"We have a complicated computer-controlled manufacturing system for our product," explained our GMP Agent-In-Place. "It includes several vessels, computer-controlled valves, and computer-controlled clean-in-place (CIP) systems. With low production levels at startup, everything ran fine—there was only one batch in process at a time. But when product demand increased, we began to process two batches at once. We didn't think it would be a problem because they would be in separate vessels. However, when the first batch was done, we needed to clean the last vessel prior to processing the second batch. Apparently, during validation we never checked the upstream valve status during the last vessel's CIP procedure, because the computer opened the bottom valve to drain on the penultimate vessel. The entire second batch was drained before we realized there was a problem—a $200,000 problem."
Double-duty filter"Sometimes processing aides do more than we think," began our GMP Agent-In-Place. "For instance, we used a depth filter in our production to help separate the proteins. We changed to a different manufacturer and the first lots produced all met the release criteria. The test results looked normal. It was only when the stability data showed faster degradation that we saw a problem. The degradation was clearly associated with the change of filter, and caused a recall of several batches. The customer was not pleased.
"After an enormous amount of investigation and testing, it turned out that the original filter had also removed protease from the resultant filtrate that becomes our product. The protease would proteolyse our protein product over time, showing up as degradation in our stability studies," sighed our Agent.
"A lot can happen with filtration," our GMP Agent-In-Place said. "We use a filter-aid during the production process of our sterile-liquid product. Filter-aid is also known as diatomaceous earth and is mined from the ground—its chemical composition can vary.
"During a stability study, when the samples were initially retrieved, we'd sometimes see a slight haze wafting from the bottom of the vial. We would only see it upon first movement; the haze would dissipate into the air. Exploratory testing was conducted to identify the haze. The haze could not be collected, so we tested the liquid for a variety of compounds. We identified some cadmium in our products. It turned out that cadmium was in the filter-aid we had used. Because filter-aid is mined, it is variable, and at least one location had some cadmium contamination.
"The identification of cadmium in our product resulted in a long medical assessment, including testing every batch that's still in-date. We found detectable levels in hundreds of batches. The release of filter-aid now includes a test for heavy metals. The search for the source of the haze is still ongoing."
"Sometimes it's the simple stuff that gets you," grumped our GMP Agent-In-Place. "Someone forgot to place the temperature probe in a vessel. The computer-controlled heating step continued to pump heat into the vessel's jacket and thereby into the product waiting for the probe to report the proper temperature. Because the probe wasn't in the vessel, it never sensed the correct temperature. When this was discovered hours later, the product was overcooked and had to be rejected.
Pharmaceutical Technology's monthly "Agent-in-Place" column distills true industry tales from the files of Control, a senior compliance officer. If you have a story to share, please email it to Control at AgentinPlace@advanstar.com