Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journal homepage: www.elsevier.com/locate/jpba

Determination of some psychotropic drugs in serum and saliva

samples by HPLC-DAD and HPLC MS
A. Petruczynik a , K. Wróblewski a , M. Szultka-Młyńska b , B. Buszewski b ,
H. Karakuła-Juchnowicz c , J. Gajewski c , J. Morylowska-Topolska c ,
M. Waksmundzka-Hajnos a,∗
a
Department of Inorganic Chemistry Medical University of Lublin, Chodźki 4a 20-093 Lublin, Poland
b
Department of Environmental Chemistry and Bioanalytics, Nicolaus Copernicus University, Faculty of Chemistry Gagarina 7, PL-87-100 Torun, Poland
c
Department of Clinical Neuropsychiatry, Medical University of Lublin, Głuska 2, 20-439 Lublin, Poland

a r t i c l e i n f o a b s t r a c t

Article history: A simple, rapid and sensitive HPLC-DAD method has been developed and validated for the simultaneous
Received 20 October 2015 determination of seven psychotropic drugs (risperidone, citalopram, clozapine,quetiapine, levomepro-
Received in revised form mazine, perazine and aripiprazole) in human serum or saliva samples. The chromatographic analyses
29 December 2015
were performed on a XSELECT CSH Phenyl-Hexyl column with a mobile phase containing methanol,
Accepted 1 January 2016
acetate buffer at pH 3.5 and 0.025 mL−1 diethylamine. The influence of concentration of methanol in
Available online 8 January 2016
injection samples and injection volume on peak symmetry and system efficiency was examined.The full
separation of all investigated drugs, good peaks’ symmetry and simultaneously high systems efficiency
Keywords:
Psychotropic drugs
were obtained in applied chromatographic system. The method is suitable for the analysis of investigated
Human serum drugs in human plasma or saliva for psychiatric patients for control of pharmacotherapy, particularly in
Human saliva combination therapy. HPLC–MS was applied for verification of the presence of drugs and their metabolites
HPLC–DAD in serum and saliva samples from patients.
HPLC–MS © 2016 Elsevier B.V. All rights reserved.

1. Introduction The general methods for analyzing psychotropic drugs in dif-

ferent biological fluids are based on a combination of efficient
Therapeutic drug monitoring is a useful tool for the clinical man- separation method with a sensitive detection technique. At present,
agement of patients receiving a pharmacotherapy, particularly in numerous separation techniques, including high-performance liq-
psychiatry. Therapeutic drug monitoring of antidepressants is nec- uid chromatography (HPLC) [1], gas chromatography (GC) [2,3] and
essary for an optimal supervision of patient drug regimen to avoid capillary electrophoresis (CE) [4,5], have been employed for the
medical complications, intoxication, non responsiveness or non- analysis of psychotropic drugs. Among those methods, HPLC has
compliance. A significant percentage of psychiatric patients who been considered as the most efficient and robust specific technique
are treated with psychotropic drugs are treated with more than to some advantages including convenience, high separation ability
one drug. Thus, it is advantageous to use a rapid and reliable assay and simple operation.
that is suitable for determination of multiple antipsychotic drugs in Various detection methods have also been applied to accurate
biological samples in a single run. The concentrations of the drugs determination of psychotropic drugs in different samples such as:
in biological samples are often very low. For these reasons, to mon- voltammetry, chemiluminescence [6], fluorescence [7], UV adsorp-
itor psychoactive drugs, the analytical methods have to be highly tion [8–11], and mass spectrometry [12–20]. Nowadays most HPLC
sensitive and selective for accurate and precise quantification. methods are developed with mass spectrometric detection tech-
niques.
Psychotropic drugs were quantified in different biological sam-
ples e.g., in plasma [7,11,21,22], serum [10,13,23,24], whole blood
[16], urine [1,7], saliva [3,25].
∗ Corresponding author. Fax: +48 81 5357378. Most of the published methods allow quantification of a single
E-mail addresses: [email protected], [email protected] drug, sometimes with its related metabolites [6,11,18]. Simulta-
(M. Waksmundzka-Hajnos).

http://dx.doi.org/10.1016/j.jpba.2016.01.004
0731-7085/© 2016 Elsevier B.V. All rights reserved.
 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80 69

