Development and application of a high-throughput sample cleanup process based on 96-well plate for simultaneous determination of 16 steroids in biological matrices using liquid chromatography–triple quadrupole mass spectrometry

A novel high-throughput sample pretreatment system was developed by the integration of protein precipitation (PP), phospholipid removal (PPR), and hollow fiber liquid-phase microextraction (HF-LPME) into two simple 96-well plates and a matching 96-grid lid. With this system, 16 steroids were separated from biological matrices of plasma, milk, and urine and analyzed by liquid chromatography–triple quadrupole mass spectrometry. In the tandem sample cleanup process, the prepositive PP and PPR step preliminarily removed some of the interferences from the biological matrices. The following HF-LPME step kept the residual interference out of the hollow fiber and enriched the steroids in the hollow fiber to achieve high sensitivity. By a series of method optimizations, acetonitrile was chosen as the crash solvent for PP and PPR. A mixture of octanol and toluene (1:1 v/v) was used as the acceptor phase for HF-LPME. The extraction was conducted at 80 rpm for 50 min in a donor phase containing 1 mL 20 % sodium chloride at 25 °C. Under these conditions, the limits of detection for the 16 steroids were 3.6–300.0 pg.mL⁻¹ in plasma, 3.0–270.0 pg.mL⁻¹ in milk, and 2.2–210.0 pg.mL⁻¹ in urine. The recoveries of the 16 steroids were 81.9–97.9 % in plasma (relative standard deviation 1.0–8.0 %), 80.6–97.7 % in milk (relative standard deviation 0.8–5.4 %), and 87.3–98.7 % in urine (relative standard deviation 1.0–4.9 %). Further, the integrated 96-well platform of PP, PPR, and HF-LPME enabled us to run this assay in an automatic and high-throughput fashion. The reliability of the method was further corroborated by evaluation of its applicability in plasma and urine samples from volunteers and fresh bovine milk from local dairy enterprises. Graphical Abstract The combined 96-throughput clean-up system for biological samples detection