Re: Tiny peak cGMP question
Posted: Tue Aug 16, 2011 3:35 pm
No, the fragrance peaks are too small to influence the quantitation of the active peak. Our QA agrees with you that such peaks are negligible, but my supervisor says "a peak is a peak, no matter how small".
We've discussed dilution of samples and standards so such peaks are not larger than 3x the noise, but I feel that doesn't help the science either, and would require dilution of each sample (another step for QC to potential screw up, plus more solvent waste). We've also discussed using a less sensitive wavelength for this, but again, doesn't help the science, just a way to "skirt" the regulations. In the one instance we can saponify the fragrance component and then that area is flat, but that's also an extra step (plus neutralization) for QC, and would also require some validation, tech transfer, etc.
We also have same preparation but different set of conditions that had been previously validated but that requires a column change and solvent change (uses THF), not really something we want to throw onto QC, like running one "flavor" by a different set of conditions.
Our test methods do not state anything about peaks that are below a certain threshold need not be integrated or determined either.
We've discussed dilution of samples and standards so such peaks are not larger than 3x the noise, but I feel that doesn't help the science either, and would require dilution of each sample (another step for QC to potential screw up, plus more solvent waste). We've also discussed using a less sensitive wavelength for this, but again, doesn't help the science, just a way to "skirt" the regulations. In the one instance we can saponify the fragrance component and then that area is flat, but that's also an extra step (plus neutralization) for QC, and would also require some validation, tech transfer, etc.
We also have same preparation but different set of conditions that had been previously validated but that requires a column change and solvent change (uses THF), not really something we want to throw onto QC, like running one "flavor" by a different set of conditions.
Our test methods do not state anything about peaks that are below a certain threshold need not be integrated or determined either.