Selected reaction monitoring LC-MS determination of idoxifene and its pyrrolidinone metabolite in human plasma using robotic high-throughput, sequential sample injection

The generation of large numbers of samples during early drug discovery has increased the demand for rapid and selective methods of analysis. Liquid ch romatography- tandem mass spectrometry (LC-MS-MS), because of its sensitivi ty, selectivity, and robustness, has emerged as a powerful tool in the phar maceutical industry for many analytical needs. This work presents;a high-th roughput selected reaction monitoring LC-MS bioanalytical method for the de termination of idoxifene, a selective estrogen receptor modulator, and its pyrrolidinone metabolite in clinical. human plasma samples. The described m ethod uses short, small-bore columns, high now rates, and elevated HPLC col umn temperatures to perform LC separations of idoxifene and its metabolite within 10 s/sample. Sequential injections were accomplished with a 215/889 multiple probe liquid handler (Gilson, Inc.), which aspirates eight samples simultaneously and performs its rinse cycle parallel to sample injection, resulting in minimum lag time between injections. This high-throughput meth od was applied to the determination of idoxifene and its metabolite in clin ical human plasma samples. Sample preparation employed liquid/liquid extrac tion in the 96-well format. Method validation included determination of int ra- and interassay accuracy and precision-values; recovery studies, autosam pler stability, and freeze-thaw stability. The:LOQ obtained was 10 ng/mL fo r idoxifene and 30 ng/mL for the metabolite. Using idoxifene-d(5) as an int ernal standard, idoxifene showed acceptable accuracy and precision values a t QC level 1 (QC1, 15 ng/mL), level 2 (QC2, 100-ng/mL) and level 3 (QC3, 18 0 ng/mL) (85.0% accuracy +/- 12.0% precision, 95.1 +/- 4.9%, and 90.3 +/- 4 .7%, respectively). The pyrrolidinone metabolite;also showed acceptable acc uracy and precision values (using no internal standard for quantitation) at QC1 (60 ng/mL),: QC2 (100 ng/mL), and QC3 (180 ng/mL) (104.9 +/- 14.4%, 91 .1 +/- 13.0%, and 90.8 +/- 12.2%, respectively). The validated method was a pplied to the; analysis of 613 human clinical plasma samples. An average ru n time of 23 s/sample (similar to 37 min/96-well plate or over 3700 sample/ day) was achieved. The successful validation presented indicates that rapid methods of analysis can efficiently and reliably contribute to the fast sa mple turnaround required for high sample number generating processes.