Journal of Analytical Toxicology,Vol.
31, July/August2007
Quantification of Eight New Antidepressantsand
Five of their Active Metabolites in Whole Bloodby
High-Performance Liquid Chromatography-Tandem
Mass Spectrometry
Nad~ge Castaing, Karine Titier*, Mathilde Receveur-Daurel, Mait~ te-D~odic, Delphine te-bars,
Nicholas Moore, and Mathieu Molimard
Department of Clinical Pharmacologyand Toxicology, Pellegrin Hospital and University Victor Segalen,
33076 Bordeaux, France
Abstract I
A liquid chromatography-tandem mass spectrometry method is
described for the blood determination of selective serotonin
reuptake inhibitors (fluoxetine, paroxetine, sertraline,
fluvoxamine, and citalopram), serotonin noradrenaline reuptake
inhibitors (milnacipram and venlafaxine), a noradrenergic and
specific serotoninergic antidepressant (mirtazapine) and five of
their active metabolites (norfluoxetine, desmethylcitalopram,
didesmethylcitalopranl, desmethylvenlafaxine, and
desmethylmirtazapine). After a liquid-liquid extraction from
blood, the compounds and the internal standard
(methylrisperidone) were eluted on a XTerra| RP18 column with a
gradient of acetonitrile/ammonium formate buffer 4 mmol/L pH
3.2. They were then detected by electrospray ionization mass
spectrometry with multiple reaction monitoring mode.
The calibration curves were linear over the range 5-500 ng/mL
(20-2000 ng/mL for venlafaxine and desmethylvenlafaxine).
The limit of quantification was set at 5 ng/mL for each compound
(except for venlafaxine and desmethylvenlafaxine: 20 ng/mL).
The bias were lower than 12%. Intraday and interday precisions,
expressed as variation coefficient, were lower than 11%.
The extraction recoveries were between 70 and 90% except for
desmethylmirtazapine, desmethylvenlafaxine, milnacipram, and
didesmethylcitalopram. This specific and sensitive method allows
management of intoxication and is suitable for the routine
determination of antidepressants in forensic investigations.
Introduction
Selective serotonin reuptake inhibitors (SSRIs) and sero-
tonin noradrenergic reuptake inhibitors (SNaRIs) are widely
used for the treatment of depression and other psychiatric
* Author to whom correspondence should be addressed. Laboratoire de Pharmacologie Clinique
et de Toxicologie, H6pital Pellegrin, Place Am~lie Raba-L~on, 33076 Bordeaux cedex, France.
E-mail: karine.titier@pharmaco.u-bordeaux2.fr.
334 Reproduction(photocopying)of editorialcontentof thisjournalis prohibitedwithoutpublisher'spermission.
Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
disorders because of their better-tolerated adverse effects in
comparison with tricyclic antidepressant drugs (1-3). However,
they are frequently used in deliberate self-poisoning and they
can lead to major intoxication (4-8). Analysis of antidepres-
sants could be also necessary in forensic cases such as driving
under the influence of drugs, cases of violent crime, cases of
drug-facilitated sexual assault and cases of unknown cause of
death. Consequently, a specific analytical method may be in-
dicated in clinical and forensic toxicologyto screen and quan-
tify these drugs in whole blood.
Some methods for the determination of several antidepres-
sants, tricyclic and "new antidepressants", in biological sam-
ples have been described. They involve high-performance
liquid chromatography (HPLC) with ultraviolet detection
(9-11), mass spectrometry (MS) (12,13) or tandem MS (14,15),
and gas chromatography with mass spectrometry (GC-MS)
(16,17). However, none of these methods allows the simulta-
neous determination of SSRIs, SNaRls and mirtazapine by
LC-MS-MS with liquid-liquid extraction.
We have previously developed a method for the determina-
tion of tricyclic antidepressants and MAOIsby LC-MS-MS in
whole blood (18). The aim of this study was to develop a
LC-MS-MS method, adapted to routine application, allowing
specific and sensitive determination of SSRIs, SNaRIs, mir-
tazapine (a noradrenergic and specific serotoninergic antide-
pressant), and five of their active metabolites (norfluoxetine,
desmethylcitalopram, didesmethylcitalopram, desmethylven-
lafaxine, and desmethylmirtazapine) for forensic toxicology.
Materials and Methods
Chemicalsand reagents
Paroxetine, fluvoxamine, fluoxetine, and sertraline solu-
tions at 1 mg/mL in methanol were obtained from Promochem
(Molsheim, France). Norfluoxetinewas purchased from Sigma
Journal of AnalyticalToxicology,Vol. 31, July/August2007

Table I. MRM Transitions for the Detection of Antidepressants and Internal Standard (IS) by LC-MS-MS

Retention Parent Daughter Relative Cone Collision Dwell

Compound Time (min) Ion (m/z) Ions* (m/z) Intensity (Volts) Energy(eV) Time (ms)

Desmethylvenlafaxine 1.7 264.1 246,3 2.6 25 10 100

Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
107,0 100
Desmethylmirtazapine 2,2 252.0 195.2 4.9 30 20 100
209.2 100
Mirtazapine 2.5 266,1 195.2 9.0 25 25 100
209.2 100
Milnacipran 3.1 247.1 230.3 2.5 15 15 100
100 100
Venlafaxine 3.8 278,1 260.3 1.5 20 10 100
121.1 100
I,S 7.8 422,0 201.2 - 35 35 100
Didesmethylcitalopram 7.9 297.0 109,0 1,8 25 20 100
262.2 100
Desmethylcitalopram 8.2 311.0 109.0 2.4 30 20 100
262.2 100
Citalopram 8.5 325,0 109,0 3.7 30 25 100
262,2 100
Fluvoxamine 10,6 319.0 258.2 1,7 20 10 100
200.2 100
Paroxetine 10.6 330,0 192.2 2.9 30 20 100
151,1 100
Noffluoxetine 12.0 296.0 134,1 6,7 15 6 100
30.0 100
Fluoxetine 12.4 310.0 148.1 0.2 15 8 100
43.9 100
Sertraline 12.8 306,0 159.0 0.9 15 25 100
275,1 100

* The transition for quantification is underlined.

s~150
12-08-~IRS- SAt4G-MedLeg.05 Sm (Mn, I ~ ) 2: MRM of 11 Cltanr,~ES+
12~064flE;- SAnG JdedLeg-05 3: MRM of 80lalm~sE S+ "r 422 ~ 201.2
" ~ 0 sefl~81~rle I~ 3 ~ 15~1121)4
. j ~ . . . . , . . . . , . . . . , . . . . , . . . . , . . . . , . . . . , . . 7.~. IS 7.39113
, . . . . . . . . . . . . . .

