CPZ, PMZ, TPZ, PPZ, CPT and chloropromazine-d₃ (CPZ-d₃) were purchased from the National Institute for the Control of Pharmaceutical and Biological Products. All reagents were HPLC grade. Acetonitrile and methanol were purchased from Fisher Scientific (Fairlawn, NJ). Formic acid was obtained from Sigma-Aldrich (USA). Deionized water was generated with an Elix water purification system (Millipore Corp., Molscheim, France).

Analyses were performed on a Waters ACQUITY UPLC system with the Xevo TQ-S MS as the detector. Chromatographic data were recorded and analyzed by Masslynx^TM software (Version 4.1).

Separations were carried out on an ACQUITY UPLC HSS SB C18 column (2.1 mm x 50 mm, 1.7 μ m; Waters, USA), being maintained at the temperature of 35 ° C and a constant flow rate of 0.5 mL/min. The mobile phase was composed of acetonitrile (A) and 0.1 % formic acid (B) in purified water. The gradient program was selected as shown in Table 1. The sample injection volume was 5 μ L.

Table 1

Table 1

Table 1 Gradient program of the established method Time/min Flow/(mL/min) A/% B/% Curve 0 0.5 30 70 - 3.0 0.5 80 20 6 3.1 0.5 90 10 1 4.0 0.5 90 10 6 5.0 0.5 30 70 1 6.0 0.5 30 70 1

Table 1 Gradient program of the established method

The quantitative analysis of 5 analytes was performed in the MRM mode. Positive ionization mode (ESI+) was used for all analytes. The MS parameters were: capillary voltage 0.5 kV, desolvation gas temperature 400 ° C, desolvation gas flow 800 L/h, cone gas flow 150 L/h and collision gas flow 0.16 mL/min. Each analyte was identified by two characteristic MRM transitions as shown in Table 2.

Blank human whole blood samples were obtained from volunteers with informed consent. The samples were stored at -20 ° C. The project was approved by the Ethics Committee on Human Research of the Faculty of Health Science of Ministry of Health. Standard stock solutions of each analyte (2 mg/mL) were prepared in methanol. A mixed standard solution containing the 5 analytes at 1 μ g/mL of each was prepared from the standard stock solutions. Spiked blood samples were prepared by the mixed standard stock solution adding into blank blood at appropriate concentrations in the range of 0.2-20 ng/mL.

Table 2

Table 2

Table 2 The two resulting MRM transitions per analyte and corresponding settings for UPLC-ESI-MS/MS system Analyte PI/(m/z) DI/(m/z) CV/V CE/V CPZ 319.13 58.17 86.13* 6 26 18 PMZ 285.17 86.12 198.07* 2 16 26 TPZ 408.20 113.17 141.13* 40 28 23 PPZ 404.12 143.13 171.15* 6 28 22 CPT 316.06 230.97 271.08* 12 26 20 CPZ-d3 322.16 61.15 89.12* 26 26 18 PI: parent ion; DI: daughter ion; CV: cone voltage; CE: collision energy; * : Quantitative ion

Table 2 The two resulting MRM transitions per analyte and corresponding settings for UPLC-ESI-MS/MS system

Captiva ND Lipids (Varian, USA) plates were used for blood sample preparation. Aliquots of 100 μ L spiked blood sample (with 1 ng/mL CPZ-d₃ as internal Standard, IS) were transferred into a 96-well filtration plate. Then, 400 μ L of 0.1 % formic acid in methanol was added to each sample. Subsequently, the 96-well filtration plate was placed on a vacuum manifold and filtered under 20 psi pressure to remove any precipitated proteins. The filtrate was collected into the 96-well receiving plate. Thereafter, the 96-well receiving plate was capped and placed in the sample manager and 5 μ L of each filtrate was used for UPLC-MS/MS analysis.

Selectivity and specificity were evaluated by analyzing six different sources of blank blood samples, spiked blood samples and neat standard solutions (1 ng/mL), examining whether there were any interferences between the matrix signals and those from the analytes and the deuterated standard. A representative chromatogram was given in Fig. 1, showing that all the analytes of interest were free of any interference from the matrix and separated on the baseline. Conventional protein precipitation of blood samples usually leaves major interferences from the matrix. However, in samples prepared with Captiva ND Lipids under identical precipitation conditions, over 97% of phosphatidylcholines, lysophospholipids included, are removed. This lipid depletion significantly reduces the interferences from the compounds by the method itself.

