OR WAIT null SECS
Failure to disclose info may work sometimes, but eventually every question will be answered.
"It seems simple, as do many deviations," began our GMP Agent-In-Place. "The deviation was for an autoclave vent-filter integrity test failure. The vent filters are tested 'in-line,' meaning that we don't remove them from the autoclave, and the test equipment is built into the system. The root cause was found pretty quickly—a failure of an air-line valve, which allowed more air flow. This added air flow looked, to the testing equipment, like a filter leak, thereby failing the test.
"The manufacturing department did the obvious things. They first fixed this valve and retested the filter, which passed. This was great because it means we didn't need to recall a bunch of batches. Second, the valve was added to the preventive maintenance (PM) program. The department even added its other autoclave to the PM program in a similar manner. However, we had subsequent similar deviations on other autoclaves at this site—in other departments," our Agent groused. "No one had thought to assess them similarly. When I mentioned this to a colleague, the first thing she said was 'What about the lyophilizers? They have vent filters too!' Ouch."
"Prior to having statistical analysis systems (SAS) on every PC, or even on the site mainframe, we used a facility in New York City to run our programs," noted our GMP Agent-In-Place. "We would type up the data cards (yes, it was a LONG time ago) locally, and check them thrice. At the time, I was working on clinical efficacy data for a veterinary product. Since I worked in upstate NY, it was a plane ride to New York City with my card deck. One time, I worked with our local SAS expert for the program, and it was actually pretty simple. We walked the cards over to the contract SAS facility, and the next day picked up the output and flew back home. I should have had a second (and maybe a third) person check the data on the cards because, when I got back with the printout to write the report, I found a data-input error. Luckily I had left the card set with my colleague in New York City who ran the program for me again after correcting the erroneous piece of data."
"We had a new product that was filled into several-sized containers," said our GMP Agent-In-Place proudly. "So we had a label and a unit carton for each size. It turns out that we had FDA license approval based on draft labeling. A final labeling submission was allowed to take place after product sales started. During assembly of the label submission, it was noted that one of the unit cartons was missing the required word 'Liquid' in the generic name. All the other unit cartons did, but not this one. We fessed up to FDA, and luckily they did not require a recall. Whew."
Far from quarantined
"Our third party distributor was trusted to handle quarantine material," our GMP Agent-In-Place noted. "They were to segregate the quarantine material and only ship the batches we specified. It was our responsibility to check the release status before requesting shipment. On one occasion, however, they had some pallets backwards in the aisle, so the quarantine placard couldn't be seen. So when they were asked to ship eight pallets of batch A, they picked them out of the aisle on the assumption that the pallets were not intermixed. Unfortunately, the pallets were intermixed and with the backwards pallet, one pallet of batch B was shipped. Batch B was not yet released. Neither the picker nor the checker caught the change in batch number nor the quarantine placard. When the mistake was discovered, we sent a representative to the customer to have it shipped back—luckily before it was used.
Pharmaceutical Technology's monthly "Agent-in-Place" column distills true-life cautionary 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. We won't use any names, but if we do use your experience in the column, you'll receive a Pharmaceutical Technology t-shirt.