Bioburdens and Validation

John Broad, director of microbiology at NAMSA (Irvine, CA), answers questions about validating the sterilization process. Ed Arscott, manager of microbiology (NAMSA, Northwood, OH); Susanne Anderson, technical specialist in marketing (NAMSA, Irvine, CA); and Dave Parente, manager of operations (NAMSA, Kennesaw, GA), contributed to these replies.

If a product supports microbiological growth, will validation be affected?

Products known to support microbial growth include hydrogel-, water-, and saline-based product lines. For these items, the length of time between final packaging and processing for sterilization, including any process interruptions, will need to be considered when determining a sterilization dose. Validation of this phase of the presterilization control of bioburden can be tested by conducting bioburden tests at multiple increments of time. For example, if the product is processed within three days of its manufacture, testing may be performed at 0, 24, 48, 72, and 96 hours. Sampling at 96 hours provides information about the bioburden's growth characteristics should the processing for sterilization exceed three days.

After completion of the validation of growth-supporting products, the samples for dose verification and audits should be irradiated at the maximum time validated (three days in this example) to maintain the schedule.

Assembling standard lots of our product takes several weeks. Will this cause any problem for completing a Method 1 validation?

Minimizing the time between lot production and bioburden testing is important. A delay could result in an artificially low or high bioburden assessment that does not reflect actual bioburden numbers at the time of production. In this case, adjusting the definition of lot for the validation study may be necessary. The purpose of using three lots for bioburden testing is to avoid basing the validation on limited data collected from only one set of production circumstances. If a lot can be redefined to contain samples representing a broad range of production factors, then it can be used for the validation. For instance, in a given day, different production shifts, use of materials or packaging from different lots, or other changes can influence the definition of a lot.

What are the differences between Method 3 and AAMI 13409 substantiation of 25-kGy validations?

Both validation methods target substantiation of 25-kGy sterilization doses for a single batch of products, for initial production of products, for field/clinical trials, and for routine production of small batches of fewer than 1000 devices. In each, bioburden testing and dose-verification experiments are conducted on samples from three production lots. However, Method 3 allows test sample sizes to be based on production lots ranging from 7 to 1000 devices. AAMI 13409, on the other hand, takes into account the pitfalls associated with taking samples from small batches by imposing a minimum sample size of 10 devices for each bioburden and sterility experiment. AAMI rationalizes that the distribution of the natural bioburden on products in small batches (lot sizes of less than 20) may vary and be insufficiently represented if fewer than 10 units are tested, potentially leading to validation failures.

Another significant difference is the AAMI 13409 acceptance criterion when testing 30 or more samples in the verification experiment. AAMI says the statistical verification is acceptable if there are no more than two positive tests. By contrast, Method 3 accepts only one positive test.

In a final variance, Method 3 recommends four successful validation runs (combined Method 3A and 3B) to routinely dose at 25 kGy without substantiation on a batch-by-batch basis. However, the AAMI 13409 substantiation method requires only three successful verifications, provided more than one batch is produced every three months. In AAMI 13409, it is expected that the verification dose will be recalculated and tested for each subsequent dose audit after validation is completed.


No comments: