Residual Solvents Validation - QL and DL assessment

[HJM]

We are working on a residual solvents method by GC but have a very stable baseline.
How would you typically perform QL and/or DL analysis for residual solvents. We are looking at determining a practical QL/DL rather than true QL/DL as we would need to go down to very small concentrations to achieve the required signal to noise ratios for each solvent.
What would be an acceptable level to go down to if we are using the ICH specification limits

Any thoughts?

[krickos]

Hi

Well thing is that it depends on the case (development/commersial/which classes of solvents used/going to argue that a solvent is not likely to be present in final substance/product?...).
Today most regulatory agencies tends to accepts that you do not have to set specifications limits for class 1 and 2 solvents below those listed in ICH Q3C, but at the same time some will have problems with a specification of mixed class 3 and 2 solvents where all class 3 solvents have a limit of 5000ppm depending on where in process they are used.

Guidelines talk about likelyhood of solvents being present in final substance/product, however to my knowledge only europe (EMA/EMEA) has defined a requirement if you want to exclude a class 2 solvent from specification, one requirement being able to quantify/limit test down to 10% of ICH Q3C limit to prove that levels do not exceed 10% of limit. None the less I/we have been ableto use europe guide on other markets as well.

So using the default ranges in ICH Q2 for residul solvents is generally in my opinion not so good regarding the lower end. I entirely agree with you that is somewhat academic to determine QL/LOQ-DL/LOD (signal to noise based) as it tends to be quite lower than the desired QL/LOQ (where linearty/precision/accuracy is fit for purpose/specification).

A general range approach could be:
Class 1 solvents, aiming for 50% of ICH limit for lower end of range.
Class 2 solvents, aiming for 50% of ICH limit for lower end of range. If you have a case where you can argue that the solvent is not likely to be present and could be removed from specification go down to 10% of ICH limit.
Class 3 solvents. As indicated can be tricker than expected. If those only are used you can get away with a simple loss of drying test (not more than 0,5%). If you have a mix of classes it may become tricker, if class 3 solvent is used in last step in process then there are no issues with a 5000ppm limit, but other class 3 ones used way earlier in process can be hard to argue for that they should have the same limit. So its harder to give advice here but would aim at 500-1000ppm, seen suppliers who has been forced to drop limits on early used class 3 solvents to that area, sure not so harmful but they do not benefit patients.

Also remeber that limit testing rather than quantification may be an option depending on your case.