OR WAIT null SECS
Keith Bader is senior director of technology at Hyde Engineering + Consulting, 6260 Lookout Rd, Suite 120 Boulder, Colorado 80301.
Kelly Jordan is an engineer, at Hyde Engineering + Consulting, 6260 Lookout Rd, Suite 120 Boulder, Colorado 80301.
Richard J. Forsyth is principal consultant with Forsyth Pharmaceutical Consulting, 907 Shamrock Ct, Royersford, PA 19468, tel. 484.535.1688, email@example.com.
The authors look at the cleanability of pharmaceutical soils from a variety of materials of construction to determine the relative ease of cleaning and explore potential grouping strategies as part of a comprehensive cleaning validation program.
A cleaning procedure is expected to remove soil from product-contact equipment surfaces. A validated cleaning procedure has been shown to reliably remove soils from these surfaces. One of the accepted approaches to cleaning validation is to identify a worst-case soil for validation. A worst-case soil is one that is the most difficult to clean in relation to all other soils manufactured in a pharmaceutical facility. If the worst-case soil can be cleaned to an acceptable level, it can be concluded that the other soils in the facility can also be cleaned to an acceptable level using the validated cleaning procedure.
Identification of a worst-case soil can be accomplished through equipment-washer and formulator experience. Those involved in manufacturing formulations, cleaning, and maintaining the equipment are in the best position to identify the hardest-to-clean soil. This approach is a practical but subjective determination and would leave a facility open to question until ongoing data supports the initial conclusions on the worst-case soil.
A second approach is a comparison of API solubility. Solubility data for APIs are typically generated in water and organic solvents as part of the physical and chemical characterization workup for the API. Those APIs that are least soluble in the cleaning solvent are considered the hardest to clean. This approach neglects the formulation excipients, which are often insoluble, comprise a much greater percentage of the formulation, and can be more difficult to clean than the API.
A more empirical method to determine a worst-case soil is to spot coupons with the soils, allow them to dry, and clean them using the identical conditions encountered in the cleaning cycle. Equipment surfaces, however, encounter different types of cleaning action depending on their location in the manufacturing equipment. Soils could be subjected to manual scrubbing, impingement under the cleaning solvent, turbulent flow from a pump forcing cleaning solvent through piping, or a cascade action as the cleaning solvent moves across equipment surfaces.
Cleanability studies are typically conducted using dried residue spotted on a coupon dipped into a beaker containing water or the cleaning solvent or solution. The cleaning solvent is stirred, creating a less rigorous cleaning action than encountered during actual cleaning, and would be considered a worse-case condition. The experiment can be conducted at room temperature or an elevated temperature. The soil with the longest cleanability time can be considered the worst-case soil for cleaning validation within the variability of the test parameters. Numerous studies have been conducted to demonstrate cleanability. Studies range from a paper-based evaluation of a facility’s soils (1) to basic laboratory conditions, such as suspending the coupons in a beaker of cleaning solvent, all the way up to a sophisticated cleanability bath (2). Cleaning parameter variations have been characterized using cleaning process design on bench-scale studies (3), but for this study the cleaning parameters were held constant. Statistical equivalence testing for assessing cleanability (4) can show that soils are cleaned to an equivalent extent. If potential worst-case soils demonstrate comparable cleanability, it would be prudent to use more than one worst-case soil for further testing.
A well-executed cleanability study is one part of a comprehensive cleaning validation program. The experimental cleanability can be preceded by an evaluation of the formulation components to narrow the number of soils tested. The selected soils and materials of construction can be tested using a matrix approach to determine a worst-case soil and a hardest-to-clean material of construction.
Complementary studies can include—rinse recoveries and swab recoveries for analytical testing, and visible residue limits (5) for inspection of cleaned equipment. The studies can be conducted in parallel, often using the same coupon samples for multiple studies. All studies might not be necessary, based on the number of formulation soils and the manufacturing equipment involved, but a comprehensive picture of the physical properties of the soils during cleaning, and the ability to test and detect residual soils after cleaning, form a sound basis for a validated cleaning program.
The cleanability study was conducted to determine ease of cleaning for a variety of soils on a range of material-of-construction coupons using the following parameters. Table I lists 10 material-of-construction coupon types, which are among the commonly used materials of construction in pharmaceutical and biopharmaceutical manufacturing. All 10 materials of construction were tested during the study.
