Log in
Register|
sarada06884 ☆ India, 2014-03-20 10:55 (4084 d 17:30 ago) Posting: # 12675 Views: 6,330 |
|
|
Hi.. We are writing a SOP for Bioanalytical method validation. As per US FDA's guideline on Bio analytical method validation "Recovery experiments should be performed by comparing the analytical results for extracted samples at three concentrations (low, medium, and high) with unextracted standards that represent 100% recovery" During validation we demonstrate QCs at HQC, MQC1, MQC2 (lower conc. than MQC1), LQC levels. Now my doubt is if an additional QC (say MQC3 (lower conc. than MQC2)) is introduced into the run, so that the subject samples meet at least two levels of QCs as per EMEA's guideline recommendation, is there a chance that the regulator may ask us to demonstrate the recovery at that additional QC level also? Is a bracketing approach valid in this case? Regards P.S.Srinivas Edit: Category changed. [Helmut] |
|
Ohlbe ★★★ France, 2014-03-20 12:19 (4084 d 16:06 ago) @ sarada06884 Posting: # 12677 Views: 5,653 |
|
|
Dear Srinivas, ❝ As per US FDA's guideline on Bio analytical method validation "Recovery experiments should be performed by comparing the analytical results for extracted samples at three concentrations (low, medium, and high) with unextracted standards that represent 100% recovery" Which is actually incorrect: you should compare with samples spiked post-extraction, not with unextracted standards. Otherwise in LC-MS/MS methods what you will calculate is a mix of recovery and matrix effects, not just recovery. ❝ Now my doubt is if an additional QC (say MQC3(lower conc. than MQC2)) is introduced into the run, so that the subject samples meet at least two levels of QCs as per EMEA's guideline recommendation, is there a chance that the regulator may ask us to demonstrate the recovery at that additional QC level also? I think not. There were even discussions at some bioanalytical conferences lately, where the FDA was present, that checking recovery at 2 levels of concentration (low and high) should be enough. The draft guidance asks you to check at 3 levels of concentration, not at each QC level. — Regards Ohlbe |
|
Ken Peh ★ Malaysia, 2014-03-23 08:23 (4081 d 20:02 ago) @ Ohlbe Posting: # 12688 Views: 5,529 |
|
|
Dear Ohlbe, ❝ Which is actually incorrect: you should compare with samples spiked post-extraction, not with unextracted standards. Otherwise in LC-MS/MS methods what you will calculate is a mix of recovery and matrix effects, not just recovery. The unextractd standard is drug solution. Right ? Comparing the response of sample after extraction and unextracted standard which is drug solution of same concentration. Do you mean instead of using drug solution, we should compare with sample spiked post-extraction. Kindly elaborate. What about samples that do not go through extraction step ? Plasma samples treated with just the addition of ACN or methanol before injecting into LCMSMS. We have been doing recovery for sample treated in this way. Is it necessary to have recovery ? We also compare the response of treated sample and drug solution (untreated sample) having similar concentration. Appreciate your comment. Thank you. Regards, Ken |
|
nobody nothing 2014-03-23 12:01 (4081 d 16:24 ago) @ Ken Peh Posting: # 12689 Views: 5,502 |
|
|
❝ What about samples that do not go through extraction step ? Plasma samples treated with just the addition of ACN or methanol before injecting into LCMSMS. ❝ We have been doing recovery for sample treated in this way. Is it necessary to have recovery ? You might precipitate in real samples some of the analyte with the protein in your sample when adding ACN or methanol. But in general, as long as you treat standards, QCs and verum samples identically (and have representative empty matrix), in most cases recovery is not very interesting. More a parameter to take care for while debugging, in my opinion... Kind regards — Kindest regards, nobody |
|
Ohlbe ★★★ France, 2014-03-23 17:28 (4081 d 10:57 ago) @ nobody Posting: # 12690 Views: 5,507 |
|
|
Dear Nobody, ❝ [...] in most cases recovery is not very interesting. More a parameter to take care for while debugging, in my opinion... I'd say recovery is more something to check during method development, to optimise your extraction method. But indeed it is of limited interest in method validation. That is, unless indeed the FDA requests it to be done on 6 different sources of matrix, as planned in their draft BMV guideline. I've seen a poster recently mentioning a variable recovery from one lot to the next. — Regards Ohlbe |
|
Ohlbe ★★★ France, 2014-03-23 17:37 (4081 d 10:48 ago) @ Ken Peh Posting: # 12691 Views: 5,562 |
|
|
Dear Ken, ❝ Do you mean instead of using drug solution, we should compare with sample spiked post-extraction. Kindly elaborate. Yes. Actually you can easily combine matrix effects and recovery experiments: - extract spiked samples in 6 different lots of matrix - spike the same 6 lots of matrix post-extraction, at the same level of concentration - prepare neat solutions, at the same level of concentration - inject all these samples - compare the response in spiked samples to samples spiked post extraction: recovery - compare the response to samples spiked post-extraction to neat solutions: matrix factor - if you compare the response in spiked samples to neat solution: what you see is a kind of overall efficiency, which takes into consideration recovery + matrix effects. ❝ What about samples that do not go through extraction step ? Plasma samples treated with just the addition of ACN or methanol before injecting into LCMSMS. I agree with Nobody: part of your analyte and IS can get trapped in the precipitate and the recovery may not be 100 %. It depends on the precipitation reagent, the way it is added, how you mix the sample afterwards, etc. — Regards Ohlbe |
|
Ken Peh ★ Malaysia, 2014-03-27 20:16 (4077 d 08:09 ago) @ Ohlbe Posting: # 12732 Views: 5,427 |
|
|
Dear Ohlbe, Thank you very much for your elaboration. Regards, Ken |
Ing. Helmut Schütz