Log in
Register|
Loky do ★ Egypt, 2025-11-19 12:40 (196 d 10:32 ago) Posting: # 24506 Views: 3,104 |
|
|
Hello everyone During bioanalysis we noticed samples with potential samples switch (two to three samples only), hence we performed reanalysis for this reason, the authority raised concerns about this, we proposed to the authority to use the original concentration as an evidence for Bioequivalence, is this sufficient Thanks in advance Our question is: Is the use of the original, pre-reanalysis concentration results – supported by available documentation and investigation records – considered sufficient and acceptable evidence for the bioequivalence conclusion, given that the reanalysis was triggered due to suspected sample switch in only a limited number of samples? Furthermore, what additional justification, documentation, or procedural steps would be required to ensure regulatory acceptance in this situation? We currently have an ongoing authority audit, and the auditor has raised this question and we proposed this solution. As this point is under active discussion, we kindly request your guidance and a official response. Edit: Instead of replying to your own post, please edit it. I deleted your follow-up posts and merged them into this one. [Helmut] |
|
dshah ★★ India, 2025-11-20 12:04 (195 d 11:09 ago) @ Loky do Posting: # 24511 Views: 2,609 |
|
|
Hi Loky do! I believe that SOP is already there and the re-analysis is based on that SOP. If not, how was it identified that samples swapping had happened? You will require a CAPA. Divyen |
|
qualityassurance ★ 2025-12-05 15:06 (180 d 08:06 ago) @ Loky do Posting: # 24521 Views: 2,351 |
|
|
Hello. First of all as per ICH "reanalysis of study samples for a PK reason (e.g., a sample concentration does not fit with the expected profile) is not acceptable, as it may bias the study result." However, as a corrective action (CA) you can present BE outcome by including and excluding the subject in question. As a preventive action (PA) you need to update your repeat SOP to avoid such cases. Regards, QA |
|
Achievwin ★★ US, 2026-02-06 13:31 (117 d 09:42 ago) @ qualityassurance Posting: # 24563 Views: 1,619 |
|
|
❝ First of all as per ICH "reanalysis of study samples for a PK reason (e.g., a sample concentration does not fit with the expected profile) is not acceptable, as it may bias the study result." My favorite topic: In 2004 or so there was a round table discussion - "PK samples reanalysis - to do or not to do" a topic everyone want to talk but every one is afraid. The consensus was you can reanalyze PK samples if there is a compelling reason - sponsor or CRO MUST have an SOP before the analysis start. Another important point has been even though folks have SOP they don't enforce the SOP consistently across all subjects and all samples. Sponsor perspective: CRO send garbage data and sometimes pulling down the study, when questioned on the quality response is so complicated citing several SOPs and guidances. CRO perspective: Sponsor Cherry picks samples and ask for reanalysis our reputation may be at risk as we serve clients from different countries and we know global regulations, if an inspection occurs and 483 is issued our reputation is at risk. Regulatory perspective: are these guys are covering up anything? - this is when the problem starts answer is full transparency - present all documentation submit your analysis both ways with original values and with reanalysis values. As a sponsor if you are on an prestigious project or multiple projects you better have an well written SOP with a decision tree how you pick samples for reanalysis, how you pick PK repeats and who authorizes from the sponsor side, how you decide on reporting the sample concentrations after reanalysis. We had a criteria 1) pick one sample on either side of the suspected outlier, analyze each sample in duplicate 2) after reanalysis report 1) original value if the renalaysis value and original value don't differ no more than 15%, 2) report median or average value if the samples differ by 15-25%, 3) report suspected outlier as non reportable if the values differ by >25%. Important points are 1) having an SOP before the study start 2) communicating that SOP with your CRO or BA lab, and 3) following the SOP consistently. In your case sample switch happened, it is more a study compliance issue, you have to investigate root cause of sample switch suspicion, even if it is one or two samples if you can not assign a reason (usual case) it is only suspicion, then don't report the values (we have had an experience we suspected sample switch during pre-dose samples - as usual clinic said we are saints, CRO monitor said I did my job meticulously - a a sponsor we ended up losing 11 subjects and excluded that one entire center. but we presented the data with and without those subjects and it did not make any difference) moral of the story Full transparency - but sadly who pays for sloppiness of CRO or site staff? |
|
Helmut ★★★   Vienna, Austria, 2026-02-09 12:40 (114 d 10:33 ago) @ Achievwin Posting: # 24564 Views: 1,571 |
|
|
