A reliable and rapid tool for plasma quantification of 18 psychotropic drugs by ESI tandem mass spectrometry

Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

Contents lists available at SciVerse ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba

A reliable and rapid tool for plasma quantification of 18 psychotropic drugs by ESI tandem mass spectrometry Gennaro Vecchione a , Bruno Casetta d , Antonella Chiapparino a , Alessandro Bertolino b , Michela Tomaiuolo a , Filomena Cappucci a , Raffaella Gatta c , Maurizio Margaglione e , Elvira Grandone a,∗ a

Atherosclerosis and Thrombosis Unit, I.R.C.C.S. “Casa Sollievo della Sofferenza”, S. Giovanni Rotondo, Foggia, Italy Department of Psychiatry, University of Bari “A. Moro”, Bari, Italy c Dipartimento di Scienze biomolecolari e Biotecnologie, Università degli Studi di Milano, Via Celoria 26, 20133 Milano, Italy d AB SCIEX Viale Lombardia 218, 20861 Brugherio, MB, Italy e Medical Genetics, University of Foggia, Foggia, Italy b

a r t i c l e

i n f o

Article history: Received 6 February 2012 Received in revised form 16 April 2012 Accepted 17 April 2012 Available online 26 April 2012 Keywords: Therapeutic drug monitoring Psychotropic drugs Poly-pharmacy LC–MS/MS Protein precipitation sMRM

a b s t r a c t A simple liquid chromatographic tandem mass spectrometry (LC–MS/MS) method has been developed for simultaneous analysis of 17 basic and one acid psychotropic drugs in human plasma. The method relies on a protein precipitation step for sample preparation and offers high sensitivity, wide linearity without interferences from endogenous matrix components. Chromatography was run on a reversed-phase column with an acetonitrile–H2 O mixture. The quantification of target compounds was performed in multiple reaction monitoring (MRM) and by switching the ionization polarity within the analytical run. A further sensitivity increase was obtained by implementing the functionality “scheduled multiple reaction monitoring” (sMRM) offered by the recent version of the software package managing the instrument. The overall injection interval was less than 5.5 min. Regression coefficients of the calibration curves and limits of quantification (LOQ) showed a good coverage of over-therapeutic, therapeutic and subtherapeutic ranges. Recovery rates, measured as percentage of recovery of spiked plasma samples, were ≥94%. Precision and accuracy data have been satisfactory for a therapeutic drug monitoring (TDM) service as for managing plasma samples from patients receiving psycho-pharmacological treatment. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The variation in individual clinical response to psychotropic drugs treatment is still a critical problem in the management of serious mentally ill patients [1]. Recent studies have demonstrated that there are significantly different responses among individuals on exposure to a particular drug [2,3] due to pharmaco-kinetic [3] and pharmaco-dynamic variations. Therefore, patient- and treatment-related variables are relevant: the enzyme activity may decrease with age, may be modified by renal and hepatic diseases and may be enhanced by concurrent psychotropic or non-psychotropic drugs and smoking (by induction).

∗ Corresponding author at: Atherosclerosis and Thrombosis Unit, I.R.C.C.S. “Casa Sollievo della Sofferenza”, Poliambulatorio “Giovanni Paolo II”, Viale Padre Pio, S. Giovanni Rotondo, Foggia, Italy. Tel.: +39 0882 416273; fax: +39 0882 416273. E-mail address: [email protected] (E. Grandone). 0731-7085/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jpba.2012.04.016

In addition, other co-medications and/or some food intake may decrease the enzyme activity. It is well known that incidence of side effects during psychotropic therapy is often dose-related, and similar correlations have been found between plasma levels and therapeutic effects, at least for some psychotropic drugs [4]. Furthermore, co-administration of different psychotropic drugs can also lead to saturation of the Cytochrome P (CYP) enzyme or to some changes in its activity, especially when the same CYP subtype is involved, resulting in unexpected and possibly dangerous pharmacological interactions [5]. Therapeutic drug monitoring (TDM) of psychotropic drugs may be clinically advantageous in order to assess the metabolic capacity of patients treated with antidepressants [6], antipsychotics and mood stabilizers for the design of dosing regimens either for optimizing the therapy [7] and for avoiding medical complications, intoxication, non-responsiveness or non-compliance. Although the chemical structure of psychotropic drugs (Fig. 1) differs markedly as do their pharmaco-kinetics and metabolic fate, most psychotropic drugs are similar in chemical properties such as high lipophylicity, molecular weight range (between 200 and

