Sample Re-tests

Situations may exist in which prior knowledge is available about the sample such as trend data, client conversation of a suspect sample or physical sample observation that would indicate a high probability that the initial test result is accurate. Other situations may also lead the investigator to question the validity of the OOS result and request a re-test.

If there is no assignable cause based on the initial investigation and a re-test is requested, the sample is prepared in duplicate, a new standard is prepared with a check standard, and the sample is re-tested to confirm the OOS results.

If the result of the re-test suggests a high probability that the original result was due to laboratory error,the original result will be invalidated, and the re-test results may be reported in accordance with the established procedure to replace the original result with an identified root cause.

Accepting results that are in specification over tests results that are out of specification without clearly identified laboratory error is better known as “testing into compliance”.  This practice is unscientific and objectionable under the cGMP’s.  In the case of a clearly identified laboratory error, the retest results would substitute for the original test result.  It is expected that laboratory error should be relatively rare.

If the result of the re-test does not suggest a high probability that the original result was due to laboratory error, then, in general, all valid test results should be considered as part of the investigation and reported.

For analytical tests, content uniformity could be a factor when evaluating the difference between the original result and the re-tests result and is considered when determining whether to invalidate or report all the data.