Residual Solvent Method Validation - Accuracy

[chrome_dome]

I am working on validating a residual solvent method for an API. I am trying to find out whether or not it is acceptable to satisfy the Accuracy requirement by making multiple injections of standard solutions at three different concentrations. Then by using linear regression determine how much deviation there is between the calculated concentrations vs. the known concentration.

The way I normally perform accuracy is by the ICH Q2(R1) guideline. Take for example a method that calls for dissolving the API sample in water. To determine Accuracy, I will dissolve the API in one of three different standards of various concentrations (i.e. sample spiked with a known amount of solvent.) After the GC analysis is done, I will subtract the amount of solvent that is present in the unspiked sample and will then calculate the percent recovery.

In other words, is it ok to do accuracy testing without ever spiking an API sample? If so, can anyone point to some reference allowing this (ICH, USP, etc)? Thanks.

[Bruce Hamilton]

Given that section 4.2 of ICH Q2 indicates that accuracy should be determined on samples spiked with know amounts of impurities, and that samples of expected solvents should be available to you, I can't see how you could avoid not spiking the sample.

I suspect they would feel aggrieved if you decided that "should be" is not the same as "must be", especially if resources are available. If solvents weren't available, you would still have to verify their absence in the sample by an independent method.

Bruce Hamilton

[chromatographer1]

This question is important. But Chrome_Dome, one has to keep in mind the question that is being posed.

Is the question, "how accurate is the measurement of residual solvents injected with these conditions from a solution of water?",

OR

is the question, "how accurate is the measurement of residual solvents injected with these conditions from an aqueous solution of my API".

The first question can be answered from injections of water containing your residual solvents of interest. The instrument and std preparation accuracy (or inaccuracy) is found in the results you gather when injecting an aqueous solution of solvents.

This measurement of error should be duplicated when injecting an aqueous solution of the API containing the solvents BUT the added inaccuracy (if any) caused by the presence of the API insolution will be added to the inaccuracy of the water solution.

These two measurements need not be the same value and regulatory authorities usually wish to see the latter, not the former.

Just remember "what is the question" before you give an answer. Make certain your answer is applicable to the question asked.

best wishes,

Rod