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

500 Da) and basicity, with the exception of valproic acid, an anticonvulsant used also as a mood stabilizer. For monitoring a large panel of psychotropic drugs, the opportunity of a simultaneous quantification is very attractive also in terms of practical aspects and labor time. Although some methods have been published using gas chromatography [8,9], measurements have been mainly performed by HPLC coupled with UV detection [10–12], and with mass spectrometry MS [13] and MS–MS [14,15]. For mass spectrometry, electrospray ionization in the positive mode (ESI+) has been mainly used, more than electrospray ionization in negative mode (ESI−) or atmospheric pressure chemical ionization [16]. A development of a LC–MS/MS working in both ESI+ and ESI− should be desirable for covering in a single run either basic or acid compounds. On another side, measurements by ESI

105

have been reported to be affected by matrix effects when dealing with biological matrices such as plasma [17,18]. A sample cleanup is therefore mandatory in order to reduce this undesirable effect which can lead to unpredictable alterations of the MS signal and wrong results. Most of the published methods are targeting single compounds, sometimes accompanied by their related metabolites [19,20]. Simultaneous measurements of various psychotropic drugs have also been published [21], mostly clozapine in combination with olanzapine [22–24], or the antidepressant venlafaxine with other drugs of the same therapeutic class [25,26].Kirchherr et al. developed a multi-level, single-sample approach for the measurement of 48 antidepressants and antipsychotics [26] not including any acid drug. With the instrumentation available at the authors’ center, a protocol has been devised for speeding up

Fig. 1. Chemical structures of the investigated analytes. Captions refer to their classes: “a” typical antipsychotics; “b” atypical antipsychotics; “c2 tricyclics; “d” selective serotonine reuptake inhibitors; “e” selective serotonine/noradrenaline reuptake inhibitors; “f” mood stabilizer; “g” anxiolytic.

106

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

Fig. 1. (Continued)

the sample preparation and for extending the panel to some acid drugs.

2. Experimental 2.1. Chemical and reagents Chemical standards of psychotropic drugs (clozapine, haloperidol, risperidone, clomipramine, imipramine, fluvoxamine,

paroxetine, sertraline, ziprasidone, venlafaxine, valproic acid; structures in Fig. 1) and clonidine, used as internal standard (IS) for basic analytes, were purchased from Sigma (Taufkirchen, Germany). Other antidepressant and antipsychotic standards were supplied by different companies: lorazepam (Dorom, S.r.l., Milano, Italy), aripiprazole (Otsuka/Bristol, MyersSquibb, Munich, Germany), escitalopram (Innova Pharma, Recordati S.p.A., Milano, Italy), olanzapine (Lilly, IN, USA), quetiapine (Astra-Zeneca, Wedel, Germany), duloxetine (Eli Lilly, Italy), clotiapine (Novartis Farma S.p.A., Origgio, Italy).

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

Docosahexaenoic acid (DHA), used as internal standard for the acidic analyte, was supplied by Sigma (Milano, Italy). Acetonitrile, ethanol, methanol and acetic acid, all HPLC grade, were purchased from VWR/Merck (Darmstadt, Germany). Purified water was obtained through a Milli-Q system (Millipore). 2.2. Working solutions Each chemical standard was firstly dissolved for getting individual stock solutions at 1 mg/mL, which were stored at −20 ◦ C in the dark. A working internal standard (IS) solution containing clonitine at 5 ␮g/mL and DHA at 20 ␮g/mL was prepared in methanol. The working standard solution was further ten-fold diluted with acetonitrile in order to get the final protein precipitation solution as methanol:acetonitrile mixture (1:9, v/v). A working standard solution containing all the compounds at 100 ␮g/mL (valproic acid at 1000 ␮g/mL) was also prepared. This solution (containing all the drugs) was used either for preparing the plasma calibration samples, or for spiking unknown samples at concentrations as below detailed. 2.3. Sample preparation Plasma samples collected from 10 drug-free volunteers in sodium citrate at 3.8%, centrifuged at 3000 rpm at room temperature (RT) for 10 min, were mixed together for building a plasma pool, stored at −20 ◦ C. Each subject gave written consent before participation. Prior to analysis, plasma pool aliquots were thawed for 30–45 min at room temperature and adequately mixed before the aliquot transfer. Two aliquots of 100 ␮L of plasma from each test sample, one aliquot spiked with 1.5 ␮g/mL of the standard mixture (15 ␮g/mL of valproic acid), were transferred into a vial and added by 300 ␮L of the protein precipitation working solution [26,27]. Vials were vortex-mixed and centrifuged for 5 min at 18,000 rpm. 100 ␮L of supernatant were added by 100 ␮L of the LCelution buffer B (see below) before the injection into the LC–MS/MS system. 2.4. Instrumentation Measurements were performed on a API 3000 Tandem Mass Spectrometer (AB Sciex, Toronto, Canada) equipped with a turboionspray source. Quantification was achieved by using either the multiple reaction monitoring (MRM) and the new option “scheduled multiple reaction monitoring” (sMRM). Source was operated by switching the polarity during the chromatographic run for getting two periods, the first one in positive ion mode lasting 2.58 min with an ESI potential of +4500 V and the second one in negative ion mode lasting 2.62 min with a voltage of −4500 V. Other operating parameters were: turbo spray gun temperature at 300 ◦ C, nebulizer gas (NEB) setting at 10, curtain gas (CUR) setting at 6 and collision gas (CAD) pressure at 8 mTorr (set at 3 as arbitrary units). The declustering potential (DP), collision energy (CE) and collision cell exit potential (CXP) were optimized for each drug. The dwell time of each analyte was set to 50 ms for the measurement in classical MRM mode. For measurement in sMRM mode, a target scan time of 0.5 s was selected with a tolerance RT window of 15 s. Other specific operating parameters are summarized in Table 1. LC separation was performed on a Chromolith Speed ROD C18 column (50 mm × 2 mm, particle size 5 ␮m) preceded by a Chromolith C18 guard cartridge (5 mm × 2 mm, 5 ␮m) (VWR/Merck, Darmstadt, Germany). Sample volume of 1 ␮L was injected into

