By Anonymous on Wednesday, February 18, 2004 - 12:40 am:
I am analysing a sample at a low (near LOQ) concentration and the peak of interest is surrounded on either sides by 2 other peaks, of comparable size. The peaks are well resolved.
A blank injection showed the same surrounding peaks (of course the peak of interest was missing).
I am using a gradient so the baseline has a slight slope, but is smooth.
I need to have a signal to noise ratio of 9 (as specified in my method). How to determine the SNR ? Visual determination is subject to interpretation because of the peaks on either sides : if I consider the chromatogram very close to the peak of interest, I just have the smooth baseline, but if I consider a bit further, the surrounding peaks are there and that reduces the SNR. Are there any guidelines (ICH, FDA or other) as to how far around the peak to look to determine the SNR ?
Regards
By Jan on Wednesday, February 18, 2004 - 06:38 am:
European Pharmacopoeia : measure the noise level in your blank in a timeframe equal to 20 times the peak width at half heigth, distributed evenly left and right of where your peak would be.
Jan
By Anonymous on Wednesday, February 18, 2004 - 10:59 am:
Thanks Jan. That is exactly the kind of information I was looking for.
By dipendra on Saturday, February 21, 2004 - 09:23 pm:
I have similar problem: getting an unknown peak at the rt of drug. The peak of blank is less than peak of LLOQ. How much I can tolerate the background noise in blank sample? or I must have smooth baseline at that RT of drug, for blank? I am using LC-MS/MS, for glimepiride.
By A.Mouse on Sunday, February 22, 2004 - 10:24 am:
What is the accuracy that your assay requires? What is the nature of your assay? Blood sample? Urine sample?
By Dipendra on Sunday, February 22, 2004 - 09:43 pm:
I nead the accuracy required for the bioequivalence testing, and I am using plasma samples.
By A.Mouse on Monday, February 23, 2004 - 06:23 pm:
There are two answers to this. If the junk peak is indeed within the noise of the instrument, then there is no problem. However, you must make sure with multiple blank samples from multiple plasma blanks are all within the noise. The second answer is that if I have a blank with MS/MS detection, I am sure doing a lousy job with chromatography and sample preparation, and I am likely to encounter other problems like ion suppression as well. Consequence: I should try to improve the chromatography or the sample preparation.