2.1. Apparatus

An API BioSystem 3200 tandem mass spectrometer, equipped with turbo-ion spray interface was used for measurements. The HPLC system consisted of Perkin-Elmer 200 series autosampler and binary pump, Restek PFP Propyl precolumn (10 mm×2.1 mm) and C18 analytical column (50 mm × 2.1 mm i.d.) with particle size of 5 µm. The mobile phase was degassed using vacuum degasser (Perkin-Elmer 200 series). Data acquisition and processing were achieved using Analyst 1.6 software (Applied Bio-systems).

2.2. Reagents

Chlorpromazine, haloperidol, levomepromazine, olanzapine, risperidone, sulpiride, and repaglinide were obtained from the National Laboratory of Pharmaceutical Products of Algiers. HPLC grade acetonitrile and methanol were obtained from Panreac and Sigma-Aldrich, respectively. Ammonium formate and formic acid used as buffer system were obtained from Analar Normapur and Panreca, respectively.

High-purity water for preparative purpose was produced by double distillation for ultrapure deionised water (18.2 MΩ-cm, type I) by bidistillation apparatus (PureLab Option-Q).

2.3. Sample and Calibration Standard Preparation

The stock solution was prepared by dissolving 10 mg of chlorpromazine, haloperidol, levomepromazine, olanzapine, risperidone, and sulpiride in methanol to obtain a final concentration of 100 mg/L, which was kept at 4°C until analysis. Work solutions were freshly prepared by an adequate dilution of the stock solution with methanol. The calibration standards were obtained by diluting (1: 20) the corresponding work solution with free blank plasma.

2.4. Sample Preparation

In this study, the sample preparation was successfully used to measure antipsychotic concentrations in human plasma samples. Blood samples were collected from patients. After centrifugation at 5,000 rpm for 5 minutes, the obtained plasma samples were transferred to cleaned tubes and kept at 4°C until analysis. 20 μL of the internal standard (repaglinide 1,000 ng/ml) were added to 500 µl of calibration standards, plasma samples, and QC.

2.5. Chromatographic Conditions

An aliquot of 20 µl of each sample and calibration standard was loaded on the column. Gradient reversed phase high-performance liquid chromatography was performed by mobile phase consisting of 90% solvent A (water ammonium formiate 2 mM; pH 2.7) + 10% solvent B (acetonitrile) for 3 minutes and subsequently decreased linearly to 10% solvent A over 4 minutes. The system was reequilibrated to the initial condition over 2 minutes. The reequilibrated condition remained for 1 minute. For all separation process, the mobile phase was set at a flow rate of 800 µL/min.

2.6. MS-MS Conditions

The spectrometric measurements were made in positive mode and operated using MRM mode. The optimal instrumental settings are given in Table 1. All given values (compounds and source/gas parameters) are the averages of three measurements (Tables 1 and 2).

2.7. Validation Experiments

The proposed method was validated for recovery, matrix effect, LOD, LOQ, selectivity, specificity, carryover, linearity, and stability. In addition, an intra- and interday validation were performed to evaluate the accuracy and precision of the measurements. All these validation experiments were carried out to allow a bioanalytical application of the present method.

3.1. Extraction Experiment

In this study, the extraction efficiency (recovery) was evaluated by comparing detector signals (peak areas) obtained from extracts of QC samples at low, medium, and high levels of all tested antipsychotics with those obtained with the corresponding standard solutions added to extracted matrices (Table 3). For all tested antipsychotics, the mean recoveries were more than 80% showing the higher efficiency of the proposed SPE procedure for some analytes compared to those obtained with other methods in previous studies (Table 6). To our knowledge, this is the first work that uses SPE-HLB cartridges for antipsychotic extraction. Although this work uses a higher amount of plasma compared to other methods, the obtained extract was clear and prevented rapid clogging of injection and ESI needles often encountered with simple LLE.

3.2. Chromatogram

Chromatograms with MRM profiles obtained from human plasma containing the tested antipsychotics are shown in Figure 1. Distinct peaks appeared for all compounds with different retention times (Table 4). The total chromatographic run time for analyte separation was 10 minutes, which is suitable for routine analysis as reported in previous works [23, 27, 31]. Representative chromatogram of drug free blank plasma is shown in Figure 2(a).