107

a flow of 0.600 mL/min of a mixture of ACN (eluent A) and H2 O (eluent B), both added by 0.2% of acetic acid, according the following gradient scheme: an initial mixture of 10% A, increasing up to 46.5% in 3 min and then stepping to 90% where it is held for 1 min. Resulting overall injection interval was 5.2 min including the reequilibration time. Retention times of each drug and ISs are shown in Table 1. Data acquisition was performed by using Analyst 1.5.1 software (AB Sciex, Toronto Canada). Calibration curves for validating the method were obtained through the peak area ratios (analyte to IS) and by implementing a linear regression with 1/x weighting. As a test, some unknown plasma samples were measured either by interpolation from the calibration curves (external calibration), and by the standard additions method, in order to assess the matrix effect.

2.5. Method validation Validation runs were conducted on 3 separate days. Pooled human drug-free plasma has been used for preparing either the calibration standards or the control standards in separate sessions. Standards were prepared in triplicate at levels to cover both plasma therapeutic and sub-therapeutic ranges, 1500, 1250, 1000, 750, 500, 250, 150, 50, 25, 5 and 0.15 ng/mL. Three hundred microliters of each spiking solution were added to 100 ␮L of each standard. After vortex mixing and centrifugation, supernatants were stored at 4 ◦ C until measurement was performed. Accuracy (bias) and precision (repeatability and intermediate precision) were assessed under the same conditions.

2.6. Specificity assessment Specificity evaluation was carried out on unspiked pooled plasma, processed by the same protein precipitation extraction protocol. Extent of interferences originated by endogenous plasma components at the specific retention time of each analyte and IS was evaluated through a comparison with the spiked pooled plasma aliquots.

2.7. Matrix effect evaluation Due to the detrimental impact of ion suppression and other matrix-related effects, it is imperative to evaluate them during a method validation. In this respect the use of stable isotopically labeled ISs should be recommended for correcting these adverse effects. However, in our case, ISs for the hereby-explored analytes are in part unavailable, very expensive, and/or with a questionable isotopic purity. Therefore the protocol hereby presented relies on one IS for each chromatographic period for covering all the analytes measured within that period. The overall matrix effect was determined by comparing signals from pooled plasma samples with the ones coming from aqueous solutions, both spiked with the lowest quantifiable analyte concentration.

2.8. Recovery assessment The extraction recovery was determined by comparing peak area from pooled blank plasma samples spiked at different concentrations both before and after the extraction.

108

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

Table 1 Mass Spectrometer parameters and retention times for each analyte. Q1 (m/z)

Q3 (m/z)

Low level drugs Haloperidol Risperidone Lorazepam Duloxetine

376.2 411.4 321.1 298.2

165 191.1 275.1 154.1

Medium level drugs Aripiprazole Escitalopram Fluvoxamine Imipramine Olanzapine Paroxetine Quetiapine Sertraline Ziprasidone

448.2 325.3 319.2 281.2 313.3 330.2 384.2 306.2 413.3

High level drugs Clozapine Venlafaxine Clomipramine Clotiapine Valproic acid Internal standard Clonidine DHA

Dwell time

DP (V)

FP (V)

EP (V)

CE (V)

50 50 50 50

60 75 85 20

50 100 100 150

10 10 10 9

40 37 31 10

285.1 262.2 71.2 86.2 256.2 192.1 253.2 275 194.1

50 50 50 50 50 50 50 50 50

80 80 30 45 80 65 60 10 80

300 100 210 50 250 300 200 75 240

10 10 10 10 10 11 10 10 10

40 29 30 21.5 37 30 35 19.5 39

327.2 278.196 315.4 344.3 143.1

270.1 58.2 86.1 287.1 143.1

50 50 50 50 500

80 31 45 65 −50

330 200 150 50 −100

10 10 10 10 −10

230 327.188

44.4 283.2

50 500

50 −101

100 −280

10 −10

2.9. Lower limit of quantification based on signal to noise ratio (S/N) estimation Serial dilutions of working solutions of analytes from 1500 to 0.15 ng/mL were accurately prepared. Concentration ranges for valproic acid were 100 times higher than other analytes, since its higher therapeutic range (50–150 ␮g/mL).

