Chromatography Forum

I´m working on a reversed phase HPLC method for assay of the active pharmaceutical ingredient in our drug product.
What is the best way of doing this?
The method is linear in the interval 80-120% of declared content which we have proven in the validation. Is it necessary to perform a five point calibration curve every day of analysis or is it enough to perform a single point calibration? Are there any general recommendations in the European Pharmacopoeia?

The final decision would usually rest with your customer, and what they have specified with regard to regulatory compliance.

The EP has a general monograph called " Substances for pharmaceutical use" that gives some guidance on the tests required.

Under Assay, they say:- " Unless justified and authorised, contents of substances for pharmaceutical use are determined. Suitable methods are used."

The EP monographs use analytical procedures have been validated and approved using EDQM processes, so you can't assume that you can also use single calibration point assays. Your procedure should have been validated using the ICH guidelines.

The EP WWW site also has a free copy of the Techical Guide to the Elaboration of Monographs, which is a useful 69 page document on the EP monographs.
http://www.edqm.eu/medias/fichiers/tech ... nglish.pdf

Bruce Hamilton

the method may well be linear over this range as it will be quite narrow. for example if 100%= 100µg/mL the range is 80 - 120 µg/mL.

If, in the validation, measurement accuracy across the 80 - 120% (of nominal) range using the response factor of the single point calibration STD is demontstrated to be comparable to that obtained using a 5-point curve over the same range, then you can justify the use of a single point method for routine analyses.

Having worked in the CRO field all of my working life, our validations include investigstions of the single and multi point calibration methods with selection of the simplest method, which meets the specified acceptance criteria, used in future (routine) analyses (single point >unweighted regression>weighted regression)

Thank you for your answers, it was very useful for me.