neous quantification of various psychotropic drugs have also been Hungary), PROTERAPIA Sp z o.o. (Warszawa, Poland), HASCO-LEK
published [1,9,10], mostly in the same therapeutic class. In recent S.A. (Wrocław, Poland), Ranbaxy Sp. z o. o. (Warszawa, Poland),
years there is a trend in clinical and forensic toxicology toward Otsuka Pharmaceutical Europe Ltd. (Wexham, Great Britain).
simultaneous quantification of a various compounds in one analyt- Methanol (MeOH) of chromatographic quality and diethylamine
ical run [16]. (DEA), acetic acid, sodium acetate, phosphoric acid, ammonium
In most cases psychotropic drugs separation and determi- (25%), ammonium chloride were purchased from Merck (Darm-
nation were performed on alkyl bonded stationary phase (C18, stadt, Germany). Water was double distilled.
rarely C8) [1,8,10,12,13,15,16,21]. Protonated basic psychotropic
drugs can interact with residual silanol groups of the stationary 2.2. Apparatus and HPLC-DAD conditions
phases. Thus, besides the reversed phase retention mechanism
also an ion-exchange retention mechanism occurs, which often There were used BAKERBONDTM speOctadecyl (C18) J.T. Baker
results in asymmetry of peaks, irreproducible retention, poor effi- (Phillipsburg, USA) cartridges (100 mg/1 mL) and SPE chamber –
ciency and worse separation. The silanol ion-exchange interaction Baker SPE – 12 G J.T. Baker (Philipsburg, USA) for sample prepara-
can be reduced by using mobile phase with buffer at low pH, tion.
when the silanol ionization is suppressed, mobile phase with Chromatographic analysis was performed using liquid chro-
buffers at high pH to suppress solutes ionization, addition of an matograph LaChrom Elite (Merck) equipped with an autosampler,
anionic ion-pairing reagents, making neutral associates or addi- column oven L-7350, solvent degasser L-7612 and DAD detector.
tion of organic amines as silanol blockers. Good results were also The chromatographic measurements were carried out at 22 ◦ C with
obtained by selecting a stationary phase to minimize the inter- an eluent flow rate of 1.0 mL/min. The chromatographic separa-
action between analyte and residual silanols. The introduction of tion was performed on Phenyl-Hexyl column from Phenomenex
hydrophobic ␲–␲ active aromatic moieties to the common n-alkyl (5 ␮m,150 mm × 4.6 mm). The following eluents were used: elu-
chain RP-sites generates a concerted ␲–␲ reversed-phase retention ent (A) H2 O and 0.025 mL−1 DEA, eluent (B) MeOH and 0.025 mL−1
mechanism, which diversivies the common RP-interaction prop- DEA, eluent (C) acetate buffer at pH 3.5 and 0.025 mL−1 DEA. Sam-
erties. Analysis of some psychotropic drugs was also successfully ples were injected onto HPLC column and eluted with following
performed on CN [26], Phenyl [27], CycloHexyl [28], Polar RP [16] gradient elution program: 40% B from 0 to 15 min, from 40 to 45% B
columns. from 15 to 35 min, from 45 to 65% B from 35 to 45 min. The concen-
Described in literature mobile phases applied to analysis of basic tration of the eluent C was constant and was 20% during the whole
psychotropic drugs often contained organic modifier and addition analysis.
of salts e.g., ammonium acetate [12], ammonium formate [17,18], The DAD detector was set in the 200–400 nm range. The chro-
acids [9,15,16,21], buffer at acidic pH [9,15,16,21,24], buffer at basic matographic data was acquired and processed with EZchrom Elite
pH [29]. Mobile phases with addition of ion-pairing reagents [7] or software.
amines e.g., triethylamine, tetramethylethylenediamine as silanol
blockers were also used [8,11].
In most cases, owing to complex matrix such as biological 2.3. HPLC–MS conditions
fluids interference and insufficient instrumental detection limit
for trace psychotropic drugs in real biological samples, direct Determination of target compounds was carried out using an
HPLC determination of those compounds is difficult.Therefore, HPLC system equipped with the Phenyl-Hexyl analytical column.
a separation from matrices and preconcentration is often The column was maintained at 20 ◦ C. The injected sample volume
required prior to chromatographic analysis of the analytes. Dif- was 20 ␮L, while the mobile phase was composed of MeOH, acetate
ferent sample preparation methods for analysis of psychotropic buffer at pH 3.5 and 0.025 mL−1 mixed and dosed at a flow rate of
drugs in various biological samples e.g., protein precipitation 0.6 mL/min.The mass spectral analysis was performed on a 6410
[14,16,20,21], liquid–liquid extraction (LLE) [8,9,17], dispersive triple quadrupole mass spectrometer from Agilent (Santa Clara,
liquid–liquid microextraction (DLLME) [1], Ultrasound-assisted USA) equipped with an ESI interface operating in positive ion mode,
emulsification microextraction [30], solid-phase extraction (SPE) with the following set of operation parameters: capillary voltage,
[11,12,18,24], stir bar sorptive extraction (SBSE) [7], some meth- 4000 V; nebulizer pressure, 40 psi; drying gas flow, 9 L/min; dry-
ods using on-line SPE with a column-switching system or ing gas temperature, 310 ◦ C; LC–MS mass spectra were recorded
methods with a simple protein precipitation have also been devel- across the range 50–1000 m/z. Quadrupole 1 was fixed at a set par-
oped. ent ion, quadrupole 2 was used as a collision chamber to induce
The aim of this study was to develop and validate a method fragmentation, and quadrupole 3 was fixed at a set daughter ion.
for the simultaneous quantification of some psychotropic drugs The HPLC–MS data were collected and processed by MassHunter
in serum and saliva. The HPLC-DAD method was fully validated software (Agilent). The data were further processed using Microsoft
including function response, linearity, limit of detection and quan- Excel.
tification, recovery, matrix effects, process efficiency, repeatability The instrument was operated in selected ions monitoring mode
and intermediate precision. The presence of some investigated (SIM) and multiple reaction monitoring (MRM) as well. The moni-
drugs in samples from patients receiving medication were con- tored pseudomolecular ions [M + H]+ are presented in Table 1 .
firmed by HPLC–MS.
2.4. Serum and saliva sample collection

2. Experimental Serum and saliva samples for SPE procedure optimization and
validation were taken from healthy volunteers. After blood coag-
2.1. Chemicals ulation the samples was centrifuged for 10 min at 1500 × g. The
serum was separated and stored at −20 ◦ C until analysis. The saliva
Drug standards: Risperidone, Quetiapine, Clozapine, Levome- was also stored at the same temperature.
promazine, Perazine, Citalopram were obtained respectively from Human body fluid samples were collected from psychiatric
Janssen-Cilag International NV (Beerse, Belgium), Adamed Sp. patients at Autonomous Public Clinical Hospital No. 1 in Lublin
z o.o. (Pieńków, Poland), EGIS Pharmaceuticals PLC (Budapest, (Poland). The study was approved by Bioethical Commission. The
 70
Table 1
MRM transitions and conditions for the measurement of target compounds.

Name of compound Abbreviation Chemical structures Precursorion m/z Production m/z Fragmentor [V] Collision energy [eV]

A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80
Risperidone R 411 191 105 29

Citalopram Ci 325 109 105 35

Clozapine Cl 327 270 135 22

Quetiapine Q 384 253 115 29

A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80
Levomepromazine L 329 100 105 35

Perazine P 340 141 135 15

Aripiprazole A 448 285 105 35

71
72 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80

Fig. 1. Chromatogram obtained for separation of psychotropic drug stnadards’ mix-

ture. Fig. 2. Chromatograms obtained for human serum spiked of 200 ng of seven psy-
chotropic drugs (1) and for blank human serum (2).  = 250 and 277 nm.

samples were collected two hours after drug administration and

next were prepared and stored as mentioned above.
tilled water, 20% (v/v) acetate buffer at pH. 3.5 and 0.025 mol/l
2.5. Extraction procedure for isolation of investigated drugs from DEA. 95 ␮L of the eluate was injected directly into the HPLC col-
serum and saliva umn.