2.10. Evaluation of limit of detection A further dilution was carried out for getting a concentration of 0.05 ng/mL (1/3 of the estimated LLOQ) in order to evaluate the LOD, expected at 3 times lower than LLOQ.

3. Results and discussion 3.1. Method development and optimization 3.1.1. MS/MS measurement optimization Method development started with the tuning of MS parameters in both positive and negative ionization modes according to physico-chemical properties of each compound. All the drugs, except valproic acid and its IS DHA, are basic; therefore the mass spectrometer was operated in positive ion mode. Use of acetic acid (0.2%) in the LC mobile phase has demonstrated to enhance sensitivity. Basic analytes and clonidine showed the singly charged protonated precursor [M + H]+ as the prominent ion in the full scan single-MS spectra. Fragmentation optimization for all precursor ions led to the most abundant product ion as detailed in Table 1. Acidic nature of valproic acid made its sensitivity higher in negative ion mode. Since the lack of any exploitable fragmentation ion, MRM transition was tuned on the precursor ion itself with minimal collision energy for achieving a sensitivity level comparable with the other compounds, nonetheless with a higher background due to a reduced selectivity offered by a precursor–precursor transition.

CXP (V)

Rt (min)

Internal standard

10 12 11 14

1.9 1.4 2.5 2.2

Clonidine Clonidine Clonidine Clonidine

9 11 11 13 20 14 12 10 12

2.2 1.9 2.1 2.1 0.6 2.1 1.7 2.3 1.8

Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine Clonidine

35 47 26 32 −5

15 10 9 12 −12

1.6 1.4 2.4 2.1 2.8

Clonidine Clonidine Clonidine Clonidine DHA

47 −20

9 −7

0.5 3.9

3.1.2. Liquid chromatography optimization Chromatographic conditions were optimized through several steps in order to achieve good chromatographic resolution, symmetric analyte peak shapes within a short run time. A multiple extracted ion current (XIC) chromatogram of the 18 investigated compounds is presented in Fig. 2: all peaks are resolved within a 5.2 min chromatographic run. The low concentration (0.2%) of acetic improved either the sensitivity by promoting the analyte ionization and the chromatographic peak shape. Under these conditions, traces of each analyte and IS did not show any isobaric interference due to matrix components. Choice on ISs, clonidine for the analytes measured in positive ion mode and DHA for the negative ion mode operation, demonstrated to fit the purpose since comparable with the targeted analytes in terms of molecular structure, extraction efficiency, chromatographic behavior and ionization yield. The strategy secured the results consistency from any impact generated by slight changes in the LC-eluent composition/flow rate or inaccuracies in the upstream sample preparation step.

3.1.3. Data acquisition strategy MRM assay enables multiplexing measurements (multiple analytes measured along one single analytical run) and is regarded as the most performing in terms of sensitivity. In this frame, attainable detection limits are dependant on the number of transitions simultaneously monitored being the cycle time split between the monitored MRM channels. In those cases where a large number of transitions are required to be monitored in a single chromatographic period, some instruments embed the functionality, called scheduled MRMs (sMRM), by which just the analytes eluting in the vicinity of the expected retention time share the allocated cycle time. The software manages as well the partial overlapping of the time windows in order to provide undisrupted chromatographic traces. Benefit of this functionality resides on the fact that the cycle time is now shared by a significantly smaller number of transitions monitored at that single time with an improvement on S/N ratio for each analyte.

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

109

Fig. 2. (a) Extracted ion current (XIC) chromatogram of 17 drugs plus their IS clonidine, eluting in the first chromatographic period in positive MRM mode; (b) XIC chromatogram of valproic acid plus its IS DHA, eluting in the second chromatographic period in negative MRM mode; (c) three dimensional XIC containing all 18 drugs and their two IS.

3.2. Method validation 3.2.1. Specificity Dealing with biological fluids, sample pre-treatment is needed for removing proteins and other potential interfering components prior the LC–MS/MS analysis. Clean samples are essential not only for minimizing ion suppression and matrix effect, but also for

preventing any possible chromatographic column deterioration. Most of the published procedures [26,27] have employed solid phase extraction or liquid–liquid extraction for achieving good recovery on plasma samples. Since protein precipitation (PPT) procedure is accounted for simplicity and universality, it has been adopted for the herebydescribed assay by using a methanol:acetonitrile mixture (1:9, v/v).