3.3. Linearity

The calibration curve was established with six points of standard solutions with all tested antipsychotics. Each point was determined by five calibration runs. The first-order regression equations with correlation coefficients are shown in Table 4. The linearity of the calibration curves was evaluated by the lack of fit test at 5% level of significance. According to F_reg values, the regression explains the observed variations. The F_nl values indicate nonsignificant lack of fit of the calibration curve. Therefore, the regression equations established a linear relationship between antipsychotic concentrations and detector signals in the tested ranges. Thus, subtherapeutic, therapeutic, and toxic levels of all tested psychotics can easily be determined without constantly needing of dilution procedures (Table 6).

3.4. Accuracy and Precision

In this study, precision and accuracy were determined by intraday and interday validation using QC at four levels (LOQ, 3 LOQ, 50% and 75% of the calibration curve). The intraday validation was performed by five replicate analysis of QC on the same day. For interday validation, five replicate measurements on three different days were performed. Analyte concentrations in QC were calculated using the regression equation of the calibration curve. Accuracy and precision of the analytical method were calculated and the expressed values are summarized in Table 5.

All the coefficients of variation (CV%) of intraday and interday measurements were not greater than 9% and 11%, respectively. These results indicate a multilevel high precision of the present method regardless of time factor. All accuracy measures of intraday and interday validation were less than 10% and 12% (absolute values), respectively, which indicate a multilevel high accuracy of the present method regardless of time factor. Moreover, the intraday precision in all calibration range (repeatability) was acceptable at 5% level of significance according to Cochran test (C (5%)=0.727). The interday precision (reproducibility) in all calibration range was also acceptable at 5% level of significance according to Grubbs test (G (5%)=1.115).

3.5. Limits of Detection and Quantification

The LOD and LOQ were calculated using S/N (S/N ≥ 3 for LOD and S/N ≥ 10 for LOQ (Figure 2(b))) as well as the slope and the standard deviation of calibration curve intercepts (3 s(b0)/b and 10 s(b0)/b values, respectively) (Table 4). The highest values between the two approaches were considered as LOD and LOQ for routine analyses. In general, the values of LOD and LOQ using S/N approach were below of those obtained in previous studies and demonstrate that concentrations below therapeutic ranges can be reached and determined (Tables 4 and 6).

3.6. Matrix Effect, Selectivity, Specificity, and Carryover

With respect to matrix effect and selectivity, no major interferences (0.2 times the response of the LOQ) were detected at the retention times of all tested antipsychotics or the internal standard in 10 batches of free blank plasma and no suppression effect was found to all tested antipsychotics. In addition, common drugs, when injected into the mass spectrometer, do not generate any interfering ions with those selected for antipsychotic quantification.

Due to the large range of the calibration curve in this method, the carryover was assessed by measuring detector signals (peak areas) of blank samples after the higher calibration point. The accepted limit for carry over was that the detector signals of blank samples must be less than 20% of the LOQ signal. The obtained carryover in this method complied with the acceptable limits.

3.7. Stability

The stability of the tested antipsychotics in plasma matrix was estimated using QC stored in different conditions (Table 7). The stability was expressed by the relative bias of the found concentrations to the nominal concentrations. The tested antipsychotics were considered stable when less than 15% of the nominal concentration is obtained using relative bias.

The results indicate that the signs of deterioration of all tested antipsychotics were within the acceptance limits in different conditions (Table 7).

3.8. Application

The present method was successfully used in our laboratory for routine analysis of plasma samples taken from patients receiving these antipsychotics in TDM context. The method was also used for clinical analyses of plasma samples obtained from intoxicated individuals. Overall, plasma samples from 71 patients (48 males and 23 females) were analyzed and some of the patients were examined repeatedly. A summary of the obtained results is given in Table 8. Three samples which gave results above the calibration range were diluted with calf serum and reanalyzed.

The average age of the patients was 29.3 ± 7.4 years (14-69) for males and 37.1 ± 6.9 years (16-72) for females. Examples of representative MRM chromatograms of plasma samples from six different patients taking one of the studied drugs are shown in Figures 3(a)–3(f).

According to our experience, TDM could help to enhance the therapeutic response, design optimal dosing regimens, and avoid the build-up of excessively high and potentially toxic drug concentrations, as well as monitor patient’s adherence to treatment.