. . , . . . . , . . . . . . . . . , . . . . , . . . . . . . . . . . . . , . . . . , . . . . ,

8.00 lOW 1200 14,00 16.00

12~m~064RS-~ ~ S m (M~ t~Q) 3: MRM ol8 Ch~nd~ES~
200 4.aO 6.00 BOO I@00

.................. . .... t2 24"/I >230.3

8.0~ 10.~0 1200 1400 16.00
t 2~064RS- SANG-MedLeg..05Sm (M IX 1]~) 3: MRM of8 C h e ~ ES* 1~ A rnliflaC~afl 4.=
11 2 ~ ~ 1341
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 ] . . . . ~ . . . . I . . . . ~ " ~ I " , . . . . ' . . . . ' . . . . I . . . . ' . . . . ' . . . . I . . . .

8 00 10~0 1200 14 00 16 00 2.00 4.00 6.00 5~00 10.00

12~1R~ S~+~L~Sm (M~ 1~2) 2; MRM ~ f l l Chanr~isE$+ 12~RS- ~ ~ S m {Mi~ 1~2) 1: MRM d l 0 C h ~ E S ~
3~ 278.1 ~ 2E0.3

j ~.
12-08-06-IRS- SAJ4G-NcdLeg-05Sm (Mn, 1]Q)
,0~ ,,o=, 2:MRM 0r 014ml~lES+
mnla~Une 1.Se~4

.,~
. .

8+00
12~&O6-1R5- S ~ e g
, .
~0.~
/~
.
I~/voxarnine
, . . . . , . . . .

10.~0
~ Sm (M~ 1~)
. . . . . . . . . . ,.
120~
. , . . , .
14.00 16.00
2: MRM 01"11OI~Ili~IsES+
+ , . .
3t9 ) 258.2
I 02e3
., ! 1 2 ~ R S - S ~
Ir
2.00

1~
-0SSm (Mn, I~Q)
deSIlltelh~WeftlaP4xPilil
4 0 0 6 0 0 a m

1: ~ MI0 ~
10.00

ES-,-
284,t >2,18.39,81e3
I~ Olalol~am 31259 lO9
. . . . , . . . . . . . . . . . . . . , . . . p , . . . . , . . . . . . . . . , . . . . , . . . . , . . . .
...... +,'+~ ....................................
,o| 12oo ,,| ,.~+~'"~ 2.00 4+~ 6.00 60D 10+00
12..Oe-OEARS- ~ -OGSm (Mn, I]Q) 2:MRM ofli Chemna~ES<. 12~S~64R~ ~ ~ S m (M~ lx2) 1: MRM of 10 ChaN)ds ES~
........ . . . . . . . . . . . . . . . . . . . . . '7.1+,=
8.00 t0fl0 1200 14.00 1600
12~RS- SA~G-MedLo~~ S m {M~ I]Q) 2:MRM o#11 C ~ E S +
.... , .... , , , . . . t , . . . . , . . . . , . . . . , . . . . , . . . .

'<'+I.
. . . , . . . .
~._ +,,,,,,,,.,,,~o,~
, .... , . . . . , . . . . , . . . . , . . . . + . . . . , . . . . , . . . .
~ i,,.+
=
, 12-06.,06JRS- SANG-.MedLeg..4D55m {Mn, 1x2)
2.00 4.00 6.00 800 10.00
1 MRM (11'10ChamPs ES+
8.0~ 10.~0 1200 14.00 1@00 / ~ 252 > 196.2
t2..08-~4P....%-S,A,NO-Ik~L~g,,,D5Sm (Mrl, 1~) ~: MRM 0;'11 Chann.l~SES+
~, 42'2 ~ 201 ?
iS 1.3Ge3
v--,- , ., .... ,. , ..-, .... ,..- , .,
B.00 1000 1200 14 00 1600 . . . . , . . . , . . . . , . . . . , . . . . , . . . . , .... , .... , .... , ....
200 4.~ 600 800 W.~
Time (mln)
Time (mln)

Figure 1. RepresentativeMRM chromatogramsof blood sample spiked Figure 2. RepresentativeMRM chromatogramsof blood sample spiked
with 50 ng/mLof each IRS.lhe transition of quantificationis shown for with 50 ng/mLof each other antidepressant. The transition of quantifi-
each compound. (I.S:internal standard.) cation is shown for each compound. (I.S: internal standard.)

335
 Journal of Analytical Toxicology, Vol. 31, July/August 2007

(Saint Quentin Fallavier, France). Venlafaxine (V) and its
metabolite were a gift from Lederle (Pearl River, NY); citalo-
pram, desmethylcitalopram, and didesmethylcitalopram were
from Lundbeck (Copenhagen, Denmark); mirtazapine and
desmethylmirtazapine were from Organon (Oss, The Nether-
lands); and milnacipran was from Pierre Fabre (Castres,
Standard solutions
The stock solutions w e r e prepared by dissolving accurately
weighed a m o u n t in m e t h a n o l to yield a 1 m g / m L drug con-
centration. A 1 m g / m L stock solution of internal standard in
methanol was prepared and further diluted 1:1000 in methanol
to give a 1 IJg/mL w o r k i n g solution.

