Back to the Top
Dear all,
I have a question related to unknown sample analysis in HPLC.
I have validated bioanalytical method for X compound.After this validation i got mean linearity curve equation Y=m(sd) X+c(sd) from eight different calibration curves.My question is can i use this equation for unknown sample analysis rather than using the calibration curve obtained on the day of analysis (fresh CC). Fresh calibration curve may not show (not a true representative curve ) the real value but on the other hand mean calibration cure obtained from eight different linearities will be more accurate and close to real value?
Waiting for your expert comment
-- Regards,
Rahul Vats
BITS-Pilani, Hyderabad Campus,
Back to the Top
Dear Rahul
As you know, most of the the bioanalytical extraction procedures show high magnitude of variability every day and also from subject-to-subject. This may be due to variability in extraction procedure and most importantly due to matrix effect from subject-to-subject. These variabilities often leads to changes in recovery there by affecting the slope and intercept of calibration curve.
Also the biomatrix used in validation experiment may not be representative of the subject plasma, and hence it is always advised to prepare a fresh calibration curve every day (or for every analytical batch) for unknown samples. This fact is also one of the reason for using particular subject's blank biomatrix for preparing CC & QC samples for analyzing same subject's unknown samples. When all the CC, QC and unknown samples of one subject is processed together, the variability and matrix effect is reduced and there will be a high possibility for obtaining accurate results. In case of preclinical studies, it is not possible to obtain biomatrix for each animal. In this case, you may use a pooled biomatrix for preparation of CC&QC samples.
In your case, the mean calibration curve was calculated. However, instead of mean calibration curve, individual calibration curves along with regression values must be reported. To account for variability, you may calculate confidence intervals for slope and intercept.
Hope this clarifies your doubt.
Best Regards
Sivacharan Kollipara Scientist Pharmaceutical & Analytical Development Novartis Healthcare Pvt Ltd Hyderabad India
Back to the Top
Dear Rahul
It is better to use the fresh calibration curve. From eight calibration curve you can find the slope +ve or - ve it indicates the variation in the linearity. If any error is there in the linearity on a day of analysis of samples (the same compre to previous day linearity ) there will not be any change in the concentration. You will get the real concentration. In case of the mean calibration curve the comparison of concentration between group can be done but you may not get the real concentration. NIL
Back to the Top
Dear Rahul,
Logically, calibration curves obtained on the day of analysis should not be statistically (significantly) different form the CC obtained form the 8 different analysis if the tested concentrations of compound X are the same and if tests are realized under exactly the same operational parameters. So from a technical point of view, the use of mean linearity curve equation should be possible. However, I believe that it would be more accurate to use a calibration curve obtained from fresh standards prepared in the same operational conditions that the unknown samples
Hope this will help
Regards
ALHANOUT Kamel
Back to the Top
The following message was posted to: PharmPK
Rahul,
I like your question. The answer depends on your analytical method. I
believe everyone is talking about LCMSMS assays. These suffer from many
poorly controlled sources of variation or drift - voltages, analyser
cleanliness, gas pressures, source temperatures, column age, solvent
composition, etc. On the other hand, some other analytical techniques
are so robust and precise that the main source of error is the pipetting
used to make up the calibration standards. In this situation, it will be
generally more accurate to use a consensus calibration line based on
many standards rather than a calibration run on the same day but on a
relatively small number of standards. If your system is very robust, you
could be in this situation, and well-designed QCs will tell you if you
are in specification or not in any particular run. However, I do not
think that you could win this argument for an LCMSMS assay.
Regards.
Ted Parton
Director, RDMPK, UCB, Slough
Back to the Top
The following message was posted to: PharmPK
Dear Rahul,
Based on following statements as per Guidance for Industry; Bioanalytical Method Validation Doc (May 2001):
* Calibration curve is a relationship between instrument response and known concentration of anayte.
* A calibration curve should be generated for each analyte to assay samples in each analytical run and should be used to calculate the concentration of the analyte in the unknown samples in the run.
* An analytical run can consist of QC samples, calibration standards.
We can conclude that for each analytical run, we need to establish a calibration curve. In case, you doubt your calibration curve (as raised by your concern), you should reanalyze/repeat the complete run/batch.
Hope it helps.
Kuldeep Sharma Jubilant Biosys
Back to the Top
Dear Rahual,
No doubt it is always recommended to use fresh calibration curve due to changes in various instrumental, extraction and matrix parameters. It is better to use internal standard that can compensate variations in different experimntal and instrumental parameters.
Dr Zafar
Want to post a follow-up message on this topic?
If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Calibration curve" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)