, EtO sterilization providers can be sure that their procedures are successful without using BIs. If the time, temperature, mass of water vapor per unit volume, and concentration of sterilant gas are within the process specifications, then it should be unnecessary to test the products or BIs after the procedure.

There are three basic steps necessary for implementing parametric release:

• Conducting a complete performance qualification through physical and microbiological studies.¹
  
• Extracting from the resulting data a set of physical process parameters including load configuration, which, when maintained, will consistently yield the desired sterility assurance level (SAL).
  
• Maintaining, monitoring, and recording these critical parameters through direct analysis during routine processing.

Physical and administrative controls must be sufficient to ensure that the load configuration will never deviate from validated parameters, that all procedures will be followed, and that the sterilization equipment will consistently perform within the specifications listed in its original commissioning. If the controls meet these qualifications, then real-time direct monitoring of all process parameters can replace the use of BIs for routine product release.

Even though FDA accepts parametric release and the technique obviously saves time and money, only a few EtO sterilizers have yet ventured to use it. One reason for this is a lack of regulatory guidance on parametric release implementation.

Before 1994, the principal EtO sterilization guidance document was the American National Standards Institute's Guideline for Industrial EtO Sterilization of Medical Devices (ANSI/AAMI ST-27). Although this document said that parametric release was an acceptable practice, it provided no guidance on two of the most important issues facing those who would adopt it. These issues are, how does one validate and control a parametric release process, and what are the gas analysis technologies available to monitor in-chamber process gases?

With no official answers to these two crucial questions, only a very few companies have attempted parametric release. Those that have done so have typically submitted the validation protocols with their products' 510(k)s to get FDA to approve this revolutionary release method. These firms have been able to afford the time and expense, but pursuing regulatory acceptance in this way has generally remained out of reach for smaller manufacturers and contract sterilizers.

Recently, however, regulatory agencies have begun to take steps to provide more guidance. Since 1994, both the United States and Europe have adopted International Organization for Standardization (ISO)-compliant EtO sterilization standards that contain details on validating parametric release.^2,3 In addition, in 1995 the Association for the Advancement of Medical Instrumentation (AAMI) formed an EtO task group to write a technical information report that will include a section on parametric release.⁴

Another reason that many have hesitated to use parametric release is that the technique requires sophisticated gas analysis equipment to monitor cycle parameters. Currently, however, microprocessors can ensure that each EtO phase is accurately timed, beginning only when certain variables come within required specifications. And although EtO vessels, whether operating at positive or negative pressure, require a seal as close to perfect as possible, direct temperature monitoring of both product and headspace can be easily achieved using temperature sensor leads that are passed directly through penetration points in the walls of the vessel.

Efficient and accurate monitoring of the headspace gases--water vapor and EtO--is more difficult and has been a major obstacle to widespread adoption of parametric release. However, technological innovations in this area are making even this aspect of process control practical.

Headspace Gases

Water Vapor. When process parameter integration is confirmed by the use of BIs, regulatory bodies allow water vapor to be calculated in terms of relative humidity (%Rh) rather than directly measured. Knowing the pressure at the beginning of humidification (Pi), the pressure at the end of humidification (Pt), and the chamber temperature at the end of steam injection (°Tdc), one can calculate the %Rh by referring to the pressure of saturated steam (Pss) at °Tdc in standard steam tables.⁵

But %Rh is not a completely reliable measure of water vapor. Pressure rise during the steam injection phase does not necessarily mean that pure steam has been introduced. In fact, a ruptured seal on the inlet side of the steam control valve can cause ambient air to be sucked into the vessel in place of steam. This would provide a normal pressure rise profile, but no rise in humidity. If BIs are being used afterward, they will reveal that such an error has occurred. But to ensure that this parameter is correct so that parametric release can be possible, the mass of water vapor per unit volume must be measured directly.

Sterilant Gas. EtO concentration must also be measured directly rather than calculated from other values in parametric release.

