Issue link: https://www.e-digitaleditions.com/i/1373953
6 Pharmaceutical Technology BIOLOGICS AND STERILE DRUG MANUFACTURING EBOOK 2021 P h a r mTe c h . c o m tors, and some changes incorporated. The study's first phase began in February 2019 with equipment in place at PDA Training and Researc Institute, using marginally failed filters and resources sup- plied by filter manufacturers. Setting a baseline for filter blinding The study was a first step in successfully establish- ing a baseline for blocking and blinding of margin- ally flawed 10-inch cartridge filters under extreme conditions, even though these are not typically found during commercial manufacturing. Start- ing tests at extreme conditions (i.e., high foulant concentration and 90% blocking rate) allowed the team to decide whether the masking effect were even possible and if it was, to find the edge of fail- ure, and zero in on the conditions that would be considered high risk for masking. In the first phase, 24 10-in. filter cartridges were subjected to a foulant concentration of 24 g/L of the malt chocolate drink Ovaltine (typical foulant concentration for total throughput studies is 0.8 g/L, mimicking biologic solutions) and blocked by up to 90%. Two of the pre-use test failed filters were found to have passed the post-use integrity test. This study established scientific data that, in extreme foulant concentrations and filter block- ing circumstances, a pre-use filter f law could potentially be masked. Establishing a set of extreme, non-commercial, fouling conditions where masking could occur, provided an important starting point. Now, ef- forts shifted to finding the edge of failure and to see how closely it approached typical commercial conditions. Therefore, the foulant concentration rate was reduced to levels found in typical biopro- cesses. In addition, the blocking rate of a sterilizing grade filter was adjusted to typical levels found in manufacturing, of approximately 20–30%. Test conditions were changed by using a high and average foulant f luid at different blocking rates (Table I). This matrix test condition would cre- ate a much larger data set, which would be used to determine the possible causes of masking. Because it is very difficult to gain marginally failed 10-in. filter cartridges from filter manufac- turing sites, 47-mm filter composite discs were used, which featured defects created by a con- trolled laser drill. The 10-micron filter flaw looked similar to those discovered during the first phase of testing, with an additional side benefit: such membranes can be used in future work to evalu- ate results of exposing filters to f luids that are sus- pected to produce high levels of fouling. The results showed that eight f lawed filters at 24 g/L foulant rate and high blocking all still failed the post-use test. However, of 44 filters that were tested at a foulant concentration of 0.8 g/L, two f lawed filters passed the post-use test. These two filters have a blocking rate of 81% and 97%, which is, again, an unusually high blocking rate for steril- izing grade filters. Manufacturing Table I. Foulant/blocking rate matrix for the masking trials performed. Foulant concentration Blocking rate 24 g/L 25% 50% 75% 90% 0.8 g/L 25% 50% 75% 90%