Unusual Sampling Challenge

I am currently working with a situation where the equipment to be cleaned is a mostly-closed system with a water-insoluble oil-based API in a dry powder processing train. We are having initial discussions about CV and how we would accomplish sampling. The following issues have emerged:

  1. Aqueous rinse sampling would be ineffective based on the nature of the API

  2. Non-aqueous rinse solutions are not preferred, since the system is closed and there are no candidates in the product formulation (we could, in that case, make the argument that residual rinse solution was not an issue)

  3. The worst-case locations for surface swabbing are unreachable (essentially, older transfer piping, etc.)

We are forced to think somewhat “out of the box” with this scenario. Here is the solution that we are considering:

Using a known small volume of starch and blowing it through the system as a “final rinse”, collecting it in the final blender for the operation, conducting a normal blending cycle, and then testing the starch for the API. (Starch having been chosen as we know the API binds to starch based on the process.)

My questions are these:

  1. has anything like this been done anywhere else?

  2. would there be a reference to support such a sampling method?

  3. are there any cost-effective alternatives that can be offered?

Thanks in advance.

You can probably still use TOC and rinse sampling to do your cleaning validation. How do you clean your equipment? Do you use water? There is a good article on how you can justify this (see attached)


20090923163841.pdf (208.5 KB)

[quote=entil2001]I am currently working with a situation where the equipment to be cleaned is a mostly-closed system with a water-insoluble oil-based API in a dry powder processing train. We are having initial discussions about CV and how we would accomplish sampling. The following issues have emerged:

  1. Aqueous rinse sampling would be ineffective based on the nature of the API

  2. Non-aqueous rinse solutions are not preferred, since the system is closed and there are no candidates in the product formulation (we could, in that case, make the argument that residual rinse solution was not an issue)

  3. The worst-case locations for surface swabbing are unreachable (essentially, older transfer piping, etc.)

We are forced to think somewhat “out of the box” with this scenario. Here is the solution that we are considering:

Using a known small volume of starch and blowing it through the system as a “final rinse”, collecting it in the final blender for the operation, conducting a normal blending cycle, and then testing the starch for the API. (Starch having been chosen as we know the API binds to starch based on the process.)

My questions are these:

  1. has anything like this been done anywhere else?

  2. would there be a reference to support such a sampling method?

  3. are there any cost-effective alternatives that can be offered?

Thanks in advance.[/quote]

Dear entil2001,

in my opinion the sampling method using starch is not the best solution, because of two reasons: First, it is too close to the placebo method, and actually, the references will not only not support this method but plainly reject it. Second, it will be hard to demonstrate that the target analyte is homogeneously distributed in the starch if you need to take a sample out of the starch, and even harder to carry out a meaningful recovery test, because the starch would not penetrate in the hard to reach locations, which are exactly the critical ones.

As nabichan says, the method of choice would be rinse sampling. Some questions and suggestions:

  • Is the API oil based or oily? if it is a solid dissolved in oil (oil based) rinse sampling with the same oil would be an option.

  • How do you clean the equipment? If you wash it with water based cleaning agents, probably there is a water soluble solvent that could dissolve the API (remember that there are only traces of API, so even solvents with moderate dissolving power can be used, perhaps ethanol is a candidate). After sampling, the solvent could be washed away with water, and if you could get a solvent with low toxicity, you are on the safe side.

  • If there is no such solvent, a two step approach, using first a suitable (but water insoluble) solvent to sample the API, and cleaning the equipment with a water soluble solvent with a final water rinse.

Hope this adds some ideas.

Best regards

Alfred

Dear entil2001,

Starch merely acts like a glue, it is not only going to bind with the API but with equipment surfaces also (i.e. binding API to the equipment surface). Binding is totally different from solublising. Another problem (apart from the one highlighted by Alfred) with using starch is possible interference with the analytical method, it would be difficult to extract the API from the starchy solution for analysis. There are possible cost-effective sampling methods, provided the nature of the API is known (if possible you may share with the forum). Try to understand more about the solubility of the API, for this you may contact formulation department or analytical laboratory (for example, try to find out which solvent/vehicle they are using for preparing solutions for analysis). Check out pharmocopoieas for solubility. You may try using co-solvents, acidic, alkaline and/or alcoholic solutions or changing the rinsing conditions (such as raising temperature of the rinse solvent) for the purpose of rinse sampling. For example, I faced same problem while developing a rinse sampling method for Ketoconazole (insoluble in water and soluble in more non-polar solvents). In the end I solved the problem by using acidic solution (water + citric acid), which showed good recovery results. It saved cost for the company as using citric acid for rinsing purpose is far cheaper than using any solvent (e.g. ethanol, methanol etc.) Once you are able to solve the solubility issue, carry out some simulation studies and select the best possible combination/method which could give you highest possible recovery. Again remember performing recovery studies simulating real life conditions is the key to developing effective rinse sampling methods.
Like you said, the manufacturing train is a closed system, there is a potentially high risk (as the visual inspection of the equipment is not possible) of formulations being contaminated with product residues (from previously manufactured product), cleaning agent (if used any) or microbial growth. Hence, it becomes important to device a justifiable and verifiable sampling method for the whole manufacturing train.

You can get more help provided you share all the information regarding the issue concerned.