Mininum number of Tablets required for Assay

[michaelcarolus]

All

I have an Assay and UOC method that is analysed using the same HPLC method. The sample prep is as follows:
Assay
Crush 20 tabs to a fine powder. Weigh equivalent of 1 powdered tab (1.6g) into 500 ml vol flask. Add 400 ml diluent and sonnicate for 30 min. Make up to volume with diluent. Filter through 0.45 um PVDF filter.

UOC
Transfer 1 tab to each of 10 separate 500 ml vol flasks. Add 400 ml diluent and sonnicate for 30 min. Make up to volume with diluent. Filter through 0.45 um PVDF filter.

When we compare the Assay and the UOC results, the UOC is normally 100% of label claim and the Assay is 97% of label claim. We have tried various ways of obtaining a more homogeneous sample e.g. using a mechanical mill to crush the tablets, but we still obtain an assay of 97%.
We would like to use whole tablets for the assay as this will eliminate the crushing step. For this sample we would like to use 4 whole tabs to 2000 ml for the sample prep (prepared in duplicate). Would this sample prep satisfy an auditor as being a representative sample of the batch?

Thanks in advance
Mike

[shepchemalum]

Most assays I've performed at any of my jobs, even when the USP method says to grind the tablets and take a certain amount, we actually take 10 tablets into a volumetric flask, and dilute to the requested final concentration. It is very likely that you are losing a few percent from either the blend not being homogenous. Hope this helps. Never had any issues with auditors. I would think 4 tablets might be on the low side, but since you're having to crush 20 tablets to make your blend, you should be able to use 10 and dilute down to your concentration. just do 10 tabs into 500mL, add your 400mL diluent, sonicate/shake/sonicate, cool dilute to 500, and do a 1:10 dilution with diluent.

[Peter Apps]

It is very likely that you are losing a few percent from either the blend not being homogenous.

Sample inhomogeneity gives results from replicates of the same sample that are not repeatable - a consistent 97% result is a bias, which does not come from ihomogeneity.

Peter

[gtma]

First of all, it's important to identify the likely cause for the low assay results. Is the mean of the UOC results similar to that of the assay results? If you crush several tabs individually and transfer the entire content in separate flask as per the assay procedure. Do you get results similar to that of the UOC? Segregation of the excipient and API during the crushing step can be the probable cause.

[michaelcarolus]

Hi all

Thanks for your comments/suggestions thus far. The mean UOC results is 100% of the label claim, with AV = 2.0. The assay has always been about 2-3% lower than the UOC. For the UOC we take 1 whole tab to 500ml. The diluent is 0.01M Hcl, so the whole tablet disintegrates/dissolves relatively quickly. I would prefer using whole tablets as this would eliminate the difficulty in crushing the tablets to obtain a homogenous sample.

Regards

Mike

[aceto_81]

Did you try to grind each tablet and perform UOC?
If you also find a few % lower, it must be something to do with the grinding.
Did you use an agate mortar?

HTH

Ace

[michaelcarolus]

Hi All

Thanks for the comments. We will to grind each tablet and perform UOC type testing. We are using the classic agate mortar and pestle to grind the tablets. We have also tried an electric mill (coffee grinder type) to grind the tablets. The results for both grinding techniques are usaually 2-3% lower than the UOC.

Thanks again.
Mike

[Consumer Products Guy]

Your active isn't somewhat volatile, is it?

[krickos]

Hi

Interesting topic, as Peter mentioned you seem to have a bias between UoC and Assay sample preparation.

Just to share my experiance and ongoing investigation I am involved with:

Grinding/crushing/dissolving:
Although rare, we have a case if we use the classic agate mortar and pestle to grind the tablets, we eventually push the drug substance (lipophilic, no pka to work with and used as a possible worst case in cleaning) into one of the excipients which the solvents in use can not extract 100%, so we use the mentioned grinder/crusher approach to avoid that issue.

Similar case under investigation:
A drug Product that shows similar bias for one of two strengths. In our case ~3-4% off the release result taken from UoC versus stability results for an experied batch used for investigation, degradation too low to have significant impact.
In our case we suspect that the procedure operates on the borderline of solubility of drug substance in question as the concentration is considerble higher in the intial dissolution compared to the other strenght and other Products where the DS is used, but inconclusive due to some 10-20 excipients.
I would recommended that you look at what data you have around solubility, pKas etc.
. Initially one could start with just pushing the procedure a bit with pure DS ie add 50-150% of DS to initial dilution step in in 3-6 replicates per level and see if you get a hint.

As you mentioned work with the initial dissolution step, if procedure allows increase first dissolution volume and correct the following ones resulting in final concentration as Before. If the final sample concentration for injection remains the samel have less regulatory issues/change impact