Using BIs allows the sterilization control system to depend on indirect measurement methods. The most common of these methods are pressure rise inside the vessel from sterilant addition and the change in weight of the gas supply cylinder. In these methods, minimum pressure rise and cylinder weight change are established as parameters during validation, and then verified after each production run.

However, equipment malfunctions may allow "false" pressure rises and cylinder weight changes to occur. Leaks along the sterilant gas feed lines, leaks in the gas vaporizers, and malfunctioning valves can all lead to a normal pressure rise inside the vessel and a normal decrease in cylinder weight without the introduction of the required amount of EtO to the vessel.

Load Configuration

Another obstacle to parametric release is the difficulty of accurately predicting load configuration. Validations are typically designed for load configurations that represent the most difficult products to sterilize. Rather than validate every single product manufactured (which, in the case of custom kits, for example, could require years and enormous expense), products are usually grouped according to similarities in physical construction, geometries, and bioburden test results.⁶ Those that are the most difficult to sterilize, that present the greatest physical and microbial challenges to the process, are used to build validation loads.

But there are some important variables that can remain unaccounted for in this procedure. For example, sterilization engineers often neglect to compare each component's ability to absorb water vapor and sterilant gas. If a load contains a large portion of cotton fiber material, such as contained in gowns, towels, and gauze, a sterility failure may be caused by water vapor and EtO depletion.⁷ The use of BIs would reveal such a failure, but in parametric release, water vapor and EtO must be monitored throughout the entire cycle, and the absorbency of the load must be taken into account.

More precisely metering the amount of sterilant needed by taking into account load configuration will also save expense and create safer products. By using only the required amount of EtO, sterilizers can reduce residuals on the product, aeration time, and patient and employee exposure. This optimized process will also reduce demand on the environmental control systems at the sterilization site by lowering primary and secondary emissions.⁸

Sampling the Gases

To measure the headspace gases for parametric release precisely, a system must first be configured to draw representative samples of the process gases.

Sampling points must be selected wherever it is determined that a gas sample representative of process conditions can be obtained and carried to the analyzer. Unless the analyzer is of the type that can be directly connected to the vessel (bolted in-line with the recirculation blower), or equipped with remote sensors (such as fiber optics) that can be placed inside a vessel, then sampling lines (such as 0.25-in. stainless-steel tubing) must be installed to connect the vessel to the analyzer.

These lines should be kept as short as possible and must maintain a high enough temperature that the process gases will not condense before reaching the analyzer. Maintaining this temperature, or heat tracing, actually means keeping the sample gases at a higher temperature than that of the process gases to prevent condensation.

Also, all connections must be perfectly soldered and periodically leak tested. Removable compression fittings offer an alternative to soldering, but using them may prevent the system from being approved by certain safety agencies for use in a 100% EtO sterilizer.

Depending on the technology used to perform the gas analysis, a pump may also be necessary to draw a sample from the vessel. This pump, along with all other components of the gas analysis system, must satisfy all applicable electrical safety standards for use with vessels containing flammable gas mixtures. If the analyzer is to be placed within the sterilizer area, approval for such a position must be obtained by the appropriate institutions (such as insurance companies, building inspectors, and Underwriters Laboratories).

The sampling system, especially if it is functioning continuously, must return unused gases to the vessel. Sample pumps with a capacity of more than 1 L/min can alter the process conditions if they do not return exhaust to the vessel. In addition, such a closed system provides extra protection against inadvertent violations of emission standards and employee exposure.

There should be a calibrated temperature sensor next to each sample point. The closer the sensor is to the point, the more accurate the reading, especially in larger vessels. Process temperature monitoring systems that simply measure multiple points in the vessel and then average the readings to control chamber heating should not be used to control a gas analyzer.

Provisions should also be made to allow management to empirically determine whether stratification of the process gases is possible. Depending on the type of gas analyzer chosen, this can be accomplished by extending the sample port well into the interior of the vessel and then moving it about during a series of cycles so that samples are analyzed from the top, bottom, front, and rear of the vessel. Stratification can also be verified through the use of BI incubation data during fractional studies, or through the implementation of quantitative chemical indicators placed strategically throughout a load.⁹ If stratification is verified, then the selection of the sample point should be based on the location that has the lowest levels of process gas.

