5-point Calibration curve

[bartjoosen]

Variance should be "equal" through the whole calibration line. But as yw3142 wrote, if the error is the same, the %error for the 20µg/L sample is far more than for the 800µg/L sample. So I would suggest to use the most accurate standard as possible when using single point calibration.

But before using single point, it should be pointed out that single point is acceptable, during a validation.


Bart

[guenzia]

I think some clarity should be made in this thread, since 2 totally different areas, and ways of calibrating, are overlapping here.

On one side there are posts dealing with typical QC/QA issues (range close to 100%, impurities etc.) which are regulated by a set of guidelines issued by ICH (I guess, it is not my field).

On the other hand there are posts like those of LewisC who speaks about analysis of substances in biological media (plasma, urine etc) which are performed according to the FDA guidance for industry on bioanalytical method validation (http://www.fda.gov/cder/guidance/4252fnl.pdf), which states that multi-level calibration samples (75%, or minimum 6, valid) and QC samples at minimum 3 levels should be used during routine analysis.

In this respect, I would suggest LewisC to read again the above guidance, since what he reported in his posts is absolutely not in line with the commonly accepted practices in the field and casts some doubts on the precision and accuracy of the concentrations he finds.

[LewisC]

Guenzia,

All our assays are fully validated according to FDA guidelines, I was just wondering in what respect you think these guidelines have not been followed? We use a seven point calibration to cover the expected range of results, with three independently prepared QC samples at 'low', 'medium' and 'high' concentrations. The back calculated concentration of the high QC, when using our lowest calibration standard as a single calibrator, is not within ±15%, and so cannot be accepted for routine use. If anything, the FDA guidelines completely rule out the use of single calibrators.

LewisC

[bartjoosen]

LewisC,

you wrote in a previous message about recovery ranging 60%-110%.
Is this accepted by FDA?

Ofcourse I'm not in your place, so I think from my place, which is definetely another working area, so don't take my messages personal

Bart

[LewisC]

Bart,

The concentration I found for the high IQC using single-point calibration (when the single calibrant was near LLOQ) did range from 60-110% of the nominal value. This is not accepted by FDA, and so cannot be accepted. Using multipoint calibration, all three IQC concentrations fall within FDA recommended values for routine use.

I think it really is a case of different requirements for different purposes, but it is really interesting to hear different perspectives on this issue!

LewisC