Sterility Testing: Beware of the Pitfalls

By Martin Spalding, Jr., Northview Laboratories

Contents
Introduction
The Fundamental Limitation
Protecting the Integrity of the Process
Causes of True Positives
Limitations of Testing Defined
Are Isolators the Answer?
Dealing with Sterility Test Positives
References

---

Introduction (Back to Top)

"We have a failure on your sterility test."

Every manufacturer dreads hearing those words. Your first reaction may be to say, "Lab-induced contamination!" And you could be right. But that doesn't mean your testing lab was negligent. More likely you're the victim of the inherent limitations of traditional sterility testing. Statistically, the test itself has a failure rate higher than the process it's designed to monitor.

Sterility testing is widely used in both the medical device and pharmaceutical industries. As a USP test, it is the official procedure for testing the sterility of pharmaceutical products. It is also used on terminally sterilized medical devices, both as a lot release test and for the dose audit phase of sterilization validations.

Despite its importance and its widespread use, most people are not aware of the limitations of the sterility test. The following discussion pertains to conventional sterility testing using direct transfer techniques and not to sterility testing performed in an isolator. Some points apply more to medical devices than pharmaceuticals, as devices are generally more difficult to test.

The Fundamental Limitation (Back to Top)
The fundamental limitation of sterility testing is that the sterility assurance level (SAL) for the process of conventional sterility testing is lower than that for the sterilization processes that it is used to monitor. Unfortunately, there's no escaping the fact that during the test, microorganisms are in close proximity to both the test samples and the open container of culture medium to which the samples must be transferred.

Why are these microorganisms so close to the testing process? Because sterility tests must be performed by people, and people are a prolific source of microorganisms. Humans present a huge introduction of bacteria into the cleanroom, with a total normal flora of >1014 living bacteria cells, 1012 of these residing on the skin and 1010 in the mouth.1 Besides the technicians, test sample packaging, media containers, or testing supplies can be sources of contamination. Through stringent technique, the impact of these and other sources of microorganisms can be minimized, but not eliminated.

Protecting the Integrity of the Process (Back to Top)
Laboratories take extensive precautions to protect the integrity of the sterility testing process. Test samples are removed from the cartons in which they were processed in a nonsterile area, since cardboard harbors many microorganisms and should not enter a sterility test cleanroom. Then, before the samples enter the cleanroom, their exteriors must be disinfected. Note that disinfection is not the same as sterilization. Some microorganisms may remain on the surface of a sample package and end up in the HEPA-filtered hood during the sterility test.

Swipe testing is a first-line defense against microbial contamination, but swiping only tells you after the fact if your process is contaminated.

Culture media for sterility testing must be prepared and steam sterilized in a validated autoclave cycle. It must be subjected to USP growth promotion testing, so up to several days may elapse from the time the media is sterilized until its use in testing. During this time, the exterior of the media containers may pick up microbial contamination, even if stored in a cleanroom. So, like the test samples, the media containers must be disinfected prior to being transferred into a laminar flow hood for sterility testing. Also, the work surface of the hood is disinfected before each sterility test, and the entire cleanroom each testing day.

Sterility test technicians must be fully gowned to contain the microorganisms on their skin and clothing and to protect test samples. Typically, technicians wear a hair cover, hood, face mask, goggles, coverall, shoe covers, and gloves—all sterile, of course. Technicians must be thoroughly trained in proper gowning technique, and in aseptic technique, which is absolutely essential to the sterility testing process.

Causes of True Positives (Back to Top)
Because of all of these precautions, the laboratory generally has a high degree of confidence in the results of the sterility test. True positives will occur due to a variety of causes—an inadequate sterilization cycle, inadequate delivery of the sterilization process to the sample, underestimation of product bioburden, bioburden spikes, resistant organisms on the product, or compromised packaging. These possible causes must be considered whenever a positive occurs.