Solid-phase extraction (SPE) was carried out using

BAKERBONDTM spe Octadecyl J.T. Baker cartridges, 100 mg/1 mL
2.6. Method validation
on a Baker spe-12 G apparatus.
The validation study was performed using spiked human saliva
2.5.1. Preparation of serum samples and serum according to the recommendations of the international
Blank serum samples (1.5 mL) were spiked with appropriate guidelines and include evaluation of selectivity, linearity, limit
amount of drugs. Then, the serum sample was diluted with 1.5 mL of detection (LOD), limit of quantification (LOQ), matrix effects,
of ammonium buffer at pH 8.6. extraction recovery, process efficiency, precision and accuracy.
C18 extraction columns were activated with 1 mL of methanol
and conditioned with 1 mL of mixture containing water and
ammonium buffer at pH 8.6 (5:1). Then, the serum contain-
2.6.1. Selectivity
ing the investigated drug was introduced to the columns at
Selectivity was evaluated by analyzing the saliva and serum
speed 1 mL/min. The columns were prewashed with 1 mL of
samples from different sources to investigate the potential inter-
MeOH—water solution (1:4) and dried applying vacuum for
ferences with the signals of analytes. Extent of interferences
3 min. The extracted drugs were eluted twice with 1 mL of mix-
originated by endogenous plasma components at the specific reten-
ture containing 98% methanol and 2% acetic acid. The sample
tion time of each analyte was evaluated through a comparison of
was evaporated to dryness and dissolved in 0.5 mL of mix-
blank human body fluids’ samples with the fortified serum or saliva
ture containing 40% (v/v) MeOH, 40% (v/v) double distilled
samples.
water, 20% (v/v) acetate buffer at pH. 3.5 and 0.025 mol/L DEA.
95 ␮L of the eluate was injected directly into the HPLC col-
umn.
2.6.2. Linearity
2.5.2. Preparation of saliva samples Method linearity was studied by spiking blank plasma or saliva
1.5 mL of human saliva was spiked with appropriate amount with suitable amounts of drug standards. Samples were prepared
of drugs and was filtered through a 0.22 ␮m filter. Next 0.3 mL of by SPE and determined by HPLC method described above. Calibra-
ammonium buffer at pH 8.6 was added to the sample. SPE C18 tion curves were constructed by analyzing fortified samples at six
extraction column were activated with 1 mL of methanol and con- concentrations, ranging from 10 to 1000 ng/mL in six replicates.
ditioned with 1 mL of mixture containing water and ammonium The calibration curves were obtained by means of the least square
buffer at pH 8.6 (5:1). Then, the saliva sample was introduced method.
to the column at speed 1 mL/min. The column was prewashed
with 1 mL of MeOH—water solution (1:4) and dried applying
vacuum for 3 min. The extracted drugs were eluted twice with 2.6.3. Limit of detection (LOD) and limit of quantification (LOQ)
1 mL of mixture containing 98% methanol and 2% acetic acid. Limit of detection (LOD) and limit of quantification (LOQ) values
The sample was evaporated to dryness and dissolved in 0.5 mL were calculated by determination of signal/noise ratio. Signal/noise
of mixture containing 40% (v/v) MeOH, 40% (v/v) double dis- ratio 3:1 and 10:1 responds to LOD and LOQ respectively.
 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80 73

Table 2
Comparison of process efficiency obtained for psychotropic drugs by various extraction procedures from human serum samples. Nominal concentration of investigated drugs
in spiked serum samples was 200 ng/mL.

Name of drugs Process efficiency (%)

Procedure 1 Procedure 2 Procedure 3 Procedure 4 Procedure 5 Procedure 6

Risperidone 99.35 98.52 87.52 96.68 81.53 83.98

Citalopram 119.25 106.21 100.22 97.73 37.53 97.64
Clozapine 69.79 87.64 95.24 100.78 31.67 88.46
Quetiapine 134.93 119.27 115.74 149.41 83.55 88.07
Levomepromazine 85.15 94.81 81.84 82.90 4.648 83.36
Perazine 34.41 51.50 18.04 46.04 7.272 78.46
Aripiprazole 100.95 82.08 104.67 128.16 9.685 82.60

Table 3
Matrix effect obtained for investigated psychotropic drugs in serum and saliva samples. Values were measured at 250 or 270 nm (marked *).

Name of compound Matrix effect for serum samples Matrix effect for saliva samples

20 ng/mL 200 ng/mL 800 ng/mL 20 ng/mL 200 ng/mL 800 ng/mL

Risperidone 72.69 90 95.74 99.66 102.65 100.9

98.27* 98.42* 99.84*

Citalopram 120.29 97.16 103.04 124.2 105.48 102.01

Clozapine 101.93 102.9 99.48 105.5 104.72 100.74
Quetiapine 113.76 93.46 97.41 103.7 102.9 97.84
Levomepromazine 105.84 98.62 99.68 100.6 103.4 101.4
Perazine 118.93 95.02 100.69 110.5 103.59 101.73
Apiprazole 102.71 99.46 98.12 94.79 101.97 100.57

A = external solution peak area, B = post-extraction sample peak

area, C = extracted matrix peak area.

2.6.5. Precision and accuracy

Intraday precision and accuracy were performed by analyz-
ing six replicates at three concentration: 20 ng/mL, 200 ng/mL and
800 ng/mL in the same day. For interday precision and accuracy, six
replicates of blank saliva or serum samples spiked at low, medium
and high concentrations were analyzed in three different days.