110

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

Table 2 Linearity and LLOQ performances on spiked plasma samples. Concentration range (ng/mL)

Intra-day correlation coefficient (r2 )

Inter-day correlation coefficient (r2 )

Therapeutic range (ng/ml)

S/N on 0.15 ng/ml spike (LLOQ)

S/N on 0.05 ng/ml spike (LOD)

Low level drugs Haloperidol Risperidone Clotiapine Lorazepam

1–20 1–50 1–200 1–50

0.9893 0.9640 0.9769 0.9986

0.9789 0.9769 0.9762 0.9876

5–17 20–60 10–160 10–15

38.1 24.5 50.8 163.7

11.2 7.6 16.0 42.3

Medium level drugs Aripiprazole Escitalopram Fluvoxamine Imipramine Olanzapine Paroxetine Quetiapine Sertraline Ziprasidone

10–1000 10–1000 10–1000 10–1000 10–1000 5–500 10–1000 5–500 10–1000

0.9926 0.9909 0.9924 0.9946 0.9826 0.9890 0.9880 0.9876 0.9876

0.9838 0.9859 0.9812 0.9860 0.9929 0.9865 0.9908 0.9829 0.9757

50–350 15–80 150–300 175–300 20–80 70–120 70–170 10–50 50–120

50.1 46.2 16.5 26 36.7 54.6 467.8 135.2 41.6

15.6 14.8 5.2 8.1 11.9 17.3 122.3 36 11.2

High level drugs Clozapine Venlafaxine Clomipramine Valproic acid Duloxetine

100–1000 100–1000 100–1000 10–1000 2000

0.9998

0.9895

0.9902 0.9864 0.9875

0.9911 0.8744 0.9945

350–600 195–400 175–450 50–100 ␮g/ml 20–80

34.7 190.2 76.1 18.4 ␮g/ml 175.8

9.7 60.1 22.0 5.9 ␮g/ml 52

This mixture has been shown to be effective in protein precipitation [26]. Any possible impact by incomplete protein precipitation, like column deterioration, is minimized by reducing the injection volume (1 ␮L of supernatant representing 0.25 ␮L of serum). Efficacy of pre-treatment has been verified through the absence of any detectable interference in the respective MRM channel at the retention time of each analyte when injecting different plasma samples. No distortion in chromatographic peak shape of both analytes and ISs allowed the conclusion that interferences were at least undetectable under the described experimental conditions. The same experiments enabled the carry-over estimation: injection of blank plasma after the highest validation standard showed a carryover effect lower than 0.1%. 3.2.2. Matrix effect evaluation The ion suppression caused by the plasma matrix has been extensively evaluated. The performances of the API-interface in combination with the described sample treatment demonstrated to be effective in controlling any possible ion suppression. A further contribution has been given by the optimization of the LC flow-rate at 250 ␮L/min, making negligible any possible interference generated by co-eluting matrix components. This was confirmed by the post-column addition test with the infusion of a mixture of standards and ISs while injecting plasma sample extracts: no significant MRM trace distortion was evidenced at the retention time of each analyte and IS. 3.2.3. Recovery The PPT performed with a mixture methanol:acetonitrile (1:9, v/v) as proposed by Kirchherr et al. [26] allowed us to obtain good recovery rates. Other methods of protein precipitation have shown lower yield [27]. Despite solid phase extraction and liquid–liquid extraction are proposed as alternative options for sample treatment [28–31], findings of present study have supported the choice on the described PPT procedure as the best one in terms of yield and simplicity. 3.2.4. Calibration To evaluate the linearity, calibration curves were produced in triplicate in three different days.

Linearity of each calibration was determined by plotting the peak-area ratio (y) of psychoactive drugs to IS versus the nominal concentration (x) of the calibration points. Data were obtained by a weighted (1/x) linear regression analysis. Good linearity was observed over the concentration range of 5–1500 ng/mL for all the drugs in each calibration set and no significant changes were found in the values of slope, intercept and correlation coefficient on both inter- and intra-day assays. Correlation coefficients were in the range 0.99 < r2 < 0.97 for all the compounds and for concentrations spanning between 5 and 1500 ng/mL (Table 2). Linearity demonstrated to be assured even beyond the higher concentration levels (data not shown) enabling the method to be exploited for intoxication cases, where concentrations far above the therapeutic ranges could be expected. LLOQ resulted in 0.15 ng/mL for all drugs but valproic acid for which a LLOQ of 15 ng/mL was obtained. The higher LLOQ for valproic acid collimates with its highest plasma levels and therapeutic range of the psychotropic drugs panel (Table 2). LOD was estimated in 0.05 ng/mL for all drugs and in 0.5 ng/mL for valproic acid.

3.2.5. Accuracy (bias) Accuracy was calculated according to the formula:

Accuracy % =

[Analyte]M − [Analyte]added × 100 [Analyte]added

where: [Analyte]added = effective concentration added to samples and [Analyte]M = ([Analyte]1 + [Analyte]2 + [Analyte]3 )/3. It was assessed at three different concentrations according to the type of drugs (low, medium and high concentration drugs) and by calculating three replicates for each concentration level in the same assay (intra-assay). Inter-assay precision and accuracy were obtained by running the same samples on three different days (Table 3).