Gas Analysis: Chromatography

After determining how to draw the gas samples, sterilization engineers must also decide on a gas analysis method. The two most common ways to analyze in-chamber gases are chromatography and spectroscopy.

Gas chromatography (GC) is already well known in the EtO industry for its applications in monitoring work areas and analyzing chemical residuals on products after processing. It can also be used to directly analyze headspace gases in EtO sterilization. In this application, a gas sample is injected into a heated port. This can be done either manually by collecting the sample with a syringe and injecting it through a rubber septum, or automatically with a computer-controlled vacuum pump that extracts the sample from the vessel and channels it through a heat-traced line leading to the analyzer.

After injection, the sample is vaporized with a heating element and then mixed with a carrier gas, which is usually ultrapure nitrogen or helium. This gaseous mixture then passes through a silica column. The column, which is inside a temperature-controlled oven, is packed with ground firebrick or diatomaceous earth that has been coated with a layer of a liquid that has a very high boiling point, such as a silicone or squalene.

When the gases enter the column, they adsorb onto the liquid layer. The higher the solubility of the analyte gases, the slower they will move through the column and exit into the GC detector. The gases desorb from the column and enter the detector separately, allowing identification and quantification.¹⁰

Each substance generates a concentration pulse as it enters the detector. This pulse is recorded as a chromatogram, a graph showing the detector response as a function of elution time. The number of peaks on a chromatogram reflects the number of components present in the sample. The position of the peaks, which reveal retention times when they are compared with known standards, allow identification of the sample gas components. Retention time is the time it takes for the peak to reach the detector after injection. The height and areas of the peaks show the quantity of each component. Several types of GC detectors are available: thermal conductivity detectors, flame ionization detectors, and element-specific detectors.

Unlike many other gas analysis methods, GC operates at atmospheric pressure, creating special requirements for the sampling system. Some sterilization cycles condition at very low pressures. The dynamic environmental conditioning cycle, for example, conditions at less than 1.0 psia. GC operates at a pressure of ~14.7 psia. Therefore, sampling for EtO during the sterilant dwell cycle, which usually occurs close to or above atmospheric pressure, is easier because the pressure difference between the GC and the sterilizer offers less resistance than in other cycles.

Automated sampling can be accomplished with specially designed pumps that can overcome the vacuum inside the vessel and bring the gas to the column, where it is then allowed to rise to atmospheric pressure before analysis. Naturally, the most important factors affecting this part of the system are the sample lines, which must be perfectly sealed and heat traced, and the pump, which must perform at a consistent level.

Calculating the Gas Concentrations. The initial pressure (P₁) and temperature (T₁) of the gas sample will change dramatically as it is removed from the vessel. And as the gas enters the GC injection port, these changes (P₂ and T₂) will need to be factored into the final concentration calculation.

Calculation of the analyte gas concentrations inside the GC column to determine the process concentrations inside the sterilizer is accomplished with a gas law that combines Boyle's law (on pressure and volume) and Charles's law (on volume and temperature).¹¹

Boyle's law:

Charles's law:

Combined gas law:

The combined gas law, together with Avogadro's law, which describes the relationship between volume (V) and number of moles of a gas (n),

can be used to determine density (d) or the mass per unit volume of a gas in SI units (kg/m³ or mg/L). Because gas density is inversely proportional to volume, the following can be derived:

or

where d₁ = gas density inside the sterilizer and d₂ = gas density inside the GC column.

