Abstract
In ethylene oxide (EO) sterilization processing, products can be released based on the growth responses of biological indicators or by using a parametric release (PR) process that relies on the monitoring and control of process parameters. Both methods must be used in combination with process data, in accordance with ISO 11135:2014. Process parameters can be classified as control, monitor, or both. Control parameters can be altered directly to change the readings of monitoring variables, which can't be directly controlled themselves. Currently, ISO 11135:2014 does not allow PR based on control parameters for EO concentration and humidity (calculated via pressure increment). Sterilant concentration and relative humidity (RH) can only be measured using gas analyzer probes (or similar), despite the challenges of data variance and range, calibration, redundancy, and increased sensitivity to deviations involved with use of such probes. The current article sought to experimentally determine the capacity of statistical process control to detect changes in the process, to act as an early-warning system for an out-of-specification result, and to demonstrate the use of more reliable process data for the purpose of PR. Process data from 100 routine cycles were used to trend the achieved levels of chamber RH, temperature, and EO concentration (measured by gas analyzer probes or similar), and process derived data were compared with data provided by gas analyzer probes. Process trending of routine runs was found to predict process failures, and calculating EO/water concentration via pressure increment was determined to be a viable alternative to measurement by gas analyzer probe. Further, reduced variability in key parameters enables a reduction in sterilant use.