Bioequivalence and Bioavailability Forum

Dear Shuja & all,

❝ I think we are heading towards some open ended questions with still no conclusion, […]

Well – maybe there is none?!

❝ And Yes I am still waiting for Helmut views on this whole topic !

I changed my name above from big to small because I left the lab many years ago.

Some thoughts:
Bulk vs. freshly prepared calibrators/QCs.
No specific recommendation in the FDA’s guidance, the Crystal City III conference report, and the EMA’s draft GL.
Freshly prepared calibration curves are only mentioned in the respective sections dealing with stability testing in method validation. But: in my experience inspectors prefer freshly prepared calibration curves over bulk samples (a large volume is spiked, split into vials and frozen together with QCs and study samples). As Ohlbe once mentioned:

❝ Though inspectors are paranoiacs and won't trust anything unless it is written, an inspection remains a confidence-building process.

We used bulk calibrators for many years for the following reason: Spiking small plasma volumes (with a very small volume of analyte solution) is inherently variable. Many CROs even prefer to prepare standards (and QCs) gravimetrical instead of volumetrical.
Now for a nasty example:
You have validated storage at –20 ℃ and store bulk calibrators and QCs together with study samples. After the last analyst left the lab (and checked the temperature of the freezer) you have a power outage. Temperature rises to –10 ℃ and returns to normal during the night. The analyte starts to degrade at >–15 ℃. If you have an online-warning system connected to your freezer, this must be battery-powered! Let’s assume you don’t have that. In the morning the analyst checks the temperature of the freezer – everything is fine. Responses in further batches will be lower than in previous ones. You will see lower CC slopes, but runs are not rejected because QCs are affected as well. If you use freshly prepared CCs instead, runs will be rejected (degraded QCs will fail). On the other hand bulk CCs are less variable and in BE – where we are interested in intra-individual comparisons – the ratio of test/reference will not be affected anyhow.
At the EUFEPS/EPF-workshop last April Kamal Midha and myself made an argument in this direction, but it was obvious that regulators didn’t like that at all.

Back-up sample vs. thawed original.
I would follow Ohlbe’s recommendations.

Storage of samples after the end of the study.
I partly disagree with Dan’s post. In my CRO we followed this procedure:

• State in the contract that all samples will be stored for ½ year after acceptance of the analytical report by the sponsor.
  
• After that the sponsor was notified that samples will be disposed or kept longer if required. In this case we provided a cost estimate (–80 ℃ freezers suck up energy like crazy) and also issued a warning if the long-term stability will be exhausted.
  
• In any case we kept all samples of one subject for infinite time to prove that the study was really performed – and not faked by means of a chromatogram-generator. We also kept one set of the original calibrators and QCs. For some drugs we had long-term stability data of more than five years; so re-analysis would have been possible. Even if long-term stability is exceeded, you would get ‘some’ results. Furthermore, by comparing the original CCs/QCs with freshly prepared ones you get an idea about stability.