The GC control system receives as inputs the temperature and pressure inside the sterilizer (P₁ and T₁) and the temperature and pressure inside the column (P₂ and T₂), and will detect the density of the analyte gas components injected into the port (d₂). This leaves only one unknown: the gas density inside the sterilizer (d₁). The proportional relationship between d₂ and the density of the gas inside the sterilizer (d₁) can be expressed in terms of the combined gas law:

This equation can be resolved to yield the density of the analyte gas inside the sterilizer (d₁) as:

Gas Analysis: IR Spectroscopy

Through spectroscopy, analyte gases are identified by their interaction with light. One branch of spectroscopy, infrared (IR) spectroscopy, deals with the ability of matter to absorb light from the midinfrared region of the light spectrum, or wavelengths between 2.5 and 25.0 µm.12

By identifying the wavelength at which a substance demonstrates absorbance maxima and comparing this wavelength value to known libraries of chemical compound spectral data, it is possible to identify the sample. After the absorbance characteristics of the sample have been identified, the Beer-Lambert law, which relates absorbance to concentration, can be used to determine sample concentration. The Beer-Lambert law is:

where A = absorbance (a unitless quantity), E = absorptivity (m^-1 cm^-1), l = length of the light path through the sample (cm), and c = concentration of the sample (mole/L).

Absorptivity (E) is the proportionality constant between the amount of light absorbed and the gas concentration. An IR instrument can be calibrated to yield this constant by measuring substances of known concentration. Knowing E and the path length (l), absorbance (A) can be measured for an unknown sample, and the Beer-Lambert equation can be used to compute concentration:

When IR radiation is sent through a sample it interacts with the molecules, causing the chemical bonds to vibrate as the molecules absorb the radiation. Each functional group of gases is characterized by the tendency to absorb IR radiation of a particular wavelength, regardless of the structure of the rest of the molecule. Should two different substances exhibit absorbance maxima at the same wavelength, they can still be distinguished from one another because they will differ in another aspect, namely, in their molar extinction coefficients, or absorbance of IR radiation per mole of each substance.¹³

The only substances that cannot be analyzed with IR spectroscopy are those that exist as single atoms with no chemical bonds, such as noble gases, and homonuclear diatomic molecules, because they have no dipole movement. These two types of materials will not absorb IR radiation.¹²

An IR spectrometer can be interfaced with a sterilizer to continuously monitor the water vapor per unit volume and EtO concentration during the process through direct sampling of sterilizer headspace. In the case of gas blends, such an analyzer can also monitor and record the concentration profiles of the components. Differentiating the gases is useful in determining whether there is selective gas absorption by the load. To be used with a sterilizer, the spectrometer is programmed with the spectral information of each analyte gas. This information can be considered the optical fingerprint of each gas (see Figure 1 below).

Figure 1. Infrared spectral analysis of EtO, water vapor, and CO2. (Data courtesy of Spectros Instruments, Inc.)

The gas sample is carried from the vessel to the detector site, where it flows through a tube only to return, unchanged, to the sterilizer. Gas can be conducted through the beam path, or the beam can be reflected through a window in the sterilizer so that it crosses the headspace gas and returns to the detector. The IR source is placed at one end of the tube and the IR detector at the other. The IR beam passes through the gas sample, and the detector collects the spectral reading and transmits the data to a microprocessor (see Figure 2).

The computed gas identification and concentration data can either be printed or used by the system to actually control the process. One advantage of this technology is that it permits simultaneous monitoring of more than one analyte gas. The relationships that develop between gases and between gases and product or packaging types, as well as the process itself, can be easily controlled.

IR Spectroscopy for Gas Diffusion

EtO sterilization has traditionally taken place in stainless-steel, vacuum-pressure vessels with extensive ancillary equipment. However, for more than 25 years there has been a less costly and more flexible alternative to these fixed-volume steel vessels. This method, known as gas diffusion, can also be confirmed with parametric release by using IR spectroscopy to measure the gases.

In gas diffusion, liquid EtO is vaporized directly into a low-density polyethylene (LDPE) bag that contains several finished devices, each in individual sterile-barrier packaging. Placed in a heated room or container, the EtO penetrates each sterile barrier, sterilizing the devices. With time, the EtO diffuses across the LDPE and out of the package. This sterilization method precisely delivers gases to within centimeters of each device.