The cleanability study to determine the worst-case soils was conducted using three buffers and three media under the following parameters. The coupons were weighed and soils were spotted in triplicate for each of the material-of-construction coupon types listed in Table I. Each individual coupon was spotted with 1 ml of the soil. The soil was allowed to dry for at least 4 hours, or until visually dry, but no longer than 3 days, which was the established dirty hold time. Following drying, the coupons were reweighed and the coupon weight was subtracted to determine the weight of the residue. The coupons were immersed in a 600-ml beaker containing 400 ml of room-temperature purified water. The 600-ml beaker was the smallest beaker that would hold the 2.5” x 2.5” coupons without impeding flow around the coupons during testing and 400 ml was the minimal volume necessary to completely cover the coupons. The water was agitated on a magnetic stirrer at a fixed rotation and did not generate a vortex during testing. The cleanability endpoint was defined as the time at which the soil was no longer visible to the observer under defined conditions: distance 18 inches, optimal viewing angle dependent on the material-of-construction coupon type, and 700 lux light intensity. The coupons were removed from the beaker immediately after the visual endpoint was determined.
The visual endpoint was confirmed analytically through conductivity, total organic carbon (TOC), and gravimetric testing. The cleaning solution in the beaker was tested for conductivity. A sample of the cleaning solution was taken and tested for TOC. Positive controls were tested for conductivity and TOC by pipetting 1 ml of soil directly into a beaker and testing for conductivity and TOC. After drying, the coupons were again weighed and the weight was compared to the initial weight to determine cleanability of the soils gravimetrically.
Results and discussion
The material-of-construction coupons chosen for the study were seen as representative of a wide variety of materials. A number of liquid soils were chosen to represent a sampling of pharmaceutical and biopharmaceutical soils. Implementation of this approach at a site should address all soils on all materials of construction employed at the site unless an abbreviated approach can be justified.
Instruments were checked with every use to ensure they were operating properly and within calibration dating. The soils and the cleanability results along with supporting TOC and conductivity data are presented in the following sections.
Two buffers (regeneration buffer and zinc wash buffer) were spotted and held for 3 days to match the proposed dirty hold time maximum, but these buffers did not dry due to high ethylene glycol content. It should be noted that a properly executed paper-based evaluation (1) would have eliminated these two soils from the laboratory testing and saved 2 days of waiting for each of the trials. The assays for these two buffers on 316-finish stainless steel were representative for all of the materials of construction.
The visual removal data generated for the soils as shown in Table II demonstrated complete removal of the soils from all materials of construction. The data in Tables II-V are expressed as an average of the three individual determinations. Times vary widely, with the two buffers that did not dry rinsing from all coupons within 10 seconds. The 316-finish stainless steel and nickel-steel alloy (Hastelloy) coupons were the only materials of construction that demonstrated retention of the other four soils. They provided a representative ranking of the cleanability for all the soils and comprised the vast majority of material-of-construction surface area for all automated cleaning circuits. Ethylene propylene diene monomer (EPDM), silicone, polytetrafluoroethylene (Teflon), and synthetic fluoropolymer rubber (Viton) all became visually clean quickly for all soils, making any type of soil ranking difficult on those materials. Of the remaining four material-of-contruction coupon types: acrylic, glass, polypropylene, and polyether ether ketone (PEEK), all retained some soils but not others. These inconsistent results were coupled with the fact the soils retained by these materials were also retained strongly by 316-finish stainless steel and nickel-steel alloy (Hastelloy). Visual endpoint testing is subjective, and therefore, should be done in multiples by a small, well-trained personnel group to maintain consistency for the observations.
The conductivity results shown in Table III demonstrated high percentage of removal from all of the materials of contruction for each of the soils except for the regeneration buffer with 20% ethanol, which evaporated during the coupon drying. Removals greater than 100% can be attributed to day-to-day variance in calibration of the conductivity meter. Conductivity, however, could be misleading. The conductivity readings began to plateau well before the coupons became visually clean. The conductivity levels of the fully dissolved soils were relatively low (<60 µS/cm). Use of conductivity measurements should confirm the conductivity of the soils and should graph the rate of soil removal during cleanability testing and compared to visual endpoint measurements.
