Accidents and mistakes happen, and sometimes you will need to end your dissolution run. If you see that a true error has occurred, it is important to end the dissolution run and not to collect further samples or analyze data. Continuing a run with an obvious error will usually result in failing results, lengthy investigation time, and re-running the dissolution at least once - if not several times.
What are some of these errors which should result in a run being cancelled? Some that I have seen or been asked about are: a basket falling off or dangling mid-run, dropping 2 tablets into a vessel, pouring the wrong volume of dissolution media in 1 or more vessels, etc. All of these are obvious issues which can be seen by the analyst and would result in a failed test. Only analyst and equipment errors should lead to a cancelled dissolution run.
It is not an error if you make an observation and see that the dosage form itself is behaving differently. If you see cross-linking of gelatin, or a coning issue, or a number of other things - that is important to see and reflects an actual issue with the product itself. That is important to know and the purpose of the test.
So, if you see a true analyst or equipment error, then the entire dissolution run should be cancelled. The reason for the cancellation needs to be documented, and you may want a second analyst, a photo, etc. in addition depending on your SOPs. If samples have been collected, they should be discarded and not analyzed. If samples have been analyzed - then they would need to be reported.
It is also not acceptable to partially cancel a run. By that I mean, you can't choose to only consider the run as an n=5 and then run a 2nd dissolution run to collect an n=1 and combine those 2 runs into a single n=6. This practice can appear like data is being cherry picked in order to create a passing set of data - and has been the subject of 483s in the past.
You should remember the importance of the dissolution test in ensuring safe and effective drugs are released to the public.
Cancel runs where there are obvious analyst errors or equipment malfunctions. Never cancel runs because the dosage form is behaving differently - this can be critical information about the product.