Hi Achievwin, I had the unpleasant experience of attending the “Chrystal City I” conference (Arlington 1990). For an early collection of publications see this post. Quote from the 1991 paper: Repeal Analysis – The protocol for repeat analysis should be established a priori. Some aberrant values may be identified which can be attributed to processing errors, equipment failure, poor chromatography or quality control samples outside predefined tolerance. Cautious use of ‘pharmacokinetic fit’ such as a double peak may call for repeat analysis of some samples in the study; but the reasoning should be clearly documented. And from the one of 2000 – which lead to the FDA’s 2001 guidance:Repeat Analysis: A standard operating procedure or guideline for repeat analysis and their acceptance criteria must be established a priori. This SOP or guideline should define acceptable reasons for repeating sample analysis, such as sample processing errors, equipment failure, poor chromatography, etc. Cautious use of “pharmacokinetic fit” such as double peak may call for repeat analysis of some samples in the study. The rationale for the repeat analysis and the reporting of the repeat analysis should be clearly documented. The 2001 guidance:Repeat Analysis: It is important to establish an SOP or guideline for repeat analysis and acceptance criteria. This SOP or guideline should explain the reasons for repeating sample analysis. Reasons for repeat analyses could include repeat analysis of clinical or preclinical samples for regulatory purposes, inconsistent replicate analysis, samples outside of the assay range, sample processing errors, equipment failure, poor chromatography, and inconsistent pharmacokinetic data. Reassays should be done in triplicate if sample volume allows. The rationale for the repeat analysis and the reporting of the repeat analysis should be clearly documented. (my emphases)In my CRO we had a blinded review of data by a pharmacokineticist (not the bioanalyst). We applied a similar procedure like you (until the EMA’s guideline was published in 2011). Only then we were protected against suspected sample mix-ups in the clinical phase (see this example, slides 9–11). History… BTW, the ICH M10 guideline of 2022 (overruling the FDA’s and the EMA’s) allows reanalysis of pre-dose samples if their concentration ≥ LLOQ. Well, cough, why? — Dif-tor heh smusma 🖖🏼 Довге життя Україна! ![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png) Helmut Schütz ![[image]](https://static.bebac.at/img/CC by.png) The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
Helmut ★★★ Vienna, Austria, 2026-02-09 13:41 (114 d 09:32 ago) @ qualityassurance Posting: # 24565 Views: 1,554 |
|
|
Hi QA, ❝ First of all as per ICH "reanalysis of study samples for a PK reason (e.g., a sample concentration does not fit with the expected profile) is not acceptable, as it may bias the study result." At a joint EGA/EMA workshop (London, June 2010) Gerald Beurle (TEVA/ratiopharm) presented a hypothetical example, where due to a single sample switch a study would pass BE. However, it would fail if concentrations are swapped or even the entire data of the subject would be excluded from the comparative analysis. Consensus expressed by the panel members of the PK working party (responsible for the GL): “The value must not be reanalysed. It has to be kept as it is – even if the study would falsely pass BE.” That was too much for me. I said:“The EMA is a serious risk to public health!” IMHO, this is not science anymore but a belief system: What we measure once is the truth.
— Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
ElMaestro ★★★ Denmark, 2026-02-09 14:51 (114 d 08:22 ago) @ Helmut Posting: # 24566 Views: 1,525 |
|
|
Hi all, I am reading all this with great interest, but also with a bit of apprehension. Regulators have over the years received so much garbage from CROs and companies that they have to set some rules to protect the patients. More generally, I am firmly in the camp of people who think it is absolutely necessary to restrict and regulate an applicant's ability to change reported values, simply because history has shown that certain industrial players can't handle such a privilege in a way that is favorable to patients. It sends shivers down my spine when I hear those "what if"-type of arguments (hypothetical case studies), which are from the outset designed to refute an idea, good or bad. Show me any solution to any problem in the field of BE and I can probably conjure up some kind of hypothetical example which purportedly shows that the solution is bad. I dare you, come on do it! But, and here is the absolutely salient point, my refutation that shows that the solution is bad might not apply generally. Perspective: "Linear up, linear down" vs "linear up, log down" in NCA. If someone wants to convince me that "linear up, linear down" is disadvantageous then the approach will be to show me that T/R is biased when using it. The argument is not won by showing me a profile and explaining that the area is upward biased or whatever (an upward bias, or a downward one for that matter, does not necessarily affect T/R unless it affects T and R unequally. It is the latter point that is the deal-breaker). So let's get real, and let's get general: I think that patients are well protected if CROs apply healthy rules equally to T and R because then T/R is not expected to be biased. CROs can be trusted to do so, if the rules are defined a priori. Ideally if they are applied blindedly. I will make a point out of saying that I think there is difference between recording and reporting. I think scenarios exist in which a re-analysis is a potentially useful tool