Hplc Related Substances Problem (Accuracy, Linearity & R

[seamoro]

Validating a related Substances method.We are having problems with low recovery of related impurties(Standards). Linearity (0.5%,1%,2%,4% & 5% nominal) of the related impurties is fine but our recovery for example 0.5% standard imputie is 0.1% and this recovery is representaive of all the concentrations. Different impurties have different (Low) recoveries. Do we need to show recoveries above a certain level...for example 90% ? Could we calculate a relative response factor to correct the recovery mathematically?....or does anybody have any other approaches to solving the problem....besides developing a new method!!!!!!

[Peter Apps]

Before you redevelop the method check your calculations and dilutions - a uniform recovery of 0.1% is 1/1000, and I suspect that milli has been mixed with micro somewhere.

Peter

[seamoro]

it is not a calculation error

[Uwe Neue]

Adsorption on vials or similar?

[Bruce Hamilton]

Are the recovery calculations based on simple external stds or spiking into a surrogate zero sample mixture. I assume spiking?.

What is the difference in response factor for simple external stds versus spiking?.

I believe it would be very difficult to justify recoveries below 80% for significant impurities, however lower recoveries may be acceptable for trace impuries, but you need to check what your quality group and regulators will permit.

Several possible causes, adsorption, as suggested by Uwe, degradation or reactions in sample solutions ( if spiking ) - are solutions stable over time?.

I'd be inclined to start by comparing response factors of freshly prepared sample solutions and external standards over time, - ensure suitable pH control of sample solvent.

Please keep having fun,

Bruce Hamilton