Limitations of Testing Defined (Back to Top)
Despite stringent operating procedures, good cleanroom practice, and the use of sound aseptic technique, occasional false positives will occur. A microorganism that was not on the test sample will get into a media container used in a sterility test. Experienced microbiologists acknowledge the technical limitations of sterility testing. USP 23 officially recognizes the limitations of sterility testing in two ways. First, chapter <71>, Sterility Tests, permits retesting if an investigation indicates that inadequate or faulty aseptic technique was used in the test. Second, chapter <1211>, Sterilization and Sterility Assurance, describes a sterility assurance level on the order of 10-3 for the sterility testing process. This means that for every 1000 samples tested, one false positive will occur.

This 10-3 SAL, a common reference in the pharmaceutical industry, is based on experience with the sterility testing of pharmaceutical products, typically liquid products in vials. The actual sterility assurance level for the testing of a specific product can vary significantly, depending on the difficulty of the testing procedure. Many medical devices are particularly difficult to sterility test due to the extensive sample manipulation and large media volumes required. So the sterility assurance level for testing a given medical device may be less than the 10-3 level typical for pharmaceutical products.

Are Isolators the Answer? (Back to Top)
False positives occasionally occur in even the most diligent conventional sterility testing operations. So major pharmaceutical companies are investing millions of dollars in sterility testing isolators. Isolators are completely enclosed HEPA-filtered chambers that typically are interfaced with a vapor phase hydrogen peroxide (VHP) sterilizer and/or a steam sterilizer. Sterility testing is performed from outside the unit through glove ports or halfsuits. The testing process is totally isolated from people.

Sterility testing isolators are expensive to install and validate. At the low end, a small isolator with a VHP sterilizer costs at least $250,000. Some pharmaceutical companies have invested over $4 million in more elaborate isolator set-ups.

Unfortunately, it would be difficult and expensive to test most medical devices in a sterility test isolator. A major problem is that Tyvek is permeable to VHP, so elaborate precautions must be taken to avoid undermining the integrity of the sterility test. Also, the size of many medical devices, the number of samples tested, and the volume of test supplies (including one or two media containers per test sample) would dictate the use of a relatively large isolator. To switch to a sterility test process using an isolator and a VHP cycle, manufacturers would have to validate the process for each product. The time required to perform the test would increase greatly in comparison to conventional sterility testing. This means that costs for isolator sterility tests of medical devices could be 3 to 10 times greater than for conventional testing.

Would the increased costs be justifiable for the average manufacturer? Consider the options available when the inevitable sterility test failure occurs. Unlike the aseptically filled pharmaceutical product, which must be discarded if positives are confirmed, most terminally sterilized medical devices can be reprocessed. An occasional resterilization would be far less expensive than routine testing in an isolator.

Dealing with Sterility Test Positives (Back to Top)
What's the bottom line? If you rely on conventional sterility testing, you will eventually have to deal with sterility test positives. Some positives will indicate the presence of microorganisms on the product itself. Others may be due to a contaminant introduced during the sterility testing process.

So be prepared. Write procedures detailing your plan to handle positives when they occur. Give yourself some options in addition to resterilization. In the event of a sterility test positive, the lab may have evidence that would justify invalidation of the test. Your SOPs should state what action to take if the laboratory interprets a sterility test positive as lab-induced. It's also a good idea to evaluate the effect of a second sterilization cycle on your product. Often, this is the fastest way to get beyond the problem. Many manufacturers address the issue of resterilization during their sterilization validation, packaging, and functionality testing.

Isolators have become a valuable technology for some pharmaceutical product manufacturing and sterility testing. For other products, particularly medical devices, conventional sterility testing continues to be an essential tool for validating the effectiveness of the sterilization process. But microorganisms can behave in unpredictable ways. So if you rely on sterility testing, be aware of its limitations!

References (Back to Top)

1. William Hyde, PDA Journal of Science and Technology, "Origin of Bacteria in the Clean Room and Their Growth Requirements", July/August 1998, Volume 52, Number 4, p. 154.
2. USP 23, <1211> Sterilization and Sterility Assurance, p. 1980.
3. Hank Rahe, American Pharmaceutical Review, "Implementing a Cost Effective Sterility Testing Isolator Project, Volume 1, Issue 1, 1998, pp. 34–41.