3. Results and discussion

3.1. Optimization of chromatographic condition

Psychotropic drug standards (Table 1) were chromatographed

on Pheny-Hexyl column by the use of aqueous mobile phases con-
taining MeOH, acetate buffer at pH 3.5, and addition of 0.025 M
DEA. The chromatographic system was examined in terms of reten-
tion, separation selectivity, peak shape and system efficiency in
Fig. 3. Chromatograms obtained for human saliva spiked of 200 ng of seven psy-
respect of their usefulness for analysis of investigated drugs in
chotropic drugs (1) and for blank human serum (2).  = 250 nm.
human serum or saliva. Fig. 1 shows a typical chromatogram of
standard solution. As can be seen obtained peaks are fully separated
and symmetrical which is caused by the use of the double pro-
2.6.4. Matrix effects, extraction recovery, and process efficiency tection against interaction between basic psychotropic drugs and
Matrix effects, extraction recovery, and process efficiency were free silanol groups, when are simultaneously applied phenyl-hexyl
evaluated at three concentration levels (low, medium and high) stationary phase with ␲-␲ ligands and mobile phases containing
according the following formulas: addition of DEA as silanol blocker. Application of such chromato-
graphic system lets to obtain very symmetrical peaks (all As values
B
Matrix effect (%) = × 100 in range 0.9–1.2) and system efficiency was very high (N/m > 50
A
000) for all investigated psychotropic drugs. The high performance
chromatographic system allows for separation of the analytes from
C the matrix components and decreases the limit of their detection
Recovery (%) = × 100
B and quantification.
The effect of methanol concentration in the injected psy-
C chotropic drug standards’ samples (containing mixture of methanol
ExtractionEfficiency (%) = × 100 and water) on system efficiency and peaks’ symmetry was investi-
A
gated. For most investigated compounds N/m were not significantly
 74
Table 4
Intra-day and inter-day validation parameters obtained for spiked serum samples (n = 6).

A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80
Name of compound Concentration added (ng/mL) Intra-day Inter-day

Average Standard Accuracy (%) Precision (%CV) Average Standard Accuracy (%) Precision (%CV)
concentration deviation SD concentration deviation SD
founded(ng/mL) founded(ng/mL)

Risperidone 20 21.95 0.72 109.76 3.26 21.97 0.87 109.83 3.97

200 197.22 14.07 98.61 7.13 197.29 11.44 98.64 5.80
800 785.45 21.46 98.18 2.73 785.27 18.07 98.16 2.30

Citalopram 20 22.32 4.12 111.58 18.48* 21.22 3.89 106.11 18.33

200 204.70 11.72 102.35 5.72 202.93 10.03 101.46 4.94
800 822.26 43.86 102.78 5.33 821.29 40.51 102.66 4.93

Clozapine 20 20.63 0.62 103.13 3.03 20.86 0.97 104.31 4.63

200 191.95 13.49 95.98 7.03 196.10 11.83 98.05 6.03
800 778.81 32.86 97.35 4.22 786.75 30.78 98.34 3.91

Quetiapine 20 21.13 1.19 105.66 5.61 21.25 1.05 106.26 4.96

200 186.43 8.62 93.22 4.62 191.21 11.20 95.60 5.86
800 775.94 29.11 96.99 3.75 779.09 24.09 97.39 3.09

Levomepromazine 20 21.59 2.18 107.96 10.08 21.68 1.98 108.42 9.15

200 191.83 14.17 95.92 7.39 193.51 12.48 96.75 6.45
800 792.71 37.99 99.09 4.79 802.33 33.00 100.29 4.11

Perazine 20 21.88 2.30 109.39 10.53 21.88 1.91 109.42 8.73

200 178.20 6.42 89.10 3.60 181.39 7.38 90.69 4.07
800 794.92 39.31 99.37 4.95 796.02 31.09 99.50 3.91

Aripiprazole 20 22.76 2.48 113.81 10.90 22.70 1.81 113.49 7.96

200 207.87 5.04 103.93 2.43 205.21 8.05 102.60 3.92
800 819.38 51.70 102.42 6.31 831.35 40.03 103.92 4.81
 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80 75

Table 5
Parameters of calibration curves for quantitative analysis of psychotropic drugs in human serum or saliva samples: calibration curves’ equations, regression coefficient (r),
limit of detection (LOD), limit of quantitation (LOQ). Concentration range 10–1000 ng/mL. Analytical wavelength was 277 nm for risperidone and 250 nm for other drugs.

Name of compounds Serum samples Salivasamples

Equation of calibration curve r LOD LOQ Equation of calibration curve r LOD LOQ

Risperidone y = 1821x − 8273 0.9998 4.78 15.93 y = 850.97x + 452 0.9999 4.00 13.35
Citalopram y = 1140.1x + 9532.9 0.9994 5.60 18.68 y = 1181.3x + 8903.3 0.9991 5.48 18.25
Clozapine y = 3790x − 9759.5 0.9995 3.15 10.51 y = 3848.4x + 5624.7 0.9993 5.65 18.83
Quetiapine y = 2218.1x − 2563 0.9992 5.23 17.44 y = 2325.1x + 1656.3 0.9994 2.77 9.24
Levomepromazine y = 4630.1x − 7054.9 0.9999 2.40 7.99 y = 4706.5x − 26302 0.9999 5.63 18.76
Perazine y = 5046.4x + 17551 0.9995 1.41 4.72 y = 5147.8x − 14383 0.9999 2.87 9.57
Aripiprazole y = 1370.7x + 971.9 0.9995 5.29 17.63 y = 1697.3x + 507.9 0.9996 4.97 16.58

Fig. 4. (A) Chromatogram obtained for serum sample obtained from patient treated by 30 mg of aripriprazole.  = 250 nm. (B) Chromatogram obtained for saliva sample
obtained from patient treated by 30 mg of aripriprazole.  = 250 nm.