France). Methylrisperidone, which was used as the internal
standard (I.S.),was kindly donated by Lundbeck (Copenhagen,
Denmark).
HPLC-grade solvents were purchased from Prolabo (Paris,
France), and analytical-grade chemicals were from Merck (No-
gent sur Marne, France).
Drug-free blood from healthy volunteers was provided by
Etablissement FranCais du Sang (Bordeaux, France).
Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
Routine daily calibration curves and controls were prepared
for each analytical batch in drug-free whole blood. Calibration
standards were made to yield concentrations of 5, 10, 25, 50,
100, 250, and 500 ng/mL [20, 40, 100, 200, 400, 1000, and 2000
ng/mL for V and desmethylvenlafaxine (DV)]. For the prepa-
ration of quality controls (QC) used for the validation of the
assay, an independent stock solution was prepared and further
diluted, to achieve concentrations of 10, 80, and 400 ng/mL

Table II. Intraday and Interday Precision and Accuracy Data

Intraday Interday
Added Observedconcentration Observed concentration
Compound Concentration
(ng/mL) (mean • SD, n = 4) CV (%) (mean • SD, n -- 15) CV (%) Bias (%)

Desmethylmirtazapine 10 9.34 + 0.50 5.39 9.54 + 0.79 8.26 -4.56

80 85.83 + 5.0 5.81 84.30 + 4.68 5.55 5.38
400 391.27 + 18.58 4.75 395.90 + 14.80 3.74 -1.03
Desrnethylvenlafaxine 40 39.69 + 2.30 5.81 38.88 + 2.41 6.21 -2.79
320 352.09 + 13.64 3.87 348.15 _+18.40 5.29 8.80
1600 1577.80 _+53.59 3.40 1542.73 _+85.37 5.53 -3.58
Mirtazapine 10 9.48 + 0.52 5.44 9.24 _+0.68 7.35 -7.58
80 82.73 + 5.25 6.35 80.99 + 4.51 5.57 1.24
400 384.96 + 12.59 3.27 393.25 + 9.34 2.39 -1.69
Milnacipran 10 10.47 __.1.15 10.97 10.04 _+1.00 10.13 0.41
80 85.16 __.6.12 7.18 84.60 + 5.09 6.01 5.74
400 395.51 + 9.41 2.38 382.16 __.31.05 8.13 -4.46
Venlafaxine 40 41.98 + 2.61 6.23 38.05 _+3.94 10.35 -4.87
320 361.41 _ 19.46 5.38 355.35 _+17.72 4.99 11.05
] 600 1512.10 + 47.94 3.17 1562.87 __.60.78 3.89 -2.32
Didesmethylcitalopram 10 11.35 _ 1.26 11.09 11.03 _+1.42 9.41 7.81
80 91.13 _+4.67 5.13 86.84 + 8.68 9.99 8.55
400 396.14 _+13.80 3.48 390.81 _+22.64 5.79 -2.30
Desmethylcitalopram 10 10.62 + 0.15 1.40 10.13 + 0.75 7.40 1.33
80 87.27 + 5.10 5.84 85.87 __.5.27 6.14 7.34
400 385.23 +_19.00 4.93 396.13 --. 16.66 4.21 -0.97
Citalopram 10 9.49 _+0.46 4.83 9.19 _+0.73 7.93 -8.15
80 83.81 + 5.16 6.16 82.47 _+4.73 5.74 3.08
400 397.75 + 11.36 2.86 398.52 _+17.57 4.41 -0.37
Fluvoxamine 10 9.47 + 0.72 7.60 9.41 __1.07 11.36 -5.92
80 88.89 + 11.82 4.94 87.00 _+9.73 11.19 8.75
400 370.98 +_15.60 4.21 378.91 _+14.64 3.86 -5.27
Paroxetine 10 9.64 + 0.69 7.11 9.87 _+0.89 9.00 -1.27
80 90.96 + 7.26 8.00 89.56 _+8.78 9.80 11.95
400 381.93 + 9.19 2.41 389.96 _+12.80 3.29 -2.51
Norfluoxetine 10 8.81 + 0.51 5.75 9.07 _+0.83 9.16 -9.33
80 86.78 _+5.64 6.50 86.68 + 4.97 5.73 8.35
400 379.12 _+24.16 6.37 354.47 _+37.25 10.51 -11.38
Fluoxetine 10 9.02 + 0.18 2.01 9.31 + 0.86 9.28 -6.95
80 87.60 + 3.40 3.88 87.89 _+4.11 4.67 9.86
400 383.87 + 16.88 4.40 388.15 --. 15.43 3.98 -2.96
Sertraline 10 9.44 + 0.51 5.35 9.54 + 0.74 7.80 -4.65
80 89.45 +_.3.62 4.05 88.46 + 4.06 4.59 10.58
400 377.53 + 23.32 6.t8 386.30 + 22.89 5.93 -3.42

336
Journal of Analytical Toxicology, Vol. 31, July/August 2007

Table III. Extraction and Matrix Effect Recoveries Data (40, 320, and 1600 ng/mL for V and DV).