Typically, BIs are used with gas diffusion in accordance with the ANSI/AAMI/ISO 11135-1994 sterilization standard. However, as mentioned previously, they make the process more expensive and time-consuming.

An IR beam can cross multiple layers of transparent LDPE without interference and without danger to the contents of the bag. Avoiding physical invasion of the sterile device packaging during gas analysis means that any number of packages can be held in the IR beam for confirmation of their water vapor and EtO contents without sacrificing a single sample. The use of either GC or BIs requires the sterile barrier to be compromised to obtain results.

Figure 2. Diagram showing basic components and gas flow of a nondispersive infrared gas analyzer. (Diagram courtesy of Spectros Instruments, Inc.)

To use IR spectroscopy on an LDPE sterilizer bag, a device must be constructed to hold the IR source and detector at a fixed distance so that packaged products can be placed in the beam path. With this system, accurate analysis of gas content can be achieved in seconds. In one experiment, for example, the IR spectra of EtO, water vapor, and three different types of LDPE bags were measured. The goal was to verify the ability to analyze the contents of an LDPE bag without piercing it. The results demonstrated that LDPE poses no interference to the beam, and that the bands for EtO and water vapor are clearly identifiable (see Figure 3).

For device manufacturers who require EtO sterilization on-site but cannot install the traditional vacuum vessels, or who seek a level of accuracy and process control that is not possible with industrial vessels, IR spectroscopy and the current ISO-compatible EtO sterilization standard allow parametric release following gas diffusion sterilization.

Figure 3. Infrared spectral analyses of water vapor, EtO, and three different types of LDPE sterilization bags. (LDPE sterilization bags courtesy of H. W. Andersen Products, Inc. Spectral analyses courtesy of Nicolet Analytical, Inc.)

With gas diffusion, instead of just over gassing the headspace of a large vessel and then verifying the success of the cycle through incubation of 20 or more BIs, a manufacturer can realistically verify the gas concentration at a statistically significant number of product sites. For the most demanding sterilization protocols, a manufacturer can analyze every product to confirm proper levels of sterilant and water vapor. Once the proper amount of EtO per product is determined through process validation, EtO residuals can also be controlled, protecting the environment, the worker, and, ultimately, the patient.

Gas Analysis: MW Spectroscopy

Microwave (MW) spectroscopy functions like IR analysis--through the absorption of electromagnetic radiation. The MW spectrum contains the wavelengths between 0.1 and 30 cm.

The object of the MW technique is to determine the MW frequencies that are selectively absorbed by certain materials.¹⁰

The data allow insight into the atomic, molecular, and crystalline structure of samples. At low pressures, there is a linear relationship between absorbance maxima and molecular concentration.

Like IR, MW radiation is not energetic enough to induce an electronic transition, but it can cause changes in the vibrational and rotational dipole motion of the molecule, which is the method used by this technology. The MW spectra of analyte gases at low pressures are characterized by very narrow bandwidth absorption lines.¹⁴

In the 1980s, with a grant from the Medical Devices Directorate of the Department of Health in the United Kingdom and from private industry, a group of researchers at the University of Wales College of Medicine (Cardiff) began investigating MW spectroscopy for direct analysis of EtO concentration and water vapor during sterilization. In the past decade, they have successfully designed, prototyped, demonstrated, and patented an MW molecular rotational spectrometer that not only measures sterilant and water vapor inside a vessel using 100% EtO or EtO/diluent mixtures, but, like the IR spectrometer, also measures multiple analyte gases at the same time during the entire cycle.^15,16

The researchers have documented and confirmed that the system provides accurate, reproducible, and unambiguous determination of EtO sterilization process parameters, making it suitable for parametric release. Interfaced with an automated sterilizer control system, it represents another viable solution to gain parametric release. The group has obtained a U.S. patent for their spectrometer.¹⁷

