Chromatography Forum

Hello. I would like your opinion.
I have to evaluate a new method of cleaning validation. My 5 standard readings about 4339 (mAU) Very small.
According to you what I have to set the RSD to check the system suitability?
<2%?
<10%
<15?
.....

Thank you

The analytical method should have a %RSD of 2.0% since you are making 5 injections (5.0% for 6 injections) as required by USP <621>. You can try increasing your injection volume since chromatography is mass related (and therefore the peak area increases).

The analytical method should have a %RSD of 2.0% since you are making 5 injections (5.0% for 6 injections) as required by USP <621>. You can try increasing your injection volume since chromatography is mass related (and therefore the peak area increases).

Right but.. USP said that " This
requirement does not apply to tests for related substances".

A cleaning validation is comparable to a Assay or a purity?

Sorry...
This is the USP:



For 5 inj is 0.73?

Don't worry about that table you're looking at from USP <621> with respect to your system suitability requirements. Take a quick look at the paragraph above it:

"Replicate injections of a standard preparation or other standard solutions are compared to ascertain whether requirements for precision are met. Unless otherwise specified in the individual monograph, data from five replicate injections of the analyte are used to calculate the relative standard deviation, %RSD, if the requirement is 2.0% or less; data from six replicate injections
are used if the relative standard deviation requirement is more than 2.0%."

Again, this is specific guidance for USP monographs. Generally speaking you need to let your data drive this discussion. Justifications driven by data are fairly difficult to argue against. I presume your standard area of ~4339 is low because the standard is prepared at a fairly low concentration.

If your HPLC is well maintained, I don't see why a 2% RSD couldn't be achieved even at that area count, but again, let your data drive that discussion. If necessary, you may be able to justify a 3% RSD. Let's try a simple study: how many times have you run this method? What have been your RSDs to date? Have you tried this across different HPLCs (if available)?

Mehtod is new.
Actually i have rsd = 4,41%.


I did a second test and was released = 4.52%

The concentration is very very low because the result of cleaning must be ZERO. (At least that's my idea).

It is not my method. but the method of one of my clients. I suffer as it is.

If the method belongs to one of your clients, they should have a system suitability requirement for %RSD already in place, no? Can't you just use that requirement?

With two sets of data showing %RSDs of 4.4% and 4.5%, you clearly won't meet a specification of 2.0% RSD.

Depends on the concentration of the standard. If it is 100 ppm then 2.0 %RSD is easy. 100 ppb is a different story.

Is your carryover acceptance criteria based on the molecules toxicity?

Supplier imposes a limit of 2%. For me, at these concentrations, it is too low.
Preparation std= 20mg to 100ml. Than 1ml to 100.
Volume inj 10

The easiest thing to do is increase your injection volume to 100 uL. Your HPLC should be able to perform the %RSD requirement.

Be sure your wavelength detector is on peak and your injector is injecting the right amount. This may require calibrating the wavelength detector with a caffeine solution. The injector can be calibrated by injecting water multiple times from a tared vial.

Dear etaxene


In the case of cleaning validation by hplc I think the first thing you have to find is the LOD and LOQ of the drug your are working with. In this cases involving a few µg/ml to expect RSD of less than 2 % for 5 injections is very rare!

Best wishes

Ferenando

Hi Fernando.
I have LOQ = 0.3 ug/ml

If you are doing replicate injections of the same standard solution then the only sources of variation are injection volume and peak integration.

The variability in injection volume is independent of solution concentration and peak size, but can be changed by injection technique; full loop vs partial fill for example.

Relative variation in integration gets bigger as peaks get smaller, ideally you want to reduce baseline noise with e.g. cleaner mobile phase, cleaner cell windows, new lamp or whatever. Failing that change the integration parameters to make peak detection more consistent. Judicious use of smoothing can help a lot.

By definition the rsd at the LOQ is 10%, so you will never get a 2% repeatability at levels close to or below the LOQ.

Peter

By definition the rsd at the LOQ is 10%, so you will never get a 2% repeatability at levels close to or below the LOQ.

Peter

Sorry Peter but i don't understand.
what do you mean?

just for completeness I send the chromatogram of the standard