Extraction Matrix Effect: Chromatography

Recovery (%) Relative Recovery (%) The HPLC unit consisted of an Alliance|
Nominal 2690 separation module (Waters, Milford, MA)
Concentration Mean + SD Mean • SD controlled by the Masslynx| software. Sepa-
Compound (ng/mt) (n = 4) CV (n = 3) CV

Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
rations were performed at 25~ on a XTerra
Desmethylmirtazapine 80 69.9 • 2.3 3.0 98.0• 3.0 RP18 100- x 2.l-ram column (Waters). The
400 69.88 _+1.3 1.8 mobile phase consisted of a gradient of ace-
Desmethylvenlafaxine 320 53.00 + 1.8 3.4 84.4• 2.1 tonitrile/ammonium formate buffer (4
1600 58.1 • 4.3 mmol/L, pH 3.2). First, acetonitrile was set at
Mirtazapine 80 80.3 • 1.5 1.9 94.2• 1.1 15% (v/v) for 1 rain; then itwas increased lin-
400 77.8 • 5.7 7.3 early to 35% over 12 rain and set at 35% for 1
Milnacipran 80 19.9 • 6.4 3.2 96.4• 2.3 min. The system was then set at the initial
400 10.7 _+1.3 12.6 conditions in 30 s and was re-equilibrated over
I.S. 117.8• 6.5 101.8• 1.5
2.5 rain before the next injection. The total
Venlafaxine 320 85.4 • 1.9 2.3 102.0• 2.0
run time of the analysis was 17 min at a flow
1600 81.4 + 3.9 4.8
Didesmethylcitalopram 80 39.3 • 1.8 4.7 93.9• 5.1
rate of 0.3 mL/min.
400 31.5 • 2.3 7.2
Desmethylcitalopram 80 70.8 • 1.5 2.2 98.2• 1.0 MS
400 75.2 • 1.6 2.2 A QuattroMicro| triple quadrupole detector
Citaloprarn 80 81.5 • 2.3 2.8 100.4• 2.3 (Waters) with electrospray ionization (ESI) in
400 84.0_+7.1 8.5 positive ion mode was used for detection. The
Fluvoxamine 80 74.7 + 5.3 7.2 101,5• 3.3 MS was operated in the multiple-reaction
400 80.4 + 4.1 5.1 monitoring (MRM)mode.
Paroxetine 80 76.3 _+3.0 3.9 91.0• 4.2 The ESI source was operated at a tempera-
400 73.9 _+3.7 5.0
ture of 115~ The desolvation temperature
Noffluoxetine 80 70.7 + 3.1 4.3 104.0• 9.1
was set at 280~ The cone gas flow was set at
400 73.0 -+5.3 7.2
Fluoxetine 80 71.9 -+ 1.0 1.4 99.6• 3.5
50 L/h, and the desolvation gas flow was at
400 76.2 • 4.8 6.3 500 L/h. The capillary voltage was 3.2 kV.
Sertraline 80 73.4 • 1.0 1.3 99.6• 1.0 The cone voltages were optimized for each
400 74.6 -+7.3 9.7 compound. The MS collision gas was argon at
2.96 x 10-3 mbar, and the collision energies

blanc bllrN=

723 31/ 9 IO0 1 2 - 0 7 - 0 6 - 1 ~ 8 - effe~ ..O2 3 M R M Ot 0 C n a n n ~ t s E S +

11.~ 3015 9 15g
9~ 5 . . . . . . . . . . . . . . , . . . . , . . . . . . . . .
~ ~ 11 1 .19 1347 16.04
2 00 .1 O 0 6.00 0 00 I O O0
12-07-06-l~S 8~te1-02 2 MPtM 0~ 11 C h a n n e l s E S +
S k .~ I -- ibpO0 . . . . . . . 12:Oh . . . . . . ~~ l o b . . . . . 1"S:00

12-07-0~-I~- e~ ~]2 3 MFtM of @ C ~ n n e m E S +

|5.42 I 0 > 140 1
i2-07-0G-1~C- c f f c t - 0 2 2 M ~ M of 11 C h a n n e l o EC i
[~ 11 4 = 2 ~ 201 ~ 11~ 206 13191389 15. 15.4~ 45

1oo, ' o TM ....

. . . . . . ~l.l~O . . . . . . . i'O!Ob . . . . . . . 1'2!~ ....... 11'.o'o . . . . . . . 1~:~
2 o0 4 .OO 6.00 000 10.oo
2 O7 0 6 IIqS ef~et 02 i M R M Ot I0 Ctqa~nels E S § 12-07-06-1RS- ~ -02 3; M R M ot @ Cfle=ni~ts E@*

l ........
12-07-0C-IR.~- ~f'tr~ 00
*02 . . . . . . . ~ DO . . . . . . . d ~0 ....... 0 ~0 ..R
1 M . . .M. . or 101 0C' 0h0a n n" e"l s tZG+ i " " " " " " 8 D() " " " " . . . . . I O O O
" " '" " 1 1 ~ 2 i2003 '1 '1 4 0 0 ~ 1G:O0

4 1 1 5 9 118 9 2 ~3 1 4 ~ { ; 2 71 6 1 1 ~ 2b .13 ~.~ 12-07.06-.l@~.-efte[-02 2 M@MOf I1CllanNtIESt

85 9,~ 330 9 I ~

-o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .oo
T2-OT-O6-1R.~- e ~ e t - 0 2
......... .oo .oo ,ooo
1 M R M or 10 C t l a n n e l s E S +
4.gv3 264 1 > 246 3
9 5 1 0,4 3a , o1, r~#4 3 1 9 =~ ~ 3"6J5 4 ~ ~3u 03
~ 4 24 583 / 12-07-Or:3-lRg- ~ 1 - 0 2 2 ~M o f 11 C l ~ a n r ~ t s E S §
31 g 9 2 5 8 2
2 00 4.00 6 00 8 00 1Q.00 I~ 8 " W 7.315 205
12-07-0E,-IRS e f t e t - 0 2 1 M ~ M 0f 10 C h a n n e l s E S +
90 .17 252 > 195 2
IO0 7 40 O,~ 9 9.94 10.81.. , t2;0() . . . . 1.4;0()" 1e . o 0
. 163 204 3.74 f~?4
..... . ......... . .... , .... , .... , .... , .... . .... , .... , ....
2 O0 4.00 6 00 8.00 I 0 00 1;~.CI7.06-11R5- Ml'et . 0 2 2 h4~M o f 11 C n a n ~ e t i I~S,r
] 2 - 0 7 - 0 6 - 1 1 ~ - Eqte;-02 1 M R M or 10 C h a n n e l ~ E S + 9 2 5 9 109
~5 ?" 7,04 @ 1S 188
5 57 ~47 1 - 23O 3
a4 9 .229.51 0.70 10.~

. . . . . . . . , . . . . . . . . . . . . . . , . . . . . . . . . , . . . . , . . . . . . . . . . . . . .
-5
200 400 000 B00 1000 800 1000 " " "12'00 ....... 1"4:50 . . . . . . . 1~'.00

'lime (rain) Time (mini

Figure 3. RepresentativeMRM chromatogramsof a blank human blood without internal standardand without analyte. The transition of quantification is
shown for each compound.

