抄録
It is no doubt that media fill (simulation) test is the best method to evaluate aseptic processing. However, there are some problems concerning the acceptable criteria of “a contamination rate of less than 0.1% with 95% confidence limit”, which is specified in official compendia, such as JP,USP, EU GMP, WHO GMP, and ISO 13408 part 1 (General). In performing the media fill test, critical issues in performing media fills are the number of fills. Statistical validity of the observed contamination rate for the process has not been well explained hitherto. Problems associated with the “number” of criteria may not necessarily be treated with statistics, and personnel training and operation are sometimes major factors to cause contamination. However, dogma of “a contamination rate of less than 0.1% with 95% confidence limit” should be discussed further. Operation characteristic curves, which show the relationship between percent defectives and acceptance probability, clearly indicate that the more the number of fills, the higher the acceptable probability of contaminated units becomes in the range of contamination rate less than 0.1%. This means that the larger number of fills accept the worse average quality at the range of percent defectives less than 0.1%. According to a table of ISO 13408-1 (1998), which shows the relationship between the number of fills and the number of acceptable contaminated units, 10 positive units are accepted by 16970 fills' as the rate of contaminated units are less than 0.1% with a 95% confidence limit. However, the 95% confidential range of 4 positive units in 16970 fills is calculated to be 1/1970-6/16970. This means that zero (0) positive unit in 16970 fills, that is sterile prduct, is out-of control with a 95% confidence level. Even though the product meets the requirement of “less than 0.1% with a 95% confidence limit”, zero contamination unit in 16970 fills or asepsis must be very rare case. It means that sterile products become out-of-control according to the existing criteria. It is recommended that the dogmatic criteria such as less than 0.1% with 95% confidence limit should be revised. An example of a proposed new criteria is expressed as follows: “The result of media fill should show that the aseptic processing is consistently within the control limit (with a confidence limit of 95%) by using 3000 or more fills, and show the asepsis of the process or the validity of sterility of the product” in stead of “less than 0.1% with a 95% confidence limit.”
