Puzzling Results Across Dilution Factors
Hello All, reaching out with what I'm sure is a common occurrence in food micro, but did not see it asked on this board. Here is the situation:
A food sample is tested for e.coli via petrifilm at a 10^-2 and 10^-3 dilution. The -2 plate (less diluted) shows no e.coli growth, but on the -3 plate (more diluted) there is a single e.coli colony. Our specification is that e.coli should not be detected in the sample. This is puzzling, because you would expect that a single colony on a -3 plate would be backed up by at ~10 colonies of e.coli on the -2. I would think that lab error would be the root cause of this single colony. However, I do know resampling in the face of detected pathogens is a sensitive issue. Would this case be considered unique from a regulatory perspective, where the spread of results from initial testing just don't seem to make scientific sense, so retesting would be in order?
Hello All, reaching out with what I'm sure is a common occurrence in food micro, but did not see it asked on this board. Here is the situation:
A food sample is tested for e.coli via petrifilm at a 10^-2 and 10^-3 dilution. The -2 plate (less diluted) shows no e.coli growth, but on the -3 plate (more diluted) there is a single e.coli colony. Our specification is that e.coli should not be detected in the sample. This is puzzling, because you would expect that a single colony on a -3 plate would be backed up by at ~10 colonies of e.coli on the -2. I would think that lab error would be the root cause of this single colony. However, I do know resampling in the face of detected pathogens is a sensitive issue. Would this case be considered unique from a regulatory perspective, where the spread of results from initial testing just don't seem to make scientific sense, so retesting would be in order?
Hi Brothbro,
A few comments.
I assume the observations relate to a Procedure for evaluation of generic E.coli.
A MPN-type Procedure is preferable for the numerical assessment of low levels of generic E.coli which is not, per se, regarded as a pathogen..
IMEX It is usual to seed duplicate plates for each dilution. I deduce the other 3 plates involved were all negative.
You have probably discovered a statistical aberration but which could reflect a genuine low level of generic E.coli.
I suggest to repeat the evaluation using more than one sample / a more statistically precise Procedure if the Specification requirement is significantly important to you.
Hi Brothbro,
A few comments.
I assume the observations relate to a Procedure for evaluation of generic E.coli.
A MPN-type Procedure is preferable for the numerical assessment of low levels of generic E.coli which is not, per se, regarded as a pathogen..
IMEX It is usual to seed duplicate plates for each dilution. I deduce the other 3 plates involved were all negative.
You have probably discovered a statistical aberration but which could reflect a genuine low level of generic E.coli.
I suggest to repeat the evaluation using more than one sample / a more statistically precise Procedure if the Specification requirement is significantly important to you.
Hi Charles.C, I appreciate the quick response! Re-evaluation with a greater sample size would indeed give us more insight into whether the single colony reflects lab error, an aberration, or an issue with the manufacturing process.