Determination of LOD, LOQ of HPLC method

[zanio]

Hi ,

I have some questions about determination of LOD and LOQ of HPLC method.

I want to determine LOQ and LOD ( for chlorogenic acid, caffeec acid, rutin, myricetin quercetin, naringenin and kaempferol) , but I don't know how to do it.
When I put my blank sample ( H2O) for determination of noise , two big peaks, with 4 mAU , being appeared at 19.5 and 20.3 min. I think , this is not signal of noise. Moreover , many others peaks being appeared from 26 min to 32min. What is this? Is it possibly that signals come from solvents?
I use Methanol ( Chromosolv HPLC grade >99.9 %) for mobile phase A, and 0.5 % acetic acid (for HPLC ) for phase B. Everyday after analysis , I wash my column with pure methanol for 40 min.

Thank you !!!

[JMcK]

Zanio,
It's unclear from your post whether the peaks that you see in your blank are at the retention times of the analytes of interest. If not, these can be considered background and may be ignored. In regards to determination of LOQ and LOD, there are several ways to assess these. I generally create a calibration curve. The LOD is 3 x std deviation of the curve/slope. LOQ is 10 x std dev/slope.

[JGK]

the ICH Q2(R1) document http://www.ich.org/fileadmin/Public_Web ... deline.pdf lists a number of ways of calculating LOQ and LOD (sections 6 and 7).

I used to calculate from the SD of the response of my lowest Cal STD and the slope of the calibration line

LOQ =10 x SD (lowest STD)/slope
LOD =3.3 x SD (lowest STD)/slope

[zanio]

What does standart dev. of curve mean?

The calibration curve for quercetin, for example, is y= 47237 x .

LOD= 3 x SD/ 47237 Where can I get SD of curve from?

Excuse me for stupid questions, but I can't understand it.

[JMcK]

This site shows the formula: http://chemistry.beloit.edu/classes/Che ... quares.pdf (I just googled this to find you a reference).
It's not as hard to calculate as it looks. I have a spreadsheet set up to calculate. I'll send to you if you send me your email. Regards,
John

[zanio]

Thank you!!!!

my e-mail is ani76_1976@abv.bg

[Kristof]

You can calculate slope and y-intercept in Excel without any problem. Some of the other functions are also available e.g. sum of squares.

[lmh]

Note: if you are doing this in Excel, and your calibration curve is linear, you can use the LINEST function. This is an array formula function, which means that after you have entered it in one cell, you must select a block of cells with the LINEST in the top left corner, then click on the place where you can edit the formula, and press ctrl-alt-enter simultaneously. This will populate the block with all the constants and statistical things relating to the fit.

One of those things is the "standard error of the Y-estimate" in Excel-help-file-language. This is the same as the standard deviation of the curve, i.e. what LC-Chemstation refers to as the "residual standard deviation". And that's an important fact too: Many chromatography software systems will do the calculation for you, but in most it seems quite well hidden!

For further information, my two favourite references (apart from ICH) are LC-GC Europe Feb 2009, 22:82-85, and Pharmaceutical Chemistry Journal 38:212-225.

[syedsohail]

This site shows the formula: http://chemistry.beloit.edu/classes/Che ... quares.pdf (I just googled this to find you a reference).
It's not as hard to calculate as it looks. I have a spreadsheet set up to calculate. I'll send to you if you send me your email. Regards,
John

Hello John

Can you please send me the excel sheet for LOD/LOQ
My email is felix_leo99@yahoo.com
Thanks
Sohail