Significant Change After Photostability Testing

We have a 2-year-old photostability chamber that has been performing well. It is well mapped, temperature and humidity controlled, and the bulbs are relatively fresh—just changed out a month ago. To perform confirmatory testing, in three petri dishes we place 20 tablets of our new drug product, each in a single layer. We left the dishes uncovered for the relatively short Option 2 exposure. Since the chamber is temperature controlled, we opted not to run a dark control. According to the sampling plan, after exposure, 10 tablets were removed from each, placed in a sample container and sent to the lab for assay, degs and appearance and color. The other 10 tablets were placed in a separate container and not sent to the lab.

Within three days the lab reported that we had assay results of (1)99.8%, (2)97.9% and (3)94.3% vs. initial results of 99.6% of the label claim (specifications are 95% to 105%). After an initial investigation the remaining sample was sent for testing per a pre-approved retest plan, with those results at (1)99.6%, (2), 98.5% and (3)95.2%, confirming the initial findings. Mass balance checked out, as well, with total degs of sample (3) at close to 4%. We suspect variability within the chamber, but we did not record the exact location of the petri dishes during exposure in order to recreate the exposure conditions.

Here are some possible courses of action:

1. Run with the confirmed results and either add light protection to the product or modify the label to warn against light sensitivity. Clearly the sample results show the product is light sensitive and this can’t be ignored.

2. Re-map light exposure in the chamber using enough sensors to cover the diameter of the petri dishes and hopefully pinpoint the suspected light hotspot. Use those findings (hopefully) to repeat the study.

3. Repeat the study with sample in the primary package. For good measure, given these strange results, also include sample in the proposed market package.

4. Portions of two and three combined.