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

111

Table 3 Precision and accuracy data on spiked plasma samples. Concentration added (ng/ml)

Low level drugs Haloperidol

Risperidone

Clotiapine

Lorazepam

Medium level drugs Aripiprazole

Escitalopram

Fluvoxamine

Imipramine

Olanzapine

Paroxetine

Quetiapine

Sertraline

Ziprasidone

High level drugs Clozapine

Venlafaxine

Clomipramine

Duloxetine

Valproic acid

Imprecision (CV%, n = 6)

Inaccuracy (DEV%, n = 6)

Intra-assay

Inter-assay

1.9 9.3 11.2 1.6 8.9 9.9 3.1 14.4 13.8 2.1 4.7 2.1

1.0 3.9 7.8 1.1 4.7 9.0 3.9 5.2 11.2 1.3 5.1 7.0

1.1 −5.0 13.2 −0.4 0.3 11.2 −2.2 12.2 1.5 −2.0 9.7 −9.1

0.3 −1.9 2.2 0.6 −4.0 9.3 1 −9.3 3.2 −1.7 9.2 −4.7

100 50 10 100 50 10 100 50 10 100 50 10 100 50 10 100 50 10 100 50 10 100 50 10 100 50 10

3.8 4.8 11.3 5.6 8.7 11.8 8.6 3.3 13.9 0.3 0.9 6.5 3.6 4.0 13.5 1.95 2.5 13.4 1.5 3.7 13.8 2.9 2.8 12.6 2.2 3.2 5.9

1.3 7.2 15.6 1.5 9.2 10.7 1.1 4.0 10.1 0.8 2.3 13.3 2.5 11.8 14.1 1.2 6.1 14.8 1.8 6.1 8.2 3.2 11.1 14.6 1.2 2.1 2.8

1.5 2.9 −12.6 −4.2 12.1 −13.8 −6.0 6.8 14.5 0.5 3.2 −9.8 −5.0 13.1 −11.0 −1.8 7.8 −12.3 −1.3 2.0 5.6 2.0 7.5 −13.7 −2.2 6.4 −11.1

−0.4 3.1 −0.5 0.9 0.8 −2.1 −1.1 2.5 −3.2 1.6 4.6 −10.6 1 3.8 4.2 −0.9 1.9 −7.3 −1.0 4.2 −2.0 1.1 8.3 −2.5 −0.9 3.4 −5.5

100 500 1000 100 500 1000 100 500 1000 100 500 1000 1000 5000 10,000

0.9 5.1 1.5 1.5 4.1 2.7 0.9 9.1 1.2 0.10 1. 4 0.2 2.2 5.0 8.5

1.2 6.2 6.3 0.6 0.7 2.2 3.9 6.5 1.5 1.1 2.0 2.0 1.7 4.1 11.5

10 5 1 10 5 1 10 5 1 10 5 1

3.2.6. Precision As shown by Table 3, repeatability and intermediate precision were within the acceptance criteria for the evaluated assay range. We applied the formula:

Coefficient of variation % =

SD × 100 [Analyte]M

where SD is the standard deviation.

Intra-assay

2.3 11.5 −7.9 −2.8 6.0 −9.8 3.1 10.5 −2.8 2.0 −0.7 −1.55 0.6 7.3 6.4

Inter-assay

3.1 9.2 −4.6 −3.2 9.1 −9.2 2.9 12.9 −6.2 1.2 0.5 −3.1 2.5 14.9 9.7

3.3. Application to plasma samples from psychiatric patients After validation, the method was exploited for the analysis of patients’ plasma samples suffering psychiatric disorders and referred to the Psychiatric Department of University of Bari “A. Moro”. The patients were administered with two or more psychotropic drugs. Fig. 3 shows the traces related to a patient undergoing a poly-pharmacy treatment with two different antipsychotic drugs, quetiapine and valproic acid.

112

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

Fig. 3. Chromatograms of plasma samples from one polymedicated patient with two different antipsychotics quietapine and valproic acid.