Calibration curves were prepared with seven concentration points (0.2, 0.5, 1.0, 2.0, 5.0, 10.0 and 20.0 ng/mL) for each drug in blood. Calibration was performed by linear regression of peak-area ratios of the drugs to the internal standard versus the respective standard concentration. The deuterated internal standards remained at a concentration of 1 ng/mL. Method sensitivity was systematically evaluated with a signal-to-noise ratio of at least 3. The regression equations, correlation coefficients and LODs for each analyte are listed in Table 3. All of the calibration curves were linear in the range of 0.2~20 ng/mL with correlation coefficients (r²) higher than 0.999. The LODs were 3-10 pg/mL and were far less than those reported in the literature (Table 6). When blood samples being prepared with Captiva ND Lipids, the relative responses were higher than the conventional protein precipitation. This increased sensitivity results from a significant reduction in ion suppression since Captiva ND Lipids virtually removes lipids, proteins and surfactants.

Table 3

Table 3

Table 3 Regression equations, correlation coefficients and LODs for the analytes in blood Analyte Regression equations r2 LOD/ (pg/mL) CPZ Y=1.40C+0.009 0.9991 3 PMZ Y=1.45C+0.139 0.9994 3 TPZ Y=1.15C+0.008 0.9997 8 PPZ Y=0.86C+0.049 0.9994 10 CPT Y=0.45C+0.019 0.9993 7

Table 3 Regression equations, correlation coefficients and LODs for the analytes in blood

Measurements of precision and accuracy were carried out based on three concentration levels: 0.2, 2.0 and 20.0 ng/mL. Intra-/inter-day precisions were evaluated by analyzing five replicates of each level for three consecutive days. Precision was measured as the relative standard deviation (RSD). Accuracy was calculated according to the following formula: accuracy (%) = [(calculated value - nominal value)/nominal value] × 100. From the data listed in Table 4, it can be seen that the accuracy was within -2 to 2 % and the RSDs were below 10 %. With Captiva ND Lipids, the integrity of the peak shapes was maintained. The smoother aspect of the peak made integration easier and more precise, and the reproducibility of quantitation also being improved.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Quantitative ion chromatograms of each analyte in neat standard solution (A), blank blood sample (B) and spiking blood sample (C)

Table 4

Table 4

Table 4 Accuracy, intra- and inter-day precision data (values given in %) of the UPLC-ESI-MS/MS assay for each analyte (n=5) Analyte Accuracy Intra-day precision (RSD) Inter-day precision (RSD) 0.2 2.0 20.0 0.2 2.0 20.0 0.2 2.0 20.0 CPZ -0.8 1.0 1.8 1.5 2.1 6.0 2.7 6.9 8.8 PMZ -1.3 0.5 1.7 2.9 1.7 4.5 3.4 5.9 9.1 TPZ -0.9 0.1 1.0 3.5 0.9 3.6 6.0 5.8 9.7 PPZ -1.1 0.2 1.3 5.7 0.8 6.9 5.6 7.4 9.6 CPT -1.0 0.6 1.5 2.0 4.3 5.6 4.9 7.0 9.2

Table 4 Accuracy, intra- and inter-day precision data (values given in %) of the UPLC-ESI-MS/MS assay for each analyte (n=5)

The extraction efficiency was estimated at three levels with five replicates at each concentration. Extraction recovery was calculated by comparison of the peak area obtained when the analytes and the IS were added before filtration (n=5), with those obtained when both the analytes and IS were added after the filtration step (n=5).

Matrix effects were determined by measuring five replicates of neat standard at three concentration levels and the filtrates from five spiked blank blood samples, and were calculated as the percentage of peak area in the presence of matrix relative to the peak area without matrix. Data are listed in Table 5. The extraction recoveries of samples containing the analyts at three concentration levels were 73.9 % - 103.3 % and the results of matrix effects were also satisfactory. The ion suppression effects from non-removed phosphatidylcholine interferences are obvious in a post-column infusion experiment with albuterol. When using Captiva NDLipids, the cleanliness of the extract is significantly improved compared to conventional protein precipitation. Consequently, ion suppression is dramatically reduced with no noticeable loss in MS response.

