Quantitation of midazolam in human plasma by automated chip-based infusion nanoelectrospray tandem mass spectrometry.

An automated chip-based infusion nanoelectrospray ionization (nanoESI) platform was used to demonstrate reproducible quantitation of drug molecules from biological matrices. Three sample preparation strategies were explored including protein precipitation of plasma with acetonitrile, de-salting of the plasma, and a combination of protein precipitation with subsequent de-salting of the dried and reconstituted extract. The best results were obtained when fortified human plasma samples containing midazolam were precipitated with acetonitrile containing alprazolam as the internal standard (IS). The supernatant was concentrated to dryness, reconstituted in aqueous acid, and de-salted by automated reversed-phase solid-phase extraction (SPE) prior to infusion nanoESI-MS/MS. Analyses employed a triple quadrupole mass spectrometer operated in selected reaction monitoring (SRM) mode. Each sample was infused for approximately 10 s and the resulting ion current profiles were integrated. Area ratios were used for regression analysis of standard samples (1.5-500 ng/mL). Quality control samples (3, 250, and 400 ng/mL) in five replicates from three different analysis days demonstrated intra-assay precision (< or =16%), inter-assay precision (< or =5%), and overall accuracy (+/-9% deviation). Infusion reproducibility of the assay was established by analyzing extracts after storage for 24 h at ambient temperature. Control plasma samples from six different sources probed the potential utility of this technique for the analysis of clinical samples. At the lower limit of quantitation (LLQ), variability and mean overall accuracy were < or =13% CV and +/-3% deviation, respectively, while at the upper limit of quantitation (ULQ) variability and mean overall accuracy were < or =9% CV and +/-9% deviation, respectively. Inter-chip variability was established by determining standard sample extracts across five different chips (< or =12% CV). Throughput for the assay was 55 s per sample, although this time may be shortened to 40 s per sample with recent improvements in the automated nanoESI system. No contamination or carryover was observed using this promising automated nanoESI-MS/MS platform. Copyright 2003 John Wiley & Sons, Ltd.