This plasma sample has been measured in both MRM and sMRM modes in order to appreciate the real sensitivity improvement of the latter. Peak shapes and chromatographic resolution were comparable with those displayed by spiked control plasma, and no interferences were revealed. 4. Conclusions As it is widely recognized as a sensitive and specific technique for analysis of pharmaceutical drugs and for studies of drug metabolism [32], LC–MS/MS has proved to be a compelling tool in the development of the present method for a simultaneous determination of basic and acid psycho-tropic drugs. The short analytical time in combination with a simple sample preparation offer the high throughput as required by our clinical studies. Validation data have assessed its sensitivity, its immunity from matrix effects, and good LLOQ, the latter comparing well with the ones of a similar method as previously devised by Kirchherr et al. [26]. The protocol proved to be easily implementable for a daily practice in clinical laboratories. References [1] A.K. Malhotra, G.M. Murphy, J.L. Kennedy, Pharmacogenetics of psychotropic drug response, Am. J. Psychiatry 161 (2004) 780–796. [2] K. Titier, S. Bouchet, F. Péhourcq, N. Moore, M. Molimard, High-performance liquid chromatographic method with diode array detection to identify and quantify atypical antipsychotics and haloperidol in plasma after overdose, J. Chromatogr. B: Analyt. Technol. Biomed. Life Sci. 788 (2003) 179–185. [3] Y. Higashi, M. Kitahara, Y. Fujii, Simultaneous analysis of haloperidol, its three metabolites and two other butyrophenone-type neuroleptics by high performance liquid chromatography with dual ultraviolet detection, Biomed. Chromatogr. 20 (2006) 166–172. [4] M.A. Raggi, Therapeutic drug monitoring: chemical–clinical correlations of atypical antipsychotic drugs, Curr. Med. Chem. 9 (2002) 1397–1409. [5] F. Bugamelli, R. Mandrioli, E. Kenndler, C. Bartoletti, G. Boncompagni, M.A. Raggi, Possible levomepromazine–clozapine interaction: two case reports, Prog. Neuropsychopharmacol. 31 (2007) 567–570. [6] J. Kirchheiner, K. Brøsen, M.L. Dahl, L.F. Gram, S. Kasper, I. Roots, F. Sjöqvist, E. Spina, J. Brockmöller, CYP2D6 and CYP2C19 genotype-based dose