The gravimetric results in Table IV show complete removal of the residues. Other than 316-finish stainless steel, glass, and nickel-steel alloy (Hastelloy), the material-of-construction coupon types demonstrated some results where the final weight was less than the initial weight, indicating that the coupons continued drying over a number of days at ambient temperature. This observation makes all other materials of construction poor candidates for the initial cleanability unless they are oven dried and could affect the speed and accuracy of a study. Continued use of gravimetric measurements for the 316-finish stainless steel and glass coupons would provide data of limited value, given that gravimetric testing demonstrated complete soil removal but did not provide an endpoint determination. Other testing is necessary in conjunction with gravimetric determinations.
TOC testing results, as shown in Table V, were reasonably consistent with the conductivity and gravimetric. One soil, phosphate buffered saline, had no organic carbon to measure and was, therefore, not applicable for this test. The results for the first media were consistently high, as were the conductivity results. The most likely cause would be a low positive control, but because the gravimetric results were consistently good, a repeat of the experiment was not necessary. Most soils required multiple dilutions prior to TOC testing, and testing was time-consuming compared to conductivity and gravimetric testing making TOC a less attractive soil removal confirmation test.
The evaluation of the data to identify the worst-case soil on the hardest-to-clean material of construction should be done in tandem. Table VI ranks the four retained soils on each material of construction from 1 to 4, with 1 being the hardest to clean. The rankings were then averaged across the soils and Media A was determined to be the hardest to clean soil by a small margin over the media with phenol red.
Table VII ranks retention on the six out of 10 materials of construction that demonstrated retention characteristics. The four soils were ranked from 1 to 6, with 1 being the hardest-to-clean material of construction. The rankings were then averaged across the materials of construction and 316-finish stainless steel and nickel-steel alloy (Hastelloy) were determined to be the hardest-to-clean materials of construction by a wide margin.
The worst-case combination of soil and material of construction was Media A on 316-finish stainless steel or nickel-steel alloy (Hastelloy) for the soils and materials of construction tested. The soil and material-of-construction matrix at a facility can be combined into groups based on data and identification of a worst case. A similar analysis for swab recoveries (6, 7) identified stainless steel as a representative material of construction for most equipment and for a large majority of soils. The present study indicates a similar conclusion, (i.e., 316-finish stainless steel is a representative material of construction for cleanability), but the data set is too small to be conclusive. This type of grouping strategy should be explored for each facility to streamline experimentation while still presenting a comprehensive picture of the soils and materials of construction in manufacturing operations.
Cleanability testing of the matrix of soils and materials of construction at a given site can cover all soils on all materials of construction or can be focused based on an evaluation of the components of the soils (1), coupled with preliminary testing over the range of materials of construction. The testing plan should be done within the overall cleaning evaluation including cleaning cycle development, swab and rinse recoveries, and visible residue limit determinations.
Cleanability studies can provide objective data on the hardest-to-clean soil at a facility. Conclusions could differ based on the soils and materials of construction at a given facility. A similar survey of soils and materials of construction should, therefore, be executed. Measuring the cleanability endpoint should be done consistently, including visual observation and at least one other confirmation test. The resulting cleanability data can vary for different materials of construction depending on the soil, and an evaluation of the matrix of soils and materials of construction will provide a path forward for cleaning validation execution.
For this study, based on the materials of construction and soils tested, it can be concluded that performing cleaning validation of the worst-case soils focusing on the 316-finish stainless steel or nickel-steel alloy (Hastelloy) surfaces with conductivity results to confirm visual results is an approach that will provide the most conclusive results. These materials have the added advantage of being the most frequently used materials of construction for all automated cleaning circuits.
Although important, cleanability is only one aspect of a comprehensive cleaning validation program. Personnel cleaning experience, rinse recoveries, swab recoveries, visual residue limits, cleaning cycle development, and cleaning validation are all crucial elements of a well-defined, documented, and defendable cleaning program.
About the Author
Richard J. Forsyth is a senior consultant with Forsyth Pharmaceutical Consulting, 907 Shamrock Ct, Royersford, PA 19468, firstname.lastname@example.org; Keith Bader is senior director of technology and Kelly Jordan is an engineer, both with Hyde Engineering & Consulting, 6260 Lookout Road, Suite 120, Boulder, Colorado 80301.