changed in range of 0–40% of methanol and little increased in range sample injection volume did not cause changes in the number of
40–80% of methanol concentrations. However, symmetry of peaks theoretical plates obtained for all test drugs. As values were also
significantly changed with increased methanol concentration in similar in whole range of injected sample volumes (1.0 < As < 1.3).
injected samples. In range 0–60% of methanol asymmetry factors Based on obtained results, the serum or saliva fortified samples and
were similar (in range from 1.1 to 1.4) for all injected samples, how- samples from patients, after proper preparation, were dissolved in
ever with further increase of methanol concentration As values the mobile phase.
decreased (for 80% of methanol 0.7 < As < 1.2). For most investi-
gated compounds most symmetrical peaks were obtained when 3.2. Optimization of SPE conditions
concentration of methanol in injected sample was 60% of methanol
(1.1 < As < 1.2). Different extraction procedures were tested for isolation of
The influence of the injection volume on system efficiency and investigated psychotropic drugs from both biological fluids—serum
symmetry of peaks was also investigated. For samples dissolved in and saliva. Extraction was performed on BAKERBONDTM speOc-
methanol increase of injection volume caused significant decrease tadecyl J.T. Baker cartridges (100 mg/1 mL or 500 mg/mL) with
of N/m values obtained for most investigated drugs especially for application of different solvents, pH, volume of solvents and dif-
apiprazole, perazine and levomepromazine. When injection vol- ferent concentration of solvent mixtures. The aim of this study
ume was 10 ␮L N/m values were in range 50,000–450,000 for all was selection an extraction procedure to obtain the best recov-
investigated compounds, but for 100 ␮L samples N/m were lower ery and process efficiency, the possible lowest matrix effect
(10,000–110,000). Great differences were also observed in As val- and good purification from endogenous compounds. In this
ues. Most symmetrical peaks were obtained for samples volume 10 procedure SPE columns were activated with methanol and con-
or 20 ␮L, all peaks were of excellent symmetry (0.9 < As > 1.2), but ditioned with mixture containing water and ammonium buffer
for 100 ␮L samples As for all compounds were out of the optimal at pH 8.6. Then, the samples containing investigated drugs were
range. introduced to the columns which were prewashed with MeOH
In the next part of experiments psychotropic drug standards —water solution. After drying the extracted drugs were eluted
were dissolved in mobile phase of the composition such as ini- twice with mixture containing 98% methanol and 2% acetic
tial gradient elution mobile phase composition.The increase of the acid.
 76
Table 6
Intra-day and inter-day validation parameters obtained for spiked saliva samples (n = 6).

A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80
Name of compound Concentration added (ng/mL) Intra-day Inter-day

Average Standard Accurancy (%) Precision (%CV) Average Standard Accurancy (%) Precision (%CV)
concentration deviation SD concentration deviation SD
found(ng/mL) found(ng/mL)

Risperidone 20 17.00 0.80 84.98 4.69 16.77 0.85 83.84 5.05

200 212.25 7.94 106.13 3.74 208.90 10.58 104.45 5.07
800 797.53 38.96 99.69 4.88 811.31 34.88 101.41 4.30

Citalopram 20 20.72 3.49 103.62 16.82 19.95 3.80 99.77 19.07

200 216.00 10.71 108.00 4.96 214.44 9.82 107.22 4.58
800 825.75 48.06 103.22 5.82 815.42 48.91 101.93 6.00

Clozapine 20 17.57 0.43 87.83 2.47 17.51 0.93 87.55 5.29

200 222.25 16.43 111.13 7.39 219.78 12.69 109.89 5.77
800 830.47 45.42 103.81 5.47 831.31 35.55 103.91 4.28

Quetiapine 20 18.30 1.60 91.48 8.73 17.58 1.56 87.92 8.89

200 209.96 2.11 104.98 1.00 219.78 12.69 109.89 5.77
800 775.07 29.11 96.88 3.76 211.16 4.16 105.58 1.97

Levomepromazine 20 21.53 0.38 107.63 1.77 21.54 0.84 107.70 3.89

200 195.52 2.60 97.76 1.33 195.92 3.51 97.96 1.79
800 793.99 38.10 99.25 4.80 800.26 31.37 100.03 3.92

Perazine 20 22.21 0.96 111.06 4.34 22.15 0.99 110.75 4.45

200 193.41 3.64 96.71 1.88 192.88 3.85 96.44 2.00
800 801.08 33.80 100.14 4.22 797.98 27.02 99.75 3.39

Aripiprazole 20 17.50 1.51 87.51 8.62 18.25 1.96 91.25 10.76

200 204.83 8.22 102.42 4.01 203.90 14.39 101.95 7.06
800 816.16 39.46 102.02 4.83 805.26 40.20 100.66 4.99
 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80 77

MeOH and acetic acid was applied for elution of drugs. Recov-
ery and process efficiency obtained by the procedure was higher
for all investigated drugs, for this reason, the above procedure
was used for extraction of psychotropic drugs from patient sam-
ples.
The matrix effect was also investigated at different wavelength.
Based on the obtained results the optimum wavelength for most
investigated drugs in serum and saliva samples was 250 nm, only
for risperidone in serum lowest matrix effect was at  = 227 nm.
Values of % matrix effect is presented in Table 3. All drugs in saliva
and most drugs except risperidone in serum samples from patients
were quantified at 250 nm. Risperidone in serum samples was ana-
lyzed at 227 nm.

