Bioequivalence and Bioavailability Forum • EMA Guideline on Bioanalytical Method Validation

sevim
●

2011-09-19 11:04
(5399 d 06:17 ago)

Posting: # 7369
Views: 3,452

EMA Guideline on Bioanalytical Method Validation [Regulatives / Guidelines]

Post reply

Dear all,

I have some questions about the new guideline and I will be very appreciated if you can help me.

In the chapter of Method Validation- Calibration Curve (page 6), it is written that "Each calibration standard can be analysed in replicate." Based on this sentence, the criteria for the acceptance of the calibration curve changes. By the way, if one of the calibration standards(except LOQ and ULOQ) does not comply the criteria I can omit it. What about if the other value of the same calibration standard also does not comply the criteria, do you think that I can also reject this value? And how many times can I apply this process? (There are eight calibration standards). I need at least six valid calibration standards, so I understand that I can reject two of the invalid calibration standards from the curve. Will the batch be valid, if only one of
the LOQ or ULOQ values are in the range or should both of the data be in the range?
How can be the sequence of this replicate calibration standards? Should the same calibration standards be applied successively or can they be analysed separately in the batch?
If the replicates of the calibration standard comply the criteria, should I take the average of these two values and draw the calibration curve according to this average values?

Thanks for your interest and informing.
Best Regards,
Sevim.

Ohlbe
★★★

France,
2011-09-19 22:33
(5398 d 18:49 ago)

@ sevim
Posting: # 7373
Views: 2,981

EMA Guideline on Bioanalytical Method Validation

Post reply

Dear Sevim,

❝ By the way, if one of the calibration standards(except LOQ and ULOQ) does not comply the criteria I can omit it. What about if the other value of the same calibration standard also does not comply the criteria, do you think that I can also reject this value?

No. See the last paragraph in page 6:
In case replicates are used, the criteria [...] should also be fulfilled for at least 50% of the calibration standards tested per concentration level.

❝ And how many times can I apply this process? (There are eight calibration standards). I need at least six valid calibration standards, so I understand that I can reject two of the invalid calibration standards from the curve.

Same paragraph:
At least 75% of the calibration standards, with a minimum of six calibration standard levels, must fulfil this criterion.

If you have 8 samples as singlets you can reject 2 of 8. If you run them in duplicate i.e. 16 results: you can reject 4 of 16, at 4 different levels of concentration.

❝ Will the batch be valid, if only one of the LOQ or ULOQ values are in the range

Yes. Still the same paragraph:
In case all replicates of the LLOQ or the ULOQ calibration standard are rejected then the batch should be rejected from the validation

Meaning that if one of the 2 replicates is valid, the run can be accepted.

❝ 2. How can be the sequence of this replicate calibration standards? Should the same calibration standards be applied successively or can they be analysed separately in the batch?

Usually one set of standard samples at the beginning of the run, one set at the end. The idea is to monitor (and try and compensate for) any drift in response over time. Personally if there is a drift that needs to be compensated I prefer to see shorter runs...

❝ 3. If the replicates of the calibration standard comply the criteria, should I take the average of these two values and draw the calibration curve according to this average values?

Usually both replicates are used, without any average.

Regard
Ohlbe

—
Regards
Ohlbe