Chromatography Forum

Hi,
my work area is determination of residual solvents in drug substances. My question is: what size of the unknown peaks I obliged to identify and quantify in my report? RS of ICH Class 1 are not presented in our drugs.

Hi

Well there is not a straigh forward answer except that class 2 solvents used in the last manufacturing step has to be quantified and total solvents typically has to be within 0,5%.

You typically build a case on what solvents are likely to be present in the drug substance. The data needed in EU to rule out class 2 solvents from specification can be found here:
http://www.emea.europa.eu/pdfs/human/qwp/045003en.pdf
Note that it may not work outside EU but it is a good start point I would say.


I would recommend building a table with all used solvents in process, last step in manufacturing where it is used, then add data to it (from 6 consecutive pilot scale batches or 3 consecutive industrial scale batches of the suitable intermediate or the final active substance.Taken from EU guidance above).
You can also discuss around which process steps that removes solvents.

So intially you would typically screen for all solvents, then build your case based on data and process steps to propose a final specification for those solvents that you consider likely to be present.

Hi,

There is another problem. All solvents used in a process are described in the product specification. Our substances (e.g. dry extracts) are often of natural origin and therefore they may contain various volatile compounds in addition to those listed in the specification (essential oils, alkohols, ketones, alkanes etc.). We have to prove that the peaks of these compounds are not solvents used in other processes or in cleaning of the manufacturing equipment or some other contaminations. Compositions and amounts of residues of natural origin vary during the year’s seasons. We have to know from which size the unknown peaks are necessary to be identified and quantified in the report, because evaluation e.g. of all peaks over S/N = 10 in all samples would be killing for our QC department, and simultaneously we have to comply with requirements of FDA or customer’s audits.