Sterility

Membrane filtration is preferred wherever the product can be filtered. The preparation is passed through a 0.45 micron membrane which retains any organisms present, the membrane is rinsed to remove residual antimicrobial activity, then divided between fluid thioglycollate medium for anaerobes and facultative organisms and soybean casein digest medium for aerobes and fungi.

Direct inoculation is used where filtration is impossible, with the product transferred straight into the media. It is more vulnerable to inhibition by the product itself, since the antimicrobial substance is never rinsed away.

Incubation runs fourteen days, at 30 to 35 degrees for thioglycollate and 20 to 25 degrees for soybean casein digest, with the containers examined for turbidity throughout rather than only at the end.

Sterility testing is a sampling operation. Only a defined number of units from a batch are examined, and the probability of detecting contamination depends on how widespread it is. Sparse contamination in a large batch is very likely to be missed.

A pass therefore means that no growth was observed in the units tested under the conditions used. It does not license the statement that the batch is sterile. Sterility assurance is a property of a validated manufacturing process; the test is a confirmatory check on it, not a substitute for it.

Regulators have been explicit on this point for decades, and any certificate implying that a sterility pass certifies a batch is overstating what the method can deliver.

The dominant risk in sterility testing is contamination introduced during the test itself. That is why the work is done in an isolator or a restricted-access barrier system with continuous environmental monitoring, and why the facility requirement is as much a part of the method as the media.

Where growth is observed, an investigation follows: the organism is identified and compared against the environmental monitoring flora of the suite. A skin commensal appearing in a single container with no corresponding process deviation points to the test; an organism consistent with the product or its manufacturing environment points to the batch.

A laboratory without isolator or barrier capability cannot perform this determination to standard, and should say so rather than approximate it with an enumeration test.