The work of the researchers was an extension of a method for MW gas analysis that had been proposed in 1967.¹⁸ In this original procedure, a gamma coefficient was monitored while a sample of analyte gas was exposed to an increasing incident power until the maximum signal (S_max) was determined. S_max was found to be directly proportional to the concentration (N) of the analyte gas (the number of molecules). In the new system, a constant flow of gas from the sterilization vessel is circulated through a continuous loop. First, the gas flows from the vessel to a gas cell (a cavity with a preset resonant frequency) and then, once the analysis is complete, it is returned to the vessel.¹⁸ The pressure of the sample inside the spectrometer cavity is linearly proportional to the headspace pressure of the sterilizer. Therefore, the concentrations of EtO and water vapor inside the spectrometer cavity are also proportional to the concentrations of EtO and water vapor inside the sterilizer.

The prototype unit was programmed to emit an MW source frequency range, which the researchers called a scanning window.¹⁸ This window included the full linewidth of the absorption lines of the analyte gases: EtO (scanning window, 4 MHz; absorption line, 23.123 GHz) and water vapor (scanning window, 1 MHz; absorption line, 22.235 GHz).¹⁴ Studies of the system revealed that the linear correlation between the output signal and the concentration of EtO in the spectrometer cavity yielded results of ±4% accuracy. Considering the high concentrations inside a vessel, which can range from 350 to 1200 mg/L, such a level of accuracy is more than acceptable for the control and monitoring of a parametric release sterilization system.¹⁸

Several qualities of this system make it suitable for monitoring and controlling a sterilization cycle. First, because the partial pressures of EtO and water vapor in the analyte sample and the partial pressures of these gases inside the vessel are proportional, and the conductance of the inlet and outlet valves of the gas cell are constant, this system can relate the concentration inside the analyzer to that of the vessel without having to monitor the pressure inside the vessel.¹⁵ Second, the research team reports that in their instrument, the frequency of the resonant cavity does not require an identical source frequency, as other types of spectrometers do. The ability to vary the source frequency means that the concentration of several gases in a mixture could be measured effectively.¹⁸

Conclusion

There are still many challenges to face in EtO sterilization. For example, it is difficult to determine the amount of sterilant to use, and so companies must estimate based on worst-case predictions. Because of variations in load configuration, as well as condensation of water vapor in low-temperature areas within the vessel, it is impossible to decide in advance exactly how much sterilant and water vapor will be needed in a given cycle. Therefore, EtO cycles typically employ more sterilant and water vapor than is actually necessary. In this way, fluctuations in the load's capacity to absorb gas are compensated for and the SAL is consistent with validated levels.

However, the use of excessive quantities of sterilant and water vapor will also yield higher than necessary residual levels of EtO and by-products, such as ethylene glycol and ethylene chlorohydrin (ECH). Patients and health-care workers are later exposed to these residuals, which can cause health problems. Also, water vapor condensing on the product and packaging traps EtO, which, during storage of the product, can once again become airborne, representing a danger to facility workers.

But many of the difficulties of EtO sterilization are already solvable using existing technology. Installation of a direct gas analysis system on an EtO sterilizer allows the user to map the kinetics of water vapor and EtO throughout the entire cycle, constantly access the leak rate of the chamber, quantify the gas- absorption characteristics of each load to reveal the relationship that exists between individual product types and the process, and achieve parametric release. The time-consuming step of BI incubation can be avoided. Releasing products parametrically will help most EtO sterilizers accelerate market response times.

The benefits of parametric release are sometimes limited by extended product aeration times to remove residuals. But once parametric release becomes common practice, reducing this necessary aeration phase will surely be the focus of industry attention and resources.

MW radiation has already been applied to accelerate desorption of EtO from medical supplies.^16,17,19 With respect to traditional desorption in a heated aeration area, a 400% increase in desorption efficiency for polyvinyl chloride when using MW radiation to energize the EtO molecules has been documented. The effect of injecting steam during postevacuation phases to reduce residuals of both EtO and ECH has also been studied. These two methods are commonly referred to as MW and steam distillation.²⁰

Today, EtO holds approximately 52% of the medical device sterilization market.²¹ EtO compatibility with nearly every device or component material, recent progress in the field of parametric release, and future developments in more efficient product conditioning and aeration techniques will soon make EtO sterilization even more widespread.

