Carry over - qualification

[greg07]

Hello, can I ask for an explanation how I should understand an entry from Annex1:
VOLUME PRECISION AND CARRY-OVER
Solutions: Solvent A: methanol/ water R, 60:40.
Settings: Column: Lichrospher 100 RP8, 5 µm, 125 x 4 mm, without precolumn Mobile,
flow rate: 1.0 ml/min, λ= 254 nm, injection volume=20 µl
Reference solution (a): dissolve 15.0 mg methyl-, ethyl-, and propylparabene in 100.0 ml of solvent A.
Reference solution (b): Dilute 1.0 ml of reference solution (a) to 10.0 ml of solvent A.
Reference solution (c): Dilute 1.0 ml of reference solution (b) to 100.0 ml of solvent A.
Injection scheme:
1. 6x reference solution (b)
2. 1x reference solution (a)
3. 1x solvent A (blank injection 1)
4. 1x reference solution (b)
5. 1x solvent A (blank injection 2)
6. 1x reference solution (c)
Limits: Repeatability of peak areas: The relative standard deviation of the peak areas of all peaks in the chromatogram obtained with the reference solution (b) should be ≤ 1.0 %.
Carry-over:
The percentage of the peak area corresponding to propylparabene in the blank injection 1 does not exceed 0.5% of 10 times the peak area of the propylparabene peak in the chromatogram obtained with the reference solution (b) injected after the blank injection.

The percentage of the peak area corresponding to propylparabene in the reference solution (c) is 0.9 – 1.1% of the peak area of the propylparabene peak in the chromatogram obtained with the reference solution (b) injected after the blank injection.


Why on the end are 2 conditions? Sufficient to meet one or both?

BR
greg07

[HPLCaddict]

You should meet both, of course! These are two completely different tests. The first one (injection of a blank solution after a concentrated sample) is the actual carry-over test. The second one (injection of a solution with nominal concentration of 1%) is a different thing - I'd suppose this is a kind of detector linearity test.

[greg07]

That is, strictly speaking, only the first test determines the level of carry over - only this interests me? I can skip the second test? Linearity test I do in another test.

[itspip]

I agree with HPLC Addict, you should be able to meet both. The first test is your carryover. The second is a different type of repeat-ability, i.e., can the system achieve the same peak after repeat injections and after blank injections.

[HPLCaddict]

Well, it all depends on what you're trying to do (and your SOPs, of course ). If you want to perform a qualification of your HPLC according to that document, you should stick to it and perform all the tests described and fulfill all the specifications.
If you only want to perform a carry-over test, then just perform the carry-over test. In that case, you'll only need reference solution (a) - blank (solvent) - reference solution (b)

[greg07]

Thank you for all answers, . Finally I ask, the following two sentences are the same:

1) The percentage of the peak area corresponding to propylparabene in the blank injection 1 does not exceed 0.5% of 10 times the peak area of the propylparabene peak in the chromatogram obtained with the reference solution (b) injected after the blank injection.

2) The percentage of the peak area corresponding to propylparabene in the blank injection 1 does not exceed 5% of the peak area of the propylparabene peak in the chromatogram obtained with the reference solution (b) injected after the blank injection.

[lmh]

I think you are right that those two sentences are identical. I think that this sort of "change" in a protocol isn't really a change, because you are carrying out exactly the same tests, looking for exactly the same limits, and the sentences are logically identical. The difference between them is purely an indication of the motive: it reflects the fact that standard (b) is a 1/10th dilution of standard (a), and although you're using (b) to create a (one-point) calibration curve (because linearity isn't guaranteed up to the level of "a"), the carry-over is naturally expressed in terms of "a", the injection before the carry-over blank. You measured "b", but what matters is c/a, so the wording calculates first "a", and then c/a...

On your first question, there are two thought-processes that are necessary. The first is the question: "Am I carrying out this SOP, or developing my own approach to measuring what I believe the SOP aimed to measure?" HPLCAddict is right that if you are carrying out the SOP then you must stand on your head and whistle your national anthem while eating a Pizza, if that is in the SOP, and even if you don't believe it will affect the results. That's the sad truth of SOPs.

If you're free to develop your own, then before abandoning the second part of the test, it's worth thinking why it is there. It's actually there for a very good reason, and you absolutely should carry out both parts of the test. This is the reason:
Injection "a" is so that you have something to carry over.
The next blank is to assess the carry-over
Injection "b" is so that you have a calibration "curve" on which to base your measured carry-over, given that "a" is too concentrated
The next blank is to convince yourself that "b" itself isn't now carried over...
And the final standard "c" is a vital positive control to convince yourself that if you had a carry-over of only 0.1%, you would be able to detect it, with reasonable accuracy.

The problem with missing out this second test is that if your limit of detection is actually (for some weird reason) too high to see a 0.5% carry-over then your "pass" in the first test is meaningless. If you're lucky you could guess this from the signal-strength of the run for standard "b", but if you're unlucky you might have one of those methods where binding-sites in the system titrate out analyte at low concentration, so the calibration curve is linear but crosses the y-axis at a negative value. In this case the "b" run looks fine, but you will never see very dilute standards.

[greg07]

Imh - thank you, I think you right to the point. Let me summarize, do I good understand.
Injections:
1. Check that you have at all a general repeatability - if "not" finish tests now,

2,3 Appropriate measuring of carry over

4,5-cleaning,6 Prove that your HPLC can good measure concentration 5%(x20 of(a)(or less) by measuring concentration in a wider range: (b)10%(x10 of (a)) and (c)0,1%(x1000 of (a).

Right?

[lmh]

yes, except that the protocol doesn't expect you to look at the peak area of the "a" standard at all. This is just a wildly over-concentrated injection with the aim of maximising the carry-over peak so you can see it more accurately (or be sure it's very small indeed if you can't see it).

At the moment I'm trying to sort out in my head whether it's really necessary to inject an over-concentrated sample for a carry-over test, or whether an undetectable carry-over following an injection at the top end of the calibration curve already indicates that carry-over can't be having a significant effect on the results. I'm not sure.