337
 Ioumal of Analytical Toxicology, Vol, 31, luly/August 2007

were optimized for each compound. 0.5M). Seven milliliters of hexane/isoamylic alcohol (99:1, v/v)
Parent ions, the corresponding daughter ions, retention was added to the mixture, which was then shaken for 10 rain
times, cone voltages, collision energies, and dwell times opti- prior to being centrifuged at 5000 rpm for 10 rain. Then, 200
mized for each compound and the I.S. are presented in Table I. pL of HCI (0.05N) was added to the organic layer. The mixture
was further shaken for 10 min and centrifuged at 5000 rpm for
Sample preparation 10 min, following which 5 pL of the aqueous phase was in-

Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
Whole blood (500 pL) was spiked with 100 pL of internal jected into the HPLC unit.
standard (1 pg/mL) and 500 pL of carbonate buffer (pH 9.7,
Assay validation procedures
Itd 0
12,46.064RS- S,N'tG-MedL~g -02 3: MRM 0IS C~lnnelI E S~
Ion suppression. Ion suppression was investigated through
the addition of the analytes to blank matrix extracts. If the re-
................... >.'it
9
800 10C0 t20~ t400 1600 sponse of the analytes is compared to an unextracted standard
12.0~0~-~[~- SANG-Mc'~L6~ -02 3: MRM 018 (~16rln~li E ;Trr
solution at the same nominal concentration, any difference
8,00 1000 lZ~O 1400 18.00
from 100% recovery could be attributed to matrix effect. One
12-~II~ S,~NC~MeldLe9 -02 3: MRM o f ~ ChAnr~lI E S~
I
11~ 29G) 1~4 1 concentration (80 ng/mL or 320 ng/mL for V and DV) was an-
1144~ ~ 11~NI '~3.~4 3t75 14153.I/4"1615 7'2
V -,.. , ....
6.~
, ....
1000
.-- -.' "''~'"~':,"':',''':r
1200 14.00
'~: ,
16.0~
alyzed in replicates: three different human blood samples have
12~RS- S~G-MedLeg ~ 2 2: MRkl of 11 Cll14/l~dI E ~ -
been tested. To be acceptable, relative recoveries should be be-
7 1 7 -I . 31114~ 9{ ~i1141Q441 330 >
tween 80 and 120%. This acceptance criterion was recom-
800 1000 120Q 14.0~ 1600
12-O~B~IRS- SANG*I/4~L~ ~ 2
"~ 4 7 I e" ''A
2: MRM o f l l C I ~ I ~ I I E S +
3 1 9 > 2'S8 2
mended by the FDA Guidance (19).
.~j:_~'
r
~" A.'.~'~= "~ ~t,,~ .... , .... ,. ., .... ,
Linearity. A weighted least-squares quadratic regression
800 1000 IZ0,0 1400 16,00
12~-~RS- S/I~?~-MedlJ~ ~0~ 2: MRM 01'I I O h I V M l l E ~ * model for V and DV and a weighted least-squares linear re-
1 B I 9Jl
gression model for the others compounds were used to calcu-
12~6~1RS- S/~GJ~dlJ~ ~2
8.0# 10.(]0 1ZOO 14,~
2: MR u Ofll (~LMItl/I~IIE~
16.~
late the equation relating the peak-area ratio (drug vs. internal
77 2 . 8 ~ 311 9 1@9
standard) and the concentration of each compound in the cal-
124~0~-IR5-$ANG-MedLeQ-02 2: MRM 0 f l l C h a n ~ l s E S *
ibrators. The inverse of the concentration (l/x) was used as the
I 1 7.21
7
" 0,43 9 9.
10.74
10~
~7 > 10~
192 weighting factor. Three standard curves were analyzed.
800 1000 12~ 1400 1600
I2-Og~,EE~Iq,C~ ~A~4C=t;edt erl -~2 S m 0,I~ 1 ~ 2 M~M ~f ~ ' Chann~is~'S,
422 -201 2
1=00~ 7A7 IS lfe.~
s ~a~l~v~l;i;an~
v,-r .... ,- ,- ,- . -., .... ,. , ' , .... ,
800 1000 1200 1400 1600
Time (min)

Figure 4. RepresentativeMRMchromatogramsof a zero sample(blank (JPe i~e unEsI,CFQ n 6~.(II~

C~,V~ei~hOllgtry,A)IS!r~rs 1on~

human blood with the internal standard). The transition of quantification

is shown for each IRS.

std 0
12 03LI6 IRS S/4~G ~,!e'ILeg 02 S'r{,~tn, I~2; 2 MRhl ofll ChannesES+ I 7q ..--"

200 4~ 699 ~~ 10.00 :i] -'"