3.3. Analysis of fortified human fluid samples

Fig. 2 presents chromatograms obtained for blank human serum

and for serum samples fortified by seven psychotropic drugs at
two wavelengths. All peaks of analytes are symmetrical and fully
separated from components of matrix.
Validation of the method was performed for human serum sam-
ples spiked by 20, 200 or 800 ng/mL of all investigated psychotropic
drugs. Data on intra- and inter-assay validation parameters are
summarized in Table 4. The values of the intra-day accuracy
were obtained in range 109.39–113.81% for samples spiked by
20 ng/mL of psychotropic drugs and in range 89.10–103.93% or
96.99–102.78% for samples spiked by 200 or 800 ng/mL respec-
tively. The inter-day accuracy were between 104.31–113.49%;
90.69–102.60% and 97.39–103.92% for serum samples fortified
by psychotropic drugs at 20, 200 and 800 ng/mL concentra-
tions. Intra-day CV with low level drugs were from 3.03 to
Fig. 5. Chromatogram obtained for serum sample obtained from patient treated by
3 mg of risperidone and 400 mg of quetiapine.  = 277 nm. 18.48% and inter-day from 3.97 to 18.33%. For serum samples
containing addition of 200 or 800 ng/mL intra-day CV were in
range 2.43–7.39% and 2.73–6.31 respectively. Inter-day CV were
from 3.92–6.45 for samples spiked by 200 ng/mL and from 2.30
Table 2 presents comparison of process efficiency obtained to 4.93 for samples spiked by 800 ng/mL of each psychotropic
for psychotropic drugs by various extraction procedures from drugs.
human serum samples. The optimal values of process efficiency To evaluate linearity, calibration curves with six concentra-
for most investigated drugs were obtained when procedure 6 tion points for each drug were prepared. The calibration curves
was applied. The procedure was chosen for extraction of inves- were constructed by analyzing human serum samples con-
tigated drugs from serum samples of patients treated with taining different concentrations of the target drugs previously
therapeutic doses of the psychotropic drugs. Extraction recov- treated by the optimized SPE procedure. Equation of calibra-
ery, process efficiency and %CV for psychotropic drugs in human tion curves, regression coefficient, limit of detection (LOD) and
serum samples spiked by 2000 and 800 ng/mL were also inves- limit of quantitation (LOQ) were calculated (Table 5). For all
tigated. The extraction recovery was in range from 94.23 to drugs in investigated range of concentrations linear depen-
101.21%, process efficiency in range 94.65–121.74% and %CV in dencies were obtained. The lowest LOD and LOQ values in
range 3.46–10.85 for samples spiked of 20 ng/mL of psychotropic serum samples were obtained for perazine (LOD = 1.41 and
drugs. For serum samples spiked by 200 ng/mL of investigated LOQ = 4.72 ng/mL).
drugs extraction recovery values were obtained from 80.10 to The procedure for determination of seven psychotropic drugs
93.66%, process efficiency values were in range 78.60–95.72% and in saliva samples was also developed. Typical chromatograms
CV in range 2.51–4.82% and for samples spiked by 800 ng/mL of of blank saliva and spiked saliva samples (after SPE) are given
drugs extraction recoveries were from 82.157 to 100.49%, pro- in Fig. 3. HPLC procedure for quantification of psychotropic
cess efficiency values were in range 78.46–97.64% and CV in range drugs in saliva samples was validated in a similar manner as
2.42–7.450%. serum samples. Intra-day accuracy were in range 84.98–107.63%
Optimization of SPE procedure for psychotropic drugs in saliva for samples spiked by 20 ng/mL of drugs, 96.71–111.13% and
has also been carried out. The use of saliva for determination of 96.88–103.81% for samples fortified by 200 and 800 ng/mL
psychotropic drugs has the advantage of being non-invasive and respectively (Table 6). Inter-day accuracy were 83.84–107.70%,
easily acquired. Different solvents, pH, volume of solvents and dif- 96.44–109.89% and 99.75–105.58% for samples spiked by 20,200
ferent concentration of solvent mixtures were tested. Matrix effect and 800 ng/mL of drugs. Intra-day %CV were in range 1.77–16.82,
was lower for most investigated compounds (excepted clozap- 1.00–7.39 and 3.76–5.82, inter-day %CV were from 3.89 to 19.07,
ine and quetiapine) for following SPE procedure: columns were 1.79–7.06 and 1.97–6.00 for saliva samples spiked at three con-
conditioned by MeOH, mixture of water and ammonium buffer centration levels. For saliva samples linearity was also obtained
at pH 8.6 was applied for column pre-equilibration, sample was for all tested concentration levels (10–1000 ng/mL). The lowest
mixed with ammonium buffer at pH 8.6, 20% MeOH in water LOD and LOQ in saliva were obtained for quetiapine (2.87 and
was used for washing of adsorbent and mixture containing 98% 9.24 ng/mL).
78 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80

Table 7
Drugs and their metabolites founded in human serum and saliva samples.

Sample Drugs detected by HPLC–MS Dose (mg) Measured concentration by HPLC-DAD (ng/m Metabolites detected by MS (precursor ion m/z-product
ion m/z)

Serum A Quetiapine 750 334.62 Norquetiapine: (296–210)

(7-hydroxy-quetiapine: 400–269)
7-hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)
Clozapine 100 – Norclozapine: (313–192)
Clozapine N-oxide: (343–256)

Saliva A Quetiapine 750 107.16 Norquetiapine: (296–210)

(7-hydroxy-quetiapine: 400–269)
7-hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)
Clozapine 100 14.23 (<LOQ) Norclozapine: (313–192)
ClozapineN-oxide: (343–256)
Serum B Aripiprazole 30 742 Dehydroaripiprazole: (446–285)

Saliva B Aripiprazole 30 93.06 Dehydroaripiprazole: (446–285)

Serum C Clozapine 150 86.51 Norclozapine: (313–192)

Clozapine N-oxide (343–256)

Saliva C Clozapine 150 50.53 Norclozapine: (313–192)

Clozapine N-oxide: (343–256)

Serum D Clozapine 150 97.35 Norclozapine: (313–192)

Clozapine N-oxide: (343–256)
Risperidone 3 16.36 9-Hydroxyrisperidone: (427–207)
7-Hydroxyrisperidone

Serum E Perazine 150 32.63 –

Serum F Aripiprazole 15 78.48 Dehydroaripiprazole: (446–285)

Serum G Risperidone 3 54.85 9-Hydroxyrisperidone: (427–207)

7-Hydroxyrisperidone
Quetiapine 400 304.95 Norquetiapine: (296–210)
(7-hydroxy-quetiapine: (400–269)
7-hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)

Saliva H Aripiprazole 30 93.06 Dehydroaripiprazole: (446–285)

Saliva I Quetiapine 400 107.5 Norquetiapine: (296–210)

7-Hydroxy-quetiapine: (400–269)
7-Hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)
Saliva J Citalopram 40 34,23 Desmethylcitalopram: (311–109)
N,N-Didesmethylcitalopram: (297–109)
Risperidone 4 1,1 (<LOD) 9-Hydroxyrisperidone: (427–207)
7-Hydroxyrisperidone

Saliva K Quetiapine 100 64.72 Norquetiapine: (296–210)

(7-Hydroxy-quetiapine: 400–269)
7-Hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)

Saliva L Quetiapine 400 9.69 (<LOQ) (7-Hydroxy-quetiapine: (400–269)

7-Hudroxy, N-desalkyl quetiapine: (312–226)
Quetiapine sulfoxide: (400–221)
Aripiprazole 15 – Dehydroaripiprazole: (446–285)

Saliva M Risperidone 3 54.85 9-Hydroxyrisperidone (427–207)

