I am analysing rat plasma samples for SN38 by LC-MS using protein precipitation with ACN (3× plasma volume, CPT as IS) on a C18 column. The method works fine for calibration (LOQ 0.1 ng/mL, linear range) and used to work for unknown plasma samples as well.

Now, unknown samples show multiple peaks (1.2, 2.6, 2.8 min) instead of a single sharp SN38 peak at 2.8 min. This does not occur with spiked plasma standards. Acidification to favour the lactone form only improves it slightly.

Has anyone encountered this issue before?

Thanks,
Ali