1 2 . 0 ~ 1 R S - SA~IC.~MedLeg 4}2 1 MRM of 10 Channels ES*
3 II 94. -,4<;: . . . . .
0 2 1~ 2 2 72 9 4 ~ 5 gl~
Concentra|km
.... i .... , .... r .... , .... i , '~ .... I .... i . . . . i . . . .
2 00 400 6 00 8 00 1000
12~R~- SAi~lG~MedLeg .02 1 : MRI~ of 10 Ch~'u'i~Is E S+
0fC,ePm111~t=on ~"2:0 J!(,)96
C ~effcier,t
,;~lid,at10r t,.~'~ q ~24)Ie-065"x%~ - 00H055 % + .0 {~3i6~-'
075 1 8 7 Z ~ ~ ~,~374 R~$~C%~.~e Ir lerr,~1~P. : r~,f .j :, C,eJ ' ,:I$ (::r,: ,IS Area:
: r,e t'T~ rd : I e ' :rl~;Ir EldJ?,e'Aei]l'trg Io ,~15~'ar1! !l(,l(-
! ......... 2~ . . . . . . 41~ . . . . . . 61~ . . . . . . . oi~ . . . . . . lYOO "
.~ 12-08.06JRS- SANG.hcledLeg -02 1 : MR M of 10 Chanl~els E S+ jJJ
-~ 264 1 > 2 4 6 3
i 0 ( ~ 0,i0 ~ 4OO 437 166
56 I J" *

200 400 6~0 800 t000

12-O~-IR,,~- ~.ANG-MC-'dLeg-02 I MRM of 10 Charlr~e~;I:S*
:0e .--"

:~: .i .:~a~ ~. ~/"..................

200 400 600 800 1000
5(' /." "
12~-~ IRS S/~IG ~edLeg 02 I MRM of 10 Ch~'lindSES*
1 :i~ 1. '~ 5 252 >
1962
92
25 ~ "j

~ . , ~ L _ ~ , .................... O~ ........ F. . . . . . i . . . . . . . . . . i . . . . . . . . . . i ,H ....... ~ , v,,~ , ~ ....... i,,, i ....

200 400 600 800 1000 O 200 I06 600 BOO IC!06 ii00 14DO 1660 1800
Time (mm) Concentration

Figure 5. Representative MRM chromatograms of a zero sample (blank Figure6. Calibrationcurvesof venlafaxineevaluatedusinga weighted
human blood with the internal standard).The transitionof quantification least-squareslinear regressionmodel and a weighted least-squares
is shownforeachantidepressant. quadraticregressionmodel,

338
Journal of Analytical Toxicology, Vol. 31, )uly/August 2007

To determine the adequate fit linear model, the difference be- variation and bias lower than 20%.
tween the observed values and the fitted y-values expressed as Extraction recovery. Extraction recovery from human whole
coefficient of variation (CV) was examined for each standard blood was determined by comparing the response of extracts
concentration. The CV should be less than 15% for each plot of with the drug added before extraction against extracts with the
all calibration curves (19). drug added after extraction. Two quality control samples, con-
Precision. The precision of the developed method was de- taining 80 and 400 ng/mL of each compound (320 and 1600

Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
termined by analysis of three quality control samples con- ng/mL of V and DV),were analyzed four times.
taining 10, 80, and 400 ng/mL (40, 320, and 1600 ng/mL of V Specificity. The lack of interference from endogenous blood
and DV). Intraday variation of the assay was assessed by in- compounds was examined by analyzing different blank human
jecting four quality controls at each concentration on the same blood extracts with or without internal standard. Three dif-
day. Interday variation was assessed by injecting a further four ferent blank human blood samples from hospitalised patients
samples of each concentration on three subsequent days. Sub- were tested to check for the lack of interferences with each
sequently, the mean of each set of the concentrations, the drug and the internal standard.
standard deviation (SD), and coefficient of variation (CV)were The lack of interference from each drug and internal stan-
determined. The precision expressed as CVshould be less than dard to the MRM channels of the other drugs was checked by
15% for each QC sample (19). injecting pure solutions of each drug separately and searching
Bias. Biaswas measured as the percentage of difference from for the total absence of signal in the MRM channels of the
theoretical concentration according to the equation: other drugs.

Bias (%) = [(concentrationmeasured- concentrationtheoretica~)/

Citalopram Citiiloprim
concentrationtheoreticam]x 100% F~:MRM of 11 chemnels,ES+ F2:MRM ol 11 channels,ES+
325 > 109 325 > 262.2
The bias should be less than 15% for all QC samples (19). 2.858e+006 5.646e+005

Limit of quantification. The limit of quantification (LOQ) 9O

was determined as the lowest concentration of a given drug

giving a response that could be quantified with both interassay %
i!! .~
%-

Compoulxf r~e Dssv,e-Yenlafad~e

Cotillon coencJerr r = 0.000055,f2 = 0.980208 I
Caibrdon cu~: 000350872"x * 0.046(~5 i i~) ",
I I!
Respoese~ Inttena!~ ( Ref 6),/veJ '( I~ ConcJl~ .~ea ', -10 .,,... ,., ,,, . -10 . , ., . . . , ,
Cue,~tCe.bnei-, O~rn: F.~u~,~tt~ Ik ~s tmos None
7.50 10 .IX} 7.50 10.00
TOO Time (ndn) Time (rain)
~.0:1
DemM~Citalopram DesMe-Citalopram
5.0~ F2:MRM of t 1 channels,ES+ F2:MRM of 11 channels, ES+

~ 4.00

3.0~ g0
311 > 109
4,505e+O04
311 > 262.2
1.437e+~04

20~

1.0:l %
0.0D
O
l' r
i ........

;l~

Comgoundn~e: DesMe-Venlafa~ne
j lr l.i,..,=,,.r~

400
=l,

EGO
~eer.ie,.rlr

800
Time (min)
1000
r ~r e . e ~ l r

1~0
l,l,l,rl,l

14~0
e~r,, ii,

1~0
l,ir .,r~,

11~0
, i ....