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

recommendations for antidepressants: a first step towards subpopulationspecific dosages, J. Acta Psychiatr. Scand. 104 (2001) 173–192. M.A. Raggi, R. Mandrioli, C. Sabbioni, V. Pucci, Atypical antipsychotics: pharmacokinetics, therapeutic drug monitoring and pharmacological interactions, Curr. Med. Chem. 11 (2004) 279–296. S. Ulrich, Assay of olanzapine in human plasma by a rapid and sensitive gas chromatography-nitrogen phosphorus selective detection (GC-NPD) method: validation and comparison with high-performance liquid chromatographycoulometric detection, Ther. Drug Monit. 27 (2005) 463–468. H.C. Huang, C.H. Liu, T.H. Lan, T.M. Hu, H.J. Chiu, Y.C. Wu, Y.L. Tseng, Detection and quantification of aripiprazole and its metabolite, dehydroaripiprazole, by gas chromatography–mass spectrometry in blood samples of psychiatric patients, J. Chromatogr. B 856 (2007) 57–61. L. Mercolini, R. Mandrioli, R. Cazzolla, M. Amore, M.A. Raggi, HPLC analysis of the novel antidepressant duloxetine in human plasma after an original solid-phase extraction procedure, J. Chromatogr. B 856 (2007) 81–87. C. Patel, M. Patel, S. Rani, M. Nivsarkar, H. Padh, A new high performance liquid chromatographic method for quantification of atomoxetine in human plasma and its application for pharmacokinetic study, J. Chromatogr. B 850 (2007) 356–360. Y. Shimokawa, H. Akiyama, E. Kashiyama, T. Koga, G. Miyamoto, High performance liquid chromatographic methods for the determination of aripiprazole with ultraviolet detection in rat plasma and brain: application to the pharmacokinetic study, J. Chromatogr. B 821 (2005) 8–14. W. Liu, F. Wang, H.D. Li, Simultaneous stereoselective analysis of venlafaxine and O-desmethylvenlafaxine enantiomers in human plasma by HPLC–ESI/MS using a vancomycin chiral column, J. Chromatogr. B 850 (2007) 183–189. M. Kubo, Y. Mizooku, Y. Hirao, T. Osumi, Development and validation of an LC–MS/MS method for the quantitative determination of aripiprazole and its main metabolite, OPC-14857, in human plasma, J. Chromatogr. B 822 (2005) 294–299. P. Senthamil Selvan, K.V. Gowda, U. Mandal, W.D. Sam Solomon, T.K. Pal, Determination of duloxetine in human plasma by liquid chromatography with atmospheric pressure ionization-tandem mass spectrometry and its application to pharmacokinetic study, J. Chromatogr. B 858 (2007) 269–275. M.J. Bogusz, K.D. Krüger, R.D. Maier, R. Erkwoh, F. Tuchtenhagen, Monitoring of olanzapine in serum by liquid chromatography–atmospheric pressure chemical ionization mass spectrometry, J. Chromatogr. B 732 (1999) 257–269. P.J. Taylor, Matrix effects: the Achilles heel of quantitative high-performance liquid chromatography–electrospray-tandem mass spectrometry, Clin. Biochem. 38 (2005) 328–334. W. Xie, J. Pawliszyn, W.M. Mullett, B.K. Matuszewski, Comparison of solidphase microextraction and liquid–liquid extraction in 96-well format for the determination of a drug compound in human plasma by liquid chromatography with tandem mass spectrometric detection, J. Pharm. Biomed. Anal. 45 (2007) 599–608. J. Bhatt, A. Jangid, G. Venkatesh, G. Subbaiah, S. Singh, Liquid chromatography–tandem mass spectrometry (LC–MS–MS) method for simultaneous determination of venlafaxine and its active metabolite Odesmethyl venlafaxine in human plasma, J. Chromatogr. B 829 (2005) 75–81. H.A. Niederländer, E.H. Koster, M.J. Hilhorst, H.J. Metting, M. Eilders, B. Ooms, G.J. de Jong, High throughput therapeutic drug monitoring of clozapine and metabolites in serum by on-line coupling of solid phase extraction with liquid chromatography–mass spectrometry, J. Chromatogr. B 834 (2006) 98–107. G. Zhang, A.V. Terry Jr., M.G. Bartlett, Liquid chromatography/tandem mass spectrometry method for the simultaneous determination of olanzapine, risperidone, 9-hydroxyrisperidone, clozapine, haloperidol and ziprasidone in rat plasma, Rapid Commun. Mass Spectrom. 21 (2007) 920–928. K. Titier, S. Bouchet, F. Péhourcq, N. Moore, M. Molimard, High-performance liquid chromatographic method with diode array detection to identify and quantify atypical antipsychotics and haloperidol in plasma after overdose, J. Chromatogr. B 788 (2003) 179–185. Z. Zhou, X. Li, K. Li, Z. Xie, Z. Cheng, W. Peng, F. Wang, R. Zhu, H. Li, Simultaneous determination of clozapine, olanzapine, risperidone and quetiapine in plasma by high-performance liquid chromatography–electrospray ionization mass spectrometry, J. Chromatogr. B 802 (2004) 257–262. H. Juan, Z. Zhiling, L. Huande, Simultaneous determination of fluoxetine, citalopram, paroxetine, venlafaxine in plasma by high performance liquid chromatography–electrospray ionization mass spectrometry (HPLC–MS/ESI), J. Chromatogr. B 820 (2005) 33–39. C. Frahnert, M.L. Rao, K. Grasmader, Analysis of eighteen antidepressants, four atypical antipsychotics and active metabolites in serum by liquid chromatography: a simple tool for therapeutic drug monitoring, J. Chromatogr. B 794 (2003) 35–47. H. Kirchherr, W.N. Kuhn-Velten, Quantitative determination of forty-eight antidepressants and antipsychotics in human serum by HPLC tandem mass spectrometry: a multi-level, single-sample approach, J. Chromatogr. B 843 (2006) 100–113. S. Souverain, S. Rudaz, J.L. Veuthey, Protein precipitation for the analysis of a drug cocktail in plasma by LC–ESI-MS, J. Pharm. Biomed. Anal. 35 (2004) 913–920.

G. Vecchione et al. / Journal of Pharmaceutical and Biomedical Analysis 67–68 (2012) 104–113 [28] J. Bhatt, A. Jangid, G. Venkatesh, G. Subbaiah, S. Singh, Liquid chromatography–tandem mass spectrometry (LC–MS-MS) method for simultaneous determination of venlafaxine and its active metabolite Odesmethyl venlafaxine in human plasma, J. Chromatogr. B 829 (2005) 75–81. [29] R. Mandrioli, L. Mercolini, R. Cesta, S. Fanali, M. Amore, M.A. Raggi, Analysis of the second generation antidepressant venlafaxine and its main active metabolite O-desmethylvenlafaxine in human plasma by HPLC with spectrofluorimetric detection, J. Chromatogr. B 856 (2007) 88–94. [30] W. Liu, H.L. Cai, H.D. Li, High performance liquid chromatography–electrospray ionization mass spectrometry (HPLC–MS/ESI) method for simultaneous

113

determination of venlafaxine and its three metabolites in human plasma, J. Chromatogr. B 850 (2007) 405–411. [31] P. Baumann, C. Hiemke, S. Ulrich, G. Eckermann, I. Gaertner, M. Gerlach, H.J. Kuss, G. Laux, B. Müller-Oerlinghausen, M.L. Rao, P. Riederer, Zernig G, The AGNP-TDM expert group consensus guidelines: therapeutic drug monitoring in psychiatry, Pharmacopsychiatry 37 (2004) 243–265. [32] E.J. Oliveira, D.G. Watson, Liquid chromatography–mass spectrometry in the study of the metabolism of drugs and other xenobiotics, Biomed. Chromatogr. 14 (2000) 351–372.