Table 5

Table 5

Table 5 Extraction efficiency and matrix effects (values given as %) of the UPLC-ESI-MS/MS assay for each analyte at low, medium and high concentrations (ng/mL, n=5) Analyte Extraction efficiency Matrix effect 0.2 2.0 20.0 0.2 2.0 20.0 CPZ 93.9 76.2 85.8 61.5 117.4 63.3 PMZ 90.9 79.2 86.8 38.6 115.1 88.8 TPZ 91.6 73.9 82.4 53.4 112.0 107.2 PPZ 103.3 76.2 91.8 41.6 118.3 133.6 CPT 78.0 74.2 88.2 51.2 105.8 94.1

Table 5 Extraction efficiency and matrix effects (values given as %) of the UPLC-ESI-MS/MS assay for each analyte at low, medium and high concentrations (ng/mL, n=5)

Blood samples were prepared of containing the analytes at three different concentrations (n=5 at each level) and kept at a temperature of 18 ° C for 24 h and in one freezer at -20 ° C for a month to evaluate the short- and long-term stability. To evaluate the stability of processed blood samples, filtrates of simulated blood samples containing the analytes at three different concentrations (n=5 at each level) were kept for 24 h in an autosampler at 10 ° C as well as in a freezer at -20 ° C. For the evaluation of the stability after three freeze-thaw cycles, spiked blood samples were frozen at -20 ° C for 24 h. After this period, the samples were again thawed and frozen for 24 h and this process was repeated until the third thawing cycle when the samples were processed and analyzed. All the stability-evaluating samples were determined by freshly-prepared calibration curves and compared with the reference concentrations. It was found that all measurements were within 15 % of the nominal concentrations, indicating that the samples were stable in each of the above storage conditions.

A comparison of the validation results for the proposed method with those of methods reported in the literature is given in Table 6. The established method has a wide range of analytes, tiny sample volume, time-saving for sample preparation, and much lower detection limit compared to the reported methods. These satisfactory results should be attributed to Captiva ND Lipids which uses a simple, streamlined and easy-to-follow 3-step process, making the quantification for trace analysis more precise and reproducible, reducing costs and increasing productivity. The described method was especially suitable for rapid analysis of trace analytes in small volume of specimens.

The established method was applied to several real forensic cases. Bloods from six fatalities were collected at autopsy and analyzed by the proposed method after appropriate dilution. Case details and analytical results are shown in Table 7. CPZ, PMZ and CPT had been each detected in two different cases, respectively. The proposed method was proved to be a robust and fast alternative in forensic toxicology analysis.

Table 6

Table 6

Table 6 Comparison of the proposed method with reported choices Method Analyte Sample volume /μ L Processing time /min Linear range /(ng/mL) LODs/(ng/mL) Ref. PPT-UPLC-MS/MS PPZ 100 > 12 0.2~12 0.2 [2] SPE-GC-MS CPZ 1000 > 20 1~2000 - [3] MSPE-HPLC-UV CPZ 5000 > 15 10~300 10 [4] SPE-GC-MS PPZ 500 > 60 2~64 2 [5] SPE-HPLC- fluorescence PMZ 3000 > 15 250~10000 400 [6] DLLME-CZE-UV TPZ PPZ 1000 > 30 1.5~100 1.0~100 0.5 0.3 [8] 96 well plate- UPLC-MS/MS CPZ PMZ TPZ PPZ CPT 100 < 5 0.2~20 0.003 0.0033 0.008 0.010 0.007 This paper

Table 6 Comparison of the proposed method with reported choices

Table 7

Table 7

Table 7 Details of forensic cases and detection results using the established method Case number Sex Age Sample Detected compound Concentration/(ng/mL) 1 Male 61 Blood CPZ 4400 2 Male 51 Blood CPZ 6700 3 Male 59 Blood PMZ 7700 4 Female 48 Blood PMZ 10000 5 Female 41 Blood CPT 250 6 Male 52 Blood CPT 330

Table 7 Details of forensic cases and detection results using the established method