-10
I I1~
III
.J I
I I
. , . . . . ,, . . , , . . -10 "'1,
CoeMctent OfDetefr~na~on:I~2 = 0.9~g234 7.50 10.00 7.50 1084
Calibra~0n r~ws -7.34183@007'x~2 + 0.00465345"x *-0,0111034
Resp onoe ~e: Ir~smalb~d ( Ref 6 ),/Yea * ( 18 Conc f IS Area ) Time (min) Time (min)
Curvetype:.~ndOrder.Ongin: Exclocle,Weighbng: ~/~s bins; None
DIDesMe-Citalopram DiDesMe-Citalopram
eooI .....---J"~ F2:MRM of 11 chanr'41s,E$+ F2:MRM of 11 c~annels, E$+
297 > 109 297 > 262.2.
Goo] / / / - / 6.154e+002 3.211 e+0O2
90 [OI 90"

!;i. ~, 10.50
.~ J ~j,'~__.._/-~
t~ 200 400 GOO 000 1000 1200 1400 1GO0 1800 -10 '"l ........ [' ' -t0
Time (rain) 7.50 10.0(] 7.50 10.50
Time (ndn) Time (min)
Figure 7. Calibration curves of desmethylvenlafaxineevaluated using a
weightedleast-squareslinearregressionmodeland a weightedleast- Figure 8. Chromatogram of a forensic peripheral blood with citalopram,
squares quadratic regression model. desmethylcitalopram, and didesmethylcitalopram.