as part of an investigation (they are rare, though), but that should not necessarily affect the reported value. A PK-value for reporting should only be changed if there is proof that the original value is tainted. The proof is not by itself just that the value looks odd. I have some degree of sympathy for people who say that investigations almost never give rise to identification of errors, they are for display only. For the same reason I see why some high-quality labs by principle simply do not do re-analyses as part of an investigation. They invest in doing the analysis well from the outset in stead. Got a weird value, no root cause, and no SOP or healthy rule for re-analysis? Well, it simple then. Report the value, submit the dossier, and let the regulator decide what is fair in that situation. In the meantime, spending the energy on writing an SOP and helping colleagues perform in the lab so that their work does not produce funky values in the future, those are are good investments. — Pass or fail! ElMaestro |
|
Helmut ★★★ Vienna, Austria, 2026-02-10 13:17 (113 d 09:56 ago) @ ElMaestro Posting: # 24567 Views: 1,459 |
|
|
Hi ElMaestro and all, ❝ I am reading all this with great interest, but also with a bit of apprehension.  ❝ It sends shivers down my spine when I hear those "what if"-type of arguments (hypothetical case studies), which are from the outset designed to refute an idea, good or bad. Show me any solution to any problem in the field of BE and I can probably conjure up some kind of hypothetical example which purportedly shows that the solution is bad. I dare you, come on do it! All case studies (not to use the buzz term “real world evidence”) are interesting. Given. But they prove nothing, since we don’t know the “truth”. If we are interested in assessing a potential bias, we need simulations. ❝ But, and here is the absolutely salient point, my refutation that shows that the solution is bad might not apply generally. ❝ Perspective: "Linear up, linear down" vs "linear up, log down" in NCA. If someone wants to convince me that "linear up, linear down" is disadvantageous then the approach will be to show me that T/R is biased when using it. The argument is not won by showing me a profile and explaining that the area is upward biased or whatever (an upward bias, or a downward one for that matter, does not necessarily affect T/R unless it affects T and R unequally. It is the latter point that is the deal-breaker). But:
![[image]](https://static.bebac.at/img/Rlogo_15_12.svg) , Julia, Python, C). Or do you think that assessors recalculate NCA in a spreadsheet?❝ I think that patients are well protected if CROs apply healthy rules equally to T and R because then T/R is not expected to be biased. CROs can be trusted to do so, if the rules are defined a priori. Ideally if they are applied blindedly. ❝ I will make a point out of saying that I think there is difference between recording and reporting. I think scenarios exist in which a re-analysis is a potentially useful tool as part of an investigation (they are rare, though), but that should not necessarily affect the reported value. A PK-value for reporting should only be changed if there is proof that the original value is tainted. The proof is not by itself just that the value looks odd. ![[image]](img/uploaded/image271.jpg) I know, the result of the first measurement is – since the EMA’s GL (2011), the FDA’s revised guidance (2018), and the ICH’s M10 (2022).Let’s return to my case study from above giving more details. The barcode system stopped working after three hours and the CRO had no bail-out procedure (e.g., a four eye principle, recording on paper) in place. I suspect there was a sample mix-up between subjects 1 and 2 in the first period in pipetting plasma after centrifugation. We all know the common setup of vials in racks… Was anything documented – which would allow to deal with the problem? No. That’s how true errors occur. The technician couldn’t record an error he was not aware of. Of course samples (the suspected and the samples before and after in profiles) were reanalized4 to exclude an error in bioanalytics. Not only the refrozen ones but also the backups. Results were confirmed. If errors in the clinical phase are rare, in the bioanalytical phase they are extremely rare. According to the GLs, that’s it. No documentation, bad luck. To confirm the mix-up we measured also γ-GT and albumine in the samples (plus pre-dose and post-study). We observed the same pattern, supporting what we saw with the analyte. Question: Do you think that would suffice as “proof”? ❝ I have some degree of sympathy for people who say that investigations almost never give rise to identification of errors, they are for display only. For the same reason I see why some high-quality labs by principle simply do not do re-analyses as part of an investigation. They invest in doing the analysis well from the outset in stead. ❝ Got a weird value, no root cause, and no SOP or healthy rule for re-analysis? Well, it simple then. Report the value, submit the dossier, and let the regulator decide what is fair in that situation. In the meantime, spending the energy on writing an SOP and helping colleagues perform in the lab so that their work does not produce funky values in the future, those are are good investments. 
— Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
Ing. Helmut Schütz