ASEPTIC PROCESSING: TWO - Risk Management Applications in Pharmaceutical and Biopharmaceutical Manufacturing - page 261

TABLE 10.7 The Relationship Between Mean Microbial
Load in a Batch (Described as Microorganisms/Unit) and
the Frequency of Actual Contaminated Unit [59]

Contamination Distribution Frequency of

Mean Microorganisms/Unit Contaminated Units

1.0 63.2%

0.1 9.5%

0.01 1%

0.001 0.1%

TABLE 10.8 The Relationship Between the Frequency
of Contaminated Units Within a Batch and the Number
of Units Required to be Tested to Identify Contamination [59]

Units Needed to Test Positive Number of Units

Frequency of Contaminated Units Needed to be Test

0.1 (10%) 44

0.01 (1%) 458

0.001 (0.1%) 4603

P = 1 − e − l

where:

P = probability of failing the sterility test

=

e 2.7182818

λ = likelihood of a contaminated unit

Odlaug [61] used this equation to report the magnitude of insufficiency
of final product sterility testing and to define the superiority of parametric
release (for terminally sterilized product). Tables 10.9 and 10.10 summarize the
probability of passing a sterility test (i.e., reporting a type 2 error) with different
percentages of a batch contaminated with at least one microorganism. If 0.1% of
a batch is contaminated, the risk (probability) of failing to identify nonsterility
is 0.98 or 98%. Consider this statistical likelihood of assuring product quality
when applying the mandated sterility test in counterpoise to CFR 211.165,
which also states “Acceptance criteria for the sampling and testing conducted
by the quality control unit shall be adequate to assure that batches of drug
products meet each appropriate specification and appropriate statistical quality
control criteria as a condition for their approval and release.” The aim of the
CFR for establishing statistical quality control criteria for sterile products might
indeed appear contradictory when we fully appreciate the statistical constraints
and culturability limitations of the sterility test.