7-Hydroxyrisperidone

Lower LOD and LOQ values were obtained for risperidone, citalo- presence of drugs in samples from patients were additionally con-
pram, quetiapine and aripiprazole when were analyzed in saliva firmed by MS spectra. Patients were treated with different doses of
samples, for clozapine, levomepromazine and perazine in serum one or two psychotropic drugs. Sample chromatograms obtained
samples. for serum and saliva from the same patient treated by 30 mg of
aripriprazole (both samples obtained 2 h after administration) are
presented in Fig. 4A and B. Fig. 5 presents chromatogram obtained
3.4. Analysis of fluid samples from patients for saliva sample from patient treated by 3 mg of risperidone and
400 mg of quetiapine obtained also two hours after administration.
The proposed method was successfully applied to determine of There was no significant interference for any of the analytes from
the psychotropic drugs in human serum and saliva samples from matrix components in the real serum and saliva samples. In some
patients receiving medication. For identification of psychotropic analyzed samples main metabolites of investigated drugs were also
drugs in samples from patients HPLC–MS method was used. The
 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80 79

identified. Results of quantification of some psychotropic drugs and [9] W.R. Malfara, C. Bertucci, M.E.C. Queiroz, S.D. Carvalho, M.L.P. Bianchi, E.J.
determination of their metabolites in real serum and saliva samples Cesarino, J.A. Crippa, R.H.C. Queiroz, Reliable HPLC method for therapeutic
drug monitoring of frequently prescribed tricyclic and nontricyclic
are summarized in Table 7. antidepressants, J. Pharm. Biomed. Anal. 44 (2007) 955–962.
[10] S.A. Mohajeria, G. Karimi, M.R. Khansari, Clozapine imprinted polymers:
synthesis, characterization and application for drug assay in human serum,
4. Conclusions Anal. Chim. Acta 683 (2010) 143–148.
[11] L. Mercolini, F. Bugamelli, E. Kenndler, G. Boncompagni, L. Franchini, M.A.
Raggi, Simultaneous determination of the antipsychotic drugs
A simple HPLC-DAD method was developed and validated for
levomepromazine and clozapine and their metabolites in human plasma by a
the quantification of seven basic psychotropic drugs in human HPLC-UV method with solid-phase extraction, J. Chromatogr. B 846 (2007)
serum and saliva. The SPE process allows to efficiently remove 273–280.
interfering substances from serum and saliva matrices while [12] N. Ansermot, M. Brawand-Amey, C.B. Eap, Simultaneous quantification of
selective serotonin reuptake inhibitors and metabolites in human plasma by
the HPLC-DAD method permits the quantification of investigated liquid chromatography–electrospray mass spectrometry for therapeutic drug
drugs and some their metabolites over the wide concentra- monitoring, J. Chromatogr. B 885–886 (2012) 117–130.
tion range typically measured in psychiatric patients. Optimized [13] R. Urinovska, H. Brozmanova, P. Sistik, P. Silhan, I. Kacirova, K. Lemr, M.
Grundmann, Liquid chromatography–tandem mass spectrometry method for
HPLC–MS method can be used for identification of drugs and determinationof five antidepressants and four atypical antipsychotics and
metabolites in samples of serum and saliva from medicated their mainmetabolites in human serum, J. Chromatogr. B 907 (2012)
patients. 101–107.
[14] J. Hasselstrøm, Quantification of antidepressants and antipsychotics in human
The use of the double protection against interaction between serum by precipitation and ultra high pressure liquid
basic compounds and free silanols, when are simultane- chromatography-tandem mass spectrometry, J. Chromatogr. B 879 (2011)
ously applied stationary phase with ␲–␲ ligands and mobile 123–128.
[15] J. Gradinaru, A. Vulliouda, C.B. Eap, N. Ansermot, Quantification of typical
phase containing addition of DEA as silanol blocker, allows
antipsychotics in human plasma by ultra-high performance liquid
to obtain symmetrical peaks and good system efficiency. Fur- chromatography tandem mass spectrometry for therapeutic drug monitoring,
ther improvement of peaks symmetry and system efficiency J. Pharm. Biomed. Anal. 88 (2014) 36–44.
[16] A.E. Steuer, M. Poetzsch, M. Koenig, E. Tingelhoff, S.N. Staeheli, A.T. Roemmelt,
was obtained by dissolving of injection samples in mobile
T. Kraemer, Comparison of conventional liquid chromatography–tandem
phase. massspectrometry versus microflow liquid chromatography–tandem mass
Proposed method has been successfully applied for quantifi- spectrometry within the framework of full method validation for
cation of investigated drugs in real serum and saliva samples. simultaneous quantification of 40 antidepressants and neuroleptics in whole
blood, J. Chromatogr. A 1381 (2015) 87–100.
The method allows for therapeutic drug monitoring, not only [17] M.L.D. Amundsen Å. Øiestad, L. Ekeberg Kristoffersen, Quantitative
for a single compound but also for mixtures. The good sensitiv- determination of fifteen basic pharmaceuticals in ante- and post-mortem
ity, precision and the high accuracy of the assay of psychotropic whole blood by high pH mobile phase reversed phase ultra high performance
liquid chromatography–tandem mass spectrometry, J. Chromatogr. B 927
drugs make the methods a useful tool in clinical laboratories (2013) 112–123.
for therapeutic drugs monitoring as well as in forensic laborato- [18] D.P. Patel, P. Sharma, M. Sanyal, P.S. Shrivastav, SPE–UPLC–MS/MS method for
ries for their determination at therapeutic and higher levels in sensitive and rapid determination of aripiprazole in human plasma to support