339

Results
Chromatography
A representative chromatogram of a blood sample spiked
with 50 ng/mL of each drug is presented in Figures 1 and 2.
Under the described chromatographic conditions, all com-
pounds were separated in 17 rain.
tinearity, precision, bias, and recovery
Calibration curves were linear in the range 5-500 ng/mL
(20-2000 ng/mL for V and DV) (r2 > 0.995) with a slope varia-
tion coefficient (n = 3) lower than 12%.
The results obtained for precision and bias are presented in
Table II. The method showed intraday and interday precision,
expressed as CV, of less than 10%. A bias of less than 12% was
achieved. The lower limit of quantification, determined as the
lowest concentration of a given drug giving a response that
could be quantified with both interassay variation and bias
lower than 20%, was 5 ng/mL (20 ng/mL for V and its metabo-
lite).
The extraction recoveries obtained at the different tested
concentrations are presented in Table III. The extraction re-
coveries were between 70 and 90% except for desmethylmir-
tazapine, DV, milnacipram, and didesmethylminalcipram.
Specificity
As shown in Figure 3, no interference was seen from en-
dogenous substances in drug-free human blood at the times of
interest. Figures 4 and 5 show the lack of interference from the
internal standard to the MRM channels of the drugs.
Ion suppression
The ion suppression assay did not show any significant re-
duction in the response of each drug. As shown in Table III, the
relative recoveries were more than 80%. No matrix effect was
observed.
Discussion
The method presented here allows the specific and sensitive
determination of SSRIs, SNaRIs, and mirtazapine (a nora-
drenergic and specific serotoninergic antidepressant) and five
of their active metabolites. To our knowledge, this is the first
method reported in the literature for screen and quantify all
these drugs simultaneously, in whole blood by LC-MS-MS
with a liquid-liquid extraction. LC-MS-MS enables laborato-
ries to perform analyses with the sensitivity and specificity re-
quired for forensic toxicology.
We chose to use a liquid-liquid extraction instead of a pre-
cipitation (15) to reduce matrix effect (20,21). Moreover,
liquid-liquid extraction is easily available in all laboratories,
unlike automated solid-phase extraction or turbulent flow
chromatography, which require instruments available only in
a few laboratories. Although we injected samples from HCI
(0.05N), which is far more aggressive than the solvents typi-
340
Journal of Analytical Toxicology, Vol. 31, July/August 2007
cally used for classical reversed-phase HPLC methods, we did
not see any unusually rapid degradation of column perfor-
mance, so this method is adapted to routine application and
can be used easily in clinical practice.
The present method showed good intra-assay precision and
accuracy for all compounds, with CV values lower than 10%
Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
and bias lower than 12% (Table II). The LOQ (5 ng/mL or 20
ng/mL for V and DV) and calibration range (5 to 500 ng/mL or
20 to 2000 ng/mL for V and DV)of the assay make our method
convenient for analysis of antidepressants in case of unknown
cause of death but also in cases of chemical abuse. Linearity
was evaluated using a weighted least squares linear regres-
sion model except for V and DV, a weighted least squares
quadratic regression model was used. For these molecules,
both linear and quadratic regression models were evaluated.
However, the quadratic model provided a better fit in the range
20-2000 ng/mL (Figures 6 and 7). The correlation coefficient
r2 was always higher than 0.995. Furthermore, the CVfor each
plot of all calibration curves is less than 12%, and the low
values were less underestimated.
In case of intoxication, concentrations higher than 500
ng/mL (2000 ng/mL for V and DV) could be found, and there-
fore, samples must be diluted with blank whole blood. To val-
idate the dilution with blank matrix, a quality control sample
containing 1000 ng/mL (4000 ng/mL for Vand DV)was diluted
1:10 in triplicates. The bias was lower than 15% (data not
shown).
The present assay is used routinely in our laboratory. As an
example, we describe the case of a woman who was found dead
in her residence by her husband. Deliberate self-poisoning was
suspected. First, a general drug screening by GC-MS and LC
with diode-array detection was carried out on the cardiac blood
and revealed the presence of olanzapine, alprazolam, zolpidem,
citalopram, and viloxazine. No alcohol was detected. Therefore,
a quantitative analysis of citalopram and the other drugs was
carried out on the peripheral blood. The concentrations were
6920 ng/mL for citalopram, 170 ng/mL for desmethylcitalo-
pram, and 18 ng/mL for didesmethylcitalopram. For citalo-
pram, the sample was diluted 1:20 (v/v) and analyzed again. The
chromatogram is shown in Figure 8.
Conclusions
The LC-MS-MS method described here allows the sensitive
and specific identification and quantification of SSRIs, SNaRls,
mirtazapine, and five of their active metabolites. This method
is currently used for forensic toxicology investigations.
References
1. C. Serna Arnaiz, L. Galvan Santiago, E. Gasco Eguiluz, P. Santafe
Soler, E. Martin Gracia, and T. Vila Parrot. Evolution in con-
sumption of anti-depressants during the years 2002 to 2004.
Aten. Primaria 38(8): 456-460 (2006).
Journal of Analytical Toxicology,Vol. 31, July/August2007
2. M. Cvjetkovic-Bosnjak, A. Knezevic, and B. Soldatovic-Stajic.
Comparison of the efficacy of traditional antidepressive agents
and the new generation of antidepressives in the treatment of de-
pressive disorders. Med. Pregl. 52(3-5): 108-111 (1999).
3. A. Broocks, K. Junghanns, A. Thiel, C.H. Gleiter, and B. Ban-
delow. New antidepressive drugs in comparison with classical tri-
cyclics. Mechanism of action and clinical evaluation. Fortschr.
Neurol. Psychiatr. 68(1 ): 17-24 (2000).
4. T. Vermeulen. Distribution of paroxetine in three postmortem
cases. J. Anal. Toxicol. 22(6): 541-544 (1998).
5. T.C. Lu, P.L. Chu, C.S. Wu, K.C. Tsai, and W.J. Chen. Neuroleptic
malignant syndrome after the use of venlafaxine in a patient with
generalized anxiety disorder. J. Formos. Med. Assoc. 105{1):
90-93 (2006).
6. S. Wenzel, R. Aderjan, R. Mattern, I. Pedal, and G. Skopp. Tissue
distribution of mirtazapine and desmethylmirtazapine in a case of
mirtazapine poisoning. Forensic Sci. Int. 156(2-3): 229-236
(2O06).
7. D.H. Brendel, J.A. Bodkin, and J.M. Yang. Massive sertraline
overdose. Ann. Emerg. Med. 36(5): 524-526 (2000).
8. H. Druid and P. Holmgren. A case report. Fatal case of acute flu-
oxetine poisoning. Lakartidningen 93(46): 4149-4150 (1996).
9. K. Titier, N. Castaing, E. Scotto-Gomez, F. Pehourcq, N. Moore,
and M. Molimard. High-performance liquid chromatographic
method with diode array detection for identification and quan-
tification of the eight new antidepressants and five of their active
metabolites in plasma after overdose. Ther. Drug Monit. 25(5):
581-587 (2003).
10. G. Toumel, N. Houdret, V. Hedouin, M. Deveau, D. Gosset, and
M. Lhermitte. High-performance liquid chromatographic method
to screen and quantitate seven selective serotonin reuptake in-
hibitors in human serum. J. Chromatogr. 8 761(2): 147-158
(2001).
11. C. Duverneuil, G.L. de la Grandmaison, P. de Mazancourt, and
J.C. Alvarez. A high-performance liquid chromatography method
with photodiode-array UV detection for therapeutic drug moni-
toring of the nontricyclic antidepressant drugs. Ther. Drug Monit.
25(5): 565-573 (2003).
12. H. Juan, Z. Zhiling, and L. Huande. Simultaneous determination
of fluoxetine, citalopram, paroxetine, venlafaxine in plasma by
high performance liquid chromatography-electrospray ioniza-
tion massspectrometry (HPLC-MS/ESI). J. Chromatogr. B 820(1):
33-39 (2005).
13. Z. Zhou, X. Li, K. Li, Z. Xie, Z. Cheng, W. Peng, F. Wang, R. Zhu,
and 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(2): 257-262 (2004).
14. F.L.Sauvage,J.M. Gaulier, G. Lachatre, and P. Marquet. A fully au-
tomated turbulent-flow liquid chromatography-tandem mass
Downloaded from https://academic.oup.com/jat/article-abstract/31/6/334/682846 by Benemérita Universidad Autónoma de Puebla user on 17 December 2018
spectrometry technique for monitoring antidepressants in human
serum. Ther. Drug Monit. 28(1): 123-130 (2006).
15. H. Kirchherr and 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(1): 100-113 (2006).
16. S.M. Wille, K.E. Maudens, C.H. Van Peteghem, and W.E. Lambert.
Development of a solid phase extraction for 13 "new" generation
antidepressants and their active metabolites for gas chromato-
graphic-mass spectrometric analysis. J. Chromatogr. A 1098(1-2):
19-29 (2005).
17. C.B. Eap, G. Bouchoux, M. Amey, N. Cochard, L. Savary, and P.
Baumann. Simultaneous determination of human plasma levels of
citalopram, paroxetine, sertraline, and their metabolites by gas
chromatography-mass spectrometry. J. Chromatogr. 5ci. 36(7):
365-371 (1998).
18. K. Titier, N. Castaing, M. Le-D~odic, D. Le-bars, N. Moore, and
M. Molimard. Quantification of tricyclic antidepressants and
MAOIs by high-performance liquid chromatography-tandem
mass spectrometry, in whole blood. J. Anal. Toxicol. 21(4):
200-207 (2007).
19. Food and Drug Administration. Guidance for Industry, Bioana-
lytical Methods Validation. http://www.fda.gov/cder/guidance.
20. J.X. Shen, R.J. Motyka, J.P. Roach, and R.N. Hayes. Minimization
of ion suppression in LC-MS/MS analysis through the application
of strong cation exchange solid-phase extraction (SCX-SPE).
J. Pharm. Biomed. Anal. 37(2): 359-367 (2005).
21. R. King, R. Bonfiglio, C. Fernandez-Metzler, C. Miller-Stein, and
T. Olah. Mechanistic investigation of ionization suppression in
electrospray ionization. J. Am. Soc. Mass Spectrom. 11(11):
942-950 (2000).
Manuscript received February 23, 2007;
revision received March 21,2007.
~41