AS ACCORDING TO THE P METHODS WE INJECT 5 STDS AND ONE SAMPLE INJECTION AND FINDING OUT THE ASSAY EVEN WE FOUND THAT WITHIN THE FIVE STD S
IF WE TAKE OUT THE DIFFERENCE BETWEEN THE LOWEST AND HIGHEST AREA IT IS ALWAYS AROUND 00.5 TO 1 PERCENT AND THAT IS GOING TO GIVE YOU AVERIATION OF ABOUT 1-2 PERCENT IN THE ASSAY DETERMINATION BY EXTERNAL STD METHODS


ALSO WE INJECT THE SAMPLES AND STD IN A SEQUENCE
LIKE 2 STD
TWO INJECTION OF TEST 1
TWO INJECTION OF TEST2 DIFFERENT WT OF SAME SAMPLE
AND TWO INJECTION OF STD
AND WE COMPARE THE T1 AND T2 WITH THE FOUR STDS AND GIVING THE RESULT
WHAT IS THE LOGIC BEHIND DOING ALL THIS TIME CONSUMING THINGS EVENTHOUGH THIS DOES NOT SETS YOU THE RESULT WHICH IS REPRODUCIBLE WITHIN 1PERCENT .

ANY EXPERT IS THERE TO ADVICE ON THIS ?
WHY WE TAKE OUT THE RSD ?
WAHT IS THE LOGIC BEHID THE SEQUENCIL INJECTIONS OF STD AND SAMPLES

The 5 Std injections gives you %RSD for the standards, or the system precision. This is a simple measure of the system's ability to reproducibly inject a known volume (i.e.25 uL injection volume) as specified in the monograph or method. Typically this is the heart of "system suitability". You have to prove the system is suitable....imagine a Stability Laboratory telling an FDA inspector, "Hey, trust me...it's a good system!"

The standards above and below the sample injections provide you with a "bracket". Sometimes, the overall %RSD of ALL standard injections in the run is calculated as part of the system suitability. If this fails, you can "ignore" the brackets, beginning at the end of the run, until your overall %RSD passes the suitability criteria....thus, saving part of your run.

As for using duplicate injections, hey...I'd rather report a mean than a single value. It's much easier to justify OOL results if there are duplicate values on the same sample which agree within the confidence interval. This is just another way of proving the system and sample are viable during the run.

Remember, the FDA is a "Prove it" organization. The analytical chemists job is the prove, beyond all reasonable doubt, that methods work...reproducibly...at any location on the face of the planet! Monographs have already been proven. (Supposedly....actually, I've come across some pretty shabby monographs.)

Hope that helps.

I notice that the US Pharmacopoeia states that 5 injections of a standard are to be made if the limit is %RSD less than 2, and 6 injections are to be made if the limit is greater than 2. I find this logic odd since the number of determinations should not affect the %RSD. (each std area has as much chance as being above or below the mean ie it is non-biased)It calls into question the practise of using all injections throughout the run to calculate the %RSD as part of system suitability. Any thoughts?

itseems adquate we are loosing our time and money behind this .is it so?

Margaret,

Standard deviation is a function of degrees of freedom. i.e. although two means may be equal, their variances may not be because of N value. To illustrate, take set "A" of numbers 4,5,4,5,4,5 and set "B": 4,5,4,5,4,5,4,5,4,5.

set A
N=5, DOF=4
mean=4.5
stddev=0.548
%RSD=12.17

set B
N=10, DOF=9
mean=4.5
stddev=0.53
%RSD=11.71

It only seems strange to me that a standard would be written to only require ONE more injection to yield a "more precise" determination. In real QC life, not only injections of a standard throughout a sequence should be accounted for, but day to day and operator to operator. Tedious, but in QC labs, it's done.

Spencer

is there any guide line for doing assay by external std. in which we can have aspecific things that for which method we inject in sequence or for which method we can inject just there std and one sample

why not five stds and three samples