a bioequivalence study, J. Chromatogr. B 925 (2013) 20–25.
human fluids. The use of saliva as opposed to serum or plasma [19] S. Park, C.-S. Park, S.J. Lee, B. Cha, Y.A. Cho, Y. Song, E.A. Yu, G.-S. Kim, J.S. Jin,
for determination of psychotropic drugs has been shown to be an A.M.A. El-Aty, H.A. El-Banna, A. Hacımüftüoğlu, J.-H. Shim, S.C. Shin,
attractive alternative for therapeutic drug monitoring in psychi- Development and validation of a high performance chromatography–tandem
mass spectrometric method forsimultaneous determination of bupropion,
atric patients because of its collection are simpler, non-invasive
quetiapine and escitalopram in human plasma, Biomed. Chromatogr. 29
and painless. (2015) 612–618.
[20] P. Sistik, R. Urinovska, H. Brozmanova, I. Kacirova, P. Silhan, K. Lemr, Fast
simultaneous LC/MS/MS determination of 10active compounds in human
References serum for therapeutic drug monitoring in psychiatric medication, Biomed.
Chromatogr. (2015), http://dx.doi.org/10.1002/bmc.35380.
[1] C. Xiong, J. Ruan, Y. Cai, Y. Tang, Extraction and determination of some [21] N. Ansermot, M. Brawand-Amey, A. Kottelat, C.B. Eap, Fast quantification of
psychotropic drugs in urine samples using dispersive liquid–liquid ten psychotropic drugs and metabolites in human plasmaby ultra-high
microextraction followed by high-performance liquid chromatography, J. performance liquid chromatography tandem mass sspectrometry for
Pharm. Biomed. Anal. 49 (2009) 572–578. therapeutic drug monitoring, J. Chromatogr. A 1292 (2013)
[2] C. Sasaki, T. Shinozuka, C. Murakami, W. Irie, K. Maeda, T. Watanabe, N. 160–172.
Nakamaru, M. Furukawa, S. Nakamura, K. Kurihara, Simultaneous [22] G. Vecchione, B. Casetta, A. Chiapparino, A. Bertolino, M. Tomaiuolo, F.
determination of 5 psychotropic drugs of various types in an autopsy case of Cappucci, R. Gatta, M. Margaglione, E. Grandone, A reliable and rapid tool for
acute multiple drug poisoning, Forensic Sci. Internat. 227 (2013) plasma quantification of 18 psychotropic drugs by ESI tandem mass
90–94. spectrometry, J. Pharm. Biomed. Anal. 67–68 (67) (2012)
[3] H.-C. Huang, C.-H. Liu, T.-H. Lan, T.-M. Hu, H.-J. Chiu, Y.-C. Wu, Y.L. Tseng, 104–113.
Detection and quantification of aripiprazole and its metabolite, [23] M. Caloro, L. Lionetto, I. Cuomo, A. Simonetti, D. Pucci, S. De Persis, B. Casolla,
dehydroaripiprazole, by gas chromatography–massspectrometry in blood G.D. Kotzalidis, A. Sciarretta, S. De Filippis, M. Simmaco, P. Girardi, An
samples of psychiatric patients, J. Chromatogr. B 856 (2007) 57–61. improved simple LC–MS/MS method for the measurement of serum
[4] S. Dziomba, P. Kowalski, T. Baczek, Micelle to solvent stacking of tricyclic aripiprazole and its major metabolite, J. Pharm. Biomed. Anal. 62 (2012)
psychiatric drugs incapillary electrophoresis, J. Pharm. Biomed. Anal. 62 135–139.
(2012) 149–154. [24] R. Wietecha-Posłuszny, A. Garbacik, M. Woźniakiewicz, A. Moos, M.
[5] J. Li, F. Zhao, H. Ju, Simultaneous determination of psychotropic drugs in Wieczorek, P. Kościelniak, Application of microextraction by packed sorbent
human urine by capillary electrophoresis with electrochemiluminescence to isolationof psychotropic drugs from human serum, Anal. Bioanal. Chem.
detection, Anal. Chim. Acta 575 (2006) 57–61. 402 (2012) 2249–2257.
[6] S.A. Bellomarino, A.J. Brown, Xavier A. Conlan, N.W. Barnett, Preliminary [25] R. Das, Y.K. Agrawal, J. Chromatogr. Sci. 51 (2013) 146–154.
evaluation of monolithic column high-performance liquid chromatography [26] C. Greiner, C. Hiemke, W. Bader, E. Haen, Determination of citalopram and
with tris(2,2 -bipyridyl) ruthenium(II) chemiluminescence detection for the escitalopram together with their active main metabolites desmethyl(es-)
determination of quetiapine in human body fluids, Talanta 77 (2009) citalopram in human serum by column-switching high performance liquid
1873–1876. chromatography(HPLC) and spectrophotometric detection, J. Chromatogr. B
[7] N. Unceta, A. Ugarte, A. Sánchez, A. Gómez-Caballero, M.A. Goicolea, R.J. 848 (2007) 391–394.
Barrio, Development of a stir bar sorptive extraction based HPLC-FLD [27] M. Song, X. Xu, T. Hang, A. Wen, L. Yang, Development of an LC–MS/MS
methodfor the quantification of serotonin reuptake inhibitors in plasma, method for the simultaneous quantification of aripiprazole and
urine and brain tissue samples, J. Pharm. Biomed. Anal. 51 (2010) dehydroaripiprazole in human plasma, Anal. Biochem. 385 (2009)
178–185. 270–277.
[8] G. Zhang, A.V. Terry, M.G. Bartlett, Simultaneous determination of five [28] M. Cutroneo Beljean, R. Phan Tan Luu, A.-M. Siouffi, Optimization of the
antipsychotic drugs in rat plasma by high performance liquid separation of some psychotropic drugs and their respective metabolites by
chromatography with ultraviolet detection, J. Chromatogr. B 856 (2007) liquid chromatography, J. Pharm. Biomed. Anal. 41 (2006)
20–28. 333–340.
80 A. Petruczynik et al. / Journal of Pharmaceutical and Biomedical Analysis 127 (2016) 68–80

[29] F. Lancelin, K. Djebrani, K. Tabaouti, L. Kraoul, S. Brovedani, P. Paubel, M.-L. [30] H. Ebrahimzadeh, Z. Saharkhiz, M. Tavassoli, F. Kamarei, A.A. Asgharinezhad,
Piketty, Development and validation of a high-performance liquid Ultrasound-assisted emulsification microextraction based on solidification of
chromatography method using diode array detection for the simultaneous floating organic droplet combined with HPLC-UV for the analysis of
quantification of aripiprazole and dehydro-aripiprazole in human plasma, J. antidepressant drugs in biological samples, J. Sep. Sci. 34 (2011)
Chromatogr. B 867 (2008) 15–19. 1275–1282.