Bad accuracy

[chicay]

Hi,
i've been trying to validate method of assay for Pregabalin capsules and have been facing problems with the accuracy.
we tried using different solvents for diluting the analyte, including NaOH 0.1 N, HCl 0.1 N, H3PO4 0.1%, water, phosphate buffer pH 7.0, with mobile phase (phosphate buffer : methanol = 9:1). We use different ways to dilute the analyte: sonication, vortex. the accuracy that we obtained ranges from 96.0% to 98.0%, however we need to reach 98.0% to 102.0%.
Do anyone have any idea how to solve this?
Thank you.

[GaryR]

Try using water at 50 - 55C to extract the capsules with vigorous shaking until the capsules break up, then shake a further 10 minutes, sonicate 5 minutes and cool to room temp before making to volume. Centrifuge and dilute.

[gtma]

I suggest you start with GaryR's suggestion.....if the drug does not degrade significantly at high temp. Your low assay results could also be associated with issues other than extraction of the drug from the matrix. How are you conducting your accuracy expt? What instrument used? You should be able to search this forum for topics related to low accuracy/recovery. Good Luck!

[chicay]

HI All,
I tried conducting the accuracy experiment using the method from Gary but I am still getting the result around 96-97%. I used HPLC Shimadzu 2010CHT system with C8 column with phosphate buffer pH 7: methanol (9:1) as the mobile phase. If anyone know what could have been the problem or any possible solution. Thank you very much

[paulw]

If you are consistently returning results of 96-97% with a good %RSD from the preparations, then perhaps the sample you are using is really ~97%. This could be related to a few different things. It could be related to the purity of the material used to make the preparations. It could be related to a difference in purity between the sample preparation material and the standard material. It could be related to the preparation of the capsules. Having done some ANDA work regarding capsule (and tablet) preparation, I know how tricky that can be. To see if it truly is a matrix issue, I would get some placebo capsules, then spike them with the standard solution you are testing to see if you recover at 100% of the analyte that you know you just introduced. Not perfect as you already have the analyte in solution, but it will show you if something in the capsule matrix has the ability to tie up the analyte.

A second way to determine if it is dispersion is to make a small lab batch, perhaps ~5 capsules worth in total then disperse the whole of the sample in 5 times the volume you would normally use (maybe a 500 mL or 1000 mL volumetric), washing the container you made it in etc, that way you don't have issues with loss in container or non-homogenous sample mixture.

Hopefully you get your issue resolved.

Paul

[gtma]

paulw mentioned a few suggestions that should help in understanding what may be happening. The first expt I would try is;

1. Prepare a placebo capsule spiked with the API powder at 100%.....if the amt is more than 10mg per capsule or if you're confident with your weighing accuracy. Prep nominal conc. sample soln.
2. Use the same API lot for std (neat std) .
3. Use the stock std soln (neat) and spiked with excipient and capsule to prep. nominal conc. sample soln.
4. Analyze filtered and unfiltered solutions from 1 to 3
5. Calc. recovery for filtered/unfiltered spiked API as well as std and evaluate the results.

6. To evaluate solubility, prep. spike placebo as in 1. at 80% and 120%.
7. Obtain UV spectra (DAD) for above API peak and evaluate for potential shift leading to bias low results.

Note: prep. at least duplicate for each. I hope the API pka is not within 1.5 pH unit of the buffer used and the method is stability indicating in the event there is degradation.