Ladi ☆ Thailand, 2022-11-08 04:07 (763 d 07:16 ago) Posting: # 23357 Views: 3,555 |
|
Dear forum members, What should be the acceptance criteria for the blank and zero samples that we inject before calibration curve standards in analytical run? According to the ICH M10 guideline on bioanalytical method validation, section 3.3.1 Analytical run, states that "An analytical run consists of a blank sample (processed matrix sample without analyte and without IS), a zero sample (processed matrix with IS)". However, I couldn't find the acceptance criteria for the blank and zero sample in the guideline. Our current SOP is to the accept the analytical run if interference at RT of analyte in blanK and zero sample is less than 20% compare to LLOQ and if interference at RT of IS is less than 5% compare to IS response in LLOQ (basically follow Selectivity criteria). I have seen many cases where some runs are rejected because the blank sample is contaminated from sample preparation error (zero sample and pre-dose samples are clean). I personally don't think it make sense to reject a whole run from one contaminated sample! Any suggestion or guidelines, papers I can use to use as reference to modify our SOP? Thank you for any suggestion or comment, Ladi |
ElMaestro ★★★ Denmark, 2022-11-13 21:57 (757 d 13:27 ago) @ Ladi Posting: # 23362 Views: 2,739 |
|
Hi Ladi, ❝ What should be the acceptance criteria for the blank and zero samples that we inject before calibration curve standards in analytical run? ❝ ❝ According to the ICH M10 guideline on bioanalytical method validation, section 3.3.1 Analytical run, states that "An analytical run consists of a blank sample (processed matrix sample without analyte and without IS), a zero sample (processed matrix with IS)". ❝ ❝ However, I couldn't find the acceptance criteria for the blank and zero sample in the guideline. Technically that seems correct. ❝ Our current SOP is to the accept the analytical run if interference at RT of analyte in blanK and zero sample is less than 20% compare to LLOQ and if interference at RT of IS is less than 5% compare to IS response in LLOQ (basically follow Selectivity criteria). Sounds good to me. ❝ I have seen many cases where some runs are rejected because the blank sample is contaminated from sample preparation error (zero sample and pre-dose samples are clean). ❝ ❝ I personally don't think it make sense to reject a whole run from one contaminated sample! I guess this is your major point. On the other hand, one can also argue it does not make sense to inject a run if there are contaminated samples. Much of it boils down to the time by which you conclude there is contamination. If you observe interference (unwanted signal in the blank or zero) at the RT, and you do an investigation and conclude that there was contamination in the blank / zero but certainly not in any other injection then you are almost begging for an inspection that is very difficult for you from the very beginning. Secondly, if you conclude there was contamination, how is that conclusion actually reached? If there is a source document stating that someone put ref std in the blank then we're back to the first counter-argument. If you conclude that contamination was present because you think you have ruled out all other potential explanations then we're back at the second counter-argument. To me, it makes perfect sense to reject a run if there is "contamination" in the blank / zero. The alternative is much worse. ❝ Any suggestion or guidelines, papers I can use to use as reference to modify our SOP? Your SOP is good. If you too often can't comply with it then it may be something else entirely that is your real problem. Along the same lines, bonus info: Quite a few CROs are currently experiencing sporadic carry-over issues with analytes measured in the the pg/mL range. I recommend sponsors to always insist on also having a double blank right after the highest calibrator in the PK-runs and to have some rules in-house for handling situations when there is CO in this injection. CROs really do not like it, simply because they realise their assays at study-time is not behaving as well as it may have been at the time of validation. The solution is certainly not to only insist that CO is assessed as part of the validation, not as part of the study. — Pass or fail! ElMaestro |
Ladi ☆ Thailand, 2022-11-15 05:48 (756 d 05:35 ago) @ ElMaestro Posting: # 23366 Views: 2,696 |
|
Hi ElMaestro, Thank you for your suggestions. ❝ I guess this is your major point. Yes it is. ❝ On the other hand, one can also argue it does not make sense to inject a run if there are contaminated samples. Much of it boils down to the time by which you conclude there is contamination. ❝ To me, it makes perfect sense to reject a run if there is "contamination" in the blank / zero. The alternative is much worse. In my opinoin, the guideline do give some room for some imperfections. For example, one of my LQC can have %accuracy > 115% and fail. It could be because of contamination from sample prep. The run still pass as it passes the QC acceptance criteria. For an extreme example, my LLOQ can fail from sample prep. in which the guideline allow me to use 2nd calibration point as LLOQ for the run. I do not have to reject the run. But why such rigid acceptence criteria is applied to the BK and Zero samples eventhough is not explicitly required in the guideline. What about pre-dose samples? If you do both pre-dose without internal standard and with internal standard. Do you reject the run if only pre-dose without internal standard contained some amount of analyte at about LLOQ? ❝ Your SOP is good. If you too often can't comply with it then it may be something else entirely that is your real problem. I was planning to change it to do duplicate BK and Zero samples. The run fail if both fail as the contamination is consistent. If only one fail, run is accepted as contimination is random. What do you think? However, I won't do duplicate for calibration standards, just BK and Zero. ❝ Quite a few CROs are currently experiencing sporadic carry-over issues with analytes measured in the the pg/mL range. I recommend sponsors to always insist on also having a double blank right after the highest calibrator in the PK-runs and to have some rules in-house for handling situations when there is CO in this injection. CROs really do not like it, simply because they realise their assays at study-time is not behaving as well as it may have been at the time of validation. The solution is certainly not to only insist that CO is assessed as part of the validation, not as part of the study. Thank you, I like your suggestion. We do have a double blank right after the highest calibrator in every run, but we do not apply any criteria to it yet. Do you suggest same criteria as BK before calibration is applied to it? Thank you, Ladi |