References

1. Medical Devices--Validation and Routine Control of Ethylene Oxide Sterilization, ANSI/AAMI/ISO 11135, Arlington, VA, Association for the Advancement of Medical Instrumentation (AAMI), 1994.

2. Medical Devices--Validation and Routine Control of Ethylene Oxide Sterilization, ANSI/AAMI/ISO 11135, Section 7.2, Parametric Release, Arlington, VA, AAMI, 1994.

3. Sterilization of Medical Devices--Validation and Routine Control of Ethylene Oxide Sterilization, EN 550:1994, Section B.5.2, Parametric Release, Brussels, European Committee for Standardization, 1994.

4. Sordellini PJ, "Parametric Release Comes to EtO Sterilization," Med Dev Diag Indust, 18(2):66-69, 1996.

5. AAMI Technical Information Report on Engineering Aspects of Industrial EtO Sterilization, 4th working draft, Arlington, VA, AAMI, June 1996.

6. AAMI Technical Information Report on Contract Sterilization for Ethylene Oxide, 3rd working draft, Arlington, VA, AAMI, March 1996.

7. Hoborn J, "Ethylene Oxide Sterilisation--A Proven Method," in Ethylene Oxide Sterilisation Conference Proceedings, London, EUCOMED, pp 33-50, 1989.

8. Code of Federal Regulations, 40 CFR 9 and 63, "National Emission Standards for Hazardous Air Pollutants for Ethylene Oxide Commercial Sterilization and Fumigation Operations."

9. Manning CR, "Validating EtO Packaging/Sterilizer Configurations," Med Dev Diag Indust, 11(1):130-136, 1989.

10. Harris DC, Quantitative Chemical Analysis, New York, W. H. Freeman, pp 551-605, 1982.

11. Segal BG, Chemistry, Experiment and Theory, 2nd ed, New York, John Wiley, 1989.

12. Smith BC, Fundamentals of Fourier Transform Infrared Spectroscopy, Boca Raton, FL, CRC Press, 1996.

13. Efiok BJS, Basic Calculations for Chemical and Biological Analyses, Gaithersburg, MD, AOAC International, 1995.

14. Zhu Z, Gibson C, Samuel AH, et al., "A Microwave Spectrometer with a Frequency Control System Employing a Frequency Scanning Window Locked to the Rotational Absorption Peak," Rev Sci Instrumentation, 66(10), 1995.

15. Zhu Z, Gibson C, Samuel AH, et al., "A Gas Monitoring System for Ethylene Oxide Sterilizers with Constant Sample Flow through a Microwave Cavity Spectrometer," J Med Eng Tech, 17(4):147-151, 1993.

16. Samuel AH, and Matthews IP, Sterilizing Procedures and Equipment, U.S. Patent No. 5,209,902, 1993.

17. Samuel AH, and Matthews IP, Sterilizing and Desorbing Equipment, U.S. Patent No. 5,399,314, 1995.

18. Zhu Z, Gibson C, Samuel AH, et al., "Microwave Cavity Spectrometer for Process Monitoring of Ethylene Oxide Sterilization," Rev Sci Instrumentation, 64(1), 1993.

19. Matthews IP, Gibson C, and Samuel AH, "Enhancement of the Kinetics of the Aeration of Ethylene Oxide Sterilized Polymers Using Microwave Radiation," J Biomed Mat Res, 23:143-156, 1989.

20. Shugar GJ, Shugar RA, Bauman L, et al., Chemical Technician's Ready Reference Handbook, New York, McGraw-Hill, 1981.

21. Announcement from the Ethylene Oxide Sterilization Association, Washington, DC, June 5, 1996.