Determination of citalopram in fish brain tissue: benefits of coupling laser diode thermal desorption with low- and high-resolution mass spectrometry.

Recent state-of-the-art methods developed for the analysis of polar xenobiotics from different types of biological matrices usually employ liquid chromatography with mass spectrometry. However, there are limitations when a small amount of sample mass is available. For example, individual benthic invertebrates or fish tissue samples often weigh less than 100 mg (e.g., brain, liver) but are necessary to understand environmental fate and bioaccumulation dynamics. We developed ultra-fast methods based on a direct sample introduction technique. This included coupling laser diode thermal desorption with atmospheric pressure chemical ionization mass spectrometry (LDTD-APCI-MS). We then quantitated a common selective serotonin reuptake inhibitor (citalopram) in brain tissues of individual juvenile fish after in vivo exposure to environmentally relevant concentration. Two mass spectrometric methods based on low (LDTD-APCI-triple quadrupole (QqQ)-MS/MS) and high (LDTD-APCI-high-resolution product scan (HRPS)) resolutions were developed and evaluated. Individual instrument conditions were optimized to achieve an accurate and robust analytical method with minimum sample preparation requirements. We achieved very good recovery (97-108%) across the range of 1-100 ng g for LDTD-APCI-HRPS. LDTD-APCI-QqQ-MS/MS showed poorer performance due to interferences from the matrix at the lowest concentration level. LDTD-APCI ionization was successfully validated for analysis of non-filtered sample extracts. Evaluation of final methods was performed for a set of real fish brain samples, including comparison of LDTD-APCI-HRPS with a previously validated LC-heated electrospray ionization-HRPS method. This new LDTD-APCI-HRPS method avoids the chromatographic step and provides important benefits such as analysis of limited sample masses, lower total sample volume (typically μL), and reduction in analysis time per sample run to a few seconds. Graphical abstract.