Home › Forums › General Discussion › Situation Room: Chromatographically Contaminated Cosmeceutical › Reply To: Situation Room: Chromatographically Contaminated Cosmeceutical
Reporting a new unknown impurity at the specification limit after just 1 month in the accelerated stability chamber (40℃/75%RH) would be a big worry for any company and it’s particularly worrying when it turns up as a complete surprise in a registration stability program. If a new degradation product truly developed from “zero” to a level of 0.2% after just 1 month at 40℃/75%RH, it is highly likely that it will also exceed the specification limit after 12 months at 25℃/60%RH so registration strategy option 2 may not be a good route to follow. However, there could be other factors to consider here:
• As soon as the atypical result is observed an investigation should be initiated. The first objective would be to establish whether there had been a lab error followed by some additional steps to establish that the atypical result can be confirmed. (We don’t tend to call atypical results with accelerated stability samples Out of Specification (OOS) as they are not being stored as per label conditions.)
o I will assume that this was all done correctly in which case we have a confirmed Significant Change as per requirements of ICH Guideline Q1E (Evaluation of Stability Data)
o This investigation should also include review of T=0 data to establish whether this impurity was observed during initial testing but also making special note of analytical method parameters Detection Limit (DL) and Quantitation Limit (QL) as it is very important to get a handle on its rate of formation. Any subsequent product shelf life calculation will lead to quite different results depending on that initial value.
o We should also plan to commence testing of samples being held under intermediate conditions (30℃/65%RH) which is the fallback when significant changes are observed in samples held under accelerated conditions. Generally our protocols do not consider testing of real time nor intermediate condition samples before the 3 month time point although it may be helpful in this case. Since that is likely to represent a protocol deviation, it is recommended to document this in the atypical result investigation provided this is conducted in a quality system such as TrackWise or apply a Planned Deviation. If we can achieve levels of this impurity below 0.2% up to 12 months, this will still allow us to apply some extrapolation of shelf life at time of NDA filing.
• Since this is an unknown impurity, it is probably being quantified either by peak area % or external standardization against a reference standard of the active molecule. Both of these approaches make an assumption that the responses of the active and impurity components to the method of detection are equal. So if this method uses a UV detector (which is the usual situation), it is being assumed that both components respond equally to the wavelength being employed. Responses of different molecules to UV detection can vary by orders of magnitude so it is quite possible that we are vastly overestimating the content of the impurity. Application of a photodiode array detector (PDA) set to collect full UV spectral scans of all components will provide some help with this. It won’t provide definite relative response factors but will provide some indications. Of course, this could also provide even worse news as it is equally possible that we are underestimating the impurity!
• The situation indicates that the company believes that this impurity is “…most likely a product of the astringent reacting with one of the excipients.” So is this really a complete surprise? Perhaps it was observed during an earlier experimental stability study or during an excipient compatibility study. If it really is known which excipient is causing the problem, surely the best solution would be to replace that excipient with something else that carries out the same function but does not interact with the active molecule. This might add a few months to registration but if it would have been better if Company responded to it when it was first seen in experimental samples.
• If this really is the first time this impurity is being observed and it was not seen in earlier experimental stability studies, then it may just be a one off. This is supposed to be the first batch of three in a registration stability program. I assume that for the other two batches, different lots of the active ingredient and perhaps excipients will be used. If the impurity was not seen in previous stability studies perhaps it is a feature of one of the batches of ingredients that were used in the first registration batch. So I would like to delay manufacture of batches 2 and 3 by a few weeks and conduct a LC-MS investigation. This will require some further analytical method development and further manipulations so could take a few weeks but in the hands of an experienced organic chemist it can also generate a lot of very valuable chemical structure information for an unknown impurity. If it turns out that the unknown impurity results from an issue with a specific ingredient batch, this would explain why it was not observed in early experimental stability studies and should not be too difficult to manage moving onto the outstanding two registration batches and subsequent commercialization. If on the other hand this is due to an interaction between the active ingredient and one of the excipients which is not dependent on any specific ingredient batch characteristic, it should have been dealt with during early development and will now probably lead to longer delays as a reformulation is likely to be required
• It is unlikely that registration strategy option 3 will be acceptable unless it can be shown that the new impurity was due to an issue with a specific ingredient batch and this is not representative of the material that will be used for manufacture of routine commercial product batches.
o It may allow the company to achieve the initial NDA filing date but what you should really want to achieve is an acceptable NDA approval date and if we choose to go ahead with what we’ve got this is likely to lead to a lot of difficult questions that are likely to end in regulatory rejection of the current product
