Has anyone developed & validated NDSRI in Ritonavir tablet? The limit for this impurity is 15 ppb (AI 18ppb and MDD is 1200mg). We could achieve the accuracy for same impurity in the API by LCMS but the following challenges were observed while analyzing drug product (tablet) -
Placebo interference is observed.
Placebo is slightly soluble in most of the organic solvents used during development. The dissolved placebo interferes in quantitation.
The chromatographic separation between placebo and impurity is challenging (NDSRI is polar which is co-eluting with placebo)
Due to stringent limit, high load of formulation is injected which causes contamination of mass source.
The % accuracy of spiked impurity in placebo is 0% to 20%.
We attempted to develop the method on Sciex 5500, Thermo HRMS and Shimadzu 8060NX but were not successful due to above reasons.
If possible I would check when the API eluates from the column. We use a UV detector for that, which sits between the LC and the MS. If those peaks (API and Nitroso) are not separated, I would try to use a different column or gradient.
You could also make the signal suppression visible, by using a syringe pump. Inject placebo and over a T-adapter you could infuse Nitroso via the syringe pump. You would see what´s really going on.
Since I don´t know your method, but have you tried liquid/liquid or SPE clean up in some way or another ?
Q you can set flow rate to 1.0 ml/min. And then by applying T-joint with setting 0.2 ml/min flow rate to be entered in MS-ION source. Hence the flow will be devided in to 5 parts. And placebo will suppress lesser on ionisation impact. Temperature and gas parameters should be also optimised as per given flow rate . Hence You can reliably quantite your analyte. But you have to open and clean ion source on daily basis .
Hi Phil, Please note that the chromatographic separation between API and NDSRI is not a concern but similarity in the polarity between placebo and the impurity poses the challenge in getting the desired accuracy and adopting SPE system. Even liquid-liquid extraction was not successful.
I would like to know whether any of the Ritonavir drug product manufacturer also encountered similar issues and how did they resolved it ?(Ritonavir table is in market for several years and I believe that since there are many generic manufactures so obviously someone must have attempted detecting the NDRSRI)
Yes Viraj we are using splitter or T join and hence effective flow rate entering the mass spec is about 40% less. Temperature and gas parameters are also optimized while tuning.
You could try to make single ingredient “Worst Case Placebo” and perform accuracy again. So then you will see which compound is really making trouble.
Maybe this will help to get a better understanding and optimize extraction.
A different approach would be after successfully identifying the problematic excipient to prepare a reference solution containing this excipient. With this approach matrix effects would be compensated.
“The % accuracy of spiked impurity in placebo is 0% to 20%.” → Do you find residues of NDSRI in the tablets ? If not maybe stability could be a problem. We already observed quite rapid decomposition in different solvent systems.
